Crosstalk between dihydroceramides produced by Porphyromonas gingivalis and host lysosomal cathepsin B in the promotion of osteoclastogenesis

Abstract Emerging studies indicate that intracellular eukaryotic ceramide species directly activate cathepsin B (CatB), a lysosomal‐cysteine‐protease, in the cytoplasm of osteoclast precursors (OCPs) leading to elevated RANKL‐mediated osteoclastogenesis and inflammatory osteolysis. However, the possible impact of CatB on osteoclastogenesis elevated by non‐eukaryotic ceramides is largely unknown. It was reported that a novel class of phosphoglycerol dihydroceramide (PGDHC), produced by the key periodontal pathogen Porphyromonas gingivalis upregulated RANKL‐mediated osteoclastogenesis in vitro and in vivo. Therefore, the aim of this study was to evaluate a crosstalk between host CatB and non‐eukaryotic PGDHC on the promotion of osteoclastogenesis. According to a pulldown assay, high affinity between PGDHC and CatB was observed in RANKL‐stimulated RAW264.7 cells in vitro. It was also demonstrated that PGDHC promotes enzymatic activity of recombinant CatB protein ex vivo and in RANKL‐stimulated osteoclast precursors in vitro. Furthermore, no or little effect of PGDHC on the RANKL‐primed osteoclastogenesis was observed in male and female CatB‐knock out mice compared with their wild type counterparts. Altogether, these findings demonstrate that bacterial dihydroceramides produced by P. gingivalis elevate RANKL‐primed osteoclastogenesis via direct activation of intracellular CatB in OCPs.

sphingolipids serve as second messenger molecules and may also contribute to bone tissue homeostasis and inflammatory osteolysis. [5][6][7] Ceramides are ubiquitous structural components of eukaryotic cell membranes and serve as precursors for other bioactive sphingolipid species, including dihydroceramides, sphingosine, sphingosine-1-phosphate (S1P) and ceramide-1-phosphate. 8,9 While ceramides are constantly produced by eukaryotic cells, only a limited number of gut and oral bacterial species represented by the genera Bacteroides, Prevotella and Porphyromonas, synthesize long-carbonchain ceramides (>C16) termed phosphoethanolamine dihydroceramide (PEDHC) and phosphoglycerol dihydroceramide (PGDHC). [10][11][12] Whereas the cell membrane of all Bacteroidetes contain PEDHC, the keystone pathogenic bacteria for periodontitis, Porphyromonas gingivalis, and Tannerella forsythia, additionally contain PGDHC sphingolipid. 11 It has been reported that the gut bacteria-derived PEDHC is an endogenous source of sphingolipids that impacts lipid homeostasis in the host colon and promotes the activation of invariant natural killer T (iNKT) cells in mice. 13,14 However, PEDHC and most of the host-derived long-chain ceramides cannot penetrate the mammalian cell membrane. [15][16][17] We previously discovered that PGDHC penetrates the cell membrane of osteoclast precursors (OCPs) and leads to the significant upregulation of RANKL-induced osteoclastogenesis by acting on the cytoplasmic non-muscle myosin IIA heavy chain (Myh9). 18 Furthermore, a previously published report provided evidence that temporal degradation of Myh9 in OCPs at early stages of RANKL-primed osteoclastogenesis is exclusively regulated by cathepsin B (CatB), but not cathepsin L, leading to their fusion. 19 CatB is a lysosomal cysteine protease that is ubiquitously expressed in mammalian cells and most active under acidic endosomal conditions. 20 Published evidence clearly demonstrated that CatB is released into the cytosol following lysosomal membrane permeabilization (LMP) which in turn mediates a variety of homeostatic and pathogenic processes, including Alzheimer's disease, extracellular matrix degradation, cancer metastasis to bone and osteoporosis. 11,21,22 Furthermore, CatB plays a critical role in the fusion of osteoclast precursors (OCPs) induced by RANKL. 19 While it has been demonstrated that intracellular host ceramides increase LPM and directly activate CatB in the cytoplasm of myeloid cells, 23 there is no clear evidence for the association between CatB activity and PGDHC in osteoclastogenesis and osteolysis. However, we recently demonstrated that PGDHC significantly elevates the expression of CatB in neurons, in vitro. 24 Thus, these observations raise an intriguing question of whether PGDHC plays a role in CatB release into the cytosol leading to dramatically elevated RANKLinduced osteoclastogenesis.
Our results indicated that PGDHC sphingolipid isolated from the key periodontal pathogen P. gingivalis promotes lysosomal leakage and activation of catB in the cytoplasm of RANKL-primed OCPs.
Furthermore, we also demonstrated that PGDHC promotes osteoclast differentiation and activation in a CatB-dependent manner using male and female CatB-knock out (CatB −/− ) mice and their wild type counterparts (CatB +/+ ).

