A Pilot Study of Mitochondrial Response to an In Vivo Prosthetic Joint Staphylococcus aureus Infection Model

Abstract

Prosthetic joint infections (PJI) are associated with orthopaedic morbidity and mortality. Mitochondria, the “cell’s powerhouses,” are thought to play crucial roles in infection response and in increased risk of sepsis mortality. No current research discusses PJI’s effect on mitochondrial function and a lack of understanding of immune-infection interactions potentially hinder patient care. The purpose of this pilot study was to evaluate the impact of simulated PJI on local tissue mitochondrial function. Using an established prosthetic implant-associated in vivo model, tissues were harvested from the surgical limb of a methicillin-sensitive Staphylococcus aureus implant-associated infection group (n=6) and compared to a non-infected group (n=6) at postoperative day (POD) 21. Using mitochondrial coupling assays, oxygen consumption rate and extracellular acidification rate were assessed in each group. Electron flow through mitochondrial complexes reflected group activity. Electron Paramagnetic Resonance (EPR) spectrometry measured the oxidizing potential of serum samples from infected versus non-infected groups. On POD21, colony forming units per gram of tissue showed 5×10⁹ in the infected group and 10¹ in the non-infected group (p<0.0001). Maximal respiration and oxygen consumption due to ATP synthesis were significantly lower in isolated mitochondria from infected limbs (p=0.04). Both groups had similar complex I, III, IV, and V activity (p>0.1). Infected group EPR signal intensity reflecting reactive oxygen species levels was 1.31±0.30 compared to 1.16±0.28 (p=0.73) in the non-infected group. This study highlights PJI’s role in mammalian cell mitochondrial dysfunction and oxidative tissue damage, which can help develop interventions to combat PJI.

Introduction

A successful joint replacement alleviates pain, restores function, and enhances patient quality of life, thereby improving the lives of millions of people each year. Nearly one million total hip arthroplasties (THAs) and total knee arthroplasties (TKAs) are performed annually in the US [1] which is expected to double by 2030 [2].

Although a common and effective surgical procedure, joint arthroplasty has risks, including prosthetic joint infection (PJI) which affects the prosthetic joint and surrounding tissue. In fact, PJI is a leading cause of orthopaedic morbidity (12% major complication rate), mortality (7–10% at one year) [3], and significant healthcare expenses (over $1.85 billion). PJI numbers are expected to increase in parallel with the projected increase in the number of THA and TKA procedures [2].

Approximately 80% of PJI are caused by Gram-positive organisms like Staphylococcus aureus (S. aureus); treatment failure rates can approach 50% in some cases[4]. Treatment failure is likely multifactorial but there is a limited understanding of PJI’s effects on surrounding tissues such as its effect on mitochondrial function.

Mitochondria are referred to as the “cell’s powerhouses” and are regarded as significant organelles that produce energy, oxidize fatty acids, regulate cell cycle, and help regulate cell death. When not functioning properly, multiple reports have demonstrated an increased risk of mortality from sepsis [5, 6]. Mitochondria’s role in the setting of PJI may be important due to the dynamic nature of this organelle, which is believed to be descended from endosymbiotic Rickettsia-type α-proteobacteria [7]. Thus, this hypothesized bacterial origin of mitochondria carries unique implications when considering PJI and the impact treatments may have on local tissues. Several studies have reported alterations of mitochondrial function due to infection. In fact, various pathogens have evolved methods of targeting the mitochondria to affect intracellular survival, dispersal of bacteria by mediating mitochondria-induced cell death, or evasion of host immunity [8].

Given high one- and five-year mortality rates associated with PJI, as well as the severe pain, functional difficulties, and poor quality of life associated with this condition, better treatments than those currently available are needed [9–11]. It is crucial to first comprehend mitochondrial function in the context of PJI given its central role in cellular regulatory functions and in generating adenosine triphosphate (ATP) in order to potentially mitigate changes occurring at the level of this organelle and thereby ameliorate the condition and its associated poor quality of life outcomes by enhancing the function of the cell’s central energy machine.

