Biomechanical, biochemical, and near infrared spectral data of bovine knee ligaments and patellar tendon

Knee joint ligaments and patellar tendon are rope-like tissues that enable the proper function of the knee by connecting the bones that form the joint. A better understanding of ligament structure-function relationships is needed to develop objective and reliable diagnostic methods for ligaments. Recently, arthroscopic near infrared spectroscopy (NIR) has shown the potential to quantitatively evaluate the health of the cartilages and menisci of the knee. In this dataset, we present a unique combination of NIR spectral data, biomechanical properties, and biochemical composition of bovine primary knee ligaments and patellar tendon (10 knees, 50 tissue samples). NIR spectral data were measured at 5 locations in each sample, biomechanical properties were obtained with tensile testing, and biochemical composition was quantified using colorimetric biochemical methods. The data can be reused for investigations of structure-function relationships of knee ligaments and patellar tendon, for the development of NIR spectroscopic methods to quantify the health of these tissues, and to develop new computational models to describe ligament and tendon biomechanics.

for the development of NIR spectroscopic methods to quantify the health of these tissues, and to develop new computational models to describe ligament and tendon biomechanics. ©

Value of the Data
• This dataset is useful in ligament and tendon basic research, as their structure-function relationships are not yet fully understood. • The dataset is intended for researchers investigating the biomechanical behavior and biochemistry of ligaments and tendons. The included NIRS measurements are useful for biophotonics researchers interested in quantitative evaluation of connective tissues. • Biomechanical data can be used in the development of new computational models describing their mechanics. NIRS data can be utilized to develop multivariate models for quantitative evaluation of various ligament properties.

Data Description
The data are stored in several Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) files, stored in Fairdata Research Data Storage Service [1] . Alternatively, the same data are given in comma separated value (.csv) files. Table 1 presents an overview of the methods, number of measurements and generated data.

Near infrared spectral data
Near infrared spectral data are stored in an Excel file Data_NIRS.xlsx , which contains the measured spectrum of each sample. Example codes nirs_plsr_analysis.m and nirs_visualize_spectra.m showing good practices on preprocessing, analysis and visualization of NIR spectral data were created with MATLAB R2016b (Mathworks Inc., Natick, MA, USA) and it is available publicly ( https://github.com/UEF-BBC/ristaniemi-data-in-brief-2021 ).   Additionally, the raw data of the tensile tests are provided (Files 3-52) to enable the reuse of the experimental biomechanical measurements. For each sample, there is an Excel file containing the time, upper clamp displacement, and load signals of the relaxation, sinusoidal and ultimate tests. Moreover, one sheet in that excel file indicates the zero-load length, thickness, and width in millimeters. An example code Biomechanical_analysis_example.m to calculate the mechanical properties from the raw data was created with MATLAB R2018B, and it is available publicly ( https://github.com/UEF-BBC/ristaniemi-data-in-brief-2021 ).

Sample preparation
Ten bovine stifle joints (anatomically corresponding to the human knee joint), aged 14-22 months, were collected at an abattoir (Atria Oyj, Seinäjoki, Finland) ( Fig. 1 a). The ACL, PCL, MCL, LCL, and PT were carefully extracted, placed in plastic containers filled with phosphate-buffered saline solution (PBS), and put in a freezer ( −20 °C) ( Fig. 1 b). The PBS contained two enzymeinhibitors to enhance sample preservation, the ethylenediaminetetraacetic acid disodium salt (1.86 g/L, EDTA VWR International, Radnor, PA, USA) and benzamidine hydrochloride hydrate (0.78 g/L, Sigma-Aldrich Co., St. Louis, MO, USA). Prior to the measurements, the samples were thawed at room temperature (approximately 21 °C), and a dogbone-shaped test piece was cut from the ligament mid-substance, with collagen fibers directed to the longitudinal axis ( Fig. 1 c). First, two parallel razor blades, fixed to a separation of 1.81 mm, were used to cut a slice from the tissue. Then, the desired dogbone-shape was cut from the slice using a custom punch tool. The central part of the dogbone shape had a width of approximately 2 mm, and a length of 10 mm. This type of shape and size has been used earlier when studying mechanical properties of ligaments and tendons [2][3][4][5][6][7] .

Near infrared spectroscopy
NIRS measurements were conducted using a dual-detector NIRS system (AvaSpec Multichannel Spectrometer, Avantes BV, Apeldoorn, Netherlands) with a 10 W Tungsten Halogen light source. The two detectors of the system covered wavelength regions of 320-1100 nm (AvaSpec-ULS2048L-USB2, resolution 0.6 nm) and 950-2500 nm (AvaSpec-NIR256-2.5-HSC, resolution 3.2 nm), respectively ( Fig. 2 ). The system was connected to a custom stainless-steel probe shaped like an arthroscopic hook (outer diameter = 3.25 mm, inner diameter = 1.90 mm). The probe contained a total of 114 optical fibers (fiber diameter = 100 μm); 100 fibers for emitting and 14 fibers for collecting light to the spectrometers (i.e., seven fibers per detector). NIR spectra were acquired with 100 coadded scans using an integration time of 1.5 milliseconds for AvaSpec-ULS2048L-USB2 detector and 20 milliseconds for AvaSpec-NIR256-2.5-HSC detector. Measurements were taken from five equispaced and non-overlapping sites (4.0 mm spacing between measurement center points) along the longitudinal axis of the sample (250 NIR spectra in total, Table 1 ). Prior to each measurement, the system was calibrated with a diffuse reflectance standard (99% ± 4% reflectance factor in the 20 0-250 0 nm range, Spectralon SRS-99, Labsphere Inc., North Sutton, USA) and a non-reflective piece of black rubber. The non-reflective rubber corresponded to the background noise in the absence of any incoming light, while the reflectance standard represents the maximum signal-to-noise ratio the detector can reach without saturating the signal. Absorption values ( A ) for each individual wavelength ( λ) were computed with the following equation, where S(λ) is the measured spectrum, D (λ) the spectrum of the non-reflective standard, and R (λ) the spectrum of the reflectance standard.

