Polydatin administration attenuates the severe sublesional bone loss in mice with chronic spinal cord injury

Animal care and study design

Male C57BL/6J mice (weighing 20-25 g) were purchased from the Laboratory Animal Center of Guangzhou University of Chinese Medicine, and housed in a humidity- and temperature-controlled environment with a 12-hour light-dark cycle (8 A.M. -8 P.M.) with free access to tap water and standard pellet diet. After one week of acclimatization, the mice were randomized into the placebo (saline) or PLD-treated sham-operated and SCI groups (n = 25-27 each). At the end of the experiment, blood samples and bilateral femora and tibiae devoid of soft tissue were collected for subsequent analysis. All experimental procedures conformed to the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and USE Committee of Guangzhou University of Chinese Medicine (No. 44005800013161).

Surgical procedures and treatment protocol

Severe thoracic SCI was induced under deep anesthesia (pentobarbital sodium, 50 mg/kg body weight, intraperitoneal injection) according to our previously published protocol [22]. Briefly, the spinal cord with dura mater was exposed by laminectomy at the T_8-9 level under sterile conditions, and a controlled impact was made at the velocity of 0.6 m/s and depth of 0.6 mm for 90 ms. Bladder was cleared manually twice daily after the operation until spontaneous urination was restored. The sham-operated mice underwent laminectomy with no contusion injury. Daily intramuscular penicillin injections were given for the first 5 days after surgery to prevent wound infection. Eight weeks following injury, the mice were given 40 mg/kg PLD (PubChem CID: 71311798, suspension in 0.5% CMC-Na) or 0.9% saline daily by oral gavage for another consecutive 8 weeks. The duration of treatment was determined by the fact that chronic SCI-induced bone loss is considerably more severe compared to osteoporosis caused by aging, low calcium diet, menopause and ovariectomy.

Measurement of calcium (Ca²⁺), osteocalcin (OCN) and C-terminal telopeptide of type I collagen (CTX-1) levels

Blood samples collected at termination were centrifuged at 2000 rpm for 10 min at 4° C, and the serum was separated and then kept frozen at -80° C until analysis. Serum concentration of Ca²⁺ was measured spectrophotometrically using an autoanalyzer (Hitachi, Japan). Serum concentration of OCN (a specific product of the osteoblast) was measured using a mouse OCN immunoassay kit (Jiancheng Bioengineering Institute, China), and serum CTX-1 level was evaluated using a RatLaps™ ELISA kit from BioVendor (Czech Republic), by following the manufacturer’s directions.

Measurement of urinary deoxypyridinoline (DPD) levels

Urinary DPD (the breakdown product of collagen during bone resorption), which reflects systemic bone resorption, is considered useful for assessing the effects of osteoporosis treatment. DPD excretion was quantified by using an ELISA kit (Xinqidi Biological Technology, China), and the results were corrected for the urinary creatinine concentration, which were determined with a commercially available kit (Western Bio-Tech, China).

Dual-energy X-ray absorptiometry (DEXA)

The femurs and tibiae were removed from each animal and soft tissues (as cartilage, tendon and ligament) were cleaned. Areal BMD were measured by DEXA (Hologic, USA) using the small-animal program. Samples were placed on an acrylic platform of uniform thickness to obtain the BMD of distal femurs, proximal tibiae and spine (L3-L5). The coefficient of variation was 1% for BMD.

Micro-CT (μCT) analysis and mechanical testing

The morphometry of the femoral samples was measured by high-resolution μCT (Bruker, Belgium). The scanner was set using the following parameters: 59 kVp energy, 100-μA intensity, and a resolution of 9μm per pixel. Images reconstruction and three-dimensional (3D) quantitative analysis were performed with the software provided by the manufacturer. The regions of interest (ROI) of femur were defined using CTAn v1.9 (Bruker) as shown in Figures 2A, 2B, 3A. The trabecular microstructural indices were calculated, including trabecular bone volume per total bone volume (BV/TV), trabecular number (Tb.N, mm^-1), trabecular thickness (Tb.Th, mm), trabecular separation (Tb.Sp, mm), connectivity density (Conn.D, mm^-3), and structure model index (SMI, range from 0 to 3, with 0 = plate like and 3 = rod like). To analyze cortical bone, total bone and medullary tissue area (mm²), the periosteal and endosteal perimeters (mm), cortical bone volume versus total bone volume (Ct.BV/TV) and cortical thickness (Ct.Th, mm) were measured. To assess mechanical strength, the femurs were subjected to a three-point bending test at the constant velocity of 3 mm/min until complete fracture using an ElectroForce biological material testing system (TA Instruments, USA) as described by Nakagaki et al. [23].

