Bioactive silica nanoparticles target autophagy, NF-κB, and MAPK pathways to inhibit osteoclastogenesis

Abstract

Spherical 50nm silica-based nanoparticles (SiNPs) promote healthy bone homeostasis and maintenance by supporting bone forming osteoblast lineage cells while simultaneously inhibiting the differentiation of bone resorbing osteoclasts. Previous work demonstrated that an intraperitoneal injection of SiNPs in healthy mice -both young and old -increased bone density and quality, suggesting the possibility that SiNPs represent a dual action therapeutic. However, the underlying mechanisms governing the osteoclast response to SiNPs have yet to be fully explored and defined. Therefore, the goals of this study were to investigate the cellular and molecular mechanisms by which SiNPs inhibit osteoclastogenesis. SiNPs strongly inhibited RANKL-induced osteoclast differentiation within the first hours and concomitantly inhibited early transcriptional regulators such as Nfatc1. SiNPs simultaneously stimulated expression of autophagy related genes p62 and LC3β dependent on ERK1/2 signaling pathway. Intriguingly, SiNPs were found to stimulate autophagosome formation while inhibiting the autophagic flux necessary for RANKL-stimulated osteoclast differentiation, resulting in the inhibition of both the canonical and non-canonical NF-κB signaling pathways and stabilizing TRAF3. These results suggest a model in which SiNPs inhibit osteoclastogenesis by inhibiting the autophagic machinery and RANKL-dependent functionality. This mechanism of action defines a novel therapeutic strategy for inhibiting osteoclastogenesis.

INTRODUCTION:

The skeleton is critically important to human health and undergoes a constant state of dynamic remodeling in adulthood. The process of healthy skeletal homeostasis requires a highly regulated equilibrium between bone forming osteoblasts and bone resorbing osteoclasts. However, bone disease, as well as aging, generates an imbalance in which osteoclast mediated resorption outpaces osteoblast formation, and results in decreased bone quality and increased fracture risk. Inhibition of osteoclast differentiation and/or function has become a common therapeutic target for treatment of bone disease such as osteoporosis. Previous studies have suggested that spherical silica nanoparticles (SiNPs), particularly in the 50–100nm range, potently inhibit osteoclastogenesis while stimulating matrix production and mineralization by osteoblasts [1, 2]. Further, cell viability assays and Annexin V staining demonstrated that a dose range of SiNPs (up to 100μg/ml) for up to 10 days failed to demonstrate any direct negative effects on the viability or proliferation rates of murine preosteoclast RAW264.7 cells [1] or murine pre-osteoblast MC3T3-E1 cells [3, 4]. The cellular mechanisms by which these particles so strongly inhibit osteoclast differentiation are not well understood. Understanding the underlying mechanisms governing the inhibition of SiNPs on osteoclastogenesis would provide insight towards the optimization of these particles for therapeutic modulating of skeletal health. We therefore investigated the molecular and cellular mechanisms by which SiNPs alter osteoclastogenesis.

Osteoclasts originate from the hematopoietic lineage, specifically from monocyte/macrophage cells [5], and are stimulated to differentiate by activation of the cell surface receptor RANK (Receptor Activator of Nuclear factor Kappa-B) [6]. Binding of RANK by the soluble RANK ligand (RANKL) initiates a cascade of intracellular events leading to cell fusion of monocyte/macrophages and changes in gene expression, ultimately producing multi-nucleated bone resorbing osteoclasts. Activation of the RANK receptor stimulates NF-κB (nuclear factor kappa-B)[7–9], TRAF (TNF (Tumor Necrosis Factor) Receptor-Associated Factor) and MAPK (Mitogen Activated Protein Kinase) signaling to promote the stimulation of transcriptional regulators including Nfatc1 (Nuclear factor of activated T cells cytoplasmic-1) and members of the AP-1 family including c-fos and c-jun [10–13] and down regulation of differentiation suppressors such as MafB (V-maf avian musculoaponeurotic fibrosarcoma oncogene homolog B) and Irf8 (interferon regulatory factor 8) [14, 15]. Importantly, expression of Nfatc1 is strongly increased by RANKL. While previous work identified the capability of SiNPs to reduce this increase in expression [2], the cellular mechanisms remain unknown.

Autophagy is a catabolic cellular process required for both normal physiological cell functions as well as stress responses to environmental conditions such as nutrient deprivation with the goal of enhancing cell survival [16]. Autophagy involves the degradation of organelles, macromolecules, and pathogens by a coordinated process involving the formation of a double membraned autophagosome which ultimately merges with the lysosome for cargo degradation [17, 18]. It is estimated that at least twenty autophagy related (ATG) genes/proteins are involved in the cascade of events associated with autophagy [19]. Additionally, microtubule associated protein light chain 3 (Map1lc3B, commonly referred to as LC3β) and sequestosome 1 (SQSTM1, commonly referred to as p62) are key regulators of autophagy. During autophagy, LC3β I is lipidated to LC3β II to promote the attachment of cargo to the autophagosome membrane while p62 is responsible for cargo delivery [20, 21]. Recent work suggests the general importance of macroautophagy/autophagy in bone homeostasis [22–27] but also specifically for the process of osteoclastogenesis [28–30] including individual autophagic proteins on osteoclast differentiation and function [31–37]. Indeed, the pharmacologic inhibition of autophagy has been demonstrated to block osteoclast differentiation by preventing TRAF3 degradation [38]. Nanoparticles have been identified as regulators of autophagy [39] with studies suggesting that SiNPs, specifically, can enhance autophagy [40–46] or cause autolysosome dysregulation [47–49] in various cell types including RAW264.7 cells [50]. Taken together, we hypothesize that autophagy is uniquely positioned to regulate the SiNP inhibition of osteoclastogenesis.

As such, we investigated an autophagy-based mechanism through which 50nm SiNPs inhibit osteoclastogenesis. A cell-based osteoclast differentiation model, where pre-osteoclast RAW264.7 cells were treated with RANKL, demonstrated that rapidly internalized SiNPs are most effective at inhibiting osteoclastogenesis when applied within 24h of RANKL treatment. After internalization, SiNPs altered RANKL-induced ERK1/2 and JNK signaling, while inhibiting the NF-κB pathway. Downstream analyses demonstrated that the SiNPs inhibited RANKL-induced osteoclastic gene expression, including the key early transcriptional regulator Nfatc1, while simultaneously stimulating expression of the key autophagy genes p62 and LC3β. In addition to stimulating autophagosome formation, SiNPs were found localized within double membrane structures via TEM, co-localized with autophagosomal proteins by fluorescent microcopy, and physically associated with autophagy proteins LC3β and p62. Collectively, the results suggest a model whereby internalization of SiNPs by pre-osteoclasts inhibits osteoclastogenesis by physically diverting the autophagic machinery and functionality necessary for RANKL-induced differentiation including NF-κB and AP-1 signaling and TRAF3 degradation. These results provide new insights into the mechanisms underlying how both monocyte/macrophage cells respond to nanoparticle exposure and that RANKL-stimulated autophagy can be modified to inhibit osteoclast differentiation.

