The animal experimental protocol used in this study was approved by the Institutional Animal Care and Use Committee, Yonsei University Health System (IACUC approval No. 2018-0226). Five male beagle dogs (age, 12–15 months; weight, 10–13 kg; OrientBio, Seongnam, Korea) were used in this study. Only beagles with good tooth arrangements and a healthy periodontium were selected. All beagles were housed individually in cages and bred according to the guideline for the Care and Use of Laboratory Animals of medium and large size of the Department of Laboratory Animal Resources, Yonsei Biomedical Research Institute, Yonsei University College of Medicine.

For all experimental procedures, dogs were anesthetized intravenously with 1% alfaxan (5 mg/kg, Jurox, Rutherford NSW, Australia) or via the inhalation of 2% isoflurane (Sigma-Aldrich, St. Louis, MO, USA) depending on the time required for the experiment. The dogs were continuously monitored during and after the procedure to check for any side effects of anesthesia, and no notable adverse reactions were observed.

Lingual buttons (Ormco Co., Orange, CA, USA) were bonded to the labial and lingual surfaces of the mandibular lateral incisors (I² teeth) and on the labial surface of the mandibular canine (C¹ teeth) of five beagle dogs using Superbond C&B (Sun Medical Co., Shiga, Japan). The teeth were rotated by applying rotational couple force with a power chain (Clear Generation II; Ormco Co.) for 6 weeks (Figure 1A). Each of the elastic chains engaged between buttons exerted initial forces of around 50 g.²¹ The power chain was replaced every 2 weeks and dogs were inspected every day for any debonding of bonded appliances. After 6 weeks of rotational movement, we generated a steel connection between lingual buttons by passively ligating them with a 0.2-mm steel ligature to retain the teeth in their current form.

A total of 10 lateral incisors for five beagles was examined, of which the right and left incisors were assigned to the control and experimental groups, respectively. During a retention period of 4 weeks, SFN (Sigma-Aldrich, 10 mM, 0.06 mL, the concentration and dosage were referred to in Kanzaki’s study¹⁴) for the experimental group, and dimethylsulfoxide (DMSO, solvent of SFN, 0.06 mL) for control groups, were injected into the supra-alveolar gingiva of the buccal side of the rotated teeth once per week, four times total, using a 31 G-needle insulin syringe (Ultra-Fine II 6 mm, 0.3 mL; Becton Dickinson Co., Franklin Lakes, NJ, USA). The needle was inserted at right angles to the attached gingiva area located directly below the boundary between free gingiva and attached gingiva, vertically, and at both centers of half the buccal gingiva of the experimental teeth, horizontally. After this, SFN and DMSO were injected into each side of 3 units (Figure 1B and 1C). One week after the last injection, the appliances were removed to allow relapse of the rotated teeth. An intraoral impression was obtained immediately (T1) after the orthodontic appliances were removed and 2–4 weeks later (T2, T3) to create dental models, which were compared to the initial model (T0), before the application of the orthodontic force, to measure the extent of tooth rotation. The degree of relapse of the rotated teeth in the experimental and control groups was also compared and analyzed. The experimental schedule for the beagle dogs is briefly illustrated in Figure 2.

Dental models of the beagles (T0–T3) were scanned with a 3Shape Lab Scanner (Lab scanner E3; 3Shape Co., Copenhagen K, Denmark), and the amount of tooth rotation was measured using the Geomagic Control X^TM (3D SYSTEMS, Seoul, Korea) software. The angle between the line connecting the cusp tip of the second premolars and the extension line through the incisal edges of the left and right experimental teeth (I², n = 10) was measured to determine original, rotational, and relapse angles (Figure 3). With these measurements, the degree of rotation (T1–T0), amount of relapse after 2 and 4 weeks (T1–T2, T1–T3), and percentage of total relapse after 2 and 4 weeks (T1–T2 / T1–T0 × 100%, T1–T3 / T1–T0 × 100%) were calculated to determine whether SFN could reduce orthodontic relapse of the rotated teeth.

The relapse angle was measured after arranging the scanned images such that the occlusal plane of the mandibular molars was parallel to the ground. The angle was measured three times and the correlation coefficient of each measurement was obtained. The three measurements were highly correlated, with a correlation coefficient ≥ 0.99, in both the control and experimental groups, and thus, the average of the three values, rounded to three decimal places, was used for analysis.

