Therapeutic Potential of Vasoactive Intestinal Peptide and its Derivative Stearyl-Norleucine-VIP in Inflammation-Induced Osteolysis

The common use of dental and orthopedic implants calls for special attention to the immune response leading to peri-prosthetic bone loss and implant failure. In addition to the well-established microbial etiology for oral implant failure, wear debris and in particular titanium (Ti) particles (TiP) in the implant vicinity are an important trigger of inflammation and activation of bone resorption around oral and orthopedic implants, presenting an unmet medical need. Here, we employed bacterial-derived lipopolysaccharides (LPS) to model infection and TiP to model aseptic inflammation and osteolysis. We assessed inflammation in vitro by measuring IL1β, IL6 and TNFα mRNA expression in primary macrophages, osteoclastogenesis in RANKL-induced bone marrow derived pre-osteoclasts and osteolysis in vivo in a mouse calvarial model. We also assessed the trans-epithelial penetrability and safety of the tested compound in rats. Our results show that a lipophilic super-active derivative of vasoactive intestinal peptide (VIP), namely stearyl-norleucine-VIP (SNV) presented superior anti-inflammatory and anti-osteoclastogenic effects compared to VIP in vitro. In the bacterial infection model (LPS), SNV significantly reduced IL1β expression, while VIP increased IL6 expression. In the aseptic models of osteolysis, SNV showed greater suppression of in vitro osteoclastogenesis than VIP, and significantly inhibited inflammation-induced osteolysis in vivo. We also observed that expression levels of the VIP receptor VPAC-2, but not that of VPAC-1, dramatically decreased during osteoclast differentiation. Importantly, SNV previously shown to have an increased stability compared to VIP, showed here significant trans-epithelial penetration and a clean toxicological profile, presenting a novel drug candidate that could be applied topically to counter both aseptic and infection-related bone destruction.

The common use of dental and orthopedic implants calls for special attention to the immune response leading to peri-prosthetic bone loss and implant failure. In addition to the well-established microbial etiology for oral implant failure, wear debris and in particular titanium (Ti) particles (TiP) in the implant vicinity are an important trigger of inflammation and activation of bone resorption around oral and orthopedic implants, presenting an unmet medical need. Here, we employed bacterial-derived lipopolysaccharides (LPS) to model infection and TiP to model aseptic inflammation and osteolysis. We assessed inflammation in vitro by measuring IL1β, IL6 and TNFα mRNA expression in primary macrophages, osteoclastogenesis in RANKL-induced bone marrow derived pre-osteoclasts and osteolysis in vivo in a mouse calvarial model. We also assessed the trans-epithelial penetrability and safety of the tested compound in rats. Our results show that a lipophilic super-active derivative of vasoactive intestinal peptide (VIP), namely stearylnorleucine-VIP (SNV) presented superior anti-inflammatory and anti-osteoclastogenic effects compared to VIP in vitro. In the bacterial infection model (LPS), SNV significantly reduced IL1β expression, while VIP increased IL6 expression. In the aseptic models of osteolysis, SNV showed greater suppression of in vitro osteoclastogenesis than VIP, and significantly inhibited inflammation-induced osteolysis in vivo. We also observed that expression levels of the VIP receptor VPAC-2, but not that of VPAC-1, dramatically decreased during osteoclast differentiation. Importantly, SNV previously shown to have an increased stability compared to VIP, showed here significant trans-epithelial penetration and a clean toxicological profile, presenting a novel drug candidate that could be applied topically to counter both aseptic and infection-related bone destruction.

INTRODUCTION
The common use of dental and orthopedic implants calls for special attention to the immune response leading to periprosthetic bone loss and implant failure. In dentistry, periimplantitis is a recent but already major clinical concern and the main cause of long-term implant failure (Esposito et al., 2005;Heitz-Mayfield and Mombelli, 2014). Triggered by specific oral bacteria, it consists of an inflammatory process that leads to bone resorption (osteolysis) around dental implants . Once the process starts it can hardly be controlled and often results in implant loss (Esposito et al., 2012). Treatments usually begin with attempts to preserve the implant by mechanical cleaning of the surrounding oral flora using ultrasonic-scaling and local and systemic antibiotic administration Heitz-Mayfield and Mombelli, 2014). When these options fail to restrain the destructive process, the implant is surgically removed.