RESULTS:

SiNPs inhibit osteoclast differentiation marker genes but stimulate autophagy related genes.

Spherical SiNPs inhibit the differentiation of osteoclast precursor cells into osteoclasts [1, 2]; however, the molecular and cellular mechanisms underlying this inhibition remain to be fully elucidated. A common model to study osteoclastogenesis is to induce osteoclast differentiation in murine monocyte/macrophage RAW264.7 cells through treatment with RANKL which results in multinucleated osteoclasts within 3 to 4 days (reviewed in[51]). To dissect the temporal physiology and underlying mechanisms behind SiNP mediated inhibition of osteoclastogenesis, a time course of osteoclast differentiation marker genes was performed following the addition of RANKL (0, 4, 8, and 16h) with or without SiNPs. As expected, RANKL treatment increased expression of the key early transcriptional regulators Nfatc1 and c-fos (Figure 1a), as well as later stage markers tartrate resistant phosphatase (TRAP/Acp5) and osteoclast-associated receptor (Oscar) (Figure 1b); addition of the SiNPs significantly inhibited both early and late osteoclast differentiation marker genes. As recent studies have suggested a potentially important role of autophagy in osteoclastogenesis, we also analyzed genes/proteins involved in autophagosome formation. Interestingly, the addition of the SiNPs with RANKL increased the expression of the autophagy genes, p62 and LC3β, between 4 and 8h relative to RANKL treatment alone whereas Neighbor of BRCA1 (Nbr1) was only transiently increased and Beclin1 was decreased relative to RANKL alone later in the response (Figure 1c). We also assessed the expression of inhibitors of osteoclast differentiation, MafB and Irf8, and observed the expected decease in response to RANKL stimulation alone and SiNPs did not alter the response (Figure 1d). Regulation of osteoclast differentiation and autophagy protein markers by the SiNPs on RANKL induced osteoclastogenesis were confirmed via western blotting (Figure 1e). Given the diverse function of the genes/proteins and differences in temporal, spatial, and posttranslation regulation there is generally good agreement between gene expression and protein abundance. Additionally, a Live/Dead cell assay was used to assess viability and results revealed that SiNPs do not negatively impact cell health (Figure 1f). Together, these data suggest that the SiNPs can specifically target/inhibit specific osteoclast differentiation factors and markers of autophagy without impacting cell viability.

Fig.1:

Silica Nanoparticles inhibit RANKL induced gene expression during osteoclast differentiation. RAW264.7 cells were treated with SiNPs (25μg/mL) for 30 minutes prior to addition of RANKL (15ng/ml) for indicated times and cells harvested for RNA and analysis of gene expression. (a) Osteoclast transcriptional regulators, (b) Osteoclast differentiation marker genes, (c) autophagy genes, and (d) inhibitory transcription factors. RNA was analyzed by qRT-PCR and fold change calculated using ΔΔCT method. Data expressed as Mean ±SEM. Samples run in triplicate. Representative of multiple experiments. (e) Western blot of RAW264.7 cells treated as in (a) and probed with antibodies as indicated. (f) Raw264.7 cells were untreated or treated with RANKL and NPs as in (a) and a live/dead assay performed (n=3). Results are presented as percent change from untreated cells. *P<0.05, ***P<0.0005, relative to corresponding +RANKL control by student’s t test.

SiNPs are internalized rapidly and inhibit early osteoclastogenesis events.

SiNPs have been previously demonstrated to inhibit osteoclast differentiation from primary murine and human monocytes as well as RAW264.7 cells[1, 2, 52] and we sought to determine the stage of osteoclast differentiation targeted and impacted by the SiNPs. RAW264.7 cells were treated with nanoparticles at the start of RANKL stimulation (0h) or at various times after (4, 8, 16, 24, 40, 48h), and assessed for TRAP staining after 3 days. The maximal SiNP-induced inhibition occurs both with simultaneous (0h) treatment or with a 4h delay. By 8h, this inhibition is reduced with only a 40% inhibition with treatment at 24h (Figure 2a). A similar time course was observed when we assessed the inhibition of Nfatc1 gene expression (Figure 2b). The internalization kinetics were investigated using fluorescent SiNPs for visualization by microscopy. SiNPs were detected intracellularly within 15m of treatment, with continued internalization out to 24h (Figure 2c,d). To determine if internalization was necessary for the cellular and molecular response, RAW264.7 cells were pretreated with SiNPs for 2h, with non-internalized SiNPs removed by rinsing. Cells were then treated with RANKL and harvested for gene expression after 24h. Analysis of gene expression by qRT-PCR determined that the 2h pre-treatment with the SiNPs was sufficient to inhibit the RANKL induced expression of Nfatc1, RANK, and TRAP, and increase p62 and LC3β expression (Figure 2e). These results also confirm that the effect of the SiNPs on osteoclastogenesis is intracellular and not the result of a non-specific extracellular event. Taken together, these data suggest that the SiNPs are rapidly internalized and target early RANKL-induced osteoclast differentiation events.

Fig. 2:

SiNP inhibition of osteoclastogenesis is time dependent. (a) RAW264.7 cells were treated with RANKL (15ng/mL) and SiNPs (25μg/mL) added at times indicated. After 3 days, osteoclastogenesis was analyzed by TRAP staining. (b) RAW cells were treated as in (a) 4, 8, 16 h after RANKL addition and cells harvested for RNA at 24h and Nfatc1 quantified by qRT-PCR. (c) Fluorescent SiNPs were added to RAW264.7 cells and fluorescence microscopy was used to track and quantify internalization by RFI (Relative Fluorescence Intensity) (20x) (d), over time. (e) SiNPs were added to RAW264.7 cells for 2h, removed by rinsing with medium, followed by treatment with RANKL (15ng/mL) for 24h and harvested for RNA analysis. RNA was analyzed by qRT-PCR for genes as indicted and fold change calculated using ΔΔ^CT method. Data are expressed as Mean ±SD. Samples run in triplicate. Representative of multiple experiments. #P<0.05, ##P<0.005, ###P<0.0005 compared to untreated, *P<0.05, **P<0.005, ***P<0.0005 relative to RANKL by student’s t test. &&&&P<0.0001 by One way ANOVA.

SiNPs stimulate ERK/1/2 which is required for increased p62 gene expression.