| Cells
Mouse monocyte/macrophage cell line RAW 267.4 (ATCC TIB-71) cells were obtained from ATCC (Rockville, MD, USA). Bone marrow derived macrophages (BMDMs) were isolated from the femurs and tibias of male and female CatB +/+ and CatB −/− mice. The mice were euthanized by carbon dioxide inhalation followed by neck dislocation, after which the femurs and tibias were dissected, and epiphyses removed. The bone marrow was flushed out of the remaining diaphyses with sterile PBS and viable mononuclear cells were recovered using Histopaque ® -1083 (Sigma-Aldrich, Cat# 10831). Then, isolated mononuclear cells were constantely incubated with recombinant M-CSF (30 ng/ml, BioLegend Cat# 576406).

| Assessment of lysosomal leakage
Lysosomes in RANKL-primed primary BMDMs treated with PGDHC were evaluated using Acridine Orange (Immunochemistry Technologies, Cat# 6130) at a 1:1000 dilution and Hoechst 33342 (Immunochemistry Technologies, Cat# 639) nuclear counterstain at 0.5% v/v concentration, according to the manufacturer's recommendation. The images were acquired using a Zeiss LSM 800 confocal microscope and analyzed using the Zen (Black edition), Zen 3.2 (Blue edition) and Image J software.

| Assessment of cathepsin B activity in vitro
The intracellular activation of CatB in RANKL-primed primary BMDMs treated with PGDHC were evaluated using the Magic   Bis-Tris gel (Invitrogen, Cat# NW04120BOX), followed by Western blot analysis as previously described, using the Anti-Cathepsin B antibody

| RNA extraction and quantitative real time polymerase chain reaction
The total RNA from RANKL-stimulated RAW264.7 cells, as well as The injections were repeated every other day for 10 days.
The mouse calvaria from the calvaria injection experiments were carefully dissected and homogenized in Trizol (Invitrogen) followed by RNA purification using the RNeasy Micro Kit (Quiagen) and qRT-PCR was performed as described above.

| Tissue TRAP staining
Calvaria bone samples were fixed overnight in 10% neutral buffered formalin, were decalcified in 0.5 M EDTA (Millipore) pH 9.0 for 4 weeks at 4°C and, subsequently, dehydrated in graded alcohol and embedded in paraffin. Coronal sections at a thickness of 6μm were prepared for histological analysis and TRAP stained as described by Kanzaki et al. 18 Deparaffinized slides were treated with 0.2 M Tris buffer at pH 9.0 for one hour at 37℃, followed by incubation in 0.2 M acetate +50mM 1 (+) tartaric acid buffer. Finally, the slides were incubated in TRAP solution (0.2 M acetate, 0.5 mg/ml naphtol and 1.1 mg/ml Fast Red) at 37℃ for 2 h, counterstained with haematoxylin for five seconds and mounted.

| In vivo cathepsins activity assay
At Day 10 after mouse calvaria injections were initiated, 100 µl of a CatB™ 680 FAST or CatK™ 680 FAST fluorescent probes (Perkin Elemer, Cat# NEV11112 and Cat# NEV11000) was administrated intravenously, to each mouse. After 6 hours, the luminescence intensity was measured using the In-Vivo Xtreme (Bruker) animal imaging system. A circular region of interest (ROI) was defined as the area which exhibited more than 50% of maximum luminescence in the inflammatory site, for each mouse. The total flux measured in photons per second in the ROI was quantified using the In-Vivo Xtreme Software (Bruker) according to the manufacturer's instructions.

| Statistical analysis
Differences in quantitative data were determined by One-Way ANOVA followed by Tukey's post-hoc test, whereas kinetic CatB activity as assessed using repeated measures ANOVA followed by Tukey's pot hoc test, using Prism 9.0.0 (Graphpad). The p value below 0.05 was considered significant. The data in the graphs are expressed as the group mean plus standard deviation.  Figure 1C). In addition, the colocalization of LGALS3 and LAMP-1 was significantly diminished in RANKL-primed BMDMs exposed to PGDHC ( Figure 1D). Altogether these data indicate that P. gingivalis-derived PGDHC promotes LMP and destabilizes the lysosomal membrane which in turn may promote relocation of cathepsins, including CatB, to the OCP's cytosol.

| Porphyromonas gingivalis-derived PGDHC increases the enzymatic activity of CatB in the cytosol of mouse OCPs and mouse calvaria
Because our data indicate that PGDHC promotes LMP and leakage of lysosomal enzymes in OCPs ( Figure 1A), we sought to elucidate  Figure S4). Altogether, these data indicated that P. gingivalis-derived PGDHC can directly interact with cytoplasmic CatB increasing its enzymatic activity in osteoclasts in vitro and in vivo.

| Loss of CatB abrogated osteoclastogenesis promoted by Porphyromonas gingivalisderived PGDHC
To establish the role of cytosolic CatB in osteoclastogenesis elevated by P. gingivalis-derived ceramides, we investigated the possible effect(s) of CA074-ME, a cell permeable CatB inhibitor, on RANKL-primed and PGDHC-treated RAW264.7 cells ( Figure S5A).

