Nano-Scale Mechanical Properties of the Articular Cartilage Zones in a Mouse Model of Post-Traumatic Osteoarthritis

Featured Application

Nano-scale IT-AFM is a sensitive tool to monitor biomechanical changes during the course of PT-OA.

Abstract

Destabilization of the medial meniscus (DMM) surgery in mice is used to elucidate the mechanism of post-traumatic osteoarthritis (PT-OA). The study of cartilage biomechanics in PT-OA is important for understanding the pathophysiology of the condition. We used indentation-type atomic force microscopy (IT-AFM) to assess the nanostiffness of the interterritorial matrix of articular cartilage (AC) zones in the medial and the lateral tibia plateau (MTP and LTP) on native tissue sections 2 and 8 weeks after DMM or Sham surgery. At 2 weeks, pronounced stiffening of the DMM AC was observed compared to Sham, with the most marked changes occurring in the superficial zone and affecting the proteoglycan moiety rather than the collagen network. The LTP cartilage was obviously stiffer than the MTP in DMM, but not in Sham. At 8 weeks, only modest differences in nanostiffness were observed between DMM and Sham. The difference in stiffness between MTP and LTP was reduced, and the proteoglycan and collagen phases changed in a more similar manner. Interestingly, the deep zone was softer in the DMM compared to the Sham. Sham AC showed an increase in stiffness between 2 and 8 weeks, a trend that was counteracted in the DMM group. Collectively, our study demonstrates that nano-scale IT-AFM is a sensitive tool to monitor biomechanical changes during the course of PT-OA.

1. Introduction

Osteoarthritis (OA) is the leading musculoskeletal disease worldwide, and is characterized by a complex and multifactorial etiology [1,2]. Although extracellular matrix (ECM) degradation associated with structural destruction of the articular cartilage (AC) represent the hallmark of the disease, other joint tissues such as the synovium, subchondral bone, meniscus, tendons and ligaments are also affected and can contribute to the progression of the condition, cumulatively resulting in joint swelling, pain and immobility [3]. High-risk factors affecting the onset and progression of OA include increasing age, obesity and/or metabolic conditions, joint overuse and injuries of the musculoskeletal system [4]. All these factors may lead to altered biomechanical loading, which progressively contributes to the imbalance of cartilage homeostasis, eventually resulting in AC degeneration and OA. To date, no cure or disease-modifying drug is available for OA and therapeutic interventions aim to relieve symptoms or eliminate the problem via joint replacement [5]. Therefore, understanding cartilage physiology and pathophysiology is of paramount importance to develop better therapeutic modalities for the treatment of OA.

Articular cartilage is composed of a specialized ECM and chondrocytes, a unique cell type that makes up only 5–10% of the total tissue volume. The cartilaginous ECM consists primarily of proteoglycans (mainly aggrecan) and type II collagen fibrils, which are inter-connected with adaptor proteins and build networks to withstand compressive and tensile forces [6,7]. Topically, the ECM of AC is composed of a narrow pericellular matrix (PCM) immediately surrounding the chondrocytes and the extensive territorial/inter-territorial matrix (TM/ITM), each of which contribute in a unique way to the chemical and mechanical features of the AC [8]. Based on its structural and functional properties, healthy articular cartilage can be divided into four distinct vertical zones, namely the superficial, the medial, the deep and the calcified zones. Each zone is characterized by different chondrocyte shape, ECM composition and organization, which ultimately endow the tissue with specific zonal biomechanical properties [9]. The interplay between ECM composition, biomechanics and OA is still not fully understood, although it is generally accepted that impaired mechanical properties of the articular cartilage drive its degeneration. Thus, it is of utmost importance to investigate the biomechanics of the cartilaginous ECM in the different zones during the development of spontaneous or injury-related OA.

