Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Dec 7;41(2):85–93. doi: 10.1111/omi.70015

Porphyromonas gingivalis–Derived Virulence Lipids Accelerate Osteoclastogenesis Independently of High Mobility Group Box Protein‐1 Canonical Signaling

Chiaki Yamada 1,2,3, Gang Peng 2,4, James A Johnson Jr 5, Amilia Nusbaum 1,2, Natasha Sanz 1,2, Hawra AlQallaf 6, Frank Nichols 7, Alexandru Movila 1,2,3,✉
PMCID: PMC12964516  PMID: 41355213

ABSTRACT

Periodontal bacterial pathogen‐associated molecular patterns (PAMPs) and damage‐associated molecular patterns (DAMPs) accelerate inflammatory osteoclastogenesis, resulting in alveolar bone loss. The core PAMP and DAMP prototype molecules are periodontal bacterium Porphyromonas gingivalis–derived virulence lipids, for example, phosphoglycerol dihydroceramide (PGDHC) and lipopolysaccharide (LPS Pg), and the host non‐histone alarmin high mobility group box protein‐1 (HMGB1), respectively. Although it was reported that extracellularly released HMGB1 is critical for the promotion of sepsis inflammation in response to non‐periodontal bacterial LPS, our understanding of the crosstalk between HMGB1 and P. gingivalis–derived virulence lipids remains limited. Therefore, we used Hmgb1fl/fl LysM‐Cre+ mice with ablated HMGB1 mRNA and littermate Hmgb1fl/fl LysM‐Cre− controls. We observed limited Hmgb1fl/fl LysM‐Cre+ osteoclastogenesis compared to Hmgb1fl/fl in response to RANKL in vitro. Furthermore, recombinant HMGB1 protein restored osteoclast formation in Hmgb1fl/fl LysM‐Cre+ cells, indicating the pivotal role of extracellular HMGB1 in osteoclastogenesis in vitro. Using bulk RNA‐sequencing, we identified the diminished osteoclastogenesis in Hmgb1fl/fl LysM‐Cre+ cells are linked to accelerated expression of canonical osteoclast‐suppressing interferon genes. We surprisingly detected that PGDHC and LPS Pg accelerate osteoclastogenesis in Hmgb1fl/fl LysM‐Cre+ cells in vitro. Using bulk RNA‐sequencing and real‐time PCR assays, we confirmed that PGDHC diminishes the expression patterns of different interferon‐inducible guanylate‐binding proteins (GBP 3, 4, 5, 9). At the same time, LPS Pg accelerates the expression of osteoclast‐promoting matrix metalloproteases (MMP 8 and 12) mRNAs. The results suggest that the RANKL‐primed osteoclastogenesis accelerated by P. gingivalis–derived virulence lipids is mediated by different MMP or GBP signaling pathways independently from canonical HMGB1 signaling.

Keywords: guanylate‐binding proteins, high mobility group box protein 1, matrix metalloproteases, osteoclastogenesis, Porphyromonas gingivalis, virulent lipids

1. Introduction

Osteoclasts are specialized multinucleated cells derived from the monocyte/macrophage lineage of myeloid cells upon exposure to macrophage colony‐stimulating factor (M‐CSF) and receptor activator of NF‐κB ligand (RANKL) (Boyce et al. 2009; Kodama and Kaito 2020). They are responsible for physiological bone remodeling and pathological alveolar bone resorption in periodontitis lesions, which is one of the major diseases affecting the oral cavity (AlQranei and Chellaiah 2020).

In the context of periodontitis, the pathogen‐associated molecular patterns (PAMPs) and damage‐associated molecular patterns (DAMPs) jointly promote inflammatory osteoclastogenesis, leading to accelerated alveolar bone loss (Li et al. 2021). Our group and others further demonstrated that PAMPs represented by microbial‐derived virulent factors, especially by bioactive lipids (lipopolysaccharide [LPS], dihydroceramide sphingolipids, and others), play an essential role in the promotion of RANKL‐primed osteoclastogenesis (Luo et al. 2025; Kanzaki et al. 2017; Duarte et al. 2022). In response to bacterial inflammation, PAMPs accelerate the release of DAMPs from osteoclast precursors, thereby accelerating RANKL‐primed osteoclastogenesis (Pisetsky 2011; Carroll et al. 2025).

Among the known DAMPs, a non‐histone high mobility group box protein 1 (HMGB1) serves as an alarmin prototype protein, which is critical for promoting accelerated osteoclastogenesis in periodontal lesions (Pisetsky 2011; Yamashiro et al. 2020). Numerous studies have further reported that HMGB1 directly regulates RANKL‐primed osteoclastogenesis (Davis et al. 2019; Zhou et al. 2008; Zhao et al. 2025). In addition, it was emphasized that the inflammatory effect of HMGB1 resulted from the extracellular complex with bacteria‐derived LPS in sepsis and arthritis (Qin et al. 2014; Andersson and Yang 2022). However, limited knowledge is available about the impact of HMGB1 and periodontal bacteria‐derived PAMPs on the promotion of osteoclast differentiation and function. Indeed, to the best of our knowledge, only a single study demonstrated some beneficial effects of anti‐HMGB1 monoclonal antibody in ameliorating experimental mouse periodontitis mediated by a ligature and human periodontal bacteria Porphyromonas gingivalis (Yoshihara‐Hirata et al. 2018), although it lacked clear evidence of accelerated osteoclastogenesis rescue. Of note, we also reported that accelerated P. gingivalis–derived LPS (Pg) has a controversial age‐associated effect on osteoclastogenesis in HMGB1‐expressing osteoclast precursors exposed to RANKL (Akkaoui et al. 2021).

