Automated Indentation Demonstrates Structural Stiffness of Femoral Articular Cartilage and Temporomandibular Joint Mandibular Condylar Cartilage Is Altered in FgF2KO Mice

Abstract

Objective

Employ an automated indentation technique, using a commercially available machine, to assess the effect of fibroblast growth factor 2 (FGF2) expression on structural stiffness over the surface of both murine femoral articular cartilage (AC) and temporomandibular joint (TMJ) mandibular condylar cartilage (MCC).

Design

Experiments were performed using 3-month-old female homozygote Fgf2KO mice with wild type (WT) littermates. After euthanization, isolated mandibles and hindlimbs were either processed for histology or subjected to automated indentation on a Biomomentum Mach-1 v500csst with a 3-axis motion controller in a phosphate buffered saline bath using a 0.3 mm spherical tip indenter. The effect of indentation depth on normal force was characterized, then structural stiffness was calculated and mapped at multiple positions on the AC and MCC.

Results

Automated indentation of the AC and TMJ MCC was successfully completed and was able to demonstrate both regional variation in structural stiffness and differences between WT and Fgf2KO mice. Structural stiffness values for Fgf2KO AC were significantly smaller than WT at both the medial/anterior (P < 0.05) and medial/posterior (P < 0.05) positions. Global Fgf2KO also lead to a decrease in MCC thickness of the TMJ compared with WT (P < 0.05) and increased structural stiffness values for Fgf2KO at both the posterior and anterior location (P < 0.05).

Conclusions

Automated indentation spatially resolved differences in structural stiffness between WT and Fgf2KO tissue, demonstrating FGF2 expression affects femoral AC and TMJ MCC. This quantitative method will provide a valuable approach for functional characterization of cartilage tissues in murine models relevant to knee joint and TMJ health and disease.

Introduction

Degenerative disorders affecting the mandibular condylar cartilage (MCC) of the temporomandibular joint (TMJ), and the articular cartilage (AC) of the knee joint, are common chronic conditions, occurring in 80% of the population older than 65 years. ¹ Other than trauma, aging is the main risk factor for the degenerative disorders of the TMJ and the knee joint, and both conditions impair quality of life and are costly to society.^1,2 While both are cartilage, the TMJ MCC and knee AC have distinct developmental origins and are made of different types of cartilage. They are composed of unique chondroprogenitor and chondrocyte populations as well as matrices that differ in molecular composition, structure, and mineral content. The MCC is a hybrid of fibrocartilage and hyaline cartilage, with the fibrocartilage layer at the surface covering the hyaline cartilage below, ³ while the knee AC is hyaline cartilage. Moreover, the expression profiles of the chondroprogenitors and chondrocytes of the MCC and knee AC are also different, with some genes expressed by both, but some only expressed by one or the other tissue. ⁴ Finally, the chondrocytes of the MCC and knee AC exhibit different cellular behaviors like proliferation, differentiation, and matrix synthesis. These aspects of the MCC and knee AC underlie their different responses to factors such as age, injury, and mechanical loading, and provide the rationale to study of both in the context of degenerative joint disease.

Fibroblast growth factor 2 (FGF2) is a critical growth factor in maintaining the homeostasis of the knee joint^5-7; however, to date there are no studies on its role in maintenance of the TMJ. Alteration in the expression and activity of FGF2 is associated with joint-related disorders.^5,8,9 Studies in the literature ⁷ and our studies, ⁶ using global Fgf2KO ¹⁰ mice showed that these mice develop signs of degeneration, assessed histologically, in the AC of knees at 6 months of age. However, quantitative characterization of the mechanical properties of the AC in these mice has never been done, nor has the potential degeneration of the TMJ in these Fgf2KO mice ever been examined.

