The commercial 3.0 T MRI scanner (Philips Healthcare, Eindhoven, Netherlands) was employed to obtain the MRI image preoperatively. Three sagittal MR images were acquired including fat-suppressed (FS) proton density-weighted imaging (PDWI) sequence, T2 mapping sequence, and T1ρ mapping sequence. All sagittal images were obtained without oblique angulation, parallel to the magnetic static field (B0). The acquisition parameters for the sagittal T2 mapping sequence were consistent with the methods previously described20: repetition time = 1000 ms, echo time = 0 ms, optional locking time = 9.5/19.1/28.6/38.1/47.6/57.2/66.7/76.2 ms, slice thickness = 4 mm, slice thickness = 0.5 mm, scanning layers = 20 layers, field of view = 160 mm, matrix size = 320 × 256, excitation times = 1, scanning time = 8 min 46 s.
Regions of interest (ROIs) of different cartilages were outlined. Condyles were divided into ROI parts to match the histology images (HE stain): The c-line of the condyle (Whiteside’s line) was drawn during surgery for surgery consideration. The ROIs of histopathology were selected by the R-line of histopathology was drawn parallel to the c-line. P-distance was applied to determine the interval between the c-line and the R-line. The ROI was obtained and marked by the surgeon during surgery. Then the Whiteside’s line was drawn in MRI image and set as a reference line. Using c-line as a reference line and applying p-distance, histological sections of interest can be marked in MRI images. Through the above method, the histological slices were matched with ROIs of MRI. This method could refer to our previous research21.
Mechanical measurements
To thaw the frozen AC, the samples were kept for 10h at -20 oC, 4 oC, and room temperature sequentially. The AC used for the mechanical measurements was harvested using a round punch with a diameter of 10 mm, and then the prepared sample was fixed on the holder for the mechanical measurement using 3M quick-drying glue (AD119, 3M, USA). Noted: the glue was adsorbed to the interface between the subchondral bone of the sample and the holder. Then, the prepared samples were immediately used in the mechanical measurements.
The micro-mechanical properties of AC were measured by depth-sensing indentation. A Nano Indenter G200 (Keysight Technologies, Inc., Santa Rosa, CA, USA) equipment with a standard Berkovich indenter was adopted, whose elastic modulus is 1140 GPa and Poisson's ratio is 0.07. The holder with the prepared sample was fixed in a homemade sealed cell that was filled with PBS solution. The AC sample and tip are in PBS solution to simulate its mechanical properties like the natural state in the body.
The detailed instrumentation of the indentation measurement can be found elsewhere22, 23. In brief, an indentation test started with the indenter tip approaching the surface with an approach velocity of 20 nm/s. The contact point was determined by setting a setpoint of 2 µN, which means that when the indenter approached, the instrument would judge to reach the sample surface when a value of 2 µN was reached and then start loading at the velocity of 50 nm/s. At the maximum penetration depth of 3000 nm, the load on the indenter was held constant for 60 s. Then, to reduce the viscous effect of AC on its calculation of elastic modulus, the indenter was withdrawn from the surface of the sample at a high velocity of 300 nm/s and then be held at a relatively low load for evaluating the effect of thermal drift. The insert shown in Fig. 1 is the designed loading and unloading scheme. To ensure the reproduction of the results, more than 100 points were selected randomly on each sample to carry out the indentation measurements. The distance between the indentation points on the sample is 50 µm to prevent possible strain field interactions among these indentations.
To obtain a more realistic elastic modulus of the sample, the contact zero point of the load-displacement curve was examined and corrected is it is necessary. 50–95% of the unloading segment of each load-displacement curve was analyzed using Oliver-Pharr model24. The relationship between the reduced modulus for the indenter-specimen system (Er) and slope (S) of the linear stage in the upper part of the unloading curve (Fig. 1) is25:
$${E_r}=\frac{{S\sqrt \pi }}{{2\sqrt {{A_c}} }}$$
1
where AC is the contact area between the tip and the sample. The shape of Berkovich tip, which has a three-sided pyramid with a face angle of 65.3°, used in indentation testing is usually described by area function. The projected contact area Ac as a function of contact depth hc can be described as the following Eq. 26:
$${A_c} \approx {C_0}h_{c}^{2}+{C_1}h_{c}^{{}}+{C_2}h_{c}^{{\frac{1}{2}}}+{C_3}h_{c}^{{\frac{1}{4}}}+....+{C_8}h_{c}^{{\frac{1}{{128}}}}$$
2
here hc is the contact indentation depth and the coefficients C0, C1, …, C8 are determined by fitting experimental data. For an ideal Berkovich tip, C0 is 24.5 and the other coefficients are equal to 0.
The elastic modulus of the sample can be described as the following Eq. 25:
$$\frac{1}{{{E_r}}}=\frac{{1 - \lambda _{{sample}}^{2}}}{{{E_{sample}}}}+\frac{{1 - \lambda _{{indenter}}^{2}}}{{{E_{indenter}}}}$$
3
where, λ is the Poisson's ratio for the tested specimen and the indenter, respectively. Given to the fact that the elastic modulus of the indenter (Eindenter, 1140 GPa) is much larger than that of the sample (Esample). Thus, the above equation can be transformed into:
$${E_{sample}} \approx {E_r}(1 - \lambda _{{sample}}^{2})$$
4
here, the Poisson's ratio of the AC (λsample) was considered to be 0.5, according to the statement of Lewis et. al.27 and Mow et. al28..