**1. Introduction**

The individual microbial cells that constitute the human gut microbiome outnumber our cells by a factor of 10 [1]. In utero, most fetuses are free of microorganisms [2]; the first exposure babies have to microbes is during birth as they move through the birth canal, hence babies born in natural birth are inoculated with microorganisms by their mothers. The gut microbiome initiates with breast feeding and builds complexity as the baby's diet evolves from milk to other types of foods [3]. It reaches dynamic stability by the age of 3, and while a person's gut biome is relatively stable, there are many genetic and environmental factors that influence its composition and dynamic change in each person [4]. The microorganisms living in our gut that do not cause harm, and may even have a beneficial contribution to human health, are called commensals. The gut microbiome composition can be disrupted by dietary changes, antibiotic treatment or pathogenic infections, and in reciprocal interactions changes to the composition and abundance of commensals could affect the entire system by producing unwarranted gastrointestinal and immune diseases [5–7].

Published literature suggests that the gut microbiome has an indirect effect on bone through changes to the immune system and inflammatory cytokines [8,9]. Commensals aid in immune-regulation by releasing microbial associated molecular patterns (MAMPs) such as lipopolysaccharides (LPSs) that bind and activate toll-like receptors (*Tlr*); LPSs have been shown to bind to *Tlr4* [9–14]. This activation causes an inflammatory cascade that releases inflammatory cytokines and interferons which act as transcription factors to induce naïve immune cells to mature [15–17]. Studies have suggested that gut biome dysbiosis can promote aggressive bone destruction mediated by osteoclasts due to an increase in tumor necrosis factor alpha (TNF-α) [18–20]. Furthermore, we have previously shown that elevated levels of LPSs negatively impact bone by promoting bone loss and accelerate post-traumatic osteoarthritis (PTOA) development. We also showed that LPS administration prior to injury elevates *Tlr5*/*7*/*8* transcription in the joint [10,21]. TNF-α promotes osteoclastogenesis by increasing RANK-L expression in bone marrow cells and therefore elevating the number of osteoclast precursor cells [20,22–24].

In the context of PTOA, it has been shown that when the gut microbiome of obese mice is modified by supplementing oligofructose, OA phenotypes diminish, which correlates with a reduction in the levels of inflammation in the colon and cytokine levels circulating in the serum and present in the knee [25]. Cyclic compressive loading in mice on a high-fat diet have promoted more severe PTOA phenotypes than mice on a normal diet, while *Tlr5-*/*-* mice treated with ampicillin and neomycin have shown improvement in the cartilage phenotype post injury [26]. Germ-free mice have also been shown to have a better OA outcome after destabilization of the medial meniscus, and modifications to the gut microbiome have improved PTOA phenotype in obese mice [25,27,28]. Tibial compression induced injury in 20-week-old germ-free mice has shown an increase in bone volume [29]. Therefore, a precedent exists in support of the gut microbiome composition as a potential risk factor for the development of PTOA, but additional studies are required to elucidate potential mechanisms that contribute to the unwarranted PTOA phenotypes.

Most PTOA-related studies to date have examined factors likely to exacerbate or accelerate the development of osteoarthritis post injury, if administered at the time of injury or shortly thereafter [30]. Since bone and cartilage sometimes exhibit an inverse relationship to insult, such that what is anabolic for bone is catabolic for cartilage and vice versa [31], we speculated that gut biome modifications would slow down or improve PTOA outcomes. Therefore, this study aimed to examine how partial elimination of the gut microbiome through antibiotic treatment prior to injury would influence PTOA outcomes. Studying the effects of medication administered before an injury is of high biomedical importance because, clinically, most concerns are centered on side-effects due to co-administration. Currently, however, standards of medical care do not consider gut biome status, nor is gut dysbiosis a recoded clinical parameter. Research that can show the prognostic and diagnostic value of gut biome status could potentially lead to new standards of care. In addition, antibiotics are widely prescribed to teens and young adults who may be active in sports and therefor more susceptible to joint injuries. According to the CDC, in 2016, 64.9 million oral antibiotic prescriptions were issued to people under the age of 20, the equivalent of 790 per 1000 people; therefore, gut dysbiosis may be more common than expected in young athletes suffering an articular injury [32]. As the population of the USA ages there will be an increase in PTOA cases; studying how antibiotics modify PTOA phenotypes will be helpful to finding preventative treatments in the future for both young and old patients.

