Immunostimulatory nucleic acid nanoparticles (NANPs) augment protective osteoblast and osteoclast type I interferon responses to Staphylococcus aureus

Abstract

Recalcitrant staphylococcal osteomyelitis may be due, in part, to the ability of Staphylococcus aureus to invade bone cells. However, osteoclasts and osteoblasts are now recognized to shape host responses to bacterial infection and we have recently described their ability to produce IFN-β following S. aureus infection and limit intracellular bacterial survival/propagation. Here, we have investigated the ability of novel, rationally designed, nucleic acid nanoparticles (NANPs) to induce the production of immune mediators, including IFN-β, following introduction into bone cells. We demonstrate the successful delivery of representative NANPs into osteoblasts and osteoclasts via endosomal trafficking when complexed with lipid-based carriers. Their delivery was found to differentially induce immune responses according to their composition and architecture via discrete cytosolic pattern recognition receptors. Finally, the utility of this nanoparticle technology was supported by the demonstration that immunostimulatory NANPs augment IFN-β production by S. aureus infected bone cells and reduce intracellular bacterial burden.

Immunostimulatory nucleic acid nanoparticles (NANPs) augment osteoblast (OB) and osteoclast (OC) production of IFN-β capable of significantly reducing intracellular Staphylococcus aureus burden. Figure made in BioRender.com.

Graphical Abstract

BACKGROUND

Staphylococcus aureus is the most common causative agent of osteomyelitis, and this condition is associated with severe inflammation and progressive bone loss. Staphylococcal osteomyelitis is often refractory to current treatment strategies and can manifest as recurrent infections ^1–3. This may be due, at least in part, to the ability of S. aureus to invade and persist within bone cells ^4–10. S. aureus sequestered in the cytoplasm of bone-forming osteoblasts could provide a reservoir for bacteria and contribute to recurrent chronic staphylococcal osteomyelitis that occurs despite the presence of antibiotics and a seemingly adequate humoral response. However, resident bone cells including osteoblasts are now recognized to express an array of pattern recognition receptors (PRRs), including cytosolic sensors, that enable them to respond to bacterial motifs and produce numerous immune mediators and bone regulatory factors that can shape the host response to this pathogen ^4,6,11–20. Indeed, we have recently documented the ability of osteoblasts to produce the type I interferon (IFN), IFN-β, following S. aureus infection and shown that such responses may serve a protective function by limiting intracellular bacterial survival/propagation ²⁰.

Our research team has also recently described the development and characterization of novel nucleic acid nanoparticles (NANPs) composed of DNA and/or RNA. These NANPs are designed to self-assemble various shapes including globular (3D), planar (2D), or fiber-like (1D) structures, exemplified in this work by cubes, rings, and fibers ^21–26. The immunostimulatory properties of NANPs are defined by their architectural parameters (1D vs 2D vs 3D), compositioin (RNA vs DNA), and the presence of specific ligands, all recognized by cytosolic and cell membrane-associated PRRs ^{23,24,26–29}. As such, the design principles of NANPs can be tailored to induce the production of immune mediators, including IFN-β, following their introduction into mammalian cells when complexed with an appropriate carrier to permit them to cross similarly charged cell membranes ^30–33.

In this present study, we demonstrate the delivery of representative NANPs, composed of RNA or DNA with cube, ring, and fiber connectivities, into primary murine osteoblasts and osteoclasts via endosmal trafficking when complexed with lipid-based carriers. Intracellular delivery of all the NANPs fails to disrupt bone homeostatic factor production by cells, but differentially induces immune mediator production by osteoblasts and osteoclasts according to their composition and structure. Furthermore, we show that the immune responses of bone cells to immunostimulatory NANPs are mediated by cytosolic nucleic acid PRRs, with RNA cubes eliciting responses via the recognition of 5’triphosphorylated RNA motifs by the RNA sensor retinoic acid-inducible gene-I (RIG-I), while DNA cubes induces responses via both RIG-I, through a RNA polymerase III dependent manner, and the DNA sensor cyclic GMP-AMP synthase (cGAS). Finally, the potential utility of this nanoparticle technology in bone disorders is supported by the demonstration that the introduction of immunostimulatory DNA cube NANPs augments IFN-β production by S. aureus infected osteoblasts and osteoclasts, which acts in an autocrine/paracrine manner to reduce intracellular viable bacterial burden in both cell types.

