Assessment of Native Human Articular Cartilage: A Biomechanical Protocol

Abstract

Objectives

Recapitulating the mechanical properties of articular cartilage (AC) is vital to facilitate the clinical translation of cartilage tissue engineering. Prior to evaluation of tissue-engineered constructs, it is fundamental to investigate the biomechanical properties of native AC under sudden, prolonged, and cyclic loads in a practical manner. However, previous studies have typically reported only the response of native AC to one or other of these loading regimes. We therefore developed a streamlined testing protocol to characterize the elastic and viscoelastic properties of human knee AC, generating values for several important parameters from the same sample.

Design

Human AC was harvested from macroscopically normal regions of distal femoral condyles of patients (n = 3) undergoing total knee arthroplasty. Indentation and unconfined compression tests were conducted under physiological conditions (temperature 37 °C and pH 7.4) and testing parameters (strain rates and loading frequency) to assess elastic and viscoelastic parameters.

Results

The biomechanical properties obtained were as follows: Poisson ratio (0.4 ± 0.1), instantaneous modulus (52.14 ± 9.47 MPa) at a loading rate of 1 mm/s, Young’s modulus (1.03 ± 0.48 MPa), equilibrium modulus (7.48 ± 4.42 MPa), compressive modulus (10.60 ± 3.62 MPa), dynamic modulus (7.71 ± 4.62 MPa) at 1 Hz and loss factor (0.11 ± 0.02).

Conclusions

The measurements fell within the range of reported values for human knee AC biomechanics. To the authors’ knowledge this study is the first to report such a range of biomechanical properties for human distal femoral AC. This protocol may facilitate the assessment of tissue-engineered composites for their functionality and biomechanical similarity to native AC prior to clinical trials.

Introduction

Loss of mechanical integrity is an irreversible consequence of damage to articular cartilage (AC). One of the most prevalent examples of damaged AC is knee osteoarthritis, which is the main reason for knee replacement surgery and a global health burden. ¹ The pathophysiology of knee osteoarthritis is complex. Altered joint mechanics in AC engenders an abnormal pattern of stress distribution which modifies the stress and strain fields experienced by cells embedded in cartilage matrix. These trigger signaling pathways that inhibit chondrogenesis, reduce cell viability and effect biochemical alterations in the matrix. ² The net result is the cartilage softening, fibrillation and degeneration observed in osteoarthritic knees. Current therapeutic strategies to repair AC such as matrix-induced autologous chondrocyte implantation (MACI), osteochondral grafting, and microfracture have generally produced biomechanically poor fibrocartilage. ³ Although promising results have been obtained in tissue engineering approaches for cartilage repair in animal models, most tissue-engineered constructs lack the biomechanical qualities of native human articular cartilage.^4-6 Furthermore, the vast majority of preclinical studies fail to report the biomechanical properties of neocartilage as a key outcome parameter.^7,8 Whether tissue-engineered cartilage is able to withstand mechanical loading in the same way as native human articular cartilage is therefore an important question that must be answered using biomechanical testing methods prior to clinical translation.

Articular cartilage is frequently modelled as a single-phase, incompressible, elastic solid with an instantaneous elastic response under instantaneous loading, while under sustained deformation, AC exhibits viscoelastic qualities with a time-dependent, asymptotic approach to an equilibrium stress response.^9,10 Various biomechanical models have been developed for AC, such as the simple linear monophasic elastic model ¹¹ and the more complex linear biphasic,^12,13 poroviscoelastic, ¹⁴ fibril-reinforced poroelastic, ¹⁵ and transversely isotropic biphasic, ¹⁶ models. Despite the consideration of mechanical nonlinearity in sophisticated models, modeling cartilage as a homogenous, isotropic material is practically convenient in the laboratory and it allows the measurement of equilibrium and instantaneous elastic responses of a cartilage matrix to applied loads in unconfined compression, confined compression, and indentation geometries. ¹ The dynamic viscoelastic behavior of native AC may be assessed under cyclic compression to yield the storage moduli, loss moduli, and thereby the loss factor using sinusoidal wave analysis. ¹⁷ This system of mechanical testing can be used to assess mechanical properties of one native AC sample relative to another or to tissue-engineered cartilage.