The destabilization of the medial meniscus (DMM) is a surgical model in mouse, which was developed to mimic human post-traumatic osteoarthritis (PT-OA) caused by joint instability [10]. The DMM model has been widely used to assess OA pathophysiology, but little consideration has been given to the biomechanics of AC upon DMM [11]. The fine details of the ultrastructural and mechanical properties of the AC have been mostly evaluated using the gold standard, indentation-type atomic force microscopy (IT-AFM) [12]. Utilizing a nanometer-sized tip, it is possible to distinguish between the stiffness of collagen fibers and proteoglycans [13,14]. However, ambiguous results on AC stiffness during OA progression are reported in the literature, probably depending on different experimental set-up variables, e.g., the geometry and size of the indenter, the anatomical location of the indentation test, and measurements on tissue surfaces or cross-sections, as well as the diverse OA models investigated. For example, an AFM-based indentation study using a micrometer-sized spherical tip on native mouse femoral condyle cartilage reported reduced stiffness of the AC surface at the onset and during the progression of PT-OA [15]. Another study has assessed mouse tibial AC on cross-sections and observed reduced micro-scale modulus of the PCM in the middle/deep zone during DMM-induced PT-OA [16]. Furthermore, surface indentation studies on human cartilage from patients undergoing total knee arthroplasty have shown a decrease in stiffness with increasing severity of OA [17,18]. In contrast, a study using nano-scaled AFM indentation on cross-sections of human AC revealed stiffening of collagen fibrils with increasing OA grade in all zones [19], whereas another study showed a decrease in elasticity in both the ITM and PCM [20]. Of interest is the different indenter geometries used, which makes it difficult to compare the results obtained in all the above-mentioned studies [13,21,22]. Additionally, several biomechanical investigations on murine knockout models revealed a stiffness gradient depending on the articular cartilage zone when measured on cross-sections [23,24,25,26]. It is known that the superficial layer of the AC is often the first part of the joint tissue that is degenerated, and therefore exhibits altered biomechanics, but what happens to the deeper zones of AC is also of paramount importance for the fine elucidation of the pathophysiology of OA [15]. The present study aims to decipher the depth- and time-dependent biomechanical alterations of tibia plateau articular cartilage during early murine PT-OA using nano-scaled IT-AFM.

2. Materials and Methods

2.1. Destabilization of the Medial Meniscus (DMM) Surgery

Male mice on a C57Bl/6 J background (Charles River Laboratories, Sulzfeld, Germany) were used in this study. PT-OA was induced by surgical destabilization of the medial meniscus [10] as we previously described in detail [26]. Briefly, 6 mice were operated at 12 weeks of age and sacrificed for analysis at 2 and 8 weeks post-operation. Sham surgery, consisting in the solely visualization of the ligament, served as control group (3 animals per time point). The mice were kept in standard housing conditions under a 12 h light/dark cycle and had access to food and water ad libitum. All animal experiments were approved by the ethics committee of the local authorities (District Government of Lower Franconia, Bavaria, Germany; approval number: 55.2-2531-2-289; date of approval: 27 July 2016).

2.2. Tissue Processing

Immediately after sacrification, the entire knee joints of Sham and DMM mice were dissected, de-skinned and directly immersed into Tissue Tek cryomedia in a plastic mold (Sakura, Zoeterwoude, NL, USA) and gradually frozen on a chilled copper plate placed on dry ice. Coronal sections of 20 µm were cut with the aid of double adhesive tape (as described in [26]) using a CM 1950 cryotome (Leica, Wetzlar, Germany), mounted on glass slides and kept in a −20 °C freezer until histological or biomechanical analysis. Both AFM measurements and histopathological evaluation were performed on these native sections in order to (1) reduce the number of mice needed for experiments, and (2) to obtain a more reliable comparison of the biomechanical and chemical/structural properties of articular cartilage from the same sample. Our initial analysis revealed that the use of thick sections for histology did not compromise the grading of OA.

2.3. Safranin-O Staining and Histological OA Score

Frozen sections were thawed for at least 30 min. Afterwards, the slides were hydrated in distilled water for 2 min, stained with 0.2% Fast Green/H₂O (Sigma-Aldrich, Taufkirchen, Germany) for 3 min and immersed in 1% acidic alcohol for 30 s (Roth, Karlsruhe, Germany). Next, sections were stained in 0.5% Safranin-O solution/H₂O (Sigma-Aldrich) for 2 min, briefly washed in distilled water and immersed in 1% acetic acid (Roth) for 30 s. Afterwards, slides were briefly rinsed in 100% ethanol (Roth), and washed for 2 min once in 100% ethanol and once in 100% isopropanol (Roth). Finally, slides were cleared in Roti-Histol (Roth) for 5 min and mounted with Roti-Histokitt (Roth).