Currently, there is no research investigating the influence of PJI on mitochondrial function and the interplay between the immune system and the infecting organism, posing a significant barrier to improving treatment outcomes for these patients. Our hypothesis was that PJI would have deleterious effects on mitochondria, decreasing their role as a cell’s central energy machine. The purpose of this pilot study is to investigate the techniques to determine PJI’s effect on mitochondrial function and, ultimately, to asertain the effect of simulated PJI on mitochondrial function.

Materials and Methods

Study Overview

In this pilot study, we used a well-established in vivo model associated with a Kirschner wire (K-wire) (Smith & Nephew, Memphis, TN) implant based on our prior experience and previous literature [12, 13]. Sprague-Dawley rats obtained from Hilltop Lab Animals, Inc. (Scottdale, PA), with an average weight of 433g for males and 266g for females, aged 12 weeks, were randomly assigned to one of two groups: a methicillin-sensitive S. aureus (MSSA) (ATCC 25923, American Type Culture Collection, Manassas, VA) implant-associated infection group (n=6) and a non-infected implant group (n=6) based on our prior experience and previous literature [12, 13]. A total of twelve animals were included; no additional inclusion or exclusion criteria were set. Block randomization was used in this pilot study with equal distribution of male and female rats per group. Twenty-one days after surgery, tissues and blood were collected from all animals. Whole tissues were homogenized in ATP medium buffer for ATP coupling and Electron Transport Chain (ETC) Complex activity [14, 15]. Before analysis, isolated mitochondria [16] were quantified using the Bradford protein assay. Fresh tissues were used for the ATP coupling assay and frozen tissues were used for the ETC Complex activity analysis. The Animal Care and Use Committee approved the animal protocols (Approval no. 1803013294).

Animal Model and Surgical Procedures

All animals were housed in a vivarium free of specific pathogens. In this pilot study, two rats were housed together in cages with a plastic base and wood shaving bedding, with a temperature and humidity range of 20⁰ to 23⁰C, and a relative humidity of 50%, on a light-dark cycle of 12 hours. There was free access to food and water. Following acclimation, the rats were sedated with isoflurane (3–5%). During all surgical procedures, sterile techniques were utilized. As highlighted in Prince et al.[12], following anesthesia, the right hindlimb was shaved for surgery. Following the application of povidone-iodine and alcohol swabs to sterilize the surgical site and the placement of a sterile drape, a 3 cm longitudinal skin incision was made over the knee, exposing the knee joint. A hole was drilled into the medullary cavity of the femur. In both groups, a 3 cm (male) or 2 cm (female) by 1 mm stainless steel K-wire was inserted into the femoral canal. The medullary cavity of the infected group was inoculated with 20 μl of a standardized inoculum containing 10⁵ colony forming units (CFU)/ml of methicillin-sensitive S. aureus (ATCC 25923) prior to wound closure. Based on previous models, this dose was chosen to induce a chronic localized infection while avoiding systemic sepsis [13]. For all groups, the exposed joint was closed with 2–0 Webcryl (Patterson Veterinary, Devens, MA) and staples, followed by the external application of Vetbond tissue adhesive (3M, St. Paul, MN). The analgesic Buprenorphine SR (ZooPharm, Windsor, CO) was administered subcutaneously to provide 72h of pain relief [17]. Following incision closure, radiographs were taken on the day of surgery and 21 days later to ensure proper K-wire placement. Radiographs were classified based on Aktekin et al. [18]. Protocols and monitoring for postoperative pain were carried out in accordance with institutional guidelines.

Tissue Collection

Animals were anesthetized and euthanized 21 days postoperatively as described previously. The knee was prepared and draped in the same manner as the initial surgery. Skeletal muscle was collected after tissue was carefully dissected from fascial attachments. Muscle samples were obtained from the operative knee joint [12]. Once samples were collected, the animals were euthanized; cardiac puncture was performed, blood was collected, and serum was separated by centrifugation at 1500xg for 10min, followed by the administration of Euthasol (Virbac AH, Westlake, TX).