METHODS

Preparation of NANPs

All DNA templates and primers were purchased from Integrated DNA Technologies and PCR-amplified using MyTaq^™ Mix from Bioline (London, UK). NANPs were assembled as previously described and outlined in supplemental information ^25,31,34. Once prepared, all samples were stored on ice or at 4 °C until use.

Characterization of NANPs

To confirm the assembly of NANPs, all samples were run in a non-denaturing polyacrylamide gel electrophoresis (native-PAGE) followed by ethidium bromide total staining and visualization on a ChemiDoc MP Imaging System. Atomic force microscopy (AFM) was utilized to confirm the morphology of each NANP structure according to our published protocols and described in supplemental information ^29,34,35. Transmission electron microscopy (TEM) imaging was used to visualize the complexation of NANPs to lipofectamine 2000. A FEI Talos L120C TEM with a Gatan 4k X 4k OneView camera was used to image the samples on the grids ^36–38.

Isolation and culture of primary murine osteoblasts

Whole calvaria were isolated from 2–3 day-old murine neonates, maintained, and differentiated for 10 daysas described previously ^{13,18–20,39}.Differentiation was measured by assessing levels of alkaline phosphatase using a commercially available staining kit (Abcam) and microscopy, as previously described ³⁹.

Isolation and culture of bone marrow-derived osteoclasts

Bone marrow was isolated from the tibia and femur of male and female C57BL/6 mice and differentiated to osteoclasts using 100 ng/ml receptor activator of nuclear factor kappa-B ligand (RANKL) (Sigma, Cat # R0525–10UG) and 100 ng/mL M-CSF (Sigma, Cat # SRP3221–10UG), as previously described ³⁹. Differentiation was confirmed by tartrate-resistant acid phosphatase (TRAP) staining using a commercially available kit (Sigma Aldrich).

S. aureus propagation

S. aureus strain UAMS-1 was grown to mid-log phase in Luria broth (LB) and the number of colony-forming units (CFU) was determined using a Genespec3 spectrophotometer as previously described (MiraiBio Inc.) ^18,20,39.

NANP transfection of osteoblasts and osteoclasts

Cells were transfected with 10 nM nucleic acid nanoparticles (NANPs) using the lipid-based carriers, lipofectamine 2000 (Invitrogen) or DOTAP (FormuMax) as previously described for 4 hours ^{25,31,34,40,41}. At the indicated time points, cell supernatants and whole-cell protein isolates were collected for further analysis.

S. aureus infection of osteoblasts and osteoclasts

Cells were infected with S. aureus at a multiplicity of infection (MOI) of 75:1, bacteria to bone cells, in fresh antibiotic-free media for 2 hrs as previously described ^18,20,39. The media was then replaced with antibiotic-containing media to kill residual extracellular bacteria. At the indicated time points whole-cell protein isolates and cell supernatants were collected and bacterial viability was assessed by plating viable CFU/mL ²⁰.

siRNA transfection of osteoblasts

Osteoblasts were transfected according to the manufacturer’s RNAiMAX guidelines with 15 nM of the following Silencer Select siRNAs (ThermoFisher Scientific): negative control #1 siRNA, siRNA targeted against RIG-I (assay ID s106376 and 106374), siRNA targeted against RNA polymerase III subunit A (assay ID s104375 and s104374), or siRNA targeted against cGAS (assay ID s103166 and s103167) 24 hrs prior to NANP-carrier administration as described above. Cell supernatants and lysates were collected for analysis at the indicated time points.

Endosomal immunofluorescent microscopy and colocalization quantification

Osteoblasts were transfected with an Alexa Fluor 488^™-labeled DNA cubes in vitro as described above. The cells were then fixed at the indicated time points and processed for immunofluorescence microscopy for the endosomal marker, EEA1 (Invitrogen, MA5-14794, 0.04 μg/mL). The nuclei were stained using DAPI (ThermoFischer Scientific, 62248, 33.33 ng/mL). Endosomal colocalization was quantified using ImageJ (National Institute of Health, USA). The linear correlation between the endosomal and NANP thresholded selections was measured and represented by Pearson’s coefficient details are included in the supplemental information.