Figure 2. Effects of PLD on trabecular bone structure of the distal femur. (A) The schematic diagram of the coronal section of femur and representative 3D reconstructed coronal images of cancellous bone in the distal femur of each group. (B) The schematic diagram of the ROI of distal femur and the representative μCT 3D-images of cancellous bone within the ROI. Plots of the structural parameters: (C) BV/TV, (D) Tb.N, (E) Tb.Th, (F) Tb.Sp, (G) Conn.D, and (H) SMI. Data are expressed as mean ± S.D.; n=4 to 5 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.

Figure 3. Effects of PLD on the cortical bone structure of femur midshaft. (A) The schematic diagram of the ROI of femoral midshaft and typical μCT images of cortical microarchitecture are displayed. Measurements are shown for (B) total, bone, and medullary tissue area, (C) periosteal and endosteal perimeters, (D) Ct.BV/TV and (E) Ct.Th. Data are expressed as mean ± S.D.; n=4 to 5 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.

Bone histology and immunohistochemistry (IHC) analysis

The distal end of the femurs was fixed in 10% neutral buffered formalin (NBF) for 24 h and then decalcified in 10% EDTA (pH 7.4). The samples were then dehydrated through an ethanol gradient, embedded in paraffin, and cut into 4-μm-thick sagittal-oriented sections using a ultramicrotome. The sections were stained with hematoxylin and eosin (H&E). The expression of osteoblastic markers was detected by IHC. After deparaffinization and rehydration, the tissue sections were immersed in 3% H₂O₂ to block endogenous peroxidase activity, blocked with normal goat serum, and incubated overnight with the primary antibodies against OCN (1:200; Abcam, USA) and OPN (1:100; Abcam, USA) at 4° C. The following day, the sections were washed and incubated with the secondary antibodies. After developing the stain with DAB solution and counterstaining with hematoxylin, the sections were observed under a microscope (Olympus, Japan).

Von Kossa staining and tartrate-resistant acid phosphatase (TRAP) staining

The EDTA-treated sections were further analyzed by von Kossa staining and TRAP staining. The slides were stained with the von Kossa dye and continuously irradiated under a UV lamp for 4 h. After rinsing thrice with distilled water, the sections were counterstained with H&E and observed by bright field microscopy. TRAP staining was performed using a commercially available kit (Servicebio, China). Briefly, the slices were soaked in TRAP staining solution at 37° C for 20 min. The specimens were washed in tap water, counterstained with hematoxylin, and mounted. The osteoid volume (OV) and osteoid surface (OS) were measured, and the number of osteoclasts were counted using ImageJ software (NIH, USA).

Measurement of malondialdehyde (MDA) levels

MDA levels were measured by reacting with thiobarbituric acid (TBA) using a commercially available kit (Jiancheng Bioengineering Institute, China). Briefly, blood samples and bone homogenates were centrifuged to collect the supernatants, which were further processed according to the manufacturer’s instructions. The absorbance was measured at 532 nm using a microplate reader. The content of MDA in the serum and femur were calculated as nM and μM/mg protein respectively.

Measurement of 25-hydroxyvitamin D [25(OH)D] levels

Serum level of vitamin D metabolite 25(OH)D was measured with a commercially available kit (R&D System, USA) following the manufacturer’s instruction manual. The intra-assay and inter-assay precision for the ELISA were less than 10% and 15% for 25(OH)D.

