Substance P blocks ovariectomy-induced bone loss by modulating inflammation and potentiating stem cell function

OVX induces bone loss and enriches the Th17 cell pool

OVX induces estrogen deficiency and is a common method to generate an osteoporosis animal model. OVX-induced estrogen deficiency is known to cause systemic inflammation and elevate TNF-α and IL-17 levels to activate osteoclasts and inhibit repopulation and survival of osteoblasts. The consequences include bone loss and a high risk of fracture. To examine the mechanism underlying bone loss following OVX, bone density and Treg/Th17 cell pool were monitored for 16 weeks.

Bone loss and inflammation were rarely observed at 4 weeks after surgery in OVX rats (Supplementary Figures 1 and 2). Bone density started to decline from 8 weeks post-surgery, and severe deficiency in bone density was maintained until 16 weeks, as shown in the micro computed tomography (CT) data (Figure 1A–1C). Histological analysis confirmed bone loss and accumulation of the fat (Figure 1D). In comparison with the femur of sham control rats, that of OVX rats showed high levels of the adipose tissue and a low proportion of bone collagen tissue. This observation is indicative of bone loss evident at 8 weeks post-surgery. Analysis of biochemical markers of bone loss showed that the expression of osteocalcin (OCN) and osteoprotegerin (OPG) decreased and that of collagen X (CTX) increased from 8 weeks (Figure 1E, 1F). An increase in cholesterol and triglyceride (TG) levels was observed at 12 and 16 weeks post-OVX, respectively.

Estrogen deficiency promotes Th17-skewed differentiation of naïve T cells and provides IL-17–enriched conditions in peripheral tissue and circulation. Th17 cells are key contributors of bone erosion [47] and directly produce receptor activator of nuclear factor-κB ligand (RANKL) to increase osteoclast formation [48, 49]. IL-17 also enhances the expression of RANKL in osteoblasts [50, 51] and mediates osteoclast-induced secretion of TNF-α and IL-6 in the bone tissue. Therefore, monitoring IL-17 and Th17 cells during disease progression may be important to confirm bone loss.

We determined the levels of Th17 and Treg in second lymphoid organs, the spleen and mesenteric lymph node (MLN), by fluorescence-activated cell sorting (FACS). In the spleen, the population of CD4⁺Foxp3⁺ Treg cells slightly decreased and that of CD4⁺IL-17⁺ Th17 cells increased from 8 weeks after surgery. This trend was strongly evident at 12 weeks post-OVX (Figure 2A–2D, Supplementary Figures 3, 4). The Th17-biased development was distinctly detected in the MLN. Moreover, the increase in the population of Th17 cells was consistent with the elevated levels of IL-17 in the circulation (Figure 2E). IL-10 level was reduced from 8 weeks post OVX (Figure 2F).

Taken together, OVX causes bone loss, accompanied with a Th17-enriched and Treg-deficient environment in lymphoid organs. This phenomenon was likely to be initiated at 8 weeks post-surgery.

OVX-induced vascular impairment precedes bone loss

Osteoporosis is closely related to the impairment of vasculature [30]. A recent study suggested that the dysfunction of the vasculature proceeds via bone loss [30–33], and the type H vessel in the BM supports osteogenesis by protecting osteoprogenitors [30]. Xie et al. found that angiogenic factors such as PDGF-BB determine the temporal-spatial vessel formation for new bone development [32]. That is, the intact vasculature and its paracrine potential heavily influence the maintenance of the bone tissue.

In the present study, we found that the decrease in the bone density was associated with the decrease in the density of the CD31⁺ type H vessel in the trabecular bone (Figure 3A). Analysis of angiogenic factors in the serum revealed a sharp decline in PDGF-BB, VEGF, and TGF-β levels after OVX in a time-dependent manner (Figure 3B–3D). This observation was consistent with that previously report to prove the relationship between osteoporosis and deficiency of vessels [31].