Clinical Parameters

Bacterial Culture and Identification

At euthanasia, tissue samples from the operative site of the infected (n=6) and the non-infected groups (n=6) were obtained in a sterile fashion. The presence or absence of S. aureus infection was determined by culturing samples using established methods [19] to validate the comparability between groups.

Weight Change during Course of Study

All rats (n=12) were weighed at baseline, POD10, and at euthanasia (POD21) using a digital scale.

Mitochondria Evaluation

ATP Coupling Assay

At euthanasia, tissue samples (resected skeletal muscle) from the surgery site of the infected (n=3) versus the non-infected groups (n=3) were collected, stored in test tubes containing three milliliters of ATP medium, and analyzed. Due to time constraints and the need for fresh tissues, analysis was conducted on three rats from each group. Whole tissues (approximately 50 mg of resected skeletal muscle) were chopped in Proteinase Medium and homogenized in ATP medium buffer. Samples were centrifuged and supernatant was collected and centrifuged again [14, 15]. Using the Bradford protein assay, the protein content of isolated mitochondria was quantified [16]. After preparation, the pellet was suspended with MAS buffer and used for the activity assays. The ATP coupling assay was utilized to evaluate mitochondrial function and health. This assay was undertaken using the Agilent Seahorse Bioanalyzer (Seahorse Bioscience XFe 96, Agilent, N. Billerica, MA) in accordance with the manufacturer’s instructions to collect the oxygen consumption rate (OCR) and extracellular acidification rate (ECAR). Following the Wilson et al. method [20], the mitochondrial respiratory parameter data for each biological replicate were determined by averaging the OCR values at each timepoint for each biological replicate’s technical replicates. To compensate for instrument error, raw OCR data were normalized by setting the most negative value among all successful wells to zero. Basal respiration was recorded prior to ADP injection (state 2) and ADP stimulated respiration or maximal respiration (State 3) was obtained by the highest OCR value of the two timepoints immediately after the injection of ADP. FCCP-stimulated respiration (State 3u) was obtained by the highest OCR value of the two timepoints immediately after the injection of FCCP. Oxygen consumption due to ATP synthesis was calculated by subtracting non-ADP stimulated respiration (State 4o) taken by the lowest OCR value of the two timepoints immediately after the injection of oligomycin from maximal respiration. Thus, the final concentrations of seven injections were as follows: 1) ADP (4 mM), 2) oligomycin (4mM), 3) FCCP (4 mM), and 4) rotenone (4 mM) antimycin-A (4 mM). The cycle of mixing, waiting, and measuring lasted 30s, 30s, and 2min, respectively. Prior to the first injection, and after each injection, two cycles were executed.

Mitochondrial ETC Complex Activity

Tissue samples from the operative site of both groups (n=6 in each group) were collected at euthanasia. To remove blood and debris, samples were washed immediately with cold 1X phosphate-buffered saline. Using a scalpel, scar and connective tissues were removed from the surgical site and frozen at −80⁰C. At analysis, mitochondria were isolated from resected skeletal muscle tissue samples as described above to obtain the final pellet. Finally, the pellet was suspended with MAS buffer and used for the activity assays. The activities of ETC complexes I, III, IV, and V were measured using protein homogenates (ATP synthase) following the Kunovac et al. method [21]. Briefly, the activities of ETC complexes I and III were determined by measuring the reduction of decyl ubiquinone (I) and cytochrome c (III) in tissue samples. Complex V activity was determined by measuring oligomycin-sensitive ATPase activity via pyruvate kinase and phosphoenolpyruvate. Complex IV activity was determined by measuring the oxidation of reduced cytochrome c. The microplate reader was a Molecular Devices Flex Station 3 Multi-Mode microplate reader (Sunnyvale, CA). All assays were measured spectrophotometrically in accordance with previous research [22–24]. As mentioned previously, the protein content of all samples was normalized using the Bradford method and bovine serum albumin protein assay standards [16]. The final values were reported as unit/nanogram (I-IV) or milligram (V) of protein, where unit=nanomoles of oxidized substrate (minute −1).