While the common, qualitative, approach to assess the integrity of the knee AC and TMJ MCC is histological evaluation, there is value in mechanical indentation as it provides quantitative, and complementary, data. While classic histological methods can elucidate composition and tissue level structural changes of cartilage tissues, they do not assess changes in biomechanical function. Automated indentation has been employed in preliminary studies using a commercially available mechanical testing machine to assess murine knee AC ¹¹ and porcine TMJ ¹² to provide quantitative characterization and mapping of the mechanical response of the cartilage. However, this has never been attempted on the significantly smaller murine TMJ, nor combined to evaluate both AC and TMJ MCC in the same mouse model to assess potential similarities or differences in their degeneration. Indeed, a single previous study employed atomic force microscopy (AFM) nanoindentation on murine TMJ disc and MCC to characterize the micromechanical properties, and suggested conventional macroscopic biomechanical tests were not applicable to the small murine TMJ. ¹³ The objective of this study here was to employ an automated indentation technique, using a commercially available machine, to assess the effect of FGF2 expression on structural stiffness over the surface of both murine femoral AC and TMJ MCC. While previous studies did not observe histological changes in Fgf2KO AC at 2 months ⁶ or even 3 months ⁷ of age, given that automated indentation of human cartilage has been shown to be more sensitive to early degenerative changes in cartilage than histology, ¹⁴ we hypothesize automated indentation will detect diminished structural stiffness in 3 month old Fgf2KO AC along with contaminant changes in the never before studied TMJ MCC.

Methods

Mice

Fgf2KO mice were previously developed and characterized on a Black Swiss 129 Sv background. ¹⁰ All experiments were performed using 3-month-old female homozygote mice with wide type (WT) littermates. Mice were killed utilizing the approved CO₂ method for euthanasia. All procedures were approved by the UConn Health Institute of Animal Care and Use Committee.

Automated Indentation of Temporomandibular and Femoral Joint Cartilage

After the mice were euthanized, mandibles and femurs were isolated. Mandibles and hindlimbs were processed for histology as described below, or stored at −20 °C in phosphate buffered saline (PBS) until biomechanical testing. For testing, samples were thawed at room temperature then mounted for testing. Specifically, the femur was inserted into a 1 cm plastic straw filled with hot glue (Ad-Tech Hi Temp Glue Gun (299), which applies the hot glue at a temperature of approximately 193 °C) so the condyles were facing up. The femur was held by the shaft using forceps until the glue hardened, and samples were inspected to ensure no glue was on the cartilage surface. The bone in the straw base was then mounted to the testing chamber via an osteochondral core clamp, and secured to the x-y stage of the Biomomentum Mach-1 v500csst with a 3-axis motion controller, with 0.1-µm resolution, in a bath of PBS ( Fig. 1A and B ).

Figure 1.

Schematics of indentation point locations for the femoral condyle and the temporomandibular joint (TMJ) (A), and the experimental set up for the femoral condyle on the Mach-1 machine (B), an example of raw data from a normal indentation test of femoral condyle AC (C) and TMJ mandibular condylar cartilage (MCC) (D).

For the automated indentation of the femoral AC of WT and Fgf2KO mice (N = 5 for WT, N = 6 for Fgf2KO), a 70-N multiaxial load cell with an amplification module was employed and calibrated prior to each use. Isolated femurs in PBS were subjected to automated indentation at 16 predefined positions (8 on each of the lateral [Lat] and medial [Med] condyle, with 4 points each on the anterior [Ant] and posterior [Post] aspects of the condyle) using a 0.3 mm spherical tip indenter over the surface of the femoral condyles. Automated indentation was done to a depth of 0.05 mm, at a speed of 0.05 mm/s with 10 seconds relaxation, as an initial scan and to ensure sufficient signal (load) was obtained with a 100 Hz data acquisition rate. Stress relaxation curves were observed with indentation of the femoral AC ( Fig. 1C ). Structural stiffness was then calculated at an appropriate indentation depth and mapped at the 16 position, then pooled into 4 areas of interest (Med/Ant, Med/Post, Lat/Ant, Lat/Post) with 4 points in each. The size of these regions of interest on the femoral AC was ~0.4 mm².