In addition to the well-established microbial etiology for oral peri-implant and periodontal bone loss (peri-implantitis and periodontitis, respectively), wear debris and in particular titanium (Ti) particles (TiP) in the implant vicinity are an important trigger of inflammation and activation of bone resorption. Whereas Ti is a biocompatible material, which does not induce an immune response, previous reports suggested that ions and particles shedding from implant alloys may do the opposite (Revell, 2008;Eger et al., 2017). Particles of Ti (TiP), among other metals released from the implant surface, cause an inflammatory response strikingly similar to that induced by lipopolysaccharides (a model for bacterial infection) in macrophages (Eger et al., 2018). These pro-inflammatory cytokines include IL1β, IL6 and TNFα, which are known to have a key role in stimulating bone resorption around teeth and implants (Onodera et al., 1993;Weingart et al., 1994;Vallés et al., 2008a;Vallés et al., 2008b). Aseptic loosening due to particle debris is among the main causing factors for hip joint implant failure (Revell, 2008). These particles accumulate in the environment of the implant and induce a strong inflammatory response that eventually leads to bone resorption. High concentrations of TiP were also found around failed implants (Lee et al., 2012). With time, the gradual increase in the number of particles reaches a critical concentration that leads to periprosthetic osteolysis and may jeopardize implant survival (Goodman et al., 2014).
Vasoactive intestinal peptide (VIP) is a 28 amino-acid hydrophilic peptide (Gozes, 2008) that may inhibit both inflammation (Gozes, 2008) and osteoclastogenesis , thus portraying this peptide as a promising candidate in the management on inflammationinduced osteolysis. It acts as a neurohormone and neurotransmitter and is involved in multiple physiological actions such as vasodilation, bronchodilation, cell division, and neuroprotection (Gozes, 2008).
VIP derivatives and conjugates were designed to include a lipophilic moiety or a shortened VIP chain (Gozes and Fridkin, 1992;Gozes et al., 1994;Gozes et al., 1995;Gozes et al., 1996;Gozes et al., 1999). Stearyl-norleucine-VIP (Stearyl-Nle 17 VIP, SNV) is a lipophilic compound consisting of a VIP molecule with an N-terminal attachment of stearic acid to increase cell and tissue bioavailability and an exchange of the oxidation-vulnerable methionine at position 17 with the stable norleucine. As cited above, SNV is 1000-fold more potent than VIP in terms of neuroprotection and cGMP formation  with a 15 min compared to ∼30 s half-life (Gozes et al., 1994;Gozes et al., 1996), and hence tested here for its therapeutic potential in inflammation-induced osteolysis. Furthermore, the stearyl lipophilic moiety allows for better skin penetration as previously described (Gozes et al., 1994), and as shown in detail below.

MATERIALS AND METHODS
All procedures involving animals were carried out in accordance with the guidelines of Tel Aviv University and were approved by the Institutional Animal Care and Use Committee (permit number M-015-047).
Cell culture. Primary bone marrow-derived macrophages (BMDMs) were isolated from the femora and tibiae of 10week-old C57BL/6J mice (Envigo, Israel), as previously described (Hiram-Bab et al., 2015). Briefly, cells were cultured overnight at 37°C in a humidified atmosphere with 5% CO 2 in our "standard medium" consisting of alpha-modified Eagle's medium (αMEM, Life Science Technology, NY, United States) and 10% fetal bovine serum (FBS, Rhenium, Ltd., Modi'in, Israel). After 24 h, the non-adherent fraction was cultured in 10-cm non-culture-treated dishes containing standard medium and 100 ng/ml macrophage colony stimulating factor (M-CSF) (Takeshita et al., 2000). The resulting adherent BMDMs were collected after 3 days for the specific assays described below.