Electron Paramagnetic Resonance (EPR) Spectroscopy

Serum samples isolated from blood collected at euthanasia from infected (n=6) versus non-infected groups (n=6) were frozen and stored at −80°C. Our approach for measuring the oxidizing potential of rat serum samples using EPR spectrometry and the spin probe 1-hydroxy-3-carboxymethyl-2,2,5,5-tetramethyl-pyrrolidine (CMH) (Enzo Life Science, Farmingdale, NY) was comparable to that described by Majumdar et al. and Velayutham et al.[25, 26]. Serum (0.1 mL) was incubated with CMH at 37°C for 30 minutes (0.2 mM). Serum oxidants oxidize CMH (EPR inactive) to 3-carboxymethyl-2,2,5,5-tetramethylpyrrolidinyloxy radical (CM; EPR active). After incubation, samples were frozen and stored at −80°C. At the time of EPR measurements, frozen samples were thawed, loaded (50 μL) into glass capillary tubes, and placed inside a 4 mm (O.D.) EPR quartz tube which was placed in the EPR cavity/resonator. Room temperature EPR was collected; EPR spectra were acquired using a Bruker ELEXSYS E580 spectrometer (Billerica, MA) operating at X-band with 100 kHz modulation frequency. The EPR instrument parameters were 9.854 GHz, 100 G sweep width, 18.88 mW microwave power, 1 G modulation amplitude, 60 dB receiver gain, 29.3 ms conversion time, 30 seconds sweep time, and 1 scan. EPR data collection used Bruker Xepr software.

Statistical Analysis

GraphPad Prism version 9 software was used for analysis (GraphPad Software Inc., San Diego, CA). For ROS and weight outcomes, comparisons were made using an unpaired t-test assuming normal distribution to determine significant differences between groups. For weight changes across the three timepoints (POD0, 10, and 21), comparisons were made using two-way ANOVA with Tukey’s post-test to determine significant differences between groups. Data are expressed as mean ± standard error of the mean (SEM). The number of animals in each group was based on prior experience with the infection model.

Results

Clinical Parameters

Bacterial Culture and Identification

On POD21, the operative knee bacterial load was measured in the infected group (5×10⁹ ± 5.3×10⁸ CFU/g) versus the non-infected group (10¹± 10¹) (p<0.0001) to confirm the presence or absence of S. aureus infection. Bacterial load in the operative knee for the two groups can be found in Table 1.

Table 1:

Clinical Parameters at POD21

	MSSA-Infected Group	Non-Infected Group	p-value
Bacterial Load (CFU/g)	5×109 ± 5.3×108	101± 101	<0.0001
Mean Weight POD10 (g)	308.2±32.98	323±28.57	0.74
Mean Weight POD21(g)	329±34.27	341±32.67	0.80
Mean Weight Change between POD0 and POD10 (g)	34.8±7.28	33±6.77	0.0006
Mean Weight Change between POD0 and POD21(g)	14±5.99	15± 2.67	0.0114
Mean Weight Change between POD10 and POD21 (g)	20.8±1.29	18±4.1	0.0025

Note: The bacterial load is shown as colony-forming units (CFUs) per gram of tissue and reported as mean ± SEM. The weight at different timepoints is reported as mean± SEM. Comparisons within each group throughout the three different timepoints were identified by by repeated measures ANOVA. Significance is considered at p < 0.05.