For automated indentation of TMJ MCC of WT and Fgf2KO mice (N = 6 for each group), a 150-gf uniaxial load cell was employed and calibrated prior to each use. Isolated mandibles were mounted to the testing chamber by horizontally positioning them in an osteochondral core clamp. The testing chamber was then filled with PBS and the MCC of the mandibles were subjected to automated indentation at 16 predefined positions (going from posterior to anterior) using a 0.3mm spherical tip indenter over the surface of the MCC. As with the AC, automated indentation was done to a depth of 0.05 mm, at a speed of 0.05 mm/s with 10 seconds relaxation, as an initial scan with a 100 Hz data acquisition rate. Stress relaxation curves were observed with indentation of the TMJ MCC ( Fig. 1D ). Normal load was calculated based on surface angle, using a 0.02x0.02mm grid, ¹² from which structural stiffness was then calculated at an appropriate indentation depth and mapped and pooled into 2 areas of interest (Post, Ant) with 8 points in each. The size of these regions of interest on the TMJ MCC was ~0.2 mm².

Histology

For mandibles, the portion containing the MCC and associated subchondral bone were fixed for 2 days in 10% formalin, and then placed in 30% sucrose overnight and embedded in cryomedium. For knees the entire joint was harvested and fixed as above and embedded in cryomedium. Serial sagittal sections of MCC and serial frontal sections of knee joints were stained with toluidine blue to qualitatively evaluate overall histology and proteoglycan content. Thickness of the AC and TMJ MCC samples (N = 3 for AC and N = 6 for MCC for each group, WT and Fgf2KO) was then measured at 3 different points from the images and relevant scale bar.

Statistical Analysis

Indentation data were analyzed with non-parametric statistical tests. The effect of mouse genotype (WT, Fgf2KO) on measured normal force was first determined at each position, and at indentation depths of 10, 20, 30, 40, and 50 µm by a 2-tailed Mann-Whitney U test. The effect of mouse genotype, Med/Lat position, and Post/Ant position on the femoral AC structural stiffness was then assessed by Kruskal-Wallis tests, followed by a 2-tailed Mann-Whitney U test at each position between genotypes. Similarly, the effect of mouse genotype and Post/Ant position on the MCC of the TMJ structural stiffness was assessed by Kruskal-Wallis tests, followed by a 2-tailed Mann-Whitney U test at each position between genotypes. Cartilage thickness data were normally distributed and was compared between genotypes by a 2-tailed t test. Statistical analysis was completed using SPSS 25, and P < 0.05 was considered statistically significant. A priori sample size analysis for AC based on previous work ¹¹ and pilot studies indicated N = 5 (with 4 points in each region) was sufficient for 80% power to detect a difference in structural stiffness of 4.3 N/mm with SD of 4.8. A priori sample size analysis for TMJ MCC based on pilot studies indicated N = 6 (with 8 points in each region) was sufficient for 80% power to detect a difference in structural stiffness of 0.02 N/mm with standard deviation of 0.034.

Results

Normal Force During Automated Indentation of Femoral Articular Cartilage and TMJ Mandibular Condylar Cartilage

Normal force increased with increasing indentation depth on both AC ( Fig. 2 ) and MCC ( Fig. 3 ), and at certain locations and depths, differences were observed between WT and Fgf2KO mice. In the femoral AC at the Med/Ant location, significantly more normal force was measured in the WT mice compared with the Fgf2KO mice at 0.01, 0.02, 0.03, 0.04, and 0.05 mm depth indentation (P < 0.05, Fig. 2A ). At the Med/Post location, significant more force was also measured in WT mice at 0.01 and 0.02 mm depth indentation at the Med/Post location (P < 0.05, Fig. 2A ). No significant differences were observed at the Lat/Ant (P > 0.05, though it approached so at 0.04 and 0.05 mm indentation with P = 0.071 and 0.051, respectively, Fig. 2C ) and Lat/Post locations (P > 0.05, Fig. 2D ). Conversely in the TMJ MCC, significantly more normal force was measured in the Fgf2KO mice at certain locations and depths. The normal force was greater in Fgf2KO mice at the Ant location with 0.02, 0.03, 0.04, and 0.05 mm indentation (P < 0.05, Fig. 3A ) and in the Post location at 0.02 and 0.03 mm indentation (P < 0.05, Fig. 3B ). Due to the small size of the femur and TMJ samples, the standard method of measuring cartilage thickness via automated mapping with a needle was not possible. Therefore, thickness was assessed from histological sections such that an appropriate depth (and therefore mechanical strain <20%) could be chosen for calculation of structural stiffnesses. Small (<20%) strain was chosen for structural stiffness calculation because compressive modulus remains nearly constant with increasing compression, while beyond small strain response (>20%) cartilage exhibits nonlinear strain stiffening.^15-17

Figure 2.