Imaging was performed with the AxioObserver Z1 equipped with the AxioCam MRm color camera (Carl Zeiss, Jena, Germany) using a 10x objective and the mosaic stitching tool of the microscope software Zen 2.3 lite (Carl Zeiss). Scoring of osteoarthritic changes was conducted by three independent, blinded observers according to a modified Mankin scoring system described by Nicolae et al. 2007 [27]. The scoring system includes 5 categories as follows: I. Cartilage erosion (0–5: 0, normal; 1, surface irregularities; 2, cleft to transition zone; 3, cleft to radial zone; 4, cleft extending to calcified zone; 5, exposure of subchondral bone); II. Cellularity (0–3: 0, normal; 1, diffuse hypercellularity; 2, clustering; 3, hypocellularity); III. Tidemark integrity (0–1: 0, normal; 1, loss of integrity); IV. Osteophyte formation (0–2: 0, none; 1, formation of cartilaginous tissue; 2, endochondral ossification); V. Safranin O staining (0–3: 0, normal; 1, reduced staining; 2, focal patchy loss of staining; 3, no staining). The total scale (0–16) and OA severity were as follows: 0–1, normal; 2–5, mild OA; 6–11, moderate OA; 12–16, severe OA. The medial and lateral tibia plateau were independently scored on 3 to 4 sections spaced approximately 100 µm apart, thus representing the whole knee joint, including regions adjacent to the sections used for IT-AFM measurements. The maximum score recorded among all the stained sections of each specimen was averaged with the values of all observers. The results shown in the graph represent the media ± standard deviation of three independently scored mice.

2.4. Picrosirius Red Staining

Frozen sections were thawed for at least 30 min and rehydrated/washed three times for 5 min in PBS. Nuclei were counter-stained with Weigert’s Haematoxylin solution for 2 min and then washed for 5 min in tap water. Afterwards, sections were stained for one hour at room temperature in 1.3% Direct Red 80 (also known as Sirius Red, Alfa Aesar, Kandel, Germany) dissolved in a saturated aqueous solution of picric acid (Sigma-Aldrich) for visualization of collagen. Sections were then washed twice in 0.5% acetic acid, dehydrated in three changes of 100% ethanol cleared in two changes of Roti-Histol and mounted in Roti-Histokitt. Polarized light images were taken with the AxioObserver Z1 using 10× objective and the AxioCam MRm color camera.

2.5. Indentation-Type Atomic Force Microscopy (IT-AFM)

IT-AFM was carried out as described in Muschter et al., 2020 [26]. A NanoWizard I AFM (JPK Instruments, Berlin, Germany) was used in combination with an inverted optical microscope (Axiovert 200, Carl Zeiss Micro Imaging GmbH, Göttingen, Germany) for accurate lateral positioning of the AFM-tip. The entire set-up was placed on an active vibration isolation table (Micro 60, Halcyonics, Göttingen, Germany) inside a 1 m³ soundproof box to reduce external noise. Indentation experiments were performed using silicone-nitride cantilevers (MLCT, Cantilever E, Bruker, Karlsruhe, Germany) with a nominal spring constant of 0.1 N/m, and sharp pyramidal tips with a nominal tip radius of 20 nm. For each cantilever, the spring constant was determined individually using the thermal noise method [28]. During measurements, the tissue sections were immersed in PBS (Dulbecco PBS w/o Mg²⁺/Ca²⁺, pH 7.4, Biochrom, Berlin, Germany). The vertical tip velocity during the indentation experiments was 10 µm/s. The elasticity of the ITM of each cartilage zone (superficial, middle and deep) was assessed on two different tissue sections with a total of 1875 force-indentation curves distributed over 3 different 9 µm² areas. Subsequently, a modified Hertz–Sneddon model was fitted onto the approach curve up to 500 nm using the JPK Data Processing Software (Version 5.0.96, JPK Instruments), and the Young’s modulus (E) was extracted. The resulting Young’s modulus values were summarized in histograms using the Igor Pro software (Version 6.3.7.2, WaveMetrics, Portland, OR, USA) and the two maxima of the bimodal distributions were determined by fitting a linear combination of two Gaussian distributions to the data. Previous studies observed similar bimodal nano-stiffness distributions in cartilaginous structures, where the first peak was attributed to the proteoglycan phase and the second peak to the collagen fibrils [13,14].