Given that bacterial PAMPs promote the extracellular release of HMGB1 from the host cells, thereby accelerating inflammation through a complex interaction between HMGB1 and bacterial PAMPs, we generated Hmgb1fl/fl LysM‐Cre+ mouse strain with HMGB1‐loss‐of‐function in myeloid cells. Our group has previously reported that myeloid cells serve as a source of osteoclast precursors in the context of inflammatory osteolysis (Movila et al. 2016). Nevertheless, to the best of our knowledge, no study has addressed the axis between HMGB1 and virulent lipids produced by P. gingivalis, for example, phosphoglycerol dihydroceramide (PGDHC) and LPS Pg, in RANKL‐primed osteoclastogenesis.

In the present study, we report that Hmgb1fl/fl LysM‐Cre+ osteoclast precursors with HMGB1 ablation have limited ability to differentiate into osteoclasts compared to control Hmgb1fl/fl LysM‐Cre− cells with normal function of HMGB1. In contrast, we established that PGDHC and LPS Pg restored osteoclast differentiation of HMGB1‐loss‐of‐function myeloid cells in response to RANKL in vitro. The bulk RNA‐sequencing and real‐time PCR assays were further employed to establish transcriptomic changes accelerated by PGDHC and LPS Pg in RANKL‐primed osteoclast precursors with ablation of HMGB1 mRNA.

2. Materials and Methods

2.1. Mice

To determine whether osteoclast precursors lacking HMGB1 have a reduced capacity to differentiate into osteoclasts, we used Hmgb1fl/fl LysM‐Cre+ mice. HMGB1 floxed mice (Yanai et al. 2013) were obtained from Riken (B6.129P2‐Hmgb1〈tm1Ttg〉; BRC No. RBRC06240), and LysM‐Cre (B6.129P2‐Lyz2tm1(cre)Ifo/J; RRID: IMSR_JAX:004781) mice from Jackson Laboratory. To generate Hmgb1fl/fl LysM‐Cre+ mice, we crossed Hmgb1fl/fl with LysM‐Cre+/+ mice. Hmgb1fl/fl LysM‐Cre− littermates served as controls in accordance with previously published protocols (Greve et al. 2023; Ahmed et al. 2024). All mice were genotyped to confirm Cre recombinase status. Mice were housed individually in HEPA‐filtered, ventilated polycarbonate micro‐isolator cages in an AAALAC‐accredited facility (20–24°C; 12 h light/dark cycle). Food and water were provided ad libitum. Mice were acclimated for at least 1 week before experiments. All experiments were conducted in strict accordance with protocols approved by the Indiana University School of Medicine Institutional Animal Care and Use Committee (#24046 and 20092) and followed National Institutes of Health (NIH) guidelines for housing, breeding, and experimental use. Humane care was provided to minimize suffering.

2.2. Isolation of PGDHC and LPS From Porphyromonas gingivalis

PGDHC was isolated from P. gingivalis (ATTC strain #33277) as previously described (Yamada et al. 2020). For biological experiments, PGDHC was sonicated (2 min, 3 W) in phosphate‐buffered saline (PBS) to achieve a stock concentration of 1 mg/mL. PGDHC was further diluted using cell culture medium before being exposed to cells. Ultrapure LPS isolated from P. gingivalis (LPS Pg) was purchased from InvivoGen and prepared according to the manufacturer's recommendation.

2.2.1. RANKL‐Primed Osteoclastogenesis In Vitro

Bone marrow–derived macrophages (BMDMs) were isolated from the femurs and tibias of 6‐ to 8‐week‐old male Hmgb1fl/fl LysM‐Cre+ mice and the littermate Hmgb1fl/fl LysM‐Cre− control mice (Hmgb1fl/fl) as described (Yamada et al. 2021). Briefly, the femurs and tibias were dissected, and epiphyses were removed. The bone marrow was flushed out from the remaining diaphysis with sterile PBS, and viable mononuclear cells were collected using Histopaque‐1083 (Sigma‐Aldrich), following the manufacturer's recommended instructions. Then, isolated BMDMs were plated at a density of 1 × 105 cells per well in the proliferation medium, alpha‐MEM (Corning) supplemented with 10% FBS (Atlanta Biologicals), 1% antibiotic–antimycotic, 1% l‐glutamine, 1% MEM‐NEAA (all from Gibco), in the presence of 20 ng/mL mouse recombinant M‐CSF (BioLegend) for 3 days. Cells were cultured in a humidified incubator (5% CO2 in air) at 37°C. To induce osteoclastogenesis, the proliferation media described above were additionally supplemented with 10 ng/mL of recombinant mouse RANKL (BioLegend) proteins (differentiation medium) in the presence or absence of several concentrations of PGDHC or an ultrapure LPS Pg. Five days later, cells were stained for tartrate‐resistant acid phosphatase (TRAP) using a leukocyte acid phosphatase kit (Sigma). TRAP‐positive (TRAP+) cells with more than three nuclei were considered osteoclasts. TRAP+ multinuclear cells were counted, and the results were expressed as numbers per well.

2.3. Total RNA Extraction

BMDMs were cultured in a 12‐well culture plate in the proliferation medium supplemented with M‐CSF 20 ng/mL for 3 days. Then, BMDMs were incubated with M‐CSF 20 ng/mL and RANKL 10 ng/mL, respectively, for 24 h. BMDMs were cultured in the presence or absence of several concentrations of PGDHC or LPS Pg for 24 h in differentiation medium. Total RNA was extracted with PureLink RNA Mini Kit (Invitrogen), following the manufacturer's instructions. The RNA quality was evaluated by the Center for Medical Genomics and was used for bulk RNA‐sequencing and real‐time PCR.