Normal force from automated indentation of articular cartilage of wild type (WT) and Fgf2Ko mice femur articular cartilage (AC) at medial/anterior (Med/Ant, A), medial/posterior (Med/Post, B), lateral/anterior (Lat/Ant, C), and lateral/posterior (Lat/Post, D) positions. Data shown as the normal force in newtons over the depth of indentation into the cartilage tissue, with 0.0 mm depth being the tissue surface, for WT and Fgf2KO mice (N = 5 mice for WT, N = 6 for Fgf2KO, individual data points shown, *P < 0.05).

Figure 3.

Normal force from automated indentation of temporomandibular joint (TMJ) mandibular condylar cartilage of the temporomandibular joint at anterior (A) and posterior (B) positions. Data shown as the normal force in newtons over the depth of indentation into the cartilage tissue, with 0.0 mm depth being the tissue surface, for wild type (WT) and Fgf2KO mice (N = 6 mice, individual data points shown, *P < 0.05).

Histology and Structural Stiffness of the Femoral Articular Cartilage

Histological analysis demonstrated Fgf2KO and WT mice had similar AC thickness (Fgf2KO = 106 ± 9 µm vs. WT = 112 ± 6 µm, mean ± SD, P > 0.05) ( Fig. 4A and B ). Based on the measured thicknesses, an indentation depth of 0.02 mm was chosen to ensure small (<20%) compressive strain (~18% and ~19% for Fgf2KO and WT mice, respectively) to calculate and map the structural stiffness of the AC.

Figure 4.

Comparison of femoral articular cartilage of 3-month-old wild type (WT) and Fgf2KO mice. Toluidine blue staining showing cartilage in WT mice (A) and in Fgf2KO mice (B), with similar thickness, scale bar 150 µm. Structural stiffness (C) from 20 µm deep automated indentation of articular cartilage in WT and Fgf2KO mice femurs at anterior (Ant)/posterior (Post) and medial (Med)/lateral (Lat) positions (N = 5 mice for WT, N = 6 for Fgf2KO, individual data points shown, *P < 0.05) and associated heat map for WT (D) and Fgf2KO (E) mice.

Structural stiffness of the femoral AC varied both regional and between WT and Fgf2KO mice ( Fig. 4C ). Structural stiffness was significantly affected by mouse genotype and by Med/Lat position (both P < 0.05), but not Ant/Post position (P > 0.05). Median values for Fgf2KO were significantly smaller than WT at both the Med/Ant (65% smaller, P < 0.05) and Med/Post (53% smaller, P < 0.05) positions. Values on the lateral side were not significantly different from each other, either at Lat/Ant or Lat/Post (both Ps > 0.05) versus WT. These results are qualitatively represented in a heat map ( Fig. 4D and E ).

Histology and Structural Stiffness of the TMJ Mandibular Condylar Cartilage

Counter to that which was observed with the AC, global Fgf2KO lead to a decrease in MCC thickness of the TMJ compared with WT (Fgf2KO = 137 ± 18 µm vs. WT = 157 ± 14 µm, mean ± SD, P < 0.05, Fig. 5A and B ). Given that decreased cartilage thickness is also an established early signs of cartilage degeneration, these results demonstrate OA-related changes can also be observed in MCC as early as 3 months of age. As above with the femoral AC, based on the measured thicknesses assessed histologically an indentation depth of 0.02 mm was chosen to small (<20%) compressive strain (~13% and ~15% for Fgf2KO and WT mice, respectively) to calculate and map the structural stiffness of the MCC.

Figure 5.

Comparison of the mandibular condylar cartilage of the temporomandibular joint in 3-month-old wild type (WT) and Fgf2KO mice. Toluidine blue staining showing cartilage in WT mice (A) and decrease in cartilage thickness in Fgf2KO mice (B), scale bar 200 µm. Structural stiffness (C) from 20 µm deep automated indentation of mandibular condylar cartilage of the temporomandibular joint at posterior (Post) and Anterior (Ant) positions in WT and Fgf2KO articular cartilage (AC) (N = 6 mice, individual data points shown, *P < 0.05) and associated heat map for WT (D) and Fgf2KO (E) mice.