The mesio-buccal supra-alveolar gingiva of the experimental teeth (n = 10) was dissected for real-time quantitative polymerase chain reaction (PCR) before sacrificing the beagles. We chose collagen type I alpha 1 (COL1A1), lysyl oxidase (LOX), elastin (ELN), MMP 1, MMP 12, and tissue inhibitor of metalloproteinases-1 (TIMP 1) as ECM components associated with the elasticity of the gingiva. To identify changes in the expression level of genes after SFN injection, total RNA was isolated from each mesio-buccal supra-alveolar gingiva of the experimental teeth injected with either DMSO or SFN using TRIzol reagent (Gibco BRL, New York, NY, USA). Three biologic replicates were used. cDNA was synthesized from total RNA with Maxime RT PreMix (iNtRON, Seongnam, Korea). Quantitative PCR was performed using the StepOnePlus Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific, Beverly, MA, USA) and Power SYBR green master mix (Thermo Fisher Scientific), according to the manufacturer’s instructions. Quantitative PCR products were normalized using β-2-microglobulin (B2M) levels as internal controls. Primer sequences used for real-time quantitative PCR are shown in Table 1. Data were analyzed using the StepOne (Applied Biosystems, Thermo Fisher Scientific) analysis software and the 2^−∆∆Ct method. All PCRs were performed in triplicate.

After the beagles were sacrificed via an intravenous injection of KCl, distobuccal half segments of the experimental teeth were dissected (n = 10), fixed in 4% paraformaldehyde solution at 4°C for 2 days, and decalcified in ethylenediaminetetraacetic acid (EDTA) (10%, pH 7.4) solution for 4 months at 4°C. This EDTA solution was replaced twice per week. Fully decalcified samples were dehydrated in a graded ethanol series (70–100%), cleared in xylene, and then infiltrated and embedded in paraffin. Buccolingual serial sections, 3-µm thick, were cut parallel to the longitudinal axis of the experimental teeth. The qualified ones among the sections cut near the center of the tooth were selected and mounted.

Masson’s trichrome staining was performed to detect collagen fibers in gingival tissues. The distal images of Masson’s trichrome staining were observed using a Pannoramic 250 Flash III (3DHISTECH, Budapest, Hungary) and Caseviewer (ver. 2.0; 3DHISTECH).

All data are represented as means and standard deviations. Intergroup comparisons of mean values comprising the amount and percentage of relapse between the two groups followed a normal distribution and were analyzed with a paired t-test. Intergroup comparisons of the mean values of real-time quantitative PCR data were analyzed with a Wilcoxon signed rank test to compare DMSO and SFN groups. p-values less than 0.05 were considered statistically significant. IBM SPSS Statistics (ver. 23.0; IBM Corp., Armonk, NY, USA) was used for statistical analysis.

Table 2 shows intergroup and intragroup comparisons of the mean amount and percentage of rotation in the SFN and DMSO groups. There were no significant differences in the mean degree of rotation after 6 weeks between the groups, whereas the mean amount and percentage of relapse at 2 and 4 weeks after injection were significantly reduced in the experimental group compared to those in the control group (p < 0.05).

In the control group, no significant differences were found between the mean amount and percentage of rotational relapse at 2 and 4 weeks after the retention period, indicating that most relapse occurred within 2 weeks after completion of the retention period. In contrast, in the experimental group, significant differences were found between the mean amount and percentage of rotational relapse between 2 and 4 weeks after the retention period (p < 0.05), indicating that relapse occurred at a relatively slower rate.

In the mesio-buccal supra-gingival tissue of beagles, the expression of genes encoding COL1A1, LOX, ELN, MMP 1, MMP 12, and TIMP 1 were compared after SFN or DMSO treatment by real-time RT-PCR. The gene expression of COL1A1 (Figure 4A) was not significantly different between the two groups. However, the gene expression of LOX (Figure 4B) was decreased in the SFN-treated group compared to that in the DMSO group. The gene expression of ELN (Figure 4C) was not significantly different between the two groups. In addition, the gene expression of MMP 1 (Figure 4D) and MMP 12 (Figure 4E) were increased in the SFN-treated group compared to levels in the DMSO-treated group. On the other hand, the gene expression of TIMP1 (Figure 4F) was not significantly different between the two groups.

We next observed the histological appearance, morphology, and arrangement of collagen fibers in Masson’s trichrome-stained buccal gingival tissue of the experimental and control teeth (Figures 5 and 6). Collagen fibers bundles were longer and more connected in the DMSO-treated group than in the SFN-treated group (Figure 5C and 5D). The morphology and arrangement of the principal gingival fibers are shown in detail in larger images (Figure 6). The collagen fiber bundles that form dentogingival fibers in the SFN-treated group were short and not well connected to one another (Figure 6B). However, in the DMSO-treated group, we observed a more regular and well-ordered arrangement of the collagen fiber bundles connected to form the dentogingival fibers extending from the cervical cementum into the lamina propria of the free and attached gingiva (Figure 6A). Circular fibers are fiber bundles that run their course in the gingiva and encircle the tooth in a cuff- or ring-like fashion. In the SFN-treated group, circular fibers were arranged in a thinner and less compact manner compared to the arrangement in the DMSO-treated group (Figure 6C and 6D). The dentoperiosteal fibers extended from the cementum directly over the alveolar crest and then inclined apically to the periosteum of the alveolar bone. Similarly, collagen fiber bundles connected to form dentoperiosteal fibers were irregularly arranged and not well connected in the SFN-treated group (Figure 6F), whereas collagen fiber bundles were shown more regular and well-ordered arrangement in the DMSO-treated group (Figure 6E).