Enzyme-linked immunosorbent assays

Specific capture ELISAs were conducted to quantify immune mediator production in response to NANP scaffolds complexed with lipid-based carriers. TNF, OPG, and CXCL1 concentrations were assessed using commercially available ELISA DuoSet kits (R&D Systems) according to manufacturer guidelines. Concentrations of mouse IFN-β and IL-6 were conducted as previously described ^20,39. The recombinant murine protein for each ELISA was used to generate the standard curve, and the extrapolation of the absorbance to the standard curve was utilized to determine the concentration of each protein of interest in the cell supernatants.

Immunoblot analysis

Total cell lysates were evaluated by immunoblot analysis for the presence of RIG-I, cGAS, RP3, and IFIT1 using a rabbit polyclonal RIG-I antibody (Abcepta, clone AP1900A), a rabbit monoclonal antibody directed against cGAS (Cell Signaling, clone D3O8O), a rabbit monoclonal antibody directed against POLR3A, RNA polymerase III subunit A (Cell Signaling, clone D5Y2D), or an mouse monoclonal antibody directed against IFIT1 (Novus Biologicals, clone OTI3G8), respectively. Blots were reprobed with a mouse monoclonal antibody directed against β-actin (Abcam, catalog no. 49900; 0.13 μg/mL) to assess total protein loading. Densitometric analysis was conducted using ImageLab software (BioRad) and RIG-I, cGAS, RP3, and IFIT1 protein levels were normalized to the expression of the loading control, β-actin.

Statistical analysis

Data are expressed as the mean ± standard error of the mean (SEM). Commercially available software (GraphPad Prism, GraphPad Software, La Jolla, CA, United States) was used to conduct statistical analyses including Student’s t test, one-way analysis of variance (ANOVA) with Dunnett’s post hoc test, or two-way ANOVA with Šídák’s multiple comparisons test, as appropriate. For all experiments, a P value of less than 0.05 was considered statistically significant.

RESULTS

DNA and RNA NANPs can be successfully delivered into primary osteoblasts

A focused panel of selected NANPs composed of either DNA or RNA, with various structural conformations (Figure 1A), was assembled as previously described ^25,31,34. These scaffolds included a cube composed of DNA, as well as cube, ring, and fiber structures composed of RNA. Successful assembly of the NANP structures was confirmed using native-PAGE and atomic form microscopy (AFM) (Figure 1B) and such analyses demonstrated uniformity across all preparations, supporting the reproducibility of the assembly methods consistent with previous studies ^29,34,35.

FIGURE 1.

Nucleic acid nanoparticles (NANPs) composed of DNA or RNA analyzed in this work. Schematic explanation of the assembly protocols for the NANPs tested, with the relative availability of the 5’PPP in different structures shown in black bold (when available for RIG-I recognition) versus grey for inaccessible groups (Panel A). The successful assembly of DNA cubes,RNA cubes, RNA rings, and RNA fiber NANPs, was confirmed by AFM and native-PAGE (Panel B) and their complexation with L2K was visualized by TEM (Panel C).

In the absence of a carrier, NANPs are not internalized by eukaryotic cell types due to charge repulsion ^25,29,42. As such, we employed two commercially available lipid-based carriers, lipofectamine 2000 (L2K) and DOTAP, to deliver fluorescently labeled NANPs into primary murine osteoblasts. Complexation of all NANPs with L2K was confirmed by visualization using TEM (Figure 1C), and flow cytometric analysis revealed that this lipid-based carrier delivered all NANP structures with high efficiency (greater than 75% positive) at 8 hrs following administration (Figure 2A). While NANP delivery efficiency was somewhat lower with DOTAP than L2K, successful delivery of fluorescently labeled NANPs was achieved in 35–45% percent of osteoblasts at 8 hrs following administration (Figure 2B).

FIGURE 2.