To dissect the underlying cellular mechanism(s) through which SiNPs alter gene expression during osteoclast differentiation, we examined Mitogen Activated Protein Kinase (MAPK) signaling. The three MAPK proteins ERK1/2, p38, and JNK are known to be involved in the early signaling associated with RANKL induced osteoclastogenesis, and activation of these signaling proteins can be identified by increased protein phosphorylation. RAW264.7 cells were treated with or without SiNPs followed by RANKL and harvested for Western blotting. While RANKL treatment alone stimulated phosphorylation of all three kinases, pre-treatment with the SiNPs had differing effects on the kinases. Treatment with SiNPs further stimulated ERK1/2 phosphorylation beyond RANKL alone, with no substantial changes in p38; conversely, the stimulation of JNK (p-JNK) was at least partially inhibited through the early time course (Figure 3a). To determine if activation of ERK1/2 is necessary for the SiNPs effects on differentiating osteoclasts, RAW264.7 cells were pretreated with a pharmacologic upstream inhibitor of ERK1/2, U0126. After 6h of RANKL treatment, the inhibition of ERK1/2 resulted in suppressed RANKL-induced Nfatc1 expression and produced a dose-dependent additive inhibition with SiNPs (Figure 3b). These results suggest that the inhibition of RANKL induced Nfatc1 expression by the SiNPs is independent of ERK1/2 signaling. In contrast, while ERK1/2 was not required for the small increase in p62 stimulated by RANKL alone, the dose dependent induction of p62 by the SiNPs required ERK1/2 activation (Figure 3c). Western blotting confirmed this requirement at the protein level and additionally identified a decrease in SiNP-induced LC3β and RANKL-induced NfatC1 (Figure 3d). Collectively, these results suggest that, while the repressive effects of the SiNPs on osteoclast gene expression are independent of ERK1/2 signaling, the SiNP-stimulated autophagy response requires ERK1/2 activation.

Fig. 3:

ERK1/2 signaling is required for RANKL induced Nfatc1 expression and SiNPs induced p62 expression. (a) RAW264.7 cells were treated with RANKL (15ng/ml) in the absence or presence of SiNPs (25μg/ml) and cells were harvested for Western blotting (whole cell lysate) at indicated times and probed with antibodies as indicated. (b) RAW264.7 cells were pretreated with U0126 at indicated concentrations prior to addition of RANKL (15ng/ml) in the absence or presence of SiNPs (25μg/ml) for 6hr and harvested RNA analyzed by qRT-PCR for Nfatc1 and (c) p62. (d) Cells were treated as above and harvested after 6h for Western blotting (whole cell) and probed with antibodies as indicated. Data are expressed as Mean ±SD. Samples run in triplicate. Representative of multiple experiments. #P<0.05, ##P<0.005, ###P<0.0005 compared to untreated, *P<0.05, **P<0.005, ***P<0.0005 relative to RANKL treated, and &P<0.05, &&P<0.005, &&&P<0.0005 relative to RANKL+SiNP by student’s t test.

SiNPs inhibit RANKL induced NF-kB signaling and AP-1 transcription factors levels.

Two key signaling pathways which orchestrate the transcriptional response of early osteoclast genes to RANKL treatment are the NF-kB and AP-1 protein families. Activation of the c-Jun N-terminal kinase (JNK) signaling is established as a key immediate early event in osteoclastogenesis [53]. To determine if SiNPs inhibit JNK phosphorylation, RAW264.7 cells were treated with RANKL, in the presence or absence of SiNPs, and proteins were harvested within 30m. Western blotting demonstrated that, while RANKL alone induced JNK phosphorylation and downstream induction of the transcriptional regulatory AP-1 proteins c-fos and c-jun, the addition of the SiNPs reduced both JNK activation and downstream AP-1 signaling (Figure 4a,b). Similarly, when we assessed the response of the canonical and non-canonical NF-kB pathways, we found that the SiNPs again blunted the RANKL activated phosphorylation of the regulatory kinases IκBα, IKKα, and IKKβ with little change in total protein as determined by Western blotting of whole cell lysate (Figure 4c). While the addition of SiNPs resulted in decreased phosphorylation of p-IKKα and β, an almost complete inhibition of IκBα was observed (Figure 4c). Phosphorylation of these regulatory kinases leads to the activation of the transcriptional activity of the NF-κB pathway via RelA (p65)/p50 for the canonical pathway and RelB/p52 for the noncanonical. Western blotting of nuclear lysate from RAW264.7 cells treated with RANKL revealed an increase in nuclear p50, p65, p52, and RelB in response to RANKL; addition of the SiNPs again blunted the RANKL induced response for all four factors (Figure 4d). Collectively, these results suggest that the SiNPs inhibit RANKL-induced osteoclastogenesis by repressing the key early transcriptional regulator response downstream of both the JNK and NF-kB pathways.

Fig. 4:

SiNPs inhibit RANKL induced AP-1 and NF-κB signaling. (a) RAW264.7 cells were treated with RANKL (15ng/ml) in the absence or presence of SiNPs (25μg/ml) for indicated times and whole cell lysate harvested for Western blotting of JNK pathway. (b) Cells were treated with RANKL and SiNPs as in (a) for indicated times and harvested for nuclear lysate and analyzed by Western blotting for AP-1 proteins. Representative of multiple experiments. (c) Cell lysate as in (b) and whole cell lysate was probed for NF-kB pathway proteins. (d) Cells were treated as in (a) for indicated times and harvested for nuclear lysate and analyzed by Western blotting for NF-κB proteins. Representative of multiple experiments.

SiNPs stabilize TRAF proteins.

The TNF receptor-associated factor (TRAF) proteins connect the RANKL stimulation of its receptor RANK with the initiation of the MAPK and NF-kB pathways. The intracellular portion of the RANK receptor is bound by these key adapter proteins and, in the context of osteoclast differentiation, both TRAF3 and TRAF6 play integral roles as negative (TRAF3) and positive (TRAF6) signaling regulators [54–56]. RAW264.7 cells were treated with or without SiNPs followed by RANKL for the indicated times. Western blotting revealed TRAF3 was decreased in response to RANKL early in the response (2–6h) which was blocked by the presence of SiNPs (Figure 5a). Little change was detected with TRAF2 or TRAF6 levels at these times. However, at time points out to 24h, there was little change in TRAF6 in the presence of SiNPs (Figure 5b). To confirm the ability of SiNPs to blunt the RANKL driven decrease in TRAF3, immunofluorescence was used. Whereas RANKL resulted in decreased TRAF3 abundance, the presence of SiNPs significantly reduced that decrease (Figure 5c,d). It has been demonstrated that a decrease in TRAF3 is required to permit RANKL induced osteoclastogenesis[38], and the timing of SiNPs stabilization of TRAF3 regulation correlates well with the inhibition of SiNPs on NF-kB pathways signaling.

Fig. 5:

SiNPs stabilize TRAF3 protein levels after RANKL treatment. (a) RAW264.7 cells were treated with RANKL (15ng/ml) in the absence or presence of SiNPs (25μg/ml) for indicated times and whole cell lysate analyzed by Western blotting at indicated times. (b) Cells were treated as in (a) and whole cell lysate analyzed by Western blotting. Representative of multiple experiments. (c) Cells were stained by immunohistochemistry for TRAF3 and nuclei stained with Hoescht at indicated times after addition of RANKL (15ng/ml) and +/− SiNPs (25μg/ml) (scale bar=10μm). (d) TRAF3 RFI were quantified (50 cells/condition) by ImageJ. *P<0.05, **P<0.005 relative to RANKL by student’s t test.

SiNPs are localized within specific cellular structures.

