This study was designed according to the modified Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines for preclinical in vivo experiments [28]. The research was carried out in Shanghai after receiving approval from the Medical Animal Care & Welfare Committee of Shanghai Ninth People’s Hospital affiliated with Shanghai Jiao Tong University, School of Medicine (HKDL2018225).
Animals and facilities
Six female Labrador dogs aged between 1.5 and 2 years, weighing approximately 20 kg, were used. All dogs were healthy, housed in kennels and fed a soft diet by qualified staff onsite during the entire procedure.
Experimental materials
A total of 22 cylindrical dental implants (OsseoSpeed™, Astra Tech®, Dentsply Sirona, USA; 4 mm in diameter and 10 mm long) were used. A bone substitute consisting of DBBM (Bio-Oss®, Geistlich, Wolhusen, Switzerland) and a non-crosslinked porcine CM (Bio-Gide®, Geistlich, Wolhusen, Switzerland) was used.
Experimental design
In each hemimandible, the fourth premolar (P4) and the first molar (M1) were extracted, and implants were immediately inserted into the fresh socket. After IIP, the extraction sockets were allocated to one of the following groups. Every two sites of one hemimandible were allocated to one group, which means that P4 and M1 were distributed among groups in a balanced way. One of the dogs lost its 1st molars, and therefore, there were 5 implants in groups BG and C. Due to the lack of relevant experimental data, the sample size was determined by referring to existing similar studies [29, 30]. The groups were defined as follows:
1. group BG (n = 5), blank group
2. group C (n = 5), with CM (Bio-Gide®) only
3. group D (n = 6), with DBBM (Bio-Oss®) only
4. group DC (n = 6), with DBBM (Bio-Oss®) and CM (Bio-Gide®)
The order of treatments and measurements were determined randomly to minimize potential confounders.
Surgical procedures
The animals were anesthetized with tiletamine hydrochloride, zolazepam hydrochloride (Zoletil®50, 0.1 mg/kg) and xylazine hydrochloride (Huamu®, 1~2 mg/kg). The P4 and the M1 were extracted. Every tooth was extracted with minimally invasive technology, namely, separation with a cutting drill and removal with extracting forceps. A blinded examiner carefully measured the dimensions of the socket using a periodontal probe (Hu-Friedy®, Chicago, IL, USA) and confirmed that the walls of the socket were intact. Sharp edges of alveolar bone were trimmed away so that the septal bone was suitable for implantation. The site was prepared according to the manufacturer’s instructions. A Ф4.0 × 10 mm cylindrical implant (OsseoSpeed™, Astra Tech®, Dentsply Sirona, USA) was placed with a torque of more than 35 Ncm. The shoulder of the implant was placed at the level of the buccal bone crest, and a cover screw was installed. Each implant was placed at the septal bone.
In group BG, after the implants were installed, no materials or CMs were placed, and blood filled the gap naturally; that is, natural healing began. In group C, the gap was not filled with any material. The mucoperiosteal flap on the lingual side was separated, and one side of a 30 mm × 40 mm Bio-Gide® membrane was inserted to cover the socket. Then, the other side of the membrane was covered with a buccal mucoperiosteal flap. In group D, an appropriate amount of Bio-Oss® was selected according to the size of the gap. Large particles and the small particles were mixed at a ratio of 1:1 and moistened with saline in advance. The gap was filled with Bio-Oss®, which was compacted to ensure that the gap was filled with particles. Then, the mucoperiosteal flap was reset and sutured. In group DC, the gap was filled with Bio-Oss® particles, followed by Bio-Gide®, and finally, the flaps were repositioned and sutured. Surgical photos of the different groups are shown in Figure 1, and a schematic diagram of the surgical procedure is shown in Figure 2.
Antibiotics and analgesics were injected into the animals for 7 days after the surgery, and over the next 14 days, the dogs were fed water-softened food.
