Clinical-like cryotherapy improves footprint patterns and reduces synovial inflammation in a rat model of post-traumatic knee osteoarthritis

Cryotherapy is a non-pharmacological treatment commonly used to control inflammation and improve function after acute traumas. However, there are no definitive findings about its effects on chronic joint diseases such as knee osteoarthritis (KOA). The aim of this study was to investigate the effects of clinical-like cryotherapy on functional impairment and synovial inflammation in a rat model of KOA generated by anterior cruciate ligament transection (ACLT). Thirty-two male Wistar rats were randomly divided into four groups (n = 8/group): Control, KOA, KOA + Cryotherapy and KOA + Placebo. The last two groups were submitted to the relevant interventions twice a day for five days (61 to 65), with each session lasting 20 min. Gait test, skin temperature, thermal response threshold and joint swelling were assessed in all groups before ACLT surgery, and pre (60th day) and post (66th day) intervention protocols. On day 66, the animals were euthanized and exsanguinated to remove the synovial membrane for histopathological examination and synovial fluid to determine the leukocyte count and cytokine concentration. After the intervention period (66th day), footprint area only increased in the KOA + Cryotherapy group (P = 0.004; 14%) when compared to KOA and KOA + Placebo, but did not differ from controls. Cryotherapy lowered the synovial fluid leukocyte count (P < 0.0001; ≥95.0%) and cytokine concentration (P < 0.0001; ≥55%) when compared to the KOA and Placebo groups. Synovial score and synovial fibrosis did not differ in the KOA groups. In conclusion, footprint patterns improved in rats with ACLT-induced KOA as a result of clinical-like cryotherapy, which also lowered the synovial fluid leukocyte count and inflammatory cytokine concentration in these rats.


(B).
There were no changes in the other variables related to gait analysis (Supplementary Appendix I).

Discussion
Our results show, for the first time, that clinical-like cryotherapy is a beneficial intervention for ACLT-induced KOA, since it improves footprint patterns, and its effects are mediated by downregulation of joint synovial inflammation. The reduced inflammatory process observed in the Cryotherapy group due to lower leukocyte migration to the joint cavity indicates a beneficial effect on the control of synovial inflammation. These findings demonstrate the potential of cryotherapy as a non-pharmacological treatment for joint inflammation in KOA. www.nature.com/scientificreports www.nature.com/scientificreports/ Although the anti-inflammatory effects of cryotherapy are little studied in animal models of arthritis 30 , a number of action mechanisms have been proposed to explain its effects on reducing the inflammatory process in joints. A single long-term ice-pack application (4 hours) decreased leukocyte phagocytosis in the synovial fluid of a dog's knee with urate crystal-induced synovitis. The effects were attributed to increased viscosity in the synovial fluid, which may have hindered leukocyte movement to the crystals 23 . In another study, ice packs (30 minutes, once a day for 10 days) reduced cell infiltrate and synovial hyperplasia in a rabbit zymosan-induced arthritis model 31 . In a traumatic model, pigs with radiocarpal ligament sprain experienced a decline in the number of leukocytes after two applications of crushed-ice packs (20 min each) 26 . In these studies, synovitis and inflammation were suppressed by lowering intra-articular temperature. According to previous studies, the enzymatic activity responsible for matrix degradation (i.e. collagenases) decreases at lower temperatures 22,32 . Although we did not measure intra-articular temperature, these mechanisms likely also occurred, contributing to the decline in joint inflammation. Another possible mechanism associated with cryotherapy is a decrease in the local metabolism, which was demonstrated in an earlier investigation 33 , causing less cellular infiltrate in the synovial membrane and, consequently, less activation of inflammatory mediators.
It is known that pro-inflammatory cytokines, such as IL1-β and TNF-α, exert catabolic action and contribute in a number of ways to joint degeneration in KOA, including proteinase activation and suppression of type II collagen, one of the main components of the extracellular matrix 8,9 . Other cytokines, such as IL-17, act as a critical mediator in neutrophil recruitment, migration and activation 34 . According to our results, cryotherapy was effective in reducing (≥55%) pro-inflammatory cytokine levels in the synovial fluid. An earlier study on cryotherapy applied to arthritic rat paws (30 minutes, twice a day, for 14 days) showed local and systemic anti-inflammatory effects, which were mediated primarily by genetic and protein expression of IL-6 and IL-17, independent of TNF-α 27 . Studies conducted in different pathological and physiological conditions related the beneficial effects of cryotherapy to NF-kβ-dependent gene inhibition of IL-1β, TNF-α and IL-6 30,35 . These pathways are also involved in the effects of cryotherapy observed in the present study.
Despite the significant decline in the inflammatory process observed in the KOA group submitted to cryotherapy, there was no improvement in the histopathological signs of their synovial membrane. This result shows that the experimental model used induces KOA, and that the signs of the disease in the synovial membrane are unchanged by cryotherapy.
In relation to gait analysis, the larger footprint area affected in the KOA + Cryotherapy compared to KOA and KOA + Placebo groups, and the similar findings to Controls, indicates more homogeneous weight bearing, favoring a normal gait pattern. The results of the present study with the KOA model in rats were similar to those of previous research 36 , which also reported a decline in the footprint area of mice after four and eight weeks of ACLT. Cold-induced analgesia directly affects gait control, decreased nociceptor excitability thresholds and nerve conduction velocities 37 . A decline in nociceptive information transmitted through primary afferents to the spinal cord would result in fewer behavioral signs and reduced neuronal activity of dorsal horn neurons, prompting a reduction of the expanded receptive field, which may impact gait responsiveness 25 . In contrast to that observed in patients with KOA 38,39 , we found no changes in the other variables related to gait analysis in all groups, such as stride length and width. Although our results corroborate with previous findings in rats 36,40,41 , we did not control the gait speed of the animals, an important covariate for nearly all gait parameters. This can be measured in future studies by recording of the time taken (in seconds) to reach the dark chamber (distance traveled divided by the time it took to cover this distance) 42 , statistical models 43 , speed control with treadmill 44,45 or by using modern high-speed videography (i.e. catwalk), which is more likely to provide a robust analysis of spatial parameters 46 .
Finally, the chronicity characteristics of KOA may explain the absence of a cryotherapy effect on both skin temperature and swelling of the knee joint. There were also no changes in thermal hyperalgesia in the KOA model, which is supported by previous studies of the post-traumatic KOA model 47,48 . Behavioral changes commonly reported in the literature involve the assessment of other parameters, such as gait adaptations, mechanical hyperalgesia, mechanical allodynia and weight-bearing asymmetries.
One of the limitations of this study was using the footprint test to evaluate gait patterns, since it cannot accurately assess the velocity of the animal. Another weakness was the lack of a sham group for the surgical procedure (knee surgery without ACLT); future research should include a sham group for comparison purposes.