The material properties of articular cartilage have been extensively studied using indentation and unconfined compression testing. In such studies, the influence of various factors on the derivation of cartilage mechanical properties has been thoroughly analyzed, including the effect of patient factors like aging,^18,19 tissue factors such as cartilage thickness ²⁰ and mucopolysaccharide or water content of the matrix,^19,21 biomechanical factors such as topographical variation of joint contact stresses,^9,22,23 and instrumental factors such as indenter geometry, ²⁴ indentation depth,^21,25 and the type of elastic or viscoelastic model applied.^10,26,27 However, the majority of biomechanical studies of AC have utilized animal samples, particularly bovine AC.^21,24,27 Those that have used human cartilage have evaluated either the equilibrium, instantaneous or dynamic response to loading rather than all of these on a single sample.^13,28,29

The functional mechanical testing protocol established herein aimed to measure the elastic properties under instantaneous and equilibrium loading conditions, as well as viscoelastic properties under dynamic loading conditions through a sequence of tests in indentation and unconfined compression geometry ( Fig. 1 ). In doing so, this protocol aims to facilitate an efficient systematic approach to mechanical testing of native human knee AC and tissue-engineered cartilage such that the functionality of the latter may be adequately assessed in tissue engineering studies.

Figure 1.

Schematic representation of the biomechanical testing sequence consisting of initial indentation, dynamic compression, cyclic loading, and repeat indentation tests. Four separate sites on the chondral surface (depicted by small arrows diverging from the indenter tip to the cartilage) of each sample were indented for each calculation of the instantaneous or Young’s modulus. The dynamic compression step and cyclic compressive loading steps were repeated sequentially 4 times before moving onto the final indentation step. AC, articular cartilage; SB, subchondral bone.

Methods

For the purposes of protocol validation, a preliminary set of test samples were obtained, following institutional approval (HREC-A 143/16) and informed consent from patients (n = 3; 77-year-old female, 71-year-old male, and 71-year-old male) undergoing total knee arthroplasty at St. Vincent’s Hospital, Victoria, Australia. In particular, the “distal femoral cuts” of femoral condyles were acquired during each operation.

Biomechanical Testing Protocol

Osteochondral Sample Preparation

The distal femoral condyles were carefully inspected for macroscopically normal looking cartilage with no signs of degeneration or fibrillation. Cylindrical osteochondral (OC) plugs were obtained from condyles using a core extruder of 8 mm diameter. Three plugs each were obtained from 2 patients and 2 plugs obtained from one patient, resulting in a total of 8 samples. Following extrusion, the bottom bony surface of each OC plug was immediately shaped via cutting and sanding to be orthogonal to the cartilage surface, using a Dremel 3000-2/30 Rotary Tool at 35,000 rpm. The OC plugs were stored in phosphate-buffered saline (PBS) containing 1% penicillin/streptomycin and protease inhibitor cocktail (Roche Complete Mini EDTA-free) at 4 °C for 24 to 48 hours prior to testing. These steps are outlined in the supporting information (Supplemental Figure S1).

Cartilage Thickness and Area Measurement

Cartilage thickness was measured using stereomicroscopy as per Burgin et al. ³⁰ The stereomicroscope was oriented perpendicular to the articular surface and used to capture images at four locations around the perimeter. The cartilage was identified as the white layer between articulating surface and the tidemark. The thickness of the white cartilage layer was measured using image analysis software (FIJI/Image J).

Experimental Setup

Mechanical testing was conducted using the ElectroForce Biodynamic 5500 instrument (Bose, Eden Prairie, MN) controlled with WinTest 7 software (Bose). The mover arm of the instrument was fitted with an indenter or 25 mm circular metal platen for tests under indentation or unconfined compression geometry, respectively. OC plugs were attached to the surface of a glass petri dish with glue (bonded to residual bone) and immersed in PBS. Temperature inside the dish was maintained at 37 ± 1 °C using a heated circulation system (Lauda-Brinkmann, Deltran, NJ). A 25-mm metal platen was attached to either a 250 g load cell or a 50 lb (200 N) load cell for indentation and unconfined compression respectively. This platen served as the platform for the petri dish. A photograph of the test setup is provided in the supporting information (Supplemental Figure S2).