Sequential fluorescence labeling
Triple fluorescence labeling of the alveolar bone was conducted according to previous research [30]. For the next four, eight and ten weeks, animals under anesthesia were injected with three kinds of fluorochromes: tetracycline hydrochloride (TE, 25 mg/kg), calcein (CA, 20 mg/kg), and alizarin red S (AL, 30 mg/kg).
Euthanasia
After 12 weeks, the dogs were euthanized with an overdose of anesthetics, and their bone blocks were harvested for histologic analysis. Undecalcified specimens were fixed in 10% buffered formalin.
Micro-CT analysis
The specimens were scanned by a micro-CT machine (Skyscan1076, Bruker, Belgium), and the settings were as follows: source voltage (kV)=70, source current (uA)= 141, image pixel size (µm)= 36.5200, filter=Al 1.0 mm, exposure (ms)=110, rotation step (deg)=0.700, and frame averaging=On (1). The data were processed with software (Microview, Scanco Medical AG). The 35-μm thick 2D slices were reconstructed into a 3D model. The volume of interest (VOI) of each implant was a cylinder 3 mm in diameter and 4 mm in height starting at the level of the thread in the mesial and distal area near the implant. The biopsies were approximately aligned with the implant axes. The main characteristics used to evaluate the newly formed bone were as follows: bone volume (BV), total volume (TV), bone surface (BS), bone mineral density (BMD, representing the density of bone mineral in the bone tissue), bone volume fraction (BV/TV, representing the ratio of bone tissue volume to tissue volume, which can directly reflect the changes of bone volume), bone surface volume ratio (BS/BV, representing the area of bone tissue per unit volume), trabecular thickness (Tb.Th, representing the average trabecular bone thickness), trabecular number (Tb.N, representing the mean number of bone and nonbone tissue intersections per millimeter), and trabecular separation (Tb.Sp, representing the mean width of the cavity between bone trabeculae). The micro-CT screenshots of each group are shown in Figure 3.
3D reconstruction and bone height measurement
The data obtained from the micro-CT scans were reconstructed using Mimics 21.0 (Materialise, USA) software. The height of the new bone in the mesiodistal sites of the implants was obtained, and the measurement method is shown in Figure 4.
Histological analysis
After fixation in formaldehyde solution, the specimen was dehydrated with a series of alcohol of increasing concentrations. Then, the specimens were placed into the embedding solution (methyl methacrylate: dibutyl phthalate = 4:1). When the resin was completely cured, the specimen was obtained. Each mesiodistal section representing the central area of the implant was prepared for biopsy. This cutting direction was determined to observe bone formation in the mesiodistal sockets around the implants. After the sections were reduced to a thickness of 20-25 μm by grinding and polishing, they were subjected to fluorescence microscopy analysis (confocal laser scanning microscopy, TCS SP8 STED 3X, Leica, Wetzlar, Germany). The number of pixels labeled with different colors in each image was used to calculate the percentage of mineralized bone in ImageJ software (National Institutes of Health, USA). After that, when Stevenel's blue and van Gieson’s picrofuchsin staining was complete, the sections were observed under a microscope (Olympus, Tokyo, Japan). The region of interest (ROI) was a rectangle with a length of 40 mm and a width of 10 mm located at the edge of the implant, below the shoulder. Bone-to-implant contact (BIC) was defined as the ratio of the length of contact between the implant and bone in the area to the full length of the implant thread. The percentage of the new bone area around the implants was determined by the ratio of the new bone area to the total ROI area.
Data analysis
Groups C and BG each contained 5 sites due to the lack of M1s in one dog. Group DC and D each contained 6 sites. The data are shown as the mean value ± standard deviation. The normality of the distribution was tested by the Shapiro–Wilk test, and then, the homogeneity of variance was analyzed. For the normally distributed data that met homogeneity of variance, one-way ANOVA was further performed to assess the differences among groups. If P < 0.05, then further post hoc Tukey tests were performed to evaluate the significance. For data that did not meet the normal distribution or homogeneity of variance, the Kruskal-Wallis test was used to assess the difference among groups, and Bonferroni correction was used to correct p-values. The above statistics were calculated with SPSS 25.0 software (Chicago, IL, USA).