2.4. Bulk RNA‐Sequencing and Analysis

Bulk RNA‐sequencing was performed by the Center for Medical Genomics at Indiana University Indianapolis, using three samples per condition. Total RNA samples were first evaluated for their quality using the Agilent TapeStation assay. All the samples used for RNA‐sequencing had the RNA integrity number equivalent between 9.7 and 9.9. The RNA‐sequencing reads were aligned to the GRCm39 reference genome using STAR (Dobin et al. 2013), and read counts for each gene were obtained with featureCounts (Liao et al. 2014). The raw read counts for each gene were then imported into R software environment (R Core Team 2023) and processed using DeSeq2 package (Love et al. 2014). Within DeSeq2, the data underwent normalization using the median of ratios method, and differentially expressed genes (DEGs) were identified through independent filtering. Our analysis focused on DEGs across three specific comparisons: (I) Hmgb1fl/fl LysM‐Cre+ versus Hmgb1fl/fl LysM‐Cre− exposed to RANKL alone; (II) impact of a mixture of 10 ng/mL of PGDHC with RANKL versus RANKL alone on Hmgb1fl/fl LysM‐Cre+; (III) impact of a mixture of 10 ng/mL of LPS Pg with RANKL versus RANKL alone on Hmgb1fl/fl LysM‐Cre+. Finally, we employed the R package ClusterProfiler to conduct the gene set enrichment analyses using gene ontology (GO) pathways (Yu et al. 2012). Of note, we applied more flexible criteria to identify a sufficient number of DEGs for pathway analysis due to the smaller effect size in the comparisons for Comparisons II and III relative to Comparison I. Results were validated by real‐time PCR assay.

2.5. Real‐Time PCR

Complementary DNA (cDNA) was synthesized from 1 µg of total RNA by Verso cDNA Synthesis Kit (Thermo Fisher Scientific) following the manufacturer's recommendation. Real‐time PCR was performed using Taqman Fast Gene Expression Master Mix (Applied Biosystems) in the QuantStudio 3 Real‐Time PCR Systems. The following primers were used: Acp5 (Mm00475698_m1), Oc‐stamp (Mm00512445_m1), Dc‐stamp (Mm04209236_m1), Ctsk (Mm00484039_m1), Gbp2 (Mm00494576_g1), Gbp3 (Mm00497606_m1), Gbp4 (Mm00657752_m1), Gbp5 (Mm00463729_m1), Gbp7 (Mm00523797_m1), Gbp9 (Mm07301706_m1), Mmp8 (Mm00439509_m1), and Mmp12 (Mm00500554_m1). Data were analyzed using the 2−ΔΔ Ct method normalized to Gapdh (Mm 99999915_g1).

2.6. Statistical Analysis

Data are displayed as mean ± SD. Statistical significance was evaluated using a Student's t‐test or one‐way ANOVA with post hoc Tukey's test. A p < 0.05 was considered statistically significant. Data were analyzed using PAST 2.1 statistical software.

3. Results

HMGB1 plays a pivotal role in the promotion of RANKL‐primed osteoclastogenesis (Zhou et al. 2008; Cheng et al. 2016). Because LysM myeloid cells serve as a source of osteoclast precursors (Astleford‐Hopper et al. 2025), we first tested whether selective ablation of Hmgb1 in LySM myeloid cells impacts expression patterns of genes associated with osteoclast differentiation and activity in RANKL‐primed bone marrow mononuclear cells obtained from Hmgb1fl/fl LysM‐Cre+ and littermate control Hmgb1fl/fl LysM‐Cre− (Hmgb1fl/fl) mice in vitro. We previously confirmed that the deletion of HMGB1 mRNA expression is only in the LysM peripheral myeloid cells (Greve et al. 2023). According to the bulk RNA‐sequencing data analysis, we discovered that 1810 genes with a false discovery rate less than 0.05 and an absolute log2 fold change larger than 0.585 (1.5‐fold larger or 0.667‐fold smaller, Table S1) in Hmgb1fl/fl LysM‐Cre+ BMDM cells compared to Hmgb1fl/fl exposed to recombinant mouse RANKL protein (Figure 1A,B). Among the detected genes, we further observed significant downregulation of genes associated with osteoclast formation, Nfatc1, Itgb3, Calcr, Oc‐stamp, Dc‐stamp, and function, CtsK, Fosl2, Oscar, Acp5, and Mmp9 (Figure 1B). Gene set enrichment analysis of GO analysis of the DEGs in RANKL‐primed Hmgb1fl/fl versus Hmgb1fl/fl LysM‐Cre+ identified several GO terms with accelerated expression changes in genes involved in the interferon‐beta signaling pathway (Figure 1C), which is known for negative regulation of osteoclast differentiation (Takayanagi et al. 2002; Yamaguchi et al. 2016).

FIGURE 1.

FIGURE 1

Open in a new tab

Pivotal role of high‐mobility group box 1 (HMGB1) in RANKL‐primed osteoclastogenesis in vitro. (A) Volcano plot of the number of differently regulated genes in RANKL‐primed bone marrow–derived macrophages isolated from mice with HMGB1 ablation in myeloid cells, Hmgb1fl/fl LysM‐Cre+ and littermate control, Hmgb1fl/fl. (B) Heatmap showing the normalized gene expression of DEGs in Hmgb1fl/fl LysM‐Cre+ and Hmgb1fl/fl bone marrow–derived macrophages in response to 2 days of exposure to recombinant mouse RANKL protein (left side). Heatmap of osteoclast formation and function genes in RANKL‐primed osteoclastogenesis identified in bulk RNA‐sequencing of Hmgb1fl/fl LysM‐Cre+ and Hmgb1fl/fl bone marrow‐derived macrophages (right side). (C) The top 10 gene ontology (GO) terms of the DEGs between Hmgb1fl/fl LysM‐Cre+ versus Hmgb1fl/fl bone marrow–derived macrophages after 2 days of mouse recombinant RANKL exposure. Gene ratio is defined as the proportion of differentially expressed genes (DEGs) in each GO term. (D) Real‐time PCR of selected osteoclast formation and function genes in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ and Hmgb1fl/fl bone marrow–derived macrophages. (E) Microscopic evaluation of the TRAP staining in Hmgb1fl/fl LysM‐Cre+ and Hmgb1fl/fl cells exposed to RANKL for 5 days. (F) Number of TRAP+ multinucleated osteoclast‐like cells based on the cells’ microscopic evaluation. (G) Effect of recombinant HMGB1 protein on expression patterns of some representative genes associated with osteoclast formation and function markers in Hmgb1fl/fl. LysM‐Cre+ bone marrow–derived macrophages exposed to RANKL for 48 h. (H) Microscopic evaluation and (I) quantification of the number of TRAP+ multinucleated osteoclast‐like cells in the RANKL‐primed Hmgb1fl/fl. LysM‐Cre+ bone marrow–derived macrophages exposed to HMGB1. For all studies, bone marrow–derived mononuclear cells were first proliferated into macrophages using recombinant mouse M‐CSF protein (20 ng/mL) for 3 days, followed by a mixture of RANKL (10 ng/mL) and M‐CSF (20 ng/mL). Gene expression normalized to Gapdh is represented as 2−ΔΔ Ct . TRAP+ cells with more than three nuclei were considered mature osteoclast‐like and were microscopically counted. n = 3 samples/condition. *p < 0.05, **p < 0.01, ***p < 0.001 by a student t‐test or ANOVA with Tukey post hoc.