DNA and RNA NANPs can be successfully delivered into primary osteoblasts using lipid-based carriers via endosomal trafficking. Alexa Fluor^™ 488-labeled DNA cube (DC), RNA cube transcribed or synthetic (3P-RC and RC), RNA ring (Ring), and RNA fiber (Fiber) NANPs (10 nM) complexed with either L2K (Panel A) or DOTAP (Panel B) were delivered to murine osteoblasts and uptake was assessed at 8 hrs by flow cytometry. Asterisks indicate statistical significance compared to the carrier alone (C) (mean +/− SEM, n = 3; Student’s t-test, p-value < 0.05). Immunofluorescence microscopy was also performed to assess NANP uptake into osteoblasts and to determine subcellular localization (Panel C). Representative immunofluorescence microscopy images of osteoblasts with Alexa Fluor^™ 488-labeled NANPs (green), DAPI-stained nuclei (blue), and Alexa Fluor^™ 555-labeled antibody staining for EEA1 (red), an early endosome marker, are shown at 8 hours following carrier alone (L2K) or NANP-carrier addition (Panel C). Endosomal colocalization was quantified at 2, 4, and 8-hours following treatment using Pearson’s correlation coefficient (Panel D). Asterisks and daggers indicate significance compared to the 2 and 4-hour time points, respectively (mean +/− SEM, n = 3; Student’s t-test, p-value < 0.05).

Our previous studies in other cell types have shown that NANPs complexed with lipid-based carriers are delivered into cells via an endosomal route ⁴⁰. The use of such uptake mechanisms in osteoblasts was supported by the demonstration that our panel of fluorescently labeled NANPs co-localizes with EEA1, an early endosomal marker, at 2 hrs post-delivery (Figure 2C), an association that decreases over 8 hrs as NANPs are trafficked to the cytosol (Figure 2D).

DNA and RNA NANPs differentially stimulate immune mediator production by osteoblasts via cytosolic pattern recognition receptors

The immunostimulatory activity of this NANP panel complexed with L2K in osteoblasts was determined as assessed by release of the key inflammatory cytokines IL-6 and TNF, the chemokine CXCL1, the type I interferon IFN-β, and the bone homeostatic factor, OPG following administration. As shown in Figure 3, RNA rings and RNA fibers were found to be immunoquiescent, failing to elicit significant immune mediator release and having no effect on the production of OPG even though all RNA monomers used for assembly of NANPs were produced via in vitro transcription and therefore contained 5’ triphophates. In contrast, both DNA and RNA-based cubes proved to be potent stimuli for inflammatory immune mediator production by primary murine osteoblasts but did not alter the production of the bone homeostatic factor OPG (Figure 3). Such differential responses suggest that distinct NANP compositions/structures might be selectively employed to serve either as immunoquiescent scaffolds for functional TNA delivery to osteoblasts, or as scaffolds possessing adjuvanticity, without effects on the homeostatic functions of these bone cells.

FIGURE 3.

NANP-carrier formulations differentially induce immune mediator production by primary osteoblasts according to their composition and structure. Osteoblasts were treated with DNA cube (DC), RNA cube transcribed and synthetic (3P-RC and RC), RNA ring (Ring), and RNA fiber (Fib) NANPs (10 nM) complexed with L2K and production of the immune mediators IL-6, TNF, CXCL1, and IFN-β, and the bone homeostatic factor OPG, was evaluated at 8 hrs by specific capture ELISAs. Asterisks indicate a statistically significant difference compared to carrier alone (C). (mean +/− SEM, n = 3–8; Student’s t-test, p < 0.05).

We then employed siRNA knockdown approaches to identify the mechanisms underlying the ability of DNA and RNA cubes complexed with L2K to stimulate osteoblast immune mediator production. As shown in Figure 4, knockdown of the cytosolic RNA sensor RIG-I (confirmed in Figure 4A) significantly reduced both IL-6 and IFN-β production induced by transcribed RNA cubes that possess 5’ triphosphorylated motifs, a known ligand for RIG-I, but not the TLR4 agonist LPS (Figures 4B and C, respectively). This finding is consistent with the inability of synthetic RNA cubes that lack this motif to elicit significant IFN-β production by osteoblasts (Figure 3). Interestingly, DNA cube-induced osteoblast immune responses were also reduced following RIG-I knockdown (Figure 4B and C). To determine whether this RIG-I dependency reflects the indirect identification of RNA 5’ triphosphorylated ligands generated via RNA polymerase III (RP3) from administered DNA, the effect of RP3 knockdown (confirmed in Figure 4D) on NANP-mediated osteoblast responses was determined. As shown in Figures 4E and F, RP3 knockdown failed to significantly effect RNA cube mediated responses but significantly reduced the production of both IL-6 and IFN-β, supporting a role for RP3 in osteoblast responses to DNA cube NANPs.