The above results suggested that inhibition of key RANKL induced osteoclastogenic signaling pathways occur within 4h of the addition of SiNPs. To determine if sub-cellular localization might provide insight into mechanism, co-localization studies were performed leveraging rhodamine B doped SiNPs. RAW264.7 cells were treated with fluorescent SiNPs and co-localization was assessed by live cell imaging after 2h for endosome and 4h for mitochondria, endoplasmic reticulum (ER), and lysosome associations. While negligible overlap with mitochondria or ER was observed, we found that the SiNP co-localized with both the endosome and lysosome compartments after internalization (Figure 6a). In the course of these studies, we observed that the size of the lysosomes appeared larger after treatment with SiNPs. Lysosome size was measured from images of control, SiNP treated, or treatment with CQ (Chloroquine; which inhibits autophagosome fusion to the lysosome) and results revealed that both SiNPs and CQ significantly increased the size of lysosomes after 4h of treatment (Figure 6b). These results demonstrate that SiNPs are contained within specific structures and suggest a regulated mechanism for internalization and processing of SiNPs.

Fig. 6:

SiNPs co-localize with the specific organelles. (a) RAW264.7 cells were treated with SiNPs (25μg/ml) and cells live imaged by confocal microcopy. SiNPs (red), Lysosome (Lyso-Tracker), Mitochondria (Mito-Tracker), Endosome (Tfn-FITC) or Endoplasmic Reticulum (ER-Tracker) (green) and Nuclei were stained with Hoescht 33342 (blue). Scale bars=10μm, Magnified Scale bars=5μm. (b) RAW264.7 cells were treated for 4h as indicated and were incubated with LysoTracker (green) and Hoescht 3342 (blue) before imaging as in (a). Scale bars=10μm. Quantification of the lysosomal size per cell from the immunofluorescence images are shown (50 cells/condition). ****P<0.0001 relative to control by student’s t test.

SiNPs localize in autophagosomes/autolysosomes.

The lysosome is the ultimate destination of the process of autophagy and previous studies have identified autophagy in the processing of SiNPs. Therefore, autophagosomes were visualized by immunofluorescence microscopy of LC3β, a traditional marker of autophagy stimulation [18]. We identified strong co-localization of the SiNPs and LC3β in RANKL treated RAW264.7 cells (Figure 7a). RAW264.7 cells were then treated with metal core SiNPs for 24h and analyzed by Transmission Electron Microscopy (TEM). SiNPs were identified in double membrane vesicles/structures characteristic of autophagosomes and autolysosomes (Figure 7b). Together, these results demonstrate a co-localization of SiNPs with the autophagosome which temporally coordinates with the inhibition of RANKL signaling and TRAF3 degradation.

Fig. 7:

SiNPs co-localize and are found in the Autophagosome/Autolysosome like structures. (a) RAW264.7 cells were treated with SiNPs (red) (25μg/ml) and fixed after 4h autophagosomes visualized with LC3β antibody (green). Nuclei were stained with Hoescht 33342 (blue). Scale bar=5μm. Representative of multiple experiments. (b) RAW264.7 cells were treated with metal core nanoparticles for 24h (i,ii) or control (iii) and visualized by Transmission electron Microscopy. Scale bars (200nm (i, iii) or 500nm (ii)) and insets (50nm (i, iii) or 100nm (ii)).

Autophagy is required for osteoclastogenesis and RANKL induced Nfatc1 expression.

To better understand the temporal role of autophagy in osteoclast differentiation, two established inhibitors of autophagy, 3-MA (3-Methyladenine, which inhibits autophagosome formation) and CQ were utilized in a “posttreatment” time course similar to Figure 2a. Treatment of RAW264.7 cells with either autophagy inhibitor resulted in substantial inhibition of RANKL induced osteoclastogenesis, as defined by numbers of TRAP positive cells with greater than three nuclei. The effects were most striking up to 4 to 8h with reducing efficacy out to 24 and 48h which had limited inhibition relative to control (Figure 8a). Nfatc1 gene expression was analyzed at 16h and both autophagy inhibitors (3-MA and CQ) blocked RANKL induced Nfatc1 expression (Figure 8b). However, the inhibitors had no effect on SiNP inhibition of Nfatc1 expression or induction of p62 expression (Figure 8b). Taken together, the results identify the time dependent importance of autophagy for osteoclastogenesis and the selective requirement of autophagy for the expression of certain early osteoclast differentiation genes. Further, these data suggest that SiNPs inhibit autophagic flux with a nearly identical efficacy and timing as traditional autophagy inhibitors to suppress osteoclast differentiation.

Fig. 8:

Inhibition of autophagy blocks RANKL induced gene expression. (a) RAW264.7 cells were treated with RANKL (15ng/ml) and autophagy inhibitors 3-MA (500μM) or CQ (2.5μM) at indicated times after RANKL. (b) RAW264.7 cells were treated with RANKL and SiNPs (25μg/ml) and autophagy inhibitors as indicated 3MA (500μM) and (10μM) CQ, and RNA harvested after 16h. RNA was analyzed by qRT-PCR and fold change calculated using ΔΔCT method. Data are expressed as Mean ±SD. Samples run in triplicate. Representative of multiple experiments. #P<0.05, ##P<0.005, ###P<0.0005 compared to untreated, *P<0.05, **P<0.005, ***P<0.0005 relative to RANKL by student’s t test.

SiNPs stimulate autophagic initiation.

Our initial assessment of SiNP-induced gene expression changes identified an increase in autophagy related genes p62 and LC3β (Figure 1). Based on the temporal similarities to autophagy inhibitors and co-localization with autophagosomes, we hypothesized that SiNPs might impinge on the process of autophagy.

Autophagosome formation was tracked by microscopy and visualized by antibody detection of LC3β. Interestingly, our results demonstrated a clear increase in the number autophagosomes in RAW264.7 cells at 4h after addition of SiNPs (Figure 9a). Quantification of puncta per cell demonstrated an increase between 2–4h and decrease between 4–8h in the RANKL treated cells (Figure 9b). SiNPs produced a dramatic increase in total number per cell compared to RANKL alone at 2 and 8h (Figure 9b), and we observed that the decrease in puncta between 4–8h in the SiNP treated cells was blunted in comparison to RANKL alone. To determine if the SiNPs physically interacted with proteins of the autophagosome, a pulldown experiment was performed as previously described[57]. Magnetic core silica coated NPs (M-SiNPs, either OH or PEG surface) were incubated with RAW264.7 cells, lysed, and a magnet was used to separate the M-SiNPs from the cell lysate. The M-SiNPs were washed, and the bound proteins were separated by gel electrophoresis and analyzed by Western blotting. Both p62 and the active form of LC3β (LC3β-II) were clearly detected in the cell pulldown but not with M-SiNPs incubated with FBS as a negative control (Figure 9c). Additionally, an increase in the processed form of LC3β (LC3β-II), a key marker of autophagosome formation, as well as p62, were detected by Western blotting of RAW264.7 cells treated with SiNPs (Figure 9d). Taken together, the results suggest that not only do SiNPs stimulate autophagy rapidly and localize in autophagosomes, but directly interact with autophagic proteins.