| DISCUSS ION
Our understanding of the molecular mechanisms underlying pathogenic bone resorption induced by dysbiosis has been advanced in the last decade by the finding that numerous bacterial lipid-virulence factors, including LPS and ceramides, are prominently engaged in osteoclastogenesis and inflammation. Various lines of evidence indicate that dysbiosis at the periodontal site is known to result from P. gingivalis interference with the host immune response leading to alveolar bone loss. While P. gingvalis-derived LPS is detected in negligible amounts in human periodontitis lesions several novel sphingolipids isolated from P. gingivalis, including phosphoethanolamine dihydroceramide (PEDHC) and phosphoglycerol dihydroceramide (PGDHC), were identified in substantially greater amounts in inflamed human periodontal tissues when compared with LPS. 12,18 Although the effect of PGDHC in cells of the oral mucosa has not been widely studied, it can induce the secretion of inflammatory factors by primary gingival fibroblasts. 12 In the present study, we determined for the first time that phosphoglycerol dihydroceramide Cathepsin B is a lysosomal protease released into the cytosol following LMP, 34 a mechanism of cellular homeostasis involving lysosomal and recruited proteins, including galectin 3 (LGALS3) and lysosome associated membrane protein 1 (LAMP1). 35 LGALS3 is a member of the lectin family involved in lysosomal repair 36 that is considered a marker of lysosomal and endosomal damage. 37 LGALS3 has high affinity for cells expressing LAMP-1, 38 which is known to stabilize the lysosomal membrane after LMP. 30,31 The identification of both proteins during RANKL-induced osteoclastogenesis and their dysregulation by PGDHC confirms the lysosomal leakage and disruption triggered by PGDHC in OCPs (Figure 1). In addition, LGALS3 is highly expressed in areas of bone resorption but its expression declines during osteoclastogenic differentiation, 39,40 as LGALS3 can disrupt early osteoclastogenic gene expression and hinder osteoclastogenesis. 39 Similarly, LAMP1 overexpression is associated with inhibition of osteoclastogenic activity 41 and decreased traffic of lysosomal enzymes, including CatB and CatD, into the cytosol. 30 Our data show that PGDHC inhibits the expression and co-localization of Lgals3 and Lamp1, which may favor CatB leakage into the cytosol, as well as osteoclastogenic gene expression and differentiation.
Ceramides are bioactive sphingolipids involved in cell apoptosis, senescence, and autophagy 8 that effectively activate CatB. 23 Long chain ceramides and purified proteases produced by P gingivalis, We hereby demonstrate the direct activation of CatB by PGDHC ( Figure 2). CatB is temporally upregulated in preosteoclasts and is required for the degradation of non-muscle myosin IIA (Myh9), an actin binding protein that mediates OCP fusion, adhesion and migration. 19,43 Furthermore, reduced osteoclastogenesis as a result of CatB inhibition in vitro and in vivo have previously been reported. 19,44 Our results confirm these reports, as decreased osteoclast differentiation was observed in CatB −/− mice compared with wild type mice. Previous reports have also determined that P. gingivalisderived PGDHC can increase osteoclast differentiation and the expression of osteoclastogenic markers in vitro and in vivo. 12,18 However, the mechanisms by which PGDHC enhances RANKLinduced osteoclastogenesis were unclear. The data collected in this study confirm the increased osteoclastogenesis due to PGDHC stimulation but, more importantly, we demonstrate that the osteoclastogenic effect of PGDHC is not observed in the absence of CatB (Figures 3 and 4). The deletion of CatB in CatB −/− mice resulted in complete abrogation of PGDHC induced osteoclastogenesis, which demonstrates that the osteoclastogenic effect of PGDHC is dependent on CatB and its enzymatic activation.

F I G U R E 3 Effects of PGDHC on in vitro
In addition to our observations on the effect of the PGDHC/CatB axis on osteoclastogenesis, we have observed a greater osteoclastogenic effect in female rather than male wild type mice throughout our in vitro and in vivo studies. In contrast, PGDHC's osteoclastogenic effect was equally abrogated in both male and female CatB −/− mice (Figures 3 and 4). The intrinsic differences between male and female immune cells, which include monocytes that can differentiate into osteoclasts, have been reported in chicken and mice. 45,46 In fact, these sex-related dysmorphisms may account for increased immune responses and immunogenic gene expression in females. 45,46 This study has potential limitations including additional physiological factors that may have affected the expression of cathepsins in vivo after calvaria injections, as well as the existence of compensatory processes as a result of the absence of CatB in our CatB −/− mice, which may be sex dependent. The mechanisms by which PGDHC affected lysosomal homeostasis and by which sex dysmorphisms affected osteoclast differentiation where not extensively examined nor discussed since they fell outside of the scope of this study.
Altogether, these results are in agreement with previous reports of reduced osteoclastogenesis as a result of CatB inhibition in vitro and in vivo. 19,44 Furthermore, our results suggest that PGDHC isolated from P. gingivalis can enhance osteoclastogenesis in vivo via CatB-dependent signaling. These findings indicate that the novel class of bioactive sphingolipids produced by the key periodontal pathogen P. gingvalis has the potential to elevate osteoclastogenesis via interaction with lysosomal CatB in OCPs. We posit that our studies provide a strong rationale that non-eukaryotic ceramides remain to be further explored for their roles in inflammatory osteolysis.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.