2.6. Statistical Analysis

Statistical analysis was performed by 2-sided t-test with normally distributed datasets, such as the modified Mankin score (Figure 1) and differences in the proteoglycan and collagen peak moduli obtained by IT-AFM (Figure 5). For combined Gaussian distribution stiffness values, a Mann–Whitney-U-test was used as the dataset was bimodal and not normally distributed (Figure 4). Statistical significance was assumed at a p-value of ≤ 0.05.

3. Results

3.1. Histological Scoring of AC Degeneration after DMM

For qualitative analysis of osteoarthritic changes, Safranin-O/Fast Green staining was performed at 2 and 8 weeks after Sham or DMM surgery (Figure 1A), and a modified Mankin scoring system [27] was applied to quantitatively assess the degeneration of the medial and lateral tibial plateau (MTP and LTP) of the knee AC (Figure 1B). Two weeks after surgery, a significant difference (p = 0.002) in modified Mankin scores was observed in the MTP between the Sham- (1.33 ± 0.47) and the DMM-operated mice (4.67 ± 0.47). At this early-OA stage, significantly more severe cartilage degeneration of the MTP compared to the LTP was also detected in the DMM group (respectively, 4.67 ± 0.47 and 2.33 ± 0.47; p = 0.008). Eight weeks after surgery, focal absence of Safranin-O staining indicative of proteoglycan loss was observed in the MTP of DMM mice (Figure 1A, white arrow) but AC degradation scores were only slightly elevated compared to the 2 weeks post-operation group (Figure 1B), implicating only a mild progression of OA. The significantly different Mankin scores in the medial tibial compartment remained between the Sham- and the DMM-operated group (2.00 ± 0.00 versus 5.67 ± 1.70, p = 0.038); however, we only observed a tendency of higher Mankin scores for MTP compared to LTP in DMM mice, which did not reach the level of significance (respectively, 5.67 ± 1.70 and 2.67 ± 0.47; p = 0.121). In Sham mice, the modified Mankin scores for MTP and LTP were not statistically different in either age group. Picrosirius red staining combined with polarization light microscopy demonstrated comparable and intact collagen fibril organization in the MTP cartilage of Sham mice at 2 weeks and 8 weeks post-surgery (Figure 1C). However, in DMM mice, at 8 weeks postoperatively, we observed discontinued birefringence at the superficial zone (Figure 1C, white arrows) and reduced signal intensity in the middle and deep zones (Figure 1C), suggesting that the progression of OA is associated with degradation or structural disorganization of the collagenous network.