Structural stiffness of the TMJ MC varied both regional and between WT and Fgf2KO mice ( Fig. 5C ). Structural stiffness was significantly affected by mouse genotype and position (both Ps < 0.05). Media values for Fgf2KO were significantly greater than WT MCC at the Post location (78% greater, P < 0.05) and the Ant location (29% greater, P < 0.05). These results are qualitatively represented in a heat map ( Fig. 5D and E ).

Discussion

This study demonstrates the structural stiffness, as assessed by automated indentation, of the femoral AC and TMJ MCC in Fgf2KO mice is altered at 3 months of age compared to WT mice. The automated biomechanical testing employed here, with a commercially available device, was able to resolve spatial differences in the murine femoral AC and TMJ MCC structural stiffness, in addition to the differences between WT and Fgf2KO tissue. Furthermore, the change in structural difference observed in the Fgf2KO mice compared with WT mice was opposite in the AC (decreased) compared with the TMJ MCC (increased), further emphasizing the differences, and motivation, of studying the two different cartilaginous tissues. In this respect, our postulated hypothesis was proven partially correct in that the Fgf2KO mice had diminished structural stiffness, while the TMJ MCC increased. Collectively, these results demonstrate the effect of FGF2 deletion has on murine AC and TMJ MCC mechanical response, and that automated indentation is able to detect changes in mechanical response of cartilage at an early stage of OA development in both murine knee joints and TMJs.

Technical challenges can arise with biomechanical testing of such small joints. In both tissues tested here, harvesting of the joints and mounting of the sample were critical steps. Samples were frozen after harvesting to increase efficiency of mechanical testing, by enabling the testing of several samples in one day after the setup and calibration of the testing device. Each sample experienced a single freeze thaw cycle, which does not affect the mechanical properties of the articular cartilage. ¹⁸ The provided sample holder for the Biomomentum Mach-1 enabled mounting for both the femur and TMJ. For the femur, fixing the bone in a plastic straw with hot glue was a simple and effective way of mounting without compromising the mechanical properties of the cartilage tissue. For the TMJ, the isolated mandible was able to be mounted without any additional fixation. The 70 N multiaxial load cell with an amplification module was employed and calibrated prior to each usage for the femur testing. This load cell had sufficient range and sensitivity for the femoral AC indentation, in addition to being able to indent various locations on the curved condylar surface. Conversely, the TMJ required the more sensitive, yet less expensive, 150-gf uniaxial load cell. This was due to the lower forces being measured on the TMJ, which was also, in general, flatter than the femoral condyles. Last, stress relaxation curves were observed in all testing confirming the indentation of cartilage tissue ( Fig. 1C and D ). Given the small physical size of the samples, and consistent with previous pilot work, ¹¹ a limitation of this study is that cartilage thickness could not be mapped with the standard procedure (i.e., employing automated mapping with a syringe needle and measuring the cartilage thickness by measuring the force and inflection point in the force when the needle passes through the cartilage and the cartilage-bone interface, respectively), thus material properties could not be calculated. This was due to large relative size of the needle and bevel compared to the thickness of the mouse cartilage, and potentially also in part due to the varying cartilage-bone interface in mice (personal communications with Sotcheadt Sim, Research Scientist, Biomomentum Inc.). Future efforts could focus on alternative or specialized needles coupled with custom made holders, thus enabling thickness testing on the same sample (and not contralateral as done here histologically, which we expect to be similar geometrically given the controlled genetics and environment of the mice used) after indentation testing. Nevertheless, structural stiffness values measured here for the femoral AC are in good agreement with those reported previously on WT mice. ¹¹ Overall, the commercially available instrumentation employed here enabled measurements demonstrating Fgf2KO mice have altered structural stiffness in both the femur AC (diminished) and TMJ MCC (enhanced) compared with WT mice.

