LIPUS promotes spinal fusion coupling proliferation of type H microvessels in bone

Low-intensity pulsed ultrasound (LIPUS) has been found to accelerate spinal fusion. Type H microvessels are found in close relation with bone development. We analyzed the role of type H vessels in rat spinal fusion model intervened by LIPUS. It was found LIPUS could significantly accelerate bone fusion rate and enlarge bone callus. Osteoblasts were specifically located on the bone meshwork of the allograft, and were surrounded by type H microvessels. LIPUS could significantly increase the quantity of osteoblasts during spine fusion, which process was coupled with elevated angiogenesis of type H microvessels. Our results suggest that LIPUS may be a noninvasive adjuvant treatment modality in spinal fusion for clinical use. The treatment is recommended for usage for at least one month.

Histological assessments. At one week post-surgery, no chondrocytes or fibrous tissue could be seen in the fusion bed of the LIPUS or control groups (1 st column Fig. 3). At two weeks post-surgery, fibrous tissues could be observed in the fusion site in the control group, while a large number of chondrocytes could be observed in the fusion site in the LIPUS group (2 nd column Fig. 3). Many inflammatory cells were scattered in the bone grafts. At three and four weeks after surgery, more chondrocytes and a newly formed bone marrow could be seen in the fusion sites of both groups (3 rd and 4 th columns Fig. 3). A number of blood vessels were found in the fibrous tissue and bone marrow. Fusion sites were completely invaded by bone tissues intertwined with blood vessels in the LIPUS group ( Increased type H blood vessels in the LIPUS group. Both CD31 and Endomucin were employed to target type H microvessels. At two weeks post-surgery, CD31 hi microvessels were present in the LIPUS and   respectively. In both groups, the majority of CD31 hi microvessels spread around the allograft exterior, circling the osteoblasts; and few vessels were observed in the outlying area of chondrocytes. Vessel density of CD31 hi microvessels significantly increased from two weeks post-surgery to four weeks post-surgery, and there were significantly more CD31 hi microvessels in the LIPUS group compared to the control group (  Table 2). At two weeks post-surgery, Emcn hi microvessels in the callus were sparse in the LIPUS and control groups (2 nd column Fig. 6A). At three and four weeks post-surgery, an increased number of Emcn hi microvessels spread around the allograft and in the outlying area of chondrocytes in both groups (Fig. 6A JLNP). Mean vascular densities in the LIPUS and control groups were 3.22 ± 0.56% and 1.81 ± 0.60% at four weeks after surgery (P < 0.01), respectively. There were significantly more Emcn hi microvessels in the LIPUS group compared to the control group (  Table 2). Emcn hi microvessels were fewer than CD31 hi microvessels, but Emcn hi microvessels had the same distribution as CD31 hi microvessels.

Discussion
The results of this study suggest that LIPUS can enhance posterior spinal fusion with allograft in rats. Our histological assessments have shown that osteoblasts were found on the allograft, while chondrocytes were observed near the transverse process. The fusion process was accompanied by the formation of type H microvessels, which were located in the immediate proximity to the osteoblasts.
Ultrasound is a form of mechanical energy that can be transmitted into the body with no thermal effects. LIPUS has been approved by the U.S. FDA for the treatment of fresh fractures and existing non-unions. The study of Zura et al. revealed that the healing rate for chronic (1-10 years) non-union fractures in patients (n = 767) treated with LIPUS was 86.2% 31 . Another study carried out by Yunkawa et al. investigated the effects of LIPUS on spine fusion in patients undergoing single level lumbar interbody fusion 25 . They found that LIPUS could increase fusion rate over a 48-week follow-up period. However, this increase was not statistically significant. In animal studies, LIPUS has significantly increased the fusion rate and fusion mass in spinal fusion models 23,24,32 . Similarly, the current study revealed that LIPUS could accelerate the fusion process and increase fusion rate in rats, as evidenced by manual palpation and micro-CT, as well as histological analyses. At four weeks post-surgery, fusion rate was 40% in the control group and 100% in LIPUS-treated rats. Bone volume, Tb.N and Tb.Th significantly increased in the LIPUS group, compared with the control group. Furthermore, there was continuity between the allograft and transverse process in the LIPUS group, while a shallow gap existed in the control group. In accordance with our observations, callus in patients treated for fracture healing with LIPUS is larger compared to controls [22][23][24]33 .