Weight Change during Course of Study

At the start of this pilot study, the mean body weight for the infected group was 343±40.26g and 356±35.34g (p=0.81) for the non-infected group. At POD10, mean body weights of the infected and non-infected groups decreased to 308.2±32.98 and 323±28.57g, respectively (p=0.74). On POD21, the average body weight increased relative to POD10, but remained lower than the initial weight in both groups. On POD21, the mean body weight of the infected group was 329±34.27 compared to 341±32.67 (p= 0.80) in the non-infected group. At each of the three timepoints, the differences between groups were not statistically significant. Comparing weights across timepoints within each group was statistically significant (p<0.05). The decrease in weight at POD10 relative to the value at baseline, followed by an increase at POD21, was deemed significant in both groups (Table 1).

Infection Impact on Mitochondria

ATP Coupling Assay

To evaluate overall mitochondrial health and function in both groups, we performed an ATP coupling assay in isolated mitochondria from the surgery site to assess Oxygen Consumption Rate (OCR), Extracellular Acidification Rate (ECAR), State 2, State 3, State 3u, and oxygen consumption due to ATP synthesis. Overall OCR was lower in the infected group versus the non-infected group (Figure 1A); however, ECAR was higher in the infected group versus the non-infected group (Figure 1B). As for the respiration states, the mean of state 2 respiration was not statistically different between the infected group (23.80± 26.96) versus the non-infected group (65.15± 19.30) (p= 0.28) (Figure 1C). The mean state 3 respiration was significantly lower in the infected group (73.60± 40.23) versus the non-infected group (576.5± 172.3) (p=0.04) (Figure 1D) whereas the mean state 3u respiration was not different between groups (p=0.06) (Figure 1E). The mean of oxygen consumption due to ATP synthesis was significantly lower in the infected group (73.46± 41.45) versus the non-infected group (566.9± 165.6) (Figure 1F) (p=0.04).

Figure 1.

ATP coupling assay analysis on isolated mitochondria from the surgical site of the MSSA-infected model versus the non-infected model using the Seahorse XF bioanalyzer.

1a. Oxygen Consumption Rate (OCR) in MSSA-infected model versus the non-infected model

1b. Extracellular Acidification Rate (ECAR) in MSSA-infected model versus the non-infected model

1c. State 2, also known as Basal Respiration in MSSA-infected model versus the non-infected model.

1d. State 3 also known as Maximal Respiration in MSSA-infected model versus the non-infected model.

1e. State 3u in MSSA-infected model versus the non-infected model

1f. Oxygen consumption due to ATP synthesis in MSSA-infected model versus the non-infected model

Abbreviation: MSSA: Methicillin-sensitive Staphylococcus aureus; AKA: Also known as.

Mitochondrial ETC Complex Activity

ETC complex activity assessments were completed (Figure 2). Complex I activity was 309.1±26.98 in the infected group and 374± 55.25 in the non-infected group (p=0.39); Complex III activity was found to be 917.1 ±167.2 in the infected group and 872.4±86.58 in the non-infected group (p=0.79). Complex IV activity was found to be 500.9±99.49 in the infected group and 566.5±59.40 in the non-infected group (p=0.56) while Complex V activity in the infected group was 263.4 ±14.22 versus 334.7±33.17 in the non-infected group (p=0.14). For complexes I, III, IV, and V, analysis revealed no significant differences between groups.

Figure 2:

Electron Transport Complex Activity for complexes I, III, IV and V.

Reactive Oxygen Species (ROS)

At POD21, ROS, a subset of oxygen-containing free radicals, was measured using electron paramagnetic resonance on samples (EPR) in both groups. In the infected group, the mean EPR signal intensity (arb units) reflecting the ROS levels was 1.31± 0.30 compared to 1.16± 0.28 in the non-infected group (p=0.73). ROS levels were not significantly higher in the infected group.

Discussion

PJI is a challenging complication following joint arthroplasty and is associated with morbidity and mortality. Improving the treatment success of PJI is crucial; therefore, it is imperative first to gain a deeper understanding of the impact infection can have on local tissues to better guide these efforts. Given that mitochondria are crucial for proinflammatory signaling, it is likely they play an important role in the regulation of innate immunity and the inflammatory response in the setting of infection.