Western blotting

Frozen left distal femurs were pulverized in mortar and homogenized in ice-cold IP lysis buffer (Beyotime, China) supplemented with 1mM PMSF. Samples containing approximately 30 μg protein was electrophoretically separated by 12% SDS-PAGE and transferred to PVDF membranes (Millipore, USA). After blocking non-specific binding with 5% skimmed milk in TBST for 1h at room temperature, the membranes were incubated overnight with primary antibodies at 4° C. The membranes were then washed thrice with TBST (10 min each time), and incubated with the secondary antibodies for 2 h. After extensive washing, the bands were visualized using an ECL detection kit (Beyotime) and densitometrically quantified with Image Lab version 2.1 (Bio-Rad). GAPDH was used as an internal control.

Culture and differentiation of bone marrow progenitors

Procedures for study of osteoblast and osteoclast formation were followed the methods previously described [24]. Briefly, the bone marrow cells were extracted from the femurs and tibiae under sterile conditions. To initiate osteoblast differentiation, the BMSCs were cultured in α-MEM supplemented with 10% fetal bovine serum (Gibco, USA) and 1 mM ascorbic acid-2-phosphate, and the medium was replaced every two days. The formation of osteoblastic lineage cells was determined by staining the colonies with alkaline phosphatase (ALP) on day 10, and the number of ALP-positive colonies (colony-forming unit–fibroblastic, CFU-F) were counted. The cells producing mineralized bone matrix were stained with the von Kossa dye on day 28, the number of positive colonies (CFU–osteoblastic, CFU-OB) was counted. In addition, ALP activity assay and Alizarin red staining were performed on day 7 and day 14 of induction, respectively, as previously described [25]. The BMSCs were differentiated into osteoclasts by culturing in α-MEM supplemented with human macrophage colony-stimulating factor (M-CSF) for 2 days. Non-adherents were collected, purified by Ficoll-Plus, and incubated for another 5 days in α-MEM supplemented with M-CSF (30 ng/mL) and RANKL (60 ng/mL). Osteoclasts were identified by staining for TRAP. Total RNA was also extracted from the induced cells for RT-qPCR.

Quantitative real-time PCR analysis

The right distal femurs were snap frozen in liquid nitrogen and ground into a powder. Total RNA was extracted using Trizol reagent (Invitrogen, USA) according to the manufacturer’s guidelines, and 50 ng template was reverse-transcribed into cDNA using the PrimeScript™ RT Master Mix Kit (Takara, Japan). Quantitative real-time PCR was performed using a SYBR Green PCR Mix on the CFX96™ Real-Time PCR Detection System (Bio-Rad, USA) as previously described [22]. Each sample was run and analyzed in triplicate using the 2^-ΔΔCt method, and expressions of Runt-related transcription factor 2 (RUNX2), OCN, ALP, osteoprotegerin (OPG), Receptor-activator of Nuclear Factor κ-B Ligand (RANKL), calcitonin receptor (CTR), TRAP, sclerostin (SOST), wingless-related MMTV integration site 3A (Wnt3a), lipoprotein receptor-related protein 5 (Lrp5) and ctnnb1 were normalized to β-actin. The sequences of primers used for amplification of target genes were listed in Table 1.

Statistical analysis

All data were analyzed statistically using SPSS 24.0 software (SPSS Inc., USA), and expressed as mean ± S.D. Statistical comparisons among groups were made by one-way analysis of variance (ANOVA) or unpaired Student’s t-test. Graphs were generated by GraphPad Prism 6.0 software (GraphPad Software, USA). A probability value of P less than 0.05 was considered statistically significant.

Data availability statement

The datasets used to support the findings of this study are available from the corresponding author upon request.

Effects of PLD treatment on the general indices of chronic SCI

As shown in Table 2, although there was no significant difference in the initial body weight among the groups, mice with SCI weighed slightly less compared to the sham-operated mice at the end of the experiment. The serum levels of Ca²⁺ and OCN in the SCI group were significantly lower compared to that in the blank control group. In addition, SCI mice exhibited enhanced bone resorption, as indicated by the marked increase in serum CTX-1 and urine DPD levels. PLD treatment significantly increased serum OCN levels, decreased serum CTX-1 and urinary DPD levels, and slightly improved the body weight of the mice with chronic SCI. However, PLD treatment had no significant effect on the serum Ca²⁺ levels of the SCI model, or on any of the above parameters in the sham-operated control mice.