FIGURE 4:

RNA cube NANP-induced cytokine production is mediated by RIG-I in osteoblasts while DNA cube NANP responses are mediated by both RIG-I and cGAS. Cells were untreated or transfected using RNAiMAX with either siRNA (15 nM) directed against RIG-I (anti-RIG-I; Panels A-C), RNA polymerase 3 (anti-RP3; Panels D-F), cGAS (anti-cGAS; Panels G-I), or both RIG-I and cGAS (anti-RIG-I/cGAS; Panels J-L), or scrambled RNA (Control). These cells were then treated with NANPs (10 nM) complexed with L2K or carrier alone (C). At 8 hrs, RIG-I, RP3, and/or cGAS knockdown was confirmed by immunoblot analysis and expression quantified at the predicted molecular weights of 102, 165, and 62 kDa, respectively. Expression of RIG-I, RP3, and cGAS was normalized to β-actin levels and/or representative blots are shown in Panels A, D, G and J. Asterisks indicate a significant decrease in expression compared to the corresponding scrambled RNA control for each NANP stimulus (mean +/− SEM, n = 3–7; Student’s t-test, p < 0.05). Furthermore, production of the immune mediators IL-6 (Panels B, E, H, and K) and IFN-β (Panels C, F, I, and L) was assessed by specific capture ELISAs. Asterisks indicate a statistically significant difference compared to scrambled RNA treated cells (Control) for each NANP stimulus (mean +/− SEM, n = 3–7; Two-way ANOVA with Šídák’s multiple comparisons test, p < 0.05).

To determine whether additional nucleic acid sensors mediate the responses of osteoblasts to the NANP panel complexed with L2K, we have assessed the effect of knockdown of the cytosolic DNA sensor, cGAS. As shown in Figures 4H and I, cGAS knockdown (confirmed in Figure 4G) significantly reduced DNA cube-induced IL-6 and IFN-β production but had no effect on the responses of osteoblasts to RNA cubes or LPS. Furthermore, we have assessed the effect of double knockdown of both RIG-I and cGAS on the recognition of DNA cubes by osteoblasts. As shown in Figures 4K and L, cGAS knockdown (confirmed in Figure 4J) significantly reduced IL-6 and IFN-β production by DNA cube-treated osteoclasts to levels that were not statistically different from those seen in cells treated with the carrier alone (Figures 4K and L). Together, these results indicate that the immune responses of osteoblast to RNA cube NANPs are mediated by RIG-I while those to DNA cubes are initiated by both RIG-I via the actions of RP3 and by cGAS.

DNA and RNA cubes can successfully be delivered into primary osteoclasts and can similarly induce immune mediator production in a differential manner

We have extended these studies to assess our ability to deliver our NANP panel to primary bone marrow derived osteoclasts using a lipid-based carrier. Osteoclasts were transfected with fluorophore labeled NANPs complexed with L2K or treated with carrier alone and analyzed by fluorescence microscopy. As shown in the representative microscopy images in Figure 5A, all NANP scaffold structures were successfully delivered to osteoclasts. Similar to our results in osteoblasts, delivery of RNA ring and fiber NANPs failed to stimulate osteoclast production of TNF or IFN-β, supporting the immunoquiescent nature of these scaffolds (Figure 5B). In contrast, DNA and RNA cubes induced significant TNF and IFN-β production by osteoclasts further supporting the immunostimulatory activity of these NANPs. Such responses were significantly attenuated following treatment with BX795, an inhibitor of the downstream signaling components of cytosolic nucleic acid pattern recognition receptors, TBK1 and IKKε, supporting similar mechanisms underlying the initiation of osteoclast responses by these immunomodulatory nanoparticles to those seen in osteoblasts.