In in vitro studies, LIPUS has been found to directly affect osteogenic cells and enhance chondrogenesis 34,35 . Hui et al. and Cook et al. observed a mass of chondrocytes at the fusion site in rabbits and dogs, respectively; and suggested that endochondral ossification was the main process involved in spinal fusion 32,36 . In the current study, large quantities of chondrocytes were found at the interface of the allograft and transverse process; but were very few in the allograft bone meshwork. In addition, there was significantly more chondrocytes and a newly formed bone marrow in the LIPUS group, compared to the control group. Cook et al. found that LIPUS could promote new bone formation and remodeling in a dog spinal fusion model 36  Notably, we found that the number of osteoblasts surrounding the bone meshwork throughout the allograft increased with time after surgery in both the control and LIPUS groups. These observations were in accordance with our radiological findings. Interestingly, osteoblasts were not localized around chondrocytes, and were positioned on the allograft meshwork instead. The density of osteoblasts on the bone meshwork significantly increased in the LIPUS group, compared to the control group. Specifically, osteoblast did not appear at one week after surgery in both control group and LIPUS group. More osteoblasts were observed at two weeks post-surgery in the LIPUS group. And there was much more osteoblasts at four weeks post-surgery compared with at 2 or 3 weeks post-surgery. Based on these findings, we speculate that osteoblast-induced bone formations play a central role in spinal fusion, rather than endochondral ossification. Zhuo et al. used porous hydroxyapatite blocks as a graft for spinal fusion, and found that LIPUS increased the number of osteoblast-like cells in the porous blocks, as evidenced by a scanning electron microscope 38 . However, they did not describe the distribution of osteoblasts.
There have been few reports of clinically relevant data describing the effects of LIPUS on temporal changes in osteoblast frequency, and spatial distribution with time during spinal fusion. This is because most previous studies used a collagen sponge for the graft, which might induce a different pathological fusion process, compared to that associated with a freeze-dried allograft 32,36,37 . In in vitro studies, LIPUS increased the proliferation and maturation of osteoblasts, as well as the formation of the extracellular matrix 34,35 .
Evidence has suggested that angiogenesis is associated with osteogenesis, providing oxygen, stem cells, and growth factors that promote the osteogenic process. In this study, we aimed to investigate the role of type H microvessels in an osteogenic microenvironment 27,28 . Previous reports have shown that decreased numbers of CD31 hi Emcn hi ECs are associated with decreased osteogenic activity and osteoporotic changes 27 , and that pre-osteoblast cells are selectively positioned around type H microvessels. The study of Ralf et al. demonstrated a critical role for the Notch signaling pathway in angiogenesis. They revealed that Noggin, which is regulated by  Notch, is secreted by type H microvessels to promote the proliferation of osteoprogenitor cells 28 . Therefore, we investigated whether type H microvessels are involved in bone formation and remodeling during spinal fusion. Immunostaining has shown that the density of CD31 hi Emcn hi microvessels increased during fusion. Significantly, more type H microvessels were found in the LIPUS group compared to the control group. Solid fusion rate (determined by manual palpation), mass of the bone callus (determined by radiological assessment), the amount of bone tissues and number of osteoblasts in the fusion mass (determined by histological assessments), and the density of type H microvessels observed during the fusion process were greater in the LIPUS group compared to the control group. It was not until at two weeks post-surgery that type H microvessels were observed in both control and LIPUS groups. The quantity of type H microvessles increased as time going. At 4 weeks post-surgery, more type H microvessles were found in the fusion area compared with at 2 or 3 weeks post-surgery. Meanwhile, larger quantities of type H microvessels were in the LIPUS group compared with the control group. Importantly, the density of type H microvessels located in immediate proximity to the osteoblasts increased in parallel with the number of osteoblasts as fusion progressed. Taken together, these data suggest that type H microvessels play an essential role in the promotion of bone remodeling by LIPUS during spinal fusion. And it was not until two weeks treatment that LIPUS had effect in increasing quantity of osteoblasts as well as angiogenesis of type H microvessels.