To analyze the progression of vascular damage, we examined density of the CD31⁺ type H vasculature at 4 weeks post-OVX and found that it was seldom observed in the trabecular bone even at 4 weeks post-OVX, the time corresponding to that without bone loss and systemic inflammation (Supplementary Figures 1 and 2). Moreover, serum VEGF and PDGF-BB levels already dropped at 4 weeks in OVX rats as compared with those observed in control rats (Supplementary Figure 5).

Together these observations corroborate that the deficiency of estrogen after OVX deteriorates the vasculature and its paracrine potential, consequently leading to the loss of bone density.

SP prevents bone loss by regulating angiogenic and inflammatory responses

We have attempted to explore the therapeutic effect of SP on osteoporosis and identified the starting point of bone loss in OVX rats. We administered rats with SP at 8 weeks after surgery, which is responsible for the impairment of both the bone density and vasculature. The efficacy of SP was analyzed at 12 weeks post-surgery (Figure 4A).

Cyclic injection of SP prevented the reduction in bone density, as observed by micro-CT (Figure 4B–4D). Histological analysis revealed the ability of SP injection to decrease the accumulation of the fat tissue and reduce the population of TRAP⁺ activated osteoclasts in the femur (Figure 4E). Serum OCN and TGF-β levels reduced after OVX but these levels recovered after SP treatment (Figure 4F and 4G). SP effect was abrogated by NK-1R antagonist (Supplementary Figure 6).

The vasculature was found to be extremely impaired at 8 weeks post-OVX (Figure 4H), and SP injection could repair the vessel density in the BM to some extent. Serum VEGF and PDGF-BB levels reliably increased after SP treatment as compared with those observed after saline treatment (Figure 4I and 4J).

SP treatment obviously decreased the pool of Th17 cells and enriched Treg cells in both the spleen and MLNs at 12 weeks post-surgery (Figure 5A–5D). While OVX increased TNF-α/IL-17 levels and reduced IL-10 level in the serum, SP treatment reversed the OVX-induced inflammation by decreasing TNF-α/IL-17 and elevating IL-10 levels (Figure 5E–5G).

Thus, OVX develops a pro-inflammatory environment by modulating the profile of immune cells, accompanied by vascular damage. This condition eventually provokes the activation of osteoclasts and loss of bone. However, SP injection could promote Treg generation, rather than Th17, and reduce vascular damage under OVX-induced pathological conditions. These functions of SP may alleviate the reduction in bond density to maintain the intactness of the bone tissue.

SP is indirectly involved in the development of Treg cells

SP treatment prevented loss of bone density, reduced inflammation, and exhibited high angiogenic potential in vivo. SP was previously reported to induce immune suppression by controlling the profile of immune cells [35–37, 43].

To clarify the effect of SP on Treg development under osteoporosis, in vitro co-cultures of peripheral blood mononuclear cells (PBMCs) and spleen and MLN cells from sham or OVX rats were performed (Figure 6A). To mimic the in vivo systemic SP treatment, PBMCs were treated with SP and provide indirect effects of SP on the spleen or MLN cells through blood.

After 4.5 and 24 h, the ratio of Treg and Th17 cells in the spleen/MLN cells was analyzed by FACS. Osteoporosis condition maintained lower number of CD4⁺Foxp⁺ Treg but higher proportion of CD4⁺IL-17⁺ Th17 cells as compared to sham treatment (Figure 6B–6E). However, the level of Treg greatly increased and Th17 pool decreased after SP treatment; this effect was maintained for 24 h. Thus, SP-stimulated PBMCs may be involved in Treg development in lymphoid organs, thereby producing an immunosuppressive environment in osteoporosis (Figure 6B and 6C).