Particle generation. To obtain Ti particles that correspond to the particles shedding from oral implants during routine scaling, we subjected Ti discs that were made from Ti6Al4V (AlphaBio Tec., Petah-Tikva, Israel) to ultrasonic (US) scaling (Newtron Led, Satelec, Acteon, Marignac, France), adjusted to a frequency of 32 kHz. Particles were obtained from discs with a sand-blasted and acid-etched (SLA) surface topography as described previously (Eger et al., 2017). All particles were generated in a sterile environment. Each disc was subjected to US scaling for 60 s in distilled water (ddH 2 O), then cleaned twice with ethanol, and finally resuspended in distilled water. We previously showed that each 6 mm diameter disc generates ∼2.54 million particles on average. In all our in vitro assays and for the preparation of the fibrinogen-thrombin membranes (see below) we used a particle density of 1293 particles/mm 2 .
MTT assay. For proliferation assay, macrophages were plated (4000 cells/well) into 96-well plate with standard medium supplemented with 20 ng of M-CSF. Cells were treated as indicated with either TiP, SNV, VIP (10 −6 M) or vehicle at first day of incubation. Cells viability was determined using dimethylthiazol-diphenyl-tetrazolium-bromide (MTT) after 1, 3 or 5 days. MTT was added to a final concentration of 5 mg/ ml and incubated for 4 h at 37°C. After complete solubilization of the die in DMSO, plates were read at 570 nm in a colorimetric plate reader.
RNA isolation, and RT-qPCR. Following a 24 h incubation with Ti particles (or LPS/vehicle only/SNV/VIP in addition to Ti particles), macrophages were washed with sterile PBS, and RNA was extracted using Tri-RNA Reagent (Favorgen Biotech Corp, Kaohsiung, Taiwan). The 260/280 absorbance ratio was measured to verify the RNA purity and concentration. cDNA was produced using a high-capacity cDNA reverse transcription kit (Invitrogen, Grand Island, NY, United States), and real-time PCR was performed using Kapa SYBR Fast qPCR (Kapa Biosystems, Wilmington, MA, United States) on a StepOne real-time PCR machine (Applied Biosystems, Grand Island, NY, United States).
The following primers were used: F-GAAATGCCACCTTT TGACAGTG and R-TGGATGCTCTCATCAGGACAG for mouse IL1β; F-TAGTCCTTCCTACCCCAATTTCC and R-TTG GTCCTTAGCCACTCCTTC for mouse IL6; F-TCTTCTCATTC CTGCTTGTGG and R-GGTCTGGGCCATAGAACTGA for mouse TNFα; F-CCGTAACTGCACTGAAGA-AG and R-CTG TTGCTGCTCATCCATAC for VPAC-1; F-CAGCAGACC AGGAAACAT-AA and R-GCCACACGCATCTATGAA for VPAC-2 and F-ACCCAGAAGACTGTGGATGG and R-CACA TTGGGGGTAGGAACAC for Gapdh. The reaction was subjected to 40 cycles of amplification using the following program: 95°C for 20 s, 60°C for 20 s, and 72°C for 25 s. The relative mRNA expression levels of the selected genes were normalized to the level of Gapdh.
Osteoclastogenesis assay. Preosteoclasts, prepared like the BMDMs, were plated in 96-well plates (7,000 cells per well, for TRAP staining, see below) or in 6-well plates (200,000 cells per well, for RNA) in standard medium supplemented with 20 ng/ml M-CSF and 50 ng/ml RANKL (R&D Systems, Minneapolis, MN, United States). After 48 h, the medium was replaced by the conditioned medium of BMDM, supplemented with RANKL and M-CSF. Where indicated, 10 −6 M SNV or VIP were added. After 30 h, cells were stained using a TRAP kit (Sigma-Aldrich, St. Louis, MO, United States), and multinucleated (>3 nuclei) TRAP-positive cells were defined as osteoclasts. Images were acquired at an original magnification of ×4 (Evos FLC, Life Technologies, MS, United States). The number of osteoclasts and the total osteoclast area were measured using ImageJ software (National Institutes of Health, Bethesda, MD, United States).