To confirm our RNA‐sequencing results, we further tested the expression patterns of genes associated with osteoclast differentiation and activity using real‐time PCR and TRAP staining assays. In response to 48 h exposure with RANKL, we detected significant reduction of Acp5/TRAP, Ocstamp, Dcstamp, and CtsK mRNAs in Hmgb1fl/fl LysM‐Cre+ BMDMs compared to Hmgb1fl/fl cells in vitro (Figure 1D). Furthermore, the number of TRAP+ osteoclast‐like cells was significantly diminished in Hmgb1fl/fl LysM‐Cre+ cells compared to Hmgb1fl/fl cells after 5 days exposure to recombinant mouse RANKL protein (Figure 1E,F). These data agree that HMGB1 ablation in LysM myeloid cells inhibited osteoclast differentiation in response to RANKL in vitro.

Next, we aimed to understand if recombinant HMGB1 protein restores osteoclast function in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ BMDMs in vitro. To our knowledge, expression patterns of Acp5/TRAP, Ocstamp, Dcstamp, and CtsK mRNAs, as well as the number of TRAP+ osteoclast‐like cells were significantly elevated in the presence of recombinant HMGB1 protein (Figure 1G–I), indicating the critical effect of extracellular HMGB1 on the promotion of RANKL‐primed osteoclastogenesis in vitro.

Although the extracellular secretion of HMGB1 in response to P. gingivalis is pivotal for the promotion of periodontal bone loss (Yoshihara‐Hirata et al. 2018), there is a lack of knowledge of whether different virulence factors derived from P. gingivalis affect RANKL‐primed osteoclastogenesis independently of HMGB1. Our group reported that both structurally unique PGDHC and LPS Pg accelerate osteoclastogenesis in vitro using HMGB1‐expressing osteoclast precursors in the presence of RANKL (Kanzaki et al. 2017). Therefore, we further exposed RANKL‐primed Hmgb1fl/fl LysM‐Cre+ BMDMs to different concentrations of PGDHC or LPS isolated from P. gingivalis. Surprisingly, PGDHC and LPS Pg both significantly accelerated the expression of Acp5/TRAP, Ocstamp, Dcstamp, and CtsK mRNAs in HMGB1‐ablated osteoclast precursors in the presence of RANKL (Figure 2A,B). In addition, the number of TRAP+ osteoclast‐like cells was significantly accelerated in the presence of PGDHC or LPS Pg (Figure 2C–F). Thus, these results indicated that those PAMPs derived from P. gingivalis, for example, PGDHC and LPS Pg, promote RANKL‐primed osteoclastogenesis even in the HMGB1‐loss‐of‐function osteoclast precursors. We also noted that PGDHC may have a more substantial effect on osteoclastogenesis in osteoclast precursors with ablated HMGB1 in vitro, compared to LPS Pg.

FIGURE 2.

FIGURE 2

Open in a new tab

Porphyromonas gingivalis–derived phosphoglycerol dihydroceramide (PGDHC) and lipopolysaccharide (LPS Pg) virulent PAMP lipids accelerated osteoclastogenesis in RANKL‐primed osteoclast precursors with HMGB1 mRNA ablated expression in vitro. PGDHC (A) and LPS Pg (B) affect the expression patterns of selected genes engaged in osteoclast formation in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ osteoclast precursors independently of HMGB1. Microscopic evaluation of the TRAP staining in Hmgb1fl/fl LysM‐Cre+ and the number of TRAP+ multinucleated osteoclast‐like cells exposed to RANKL alone or in a combination with PGDHC (C and D) or LPS Pg (E and F) for 5 days. Hmgb1fl/fl LysM‐Cre+ osteoclast precursor differentiation followed by real‐time PCR analysis or microscopic evaluation for TRAP+ osteoclast‐like cells was performed as described in Figure 1(D–F). Gene expression normalized to Gapdh is represented as 2−ΔΔ Ct . TRAP+ cells with more than three nuclei were considered mature osteoclast‐like and were microscopically counted. *p < 0.05, **p < 0.01, ***p < 0.001 by ANOVA with Tukey post hoc.

In order to determine the effect of PGDHC and LPS isolated from P. gingivalis on HMGB1‐independent osteoclastogenesis, we performed the bulk RNA‐sequencing of Hmgb1fl/fl LysM‐Cre+ BMDMs after 48 h exposure to RANKL in the presence of PGDHC and LPS Pg in vitro. In response to PGDHC, there were 952 DEGs (Table S2) in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ (Figure 3A,B). The gene set analysis detected that the Log2 values were significantly diminished in the guanylate‐binding protein (GBP) gene family, including Gbp4, Gbp7, and Gbp9 (Figure 3C), which are critical in the negative regulation of osteoclastogenesis in response to interferon (Place et al. 2021). Of note, a mixture of LPS Pg with RANKL affected only 182 DEGs with a p value less than 0.05 and an absolute log2 fold change larger than 0.263 (1.2‐fold larger or 0.833‐fold smaller, Table S3) compared to those cells exposed to RANKL alone (Figure 3D,E). Our RNA‐sequencing data further discovered increased levels of matrix metalloproteinase (MMP) 8 and MMP12, but not GBPs or mRNAs in HMGB1‐ablated osteoclast precursors in response to a mixture of LPS Pg with RANKL compared to RANKL alone (Figure 3F). Surprisingly, no effect of PGDHC on MMPs gene expression was detected in RANKL‐primed osteoclast precursors with HMGB1 ablation (Figure 3C).