Differentiation of naïve T cells to Tregs or Th17 occurs in an exclusive way. Naïve T cells can differentiate into Tregs or Th17 cells, and the factors that drive differentiation of native T cells into Th17 cells, including IL-6, TGF-β, and IL-21, are well known. The analysis of the conditioned medium of PBMCs showed that the PBMCs from OVX rats had higher levels of IL-6, IL-21, and TGF-β than those from sham rats, indicative of a favorable environment for the generation of Th17. However, SP treatment decreased IL-6 levels (Figure 6F and 6G) and rarely affected TGF-β and IL-21 levels (Supplementary Figure 7).

Thus, this data implies that SP could facilitate the differentiation of naïve T cells into Tregs, but not Th17 cells, by decreasing IL-6 levels in the circulation under osteoporotic stress.

SP can restore the osteoporosis-mediated impaired activity of stem cells

In addition to its anti-inflammatory function, SP is capable of restoring the impaired activity of BMSCs mediated by inflammation, oxidative stress, and senescence [40, 41]. We investigated whether SP can recover the cellular activity of BMSCs from OVX. BMSCs from OVX 8W were cultured and primed with SP once a day for 2 days in vitro. The effects of SP on the viability, paracrine function, and differentiation potential of BMSCs were evaluated (Figure 7A). Cell viability and repopulation rate were clearly impaired after osteoporosis but could be restored by SP treatment (Figure 7A, 7B). The protein kinase B (AkT) and extracellular signal-regulated kinase (Erk) signaling pathway is highly associated with cell proliferation and survival. As predicted, active Akt and Erk levels reduced after osteoporosis, but this signaling pathway was activated by SP within 10 min (Figure 7D–7F), which probably contributes the improvement in the viability and repopulation of BMSCs.

OVX conditions impair the differentiation ability of BMSCs into osteoblasts. However, SP-primed MSCs showed better osteogenic differentiation and calcium deposition (Figure 7E). TGF-β, OPG, and hepatocyte growth factor (HGF) required for bone formation were elevated after SP treatment (Figure 7F).

SP treatment restored the BM vasculature in OVX rats. BMSCs constitutively secrete angiogenic factors such as VEGF and TGF-β, contributing to the maintenance and homeostasis of the vasculature in the BM microenvironment. To determine the supportive role of BMSCs in vessel formation, endothelial progenitor cells (EPCs) from OVX rats were co-cultured with BMSCs in Matrigel in vitro (Figure 8A).

As predicted, the EPCs from sham (sham EPC) showed a tubular structure, whereas those from OVX (OVX EPC) failed to develop any vascular structure. However, the coexistence of sham BMSCs enabled the formation of a continuous vasculature in OVX EPCs. To mimic in vivo conditions, OVX BMSCs were primed with SP (OVX-BMSC^SP) and then incubated with OVX EPCs in Matrigel. OVX BMSCs rarely provided a supportive role in OVX EPC–induced vessel formation. In contrast, OVX-BMSC^SP clearly facilitated OVX EPC–induced vessel formation (Figure 8B–8E).

The formation of vasculature is governed by the profile of paracrine factors; thus, SP was thought to promote the production of angiogenic factors in OVX BMSCs. The effect of SP on the secretion of cytokines by OVX BMSCs was assessed through enzyme-linked immunosorbent assay (ELISA). SP restored the levels of VEGF and TGF-β, which were reduced by OVX (Figure 8F and 8G). Thus, SP treatment may enhance the vascular activity of BMSCs possibly by modulating their paracrine action.

Collectively, SP could restore the cellular functions of BMSCs injured by osteoporosis and thus, maintain the pool of BMSCs and osteogenic progenitor cells with vascular potential. These effects may negatively affect the progression of bone loss in vivo.

Efficacy of SP is superior to that of parathyroid hormone (PTH) therapy

Conventional therapy for osteoporosis includes the administration of calcium, bisphosphonate, or PTH. Of these, PTH is used as a bone-forming agent that necessitates repeated injections once a day for several months. However, PTH exerts opposite effects depending on dosage and is ineffective on the old population including daily alcohol consumption or bisphosphonate taker; thus, it is important to control PTH dosing by limiting injection period or determining the patient population. Despite the associated risks, PTH is the most widely used drug for osteoporosis treatment.