Poisson Ratio Calculation

In unconfined compression geometry, the mover arm fitted with a 25 mm metal platen was lowered until it came into contact with the cartilage surface and the load measured in real time using a 50 lb load-cell. The minimum load required to achieve even contact with the chondral surface was set as the zero or baseline value. Then 20% axial compression was applied at 0.01 mm/s and the compression was observed in real time using video microscopy. The sample was allowed to stress-relax until equilibrium, at which point photos of the samples were taken horizontally. The axial strain was measured between the top platen and bottom of the OC plug and the lateral strain was measured from the horizontal expansion of the cartilage layer at equilibrium. Using image processing (ImageJ software), the ratio of the average lateral strain to average axial strain was calculated to optically determine the Poisson ratio, similar to previous studies.^22,31

Mechanical Testing Sequence

Indentation

First, stepwise indentation tests were performed at 1 mm/s for strains of 0% to 5%, 5% to 8%, and 8% to 10%, with 300 seconds relaxation to equilibrium between each step. Indentations were performed either with a diameter flat-ended cylindrical indenter of diameter 0.36 mm or 0.5 mm. The indenters had polished edges and were cleaned with paraffin oil before and after each experiment. After the indenter (see Supplemental Figure S2) was lifted off the cartilage at the end of each indentation step, the portable stereomicroscope (affixed to a clamp stand) was used to zoom in to the area of indentation to ensure that the indenter had not penetrated or caused fissures in the cartilage surface. Examination of the osteochondral unit for potential damage is an important part of the protocol as a damaged sample cannot be used to accurately conduct subsequent biomechanical tests. The gradient of the instantaneous stress-strain curve at 5% strain was calculated to derive the instantaneous modulus E_i. The gradient of the equilibrium stress-strain curve from all 3 indentation steps was calculated to derive the equilibrium modulus E_eq.

Under the assumption of a monophasic, isotropic elastic material indented using a flat-ended cylindrical indenter, we used the mathematical solution used by Korhonen and colleagues ²⁷ to determine the Young’s modulus E_s at equilibrium, from the original solution of Hayes et al. ³² :

$$E_s=\frac{\left(1-v^2\right)\pi a}{2kh}{E}_{eq}$$

where v is the Poisson ratio measured optically, a is indenter radius, h is cartilage thickness, and k is a theoretical scaling factor that depends on the aspect ratio (a/h). Values for k were obtained for finite indentation according to calculations by Zhang et al. ²⁵

Viscoelasticity

Following indentation, the sample was set up in an unconfined compression arrangement, and put through a sequence involving dynamic mechanical testing, compressive modulus measurement, and cyclic compressive loading. Viscoelastic measurements were performed using sinusoidal compression with a strain amplitude of 1.25% for 50 cycles at frequencies of 0.1 Hz, 1 Hz, and 10 Hz. The storage modulus (E′) and loss modulus (E″) were calculated using viscoelastic theory. The loss factor was obtained from the ratio E″/E′.

Compressive Modulus

Compressive modulus was measured in unconfined compression where the cartilage was axially compressed between 0% and 20% strain at 0.01 mm/s. The compressive modulus was calculated as the average stress-strain gradient between 10% and 15% strain, in a similar fashion to previous studies. ¹⁷

Cyclic Loading

Cyclic compressive loading constituted axial compression between 0% and 20% strain with a frequency of 4 Hz for 2000 cycles. The three stages of dynamic testing, compressive modulus measurement and cyclic loading were conducted sequentially and repeated 4 times per sample. At the end of 4 blocks of testing, the AC had undergone 8,000 cycles of high-frequency compressive loading. Indentation measurements were then repeated at 4 sites for each plug using the initial parameters.

Rationale for the Testing Parameters and Conditions

Strain

According to a study of tibiofemoral cartilage strains measured by magnetic resonance imaging (MRI), immediately following a dynamic hopping activity the maximum compressive strain was 6% in human medial femoral cartilage and 3% in the lateral femoral cartilage. ³³ Similarly, the average medial femoral cartilage deformation after walking was 6.7% in another study. ³⁴ Some in vivo MRI studies have shown that 30 minutes of running in healthy individuals produces significant acute medial femoral cartilage deformation up to 5.3% and lateral femoral cartilage deformation up to 4.0%, when measured within 2 minutes of ending the activity. ³⁵

While it is not possible to measure the in vivo strains in real time as a human joint is loaded, the immediate postactivity and nonequilibrium cartilage deformation values captured in the above literature suggest that physiological strains to estimate cartilage stiffness ex vivo are likely to be close to or slightly greater than reported values. Hence, in this protocol, the range of strains 5% to 7.5% in the dynamic mechanical testing phase, instantaneous strain of 5% and stepwise indentation axial strains of 5%, 8%, and 10% at equilibrium are deemed appropriate.