#### **2. Results**

#### *2.1. Antibiotic Treatment Prior to Injury Delays Cartilage Resorption in Injured Joints*

Using an established, noninvasive, tibial compression PTOA mouse model [33–37], we examined whether a six-week course of antibiotics (ampicillin (1.0 g/L)/neomycin (0.5 g/L)) [26] would impact OA outcomes, post injury. C57Bl/6J mice were examined histologically at six weeks post injury. Examination of the uninjured, contralateral femoral heads revealed a more intense Safranin-O staining

throughout the articular cartilage of the antibiotic (AB)-treated group compared to untreated controls (VEH), but both AB and VEH joints displayed normal morphology (Figure 1A,C). A slightly less mineralized area, characterized by large pockets of bone marrow, was observed in the femoral condyle of AB-treated, injured joints, relative to the VEH injured group (Figure 1B,D). Consistent with the AB uninjured control, the femoral condyle of the injured AB group (Figure 1d; arrow, asterisk) appeared to also have an increase in Safranin-O staining intensity as well as a thicker articular cartilage layer than the injured VEH, suggesting higher levels of proteoglycans and reduced chondrocyte apoptosis in the injured joints of AB-treated animals (Figure 1b; arrow, asterisk). The meniscus (Figure 1bb,dd; arrow, asterisk) of injured AB joints showed a thicker hyperplastic morphology with enhanced cellular infiltration. The meniscus in the injured VEH group also showed cellular infiltration but at a significantly lower level than the AB injured group. Examination of the sagittal views of the joints by a modified Osteoarthritis Research Society International (OARSI) grading scale determined that AB-treated injured joints had a significantly lower cartilage score than VEH-treated injured joints with a *p*-value of 0.038 (Figure 1E). These results imply that modifying the gut microbiome through the administration of an ampicillin/neomycin antibiotic cocktail prior to injury was sufficient to improve the cartilage phenotype subsequent to trauma, and reduce PTOA outcomes.

**Figure 1.** Characterization of post traumatic osteoarthritis (PTOA)-associated structural changes of antibiotic-treated animals in the knee. (**A**) Histological evaluation of vehicle (VEH) contralateral, (**B**–**bb**) VEH injured, (**C**) Antibiotic treated (AB) contralateral, and (**D**–**dd**) AB injured joints conducted at six weeks post injury using Safranin-O and Fast Green staining (scale bars indicate 200 mm). High magnification images corresponding to yellow boxes (**B**,**D**) are provided b, bb, d, dd. (**E**) PTOA severity was quantified using a modified OARSI scoring system (\* *p*-value < 0.05).

#### *2.2. Antibiotic Treatment Has a Negative E*ff*ect on Bone, Post Injury*

The bone phenotypes of AB- and VEH-treated mice were characterized by micro-computed tomography (μCT) to quantify subchondral trabecular bone mass and osteophyte volume at six weeks post injury. Consistent with prior published results, VEH injured joints had significantly less subchondral bone volume (BV/TV) by ~17.28% and ~11.31% when compared to contralateral and uninjured controls, respectively (Figure 2A). Antibiotic-treated injured joints lost 13.67% when compared to contralateral AB treated and 17.42% when compared to uninjured AB-treated controls. The subchondral bone volume (BV/TV) fraction of AB group had ~0.2%, ~7.1%, and ~10.9% lower BV/TV than the VEH group when comparing the uninjured, injured, and contralateral groups; the contralateral group was the only one that was statistically significant, suggesting that AB treatment does not elevate BV/TV in the uninjured leg (Figure 2A). Trabecular number (Tb.N) of the AB group had ~0.2% and ~0.4% higher Tb.N on the uninjured and injured groups, respectively, compared to the VEH; the contralateral had ~1.1% lower Tb.N on the VEH. Tb.N was not statistically significant (Figure 2B). Trabecular thickness (Tb.Th) showed the VEH group had ~2.9%, ~7.6%, and ~10.6 higher Tb.Th than uninjured, injured, and contralateral AB cohorts; injured and contralateral were significant (Figure 2C). Trabecular spacing (Tb.Sp) AB showed ~1%, ~0.4%, and ~2.2% larger Tb.Sp than the VEH uninjured, injured, and contralateral groups, respectively; none were statistically significant (Figure 2D). The VEH group had ~39.7% significantly higher osteophyte volume (Op.V) compared to

the AB cohort (Figure 2E). Visual representations of Op.V showed a larger amount of osteophytes in VEH joints, consistent with the quantification data (Figure 2F).

**Figure 2.** Bone phenotype of antibiotic-treated injured mice. (**A**) Subchondral trabecular bone volume fraction (BV/TV) of the distal femoral epiphysis. (**B**) Trabecular number was measured using the average number of trabeculae per unit length. (**C**) Trabecular thickness was measured using the mean thickness of trabeculae assessed using direct 3D methods. (**D**) Trabecular spacing was measured using the mean distance between trabeculae, assessed using direct 3D methods [38]. (**E**) Osteophyte volume at six weeks post injury. (**F**) Osteophyte imaging using μCT. (\* *p* < 0.05).