Park et al. found that the increased rate of spinal fusion and greater bone volume was associated with denser vascularity in COMP-angiopoietin 1-treated rats that underwent bilateral posterior and posterolateral arthrodesis Figure 5. (A) Immunohistochemical staining for CD31 and type H microvessels in the fusion mass. The density of CD31 hi vessels increased. CD31 hi microvessels were spread around the allograft exterior, circling the osteoblasts. Few type H microvessels co-localized with chondrocytes. White arrows allografts; fat red arrows microvessels. Panel (A-P) is the same as described in Fig. 3. (B) CD31 hi microvessels were significantly higher in the LIPUS group compared to those in the control group. *P < 0.05. Figure 6. (A) Immunohistochemical staining for Endomucin and type H vessels in the fusion mass. The density of Emcn hi microvessels increased. Emcn hi microvessels were spread around the allograft exterior, circling the osteoblasts. Few type H microvessels co-localized with chondrocytes. White arrows allografts; fat red arrows microvessels. Panel (A-P) is the same as described in Fig. 3. (B) Emcn hi microvessels were significantly higher in the LIPUS group compared to those in the control group. *P < 0.05. with allograft, compared to rats treated with bovine serum albumin 39 . These newly formed vessels were positioned in the primary mineralizing area where osteogenesis was actively occurring. Liu et al. studied the effect of calcitonin on lumbar spinal fusion in a rabbit model, and found that calcitonin could enhance lumbar spinal fusion by increasing the expression of genes involved in osteogenesis and angiogenesis including collagen I, BMP-2, insulin-like growth factor-1, and vascular endothelial growth factor (VEGF) 40 . The study of Street et al. revealed that exogenous VEGF enhances angiogenesis, ossification, and new bone formation in a mouse model of fracture repair and a rabbit critical size defect model 41 . Treatment of mice with a VEGF receptor antagonist inhibited osteogenesis and fracture healing, while preventing angiogenesis. These data suggest that angiogenesis is indispensable for osteogenesis during spinal fusion. In in vitro studies, LIPUS causes osteoblasts to migrate and secrete VEGF and other growth factors 42,43 , and promotes the migration of ECs 44 . An in vivo study carried out by Cheung et al. revealed that LIPUS could enhance callus formation and angiogenesis during fracture healing in a rat model, and VEGF expression was significantly increased four weeks after surgery 45 . In accordance with these data, Azuma et al. found that LIPUS could accelerate fracture healing via angiogenesis in rats 46 . Lu found that LIPUS could improve bone-joint healing through the regulation VEGF expressions in the early phase, and subsequent chondrogenesis in rabbits 47 . Changes in VEGF expressions preceded endochondral ossification and bone remodeling. These studies have shown that LIPUS may affect the function of ECs, indicating that LIPUS may induce an increase in type H microvessels and promote spinal fusion.
This study is associated with several limitations. First, the mechanical strength of the fusion mass was not evaluated, because a single-level fusion procedure was employed in this study; which did not allow the measurement of the joint flexion. Second, the rat spine is small, which makes it difficult to accurately measure the fusion mass. However, manual palpation was used, which has been proven to be effective in fusion rate evaluation 39 . In the current study, manual palpation results were in accordance with micro-CT findings and histological analyses. Third, the mechanism by which type H microvessels regulate bone remodeling during spinal fusion was not investigated, because the microenvironment of allograft fusion differs from normal osteogenesis.

Conclusions
In conclusion, LIPUS can accelerate spinal fusion and increase bony mass in a rat spinal fusion model. This process is coupled with chondrogenesis, increased osteoblasts, and angiogenesis. Specifically, LIPUS significantly increased the proliferation of osteoblasts and type H microvessels. The main process of spinal fusion is osteoblast-induced bone formation, which process is coupled with angiogenesis of type H microvessels. LIPUS is recommended for enhancing spinal fusion for treatment of at least four weeks post-surgery. Further research is required to elucidate the origin of the osteoblasts that are formed during spinal fusion, and the factors that determine the spatial distribution of type H microvessels relative to osteoblasts.

Methods
Experimental design. Single level unilateral posterior spinal arthrodesis was evaluated in 40 male Sprague-Dawley rats. LIPUS was administered to 20 rats on the surgical area for 20 minutes per day post-surgery (LIPUS group), while the remaining animals received sham treatment (control group). Five animals were sacrificed at each of the following time points: one, two, three and four weeks post-surgery. Manual palpation and micro-computed tomography (micro-CT) scanning was performed on each animal after they were sacrificed at one and four weeks post-surgery to evaluate spinal fusion. The fusion mass was analyzed histologically and immunohistochemically to investigate the pattern of the fusion. This study was approved by the Ethical Committee of Changhai Hospital. All animals were handled strictly according to the Good Animal Practice requirements of the Animal Ethics Procedures and Guidelines of the People's Republic of China.