Fig. 9:

SiNPs stimulate autophagosome formation and directly interact with autophagosomal proteins. (a) RAW264.7 cells were treated with RANKL (15ng/ml) or RANKL+SiNPs (25μg/ml) for times indicated and autophagosomes visualized with LC3β antibody (green) and nuclei with DAPI (blue). Scale bar=10μm. (b) Puncta were counted and averaged (40 cells/condition/2 exp.). *P<0.05, ****P<0.00005 relative to RANKL and #P<0.05, ###P<0.0005, ####P<0.00005 relative to previous time point within condition by One-way ANOVA with Sidak’s multiple comparisons test. (c) RAW264.7 cells were treated with SiNPs (25μg/ml) followed by RANKL (15ng/ml) and whole cell lysate analyzed by Western blotting at indicated time. (d) Magnetic core SiNPs (with OH or PEGlyated surface) were used to pull-down associated proteins from RAW264.7 cells or FBS (negative control). Whole cell lysate (WCL) prior to pulldown is shown for comparison. After washing the associated proteins were analyzed by Western blotting.

SiNPs inhibit autophagic flux.

To determine if SiNPs might alter autophagic flux, we utilized a previously described fluorescence-based probe GFP-LC3-RFP-LC3ΔG [58]. The construct results in equimolar amounts of GFP and RFP with LC3-GFP being degraded by autophagy and RFP-LC3ΔG being stable, allowing the ratio shift towards RFP during autophagic flux. RAW264.7 cells were stably transfected with the probe. Transfected cells were treated with RANKL, SiNPs, CQ or combinations and images acquired after 2 and 6h. Both SiNPs and CQ increased autophagic flux within 2h relative to RANKL alone which continued through 6h (Figure 10a). Quantification of the GFP to RFP ratio identified a significant inhibition in autophagic flux in response to SiNPs alone or in combination with RANKL relative to control (Figure 10b). In contrast, we observe a decrease in the GFP/RFP ratio over time following RANKL treatment, consistent with RANKL stimulating autophagy but not inhibiting flux. The previous analysis of SiNP internalization (Figure 2) identified increasing detectable cellular puncta between 0.5 and 1hr, and a decrease in autophagic flux was also determined to correspond to this 1hr time point (Figure 10c) correlating with the inhibition of RANKL signaling. Collectively, the results suggest that the SiNPs are capable of stimulating autophagy by increasing gene expression of LC3β and p62 while impairing their degradation by inhibiting autophagic flux.

Fig. 10:

SiNPs stimulate autophagosome formation but do not alter autophagic flux. (a) RAW264.7 cells were stably transfected with an autophagic flux probe (GFP-LC3-RFP-LC3ΔG) and treated with RANKL (RL) (15ng/ml), SiNPs (25μg/ml), Chloroquine (CQ) (10uM) or a combination as indicated and for times indicated and visualized by microscopy. Scale bar=20μm. (b) Total GFP/RFD fluorescence was quantified for the 6h timepoint, and ratio calculated and plotted as a percent relative to control (untreated) cells (50 cells/condition). *P<0.05, ****P<0.0001 as indicated and &P<0.05 relative to untreated by student’s t test. (c) RAW264.7 cells were treated as in (a) with SiNPs for 1h and imaged.

DISCUSSION:

Currently, therapeutic targeting of osteoclasts represents the most commonly used strategy for the treatment of osteoporosis. Bisphosphonates are a class of drugs that promote osteoclast apoptosis by either accumulation of nonhydrolyzable ATP analogs (non-nitrogen containing) or by inhibition of a key enzyme in the mevalonic acid pathway, farnesyl pyrophosphate synthase (nitrogen containing) [59]. While bisphosphonates are relatively effective at preventing bone loss and reducing fracture risk, significant side effects exists including osteonecrosis of the jaw and atypical femoral fractures, among others [60]. Further, osteoclastic bone resorption is coupled to osteoblastic bone formation and therefore the inhibition of osteoclast activity leads to a shutdown of osteoblast activity. Taken together, there exists a need for new therapeutic strategies to shift the balance between bone resorption and formation. In previous experiments it was found that intraperitoneal injection of SiNPs generated an increase in bone mineral density in both young and old mice [1, 61]. In agreement with previous functional data suggesting SiNPs inhibit osteoclasts [1, 2], mechanistic data presented herein suggest that SiNPs work by stimulating autophagy while inhibiting the NF-κB and AP-1 signaling pathway. The results suggest a novel therapeutic strategy to inhibit osteoclast differentiation with a different mechanism of action from existing antiresorptive drugs. Combined with previous studies demonstrating a positive effect on osteoblast mineralization [1, 2, 52, 57], SiNPs represent a potential novel, dual purpose osteoporosis, drug.

Growing evidence supports an important role for autophagy in regulating skeletal homeostasis in general, and osteoblasts and osteoclasts specifically [22]. Autophagy in both cell types has generally been reported as positively influencing or at least required for the differentiation processes in vitro [28]. The importance of autophagy in the process of osteoclast function has been demonstrated by the requirement of a number of autophagy genes including Beclin1 (BECN1;Atg6), p62 (SQSTM1), Atg5, Atg7, and LC3β (Map1lc3β) for differentiation, generating the osteoclast ruffled border, secretory function, and bone resorption in vitro and in vivo [31, 34, 36, 37, 62]. Interestingly, we found divergent effects of SiNPs on autophagic genes known to be important to osteoclasts. Expression of p62 and LC3β were not responsive to RANKL but strongly upregulated in the presence of SiNPs, whereas Beclin1 was upregulated by RANKL but was not altered by SiNPs early in the response and SiNPs led to a decrease in expression at 16h. Beclin1 is known to have functions beyond autophagy, such as apoptosis and membrane trafficking [63], and although the down regulation in response to SiNPs is consistent with its positive role in osteoclastogenesis the result suggests that in this context, the primary role of Beclin1 may be a non-autophagic role as previously suggested [37]. Further, the timing of the SiNP affect 8–16h, suggests it is unlikely the primary and most upstream mechanism by which SiNPs regulate osteoclast differentiation.

The current study also identified a unique response of p62 to SiNP treatment during osteoclast differentiation. The role(s) of p62 in influencing osteoclastogenesis as well as the response to autophagic activation are complex and remain only partially understood. Whereas p62 deficiency accelerates osteoclastogenesis and generates a Paget’s like phenotype in mice [64], p62 has also been demonstrated to act as a functional scaffold between RANK/TRAF and the NF-κB and Nfatc1 signaling cascades [35, 65, 66] and necessary for RANKLinduced autophagy [34] (Figure 10). SiNPs drove a strong increase in both p62 gene expression as well as protein levels whereas RANKL generated a much more muted increase. Expression of p62 is known to be regulated by stress transcriptional regulator NF-E2-related factor 2 (Nrf2) [67] as well as other common transcription factors such as AP-1 and NF-κB [68]. Stimulation of p62 gene expression by SiNPs was blocked by inhibition of the ERK1/2 pathway identifying a mechanistic link and in agreement with previous studies connecting p62 expression and ERK1/2 signaling [68]. While p62 protein levels typically decrease as autophagy flux increases, our treatment with the SiNPs during RANKL induced osteoclastogenesis upregulated both p62 mRNA and protein levels with no subsequent decrease, consistent with the SiNPs stimulating autophagy while also inhibiting autophagic flux. Interestingly, our pulldown assay identified 2 bands of p62. P62 is known to be post-translationally modified by both phosphorylation as well as ubiquitination which increases its interaction with the autophagy process [69] and therefore the result supports that SiNPs are in fact associated with active autophagy.