Materials and Methods
The experimental protocol was in accordance with the National Guide for the Care and Use of Laboratory Animals (National Research Council, 1996) 49 . The Federal University of São Carlos Ethics Committee approved the experimental procedures (number 7949291116/2017) and the study was conducted by trained professionals blinded to the identity of the experimental groups 49 . A priori sample calculation was performed using G*Power (version 3.1; University of Trier, Trier, Germany) 50 . Based on a previously performed pilot study (n = 10), it was adopted an α = 0.05, power (1-β) = 0.95, correlation coefficient = 0.5 and effect size = 0.37. Gait test and Leukocyte migration were used in the calculation. Therefore, it was estimated that a total of 32 animals (8 per group), would be an adequate sample size.
Experimental design. The animals analyzed were two-month-old male Wistar rats (Rattus norvegicus, n = 32; 297 ± 25 g), housed at 24 °C ± 1 °C (three per cage) under pathogen-free conditions in a reverse light cycle (12/12 light/dark) and given ad libitum access to standard rat chow and water. A computer program (www. random.org) was used to randomly divide the animals into four groups (n = 8 per group): Control (composed of naïve animals not submitted to surgery or intervention); ACLT knee surgery (KOA, untreated group); ACLT knee surgery and an ice pack (KOA + Cryotherapy), and ACLT knee surgery with a sand pack similar to the ice pack (KOA + Placebo) 49 . The groups were assessed one day prior to ACLT surgery and 60 days after the procedure, performing the least stressful tests firsts and progressing to the most stressful ones, as follows: skin temperature, gait test, thermal response threshold and swelling. The interventions (ice pack or sand pack) were then carried out in the the KOA + Cryotherapy and KOA + Placebo groups twice a day for five days (61 to 65), With each session lasting 20 min. All four groups were assessed day 66, and the animals were euthanized and exsanguinated to remove the synovial membrane and synovial fluid (Fig. 1).