We compared the effect of SP and PTH therapy in OVX rats. According to the application route in the clinic, PTH was subcutaneously injected once a day for 5 days and SP was intravenously injected twice a week from 8 weeks post-OVX.

The analysis of bone density and histology showed that SP injection inhibited bone loss, consistent with previous data. Although PTH tended to exert inhibitory effects on bone loss, these effects were small and showed variations within experimental animals (Figure 9A–9E).

Serum cytokine analysis revealed that SP could reduce TNF-α and restore IL-10, whereas PTH did not suppress inflammatory factors as compared with saline treatment. Angiogenic factors, VEGF and PDGF-BB were declined by OVX, which were restored by SP or PTH treatment. PTH therapy was involved in angiogenic potential, rather than systemic inflammation.

These observations suggest that PTH treatment could prevent bone loss, but its effects were ambiguous. SP injection could inhibit bone loss and facilitate anti-inflammatory and angiogenic conditions. We injected SP twice a week, while PTH treatment was performed for 5 days a week. Considering the treatment schedule and the reported therapeutic effects, SP is expected to have superior performance to PTH therapy.

Induction of osteoporosis animal model

Seven-week-old Sprague Dawley (SD) rats (160-170 g, female) were purchased from DBL (Daehan Bio Link, Seoul, Korea). All animals were maintained under a regular light/dark illumination cycle in an animal holding room and permitted to adapt new environment for 7 days before the experiments. All animals received standard chow diet, and this study was approved by the Ethical Committees for Experimental Animals of Kyung Hee University Hospital with the approval number of KHMC-IACUC 2018-40.

To induce an osteoporosis disease model, female SD rats were anesthetized during the operation by using intraperitoneal injections of Ketamine (100 mg/kg, Yuhan, Seoul, Korea) and Rompun (1.2 mg/kg, Bayer Healthcare, Kyunggi-do, Korea). About 2 cm-long skin incisions was induced both sides of abdomen to expose the dorsolateral abdominal muscles and reached to the peritoneal cavity, then the ovaries were found below the kidney. After identifying the ovary, both the terminologia anatomica and ovarian artery were carefully uncovered and ligated with 5-0 silk and subsequently the ovaries were removed. After hemostasis process, both two sides of abdominal incisions were sutured with 3-0 silk. All of the surgical procedures were performed under sterilized condition. All rats were randomly divided into five groups: (1) Sham, (2) OVX + saline, (3) OVX + SP, (4) OVX + RP67580 (NK-1R antagonist) + SP and (5) OVX + PTH.

Administration of SP, RP67580 and hPTH (1-34)

SP (Sigma-Aldrich, St. Louis, MO, USA) was diluted in saline (JW Pharmaceuticals, Seoul, Korea) just before use and administrated intravenously twice a week at a dose of 5 nmol/kg (n = 12/group). Saline was used for the vehicle treatment (n = 12/group). RP67580 (TOCRIS, Bristol, UK, NK-1R antagonist) was diluted in saline and administrated intravenously just 5 minutes before SP injection at a dose of 10 nmol/kg (n = 6). Human parathyroid hormone-(1-34) (h-PTH (1-34) (Mimotopes, Melbourne, Australia), was also diluted in saline and administrated subcutaneously five times a week at a dose of 80 μg/kg

Ex-vivo micro-computed tomography (μ-CT) imaging

Femur was scanned on a μ-CT scanner (SkyScan 1173 X-ray microtomography; Bruker) at a resolution of 13.85 μm to evaluate the degree of osteoporosis. After three-dimensional reconstruction, the bone mineral density (BMD), the bone volume fraction (BV/TV) and the trabecular bone space (Tb.sp) in the metaphysis area were used to calculate using CT analysis software. (n = 12/group).