PLD reversed chronic SCI-induced bone loss

Mice with SCI had 34.8%, 35.1%, and 8.3% lower BMD compared to the sham-operated mice in the distal femoral metaphysis (DFM), proximal tibial metaphysis (PTM) and lumbar vertebrae, respectively. The loss of BMD in both lower limbs was similar (data not shown). Eight weeks after PLD administration, the BMD of the DFM and PTM were restored to that of the sham-operated group (Figure 1C, 1D). It is worth noting that the BMD of the lumbar spine was slightly greater for PLD-treated SCI mice compared to that of the sham-operated mice (Figure 1E). Chronic SCI also decreased the BV/TV, Tb.N and Tb.Th, and increased the Tb.Sp at this site compared to the sham-operated controls (Figure 2C–2F). In addition, trabecular Conn.D was significantly reduced after SCI, and associated with the transition from plate-like to rod-like structures (Figure 2G, 2H). Treatment of SCI mice with PLD significantly increased the BV/TV ratio, primarily by increasing Tb.Th (+ 97.6% vs. SCI) and reducing Tb.Sp (- 42.7% vs. SCI), with a lesser effect on Tb.N (Figure 2C–2F). Furthermore, a slight (statistically non-significant) increase in Conn.D was also noted in chronic SCI mice due to SMI reduction (Figure 2G, 2H).

High-resolution μCT was also used to evaluate the effect of PLD on the cortical bone structure in the femur midshaft. The cortical BMD of all 4 groups was similar 16 weeks after surgery (data not shown). However, the SCI mice had thinner bones and reduced total area, bone area, and medullary area compared to the sham-operated group (Figure 3B), resulting in a slight decrease in Ct.Th (Figure 3E). Interestingly, both Ct.BV/TV (+ 16.1% vs. SCI; Figure 3D) and Ct.Th (+ 42.2% vs. SCI; Figure 3E) were significantly increased in the PLD-treated SCI animals due to a moderate increase in periosteal perimeter (+ 10.9% vs. sham; Figure 3C), as well as significant reductions in medullary area (- 16.6% vs. sham; Figure 3B) and endosteal perimeter (- 16.5% vs. SCI; Figure 3C). The femoral diaphysis of the SCI mice had significantly lower ultimate force and stiffness, and required less energy to fracture compared to that of the sham-operated group (Figure 4). PLD treatment significantly improved the mechanical properties of the bones of SCI mice. However, 8 weeks of PLD treatment had no significant effect on the bone geometry, microstructure and mechanical properties of sham-operated mice.

Figure 4. Effects of treatment with PLD on biomechanical parameters of mouse femurs. (A) Ultimate force, (B) stiffness, and (C) energy to fracture of the femurs among the four groups of mice. Data are expressed as mean ± S.D.; n=4 to 5 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.

PLD regulated bone formation and bone resorption in chronic SCI mice

We next analyzed the effect of PLD treatment on bone formation and bone resorption in terms of histological changes, osteoblastogenesis and osteoclastogenesis markers, and calcium deposition in the femurs. H&E staining of DFM revealed smaller, thinner trabeculae with more lipid droplets and microfractures in the SCI group relative to the sham-operated controls. In contrast, PLD treatment largely attenuated the SCI-induced histo-morphological damage (Figure 5A).

Figure 5. Effects of PLD on bone morphology, bone formation, calcium deposition and bone resorption. (A) Representative photomicrographs of H&E-stained sections. IHC analysis of (B) OCN and (C) OPN protein expression in distal femurs. (D) Number of OCN⁺ and OPN⁺ osteoblasts on the bone surface. (E) Representative images of TRAP-stained distal femoral sections. Arrows indicate TRAP⁺ cells. (F) Histo-morphometric quantification of osteoclast number per bone perimeter (N.TRAP+/B.Pm), and normal osteoclasts surface per bone surface (OC.S/BS). (G) Representative images of von Kossa-stained mouse femoral trabecular bone sections. Arrows indicate the presence of mineralized bone. (H) Quantitative analysis of osteoid volume versus total volume (OV/TV), osteoid volume versus bone volume (OV/BV), and osteoid surface versus bone surface (OS/BS). Measurements were presented as mean ± S.D.; n=6 to 7 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.