FIGURE 3.

FIGURE 3

Open in a new tab

Virulent PAMP lipids isolated from periodontal bacteria Porphyromonas gingivalis impact transcriptome profile in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ osteoclast precursors with ablation of HMGB1. Volcano plot (A) and clustered heatmap (B) of normalized counts of 952 differently expressed genes in Hmgb1fl/fl LysM‐Cre+ osteoclast precursors exposed either with RANKL alone or in the mixture with phosphoglycerol–dihydroceramide (PGDHC; 10 ng/mL) isolated from P. gingivalis. (C) Gene set enrichment analyses detected downregulation of the guanylate nucleotide‐binding protein (GBP) gene family in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ osteoclast precursors exposed to PGDHC. Volcano plot (D) and clustered heatmap (E) of normalized counts of 182 differently expressed genes in Hmgb1fl/fl LysM‐Cre+ osteoclast precursors exposed to either RANKL alone or in a mixture with lipopolysaccharide isolated from P. gingivalis (LPS Pg; 10 ng/mL). (F) Gene set enrichment analyses detected upregulation of matrix metalloproteases (MMPs) in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ osteoclast precursors exposed to LPS Pg. To induce osteoclast differentiation, bone marrow–derived mononuclear cells were proliferated into macrophages with recombinant mouse M‐CSF protein (20 ng/mL) for 3 days, followed by a mixture of RANKL (10 ng/mL) and M‐CSF (20 ng/mL). After 2 days, RNA was isolated and analyzed by bulk RNA‐sequencing assay. n = 3 samples/condition. LPS, lipopolysaccharide; PGDHC, phosphoglycerol dihydroceramide.

Finally, we validated the bulk RNA‐sequencing discoveries using real‐time PCR assay. PGDHC dramatically diminished the expression patterns of Gbp4 and Gbp9 mRNA in RANKL‐primed osteoclast precursors with ablated function of HMGB1 (Figure 4A). In addition, some minor significant effects of PGDHC were observed on the levels of GBP3 and GBP5. Furthermore, neither Mmp8 nor Mmp12 mRNAs were impacted by PGDHC in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ cells (Figure 4A). In contrast, P. gingivalis–derived LPS accelerated expression of Mmp8 and Mmp12 mRNAs, but not GBPs mRNA, in Hmgb1fl/fl LysM‐Cre+ BMDMs exposed to RANKL in vitro (Figure 4B). Altogether, our real‐time PCR data confirmed RNA‐sequencing findings that PGDHC and LPS Pg accelerated osteoclastogenesis via diminished expression of GPBs and promotion of MMPs, respectively, in RANKL‐primed osteoclast precursors with ablated production of HMGB1.

FIGURE 4.

FIGURE 4

Open in a new tab

Real‐time PCR validation of the featured genes identified by bulk RNA‐sequencing. (A) PGDHC (10 ng/mL) inhibits expression patterns of Gbp3, Gbp4, Gbp5, and Gbp9 mRNA, but not MMPs, in Hmgb1fl/fl LysM‐Cre+ osteoclast precursors exposed to recombinant RANKL protein. (B) In contrast, Porphyromonas gingivalis–derived LPS (LPS Pg, 10 ng/mL) accelerated expression patterns of Mmp8 and Mmp12 and had no effect on GBPs in Hmgb1fl/fl LysM‐Cre+ osteoclast precursors exposed to recombinant RANKL protein. Gene expression normalized to Gapdh is represented as 2−ΔΔ Ct . n = 3 samples/condition. *p < 0.05, **p < 0.01, ***p < 0.001 by ANOVA with Tukey post hoc. LPS, lipopolysaccharide; PGDHC, phosphoglycerol dihydroceramide.

4. Discussion

Periodontitis is an inflammatory non‐communicable disease caused by a complex interplay between periodontal pathogenic bacteria and the host immune response, which results in alveolar bone loss due to accelerated osteoclast activity (Könönen et al. 2019; Puzhankara and Janakiram 2022). The inflammatory existence of RANKL‐primed osteoclastogenesis in periodontal lesions is regulated in part by the accelerated prevalence of periodontal bacteria‐derived PAMPs and host‐produced DAMPs, represented by various virulent lipids (LPS, PGDHC) and alarmins (HMGB1), respectively (Kanzaki et al. 2017; Duarte et al. 2022; Kassem et al. 2015). Emerging evidence suggests that a pro‐inflammatory complex between HMGB1 and LPS, isolated from non‐periodontitis bacteria, is essential for promoting inflammation in sepsis (Andersson and Yang 2022). In this study, we demonstrated for the first time that the virulent lipids isolated from periodontal bacteria P. gingivalis accelerate osteoclastogenesis independently of HMGB1.

The host‐produced alarmin HMGB1 is a nuclear protein released from activated macrophages or injured cells (Zhou et al. 2008; Taniguchi et al. 2007). It was further reported that HMGB1 is released from RANKL‐primed osteoclast precursors, playing a pivotal role in osteoclastogenesis (Zhou et al. 2008). Using Hmgb1fl/fl LysM‐Cre+ osteoclast precursors with confirmed HMGB1 ablation in myeloid cells (Ahmed et al. 2024), we confirmed diminished osteoclastogenesis in RANKL‐primed osteoclast precursors compared to control cells with normal HMGB1 function (Figure 1). We further observed that the recombinant HMGB1 protein restored RANKL‐primed osteoclastogenesis in Hmgb1fl/fl LysM‐Cre+ cells (Figure 1). Thus, our observations are in line with the fact that HMGB1 is one of the active osteoclastogenesis‐promoting cytokines (Yang et al. 2008). Using bulk RNA‐sequencing assay (Figure 1), we also detected that Hmgb1fl/fl LysM‐Cre+ cells have accelerated expression rates of interferon signaling genes, known for osteoclastogenesis‐suppressing functions (Coelho et al. 2005).