Enzyme-linked immunosorbent assay (ELISA)

Cytokine profiles in serum, bone marrow aspirates and conditioned medium were analyzed by ELISA. The concentrations of TNF-α (Biolegend, San Diego, CA, USA), IL-6, IL-10 (R&D system), IL-21 (LSBio, Seattle, WA, USA), PDGF-BB, VEGF, TGF-β (R&D system, #MB100B), Glu-osteocalcin (OCN) (Takara, Shiga, Japan), C-telopeptide of type 1 collagen (CTX) and osteoprotegerin (OPG) Cusabio) in serum or conditioned medium samples were measured by ELISA according to the manufacturer’s instructions. In brief, all reagents, standard dilutions, and samples were prepared and added to primary antibody-coated well as directed for 2h at room temperature (RT). Next, the horseradish peroixdase-conjuated secondary antibody solution was treated to each well. 1 or 2 hours later, substrate solution was added. Once the color of the solution changed to blue, the reaction was stopped, and the optical density was measured with the wavelength correction set to 450 nm using an EMax Endpoint ELISA Microplate Reader (Molecular Devices, Sunnyvale, CA).

Histological analysis

Femur was isolated and fixed in 3.7 % paraformaldehyde (Sigma-Aldrich) for 1 day. Samples were decalcified in decalcifying solution (Sigma-Aldrich) and processed with a TP1020 tissue processor (Leica Biosystems, Wetzlar, Germany) to make paraffin blocks, and 5.0-μm-thick sections were prepared.

For hematoxylin and eosin (H&E) staining, paraffin-sectioned samples were hydrated in alcohol (Merck, Germany). After washing in tap water, nuclei were stained with hematoxylin (Sigma-Aldrich) for 10 minutes and washed in tap water. To stain the cytoplasm, eosin Y (Sigma-Aldrich) was applied for 10 seconds and the samples were washed with tap water and then mounted.

Trichrome staining was performed using the NovaUltraTM Masson Trichrome Stain Kit (IHC World, Woodstock, MD). Staining for tartrate-resistant acid phosphatase (TRAP) was carried out by TRAP staining kit (Sigma Aldrich, Saint Louis, MI, USA).

For immunofluorescence staining, rehydrated samples were boiled with 0.01 M sodium citrate for antigen unmasking and then treated with 0.1 % sodium borohydride (Sigma-Aldrich) to quench activity of endogenous fluorescence and permeabilized with 0.3 % Triton-X 100. Nonspecific binding was blocked by incubating the samples with 2 % normal horse serum for 1 hour at RT. Primary antibodies against CD31 (Abcam) were treated overnight. After washing with phosphate-buffered saline for three times, the samples were incubated with a Cy3-conjugated secondary antibody (Jackson ImmunoResearch Laboratories) for 1 hour at RT. After washing, the samples mounted with VectaMount reagent (Vector Laboratories).

Isolation of splenocytes and MLN cells

The spleen and MLN of SD rat was aseptically isolated and, trimmed of all connective tissues. Single cells of spleen and MLN were obtained by mechanical force with syringe plunger. The cell suspension was centrifuged at 1500 rpm for 5 min and washed in 1 x PBS (Welgene, Daegu, Korea) containing 1 % P/S (Welgene).

Fluorescence-activated cell sorting (FACS) analysis

To detect CD4⁺Foxp3⁺ Treg cells in the spleen and MLN, 2 × 10⁷ of splenocytes and MLN cells were pretreated by True-Nuclear Transcription Factor Buffer Set (Biolegend, San Diego, CA, USA) and then incubated with allophycocyanin (APC)-conjugated CD4 antibody (Biolegend) and PE-conjugated anti-Foxp3 antibody (Biolegend).

