Two male non-human rhesus primates, Macaca mulatta were used in the study. Animal G was 15 years old and weighed 7.95 kg while Animal T was 16 years old and weighed 12.09 kg at the initial exam. Both animals were systemically healthy and their dentitions were free of caries or endodontic pathology. They were housed at the animal facility of VCU in extra-large enclosures, provided with social and environmental enrichment, and fed a diet of kibble (Monkey Chow, Purina). For the duration of the study, the diet of the animals was changed to a soft one, and means of mechanical oral hygiene were reduced to allow for plaque accumulation. The animals were treated twice daily for 14 days with 20 mg/kg/animal of amixicile encased in a marshmallow.

The animals were placed under general anesthesia by injection of ketamine (10 mg/kg) followed by intubation and administration of 2% isoflurane at 2 L/min and 100% oxygen at 1 L/min.

Clinical photographs and a comprehensive periodontal examination were performed during four visits: (1) baseline (pretreatment, visit 1), (2) posttreatment (visit 2), (3) 3-months posttreatment (visit 3), (4) 6-months posttreatment (visit 4) (Supplemental Figure 1). Periapical radiographs were taken at visit 1. Periodontal exam included recordings of pocket depth (PD), free gingival margin (FGM), bleeding on probing (BOP), and the presence of plaque at four sites (mesiobuccal (MB), buccal (B), distobuccal (DB), and lingual (L) on each tooth. Clinical attachment loss (CAL) was calculated from PDs and FGM values. Diagnosis of periodontitis was determined using CAL and supported by radiographic bone loss on both animals (Figure 1A). Photographs were used to assign a modified gingival index (GI) [11] score to each sextant for each exam. The scoring was as follows: GI 0 = pale pink to pink, knife-edge margin, positive architecture; GI 1 = slightly reddish, slight marginal edema, clear exudate, no bleeding on probing (BOP); GI 2 = red to bluish-red, glazy, marginal edema, BOP; GI 3 = markedly red to bluish, edematous, BOP/spontaneous bleeding. Photographs were randomized and each sextant was scored with a single value by three independent examiners and averaged to assess the GI of each sextant for the four exams (Figure 2).

Saliva samples were collected using five cotton swabs equally representing all areas of the mouth by swabbing the entire oral cavity. The collected cotton swabs were placed in microcentrifuge tubes containing 500 μl of phosphate-buffered saline (PBS, pH 7.2) and RNAlater solution (Invitrogen, Waltham, Massachusetts, USA).

Gingival crevicular fluid (GCF) and plaque samples for microbiological analysis were harvested from 12 selected sites in all four exams exam to correlate the findings. For subgingival plaque collection, a Nevi 2 periodontal scaler (SCNEVI29E2, Hu-Friedy, Chicago, Illinois, USA) was inserted to the base of the attachment, and the plaque was collected from the subgingival area and stored in RNA later at −80°C. For GCF collection, a Periostrip paper (Periopaper Gingival Crevicular Fluid Collection Strip, Fisher Scientific International, Pittsburgh, Pennsylvania, USA) was inserted into the periodontal sulci of interest and left for 30 s or until visibly saturated. The samples were collected into PBS and stored at −80°C.

Plaque samples were processed individually whereas aliquots of saliva samples were pooled together prior to analysis. Genomic DNA (gDNA) was isolated using 200 μl of the collected samples with the PureLink™ Microbiome DNA purification kit (Thermo-Fisher Scientific) according to instructions of the manufacturer.

Due to the limited sample size and nature of this pilot study, statistical data analysis will be primarily descriptive in nature with the intent of providing information to power future studies rather than to determine statistical significance. The average values were calculated for PDs, CAL, BOP, and the presence of plaque for each animal and each exam. Clinical indices for selected sites were compared with microbiological findings on an individual basis.

The periodontal examination of both animals revealed mild periodontitis at several sites (Figures 1A,B, 3A,B, 4A,B). Animal G presented with localized areas exhibiting 2–4 mm CAL, but mostly 0–1 mm in the four exams (Figure 1A). The values for CAL in mm ranged: 0–4 (average 0.49) at baseline, 0–4 (average 0.34) posttreatment, 0–2 (average 0.78) at 3 months, and 0–3 (average 0.08) at 6 months. Animal T presented with a higher prevalence and generalized pattern of sites with CAL and thus displayed higher average values for CAL (Figure 1A). The values for CAL in mm ranged: 0–5 (average 1.30) at baseline, 0–4 (average 1.39) posttreatment, 0–4 (average 1.04) at 3 months, and 0–4 (average 0.40) at 6 months. In addition, both animals presented with PD of 1–4 mm at the baseline exam with localized interproximal PD of 4 mm and localized slight radiographic bone loss which corresponds to mild periodontitis for NHP [1]. We thus concluded that the clinical characteristics pointed to an acceptable level of localized mild periodontitis to be used in our study.