The results in this study agree with and extend previous studies on murine AC and Fgf2KO mice. The observed diminished cartilage structural stiffness in the Fgf2KO mice are consistent with, and extend on, previous studies demonstrating OA-like changes histologically.^6,7 Interestingly, histological changes in these studies were not observed at 2 months ⁶ or even 3 months ⁷ , consistent with the lack of change of cartilage thickness observed here at 3 months. However, quantitative changes in structural stiffness were measured in the present study at 3 months, demonstrating automated indentation has the ability to detect changes in femoral AC structural stiffness earlier than changes observed histologically. While N = 5 was sufficient to detect statistically significant differences in both Med/Ant and Med/Post positions and demonstrate the study was sufficient powered, consistent with previous pilot work using a similar number of mice, ¹¹ increased variability and lack of a significant difference at other sites (e.g., Lat/Ant, P = 0.21) suggest future studies could employ additional animals. Nevertheless, to our knowledge this is the first study to employ mechanical testing to examine the effect of FGF2 expression on AC and demonstrate an effect on structural stiffness. Future studies could build off these findings to further elucidate the time course of the mechanical changes identified here, potentially measuring the material properties of the AC (if an appropriate needle and holder can be found or made) in Fgf2KO mice and determine how or if these changes correlate to and/or precede histological and molecular changes in the femoral AC over time.

The results in this study also demonstrate, for the first time, the effect of FGF2 expression on structural stiffness of the TMJ MCC. While the effects of FgFr1 and FgFr3 on murine osteoarthritis in the TMJ have been studied,^19,20 to the authors’ best knowledge, this is the first study to employ automated indentation testing on the small murine TMJ and demonstrate a role of FGF2 in MCC homeostasis. FGF2 can therefore be added to the list of other matrix molecules, including proteoglycan-4,^21,22 biglycan and fibromodulin, ²³ collagen types II ²⁴ and XI, ²⁵ previously shown to play a role in TMJ homeostasis. In terms of mechanical indentation of TMJ tissues, one previous study employed nanoindentation with AFM to characterize the micromechanical properties of the TMJ disc and MCC. ¹³ That study was able to assess special variation in the disc, but not the MCC due to its curved surface. In the present study a commercially available biomechanical testing machine was employed, and while material properties were not able to be calculated due to an inability to map cartilage thickness, the method was able to spatially map the structural stiffness of the curved surface of the MCC. Additionally, a simple yet sensitive (150 gf) uniaxial load cell was sufficient for indentation of the MCC, which had a much lower structural stiffness compared with the femoral AC.

The results from indentation of the TMJ MCC indicated Fgf2KO mice have elevated structural stiffness posteriorly and tending toward so on the anterior aspect. This is likely due in part to the decreased thickness of the TMJ MCC in the Fgf2KO mice compared with WT mice, yet based on the thickness assessed histological the 0.02-mm indentation depth equates to ~13% versus 15% compressive strain, respectively, which are similar. An increased instantaneous modulus with degeneration has previously been reported with degeneration of human (fibrocartilaginous) menisci using the same automated indentation method employed here. ²⁶ Therefore, the findings here on the fibrocartilaginous TMJ MCC are consistent in that manner. Overall the findings here provide the framework for future studies to combine automated indentation, potentially with thickness mapping pending further method development as mentioned above, with histology and molecular biology techniques to further elucidate the time course and underlying mechanism(s) of TMJ tissue alterations in Fgf2KO mice, or other genetically modified mice of interest.

One final consideration that makes this study unique is the mechanical testing of both femoral AC and TMJ MCC in the same animal. As described above, these results agree with and extend previous studies examining the effect of Fgf2KO expression in the femoral AC and show for the first time its effect on TMJ MCC. Collectively, they demonstrate the motivation and value, and potential differences, in studying hyaline and fibrocartilage tissues in different joints.

In conclusion, these results demonstrate FGF2 expression affects the structural stiffness of both murine femoral AC and TMJ AC, and that a commercially available mechanical indentation machine can be employed for automated indentation of both these joints. This quantitative method, and future potential work enabling automated thickness mapping of the cartilage thickness, has the ability to detect changes in cartilage prior to those observed histologically. Moreover, when combined with other histological and molecular biology methods, this procedure will provide a valuable multidisciplinary approach for functional characterization of cartilage tissues in murine genetic models of interest relevant to knee joint and TMJ health and disease.