New experimental methodology to evaluate lubrication properties of synovial fluid containing worn tissue particles in osteoarthritis patients

Studying the lubrication properties of osteoarthritis (OA) synovial fluid (SF) enables an understanding of the boundary lubrication joint, mobility, and friction. However, tribology has never been combined with the clinical reality of the presence of worn particles within the synovial fluid and how they affect the osteoarthritic joints. Part of the problem relates to the tribology methods studying friction by applying inadequate pin-on-disc techniques. In this study, synovial fluid with and without worn particles was studied using a customized tribometer. This method enables opening the contact at the end of each cycle and simulates better contact conditions of a natural knee joint and can thus be applied for evaluating the severity of joint OA and the treatment given to the patient.


Introduction
Osteoarthritis (OA), a worldwide disease, is considered the most common and most disabling form of arthritis, affecting 1 in 5 adults. OA increases with age, and its most common incidence affects men and women between 55 and 65. It is also considered to be an economic burden [1]. Despite these facts, there are neither reliable diagnostic tests for this disease nor reliable diagnostic means to evaluating the effectiveness of any treatment.
Although there are advanced diagnostic methods in different medical fields, there is neither a diagnostic blood test nor any other test to diagnose the early onset of OA or objectively classify its severity. Physical examination combined with imaging of the joints is used as the best available method to diagnose and treat this condition. Yet, both are very inaccurate: The diagnostic tests and the imaging interpretation may vary between individual physicians examining the same case. This also applies to the evaluation of the treatment given to the patients.
Appropriate diagnostic and treatment for OA should rely on the comprehension of the knee functioning and changes as the disease progresses. Synovial joint lubrication takes place in all synovial joints and explains the longevity of normal articular cartilage. Synovial fluid (SF) is an ultrafiltrate of the blood plasma. It comprises hyaluronan, lubricin, proteinase, collagenases, phospholipids, and prostaglandins. It also contains blood cells [2]. The level, composition, and distribution of the synovial fluid can be changed with joint disease and stage of OA [3].
The low friction property of the synovial fluid is obtained by a microscopic film present on the articular surface of the cartilage. This lubrication mechanism of articular cartilage surfaces is caused by lubricin, a mucinous glycoprotein that guarantees boundary lubrication [4]. Boundary lubrication decreases friction between surface asperities that are pressurized together www.Springer.com/journal/40544 | Friction during periods of high loads and low speeds [5]. This phenomenon mainly occurs in large weight-bearing joints when the lubricant fluid is present on the surface of cartilage but with the thickness significantly thinner than surface roughness asperities [6]. The viability of lubricin-producing cells in the superficial zone and the availability of lubricin in SF to deposit on the cartilage surface help maintain boundary lubrication by preventing cell and protein adhesion [7]. The second mechanism of joint lubrication is hydrostatic lubrication. It is caused by interstitial fluid pressurization that is responsible for reducing the friction coefficient (CoF) μ of articular cartilage under physiological conditions [8]. The tribological properties of the sliding surfaces of the articular cartilage have been measured to be very low (low CoFs) even in the presence of high physiological contact pressures. The low CoF is also explained by the molecular structure of articulating surfaces [9,10]. In the present study, we investigate the lubricating properties of synovial fluids simulating patients with OA compared to those of normal healthy knees by measuring the CoF.
Caligaris et al. [11] have investigated the osteoarthritic human tibiofemoral joint CoF and the potential beneficial tribological effect of a healthy synovial fluid. They report no statistical differences in the CoF with the increasing OA, whether in migrating or stationary contact area configurations. Caligaris et al. [11] even conclude that the CoF of human tibiofemoral articular cartilage does not increase with the naturally increasing OA for visual stages ranging from 1 to 3. Although this outcome appears counterintuitive relative to that reported in the prior studies of enzymatically or mechanically degraded cartilage, it can be explained since interstitial fluid pressurization is not necessarily defeated by advancing tissue degeneration. However, it is widely recognized that the lack of lubricin in the synovial fluid does enhance cartilage degradation due to the increase in shear stress and lack of boundary lubrication [12,13].
It is vital to answer a question to reach an acceptable understanding of the precise functioning mechanisms of a knee-Does the CoF of the synovial fluid represent the friction that occurs in the human knee? Numerous studies have been reported in Refs. [14][15][16][17], most testing the tribology effect of synovial fluid on components used in joint replacement. The tribology methods used are usually pin-on-disk simulators, reciprocating pin-on-disks [15], pin-on-plate tribometers [16], or tracking pendulum tester [17]. The tests aim to characterize the tribological properties of different materials such as ceramics-on-ceramics and ceramicson-polyethylene or the lubrication capacity of synovial fluid. In these tribology tests, the goal is to determine the effect of different (lack or deficient) synovial components on the lubrication, i.e., the effect on the CoF of the synovial fluid [14,15]. However, most of the experimental simulations using rotating pin-on-disc and reciprocating pin-on-disk or tracking pendulum testers do not sufficiently consider the effect of tissue debris in the SF in the case of increased OA. Indeed, focusing on daily activities such as walking, the articular cartilage in knee joint is repeatedly loaded and unloaded within very short time periods, during which the cartilage is assumed to be recovered by the synovial fluid containing the wear debris. This is particularly difficult to simulate using a laboratory tester. Therefore, some attempts reported in Ref. [18] try to overcome this limitation by performing continue friction tests on a reciprocating pin-on-plate tribometer with a loading phase lasting a few minutes followed by unloading of a few minutes.
In light of the above descriptions, we are introducing a new diagnostic method for objectively evaluating OA. This method combines both tribology testing and orthopedic data. We conducted an experimental study to investigate the tribological effects of synovial fluid, including all the components present within the synovial fluid: dead cells, cartilage, and bone debris that interfere with joint mobility and function. Such tribological testing does not concentrate only on the synovial fluid components for tribology evaluation of the joints.