Average CAL for Animal G decreased from visit 1 to visit 2, and then increased by visit 3 and again decreased by visit 4. Animal T saw an initial increase in CAL at visit 2, followed by a decrease at visits 3 and 4. Both animals were below their initial average CAL by the 6-month examination (Animal G: 0.49 vs. 0.08; Animal T: 1.30 vs. 0.40; Figure 1A)

Gingival index was summarized by sextant (upper left, upper right, lower left, lower right). The average GI was calculated by averaging across the three raters and across all pictures. The trends for GI by animal and sextant are given in Figure 1C. Data for the GI of Animal G for the upper left sextant was not available. GI was relatively steady on the upper and lower left. The upper right initially decreased for Animal G and then increased steadily at 3- and 6-month examinations. Animal T increased slightly from initial to 3-month and then decreased by the 6-month exam. The lower right demonstrated a decrease from the initial to posttreatment and then increased at the 3-month exam. Between the 3-month and the 6-month examination, Animal T remained relatively steady and Animal G had a slight increase in GI scores.

Clinical findings across the visits are summarized in Figure 1B. Further statistical analyses were not performed due to the limited number of animals. Average PD for Animal G increased from visit 1 to visits 2 and 3 and then decreased by visit 4. Animal T saw an initial decrease at visit 2, followed by an increase at visit 3 and a decrease at visit 4. Both animals were at or below their initial average PD by the 6-month examination (Animal G: 2.00 vs. 1.99; Animal T: 1.93 vs. 1.73; Figure 1D). Both animals had an initial increase in the bleeding sites (Figure 1D) from visit 1 to visit 2, followed by a substantial improvement back to baseline at visit 3. By the 6-month follow-up, Animal G had an additional decrease in bleeding sites, but Animal T saw a slight increase. Although BOP decreased, the number of sites with plaque simultaneously increased progressively for both the animals from initial examination through the 6-month follow-up (Figure 1F). Both animals began with 0 plaque sites at baseline and finished with 96 for Animal G and 95 for Animal T.

For selected sites with microbiological analysis, 12 of the 36 sites (33.3%) in Animal G exhibited BOP at baseline. Similarly, in Animal T, there were four sites with BOP out of the 36 tested (11.1%). Following amixicile treatment, the number of sites with BOP remained similar to the baseline. Three months posttreatment there were no sites with PD of 4 mm in either animal. There were fourteen sites in Animal G and eleven sites in Animal T with a PD of 3 mm. Significantly, all the sites previously measuring 4 mm were reduced to 3 mm or less at 3 months posttreatment. At the final exam, no sites measured 4 mm, and only 11 sites measured 3 mm in each of the animals. BOP remained similar to baseline at the 3-month posttreatment exam. A significant reduction in BOP was noted in Animal G with four sites (11.1%), whereas Animal T experienced a rebound in BOP by the 6-month exam with eight sites (22.2%). The above observations suggest that a 2-week treatment with amixicile resulted in improvement in the BOP in Animal G from 33.3 to 11.1% (Figure 1E) and stability of PDs in Animal G and a reduction in PDs in Animal T from 1.93 mm at baseline to 1.73 mm at the 6-month exam (Figures 1B,D), all while both animals experienced a continuous dramatic increase of detectable plaque biofilm (Animal G: 51–96 sites; Animal T: 47–95 sites) (Figure 1F).

A more focused examination in Animal G shows that 4 mm PDs present at teeth #14, 15, and 16 at the baseline had reduced to 3 mm at 6 months posttreatment with amixicile (Figure 3A). Additionally, BOP for these teeth decreased from three sites to one at 3 months posttreatment and was maintained through the 6-month exam in Animal G. Similarly, in Animal T at teeth #14, 15, and 16, the number of sites with 4 mm PD was reduced from two to zero from baseline through 6-months posttreatment (Figure 4A), and BOP decreased to only one site at 3-months, but Animal T experienced a rebound between 3 and 6 months ending with three sites with BOP.

Microbiological samples of saliva and subgingival plaque from 12 teeth on three facial sites (MB, B, DB) were collected and analyzed resulting in 36 sites per animal. Aliquots of pooled saliva collected during baseline and follow-up periodontal exams were used to isolate total DNA for 16S rRNA gene sequencing (Figure 2). Analysis of the data derived from Animal G at the genus level revealed that at baseline the dominant bacteria belonged to Streptococcus genus (Figure 2A). Abundant also were bacteria belonging to the Haemophilus, Porphyromonas, Gemella, and Fusobacterium genera. In Animal T bacteria belonging to the Neisseria genus were the most abundant ones (Figure 2C) along with bacteria belonging to the genera of Porphyromonas, Streptococcus, and Gemella. Having large proportions of anaerobic bacteria, and specifically, bacteria belonging to Porphyromonas and Fusobacterium justified the use of this model for testing of amixicile efficacy. Following treatment with amixicile, a reduction in anaerobic bacteria with a concomitant increase in aerotolerant ones was observed. Specifically, reduction in bacteria belonging to the genera Porphyromonas, Fusobacterium, and Alloprevotella was seen in Animal G. At the same time, an increase in Haemophilus, Gemella, and Escherichia was detected (Figure 2A). Streptococcus remained at the same level as in the pretreatment phase. In Animal T, we observed a reduction in the bacteria belonging to Porphyromonas (Figure 2C). However, a reduction in Neisseria was also observed. An increase in levels of aerotolerant bacteria belonging to the Streptococcus, Haemophilus, and Gemella genera were observed (Figure 2C). Overall, amixicile was effective in reducing the levels of anaerobic bacteria present in the salivary microbiome of an NHP model.