FIGURE 5.

DNA and RNA NANPs can be successfully delivered into primary osteoclasts using a lipid-based carrier and differentially induce immune mediator production in a TBK1/IKK-dependent manner. Alexa Fluor^™ 488-labeled DNA cube (DC), RNA cube transcribed (3P-RC), RNA ring (Ring), and RNA fiber (Fib) NANPs (10 nM) complexed with L2K (Carrier) were delivered to murine osteoclasts and uptake was assessed at 4 hrs by fluorescence microscopy (Panel A). At 8 hrs following treatment, production of the immune mediators IL-6 and IFN-β was evaluated by specific capture ELISAs (Panel B). Asterisks indicate a statistically significant difference compared to carrier alone (mean +/− SEM, n = 3; Student’s t-test p < 0.05). In Panel C, osteoclasts were treated with the TBK1/IKKε inhibitor BX795 (1 mM) for 2 hours prior to treatment with DNA-based cube NANPs (10 nM) complexed with L2K or transfection with BDNA (1 mg) and assessment of IFN-β production at 8 hrs by specific capture ELISA. Asterisks and daggers indicate a statistically significant difference compared to the corresponding carrier alone or similarly treated cells in the absence of BX795, respectively (mean +/− SEM, n = 3; Two-way ANOVA with Šídák’s multiple comparisons test, p < 0.05).

DNA cube stimulated IFN-β production by primary osteoblasts and osteoclasts reduces intracellular S. aureus burden following infection

Previously, we have documented that IFN-β treatment of osteoblasts serves to restrict intracellular bacterial burden ²⁰. Here, we have begun to evaluate the ability of immunostimulatory NANPs to limit the number of bacteria harbored by S. aureus infected bone cells via IFN-β production. We selected to employ the DNA cube for the following experiments due the high stability, cost effective assembly, and dual agonist activity of DNA cube NANPs. As shown in Figure 6A, DNA cube NANPs complexed with L2K induces IFN-β production by osteoblasts and such production is associated with a significant reduction in intracellular S. aureus burden. Importantly, we have confirmed that this effect occurs via IFNAR with the demonstration that the STAT1 activation inhibitor, fludarabine, significantly attenuates DNA cube-induced increases in the expression of the ISG product IFIT1 in both uninfected and S. aureus infected osteoblasts, and abolishes the ability of this NANP to restrict intracellular bacterial burden (Figure 6B). Similarly, administration of DNA cubes complexed with L2K induces IFN-β production by osteoclasts and such production is also associated with bacterial burden restriction in S. aureus infected cells (Figure 6C).

FIGURE 6.

DNA cube NANP-stimulated IFN-β production by primary osteoblasts (Panels A and B) and osteoclasts (Panel C) reduces intracellular S. aureus burden following infection. Panel A: Osteoblasts were either treated with carrier alone (Control) or DNA-based cube NANPs (10 nM) complexed with L2K for 2 hrs, and were then uninfected or infected with S. aureus (SA: MOI of 75:1 bacteria to bone cells). At 8 hrs, IFN-b production was assessed by specific capture ELISA and viable intracellular bacterial burden was assessed at 24 hrs by colony counting. Panel B: Osteoblasts were untreated or treated with the STAT1 inhibitor fludarabine (5 μM) for 2 hrs prior to treatment with carrier alone (C), recombinant IFN-β (IFNb: 1 ng/mL), and infected with S. aureus (SA: MOI of 75:1) with and without treatment with DNA-based cube NANPs (10 nM) complexed with L2K (DC). At 8 hrs, IFIT1 protein expression was quantified by immunoblot analysis and normalized to β-actin levels, and viable intracellular bacterial burden was assessed by colony counting. Panel C: Osteoclasts were either treated with carrier alone (Control) or DNA-based cube NANPs (10 nM) complexed with L2K for 2 hrs, and were then infected with S. aureus (SA: MOI of 75:1 bacteria to bone cells). At 8 hrs IFN-b production was assessed by specific capture ELISA and viable intracellular bacterial burden was assessed at 24 hrs by colony counting. Data is shown as the mean +/− SEM, n = 3. Asterisks and daggers indicate statistically significant differences from untreated or uninfected control cells and untreated similarly stimulated cells, respectively (Student’s t-test, p < 0.05).