As shown in Figure 5B–5D, the in-situ expression of OCN and OPN was markedly decreased in the vehicle-treated SCI animals, suggesting inhibition of osteoblastic differentiation. PLD treatment improved bone formation in DFM, which was indicated by an increased number of OCN-positive mature osteoblasts and OPN-positive pre-osteoblasts on the bone surfaces. Furthermore, von Kossa/Tetrachrome staining revealed a marginal reduction in osteoid volume (OV/TV and OV/BV) and osteoid surface (OS/BS) in the SCI group compared to the sham-operated group (Figure 5G, 5H), which were substantially elevated after PLD treatment to levels higher than that of the sham-operated mice.

The number of TRAP⁺ osteoblasts (N.TRAP⁺/B.Pm) in the SCI group was 46 ± 7.28 per bone perimeter (mm^-1), which was significantly higher compared to 15 ± 2.73/mm^-1 in the sham-operated group (Figure 5E, 5F). After PLD administration, the number of N.TRAP⁺/B.Pm cells decreased significantly. Moreover, the normal osteoclasts surface per bone surface (OC.S/BS) of vehicle-treated and PLD-treated SCI mice showed an opposite trend. PLD treatment had not significant effect on the number of osteoblasts and osteoclasts in the sham-operated animals. Taken together, PLD enhanced bone formation and suppressed bone resorption in chronic SCI mice.

PLD abated oxidative stress and attenuated MMP activity in chronic SCI mice

MDA level is an indicator of lipid peroxidation and oxidative stress levels. Compared to the sham-operated mice, the serum and femoral MDA content were higher in the chronic SCI mice (Figure 6A, 6B). PLD supplementation significantly decreased serum and femoral MDA levels, which is consistent with its anti-oxidant effects. Induction of SCI also activated femoral MMP9, which plays a key role in bone resorption (Figure 6C), compared to the control group. Treatment of chronic SCI mice with PLD significantly decreased protein expression of MMP9 in the femur.

Figure 6. Effects of treatment with PLD on oxidative stress and vitamin D deficiency in serum and femurs of mice. MDA levels in serum (A) and femurs (B) were determined using a kit. (C) The expression level of MMP-9 in femoral tissues was detected by immunoblot. (D) Serum 25(OH)D levels were determined by radioimmune assay. (E) Femoral protein expression of VDR were determined by Western blotting and quantification analysis were shown. The protein levels of MMP-9 and VDR were adjusted as relative values to GAPDH. Data are expressed as mean ± S.D.; n=5 to 7 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.

PLD had no effect on 25(OH)D and VDR in chronic SCI mice

We assessed the effect of PLD therapy on the reversal of vitamin D deficiency by detecting serum 25(OH)D levels and VDR protein expression in the distal femurs. Although the serum 25(OH)D and VDR protein expression levels in SCI group were significantly lower than that in the sham-operated group, PLD treatment had no effect on these parameters (Figure 6D, 6E).

PLD enhanced osteoblastogenesis in mice with chronic SCI

At 16 weeks after SCI induction, a marked reduction was observed in the number of ALP⁺ osteoblast-like cells (CFU-F), an indicator of the differentiation of bone marrow precursors into the osteoblast lineage (Figure 7A). Similarly, the number of mineralized nodules (CFU-OB) was also significantly reduced in the SCI group (Figure 7A). PLD treatment restored osteoblast differentiation in the affected bones. As shown in Figure 7B, PLD promoted BMSC calcification in animals with chronic SCI, as evidenced by Alizarin red staining. This result was also consistent with the outcome of the quantitative analysis of the ALP activity assay (Figure 7C).