Materials and methods
In the present study, the friction pair was composed of a face-polished cylindrical polytetrafluoroethylene (PTFE) specimen rubbed against a smooth flat and rigid glass under different lubricants.

PTFE specimen
The flat specimens were prepared from a pure PTFE 2134 Friction 11(11): 2132-2141 (2023) | https://mc03.manuscriptcentral.com/friction rod (Polimersan, Turkey) of 10±0.1 mm in diameter and 7±0.3 mm in height ( Fig. 1(a)). The surface roughness has a significant role in determining the frictional performance as it can significantly affect the fluid flow between cartilage contact gaps [19]. To overcome this effect in the present study, all PTFE samples were deliberately polished to the same roughness degree. To do so, a high finish of the specimen faces was achieved by a controlled polishing process, gradually decreasing in stages of 120, 320, 500, and 1,000 to 1,200 grit. Then, after the polishing process, the surface roughnesses of different samples were checked using an optical profilometer (NT1100, Wyko, USA) used to characterize the surface roughness of the polished flat surfaces of the PTFE specimens, in which the arithmetic average roughness R a was reduced from R a ≈ 900 nm before polishing to 240 nm < R a < 260 nm after polishing. Figure 1(b) shows the three-dimensional (3D) optical profilometry image for the topography of the polished specimen. Table 1 summarizes the main roughness parameters for the PTFE samples after polishing (the average values obtained from 4 measurements using the optical profilometer).

Counter surface
The friction tests were performed against a rigid hard counter-face made of a 76 mm × 26 mm × 1 mm glass plate (Am Woellerspfad 4, Paul Marienfeld GmbH & Co. KG, Germany). The glass plate has a smooth surface with R a ≈ 30 nm (measured using a 3D optical profilometer (NT1100, Wyko, USA)).

Lubricants
As mentioned above, the present study evaluates the capabilities of a customized linear two-axis tribometer to accurately evaluate the tribological behaviors of lubricated contact under linear, cyclic friction, and particularly the accuracy and repeatability of results under different operational conditions and parameters. This allows objectively classifying the resolution of tribological performance in repeated experiments, which is vital in the study of lubrication properties of synovial fluid containing (or not containing) worn tissue particles in OA patients. To do so, we used five different lubricants.
iii) Synthetic SF, purchased from Limbs & Things, UK, the volume of 0.5 mL. iv) Synthetic synovial fluid containing non-porous tricalcium phosphate (β-TCP) granule particles manufactured by CAM Bioceramics, the Netherlands, with sizes ranging between 50 and 150 μm with the ratios of 0.5 mL SF and 2 mg β-TCP. For future studies, the TCP particles were used to simulate worn tissue debris presented in the synovial fluid in OA patients.