A cell-based study found that treatment with CQ reduced RANKL-induced osteoclast formation and function by stabilizing TRAF3 levels in osteoclast precursor cells in vitro and in vivo [38]. Interestingly, although SiNPs stimulated autophagy, they appear to also generate a similar response to autophagy inhibitors, including the stabilization of the TRAF3 protein and inhibition of osteoclast formation (compare Figure 2a with Figure 8a). This may be explained by pre-osteoclasts deriving from the monocyte/macrophage lineage and in the presence of a foreign factor, the hierarchal default for the cells is to prioritize internalization and processing of environmental stimuli over differentiation. This leads to increased expression of autophagic machinery (LC3β, p62 etc.) by the SiNPs. The increase in p62 (Figure 1), LC3β-II, and LC3 positive-puncta over time (Figure 9) following SiNP treatment supports the inhibition of autophagic flux as observed in Figure 10. This, in turn, results in the autophagic machinery required for RANKL to no longer be processed, thereby disrupting the signaling pathways downstream of RANK. This is highlighted by the reduced degradation of TRAF3 by SiNPs in response to RANKL allowing for the continued inhibition of downstream targets by TRAF3, similar to CQ treatment.

Mechanistic studies herein also identified a unique effect of SiNPs on MAPK signaling. SiNPs had divergent effects on the RANKL induced MAPK proteins in RAW264.7 cells by stimulating ERK1/2 phosphorylation and simultaneously decreasing RANKL-induced JNK phosphorylation. The stimulation of the ERK1/2 pathway agrees with other studies on SiNPs in varied cell types. For example, human adipose tissue-derived stem cells treated with 50–120nm SiNPs increased ERK1/2 phosphorylation within 10–30m [70]. Further, human bronchial epithelial cells, Beas-2B responded to porous SiNPs (25nm) with increased ERK1/2 phosphorylation after 24h [71]. Additionally, 50nm SiNPs have been found to increase ERK1/2 in osteoblasts within 15–45m, leading to an increase in autophagy[57]. However, whereas the results reported herein found an inhibitory effect on the JNK signaling pathway, previous studies of SiNPs on the JNK pathway have found either stimulation [72–75] or no effect [57, 70, 71]. It should be noted that the results presented herein are in the context of RANKL treatment, which is a strong inducer of JNK, and other studies have mostly been performed in the context of basal levels activation.

Downstream of these cellular signaling pathways are key early-stage RANKL-induced osteoclast transcriptional regulatory complexes and factors such as AP-1, NF-kB, and the central osteoclast differentiation regulator Nfatc1, which were all reduced by treatment with SiNPs. AP-1 and NF-kB family of proteins are well established to regulate Nfatc1 in response to RANKL [11, 13] and certain members from both families have been established to be required for various stages of osteoclast differentiation [12, 76]. Interestingly, the complex signaling of NF-κB, has both a canonical (activation of p50/p65) and a non-canonical pathway (RelB/p52) [77] and it appears that SiNPs are capable of inhibiting both pathways. Both pathways are thought to potentially be regulated by autophagy and both pathways have immediate upstream proteins such as IκB-α and p100 which either need to be degraded or processed however, the details of the interaction of NF-κB signaling with autophagy need to be fully elucidated. The fact that both NF-κB pathways are inhibited by SiNPs suggest the possibility that SiNPs impinge on the pathway targeting an upstream event/factor such as degradation of the TRAF proteins. Although more work is required to full elucidate the complex interaction of autophagy with the NF-κB and AP-1 pathways, the inhibition of these two transcriptional regulatory complexes by SiNPs provides a mechanism for the downstream inhibition of Nfatc1. Taken together, the data suggests a model in which treatment of differentiating osteoclasts with SiNPs creates a competition for the autophagic machinery, including p62, and in the context of limited autophagosome formation, the RANKL signaling pathway is no longer effective at stimulating the necessary downstream signaling required for activation of key transcriptional regulators such as AP-1, NF-κB, and Nfatc1(Figure 11).

Fig. 11:

Working model of the effects of SiNPs on RANKL induced osteoclast differentiation. The internalization of SiNPs stimulated phosphorylation and activation of the ERK1/2 signaling pathway and increased expression of autophagy genes p62 and LC3β. SiNPs strongly stimulate autophagy and localize to autophagosomes and autolysosomes while simultaneously inhibiting autophagic flux (dashed red line) which is required for RANKL induced signaling of the AP-1 and NF-κB transcriptional regulators to stimulate osteoclastogenesis. The blocked degradation of autolysosome components results in the inhibition of TRAF3 degradation (stabilization) thereby inhibiting downstream RANKL signaling.

This study focused solely on the mechanisms by which SiNPs alter osteoclastogenesis, however, previous studies have also defined the positive effects of SiNPs on osteoblast mineral formation [1, 2, 52, 57, 61, 78–81]. Studies on osteoblast cells using cell culture models and primary human and mouse bone marrow stromal cells (BMSCs) [1] have demonstrated that SiNPs dramatically enhanced mineralization [1, 52]. Other SiNP variants were also capable of accelerating the differentiation of MC3T3-E1 cells into mineralizing osteoblasts [1, 52, 57] demonstrating that surface and core modifications did not negatively impact the biological activity of this nanomaterial on osteoblasts. These studies also demonstrated that SiNPs dose-dependently upregulated the expression of key osteoblast matrix proteins including bone sialoprotein, osteocalcin, and osteopontin in addition to stimulating expression of Osterix, a key transcription factor involved in osteoblast differentiation. A recent study investigating physical characteristics of SiNPs found that composition (silica) and size (50–100nm) were most influential for positive effects nanoparticles on osteoblasts [2]. Importantly, SiNPs have exhibited no significant effects on osteoblastic gene induction in non-osteoblastic cell lines including monocytes, fibroblasts, kidney, and aortic smooth muscle cells which demonstrates cell specific activity[1, 61]. When combined with data presented herein the results identify SiNPs as an agent capable of simultaneously promoting osteoblastic mineralization while inhibiting osteoclastic resorption.

SiNPs demonstrate promising drug like properties for the treatment of bone disease. A study designed to assess the therapeutic index of 50nm spherical SiNPs in cell models revealed little toxicity in any cell type analyzed and when combined with an Inhibitory Concentration 50% (IC₅₀) of 9.3μg/ml, resulted in a very favorable in vitro therapeutic index (>100) [52]. In vivo toxicity has also been assessed by histology in aged mice (20mo.) treated with SiNPs (intraperitoneal) for four months and no observable effects were detected in liver, spleen, brain, or kidney tissue. Further, serum levels of Creatinine (Kidney), ALT (Liver), and TNFα (Inflammation) were assessed with no significant differences between control and SiNPs treated mice [61]. However, not all studies agree, a number of studies have found some degree of toxicity with various forms of SiNPs in RAW264.7 cells [82]. The reason(s) for the discrepancies can be at least partly explained by a combination of size and concentration with smaller particle <~30nm and concentrations >100μg/ml resulting in some toxicity in some studies, as well as the synthesis method which also influences biological response due to surface charge, aggregation, and/or surfactants [83]. Further, consumption of dietary silica and Silica (SiO₂) are defined as a “generally regarded as safe” (GRAS) agent by the US-Federal Drug Administration (FDA) with concentrations as high as 50,000 ppm producing no adverse effects in rats [84] indicating SiNPs are a safe therapeutic option for the treatment of bone disease.