Toxicology. We performed toxicology studies for SNV that consisted of acute subcutaneous, intravenous and oral toxicity in rats, acute dermal toxicity in rabbits, skin sensitization using adjuvant and patch test in guinea pigs, and single and repeated (for 13 weeks) dose toxicity studies in rats. The latter was a 90days repeated dose toxicity study, consisting of SNV once a day at three different doses to the penis and vagina of male and female rats, respectively. The compound was tested for the treatment of erectile dysfunction. The active dose (7 µg SNV) was chosen as the lowest dose for the study. The highest dosage group in the repeated dose toxicity study received 3,500 µg (Gozes et al., 1994). For study 001 (parts 1 through 3), the 1× dose consisted of 0.5 mg SNV dissolved in 7.1 ml vehicle (3.55 ml 10% Sefsol [glycerin monocaprylate] + 3.55 ml isopropanol) and 100µl/animal were administered; the 1000X dose consisted of 91 mg SNV dissolved in 1.3 ml vehicle (0.65 ml 10% Sefsol +0.65 ml isopropanol) and 100µl/animal were administered. For studies 002 through 004, the indicated doses of SNV (Table 1) were dissolved in 0.25 ml 10% Sefsol +0.25 ml 40% isopropanol (vehicle). The methodology for each of the toxicology studies is detailed in Table 1.
Although, the studies were not carried out under Good Laboratory Practices (GLP) stipulations, SNV was synthesized under GLP conditions, and all the work was carried out at Tel Aviv University under meticulous conditions using professional staff and established methods (in a designated air-conditioned, clean room and specific luminary flow hood), which all strengthen the trustworthiness of the study results. Ames tests were carried out by Life Science Research Israel Ltd. Ness Ziona, Israel.
Rat model of transepithelial penetrability. Because of the high hydrophobicity of SNV, we employed an iodination labeling protocol based on the Chloramine T method (Markwell, 1982) using 1 mCi Na 125 I with a few modifications that included 1) the dissolution of SNV in dimethylformamide (DMF), 2) replacement of the phosphate buffer with 0.2 M HEPES (pH 7.6), a more compatible buffer, and 3) reaction termination by the addition of sodium metabisulfite and KI. The pharmacokinetic profile of 125 I-SNV absorption and distribution was evaluated following vaginal delivery of the compound to rats. Wistar rats at the estrus phase received 50 µl (7.0 µCi/rat, 15 million CPM, 100 Ci/mM specific activity) of 125 I-SNV dissolved in 5% Sefsol, applied directly in the vagina, using a P100 micro pipettor. At the indicated time points, we anesthetized the rats with chloral hydrate and collected blood before cervical dislocation. We then harvested the uterus, liver, lungs, heart, intestine, kidneys and vagina (duplicate samples were taken from each organ). Organs were weighed and radioactivity levels were measured using a gamma counter. Each point on the chart represents at least two rats.
Animal model and micro-computed tomography (μCT). We used our calvarial model, as described previously (Eger et al., 2017). Briefly, US-released TiP (from SLA-treated discs) were incorporated into a fibrinogen-thrombin degradable membrane used as a scaffold to localize the TiP, and membranes with no particles were prepared as positive and negative controls, respectively. As indicated, 2 × 10 -8 mol SNV (or saline as control) were incorporated into the membrane together with the TiP. The parietal bones of the 10-week-old C57Bl/6J female mice were exposed, and the periosteum was removed before inserting the fibrinogen-thrombin membranes to cover both parietal bones. In the control group an empty fibrinogenthrombin membrane was inserted (with no TiP). All groups comprised six animals. Animals were euthanized 5 weeks post insertion, and the skull of each mouse was removed, fixed for 24 h in 4% phosphate-buffered formalin, followed by 70% ethanol. All specimens were scanned and analyzed using a μCT system (μCT 50, Scanco Medical AG, Switzerland). Scans were performed at a As in study 001-1, except injection was intravenous Necrotic reaction was observed at the site of injection in 55% group 2, 58% group 3 and 100% group 4. Most surviving rats displayed normal body weight, and some weight loss observed initially was recovered during the study 7 rats (3 males and 4 females) out of 12 died within 3 h after administration in group 4 (SNV -1000X). One rat died in group 2 within 5-24 h after administration. No mortality occurred in groups 1 and 3 Under the conditions of this study, the acute intravenous median lethal dose of SNV was estimated to be 7 mg/male rat, and due to the higher mortality in females, it was estimated to be less than 7 mg for males and females combined. In view of necrotic reactions in group 2vehicle, isopropanol