System description
The present study performed friction tests on a customized two-axis tribometer designed and constructed in the Tribology Laboratory at Azrieli College of Engineering Jerusalem (JCE), Israel. A full description can be found in Refs. [20,21]. Unlike Ref. [22] based on a vertical moving counter-face, the present concept is based on a horizontal counter-face moving in three directions. It allows loading/unloading the mating surfaces in both vertical and tangential directions and evaluating the tribological performance of different materials accurately under dry or lubricated contact conditions according to needs.
The device, of which a general schematic illustration is provided in Fig. 2, incorporates two main units used for driving and measuring purposes. (i) The drive unit, located on the lower part of the system, contains three translation stages (two motorized and one manual), which allow moving the horizontal hard counter-face (glass in this study) in the three directions to adjust the contact with the PTFE specimen, load the system, and perform friction tests. The movements in normal (z) and lateral (x) directions are motorized using the motors (X-LSM025A-E03, ZABER, Canada; 25 mm of total travel distance and 15 μm of accuracy and X-LHM050A-E03, ZABER, Canada; 50 mm of total travel distance and 75 μm of accuracy) for normal and lateral displacements, respectively. However, adjustment of the specimen location in the third (y) direction with regard to the rigid counter-face is done manually using an accurate micrometric translation stage (TSB28M-MH2, ZABER, Canada). (ii) The stationary measurement unit, which is located on the upper part of the device, consists of two load cells (FSH00092-LSB200, FUTEK, USA) capable of providing a high-resolution measurement (0.1 mN) of force variation generated in the contact during friction. The two load cells are positioned perpendicular to each other in such a way that their action measurement lines intersect exactly at the frictional plan to avoid any unwanted moment.
The measurements were sampled with a multifunctional data acquisition board (Lab-PC-NI USB-6211, National Instruments Corporation, USA) and processed using a LabVIEW 2017 software package (11500 N Mopac Expressway, National Instruments Corporation, USA). The sampling rate of the data acquisition was adjusted at 1,000 samples per second, and then the collected data were averaged for every 100 points to minimize the noises of the collected data. The programmed interface allows choosing operational parameters such as velocities of the counter-face, normal force, sliding distance, and waiting dwell time.
It is recognized that an accurate study of tribological properties emphasizes parallelism between the mating surfaces during a test. To this end, the current study used a passive self-aligning system based on the principle of two free rotation axes (the essential "uniform normal load distribution" principle in tribology testing) to guarantee a full flat-on-flat contact between the mating surfaces to the possible maximum degree (Fig. 3). The self-alignment system consists of a circular frame printed in a 3D printer (PlasCLEAR material, Asiga, Australia), receiving two small metallic axes to allow freedom in the two axis perpendicular to the vertical axis.

Experimental procedure
This tribometer was initially designed to perform three modes of tribological tests, i.e., (i) friction,  (ii) adhesion, and (iii) peeling. During friction tests conducted in the present study, the axis perpendicular to the sliding direction of the self-alignment mechanism (y axis in Fig. 2) is intentionally locked under a loaded state to avoid any unwanted rotation of the samples during sliding. The use of this lockable self-alignment (called "LSA" hereafter) is a self-alignment mechanism, in which the rotation of one of its axis (the one perpendicular to the sliding direction) can be locked after loading, thus being suitable for the cartilage friction experiment.
Before friction tests, all specimens (PTFE pin and glass counter-face) were cleaned with 99% ethanol to eliminate any possible contamination from the surfaces before each experiment. Then, the specimens were mounted on the tribometer. The contact location between the PTFE surface and the glass counter-face was adjusted in the y direction using the micrometric stage. Once the specimens are mounted, the measurement system is calibrated by resetting the load cells to eliminate the effect of gravity (calibration was performed after each sample replacement as well). Then, the glass counter-face was covered with the lubricant liquid, moved in the vertical direction, and brought into contact with the specimen until a desired normal load P was obtained, and then the LSA was locked. Then, the stage holding the glass counter-face is withdrawn in the vertical opposite direction to open the contact and allow the lubricant to cover the surface before the test begins. This procedure allows full flat-on-flat contact between the mating surfaces during the friction test.
The friction tests were conducted following six successive stages as follows (Fig. 4). Stage 1: While the contact is open, the glass counter-face covered with the lubricant is moved in the vertical direction and brought into full contact with the PTFE specimen at a constant loading speed of 1 mm/s, leading to a gradual increase in the P, until the desired predefined value of 5 N is reached. Stage 2: The P is kept constant for a dwell time of 5 s, and the tangential load cell is rested. Stage 3 starts with the movement of the translation stage in the tangential direction at a constant sliding velocity of 0.5 mm/s for a total travel distance of 20 mm. During this stage, the P is kept constant based on programmed close loop control of 2% error of normal force, and the friction force resisting the sample motion is recorded. Stage 4: Upon completion of the tangential motion of 20 mm, a dwell time of 0.5 s transpires; then, Stage 5 begins with the translation stage withdrawing in the vertical opposite direction at a constant velocity of 5 mm/s up to a total distance of 10 mm. Stage 6 follows with the displacement in the opposite lateral direction at a constant velocity of 5 mm/s to bring the specimens to their initial position, thus marking the end of one friction test cycle. Then, a new test cycle can start. Note that the contact is open during the motion of the glass counter-face in the opposite normal and lateral directions (Stages 5 and 6). This allows the lubricant to be reintroduced in the interface covering the specimen surfaces, particularly if it is charged with micro-sized solid particulates. Each friction test is repeated 13 times (friction cycles) for each lubricant (the first three are considered running-in without recording, and the other ten are recorded and saved to be analyzed). The friction tests in this study were performed at room temperature of 25±1 °C and a  Figure 5 presents the typical variation in time of the μ (obtained by dividing the measured friction force by the applied P) during the ten successive friction cycles for each tested lubricant Section 2.3. The tribological behavior in each cycle can be divided into three characteristic stages: Stage 1 relates to the increase of friction force during pre-sliding; Stage 2 designates the maximum static CoF attained at sliding inception; and Stage 3 corresponds to the decrease and stabilization of the friction force during sliding corresponding to the dynamic CoF (the first friction    Fig. 5). Except for the test conducted with the SF-containing particles, in which the friction was high, the scale of the CoF is plotted between 0 and 0.6 in all other curves, the values of the CoF are much lower (~0.03-0.05), and their friction scales were plotted between 0 and 0.1. It can also be seen that the presence of PCT particles leads to notably elevated CoF and induces a high friction instability that appears as fluctuations in the friction curves (the bottom curves in Fig. 5). Figure 6 presents an example of the superposition of all ten friction cyclic curves for the tests conducted under "SF" and synovial fluid enriched with microparticles (SF+MP). Very good, even perfect reproducibility is obtained with the successive friction cycle repetition when the tests are conducted with a given lubricant under the same operational condition. The reproducibility of friction curves in the case of synovial fluid with microparticles indicates that opening the contact at the end of each friction cycle allows the lubricant charged with particles to be reintroduced in the contact, which seems to simulate the physiological conditions of a natural knee joint during walking more objectively.