Strain Rate or Loading Rate

For an average cartilage thickness of 2.09 ± 0.23 mm in our study, when axial strain is applied at 1 mm/s, a 5% deformation is achieved on average within 105 ms. These conditions are designed to mimic physiological loading conditions during running or drop-landing, where tibiofemoral cartilage experiences large force and frequent instances of rapid loading.^33-35

In a study of noninvasive measurements of in vivo human tibiofemoral cartilage deformation via MRI, when a cyclic, half-body-weight load was applied axially to a subject’s knee, it took less than 200 ms to apply the required force during every cycle before the cartilage tissue experienced creep. ³⁶ Higher loads (equivalent to full body weight or higher, e.g., when landing from a height) would further reduce the time to deformation and hence increase the strain rate. The chosen strain of 5% following rapid loading is well supported by in vivo human knee MRI studies, which have shown average deformations of 6% ± 2% in the distal femur and 5% ± 1% in the tibial plateau immediately following single legged hopping. ³³ Thus, our chosen rate and axial strain for instantaneous deformation falls within the range of deformations observed in vivo using radiographic techniques. It is however important to note that the reported literature values of maximal axial strain due to rapid loading lie somewhere between the true peak strain and the equilibrium strain depending on the timing of measurement following the joint-loading activity.

This protocol follows similar methods to several published studies that have utilized the Hayes’ elastic model to obtain the Young’s modulus for cartilage under indentation.^23,24,27 The rate of impact is less influential when using the equilibrium stress-strain values to determine the equilibrium modulus and subsequently the Young’s modulus of the cartilage.

While we chose to probe a regime of high force and rapid loading, researchers more interested in the response of the cartilage to more gentle activities, such as walking, may opt to choose a slower strain rate of 0.1 s⁻¹.

Viscoelastic Frequency

The frequency chosen (1 Hz) to obtain the dynamic modulus via sinusoidal compression matches that used previously ²² and approximates the average physiological loading frequency of 1.2 Hz which has been reported for the human tibiofemoral (knee) joint during running. ³⁷

Statistical Analysis

Due to a small number of samples, the nonparametric Wilcoxon matched-pairs signed rank test (statistical significance at α = 0.05) was used to analyze the difference in the modulus values before and after cyclic compression for each sample separately, as done in previous studies. ²⁷ For similar reasons, the nonparametric Spearman rank correlation test (statistical significance at α =0.05) was conducted to investigate correlations between the different elastic and viscoelastic parameters both before and after cyclic compression.

Results

The functional mechanical testing protocol was validated through preliminary indentation studies conducted on 8 osteochondral samples from macroscopically normal regions of distal femoral condyles from 3 patients (patient 1, patient 2, and patient 3). Calculations of dynamic and compressive elastic properties were performed using unconfined compression data at baseline and after 8,000 compressive loading cycles to simulate joint loading (4 consecutive steps of 20% axial compression for 2,000 cycles at 4 Hz). Final indentation studies on each sample were conducted after a total of 8,000 compressive loading cycles and compared to baseline.

The mean cartilage thickness obtained via stereomicroscopy was 2.09 ± 0.23 mm, which is similar to that reported elsewhere (2.0-3.9 mm, ²³ 1.2- 2.0 mm, ¹⁷ 1.95 mm ⁴⁴ ). The mean Poisson ratio (v) of 0.4 ± 0.1 demonstrated some overlap with literature values in the range of 0.12 to 0. 38 for human articular cartilage.^17,28

Table 1 compares the mechanical properties obtained before cyclic compression with literature values of the same properties in various animal and human joints. The values obtained using the current protocol were more similar to certain studies than others due to methodological similarities and differences as well as intrinsic variations of “normal” cartilage within the same population.

Table 1.

Comparison of Various Mechanical Properties of Human and Large Animal Articular Cartilage from Different Sites, as Measured in Previous Studies and in This Study (Expressed as Mean, Mean ± SD, or Range).