3.2. Analysis of AC Biomechanics upon DMM

Nano-scale IT-AFM measurements in the ITM were performed on the MTP and LTP to assess the biomechanical properties of the native articular cartilage 2 and 8 weeks after Sham or DMM surgery. Histograms of the Young’s moduli demonstrated bimodal stiffness distribution in each cartilage zone (deep, middle and superficial), representing the proteoglycan (peak E1) and collagen (peak E2) phases of the tissue ECM (Figure 2 and Figure 3). In all samples, the cumulative proteoglycan and collagen mean stiffness values (E) displayed a decreasing gradient from the deep towards the superficial zone (Figure 2, Figure 3, Figure 4 and Figure 5), reflecting the composition and function of the AC. Two weeks after surgery, DMM mice showed an approximately 1.4-fold increase in the mean cumulative stiffness E in all three AC zones (E_DZ: 1.40-fold; E_MZ: 1.35-fold; E_SZ: 1.43-fold) of the MTP compared to the Sham group (Figure 2A and Figure 4). The individual stiffness for the proteoglycan (E1) and collagen (E2) networks in the deep zone MTP of the DMM animals increased 2.04-fold and 1.71-fold, respectively, compared to Sham (Figure 2A and Figure 5). In the middle zone, a 1.78-fold increase for E1 and a 1.51-fold increase for E2 were observed in DMM MTP compared to Sham MTP. The superficial zone of the MTP showed the highest DMM-versus-Sham shift with increasing E1 from 127 kPa to 287 kPa (2.25-fold) and E2 from 299 kPa to 567 kPa (1.99-fold). Interestingly, the LTP showed a higher cumulative stiffening (E_DZ: 2.02-fold; E_MZ: 1.98-fold; E_SZ: 2.28-fold) in DMM mice compared to Sham than the medial side (Figure 2B and Figure 4). There was a general increase of LTP stiffness for both macromolecular networks, especially for the proteoglycans, in DMM animals compared to Sham in the deep zone (E1: 2.13-fold; E2: 1.98-fold), the middle zone (E1: 2.59-fold; E2: 2.17-fold) and the superficial zone (E1: 4.52-fold; E2: 2.76-fold). Taken together, nano-scale AFM indentation measurements demonstrated that in this very early phase of PT-OA, the articular cartilage zones in DMM mice were stiffer compared to Sham animals, the most pronounced biomechanical changes occurred in the superficial zone, the proteoglycan phase of cartilage ECM was more affected than the collagenous network, and the lateral DMM compartment was stiffer than the medial DMM compartment. On the other hand, cartilage stiffness in the corresponding articular cartilage zones of the lateral and medial cartilage compartments of Sham mice was comparable.

At 8 weeks post-operatively, overall minor changes were observed compared to those detected in the 2 weeks group. When comparing DMM to Sham in MTP, we recorded a mild softening of the deep zone (E: 0.84-fold; E1: 0.70-fold; E2: 0.78-fold) and a slightly increased Young’s modulus of the middle (E: 1.58; E1: 1.04-fold; E2: 1.35-fold) and superficial (E: 1.12; E1: 1.32-fold; E2: 1.22-fold) zones (Figure 3A, Figure 4 and Figure 5). However, in the LTP, a moderate stiffening in all non-calcified AC zones (E_DZ: 1.63-fold; E_MZ: 1.34-fold; E_SZ: 1.13-fold) was observed upon DMM surgery compared to Sham (Figure 3B, Figure 4 and Figure 5). The stiffness of the proteoglycan moiety and the collagen network in the deep zone of the DMM increased 1.42-fold and 1.50-fold, respectively, compared to the Sham. The mean individual stiffness values increased by 1.44-fold (E1) and 1.19-fold (E2) in the middle zone, and by 1.69-fold (E1) and 1.38-fold (E2) in the superficial zone relative to Sham. These data demonstrate that, in contrast to the 2 weeks, in the 8 weeks post-OP group, the stiffness of the proteoglycan and collagen networks changed more similarly; the stiffness difference between the LTP and MTP of the DMM mice was reduced (LTP/MTP 1.50-fold (E_2w) and 1.15-fold (E_8w), with E averaged over all three zones); and the MTP deep zone was softer in DMM compared to Sham. Of note, the stiffness of the LTP deep zone in DMM mice showed a 1.34-fold increase, the middle zone showed no differences, whereas the LTP superficial zone had a reduced stiffness (0.68-fold) compared to the corresponding MTP zone. Unexpectedly, we recorded softening of the deep and superficial zones of the Sham LTP compared to the Sham MTP (E_DZ: 0.68-fold; E_SZ: 0.68-fold), while the middle zone LTP slightly stiffened compared to the MTP middle zone (E_MZ: 1.19-fold).