Four sites from both animals were successfully surveyed for the composition of the oral microbiome at baseline and following amixicile treatment (Figures 3, 4). A Survey of the G3 site at the genus level in Animal G (Figure 3C) indicated that at the baseline Porphyromonas and Fusobacterium were highly dominant. Other abundant bacteria included Leptotrichia, Prevotella, and Streptococcus. Following amixicile treatment, a significant reduction in Porphyromonas and Fusobacterium levels was observed. This corresponded to an increase in levels of Leptotrichia and Prevotella. Analysis of the G14 site showed that at baseline, Escherichia, Prevotella, and Streptococcus were highly abundant genera (Figure 3D). Following amixicile treatment, levels of Escherichia and Prevotella were reduced whereas those of Leptotrichia and Fusobacterium increased. Baseline abundance of the G16 site displayed high levels of Escherichia, Veillonella, Streptococcus, Leptotrichia, Prevotella, and Fusobacterium (Figure 3E). Following treatment, the abundance of Veillonella, Streptococcus, Leptotrichia, Prevotella, and Fusobacterium were reduced whereas those of Escherichia, Actinomyces, and Aggregatibacter were elevated. The fourth site of Animal G, G22, displayed a high abundance of Escherichia, Porphyromonas, Prevotella, Fusobacterium, and Alloprevotella (Figure 3F). Following amixicile treatment, the levels of Porphyromonas, Fusobacterium, and Alloprevotella were reduced with an increase in Leptotrichia, Selenomonas, Prevotella, and Actinomyces. Interestingly, levels of Streptococcus were also reduced. Overall, in Animal G we observed a reduction in the levels of anaerobic bacteria and an increase in aerotolerant ones. There was significant variation in the microbial composition of the microbiome at the baseline and posttreatment.

Analysis of the subgingival microbiomes of Animal T, both at baseline and post-treatment showed a similar trend of reduced levels of anaerobic bacteria in favor of the aerobic ones. Specifically, the abundant genera of T3 were Escherichia, Streptococcus, Porphyromonas, and Prevotella (Figure 4C). Following amixicile treatment, the abundance of Prevotella was significantly reduced. Increased levels of Escherichia, Streptococcus, Aggregatibacter, and Peptidiphaga were observed (Figure 4C). At the second site T14, Streptococcus, Porphyromonas, and Fusobacterium were the dominant bacterial genera (Figure 4D). After amixicile treatment, the levels of Porphyromonas, Fusobacterium, and Streptococcus were significantly reduced, and this corresponded to an increase in the levels of Escherichia, Aggregatibacter, and Peptidiphaga (Figure 4D). The T15 site at baseline was abundant in Escherichia, Streptococcus, Porphyromonas, and Haemophilus. Following amixicile treatment, the levels of Porphyromonas, Streptococcus, Haemophilus, and Fusobacterium were significantly reduced with an increase in the levels of Leptotrichia and Bacteroides (Figure 4E). The final site analyzed for microbiome composition, T16, had high levels of Escherichia at baseline that got reduced with a concomitant increase in the levels of Streptococcus (Figure 4F). Similarly, as in Animal G, we observed a high variability between the microbial composition of the samples derived from different sites of Animal T. The common theme from the analyzed sites was the reduction in the abundance of anaerobic bacteria and increase in the levels of aerotolerant ones.

We followed the microbiome changes of the animals up to 6-month posttreatment; two exams (3- and 6-months) were conducted where saliva and subgingival plaque from previously examined sites were collected and 16S rDNA was sequenced. The salivary microbiome of Animal G had increased in abundance in Porphyromonas, Fusobacterium, and Leptotrichia. Reduction in Haemophilus, Streptococcus, and Gemella was observed (Figure 2A). The composition resembled the baseline microbiome; however, higher levels of Leptotrichia were observed. In Animal T, the salivary microbiome had higher levels of Haemophilus and Fusobacterium after longer posttreatment intervals (Figure 2C). In the subgingival microbiome, an increase in the proportion of anaerobic bacteria was observed after longer posttreatment intervals (Figures 3, 4).

A significant clinical improvement as determined by CAL, PD, and BOP was observed at 3-month posttreatment and continued into the sixth month. These improvements were delayed with respect to the onset of reduction of anaerobic bacteria in the microbiome. At 6 months, despite some reversal of the abundance of anaerobic bacteria to the baseline levels, the clinical improvements still persisted despite the increased plaque burden.