Cartilage Source	Instantaneous Modulus (MPa)	Young’s Modulus (MPa)	Equilibrium Modulus (MPa)	Compressive Modulus (MPa)	Dynamic Modulus (MPa)	Loss Factor
Animal
Bovine humeral head	—	0.677 ± 0.223 (Jurvelin et al. 1997) (unconfined compression) 0.80 ± 0.33 (Korhonen et al. 2002) (unconfined compression) 1.15 ± 0.44 (Korhonen et al. 2002) (indentation) 0.10-1.60 (Lyyra-Laitinen et al. 1999) (unconfined compression)	—	—	8.1 ± 6.4 a (Nieminen et al. 2004) 22	~0.25 (Lawless et al. 2017) 38
Bovine patella	20-65 b (Brown et al. 2007) (indentation) 4.0-16 (Brown et al. 2007) 21 (indentation)	0.57 ± 0.17 (Korhonen et al. 2002) (unconfined compression) 27 0.56 ± 0.19 (Korhonen et al. 2002) (unconfined compression) 27 0.72 ± 0.19 (Korhonen et al. 2002) (indentation) 0.83 ± 0.21 (Korhonen et al. 2002) (indentation)	—	—	8.1 ± 6.4 (Nieminen et al. 2004 22 ) a	—
Bovine femoral condyles	—	0.31 ± 0.18 (unconfined compression) 27 0.55 ± 0.19 (indentation) 27	0.2-0.4 (Laasanen et al 2003) 16 (unconfined compression)	~1.0 (Martin et al. 2000) 39 (confined compression)	~15-17 (Martin et al. 2000) 39 ~4-7 (Laasanen et al. 2003) 16	—
Bovine femoral head	—	—	—	—	—	~0.20 (Lawless et al. 2017) 38
Sheep femoral condyles	0.6 ± 0.4 (Di Bella et al. 2018) (indentation)	—	4.5 ± 4 (Di Bella et al. 2018) 40 (indentation)	—	—	—
Human
Tibial plateau	11.1 ± 2.61 d (Stok et al. 2010) 41 (indentation)	3.49 ± 1.06 (Stok et al. 2010) 41 (indentation)	—	—	—	—
Patella	—	0.64 ± 0.30 (Kiviranta et al. 2008) 29 (indentation)	—	3.1-10 (Shepherd et al. 1999) 9	4.67 ± 2.24 (Kiviranta et al. 2008) 29	—
Ankle (talar and tibial surfaces)	—	—	—	10.0-18.6 (Shepherd et al. 1999) 9	—	—
Femoral head	51.9-89.8 e (Burgin et al. 2014) 30 (unconfined compression)	—	0.250.7 (Boschetti et al. 2004) 28 (unconfined compression)	~4.89 (Taylor et al. 2012) 42	—	0.14-0.17 (Temple et al. 2016) 43
Femoral condyles	—	~0.70 (Lyyra et al. 1999) 23 (indentation) 0.85 ± 0.35 (Jurvelin et al. 2003) 44 (unconfined compression)	0.45 ± 0.1 (Bas et al. 2017) 17 (unconfined compression) 0.2-0.4 (Bartnikowski et al. 2015) 45 (unconfined compression) 0.2-0.58 (Jurvelin et al. 2003) 44 (unconfined compression)	1.63 ± 0.26 (Bas et al. 2017) 17 (unconfined compression) 1.70-2.75 (Bartnikowski et al. 2015) 45 (unconfined compression) 45 4.3-13.0 (Shepherd et al. 1999) 9	—	0.12 ± 0.02 (Bas et al. 2017) 17 0.16-0.17 (Bartnikowski et al.) 45
Femoral condyles (present study)	52.14 ± 9.47	1.03 ± 0.48	7.48 ± 4.42	10.60 ± 3.62	7.71 ± 4.62	0.11 ± 0.02

^aCombined population of bovine humeral head and patella samples.

^bLoaded up to 30% strain at a strain rate of 0.1 s⁻¹.

^cLoaded up to 10% strain at a strain rate of 0.1 s⁻¹.

^dLoaded at a strain rate of 0.08 s⁻¹.

^eMeasured instantaneously as the “peak dynamic modulus” on rapid loading (100 g load dropped from 8 cm above cartilage surface with impact velocity of 1.25 m/s).