Analyzing the biomechanical differences between the 2 and 8 weeks post-operative time points, we first observed a general tissue stiffening of the Sham articular cartilage, with a trend of change that was attenuated in the DMM experimental group. The three-zone averaged cumulative stiffness increases were 1.90-fold (E_MTP) and 1.48-fold (E_LTP) for Sham, and 1.55-fold (E_MTP) and 0.98-fold (E_LTP) for DMM. The largest time-dependent (8w/2w) stiffening of Sham was recorded in the deep zone of the MTP (E_DZ: 2.17-fold) with remarkable individual increases: E1: 3.19-fold; E2: 2.39-fold. The middle and superficial zones showed 1.46-fold (E_MZ)/1.88-fold (E1_MZ)/1.55-fold (E2_MZ) and 2.11-fold (E_SZ)/1.52-fold (E1_SZ)/1.67-fold (E2_SZ) changes, respectively. On the other hand, zone-dependent stiffness variation between 8 and 2 weeks Sham were more homogenous on the lateral side (E_DZ: 1.56-fold; E_MZ: 1.65-fold; E_SZ: 1.27-fold). The tendency towards stiffening observed in the Sham group was reduced in the DMM group, especially in the deep and superficial zones of the MTP. For MTP, increases in stiffness of 1.29-fold (E_DZ), 1.71-fold (E_MZ) and 1.66-fold (E_SZ) were observed. For LTP, 1.26-fold (E_DZ), 1.12-fold (E_MZ) and 0.63-fold (E_SZ) changes were recorded. Analysis of individual proteoglycan and collagen networks revealed superficial zone softening in the range of 0.89-fold (E1) and 1.07-fold (E2) (unchanged) for MTP and 0.49-fold (E1) and 0.65-fold (E2) for LTP. Collectively, IT-AFM measurements demonstrated that DMM surgery induced processes that counteracted the age-dependent stiffening of articular cartilage zones observed in Sham animals.

4. Discussion

The aim of this study was to analyze the biomechanical alterations in murine articular cartilage during the course of PT-OA induced by the DMM model. It is well accepted that during OA, changes in the biomechanical properties of the cartilage ECM precede histologically visible degradation [15,17,24,26,29,30]. In this regard, we could clearly demonstrate pronounced changes in AC tissue stiffness at the onset and progression of PT-OA in DMM animals compared to Sham-operated groups, whereas histology and the application of a modified Mankin score could reveal only mild differences. For histopathological scoring, we used 20 µm thick, native, non-decalcified sections, whereas the recommended guidelines for histology suggest 4–6 µm thin, fixed, decalcified sections [31]. Nevertheless, the OA histological scores obtained showed a good comparison with our previous study [26]. After thawing and hydration, native frozen sections of 6–30 µm are also routinely used to evaluate the biomechanical properties of cartilaginous tissues by IT-AFM [14,24,25,26,30]. Similarly to previous studies [13,14,24,25,26], we could identify the typical cartilaginous bimodal stiffness distribution, where the first peak represents the softer proteoglycan phase and the second peak the stiffer collagenous network. The high comparability achieved by analyzing adjacent sections, and the fact that the same animal could be used for both histological and AFM investigations, further justifies our methodological approach to use thick, native sections.

Articular cartilage ECM exhibits a decreasing stiffness gradient from the deep to the superficial zone, regardless of species [19,24,25,32,33]. This zone-specific biomechanical property, reflecting the composition, structure and function of AC regions, was confirmed in Sham mice by our nano-scale AFM indentation, which showed the highest Young’s modulus in the deep zone and the lowest in the superficial zone for both the proteoglycan gel and the collagen fibrillary network. Furthermore, the time-dependent increase from weeks 2 to 8 in cartilage nanostiffness in Sham animals is indicative of tissue maturation, which is consistent with the previously reported stiffening associated with skeletal development and aging, and is most likely attributable to the maturation of the collagenous network [14,17,34,35].

