Knee osteoarthritis induces atrophy and neuromuscular junction remodeling in the quadriceps and tibialis anterior muscles of rats

Knee osteoarthritis (KOA) is associated with muscle weakness, but it is unclear which structures are involved in the muscle changes. This study assessed morphological alterations and the expression of genes and proteins linked to muscular atrophy and neuromuscular junctions (NMJs) in KOA, induced by anterior cruciate ligament transection (ACLT) in rats. Two groups of rats were assessed: control (without intervention) and KOA (ACLT surgery in the right knee). After 8 weeks, quadriceps, tibialis anterior (TA) and gastrocnemius muscles were analyzed (area of muscle fibers, NMJ, gene and protein expression). KOA group showed atrophy in quadriceps (15.7%) and TA (33%), with an increase in atrogin-1 and muscle RING-finger protein-1 (MuRF-1). KOA group showed quadriceps NMJ remodeling (reduction area and perimeter) and decrease in NMJ diameter in TA muscle. The expression of nicotinic acetylcholine receptor (nAChR) γ-nAChR increased and that of α-nAChR and muscle specific tyrosine kinase (MuSK) declined in the quadriceps, with a decrease in ε-nAChR in TA. MuRF-1 protein expression increased in quadriceps and TA, with no changes in neural cell adhesion molecule (NCAM). In conclusion, ACLT-induced KOA promotes NMJ remodeling and atrophy in quadriceps and TA muscles, associated with inflammatory signs and changes in muscle gene and protein expression.

The NMJ of the quadriceps exhibited a decline in the area and perimeter (p = 0.001) in the KOA group compared to controls, while the TA only demonstrated a decline in the maximum NMJ diameter (p = 0.014) in the KOA group compared to controls. No changes in neuromuscular junctions were detected in the gastrocnemius muscle (Table 1).
In both the control and KOA groups, the NMJ showed fewer type pretzel branches, with normal morphological characteristics. This means a normal number of junctional folds that ensure the localization of nAChRs and the propagation of plaque potential to generate muscle contraction 18 . However, in the quadriceps muscle, there was a clear reduction in the spread of NMJs in the KOA group (Fig. 2B) compared to controls ( Fig. 2A). The TA muscle showed a decrease in the maximum diameter (Fig. 2D) of the NMJ.
We found a decline in pain threshold in the Von Frey test in the KOA group compared to controls (46.82 ± 0.69 g/F vs 60.63 ± 2.25 g/F, respectively; p = 0.0002), Fig. 6E. The KOA group showed a higher Mankin histological score than that of controls (4.16 ± 1.07 vs 0.66 ± 0.21, respectively; p = 0.028; Fig. 6F and Supplementary Table 3). Figure 6G,H show a difference in proteoglycan content, evidenced by dye intensity (red), in addition to discontinuity on the joint surface in the KOA group.

Discussion
This study provided a novel contribution, showing changes in NMJs associated with quadriceps and TA muscle atrophy in rats with KOA. The most significant NMJ changes (decreased area and perimeter) were observed in the quadriceps, which exhibited 15.7% atrophy in muscle fibers. Although the TA demonstrated greater muscle atrophy (33%), alterations in the NMJs were smaller than in the quadriceps and restricted to an increase in diameter. Increased MuRF-1 expression 19 and muscle autophagy 20 have been suggested as possible mechanisms associated with NMJ alterations. Our study also found an increase in MuRF-1 and atrogin-1 expression could be associated with NMJ changes. It is known that MuRF-1 and atrogin-1 are the main signalers of muscle atrophy in different experimental models 21,22 . We identified an increase in the protein expression of MuRF-1 in the quadriceps and TA muscles, in line with increased genetic expression, indicating ACLT-induced atrophy in the KOA model. In addition to being involved in muscle atrophy, recent studies suggest that MuRF-1 is an www.nature.com/scientificreports www.nature.com/scientificreports/ important mediator in regulating the amount of nAChRs in the NMJ 19,23 , and may also be linked to NCAM, a regulator of synaptic stability related to neurotransmission 24,25 . However, the KOA group showed no changes in the protein expression of NCAM, which may be more sensitive only in cases of muscle denervation 26,27 . However, the increase in MURF-1 protein expression observed in this model may be related to both muscle atrophy and NMJ changes.