Given that HMGB1 is prone to form a complex with bacteria‐derived virulent factors, thereby promoting inflammation (Yang et al. 2020), we further expected no or little effect of PGDHC and LPS isolated from P. gingivalis on Hmgb1fl/fl LysM‐Cre+ osteoclast precursors exposed to RANKL. Surprisingly, we observed that PGDHC and LPS Pg both show accelerated expression patterns of osteoclast formation and function genes, as well as accumulation of TRAP+ osteoclast‐like cells in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ cells in vitro (Figure 2). Using bulk RNA‐sequencing and real‐time PCR (Figures 3 and 4), we further confirmed that PGDHC may promote osteoclastogenesis via distinct molecular pathways. More specifically, PGDHC inhibited the expression of several interferon‐inducible GPBs genes in Hmgb1fl/fl LysM‐Cre+ cells exposed to RANKL. It was recently demonstrated that increased levels of GBPs negatively regulate osteoclastogenesis and bone loss (Place et al. 2021). Because RANKL‐primed Hmgb1fl/fl LysM‐Cre+ demonstrated accelerated expression patterns of interferon genes (Figure 2), we may speculate that PGDHC may facilitate osteoclastogenesis via downregulation of the interferon/GBP axis. In contrast, LPS Pg accelerated the expression of MMP8 and MMP12 mRNAs in RANKL‐primed Hmgb1fl/fl LysM‐Cre+ (Figures 3 and 4). The MMPs are secreted by osteoclasts and are involved in the bone tissue degradation (Zhu et al. 2020). Furthermore, MMPs are recognized as a valuable diagnostic tool in assessing the severity of periodontitis (Luchian et al. 2022). We believe these effects are related in part to the distinct chemical structure of P. gingivalis–derived PGDHC and LPS compared to other gram‐negative bacteria (Kanzaki et al. 2017). However, these findings need further clarification.

It was previously proposed that HMGB1 neutralization by anti‐HMGB1 monoclonal antibody may reduce periodontal bone loss in P. gingivalis–mediated experimental mouse periodontitis model (Yoshihara‐Hirata et al. 2018). On the basis of our findings, this approach may prove ineffective, as virulent lipids derived from periodontal bacteria can promote osteoclastogenesis independently of HMGB1. However, in vitro observations obtained in the current study using cells isolated from Hmgb1fl/fl LysM‐Cre+ and corresponding control Hmgb1fl/fl male mice are warranted to compare with samples collected from equal numbers of male and female mice in vitro as well as in vivo in the future, which represents the major limitation of the current study. Given the limited number of significant genes remaining after multiple testing corrections in Comparisons II and III, we applied a more flexible threshold to ensure sufficient gene inclusion for pathway analysis. We acknowledge that further validation in larger cohorts will be essential to confirm these findings. Another major limitation of this study relates to its preclinical stage. Therefore, an investigation using samples from patients from a wide geographical area is required for a better understanding of the crosstalk between periodontal bacteria‐derived PAMPs and host‐produced DAMPs on osteoclastogenesis via GBPs and MMPs signaling pathways.

5. Conclusions

The present work provides new evidence of crosstalk between bacteria‐derived virulent PAMPs and host‐produced DAMPs in RANKL‐primed osteoclastogenesis, using an in vitro experimental model of periodontitis. Although this study confirmed previous observations that extracellularly released DAMPs, for example, alarmin HMGB1, from osteoclast precursors are pivotal for osteoclastogenesis, we demonstrated that P. gingivalis–derived PGDHC and LPS accelerate osteoclastogenesis independently of DAMPs. Analysis of transcriptomic profiles from osteoclast precursors with HMGB1 loss‐of‐function revealed that PGDHC diminishes the expression of osteoclast‐suppressing GBPs. Conversely, in RANKL‐primed osteoclast precursors with HMGB1 ablation, LPS Pg accelerated the mRNA expression of MMPs, which are known to accelerate RANKL‐primed osteoclastogenesis in periodontal lesions.

Funding

This work was supported by Hevolution Foundation under grant number (HF‐GRO‐23‐1199172‐46); NIA under grant number (R01AG064003); U.S. Department of Veterans Affairs under grant number (I21BX006307) and NIH under grant number (R01 ES029835GW, 1RF1AG077826).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: omi70015‐sup‐0001‐tablesS1‐S3.xlsx

OMI-41-85-s001.xlsx (2.4MB, xlsx)