At 2 weeks after surgery, the nano-scale stiffness in all non-calcified AC zones of MTP and LTP of DMM mice was significantly elevated compared to Sham, which was associated with a significant increase in histological grade in MTP from normal AC (Sham, 1.33 ± 0.47) to mild OA (DMM, 4.67 ± 0.47), and with no significant histopathological change in LTP (Sham: 2.0 ± 0.00; DMM: 2.33 ± 0.47). Previous studies by surface AFM indentation showed significant changes in the stiffness of the intact medial condyle as early as 1 week post-operatively between DMM and Sham, whereas significant histopathological differences were observed only from 8 weeks post-DMM [15,36]. In contrast, Fang et al. reported significant differences in OARSI scores between DMM and Sham in MTP 2 weeks after surgery; however, concomitant changes in bone mineral density of the subchondral bone were significant only from 5 weeks [37]. Using native knee sections, micro- and nano-scale AFM indentation experiments revealed significant modulus changes from 3 days in the PCM and from 1-2 weeks in the ITM, preceding obvious histological differences at 4–8 weeks post-DMM [26,30]. In the present and the aforementioned studies, DMM surgical procedures were performed in 12-week-old male mice, reducing experimental variations between studies that may be caused by the age or gender of the operated mice [38,39]. Thus, our results, which show early changes in cartilage stiffness after DMM, confirm that biomechanical changes are among the first to occur during the onset of PT-OA [15,29].

In contrast to skeletal maturation-related stiffness alterations, changes in AC biomechanics during the onset and progression of OA are still a matter of debate, with conflicting results in the literature. Applying AFM with a micro-scale flat (100 µm) indenter on human articular cartilage specimens, a decreasing cartilage shear modulus was reported with increasing OA grade [18]. In human, an inverse correlation between stiffness and higher OA grades was demonstrated with IT-AFM by using pyramidal nano-scale tips (nominal radius = 30 nm) but not with micro-scale spherical tips (radius = 2.5 µm) [17]. Regarding PT-OA, probing the surface with spherical AFM tips, a markedly lower stiffness of the mouse medial condyle was found from 1 to 8 weeks after DMM compared to Sham [15,36]. Interestingly, at 12 weeks post-surgery, which is considered as a late OA time point, the effective indentation modulus of the DMM AC had dramatically increased compared to the earlier stages and approached the stiffness level of the Sham [15]. This unexpected stiffness increase at late PT-OA could be due to erosion of the surface exposing deeper layers of the articular cartilage with higher compressive modulus. On the other hand, in the same study, the lateral condyle appeared to be significantly softer only at 4, 8 and 12 weeks post-DMM compared to Sham [15]. More recently, micromechanical properties of the middle/deep zone of the tibial plateau cartilage were assessed on cross-sections during the course of DMM-induced PT-OA between 3 days and 8 weeks [30]. AFM-based indentation using spherical tips (with a radius of 2.25 µm) revealed reduced stiffness from 3 days in the PCM and 1 week in the ITM in DMM animals compared to Sham controls [30], suggesting that OA is potentially initiated in the pericellular matrix [40]. In contrast, our AFM indentation with a sharp, pyramidal tip (radius = 20 nm) in the ITM revealed higher stiffness of both the MTP and LTP in DMM animals compared to Sham 2 weeks after surgery. At this early stage of PT-OA, proteoglycan structural changes and GAG depletion may represent alterations that could affect cartilage biomechanics. Proteoglycan degradation through cathepsin D digestion in porcine articular cartilage samples resulted in stiffening of the matrix detected via nano-scale AFM indentation [21]. We previously demonstrated by nano-scale IT-AFM that reduced level of the major proteoglycan aggrecan in a mouse genetic model leads to stiffening of the AC zones of the tibial articular cartilage at 6 months of age, predisposing to OA-like cartilage degeneration at 12 months of age [24]. Furthermore, we showed that the loss of the small leucine-rich proteoglycan Decorin in mice increases proteoglycan nanostiffness and, to a lesser extent, collagen nanostiffness, in the articular cartilage [41]. Thus, the tissue stiffening that we observed in DMM mice 2 weeks post-surgery could be the biomechanical manifestation of the initial, subtle proteoglycan loss typical of early-OA. The contrasting results for AC biomechanics in OA studies may be related to the indenter geometry, where, as suggested previously, measurements using micro-scale spherical tip reflect integrated responses of cartilage ECM composition and structure, whereas nano-scale indentation using pyramidal tip represents the local nanomechanics of the ECM proteoglycan gel and collagen network [17,42,43].