was reduced from 50 to 20% in the other studies (002-004) N 6 males (160-350 g) N 6 females (155-285 g), for each of the experimental groups 001-3 acute oral toxicity in rats As in study 001- 10-μm resolution in all three spatial dimensions, with 90 kV energy, 88 μA intensity, and 1000 projections at a 1000 msec integration time. The volume of interest (VOI) was defined as two 3.7 mm circles in the center of the parietal bones. A custom-made algorithm, based on Image-Processing Language (IPL, Scanco Medical), was developed to isolate the resorption pits, defined as unmineralized volumes that were 10-40 μm deep on the bone surface (as in (Eger et al., 2017), Figure 1). Morphometric  parameters were determined at the 3D level inside a fixed VOI total volume (TV) and included the total volume of the bone resorption pits (Pit Resorption Volume, PRV, mm 3 ), and the calculated PRV/ TV ratio (Figure 1). We also measured the mean calvarial plate thickness using a mask that obliterated the bone marrow volume and resorption pits (Figure 1). This parameter is therefore independent of the resorption volume and it discriminates between actual osteolysis (no change in the calvarial plate thickness) and irregular bone thickening (increase in the calvarial plate thickness). The values used for each animal are the average of the right and left sides. Statistical analyses. Values are expressed as the mean ± SD unless otherwise indicated. Statistical analyses were performed using GraphPad Prism 7.0 (La Jolla, CA, United States). As all presented data typically display a normal distribution, t-test (between two groups) or analysis of variance (ANOVA) with Tukey's post hoc test (multiple group comparison) were used. Differences between groups were defined as significant at p < 0.05.

SNV Presents Anti-inflammatory Effects in Cell Cultures
IL1β, IL6 and TNFα are the main mediators of Ti particlemediated inflammation and osteolysis (Eger et al., 2017;Eger et al., 2018). The immuno-modulating role of VIP is well established (Harmar et al., 2012) and we first examined the effect of both VIP and SNV on the expression of these proinflammatory cytokines in macrophages. We treated bone marrow-derived macrophages with 10 −6 M VIP or SNV (or saline control) 1 h before adding LPS (1 μg/ml). Control cultures were left untreated (M-CSF only). After 24 h, we extracted RNA and observed a significant increase in the expression levels of the three cytokines in the LPS-treated vs. control macrophage cultures. VIP demonstrated mixed immunomodulatory effects, decreasing IL1β but increasing IL6 (-40% and +90%, respectively, p < 0.05, Figure 2) and presenting no change on TNFα expression. SNV had no effect on IL6 and TNFα expression levels but it significantly reduced the expression levels of IL1β (by 35%, p < 0.05, Figure 2), suggesting a net antiinflammatory effect.

SNV Suppresses Osteoclastogenesis
We then examined the direct effect of VIP and SNV on osteoclastogenesis. First, we asked whether the transcripts coding for VPAC-1 and VPAC-2, the classical receptors for VIP and SNV, are expressed in preosteoclasts during the differentiation process. BMDM, which are osteoclasts precursors in vitro, were cultured in the presence of M-CSF and RANKL to induce osteoclast differentiation. VPAC-1 and VPAC-2 transcripts were both expressed in BMDM before the addition of RANKL (Day 0) and these levels declined by 58 and 92%, respectively, within 2 days of osteoclastogenesis ( Figures  3A,B). We then assessed the effect of VIP and SNV on osteoclast differentiation in the absence of inflammation. In line with previous reports   , VIP significantly inhibited osteoclastogenesis as indicated by the reduced osteoclast number and total area ( Figures 3C-E). SNV had a comparable inhibitory effect on osteoclast differentiation with no difference between the effect of VIP and that of SNV in this assay. We next exposed BMDM to TiP and as previously reported by us, we found that IL1β, IL6 and TNFα were significantly elevated, similarly to the response to LPS (Figure 2, 4A). We also collected the supernatant from macrophages exposed to TiP and added it to day-2 osteoclastogenic cultures together with VIP, SNV or no treatment (Figures 4B-D). Notably, the supernatant of TiP-exposed macrophages had a significant effect on osteoclastogenesis, increasing osteoclast area by 57%. In these FIGURE 2 | VIP and SNV displayed a partial anti-inflammatory effect in LPS-stimulated macrophage cultures. Primary murine macrophages were treated with 10 -6 M of VIP or SNV (or saline) and exposed to 1 µg/ml LPS for 24 h. Expression of IL1β, IL6 and TNFα was examined using RT-qPCR and presented as fold from untreated controls (no LPS). n 5, # p<0.05 vs. Control; *p<0.05 vs. LPS, 1-way ANOVA.