Acknowledgments

This work was supported by the Hevolution Foundation grant HF‐GRO‐23‐1199172‐46, NIA R01AG064003, and VA Merit Pilot Award I21BX006307 (AM), NIH R01 ES029835GW, and NIH 1RF1AG077826 (JAJ). We thank Dr. Michelle Block for providing Hmgb1fl/fl LysM‐Cre+ and the littermate Hmgb1fl/fl LysM‐Cre− control mice for this study.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  1. Ahmed, C. , Greve H. J., Garza‐Lombo C., et al. 2024. “Peripheral HMGB1 Is Linked to O(3) Pathology of Disease‐Associated Astrocytes and Amyloid.” Alzheimer's Dementia 20, no. 5: 3551–3566. 10.1002/alz.13825. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Akkaoui, J. , Yamada C., Duarte C., et al. 2021. “Contribution of Porphyromonas gingivalis Lipopolysaccharide to Experimental Periodontitis in Relation to Aging.” Geroscience 43, no. 1: 367–376. 10.1007/s11357-020-00258-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. AlQranei, M. S. , and Chellaiah M. A.. 2020. “Osteoclastogenesis in Periodontal Diseases: Possible Mediators and Mechanisms.” Journal of Oral Biosciences 62, no. 2: 123–130. 10.1016/j.job.2020.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Andersson, U. , and Yang H.. 2022. “HMGB1 is a Critical Molecule in the Pathogenesis of Gram‐Negative Sepsis.” Journal of Intensive Medicine 2, no. 3: 156–166. 10.1016/j.jointm.2022.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Astleford‐Hopper, K. , Abrahante Llorens J. E., Bradley E. W., and Mansky K. C. 2025. “Lysine Specific Demethylase 1 Conditional Myeloid Cell Knockout Mice Have Decreased Osteoclast Differentiation Due to Increased IFN‐β Gene Expression.” JBMR Plus 9, no. 1: ziae142. 10.1093/jbmrpl/ziae142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Boyce, B. F. , Yao Z., and Xing L.. 2009. “Osteoclasts Have Multiple Roles in Bone in Addition to Bone Resorption.” Critical Reviews in Eukaryotic Gene Expression 19, no. 3: 171–180. 10.1615/critreveukargeneexpr.v19.i3.10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Carroll, K. A. , Sawden M., and Sharma S.. 2025. “DAMPs, PAMPs, NLRs, RIGs, CLRs and TLRs—Understanding the Alphabet Soup in the Context of Bone Biology.” Current Osteoporosis Reports 23, no. 1: 6. 10.1007/s11914-024-00900-3. [DOI] [PubMed] [Google Scholar]
  8. Cheng, T. L. , Lai C. H., Shieh S. J., et al. 2016. “Myeloid Thrombomodulin Lectin‐Like Domain Inhibits Osteoclastogenesis and Inflammatory Bone Loss.” Scientific Reports 6: 28340. 10.1038/srep28340. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Coelho, L. F. , Magno de Freitas Almeida G., Mennechet F. J., Blangy A., and Uzé G.. 2005. “Interferon‐Alpha and ‐Beta Differentially Regulate Osteoclastogenesis: Role of Differential Induction of Chemokine CXCL11 Expression.” PNAS 102, no. 33: 11917–11922. 10.1073/pnas.0502188102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Davis, H. M. , Valdez S., Gomez L., et al. 2019. “High Mobility Group Box 1 Protein Regulates Osteoclastogenesis Through Direct Actions on Osteocytes and Osteoclasts In Vitro.” Journal of Cellular Biochemistry 120, no. 10: 16741–16749. 10.1002/jcb.28932. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dobin, A. , Davis C. A., Schlesinger F., et al. 2013. “STAR: Ultrafast Universal RNA‐Seq Aligner.” Bioinformatics 29, no. 1: 15–21. 10.1093/bioinformatics/bts635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Duarte, C. , Yamada C., Garcia C., et al. 2022. “Crosstalk Between Dihydroceramides Produced by Porphyromonas gingivalis and Host Lysosomal Cathepsin B in the Promotion of Osteoclastogenesis.” Journal of Cellular and Molecular Medicine 26, no. 10: 2841–2851. 10.1111/jcmm.17299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Greve, H. J. , Dunbar A. L., Lombo C. G., et al. 2023. “The Bidirectional Lung Brain‐Axis of Amyloid‐Beta Pathology: Ozone Dysregulates the Peri‐Plaque Microenvironment.” Brain 146, no. 3: 991–1005. 10.1093/brain/awac113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kanzaki, H. , Movila A., Kayal R., et al. 2017. “Phosphoglycerol Dihydroceramide, a Distinctive Ceramide Produced by Porphyromonas gingivalis, Promotes RANKL‐Induced Osteoclastogenesis by Acting on Non‐Muscle Myosin II‐A (Myh9), an Osteoclast Cell Fusion Regulatory Factor.” Biochimica et Biophysica Acta—Molecular and Cell Biology of Lipids 1862, no. 5: 452–462. 10.1016/j.bbalip.2017.01.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kassem, A. , Henning P., Lundberg P., Souza P. P., Lindholm C., and Lerner U. H.. 2015. “ Porphyromonas gingivalis Stimulates Bone Resorption by Enhancing RANKL (Receptor Activator of NF‐κB Ligand) Through Activation of Toll‐Like Receptor 2 in Osteoblasts.” Journal of Biological Chemistry 290, no. 33: 20147–20158. 10.1074/jbc.M115.655787. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kodama, J. , and Kaito T.. 2020. “Osteoclast Multinucleation: Review of Current Literature.” International Journal of Molecular Sciences 21, no. 16: 5685. 10.3390/ijms21165685. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Könönen, E. , Gursoy M., and Gursoy U. K.. 2019. “Periodontitis: A Multifaceted Disease of Tooth‐Supporting Tissues.” Journal of Clinical Medicine 8, no. 8: 1135. 10.3390/jcm8081135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Li, Y. , Ling J., and Jiang Q.. 2021. “Inflammasomes in Alveolar Bone Loss.” Frontiers in Immunology 12: 691013. 10.3389/fimmu.2021.691013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Liao, Y. , Smyth G. K., and Shi W.. 2014. “featureCounts: An Efficient General Purpose Program for Assigning Sequence Reads to Genomic Features.” Bioinformatics 30, no. 7: 923–930. 10.1093/bioinformatics/btt656. [DOI] [PubMed] [Google Scholar]