While there was no difference in complex activity and EPR signal intensity reflecting ROS levels between groups, the Oxygen Consumption Rate (OCR) and Extracellular Acidification Rate (ECAR) differed significantly, with a lower OCR and higher ECAR in the infected group compared to the non-infected group. Further analysis of OCR showed significantly lower maximal respiration and oxygen consumption due to ATP synthesis in the infected group.

Mitochondria generate energy in the form of adenosine triphosphate (ATP) and heat and participate in the apoptosis-signaling pathway. Mitochondrial DNA (mtDNA) are not protected by histones [27], and therefore lack the structural protection and repair mechanisms of histones, leading to the possibility of damage. In addition, Solomon et al. showed that infection, particularly sepsis, has been linked to changes in mitochondrial shape and function [22]. This sepsis-induced metabolic dysfunction due to mitochondrial damage can cause oxidative stress, muted immune response, end organ damage, and death [5, 6]. In fact, mitochondria could be a target of oxidative stress-induced damage along with membrane phospholipids, respiratory chain complexes, proteins, and mtDNA which may lead to decreased cellular energy, increased electron leakage, and oxygen reduction in some cells [28].

In the context of PJI, our findings suggest that general mitochondrial dysfunction may be present. Initially, this mitochondrial dysfunction was hypothesized based on the clinical manifestations of PJI, including muscle weakness and pain. Given mitochondria’s role in providing energy to cells, their improper function in an infected prosthesis would explain these clinical symptoms. We evaluated how simulated PJI affects mitochondrial function using an established MSSA implant-associated infection group versus a non-infected implant group.

At POD21, the ROS levels in serum samples of the infected group were higher compared to the non-infected group; however, the difference was not statistically significant. ROS are frequently produced as a byproduct of aerobic respiration by the pathogen during an infection, but they can also be found in the host environment [23]. The “two-edged sword” [29] nature of ROS in pathogen clearance is a common theme in literature. Antibiotics and the host immune system are both thought to use ROS, which some authors have linked to damage to DNA, lipids, and proteins. Yet a number of researchers have cast doubt on this widely held belief and proposed instead that certain pathogens actually use ROS to their own survival advantage [30]. Also, prior literature has suggested that neutrophils initially phagocytose bacteria and subsequently use intracellular ROS and proteases to kill bacteria in the phagolysosome. It has long been believed that neutrophil ROS production is due to the activity of nicotinamide adenine dinucleotide phosphate (NADPH) oxidases [31]. Therefore, neutrophils with healthy mitochondria are an important part of the antimicrobial defense against S. aureus infection.

In our analysis, at POD21, the infected group had a mean CFU/g of 5×10⁹ ± 5.3×10⁸ CFU/g compared to 10¹± 10¹ in the non-infected group indicating the persistence of infection. Persistent infection demonstrates the inability of the immune system to completely eliminate the infection which agrees with findings by Snary et al. [24] about the inhibition of neutrophil mitochondria at various mitochondrial targets. Another factor explaining the lack of increase in the ROS level or limited decrease in CFU/g would be a temporal relationship. The samples were taken at POD21 at which point the host enzyme is dedicated to ROS production and NADPH oxidase and could no longer be activated as shown by Boncompain et al. [32] highlighting the production of ROS turned on and rapidly shutting down in Chlamydia Trachomatis infected epithelial cells. Garrabou et al. [33], highlighted that mitochondrial dysfunction enhances ROS production, especially when complexes I and III are impaired, possibly providing rationale why ROS production was not significantly higher in the PJI group when compared to the non-infected group given the fact that complex activities I, III, IV, and V were still intact.