Between 2 weeks and 8 weeks post-operatively, a marked stiffening of both MTP and LTP was observed in Sham animals, which can be explained by the maturation of AC tissue as described above. In DMM mice, however, there was only a modest increase in overall indentation modulus in the MTP and no change in stiffness in the LTP. The normal age-related increase in nanostiffness, especially in the MTP, was accompanied, by a stronger birefringence of the collagen network in Sham, while in DMM samples, polarized light signals were apparently reduced at 8 weeks compared to 2 weeks. As the presence and intensity of birefringence is indicative of the organization and density of collagen fibrils, polarization light microscopy is a tool to assess the structural disorganization and degradation of the collagen network, associated with AC degeneration [44]. We previously showed that limited collagenase digestion of bovine articular cartilage explants leads to reduced birefringence and nanostiffness of the surface zone [45]. Likewise, Stolz et al. demonstrated with micro-scale IT-AFM that collagen depletion with prolonged elastase treatment reduces stiffness in the surface of pig AC [21]. During the course of DMM-induced PT-OA, increased degradation of collagen was reported [46]; therefore, it is tempting to speculate that the eventual structural deterioration of the collagen network in DMM mice at 8 weeks counteracts ageing-associated stiffening, resulting in only a modest change of the nano-scale indentation modulus from 2 weeks to 8 weeks post-surgery.

Osteoarthritis is a whole joint disease in which both the lateral and medial cartilage compartments are compromised. A realistic total human knee finite element model showed that medial meniscus damage also affects the lateral AC [24], and in a DMM PT-OA model, the micro-scale stiffness of the lateral condyle cartilage was significantly reduced from 4 weeks after surgery [15]. In the present study, we found that at 2 weeks post-surgery, the Young’s moduli of MTP and LTP in Sham were in a similar range, whereas the LTP after DMM showed markedly increased stiffness compared to MTP, despite the significantly higher Mankin score in the MTP. Further investigation is needed to wisely interpret why the lateral side shows higher stiffening and degrades slower compared to the medial side in the DMM model [10].

Lastly, this study has some limitations that could be addressed in follow-up experiments. As has been widely discussed, the size and geometry of the indenter is a critical issue when measuring biomechanical properties of biological tissues with IT-AFM. Therefore, it would be interesting to directly compare nano-scale and micro-scale stiffness data, obtained by sharp pyramidal or spherical tips, respectively, on the same sample. OA may start in the PCM than spread to other ECM compartments. In our study, we focused on the ITM; therefore, it would be informative to complement our data with the biomechanical changes in the PCM. Degradation and structural modification of ECM components underlie the biomechanical changes that occur during the onset and progression of OA. For a better understanding of how changes at the molecular level are responsible for altering tissue biomechanics, a detailed immunohistochemical investigation of typical pro-teases and degradation epitopes of the major AC proteins is desirable. Finally, in this study, we only examined the early and middle stages of PT-OA development (namely, 2 and 8 weeks post-surgery) to challenge the ability of IT-AFM to detect biomechanical changes. It would also be useful to extend our analysis to later post-operative time points, when AC degradation is more advanced and, presumably, biomechanical changes are more pronounced.

5. Conclusions

To the best of our knowledge, this is the first detailed study to use IT-AFM to determine spatial and temporal biomechanical changes at the nano-scale during murine PT-OA. Our study shows that nano-stiffness of the interterritorial matrix of the articular cartilage increases in a time- and depth-dependent manner during DMM-induced PT-OA. The Young’s moduli for collagen and proteoglycans showed increasing stiffness gradients from the superficial zone to the middle and deep zones in both Sham and DMM cartilage. The increase in stiffness 2 weeks after DMM was more pronounced than in Sham, likely due to proteoglycan loss associated with the early status of OA. The age-related increase in nano-scale AC stiffness was partially abolished 8 weeks after DMM due to structural deterioration of the collagen network. Interestingly, the tendency for stiffness to increase was greater in the lateral compartment, even when little or no OA-like degeneration was observed histologically. Last but not least, with this study, we further support the importance of nano-scaled IT-AFM as a complementary methodology for understanding OA pathophysiology and diagnosing the early phase of the disease.