Conclusion:

Osteoporosis is a significant worldwide health problem that leads to loss of bone strength and increased bone fragility which often accumulates in an increased risk of fractures. Fractures have significant health consequences, particularly in the elderly and can lead to mortality rates of 17–32% in the first year alone [85, 86]. Furthermore, only 50% of hip fracture patients are discharged to home [87]. Sustained bone health is therefore an important health issue affecting both mortality but also quality of life. Previous studies have suggested the potential use of silica nanoparticles as a novel therapeutic approach to inhibiting osteoclastogenesis [1, 2] and here we define a unique mechanism of action compared to existing antiresorptive drugs in which SiNPs, target autophagy, autophagic flux, and associated signaling pathways such as NF-κB and AP-1 signaling. Further, SiNPs were demonstrated to suppress the process of osteoclast differentiation rather than reduce mature osteoclasts via cell toxicity as with certain anti-resorptives such as bisphosphonates. Collectively, the results identify SiNPs as a representative of a novel class of drugs that influence autophagy to inhibit osteoclastogenesis and promote a positive effect on osteoblast mineralization, providing potential for an exciting new avenue of osteoporosis treatments.

MATERIALS AND METHODS:

Silica nanoparticles.

The fluorescent rhodamine B doped spherical 50 nm silica nanoparticles (as prepared silanol surface (OH)) were either synthesized in-house and characterization reported in[2] or purchased from Creative Diagnostics (New York, NY). The size and shape were characterized as spherical by transmission electron microscope (TEM) (JEOL JEM-1400, Peabody, MA) (not shown). Zeta potential and Dynamic Light Scattering measurements were performed on a Zetasizer (Malvern Instruments; Malvern UK) by Nanocomposix (San Diego, CA) as follows: Zeta Potential (−45 ± 6.81 mean mV) and size (54.36 ± 17.50 mean nm). The synthesis and characterization of metal core (cobalt ferrite) silica coated magnetic nanoparticles (M-SiNP-OH = silanol, M-SiNP-PEG = PEGylated) have been previously described as ~50nm in size and spherical [57].

Osteoclast differentiation and TRAP staining.

The murine monocyte/macrophage (pre-osteoclast) cell line RAW264.7 was obtained from ATCC (Manassas, VA) and all cell experiments were performed within 10 passages. RAW264.7 cells were grown in growth medium; Dulbecco’s Modified Eagle’s Medium (DMEM; Corning (Glendale, AZ)) supplemented with 100U/ml penicillin, 100mg/ml streptomycin, 2mM L-glutamine, and 10% fetal bovine serum (FBS; Atlanta Biologicals, Lawrenceville, GA). All cells were cultured at 37°C in 5% CO₂. To induce osteoclast formation RAW264.7 cells were plated in 48-well plates and 24 hours later growth medium changed to differentiation medium; αModified Eagle’s Medium (α-MEM; Corning) with 100U/ml penicillin, 100mg/ml streptomycin, 2mM Lglutamine and medium supplemented with 15ng/mL RANKL (R&D Systems, Minneapolis, MN). Upon formation of osteoclasts (visualized by light microscopy), usually between days 3 and 4, cells were stained for tartrate-resistant alkaline phosphatase (TRAP) (Sigma-Aldrich, St. Louis, MO) and TRAP⁺ multinucleated cells (≥ 3 nuclei) were quantitated under light microscopy and normalized for size based on number of nuclei as described [52]. Samples for protein or RNA were harvested under the same differentiation conditions at times indicated from cells cultured in 10cm plates. CQ was purchased from Acros Oraganics (Geel, Belgium) and 3-MA from Selleck chemicals (Houston, TX).

Cell viability assay.

To assess cell viability, the Live/Dead Viability/Cytotoxicity assay for mammalian cells (Thermo-Fisher Scientific) was used. Per manufacturer’s protocol, Raw264.7 cells were plated at 5×10³ cells per well in a 96well plate. After 24h, nanoparticles were added as indicated and dually stained with two fluorescently labeled probes, Calcein AM (live cell indicator) and SYTOX Deep Red Nucleic Acid Stain (dead cell indicator) and fluorescence was measured on a spectrophotometer (SpectraMax iD3, Molecular Devices (San Jose, CA)). SiNP background was subtracted, and results are expressed as precent change from untreated control.

Western blot analysis.

Cells were rinsed with DPBS and total cell lysate was generated by lysis in RIPA lysis buffer (50mM Tris-HCl (pH 8.0), 2mM EDTA, 0.5% SDS, 150mM NaCl, 0.5% sodium deoxycholate, 1% Triton X-100, 2mM fresh DTT) supplemented with Halt Protease & Phosphatase Single-Use Inhibitor Cocktail (Thermo-Fisher Scientific, Waltham MA). Cytoplasmic fraction was generated by lysis in (20mM HEPES (pH 7.5), 10mM KCl, 0.1mM EDTA, 2mM fresh DTT), and nuclear fraction in (20mM HEPES (pH 7.5), 420mM NaCl, 5mM EDTA, 10% glycerol, 5mM fresh DTT). Total cell lysate (20–30μg) or nuclear fraction (10–20μg) was separated by electrophoresis on a Novex WedgeWell 10–20% Tris-Glcine Gel (Invitrogen, Waltham MA) and electrotransferred to PVDF membrane (Thermo-Fisher). After blocking in 1x TBST with 5% non-fat dry milk, blots were incubated with specific primary antibodies overnight at 4°C and visualized by horseradish peroxidaseconjugated second antibody using Immobilon Forte Western HRP Substrate (Millipore, Burlington, MA). Nonphospho antibodies were used in 5% milk in Tris-buffered saline/Tween 20 (Bio-Rad, Hercules, CA) while phospho-specific antibodies were used in 5% Bovine Serum Albumin (Thermo-Fisher) in the Tris-buffered saline/Tween 20 solution. Antibodies were purchased as follows; from Cell Signaling Technology (Danvers, MA), LC3β (D11), p62 (5114), TRAF3 (4729), TRAF2 (C192), MafB (E309X), phospho-IKK α/β (16A6), IKK β (D30C6), phospho-IκBα (14D4), IκBα (L35A5), NF-κB1 p105/p50 (D7H5M), NF-κB2 p100/p52, RelB (D7D7W), NF-κB p65 (L8F6), phospho-ERK1/2 (D13.14.4E), ERK1/2, phospho-JNK (81E11), JNK (9252), phospho-p38 (D3F9), p38 (9212), and Histone H3 (1B1B2); from Santa Cruz Biotechnologies, Inc. (Santa Cruz, CA), β-actin (C4), BECN (E8), c-fos (4), c-Jun (D-11), Nfatc1(7A6) from LSBio (Shirley, MA) NBR1 (C332925) and from Abcam (Cambridge, UK), TRAF6 (ab33915) and IKKα (ab138426).