KoA induction.
We used an adapted ACLT-induced KOA model that causes joint changes in rats similar to those observed in humans 5 . Briefly, the animals were anesthetized with an intraperitoneal injection (12 mg/kg Xylazine and 95 mg/kg Ketamine) and the right knee was shaved and prepared using an iodine solution. Next, a medial longitudinal parapatellar incision was made and joint capsule injury and ACLT were induced with ophthalmic scissors. The anterior drawer test (free anterior displacement of the tibia in relation to the femur) was performed to confirm ACLT. After the skin was sutured, the animals were returned to their cages and paracetamol (13.5 mg/100 mL) was added to their drinking water for the first 48 hours as postoperative analgesia 5 . In this model we previously described a higher Mankin histological score in the KOA group (60 days after ACLT) when compared to controls (naïve animals) 7 .
Gait test. Gait analysis was conducted using the paw print test as previously performed in an ACLT animal model 36 . 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. The measurement around the right paw was defined as footprint area (pixels), the distance between the first and fifth toe as paw width (cm), the distance between two steps of the same hind paw 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 between the fifth toe and the calcaneus and a horizontal line (°). Skin temperature. In accordance with the criteria for acquiring thermal images 51 , the animals were acclimated in a dark room (15 min; 23 °C ± 1). Thermography was used to measure the skin temperature of the right www.nature.com/scientificreports www.nature.com/scientificreports/ knee in all groups, using a FLIR T420 infrared thermal camera (FLIR Systems ® , USA), attached to a tripod placed 50 cm from the animal's knee. The images were analyzed in FLIR Tools software, and the results expressed in °C. A pilot study (n = 8 rats) was conducted to determine skin cooling in the KOA + Cryotherapy group immediately after a single session. The knee skin temperature decreased 26.8 ± 0.4 °C, from 36.9 ± 0.7 °C to 10.1 ± 1.5 °C.
Knee joint swelling. Knee joint thickness was measured with the animals in the supine position under anesthesia (2 ml/mlO 2 ; 1.5% isoflurane) 52 , using a digital caliper (Fisher Scientific, 150 mm, USA), positioned on the medial and lateral femoral condyles, at the knee joint interline level 7 . The mean of two measures was used and the data expressed in millimeters (mm). ice) and KOA + Placebo groups (20 g sand pack) were carried out twice a day in the laboratory, approximately 4 hours apart. Each intervention lasted 20 minutes 21 and was applied under anesthesia (2 ml/mlO 2 ; 1.5% isoflurane) 52 , with the animals in the supine position with raised paw (hip joint ± 45°). The ice and sand packs were placed around the knee, using an elastic band for compression. The cryotherapy protocol followed the clinical recommendations for the management of musculoskeletal injuries, according to PRICE protocol (Protection, Rest, Ice, Compression and Elevation) 29 . Despite the absence of intervention in the Control and KOA groups, they were submitted to the same anesthesia applied to the intervention groups.
Synovial fluid collection. The animals were anesthetized (240 mg/kg ketamine and 60 mg/kg xylazine; i.p.) and exsanguinated 54,55 . This last procedure was carried out to minimize the possibility of blood contamination in synovial fluid. The skin and right knee joint ligaments were removed, and the synovial cavity was washed twice with 200 µL of phosphate-buffered saline (PBS) containing 10 mM ethylenediaminetetraacetic acid (EDTA). This joint lavage fluid was used for cell counting, leukocyte differential counting and cytokine level determination 55 . In Vivo leukocyte migration. Leukocyte migration was determined using synovial fluid, as previously described 56 . The joint cavities were washed twice with 5 μL of PBS containing 1 mM EDTA and then diluted to a final volume of 50 μL with PBS/EDTA to evaluate leukocyte migration at the established time. The leukocytes were counted in a Neubauer chamber diluted in Turk's solution. The results were expressed as the number of leucocytes per joint cavity. Differential leukocyte count. For differential count, aliquots of joint lavage fluid were removed and centrifuged at 1,500 rpm for 10 min at 4 °C. The supernatant was stored at −80 °C for subsequent analysis (cytokine determination), and the cell pellet was resuspended in 200 µL of as PBS plus EDTA solution. Differential count slides were prepared using an aliquot of the washed joint fluid (50 μl) subjected to cytocentrifugation at 1,500 rpm for 10 min. The slides were then mounted, fixed for 4 min, and stained with eosin and hematoxylin. Next, they were washed in tap water and allowed to dry 55 .
One hundred cells per slide were counted under an optical microscope with a 100x oil immersion objective, in order to differentiate cell types (macrophages, lymphocytes and neutrophils). The cell count in the joint lavage was obtained by calculating the percentage of each cell type (differential count) and number of leukocytes in the joint lavage fluid 55 . The results were expressed as number of cells × 10 3 /ml. Determination of cytokine levels. TNF-α, IL-1β, IL-6, IL-17 and IL-10 concentrations were determined using a commercially available enzyme-linked immunosorbent assay (ELISA), following the manufacturer's instructions (Duo-Set kits; R&D Systems, Minneapolis, MN, USA). The optical density of the individual samples was measured at 450 nm using a spectrophotometer (Spectra Max-250, Molecular Devices, Sunnyvale, CA, USA). Results were expressed as the mean ± SD of cytokine levels in pg/mg of joint fluid 57,58 . Histopathological assessment of synovitis. Articular capsule samples were fixed in 4% (vol/vol) buffered formalin, dehydrated in ethanol, and embedded in paraffin for slide preparation. Tissue sections were stained with hematoxylin and eosin (H&E) to analyze synovitis (inflammatory cell influx and synovial hyperplasia). The severity of the synovial pathology (i.e., synovitis) was determined using a scoring system that measures the thickness of the synovial cell layer on a scale of 0-3 (0 = 1-2 cells, 1 = 2-4 cells, 2 = 4-9 cells, and 3 = 10 or more cells) and cell density in the synovial stroma on a scale of 0-3 (0 = normal cellularity, 1 = slightly increased cellularity, 2 = moderately increased cellularity, and 3 = greatly increased cellularity) 59 . Additional slides were stained using the Sirius Red staining protocol in order to evaluate articular fibrosis. Collagen deposition under the synovial membrane was measured as the Sirius Red-positive staining area in 15 random high power fields (400x magnification) using Image J software (Image J, 1.33 u, USA). The results are expressed as a percentage of collagen-positive area.