  20. Love, M. I. , Huber W., and Anders S.. 2014. “Moderated Estimation of Fold Change and Dispersion for RNA‐Seq Data With DESeq2.” Genome Biology 15, no. 12: 550. 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Luchian, I. , Goriuc A., Sandu D., and Covasa M.. 2022. “The Role of Matrix Metalloproteinases (MMP‐8, MMP‐9, MMP‐13) in Periodontal and Peri‐Implant Pathological Processes.” International Journal of Molecular Sciences 23, no. 3: 1806. 10.3390/ijms23031806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Luo, F. , Chen T., Chen S., Bai D., and Li X.. 2025. “Regulation of Osteoclast‐Mediated Bone Resorption by Lipids.” Bone 193: 117423. 10.1016/j.bone.2025.117423. [DOI] [PubMed] [Google Scholar]
  23. Movila, A. , Ishii T., Albassam A., et al. 2016. “Macrophage Migration Inhibitory Factor (MIF) Supports Homing of Osteoclast Precursors to Peripheral Osteolytic Lesions.” Journal of Bone and Mineral Research 31, no. 9: 1688–1700. 10.1002/jbmr.2854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Pisetsky, D. 2011. “Cell Death in the Pathogenesis of Immune‐Mediated Diseases: The Role of HMGB1 and DAMP‐PAMP Complexes.” Swiss Medical Weekly: Official Journal of the Swiss Society of Infectious Diseases, the Swiss Society of Internal Medicine, the Swiss Society of Pneumology 141: w13256. 10.4414/smw.2011.13256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Place, D. E. , Malireddi R. K. S., Kim J., Vogel P., Yamamoto M., and Kanneganti T. D.. 2021. “Osteoclast Fusion and Bone Loss Are Restricted by Interferon Inducible Guanylate Binding Proteins.” Nature Communications 12, no. 1: 496. 10.1038/s41467-020-20807-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Puzhankara, L. , and Janakiram C.. 2022. “Assessment of Common Risk Factors of Non‐Communicable Diseases (NCDs) and Periodontal Disease in Indian Adults: An Analytical Cross‐Sectional Study.” Methods and Protocols 5, no. 2: 22. 10.3390/mps5020022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Qin, Y. , Chen Y., Wang W., et al. 2014. “HMGB1‐LPS Complex Promotes Transformation of Osteoarthritis Synovial Fibroblasts to a Rheumatoid Arthritis Synovial Fibroblast‐Like Phenotype.” Cell Death & Disease 5, no. 2: e1077. 10.1038/cddis.2014.48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. R Core Team . 2023. R: A Language and Environment for Statistical Computing. R Core Team. [Google Scholar]
  29. Takayanagi, H. , Kim S., Matsuo K., et al. 2002. “RANKL Maintains Bone Homeostasis Through c‐Fos‐Dependent Induction of Interferon‐Beta.” Nature 416, no. 6882: 744–749. 10.1038/416744a. [DOI] [PubMed] [Google Scholar]
  30. Taniguchi, N. , Yoshida K., Ito T., et al. 2007. “Stage‐Specific Secretion of HMGB1 in Cartilage Regulates Endochondral Ossification.” Molecular and Cellular Biology 27, no. 16: 5650–5663. 10.1128/mcb.00130-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Yamada, C. , Akkaoui J., Ho A., et al. 2020. “Potential Role of Phosphoglycerol Dihydroceramide Produced by Periodontal Pathogen Porphyromonas gingivalis in the Pathogenesis of Alzheimer's Disease.” Frontiers in Immunology 11: 591571. 10.3389/fimmu.2020.591571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Yamada, C. , Ho A., Akkaoui J., Garcia C., Duarte C., and Movila A.. 2021. “Glycyrrhizin Mitigates Inflammatory Bone Loss and Promotes Expression of Senescence‐Protective Sirtuins in an Aging Mouse Model of Periprosthetic Osteolysis.” Biomedicine & Pharmacotherapy 138: 111503. 10.1016/j.biopha.2021.111503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Yamaguchi, T. , Movila A., Kataoka S., et al. 2016. “Proinflammatory M1 Macrophages Inhibit RANKL‐Induced Osteoclastogenesis.” Infection and Immunity 84, no. 10: 2802–2812. 10.1128/iai.00461-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Yamashiro, K. , Ideguchi H., Aoyagi H., et al. 2020. “High Mobility Group Box 1 Expression in Oral Inflammation and Regeneration.” Frontiers in Immunology 11: 1461. 10.3389/fimmu.2020.01461. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Yanai, H. , Matsuda A., An J., et al. 2013. “Conditional Ablation of HMGB1 in Mice Reveals Its Protective Function Against Endotoxemia and Bacterial Infection.” PNAS 110, no. 51: 20699–20704. 10.1073/pnas.1320808110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Yang, H. , Wang H., and Andersson U.. 2020. “Targeting Inflammation Driven by HMGB1.” Frontiers in Immunology 11: 484. 10.3389/fimmu.2020.00484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Yang, J. , Shah R., Robling A. G., et al. 2008. “HMGB1 is a Bone‐Active Cytokine.” Journal of Cellular Physiology 214, no. 3: 730–739. 10.1002/jcp.21268. [DOI] [PubMed] [Google Scholar]
  38. Yoshihara‐Hirata, C. , Yamashiro K., Yamamoto T., et al. 2018. “Anti‐HMGB1 Neutralizing Antibody Attenuates Periodontal Inflammation and Bone Resorption in a Murine Periodontitis Model.” Infection and Immunity 86, no. 5: e00111–e00118. 10.1128/iai.00111-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Yu, G. , Wang L. G., Han Y., and He Q. Y.. 2012. “clusterProfiler: An R Package for Comparing Biological Themes Among Gene Clusters.” OMICS 16, no. 5: 284–287. 10.1089/omi.2011.0118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Zhao, W. , Qian J., Li J., et al. 2025. “From Death to Birth: How Osteocyte Death Promotes Osteoclast Formation.” Frontiers in Immunology 16: 1551542. 10.3389/fimmu.2025.1551542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Zhou, Z. , Han J. Y., Xi C. X., et al. 2008. “HMGB1 Regulates RANKL‐Induced Osteoclastogenesis in a Manner Dependent on RAGE.” Journal of Bone and Mineral Research 23, no. 7: 1084–1096. 10.1359/jbmr.080234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Zhu, L. , Tang Y., Li X. Y., et al. 2020. “Osteoclast‐Mediated Bone Resorption Is Controlled by a Compensatory Network of Secreted and Membrane‐Tethered Metalloproteinases.” Science Translational Medicine 12, no. 529: eaaw6143. 10.1126/scitranslmed.aaw6143. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting File 1: omi70015‐sup‐0001‐tablesS1‐S3.xlsx

OMI-41-85-s001.xlsx (2.4MB, xlsx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from Molecular Oral Microbiology are provided here courtesy of Wiley

RESOURCES