DISCUSSION

Due to the biological functions of nucleic acids, therapeutic nucleic acids can be applied to a broad range of applications including aptamers to recognize and bind target molecules, RNA interference-induced gene silencing, mRNA induced gene expression, and immunomodulatory agents ²⁸. However, the development of such therapies has been hampered by off-target effects or detrimental immune responses ⁴³. Due to the physiochemical properties of RNA and DNA, these NANPs can be designed with specific thermo and enzymatic stabilities ⁴¹. In addition, the intracellular delivery of these novel agents can be optimized according to their complexation with various carrier reagents, since these NANPs are not internalized by mammalian cells in the absence of a carrier due to charge repulsion. Most importantly, the immunomodulatory properties of NANPs can be engineered via the inclusion or exclusion of known ligands for specific PRRs ²⁸. Therefore, these NANPs can be manufactured to function either as immunoquiescent scaffolds or as immunostimulatory structures that induce potentially beneficial cell responses, such as the production of type I IFNs.

Staphylococcal osteomyelitis can result in devastating bone damage and is often refractory to current treatment strategies including systemic antibiotic treatment and surgical debridement of necrotic tissue ^1,2,44,45. Such chronic infections may be due, at least in part, to the ability of S. aureus to invade and persist within the resident bone cells, osteoblasts and osteoclasts ^{4–6,8,9,46,47}. However, it has become increasingly apparent that resident bone cells are capable of detecting and responding to the presence of intracellular S. aureus via cytosolic PRRs for bacterial motifs ^15,16,48. Furthermore, we have recently shown that primary osteoblasts produce IFN-β in response to S. aureus infection, which can then act in an autocrine and/or paracrine manner to reduce intracellular bacterial burden in this cell type ²⁰. Such responses may serve to mitigate staphylococcal bone infections and their augmentation could be exploited as a therapeutic invention.

In the present study, we have demonstrated our ability to deliver a representative panel of NANPs of differing compositions and structures into primary murine osteoblasts and bone marrow-derived osteoclasts when complexed with two commercially available lipid-based carrier reagents. We show that DNA cubes and RNA cubes, rings, and fibers, complexed with either lipofectamine or DOTAP, enter the cytosol of bone cells via endosomal trafficking, consistent with previous mechanistic descriptions of these carrier agents ^22,31,36,38. Delivery of these NANPs failed to significantly impact the production of the critical bone homeostatic factor, OPG, that functions as the decoy receptor for the potent osteoclastogenic mediator RANKL ⁴⁹. Such a finding is in contrast with the effects of S. aureus infection, which is associated with a significant decrease in the production of OPG with a concomitant increase in the production of RANKL ^39,50,51, and indicates that these NANPs can be administered into primary bone cells without disrupting their homeostatic functions.

Importantly, we have determined that our NANPs are immunoquiescent or elicit immune mediator production by osteoblasts and osteoclasts according to their composition and shape, consistent with our previous studies in other cell types ^{25,29,31,40,41,52}. We show that RNA rings and fibers fail to elicit the production of key inflammatory cytokines and chemokines, or type I interferons by either osteoblasts or osteoclasts despite inclusion of the 5’ triphosphate motif. Based on our published studies we predict that due to the design principles of the RNA rings and fibers the 5’ triphosphate motif in less accessible for recognition by PRRs ²². As such, these NANP structures represent suitable scaffolds for the delivery of therapeutic nucleic acids including antisense oligonucleotides, mRNAs, siRNAs, miRNAs, and aptamers. Furthermore, such immunoquiescent scaffolds could be utilized to allow the coordinated simultaneous delivery of multiple therapeutic functional groups to each bone cell. In contrast, DNA and RNA cubes were found to be immunostimulatory, inducing the production of the inflammatory cytokines, IL-6 and TNF, the chemokine, CXCL1, and the type I interferon, IFN-β, by osteoclasts and/or osteoblasts.