Results and discussion
The maximum static CoF obtained in Stage 2, as well as the average value of the dynamic friction recorded after stabilization (steady-state) in Stage 3 (in Fig. 5(a)), is recorded for each test or repetition. Then, the average values obtained from 10 repetitions and the relative standard deviations for all the tested lubricants were calculated and are displayed in Fig. 7.
Except for SF+MP, in the case of all lubricants, the static CoF is quite similar to the dynamic CoF. It can be clearly seen that the lower dynamic CoF is obtained with the pure synovial fluid, even lower than those of PBS and GLY. In contrast, a higher value is obtained with synovial fluid containing TCP particles. It should be noted here that in the case of real damaged cartilage, the elevation of CoF due to the presence of worn micro-particulates should be added to the elevation of friction related to the cartilage lesion, which leads to rapid consolidation associated with higher CoFs [23]. However, diluting the synovial fluid with xylitol while maintaining the same TCP fraction leads to a notable friction decrease compared to undiluted SF with TCP particles. This friction reduction can be explained by the dilution and lubrication properties of xylitol and by the water sorption qualities of polyols and xylitol [24,25]. These results highlight that unlike findings reported in the literature with pin-on-disc tested technique designing [15], a new experimental technique allowing capture of the particles suspended in the synovial fluid within the interface between the mating surfaces can lead to more objective results in terms of friction. In addition, comparing the values of friction obtained in the present study with those of comparable work reported in the literature, it appears that the dynamic CoF obtained for the pure synovial fluid is of the same order as those reported in Refs. [11,16,17]. However, it was impossible to compare the results obtained for the synovial fluid charged with particles because to the best of our knowledge, there is no equivalent work reported in the literature dealing simultaneously with both the presence of suspended particles and the reopening of contact at the end of each cycle to allow the covering of the surface with the synovial fluid charged with particles.

Conclusions
A new experimental test rig has been adapted to evaluate lubrication properties of synovial fluid containing or not containing TCP particles to simulate synovial joint OA disease. The friction pair was composed of a PTFE rod with flat contact sliding against a flat and rigid glass plate. Five different lubricants were tested: (i) PBS; (ii) GLY; (iii) SF; (iv) synthetic synovial fluid containing β-TCP granule particles used to simulate the presence of worn tissue debris in the synovial fluid in OA patients; and (v) diluted synovial fluid by xylitol with TCP particles to simulate possible future treatment in OA patients. The friction experiments were conducted under a normal load of 5 N and sliding velocity of 0.5 mm/s The following observations were made.
i) Very good reproducibility is obtained with the successive friction cycle repetition.
ii) The excellent reproducibility of friction curves in the case of synovial fluid with microparticles indicates that opening the contact at the end of each friction cycle allows the lubricant charged with particles to be reintroduced in the contact, which seems to simulate the physiological conditions of a natural knee joint in OA patients during walking more objectively.
iii) Capturing the suspended worn particles in the synovial fluid inside the interface between the mating surfaces led to different results in terms of friction. This proves that the OA synovial fluid changes its lubrication properties due to the presence of worn fragments. iv) Diluting the synovial fluid with xylitol while maintaining the same TCP fraction leads to a friction decrease compared to undiluted SF.
v) This experimental setup, tribological testing method, and orthopedic data can be successfully applied to examine OA joints and objectively follow the treatment progress.
vi) Tribology testing of synovial fluid from OA joints can be an objective parameter in classifying OA joints.