Indentation

The baseline instantaneous strain 5%, compared with 58.26 ± 8.12 MPa following 8,000 high-frequency compression cycles. Values for the fluid-independent elastic properties equilibrium modulus and Young’s modulus were 7.48 ± 4.42 MPa and 1.03 ± 0.48 MPa before cyclic compression, respectively. Following compressive loading, the corresponding values for equilibrium and Young’s moduli were 7.43 ± 4.78 MPa and 1.01 ± 0.49 MPa, respectively. Wilcoxon matched-pairs signed rank test^27,44 for each osteochondral sample showed no statistically significant difference between the “before cyclic compression” and “after cyclic compression” values in terms of the instantaneous (P = 0.3), equilibrium (P > 0.9), or Young’s moduli (P = 0.8) ( Fig. 2 ).

Figure 2.

Mechanical properties of articular cartilage tested in and equilibrium loading conditions using indentation geometry before (blue) and after (red) 8000 cycles of cyclic compression at a frequency of 4 Hz. Pairwise “before” and “after” comparisons for each sample are shown for (A) instantaneous modulus, (B) Young’s modulus, and (C) equilibrium modulus. The pooled mean values for equilibrium modulus (D), instantaneous modulus (E), and Young’s modulus (F) are illustrated before and after cyclic compression.

Unconfined Compression

In unconfined compression, the mean compressive modulus was 10.60 ± 3.62 MPa at baseline and 9.47 ± 3.54 MPa after cyclic compression. The dynamic, storage and loss moduli at baseline were 7.71 ± 4.62 MPa, 7.66 ± 4.60 MPa, and 0.84 ± 0.47 MPa, respectively. Following 8,000 cycles of compressive loading, the corresponding values were 6.57 ± 3.88 MPa, 6.53 ± 3.86 MPa, and 0.76 ± 0.41 MPa, respectively. From the storage and loss moduli, the “before” and “after” values of the loss factor were calculated as 0.11 ± 0.02 and 0.12 ± 0.02, respectively ( Fig. 3 ). Similar to indentation results, there was no statistically significant difference in the values for each modulus before and after cyclic compression.

Figure 3.

Comparison of dynamic mechanical properties of articular cartilage before (blue) and after (red) cyclic compressive loading for (A) dynamic modulus, (B) storage modulus, (C) loss modulus, (D) compressive modulus, and (E) loss tangent when tested at 1-Hz sinusoidal compression (closest loading frequency of knee joint during walking and running). The “after cyclic compression” data were measured after 4 separate steps of 2,000 cycles of cyclic compression conducted at a frequency of 4 Hz (total of 8,000 cycles).

Correlations between Biomechanical Properties

According to Spearman rank correlation ( Fig. 4 ), the Young’s modulus at equilibrium correlated poorly with the instantaneous stiffness measured on rapid loading. While there was no significant correlation between the baseline moduli measured using indentation and unconfined compression, there was a significant positive correlation (P < 0.05) between the instantaneous modulus and dynamic modulus (ρ = 0.79) as well as between the instantaneous modulus and storage modulus (ρ = 0.79) after cyclic compressive loading of the cartilage.

Figure 4.

Spearman rank correlation matrix to assess the relationship between several mechanical properties of native hyaline articular cartilage before (left) and after (right) cyclic compressive loading. Values of the correlation coefficient rho (ρ) greater than 0.7 were statistically significant positive correlations (P < 0.05).

Discussion

In the present study, we developed a novel mechanical testing protocol to enable the biomechanical characterization of human articular cartilage and conduct a functional test of cartilage durability. Using a preliminary set of human distal femoral articular cartilage samples, indentation and unconfined compression tests were conducted in physiological conditions (37 ± 1 °C) to quantify elastic and viscoelastic material properties of native human articular cartilage before and after thousands of compressive loading cycles. In general, the material properties obtained from this protocol, especially the Young’s modulus, compressive modulus, dynamic modulus, and loss factor, demonstrated good agreement with values reported using similar methodologies in the cartilage biomechanics literature ( Table 1 ). The cartilage samples showed no significant mechanical deterioration in any of the 8 tested parameters following compressive loading at 4 Hz for 8000 cycles, which revealed a high degree of durability of native human femoral condylar cartilage and highlighted that a tissue-engineered construct would need to be highly durable to be clinically suitable for chondral repair in human knees. While engineered constructs may be expected to mature biomechanically over time, it is essential for the immature tissue to have a minimum level of stiffness to maintain its shape, prevent disintegration, and support chondrogenesis.