Frontiers in Pharmacology | www.frontiersin.org May 2021 | Volume 12 | Article 638128 inflammatory settings, VIP treatment reduced osteoclast area and SNV attenuated both osteoclast area and number, displaying a stronger inhibitory effect on osteoclasts than VIP. This inhibitory effect was not attributable to cytotoxicity as neither VIP nor SNV affected the viability of the cells in the culture over 5 days of treatment (MTT assay, Figure 4E).

SNV Is Skin Bioavailable
Based on these findings, and due to former pharmacological experiments indicating that the therapeutic applications of VIP have been hampered by its very short half-life and low penetration through lipidic barriers (skin and epithelium), we elected SNV as a potential therapeutic agent in the treatment of inflammation-induced osteolysis. A prerequisite for such a treatment modality is high penetrability through tissues and epithelium. We therefore conducted a transepithelial penetrability assay in a rat model. In this assay, a bolus of radiolabeled SNV was delivered to the vagina and the local levels of SNV showed a decline over time ( Figure 5A). In parallel, the levels of SNV in the blood, heart, lungs, liver, gut, uterus and kidneys steadily increased ( Figure 5B), indicating the high penetrability of SNV through the vaginal epithelium, and further tissue stability. The integrity of the radioactive SNV was previously shown (Gozes et al., 1996).

SNV Is Nontoxic
We also conducted several toxicology studies in rabbits, guinea pigs and rats as detailed in the method section, and all showed a clean toxicology profile ( Table 1). The study named 004-1, was the most protracted and detailed toxicology study. We carried it out in rats as a repeated dose toxicity study on target external organs, the penis and the vagina. The structure of the study included a detailed pathological evaluation of all body organs.
No drug-related macroscopic or microscopic changes were detected in any of the dosages used. All changes reported by the pathologist were regarded as spontaneous incidental pathological findings similar to those not infrequently found in rats of the same strain and age used as controls.

SNV Protects Again TiP-Induced Osteolysis
Next, we conducted a TiP-induced osteolysis assay using our mouse calvarial model (Eger et al., 2017;Eger et al., 2018). The topical effect of SNV was tested by incorporating SNV into membranes that were also loaded with Ti particles. We assumed that SNV will penetrate the tissues and suppress the inflammation and osteoclastogenesis on the calvarial surface. In this experiment, we observed a severe osteolysis on the calvarial surface, beneath the TiP-loaded membrane ( Figure 6). Topically administered SNV significantly suppressed the TiPinduced osteolysis leading to a 60% significantly lower pit resorption volume and pit resorption volume as percentage of the total tissue volume ( Figures 6A,B). There were no significant differences in the calvaria thickness ( Figure 6C), supporting the idea that SNV suppressed pit resorption rather than stimulated irregular bone apposition.

