**3. Discussion**

The role of chronic antibiotic treatment prior to injury on the development of PTOA has not been previously investigated. Previous studies associating gut biome changes with skeletal phenotypes have focused primarily on bone, and have shown that the gut microbiome can influence osteoclastogenesis and affect bone volume [23]. Previous studies have also described the changes in the gut microbiome and bone using the same antibiotic cocktail [26,72]; however, our study is the first to examine how depletion of the gut biome with an ampicillin/neomycin antibiotics cocktail affects the development of osteoarthritis following a traumatic joint injury. Although morphologically our data indicate increased cellular infiltration in the synovium of injured joints of antibiotic-treated cohorts, reminiscent of synovitis induced by LPS administration [10], AB treatment did not have a significant effect on the BV/TV of the injured joint, resulting in no significant changes in BV/TV between the injured VEH and injured AB groups. The only significant bone phenotype was observed when contralateral were compared, where VEH-treated mice had an increase in BV/TV relative to uninjured control AB-treated mice had a slight, but significant decrease in BV/TV. However, AB treatment improved the cartilage phenotype as reflected by a significantly lower OARSI score in these mice. There are several possible explanations for this outcome. The cellularity was examined using Ly6G and Ly6C markers, which highlighted an increase in the presence of monocytes, neutrophils, and granulocytes. In order to increase specificity, we stained with M1 and M2 markers, which showed an increase in the frequency of pro-inflammatory macrophages in the LPS-treated joints, while more anti-inflammatory macrophages were present in the AB-treated group. This correlation supports the conclusion that different macrophage subtypes could have influenced the exacerbation of PTOA in the LPS-treated cohorts, while deaccelerating PTOA progression in AB-treated cohorts.

It is also possible that chronic antibiotic treatment in juvenile mice (starting at four weeks of age) is anabolic to the articular cartilage, and during the six weeks of AB treatment prior to the injury the cartilage of these animals produced significantly more extracellular matrix. Furthermore, the cartilage of AB-treated animals may also display slightly different biomechanical properties than VEH controls. This theory is in part supported by the observation that uninjured joints of AB-treated animals stained more intensely with Safranin-O than the VEH uninjured joints (Figure 1A,C). If the cartilage of AB-treated animals acquired different mechanical properties, including increased stiffness or increased elastic modulus, cartilage degradation in these mice may have proceeded at a slower pace than the VEH, accounting for the milder phenotype in these injured joints. Additionally, we see a downregulation of gene *Rspo1* in injured AB joints when compared to injured VEH. This could be the reasoning for the decrease in OA progression, as *Rspo1* has previously been shown to have a role during OA progression [71]. Future studies examining the biomechanical properties of cartilage at different ages and in different treatment groups will have to be conducted to confirm whether significant differences in these properties exist.

Alternatively, the observed PTOA outcomes may be driven primarily by molecular changes. We observed 113 genes, including inflammatory genes *Bmper*, *Ccl2*, *Ccl7*, *Ccl8*, *Cxcl5*, *IL6*, *IL11*, *IL33*, and *Cxcl10,* that overlapped between the injured VEH, LPS, and AB groups and the uninjured groups of their respective treatments with significantly elevated expression in the injured joints; however, the blunted effect was present only in the AB injured group, suggesting that these molecules have a less potent effect in when mice are treated with AB prior to injury. Gene expression data indicate that the inflammatory genes *Ptgs2, Reg3g,* and *Serpine1* were downregulated in uninjured AB-treated mice, while *Tbx21* was found to be downregulated in injured AB joints. Two inflammatory genes found to be downregulated in both the injured and uninjured AB groups as compared to the corresponding VEH were *Tafa3* and *Cntfr*, which are both associated with the immune and nervous system. These genes have not been studied in the context of PTOA, and their influence on injury outcome would be interesting to study. When examining molecular changes in the immune system in prior reports, we have shown that elevated and persistent immune activation accelerates osteoarthritis phenotypes post injury [10]. Furthermore, we have shown that systemic LPS administration five days prior to injury negatively impacts PTOA, resulting in a more severe phonotype [10]. We have also shown that LPS-treated mice display highly elevated levels of toll-like receptors 5, 7, and 8 (*Tlr5*, *Tlr7*, *Tlr8*), and we hypothesized that the enhanced PTOA phenotype in LPS-treated mice may be due in part to increased signaling through these receptors. One complementary transcriptional result derived from the RNA-seq data examined herein is the discovery that *Tlr5* is significantly downregulated in AB-treated uninjured joints. Kim et al. showed that *Tlr5* in rheumatoid arthritis promotes monocyte presence and osteoclast formation due to the cross regulation of the *Tlr5* and TNF-α pathways [73]. If *Tlr5* similarly modulates the expression of inflammatory genes that are directly involved in cartilage degradation post injury, the observed transcriptional suppression of *Tlr5* in uninjured joints may promote a molecular resistance to inflammatory cytokines. Future studies will have to evaluate whether *Tlr5* receptor antagonists can prevent or slow down the development of OA post ACL rupture.

The current literature presents conflicting evidence on the effects of antibiotic treatment and the gut microbiome on bone, showing that modifications may have no significant changes in BV/TV after injury when compared to untreated control mice, but have a lowering effect when compared to uninjured treated mice [26]. In germ-free mice, AB treatment can increase the BV/TV after injury compared to injured controls, but these mice also have a lower BV/TV when compared to uninjured germ-free mice [29]. Our results showed no changes in BV/TV on injured AB-treated mice when compared to the injured VEH, but did show a decrease in in BV/TV when compared to the AB contralateral. These results are similar to the decrease in BV/TV observed after tibial compression (TC) injury in strains that are resistant to PTOA, like *MRL*/*MpJ* [36,74]. Our AB group showed an improvement in cartilage, and though these mice are not OA-resistant to the extent of *MRL*/*MpJ* mice, we observed similarities in the expression of T-cell markers like *Cd3d* and *Cd8b1* [36]. Although inflammation is

resolved quicker in *MRL*/*MpJ* mice, bone resorption still occurs, similar to the C57Bl6 strain, due to the presence of pro-inflammatory cytokines and matrix metalloproteinases (MMPs) that promote osteoclastogenesis and extracellular matrix degradation [75–77]. The similarity of T-cell markers could potentially enhance healing and accelerate inflammation resolution in AB-treated mice, which could diminish the PTOA phenotype.

Our study uniquely examines the impact of long-term antibiotic treatment on OA outcomes subsequent to joint trauma. Prior to this study, we did not know if there were any PTOA changes caused by the administration of antibiotics. While we found that this particular AB regime had a beneficial effect on the health of injured and uninjured joints, it still remains to be elucidated whether short term AB treatment can be prophylactic; most importantly, questions remain about whether AB treatment post injury would have the same beneficial effect. This study highlights the importance of how the body works as a system and how systemic and local factors present prior to injury can significantly impact how our body heals and responds to trauma. This study highlights the importance of the gut biome in modulating PTOA, specifically by affecting toll-like receptors transcriptional levels that may in turn influence PTOA outcome after injury.