Most studies with KOA focus on observing changes in quadriceps, since they are directly related to the knee joint and exhibit muscle atrophy [28][29][30] , reduced strength 14,31 and functional decline 15,32 . In addition to muscle atrophy, the quadriceps showed greater NMJ alterations. With respect to mRNA expression in the quadriceps, we www.nature.com/scientificreports www.nature.com/scientificreports/ observed an increase in the gamma subunit of the acetylcholine receptor (γ-nAChR), which is normally expressed in the embryonic phase 33 , but also in muscle diseases such as denervation 34 . There was also a decline in alpha subunit (α-nAChR) expression, responsible for acetylcholine flow into the postsynaptic membrane 35 , as well as MuSK, a gene that initiates post-synaptic differentiation and plays an important clustering role and maintains nAChR 36 . The results of the present study show intense remodeling in the NMJs of quadriceps associated with KOA. Muscle fiber atrophy and arthrogenic muscle inhibition, which decreases quadriceps activation by afferent inhibition of the α motor neuron 11,37 , may be important mechanisms associated with the NMJ changes identified in this muscle. Another possible mechanism may be nerve terminal withdrawal, modulating the release of acetylcholine and leading to adaptation of the NMJs 38,39 . However, other studies are needed to assess the mechanisms involved in NMJ remodeling observed in the present study.
The TA muscle is little researched in KOA. We found only one study that assessed the TA of mice 4 and 8 weeks after KOA induction, exhibiting a decline in strength, 40% decrease in relaxation rate and less ATP expression, but with no alterations in the cross-sectional area of muscle fibers 40 . On the other hand, some studies on the TA of individuals with KOA assess function 41,42 and report a decline in eccentric strength in ankle dorsiflexors 43 . However, there is no clear evidence about the TA atrophy mechanisms in individuals with KOA. We believe that the greater atrophy observed in our study may be related to disuse, as described for quadriceps 31 , due to muscle compensation 13 , or because TA is a monoarticular muscle that may be more affected by atrophy 44 . Given the significant atrophy and NMJ alterations observed in the TA muscle, we believe that this muscle deserves to be more thoroughly studied in the KOA. Future studies on TA adaptations in the KOA may provide new scientific evidence to clinical practice and rehabilitation of this muscle.
The absence of neuromuscular changes in the gastrocnemius muscle observed here indicates that the KOA affects primarily the knee extensor muscles such as the quadriceps 9,15 and those that contribute to decelerate movement during functional activities such as the TA muscle 45,46 . Our results corroborate this hypothesis.
Our findings showed impaired gait pattern in the KOA group, as observed in similar studies 47,48 . Recent studies have associated gait deficits with a decrease in the CSA of the gastrocnemius muscle of rats 49 and long extensor of mouse fingers 50 . In addition, changes in gait were also associated with increased KOA pain in mice 51 . Changes in the gait pattern of animals are similar to the functional changes observed in individuals with KOA 52 . Our results show a correlation between altered gait patterns and the neuromuscular changes identified by atrophy and fewer NMJs in the quadriceps and TA muscles of rats with KOA.
The paw withdrawal threshold declined in the KOA group. This result corroborates a recent study that found reduced pain threshold associated with muscle weakness after ACLT 53 . Another recent study showed the presence www.nature.com/scientificreports www.nature.com/scientificreports/ of persistent mechanical hyperalgesia in a KOA model 54 . Our results strengthen the evidence that pain is related to the inflammatory and neuromuscular changes observed in the KOA.
We found an increase in leukocytes (73% macrophages) in the synovial fluid of the KOA group. Our results corroborate earlier studies indicating that macrophages are the main component of synovial fluid cells in KOA, contributing to the destruction of cartilage and osteophyte growth 55,56 . Recent studies have also classified the KOA into subgroups related to macrophages 57 . In addition to increased leukocytes in the synovial fluid, we observed increases in other signs, thereby strengthening the role of inflammation in the pathogenesis of KOA, such as greater knee surface temperature and swelling. experimental design and period. Rats Wistar (two months old, average weight of 214.41 ± 20 g) were randomly distributed into two groups (n = 6 per group): a) control (without surgery) and b) KOA (submitted to ACLT). Sample size was determined based on data from a pilot study, with muscle fiber cross-sectional area and NMJ area for quadriceps and TA muscles as the primary outcomes. Considering an alpha value of 0.05 and power of 95% (Software GPower 3.1), the sample size was calculated to be 5 animals per group. In the event of a possible loss during the experimental procedures, 8 animals were allocated to each group. Two animals from each group died during the experiments and the causes of death are described in Supplementary Table 1.