DISCUSSION
VIP and its derivative SNV have been previously studied for immunomodulation properties (Abad et al., 2006;Goursaud et al., 2015). Here we show the effect of each peptide on the cytokine profile of macrophages in response to LPS and in the osteolytic response to Ti particles using in vitro and in vivo mouse models. Our data indicated that under the currently described experimental conditions, SNV had a stronger anti-inflammatory and anti-osteoclastogenic effect than VIP in vitro, in the presence of LPS and TiP, respectively. Furthermore, loading a single dose of SNV into a degradable membrane significantly inhibited bone resorption in a mouse calvarial model of inflammation-induced osteolysis. Lastly, this study also demonstrated the safety profile and high transepithelial penetrability of SNV. The presented data also shows the bioavailability of radioactive SNV ( Figure 5), in line with the previously published half-life of SNV (Gozes et al., 1996) and rapid clearance of intravenously administered SNV (10% remaining intact SNV, 30 min after injection (Gozes et al., 1994). Although SNV is derived from VIP, differences were reported in terms of potential mechanisms (cAMP vs. cGMP, see above), and biological potency in models such as neuroprotection (Gozes et al., 1995;Delgado et al., 1998;Ashur-Fabian et al., 1999;Kim , 2000). These and other previous studies showed that VIP downregulates TNFα expression in LPS-treated RAW264.7 cells and activated microglia, and in vivo in models of nerve injury (Kim et al., 2000) and endotoxemia (Delgado et al., 1999). In contrast, in our study VIP had no significant effect on TNFα expression in LPS-treated primary BMDM (Figure 2). This apparent contradiction may stem from the origin of the macrophages (BMDM vs. RAW264.7 and rat microglia) or from indirect mechanisms that only take place in vivo.
Importantly, previous studies showed that VIP may present a cAMP-dependent pro-inflammatory role (Vu et al., 2014) and induce IL6 expression (Mullol et al., 1997;Martinez et al., 1998;Brenneman et al., 2003) while SNV does not activate cAMP formation (Goursaud et al., 2015). Here also, we found that VIP increased IL6 expression in LPS-activated macrophages, whereas SNV did not (Figure 2), supporting the notion of a distinct mechanism between VIP and SNV. The superior anti-osteoclastogenic effect of SNV over VIP may therefore result from either its inhibitory effect on IL1β with no stimulation of IL6, a separate internal signaling pathway or its longer half-life (Gozes et al., 1996). Another advantage of SNV is that its chemical properties allowed its incorporation in the thrombin-fibrinogen membrane whereas VIP and its solvent perturbed the formation of the membrane (data not shown). Notably, we cannot rule out that more stable delivery methods of VIP would improve its bioactivity as reported by others (Wang et al., 2016).
Previous studies showed controversial data regarding the effect of VIP on osteoclastogenesis. VIP was shown to stimulate bone resorption in an ex vivo organ culture (Hohmann et al., 1983) but had no effect on basal osteoclastogenesis and even inhibited osteoclast formation in mouse bone marrow cultures . These opposing findings may be rooted in the fact that VIP has a pro-osteoclastogenic effect mediated by its actions on osteoblasts but also an antiosteoclastogenic effect via its direct action on osteoclasts . Our data further support the conclusion of such a direct inhibitory effect of VIP on preosteoclasts. In our inflammatory models in vitro, the observed impact of VIP on osteoclast differentiation is likely a direct antiosteoclastogenic effect that slightly supersedes the proosteoclastogenic increase in IL6. The therapeutic potential of VIP has been tested in a rat model of periodontitis and displayed a partial effect on the inflammatory status and osteoclastogenic signals but no significant positive outcome on bone loss (Gürkan et al., 2009). In contrast, we show here that SNV significantly prevented the bone loss induced by the presence of TiP. Our findings suggest that while SNV and VIP have similar effects on osteoclast differentiation (Figure 3), SNV exerts a more potent anti-osteoclastogenic effect in the context of inflammation ( Figure 4). One of our study limitations is that we did not directly demonstrate the effect of SNV in vivo on the inflammatory response and in a model of bacteria-induced osteolysis. It should be emphasized that a corroborative independent study demonstrated the significant anti-inflammatory effects of SNV in vivo in an amyotrophic lateral sclerosis (ALS) model (Goursaud et al., 2015). In our study, the strong suppression of osteoclast differentiation (in vitro) and bone resorption (in vivo) by SNV might result from the combined effect on the inflammation (decrease in IL1β, Figure 2 and (Eger et al., 2018)) as well as a direct effect on osteoclasts (Figure 3). Notably, our osteoclastogenic assay performed in the presence of conditioned medium from TiP-exposed macrophages (Figure 4), contained a host of pro-inflammatory and proosteoclastogenic signals, and SNV blocked or supplanted these signals, pointing to a strong inhibitory effect on inflammationinduced osteoclast differentiation.