Animal model and surgical procedure. Forty male Sprague-Dawley rats (age, 12 weeks; weight, 300-350 g) were kept at 22 °C with free access to water and food. Anesthesia was induced and maintained with isoflurane (0.5% to 2%) and oxygen using a coaxial nose cone. After shaving the surgical site, a posterior lateral incision was made over the lumbar spine. The transverse process of the L4 vertebra was exposed by bluntly splitting the back muscles referencing the iliac crest. Once exposed, the transverse process was decorticated with an electric bur until shallow bleeding was observed. Then, a demineralized freeze-dried bone allograft (Aorui Biological Material Co., Ltd., Shanxi, China) was implanted on the decorticated fusion beds of the L4 transverse process. Finally, fascia and skin were closed with interrupted sutures. Postoperative antibiotics were administered intramuscularly for two days (cefuroxime, 0.5 mg/kg). None of the animals died before euthanasia. Twenty-five preliminary trial operations were performed to confirm the reliability of this model. The surgery was all performed by XMX. Animals were randomly allocated to LIPUS group or Control group by XZW. LIPUS device and treatment. Rats were randomly divided into two groups (n = 20 each), and all animals were anesthetized. LIPUS was initiated in use from 3rd day post-operatively. Rats in the LIPUS group were treated for 20 minutes per day (five days a week) with an Exogen 2000 + device (Bioventus Inc., Piscataway, NJ) with the following specifications: 30 ± 5.0 mW/cm 2 spatial average and temporal average incident intensity, 1.5-MHz pulsed frequency, 200-μ s pulse duration, 1.0-kHz repetition rate. Rats in the control group were treated with the same device for the same duration and number of days, but with no power. The transducer head was centered over the fusion site. Coupling gel was carefully smeared to ensure contact between the transducer head and skin. LIPUS treatment was conducted by XMX and XYZ.
Manual palpation assessment. Fusion mass at the bone-grafted segment was assessed by manual palpation ex vivo 48 The lumbar spine was gently palpated immediately after death by two assessors (FW and YHY), who were blind to group allocation. Disagreement was resolved by a third rater ML. Bridging bone formation between the graft and the L4 transverse process was evaluated. Fusion status of each specimen was graded based on the Scientific RepoRts | 6:20116 | DOI: 10.1038/srep20116 following categories: solid union, defined as no motion detected; immature union, defined as bony continuity with slight motion; and non-union, defined as wide motion detected. Fusion rate was defined as the percentage of solid or immature unions. Manual palpation is an established method for assessing the stiffness of the fusion mass in rats 49,50 , since mechanical stiffness of the spine segment is difficult to measure due to the small size of the grafted segment and complicated shape of the rat spine. Inter-tester reproducibility was tested via calculating intra-class coefficient (ICC).
Micro-computed tomography scanning. Using a CT scanner (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA), micro-CT was performed in the axial plane to evaluate the calcified bone graft. Scans were initiated from the pelvis to the L1 vertebral body cranially in 13-μ m sections. Microstructural indices were measured using MicroView (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA). Bone volume/tissue volume (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular separation (Tb.Sp) were calculated.
Histologic evaluation. After removal, L4-L5 transverse process fusion area were fixed with 4% paraformaldehyde for 24 hours. Then the specimens were decalcified in 5% nitric acid for 3 days, washed in distilled water, and embedded in paraffin 51 . Transversal sections (5 μ m) were obtained from the midline of the transverse process area. Sections were stained with hematoxylin and eosin (H&E), and observed under light microscopy.
Immunohistochemical staining for osteoblasts and blood vessels. Osteoblasts and type H microvessels were immunostained. Briefly, sections were rinsed in phosphate buffered saline to remove traces of the fixative, and were incubated in normal bovine serum for two hours at room temperature to block nonspecific binding of antibodies. Then, sections were incubated with goat polyclonal Osterix, CD31, and Emcn primary antibodies (diluted at 1:50; Santa Cruz, CA, USA) for 24 hours at 4 °C. After rinsing in phosphate buffered saline, fluorescein isothiocyanate conjugated bovine anti-goat secondary antibody (diluted 1:100; Santa Cruz, CA, USA) was applied for one hour at 37 °C. The sections were mounted after a final rinse, and examined under an Olympus eclipse 80i (Olympus, Japan) fluorescence microscope equipped with a mercury lamp and filter system. In histological analysis, the transverse process was first to be identified. Areas between transverse process and superior allograft were the interested fusion areas. Three sections were obtained in each specimen for statistical analysis. Osteoblast numbers or vascular densities in each section were acquired by calculating the means of five randomly captured views (× 400 magnification). Osteoblast number per bone surface (Ob/BS) was calculated 52 . Mean vascular densities was calculated as (CD31 hi or Emcn hi vascular area/total image area in × 400 magnification) × 100%. Analysis of images involved ImagePro Plus 5.02 (Media Cybernetics). Statistical analysis. SPSS 16.0 was used for statistical analyses. Differences in fusion status between groups were analyzed by Chi-squared test. Micro-CT and histomorphometric data are expressed as mean ± standard deviation (SD). The density of type H microvessels across various healing stages in the fusion mass was compared by one-way analysis of variance (ANOVA) for repeated measurements. Differences of osteoblast numbers and MVD between the control group and LIPUS group were analyzed by t test. P < 0.05 was considered statistically significant.