Figure 7. PLD promoted osteoblastogenesis in bone marrow progenitor cells generated from chronic SCI mice. (A) The number of osteoblast-like cells (CFU-F) and the number of mineralized nodules (CFU-OB) in BMSC cultures were identified by ALP staining and von Kossa staining, respectively. (B) Alizarin red staining was used to evaluate BMSC calcification in the presence or absence of OIM of in groups. (C) Quantitative analysis of the ALP activity assay. (D) RUNX2, OCN, ALP, OPG and RANKL mRNA levels, and OPG/RANKL ratio in the osteoblast-like cells. mRNA levels were determined by qRT-PCR analysis. Data are expressed as mean ± S.D.; n=4 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.

Total RNA was isolated from osteoblast-like cells derived from bone marrow stromal precursors in vitro to detect the transcript levels of osteogenic differentiation markers. Compared with the sham-operated group, the chronic SCI group showed a 50%, 69% and 60% reduction in the relative expression of RUNX2, OCN and ALP, respectively (Figure 7D). PLD treatment significantly increased all transcripts to levels similar to that of the sham-operated group. OPG mRNA expression was reduced by approximately 40% in the chronic SCI mice compared to the sham-operated controls, which was accompanied by a 2-fold elevation in RANKL mRNA expression (Figure 7D). Treatment of SCI mice with PLD significantly upregulated OPG and downregulated RANKL mRNA levels, resulting in 2-fold increase in the OPG/RANKL ratio. There was no significant difference between the two control groups in terms of the above transcripts.

PLD inhibited osteoclastogenesis in mice with chronic SCI

The osteoclastic potential of bone marrow hematopoietic precursors in different groups was determined by counting the number of TRAP⁺ multinucleated cells (MNCs). Consistent with previous findings [26, 27], TRAP⁺ MNCs increased by 55% in bone marrow cell cultures from mice with chronic SCI compared to that of the sham-operated controls (Figure 8A). Administration of PLD significantly reduced the number of TRAP⁺ MNCs by 19% in animals with chronic SCI.

Figure 8. PLD inhibited the osteoclastogenic potential of bone marrow precursors isolated from chronic SCI mice. (A) The number of osteoclasts were identified by TRAP staining. (B) CTR and TRAP mRNA levels in cultured bone marrow progenitor cells. mRNA levels were determined by qRT-PCR analysis. (C) NFATc1 and c-Fos protein expression was detected by Western blotting and adjusted as relative values to GAPDH. Data are expressed as mean ± S.D.; n=4 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.

Consistent with this, the osteoclast differentiation markers CTR and TRAP were significantly upregulated by about 60% and 170%, respectively, in osteoclast cultures from animals with chronic SCI. After PLD treatment, the mRNA levels were decreased to levels similar to those detected in vehicle-treated or PLD-treated sham-operated animals (Figure 8B). In addition, the protein expression of osteoclast target genes (NFATc1 and c-Fos) also showed a similar trend (Figure 8C).

PLD activated the Wnt/β-catenin pathway in the femurs of chronic SCI mice

Wnt/β-catenin signaling was evaluated by analyzing the femoral levels of SOST, Wnt3a, Lrp5, and ctnnb1 mRNAs. SOST mRNA levels in the chronic SCI mice were approximately 2-fold higher compared to that in the sham-operated group, and unaffected by PLD treatment. Compared to the sham-operated group, Wnt3a, Lrp5, and ctnnb1 mRNAs were significantly downregulated in the SCI group, and restored to varying degrees by PLD (Figure 9A). Strikingly, Lrp5 expression in the PLD-treated SCI group was similar to that of the sham-operated controls. In addition, both the active and total β-catenin protein levels were decreased in the femurs of chronic SCI mice and reversed after PLD treatment (Figure 9B). Taken together, PLD exerted its therapeutic effects against chronic SCI by activating the femoral Wnt/β-catenin signaling.

Figure 9. PLD activated Wnt/β-catenin signaling pathway in the femurs of chronic SCI mice. (A) SOST, Wnt3a, Lrp5 and ctnnb1 mRNA levels were determined by qRT-PCR method. (B) Immunoblot showing active β-catenin and total β-catenin levels. The protein levels of β-catenin were adjusted as relative values to GAPDH. Data are expressed as mean ± S.D.; n=5 to 7 per group; *P < 0.05, vs. the Sham group; #P < 0.05, vs. the SCI group.