The rats were housed in cages (two animals per cage) in pathogen-free conditions at 24 °C ± 1 °C under a reverse light cycle (12 h light /12 h dark), with unrestricted access to standard rat chow and water. All animals www.nature.com/scientificreports www.nature.com/scientificreports/ underwent in-vivo tests one day before and 59 days after the ACLT and were euthanized the following day. The experimental design and study period are described in Fig. 1. ACLT-induced KOA. An adapted ACLT-induced KOA model was used. The rats were anesthetized with an intraperitoneal injection (12 mg/kg xylazine and 95 mg/kg ketamine). The right knee was shaved and prepared using an iodine solution. An incision on the medial side of the patellar tendon provided access to the joint space, after which the patella was dislocated laterally with the leg in extension. Afterwards, the joint capsule injury and ACLT were performed with ophthalmic scissors. Next, the anterior drawer test (free displacement of the tibia in relation to the femur) was conducted to confirm the ACLT 8,58 . The animals were returned to their cages and paracetamol (13.5 mg/100 mL) was added to the drinking water for the first 48 hours after surgery as postoperative analgesia. We used only the control group without surgical intervention because previous reports found joint wear similar to KOA in the sham groups (knee surgery without ACLT) 8,59 .
Muscle sample collection. The quadriceps, tibialis anterior (TA) and gastrocnemius muscles were isolated, removed and weighed. Each muscle was divided into three parts. The proximal fragment was immersed in Isopentane (pre-cooled in liquid nitrogen), stored at −80 °C, and used to measure the cross-sectional area (CSA) of the muscle fibers. The middle fragment was used for the nonspecific esterase technique. The distal fragment was rapidly frozen in liquid nitrogen, stored at −80 °C, and used to measure mRNA and protein levels Muscle Fiber CsA. Histologic serial cross-sections were obtained from the quadriceps, TA and gastrocnemius muscles in a cryostat microtome (Leica, CM1860, Germany). A histologic cross-section (10 µm) stained with toluidine blue was selected to measure the CSA under a light microscope (Axiovision 3.0.6 SP4 -Carl Zeiss, Germany) using morphometric analysis (Image J software, version 1.43u, National Institutes of Health, USA). The CSA of each muscle was obtained by measuring 100 fibers located in the central region of the section. The percentage atrophy index was calculated by the difference in the proportion between the cross-sectional area of muscle fibers from the control group and the cross-sectional area of the KOA muscle fibers. The atrophy index was calculated by the weight of each muscle (g) and normalized using body mass (BM; g) to obtain the MM/BM ratio of the control and KOA groups 8 . NMJ analysis -nonspecific esterase technique. The surface portions of the quadriceps, TA and gastrocnemius muscles were trimmed to the motor end-plate portion (containing the motor point), which was cut lengthwise into three or four slices. The resulting material was subjected to the nonspecific esterase technique 60 to characterize the NMJ.

Analysis of mRNA Expression by Real-Time PCR.
Total RNA was extracted from quadriceps and TA muscle samples in the control and KOA groups for mRNA analysis using TRIzol Reagent (Life Technologies, USA), according to the manufacturer's recommendations. The RNA was quantified using a Bioresearch BioPhotometer spectrophometer (Eppendorf ® Hamburg, Germany), which also determined RNA purity by measuring absorbance at 260 nm (RNA quantity) and 280 nm (protein quantity). Only samples with 260/280 ratios > 1.8 were used. RNA integrity was evaluated by ethidium bromide staining (Invitrogen) based on 28s and 18s ribosomal RNAs. Extracted RNA was treated with DNase I, Amplification Grade (Sigma Aldrich, AMPD1) to eliminate any possible contamination with genomic DNA from the samples. mRNA reverse transcription was performed using the iScript TM cDNA Synthesis Kit (Bio-Rad, CA), following the manufacturer's guidelines. The expression levels of mRNAs were assessed by quantitative real-time PCR (qPCR) using the CFX 96 Touch TM Real Time PCR Detection System, version 3.0 (Bio-Rad, CA). The cDNA samples corresponding to the mRNA of the genes analyzed were amplified by SsoFast ™ EvaGreen ® Supermix (Bio-Rad, CA) and primers were designed using Primer Express ® 3.0.11 software (Applied Biosystems, CA, USA) from sequences published in GenBank (www.pubmed.com) and synthesized by Life Technologies (USA) (Supplementary Table 5). The expression levels were normalized by glyceraldehyde 3-phosphate dehydrogenase (GAPDH), peptidylprolyl isomerase A (PPLA), peptidylprolyl isomerase B (PPLB), beta cytoskeletal actin (ACTB) and hypoxanthine-guanine phosphoribosyltransferase (HPRT), whose expression was constant among all samples. Relative quantification of gene expression was performed using the comparative 2 ∆∆ C(T) method 63 . protein expressions by western blot analysis. Protein samples (containing 80 μg of protein) were collected from the quadriceps and TA muscles. These were separated on SDS/PAGE gels (12% wt/vol) and transferred to a nitrocellulose membrane. The membranes were incubated with antibodies against MuRF-1 (IgG, polyclonal, 1:500, 40Kda, Gene-Tex, USA) NCAM (IgG, polyclonal, 1:1000, 200Kda, Merck, USA) and normalized using α-Tubulin (IgG, monoclonal, 1:1000, 50Kda, Sigma, USA). The blots were visualized in an ECL solution (Amersham Pharmacia Biotech, Little Chalfont, UK) and exposed in a ChemiDoc MP Imaging System (Bio-Rad Laboratories, Hercules, California, USA) 64,65 . Gait test. The hind paws of the rats were brushed with ink. Next, the animals were allowed to run on a 60 cm-long, 7 cm-wide track covered with white paper. A dark chamber was placed at the end of the track to entice the rats. Upon completion of the test, the paper was scanned at 300 dpi 47 . The measurement around the paw was defined as paw area (cm²), the distance between the first and fifth toes as paw width (cm), the distance of the same hind paw between two steps as stride length (cm), the horizontal distance between the left and right paw as the base (cm), the distance between the third toe and the heel as paw length (cm), and the paw angle as the angle through the hind legs (°). The measures of footsteps were quantified by ImageJ software (version 1.43u, National Institutes of Health, USA).
Knee skin temperature. The animals were acclimated in a dark room (15 min; 24 °C ± 1). Thermography was used to quantify the skin temperature of the knee in both KOA and control groups using an infrared thermal camera (FLIR Systems ® T420, USA), placed on a tripod 50 cm from the animal's knee. The images were analyzed in Flir Tools software, and the results expressed in (°C).
Ex-vivo leukocyte migration. Leukocyte migration was determined by synovial fluid, as previously described 66,67 . The joint cavities were washed twice with 5 μL phosphate-buffered saline (PBS) solution containing 1 mM ethylenediaminetetraacetic acid (EDTA) and then diluted to a final volume of 50 μL with PBS/ EDTA to evaluate leukocyte migration at the indicated time. The total number of leukocytes was counted in a Neubauer chamber diluted in Turk's solution. The results were expressed as the number of leukocytes per joint cavity. Differential cell counts were determined in cytocentrifuge Rosenfeld-stained slices (Cytospin 4; Shandon, Pittsburgh, PA, USA). Differential cell counts were performed with a light microscope, and the results were expressed as the number (mean ± SEM) of leukocytes per joint cavity. For this analysis we used only 4 animals from the KOA group 66 . paw withdrawal threshold. The articular hypernociception of the femur-tibial joint was evaluated using an Electronic Von Frey meter (Insight ® EFF-301). The animals were acclimated (30 min) in acrylic cages with a wire grid floor (12 × 10 × 17 cm high). A trained investigator (blind to group allocation) applied perpendicular pressure to the plantar surface of the hind paw. When the paw was withdrawn, the intensity of the force applied was automatically recorded. The mechanical threshold results are expressed in grams (g/F) 53,66 . Histological assessment of the knee joint. A standard histological protocol was used. Briefly, the knees were fixed (4% formaldehyde for 2 days) and decalcified (10% EDTA). The samples were embedded in paraffin blocks and histological sections were obtained (10 μm) using a microtome (Leica RM-2245, Germany). Samples were stained with hematoxylin and eosin (HE-Merck, Darmstadt, Germany) and 0.1% Safranin-O (Merck, Darmstadt, Germany) 68,69 , and examined under a light microscope (×100; Axiovision 3.0.6 SP4 -Carl Zeiss, Germany). Two experienced blinded observers evaluated cartilage damage using the Modified Mankin Score 70 , Supplementary Table 4. statistical analysis. Continuous variables were presented as mean ± standard error of the mean (SEM).
Since all variables were normally distributed (p > 0.05), according to the Shapiro-Wilk test, an independent t-test was performed. Statistical analyses were performed with SPSS, version 23.0 (SPSS Inc., Chicago, USA). The figures were plotted in GraphPad Prism software, version 5.0. A p-value < 0.05 was considered statistically significant.

Conclusion
ACLT-induced KOA in rats promotes NMJ remodeling and atrophy in quadriceps and TA muscles, associated with inflammatory signs and alterations in gait and muscle gene and protein expression.

Data Availability
The datasets generated and/or analyses conducted during the study are available from the corresponding author.