For CD4⁺IL17⁺ Th17 cells in the spleen and MLN, 2 × 10⁷ of splenocytes and MLN cells were incubated with allophycocyanin (APC)-conjugated CD4 antibody (Biolegend) and phycoerythrin (PE)-conjugated anti-CD17 antibody (eBioscience, San Diego, CA, USA).

The fractions of Treg, and Th17 cells from each tissue were analyzed using a FACSCalibur Flow Cytometer using the CellQuest software (Becton Dickinson, San Jose, CA).

Co-culture of PBMCs with splenocytes or MLN cells

For indirect co-culture of peripheral blood mononuclear cells (PBMCs) and splenocytes or MLN cells, splenocytes or MLN cells (2 × 10⁶) were seeded on upper transwell insert (0.4 μm pore size; Corning, NY, USA) and PBMCs (2 × 10⁶) were seeded on 6 well plate (Corning). SP was added to PBMCs and coculture was maintained for 24h. After termination of coculture, the ratios of both Treg cells and Th17 cells in splenocytes and MLN cells were analyzed by FACS analysis and IL-6, IL-21 and TGF-β1 expression level of PBMCs in each culture supernatant was measured by ELISA.

WST-1 assay

10 μl of WST-1 (Roche, Indianapolis, IN, USA), solution was added to each well at 10 % the total volume of the medium, and the plate was incubated for 1 hour at 37 °C in 5 % CO₂. After incubation, the optical density values for each well were measured at a wavelength of 450 nm using an EMax Endpoint ELISA Microplate Reader.

Osteogenic differentiation of rat BMSCs

Rat BMSCs were seeded at a density of 5 × 10³ cells/cm², cultured overnight in mesenchymal stem cell growth basal medium (MSCGM; Lonza, Basel, Switzerland), and then induced to undergo osteogenesis for up to 15 days in rat osteoblast differentiation medium (Amsbio, Cambridge, MA, USA). Matrix mineralization by BMSC-derived osteoblasts was determined by using Alizarin Red S (Sigma-Aldrich), and staining quantified by measuring the optical densities of extracted stain at 560 nm.

Tube formation assay

Matrigel (BD, San Jose, CA, USA) was coated on μ-Slide Angiogenesis (Ibidi, Munich, Germany) and incubated for 30min at 37 °C in 5 % CO₂. EPCs were incubated on the Matrigel with or without BMSCs for 5 hours at 37 °C in 5 % CO₂. The tube structure was observed under phase-contrast microscopy. And the quantification of tube formation was analyzed by Image J.

Preparation of protein extracts and western blot analysis

Protein was extracted by cell lysis buffer (20 mM Tris-HCl (pH=7.5), 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1 % Triton, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na₃VO₄ (Cell Signaling Technology, Danvers, MA, USA) and 1 μg/ml leupeptin (Cell Signaling Technology) containing phenylmethylsulfonyl fluoride (Sigma-Aldrich). The supernatants were collected by centrifugation at 12000 rpm for 10 min. The protein concentration was determined using a BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA).

The proteins were denatured and electrophoresed using sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. Blocking of the membrane was performed with 5 % skim milk (Bio-Rad) or bovine serum albumin (Sigma-Aldrich) for 1 h. After blocking, the membranes were incubated with the primary antibodies for p-Akt (Cell Signaling Technology), t-Akt (Cell Signaling Technology), p-Erk (Cell Signaling Technology), t-Erk (Cell Signaling Technology), TGF-β1 (Abcam, Cambridge, MA, USA), OPG (Abcam), HGF (Abcam), NK-1R (Novus Biologicals, Littleton, CO, USA) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Abcam), followed by anti-immunoglobulin G horseradish peroxidase-conjugated secondary antibody and visualized with a chemiluminescence substrate (Dogen). Expression level was quantified using the Image J program.

Statistics

All Data are presented as the mean standard deviation (SD) of three independent experiments. P values of less than 0.05 were considered statistically significant. Statistical analysis of all data was carried out by an unpaired, two-tailed Student t-test.