RNA extraction, cDNA synthesis, and qRT-PCR.

RNA was extracted using TRIzol reagent (Ambion, Austin, TX) following the manufacturer’s protocol and cDNA was synthesized using High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City CA). qRT-PCR was performed using PowerUp SYBR Green Master Mix (Applied Biosystems) on an Applied Biosystems-StepOnePlus. Primers were designed using qPrimerDepot software as described previously [88] or PrimerBank [89]. 18S: F-5’-TTGACGGAAGGGCACCACCAG-3’ and R-5’-GCACCACCACCCACGGAATCG-3’, Nfatc1: F-5’-AGTCTCTTTCCCCGACATCA-3’ and R-5’-GATCCGAAGCTCGTATGGAC-3’, C-fos: F-5’-TCCTACTACCATTCCCCAGC-3’ and R-5’-TGGCACTAGAGACGGACAGA-3’, Trap (Acp5): F-5’-GCTGGAAACCATGATCACCT-3’ and R-5’-GCAAACGGTAGTAAGGGCTG −3’, Oscar: F-5”-CCTAGCCTCATACCCCCAG-3’ and R-5-’CGTTGATCCCAGGAGTCACAA-3’, p62: F-5’-AGAATGTGGGGGAGAGTGTG-3’ and R-5’-TCTGGGGTAGTGGGTGTCAG-3’, LC3β: F-5’-GAGAAGACCTTCAAGCAGCG-3’ and R-5’-AATCACTGGGATCTTGGTGG-3’, Irf8: 5’-AGACCATGTTCCGTATCCCCT-3’ and R-5’CACAGCGTAACCTCGTCTTCC-3’, and MafB: 5’-TTCGACCTTCTCAAGTTCGACG-3’ and R-5’GAGATGGGTCTTCGGTTCAGT-3’. Beclin1 (Becn1): 5’-CCAGGAACTCACAGCTCCAT and R-5’-GGCGAGTTTCAATAAATGGC, Nbr1: 5’-TGGAGTGCGCTACCAGTGTA and R-5’-GGTTAGTGTCATGGTGTATGG. Fold change was calculated using the 2^(-ΔΔCt) method [90].

Microscopy.

RAW264.7 cells were treated with SiNPs 25 μg/mL (red) and either transferrin, fluorescein conjugated, MitoTracker, ER-Tracker, or LysoTracker (Invitrogen-Molecular Bioprobes) (green) and Nuclei were stained with Hoescht 33342 (blue) (Enzo) according to the manufacturer’s protocol after 4h. Live cell imaging was captured by Leica SP8 point scanning laser confocal microscope (63X) and cropped for enhanced magnification when noted. Visualization of LC3β and TRAF3 was by immunofluorescence with LC3β antibody (as noted above) and TRAF3 (ab36988, Abcam) following fixation at indicated times. For comparison of lysosomal size, cells were left untreated or treated with SiNPs (25ug/ml) or CQ (10uM) for 4h and then stained with LysoTracker Green and imaged as described above. The average size of lysosomes was calculated per cell and then 50 cells averaged. Quantification of LysoTracker positive puncta, LC3-positive puncta, and TRAF3 staining were preformed using image analyzing software (Fiji [91]) on a minimum of 40–50 cells per treatment.

Autophagy Flux plasmid transfection.

Autophagy flux was assessed utilizing the Sleeping Beauty transposon system to stably express a pSBbi-GFP-LC3-RFP-LC3ΔG plasmid[58]. The pSBbi-GFP-LC3-RFP-LC3ΔG plasmid was generated by inserting the GFP-LC3-RFP-LC3ΔG fragment from pMRX-IP-GFP-LC3-RFP-LC3ΔG (#84572, Addgene (Watertown, MA)) into pSBbi-Pur (#6053, Addgene). The pSBbi-GFP-LC3-RFP-LC3ΔG and the transposon expression vectors, pCMV(CAT)T7-SB100 (#34879, Addgene), were transfected into Raw264.7 cells using Fugene6 (Roche) and stable cells selected with puromycin (VWR; Radnor, PA). The resulting cells were treated with RANKL or SiNPs as indicated. For the co-treatment condition, SiNPs were added 1hr prior to RANKL. Live cell imaging was captured using a Leica SP8 point scanning laser confocal microscope (63x). Quantification of was preformed using image analyzing software (Fiji) on a minimum of 50 cells per treatment.

Transmission Electron Microscopy

The cultured cells were fixed with 2.5% glutaraldehyde in 0.1M cacodylate buffer and followed by 1% osmium tetroxide in the same buffer. After dehydration with ethanol, cells were infiltrated and eventually embedded in Epoxy resin. Ultrathin sections were counterstained with 4% uranyl acetate and lead citrate and examined on a Hitachi H-7500 transmission electron microscope.

Magnetic pulldown assay.

RAW264.7 cells were incubated with M-SiNP-OH or M-SiNP-PEG for 72h. Control samples were SiNPs incubated in Fetal Bovine Serum (FBS). Cells were rinsed with cold PBS once and lysed in E1A lysis buffer (250mM NaCl, 0.1% NP-40, and 50mM HEPES, adjusted at pH7.5) supplemented with Halt Protease & Phosphatase Inhibitor Cocktail (Thermo). Magnetic core nanoparticles (M-SiNP-OH or M-SiNP-PEG) were “pulled-down” from the lysate using Dynal magnets (Dynal Biotech ASA, Oslo Norway) ~8h initially followed by washing with lysis buffer. Two subsequent pulldown washes were performed at 4°C to further purify bound proteins. Bound proteins were analyzed by Western blotting.

Statistical analysis

Data are presented as the mean ± standard error of the mean (SEM) or standard deviation (SD). Statistical analyses were performed using Graphpad Prism (GraphPad Software, San Diego, CA). Untreated control cells were compared against RANKL treated and SiNP or drug treated cells were against the positive control osteoclasts (+RANKL) using One-way ANOVA or unpaired, two-tailed student’s t test as indicated.

Abbreviations:

3MA

3-methyladenine

CQ

Chloroquine

GFP

Green Fluorescent Protein

GRAS

generally regarded as safe

JNK

c-Jun N-terminal kinase

M-SiNPs

Magnetic core silica coated nanoparticles

MAP1LC3B/LC3β

microtubule-associated protein 1 light chain 3 beta

MAPK

Mitogen Activated Protein Kinase

Nbr1

Neighbor of BRCA1

NFATc1

nuclear factor of activated T cells 1

PEG

Polyethylene Glycol

qRT-PCR

quantitative real time polymerase chain reaction

SiNPs

spherical silica nanoparticles

RANK

Receptor Activator of Nuclear factor Kappa-B

RANKL

RANK ligand

RFP

Red Fluorescent Protein

NF-κB

nuclear factor kappa-B

TNF

Tumor Necrosis Factor

TRAP

tartrate-resistant alkaline phosphatase

TRAF

TNF Receptor-Associated Factor

DATA AVAILABILITY:

The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.