The mechanisms underlying the immunostimulatory activity of the RNA and DNA cubes was investigated using siRNA knockdown approaches directed against the cytosolic RNA and DNA sensing molecules, RIG-I and cGAS, respectively. These approaches demonstrated that RIG-I is required for immune production by osteoblasts in response to RNA cube administration. This finding was predicted since our transcribed RNA cubes are assembled with transcribed RNA strands possessing 5’triphosphorylated motifs, a known ligand for RIG-I ⁵³. Such a mechanism is further supported by the observation that our synthetic RNA cubes, which do not possess 5’triphosphorylated motifs, failed to elicit osteoblast immune responses. Somewhat surprisingly, osteoblast responses to DNA cubes were found to be mediated by both the cytosolic DNA sensor, cGAS, and RIG-I. However, this is consistent with the known ability of poly(dA:dT) to function as an indirect agonist for RIG-I via the ability of RNA polymerase III to generate RNA from the DNA containing 5’triphosphorylated motifs ^54,55. Such a hypothesis is supported by demonstration that osteoblast responses to DNA cubes was significantly reduced following RNA polymerase III knockdown.

In contrast, DNA cubes appear to function directly as a cGAS agonist in osteoblasts. Recognition of dsDNA by cGAS is sequence-independent due to electrostatic interactions with the sugar phosphate backbone, but this sensor displays preferential recognition of B-form dsDNA with a minimum of 20–40 base-pairs in length ^56–61. Consistent with this, the DNA cubes in the present study are composed of six 52 base-pair complementary strands that assemble into a 3-dimensional cube possessing B-form double-stranded DNA helices. Similarly, immune mediator production by osteoclasts in response to RNA and DNA cubes appears to occur via these sensor molecules as evidenced by the ability of a pharmacological inhibitor of TBK1/IKKε catalytic activity, which occurs downstream of RIG-I, and cGAS activation ^57,62, to significantly reduced both TNF and IFN-β production by these cells. Together, these data support the notion that our stimulatory NANPs elicit osteoblast and osteoclast responses via the specific activation of the cytosolic nucleic acid PRRs, RIG-I and cGAS, based upon their structural composition supporting their further development as immunomodulatory agents.

To further examine this possibility, we have assessed the ability of an immunostimulatory NANP to augment IFN-β production by S. aureus infected osteoblasts and osteoclasts, and determined its effect on intracellular bacterial burden. Our results indicate that DNA cube NANPs can markedly elevate IFN-β release by both cell types following infection. Importantly, we have determined that such responses significantly reduce the number of viable intracellular bacteria harbored by osteoblasts and osteoclasts, and that this occurs via autocrine or paracrine signalling through the IFN-α/β receptor. These data support a mechanism, summarized in Figure 7, in which DNA cube NANPs, complexed with a lipid-based carrier, stimulate RIG-I and cGAS dependent IFN-β production by primary bone cells that restricts intracellular S. aureus growth and/or survival in these cells. Collectively, this study supports the further development of these versatile NANPs as both immunoquiescent scaffolds for the delivery of therapeutic nucleic acids into bone cells and as novel PRR agonists capable of promoting protective host cell responses to bacterial infection.

FIGURE 7.

Immunogenic nucleic acid nanoparticles augment protective bone cell responses to bacterial infection. DNA (DC) and RNA transcribed cube (3P-RC) nucleic acid nanoparticles complexed to a lipid-based carrier traffic to the cytosol via an endosomal pathway. DC and 3P-RC NANPs stimulate production of the type I IFN, IFN-b, via activation of the cytosolic nucleic acid sensors, cGAS and RIG-I. IFN-β autocrine and paracrine signalling via the IFN-α/β receptor restricts intracellular Staphylococcus aureus burden in osteoblasts and osteoclasts. Figure made in BioRender.com.

HIGHLIGHTS.

• DNA and RNA nucleic acid nanoparticles can be delivered into primary bone cells.

• DNA and RNA cubes stimulate RIG-I and cGAS dependent production of type I IFNs by bone cells.

• DNA cubes stimulate type I IFNs by bone cells that restrict intracellular bacterial burden.

DATA AVAILABILITY

The authors confirm that the data supporting the findings of this study are available within the article and/or its supplementary materials. Any additional data that support the findings of this study are available from the corresponding author (MBJ) upon reasonable request.