Focusing on ETC complex activity, the inner mitochondrial membrane (IMM) contains the different complexes of the respiratory electron transport chain (complexes I–IV) as well as the Fo-F1 ATP synthase (complex V), which are responsible for ATP production by oxidative phosphorylation (OXPHOS) [34]. Through glycolysis, glucose is converted to pyruvate in the cytoplasm, then shuttled into mitochondria and oxidized via the tricarboxylic acid (TCA) cycle to generate ATP through the ETC by liberating electrons from reducing substrates NADH and FADH to O2 with proton (H+) pumped (41) through OXPHOS, producing 36 ATP. In our analysis, the activities of Complexes I, III, IV, and V along the ETC did not differ between groups; consequently, these complexes’ activities are regarded as intact.

However, when Oxygen Consumption Rate (OCR) and Extracellular Acidification Rate (ECAR) are compared, the infected group had a lower OCR and a higher ECAR than the non-infected group. The increase in ECAR was due to the cell’s release of H+ into the extracellular milieu. Taking into account the three primary factors of intact complex activity of the ETC, the decrease in OCR and increase in ECAR could be explained by the fact that OXPHOS is a metabolically slow but energy-efficient process. In fact, as highlighted by Liberti et al. cells with a high metabolic rate that are dividing rapidly, like cancer cells or activated immune cells, need a constant supply of new energy to keep functioning. This type of cell uses anaerobic glycolysis, a non-oxygenated pathway, to generate energy. The Warburg effect refers to the metabolic shift from aerobic glycolysis to anaerobic glycolysis, where pyruvate is converted to lactate in the cytoplasm, yielding two ATP molecules for each glucose molecule [35]. In fact, according to our analysis, the average oxygen consumption due to ATP synthesis was significantly lower in the infected group compared to the non-infected group, potentially resulting in less ATP production in the infected group overall, thus further supporting this theory. There are many bacteria known to affect cellular metabolism. Either the host’s immune response to the bacteria or the bacteria themselves can cause alterations to the host cell metabolism that help the bacteria survive inside the host cell. Regardless of the cause, the TCA cycle is typically slowed and glycolysis is induced [8]. Thus, the Infected group could be highlighted with a glycolytic phenotype exhibiting significantly higher rates of proton production reflected by the ECAR rather than using oxidative phosphorylation reflected by the OCR. The high ECAR could be the product of potential cellular damage, resulting in a plasma membrane leaking, or damage in the NAD/FAD system, resulting in the non-movement of electrons and the consequent pumping out of H+ to maintain intracellular pH. Instead of using high-energy electrons to power the ETC, the cell would release protons. The decrease in OCR despite the intact activity of the complexes in the ETC may be the result of a lack of high-energy electrons reaching the complexes potentially explaining why there is no oxygen consumption at the complex level, even though they are functioning normally in the assay. Thus, the ETC could be stalled or inactive without high-energy electron input.

Our pilot study has limitations; our analysis was conducted at a single timepoint, POD21, whereas other timepoints, particularly at the peak of infection, would be of interest to investigate. Another limitation was that only the response to one organism was obtained, whereas the response of other organisms and their impact on mitochondria may have differed. In addition, we did not investigate the potential effects of treatments on the local environment (both on infection and mitochondrial health). Due to time constraints and the need for fresh tissues, the ATP coupling assay was conducted on just three rats from each group which is another limitation. Additionally, we did not characterize the cellular origin of the mitochondria in this pilot analysis, but it is worthy of consideration in future work.

In this pilot study, we demonstrated that persistent infection modifies the mitochondrial response at the local level. The full impact of this information on the improvement of treatment outcomes will require additional evaluation.

Conclusions

In this work, we highlight PJI’s role in oxidative damage to mammalian tissues that is caused by mitochondrial dysfunction. Complex activity was maintained in the infected group, despite a switch to glycolytic pathways (lower OCR and higher ECAR) which is the first step toward comprehending how in vivo PJI impacts mitochondrial function. With a deeper understanding of the underlying mechanism, researchers may be able to investigate ways to mitigate these changes and fight the devastating orthopaedic problem of PJI and perhaps even conduct clinical trials to help these patients.