Comparisons of absolute values of cartilage mechanical properties are difficult due to the natural variation of cartilage stiffness across a joint surface, ³⁰ between different joints within the same individual ²⁰ and between joints of different individuals with similar ages. ²³ Cartilage stiffness also varies with thickness,^24,31 which differs enormously between various human joints and between human and animal joints, which are generally thinner (e.g., 1.4 ± 0.3 mm for bovine cartilage ²⁷ ). Furthermore, added variability in experimental measurements arises due to differences in testing conditions.

The rate of cartilage loading influences the instantaneous and dynamic moduli, ²¹ which reflect the load support provided by the interstitial fluid pressurization within the viscoelastic collagen matrix when cartilage is loaded dynamically. ⁹ The mean value of instantaneous modulus obtained for this study was 52.14 ± 9.47 MPa under a loading rate of 1 mm/s, which translates to a strain rate of 0.5 s⁻¹ for the average cartilage thickness of ~2 mm. This rate simulates the application of a sudden weight-bearing force on cartilage, for example, when landing from a height. Brown et al. ²¹ conducted similar indentation studies on bovine articular cartilage with a strain rate of 0.1 s⁻¹ (20% of the rate used in the current protocol), yielding values of instantaneous stiffness around ~10 ± 5 MPa, which is 20% of and in proportion to the value obtained in this study. Similarly, another study employed a slower strain rate of 0.08 s⁻¹ for instantaneous loading of human tibial plateau cartilage, resulting in an instantaneous modulus of 11.1 ± 2.6 MPa, which is another example of the proportionality observed between loading rate and instantaneous modulus of the measured cartilage sample. Moreover, Burgin and colleagues ³⁰ measured maximum stress-strain gradients in the range of ~51.9 to 89.8 MPa following the instantaneous application of a load onto cartilage samples obtained from elderly human femoral heads. The impact testing in this study better simulates high impact events such as running or jumping from a height, similar to the instantaneous testing in the present study, rather than low-impact activities like walking. Nevertheless, the instantaneous values also closely approximate those in the present study.

Evaluation of cartilage viscoelasticity involved cyclic displacement between a strain amplitude of 1.25% at 1 Hz, which is close to loading frequencies for walking and running.^37,46 The same conditions were set by Nieminen et al. ²² to test bovine humeral and patellar cartilage, resulting in a dynamic modulus of 8.1 ± 6.4 MPa, which overlaps with the values from the present protocol ( Table 1 ). Similarly, another study utilizing the same frequency obtained values of 4.67 ± 2.24 MPa for human patellar cartilage, which further validates the results of the present protocol. The loss factor was also comparable to literature values^17,43 ( Table 1 ), suggesting that this protocol can provide an effective assessment of viscoelastic damping due to sinusoidal compression.

As opposed to dynamic loading, when cartilage is subject to a constant deformation, the water in the collagen matrix exudes out until an equilibrium stress is reached. At this point, the resistance to the applied load reflects only the intrinsic compressive stiffness of the solid phase of articular cartilage, which is measured as the Young’s modulus. ⁴⁷ The present study reported a Young’s modulus of 1.03 ± 0.48 MPa, which overlaps with the range of Young’s moduli measured in equilibrium conditions (similar to the present study) for human patella and femoral condyles^29,44 ( Table 1 ). The strong correlation between the Young’s modulus and equilibrium modulus reflects the inherent dependence of the Young’s modulus on the equilibrium modulus, which is the slope of the equilibrium stress-strain curve. ²⁷ Since the assumption of a monophasic, isotropic and homogenous material is required to apply the indentation data to the mathematical solution for the Young’s modulus, variations in the values of this elastic property across studies may be further explained by different Poisson ratio measurements or incorporation of different scaling factors (k value). While a value of the Poisson ratio was determined optically by video microscopy ³¹ in the current protocol, other studies have calculated this intrinsic elastic parameter indirectly or assumed a value of 0.5, which applies to incompressible solids.^24,27 Since the formula to obtain the Young’s modulus is related to the “square” of the Poisson ratio (see Equation 1), variation in this value may explain differences in calculated Young’s moduli for articular cartilage across different studies. Similarly, adoption of the theoretical scaling factor k from Hayes et al. ³² (infinitesimal deformation) rather than Zhang et al. ²⁵ (finite indentation depth) in the formula may alter such measurements.