Both SNV and VIP bind at VPAC-1 and VPAC-2 (Gourlet et al., 1998) and macrophages from humans and mice express both receptors (Calvo et al., 1994;Herrera et al., 2009;Burian FIGURE 5 | SNV demonstrated a high transepithelial penetrability and stability using a model of vaginal delivery in rats. 7.0 microCi/rat (15 million CPM, 100 Ci/mmole specific activity) of radioactive SNV was delivered in the vagina of rats at T0. Each point on the chart represents at least two rats. Levels of radioactivity were monitored over 250 min to assess the remaining amount of SNV in the vagina (A) versus its distribution in the indicated organs (B) as a measure of its penetrability through the vaginal epithelium. Data are Mean ± SD.
Frontiers in Pharmacology | www.frontiersin.org May 2021 | Volume 12 | Article 638128 9 et al., 2010;Hauk et al., 2014). Others showed that VPAC-2 expression increases in LPS-exposed macrophages (Herrera et al., 2009) and we showed that VPAC-2 expression declines during osteoclast differentiation (Figure 3), in line with previous studies showing that osteoclasts mainly express VPAC-1 . On the other hand, our previous work  in RAW264.7 murine macrophages showed that VIP inhibited LPS-driven TNFα production via VPAC-1. Notably, SNV exhibits a differential affinity to human and rodent VPAC-1 and VPAC-2 (Gourlet et al., 1998), suggesting that the ratio of expression of the two receptors in rodent and human cells will dictate the biological outcome and mechanism of action of VIP and SNV in inflammation and osteoclastogenesis in humans. A very recent study showed that VIP further modulates human macrophage phenotype via formyl peptide receptor-like 1 (FPRL1) and activation of RhoA-GTPase and PLC pathways (Harhous et al., 2021), while both VIP and SNV act through cGMP to increase neuroprotection . These findings pave the path for future mechanistic evaluations.
As discussed above, inflammation-induced osteolysis is the common pathological outcome of various conditions including periodontitis, oral peri-implantitis and orthopedic implant loosening. In this study we showed using distinct in vitro assays that SNV blocked the two common denominators leading to osteolysis, i.e. inflammation and osteoclastogenesis, notwithstanding the cause of the inflammation, i.e. bacterial (LPS) or aseptic (TiP), clearly demonstrating a greater therapeutic potential than VIP. In both oral periimplantitis and orthopedic periprosthetic osteolysis, macrophages contribute to the inflammatory response and secretion of pro-osteoclastogenic signals (Revell, 2008;Fretwurst et al., 2020). In future studies, it would be interesting to test the therapeutic effect of SNV in gingival fibroblasts and periodontal ligament cells that also fuel the inflammation and bone loss around oral implants. Obviously, our observations obtained using in vitro and in vivo models also warrant further investigation using various bacterial strains known for their deleterious impact on bone, at different anatomical sites (e.g. jaw, hip), as well as in human cells and validation in placebo controlled double-blind randomized clinical trials.
This study attributes for the first time a therapeutic potential to SNV in the treatment of inflammation-induced osteolysis. Its high penetrability and low systemic toxicity portray SNV as a valid candidate for topical administration. Our previous studies showed a close match between the inflammatory response induced by Ti particles and that induced by LPS, suggesting that SNV presents a promising therapeutic approach in the management of peri-prosthetic and periodontal bone loss induced by wear particles and bacterial infection.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The animal study was reviewed and approved by the Institutional Animal Care and Use Committee (permit number M-015-047).

AUTHOR CONTRIBUTIONS
Conceptualization, ME, DK, IG and YG; methodology, ME, TL, SH-B, ZA, EG, DD, IG and YG; software, YG and SH-B; validation, Frontiers in Pharmacology | www.frontiersin.org May 2021 | Volume 12 | Article 638128 ME, IG and YG; formal analysis, ME, TL, SH-B, ZA, EG, and DD; investigation, ME, TL, SH-B, ZA and DD; resources, IG and YG; data curation, ME, TL, SH-B, ZA, DD, DK, IG and YG; writing-original draft preparation, ME and YG; writing-review and editing, IG and YG; project administration, IG and YG; funding acquisition, IG and YG. All authors have read and agreed to the published version of the manuscript.