In unconfined compression, the compressive modulus was measured as 10.60 ± 3.62 MPa, under slow loading, that is, a nonequilibrium state. This is very similar to the 4.3 – 13 MPa range obtained by Shepherd et al. ⁹ for the human femoral condyle ( Table 1 ). Another study ¹⁷ that used similar methods to the present protocol has reported lower values in the range of 1.6 to 2.75 MPa. However, the sample size of this study was a limitation as the compressive modulus measurements were conducted on a sample from a single donor rather than multiple donors. ¹⁷ This could explain the difference in measurements, which may be affected by the large topological variation in site-specific compressive stiffness across the distal femoral AC of different individuals.

While numerous indentation and unconfined compression methods have been established for the mechanical testing of articular cartilage, the major novelties of the current protocol are that (1) it includes a comprehensive regime of mechanical tests in physiological conditions using relatively common instruments to characterize the instantaneous, equilibrium elastic, and dynamic viscoelastic properties of articular cartilage and (2) it is the first protocol to incorporate cyclic compression to assess compressive durability. We did not find any significant difference in any of the mechanical properties measured using indentation ( Fig. 2 ) or unconfined compression ( Fig. 3 ) following compressive loading, which emphasizes the durable nature of native human articular cartilage. This negative finding is particularly significant in the development of minimum biomechanical benchmarks for cartilage tissue engineering. However, a limitation was the small patient sample size (n = 3), which may have impeded the determination of statistical significance in this study. Furthermore, since the patients were in the age range of 78 to 82 years, these data may not be truly representative of healthy, young hyaline cartilage; this limitation is commonplace in the field of cartilage biomechanics due to ethical barriers to the acquisition of articular cartilage from young patients. Furthermore, since the distribution of degeneration on the femoral condyles were slightly different for each patient, the mechanical properties of AC may differ across the different samples due to the topographical variation of cartilage stiffness in the joint. ²⁶ Although the osteochondral samples were meticulously shaped to have a flat surface, the natural curvatures of the native cartilage surface could have led to subtle inaccuracies in determination of the contact point (0% strain) during compressive testing and Poisson ratio measurement, leading to minor errors in the modulus calculations. Since 8,000 compressive cycles did not effect changes in cartilage mechanics, the protocol may be modified to incorporate a greater number of compressive loading cycles to conduct mechanical failure testing of articular cartilage.

Much of the current research in cartilage mechanics includes testing of 1 to 3 material parameters of cartilage.^24,27,31 However, in order to assess the biomechanical integrity and clinical translatability of an artificially engineered construct, a more comprehensive suite of tests performed under physiological temperature and loading conditions is essential. The need for simple laboratory tests as a biomechanical evaluation tool for cartilage tissue engineering formed the motivation for this project. The protocol developed herein seeks to guide cartilage tissue engineering colleagues to conduct in-depth mechanical characterizations of both native and tissue-engineered cartilage.

Conclusion

In the present study, we developed a functional biomechanical testing protocol for native human articular cartilage under physiologically relevant conditions. Primarily, this protocol aimed to quantify both the elastic properties under instantaneous and equilibrium loading conditions, and viscoelastic properties under dynamic loading conditions through a sequence of tests in indentation and unconfined compression geometry. A secondary aim was to test the impact of cyclic compressive loading on the mechanical properties of the cartilage. Using a preliminary set of articular cartilage samples, we successfully validated the protocol through comparisons between the material properties calculated herein and those reported in the literature, taking into account the heterogeneity in experimental parameters and tissue sources. There was no significant difference in cartilage mechanics before and after 8,000 compression cycles, highlighting the short-term mechanical durability of native cartilage. Future experiments with a higher number of loading cycles are required to elucidate the mechanical durability of native cartilage. It is hoped that this protocol will aid tissue engineers in the biomechanical comparisons between native and tissue-engineered cartilage, to evaluate the clinical suitability of the latter.