Biomechanical testing
The width and thickness at the sample center cross-section were measured with an optical microscope using 4.6x magnification, and elliptical shape was assumed to determine the crosssectional area [8][9][10] . The elliptical assumption was verified to be accurate, with a 0.15% error in the resulting cross-sectional area compared with a direct measurement of a cut cross-section, using the same microscope and magnification [11] . Double-sided sandpapers (Mirox P80, Mirka Oy, Uusikaarlepyy, Finland) were glued (Loctite Precision, Henkel AG, Düsseldorf, Germany) to both sides at both ends of the sample to ensure a good grip in tensile testing. The sample was carefully placed between jaw-type clamps and aligned with the machine axis, and the clamp screws were adjusted to a 4 Nm moment with a torque wrench. This procedure ensured constant and proper clamping for all the samples, and no slipping was observed during the tests. The samples were kept at room temperature (approximately 21 °C) in PBS during the tensile testing. At body temperature (37 °C) the samples may exhibit unaltered quasi-static behavior [12] , but altered history-dependent viscoelastic properties [13] .
A uniaxial material testing system was used to conduct the tensile testing. It consisted of a linear actuator (Newport, Irvine, CA, USA; resolution 0.1 μm) and a controller (Newport, Irvine, CA, USA), operated via a custom LabView software (version 10.0, National Instruments Corporation, Austin, TX, USA), and a 25 lb load cell (Model 31/AL311BL, Honeywell, Columbus, OH, USA). Tensile stress of 0.05 MPa was applied to determine the zero-load length [3] . The preconditioning protocol consisted of 10 loading-unloading cycles with 0.05 mm/s velocity to 2% strain. A 2-min recovery period was allowed, and the zero-load length was re-established. This 10-cycle protocol was repeated five times to obtain a stable mechanical behavior [11] . Then, a 10-min recovery was allowed, the zero-load length was verified and the measurement was started. An incremental stress-relaxation test was conducted at 2, 4, 6 and 8% strain, with ramp speed of 0.1 mm/s (i.e. ~1%/s), and 30 min of relaxation at each strain level, exhibiting relaxation behavior useful for viscoelastic material model development ( Fig. 3 a). After the relaxation, with the sample at 8% strain, a sinusoidal test was imposed with 0.5% strain amplitude [14] with 0.1, 0.5 and 1 Hz frequencies for 20 cycles at each frequency ( Fig. 3 b). The sample was then allowed to recover at 0% strain (zero-load length) for 1 h. Finally, a tensile test until failure was performed with 0.005 mm/s (0.05%/s) velocity ( Fig. 3 c).
To calculate the material parameters, the sinusoidal and ultimate test force and displacement data were used to calculate the true stress , where F is the measured force, A 0 the initial cross-sectional area, L the current length and L 0 the initial length) and logarithmic strain ). The true stress assumes a constant volume of the material. For the sinusoidal test, a sinusoidal function was fit to stress-time and strain-time data: where z is the stress or strain, A is the amplitude, f is the frequency, t is the time, ϕ is the phase angle and z 0 is the constant term. The dynamic modulus E dyn and phase difference γ were then calculated as where the subscripts σ and ε denote the parameters related to stress-time and strain-time data ( Fig. 3 b).
For the ultimate test, the data from 0.2 MPa onwards were taken for analysis, to exclude the slack portion from the analysis. To determine the Young's modulus, linear fits were made to the stress-strain curve and the fit with the maximum tangent modulus [15] was used to define the Young's modulus. More precisely, at each data point of the curve, the successive points covering an 8% strain interval were taken and a linear fit was made to determine the modulus. The maximum of these moduli, which practically occurs in the linear region, was then defined as the Young's modulus ( Fig. 3 c). Then, the toe region end ( ε toe , σ toe ) and yield points ( ε yield , σ yield ) were defined as points where the experimental stress-strain curve deviated by 0.6% strain [16] from the linear fit with the maximum modulus. The maximum stress reached during the experiment was defined as the ultimate strength ( ε ult , σ ult ). The linear region length was then calculated as ε linear = ε yield − ε toe . Toughness K, or energy density, was calculated as the integral over the stress-strain curve to the ultimate strain. In addition, we calculated this parameter also at yield point: The toe region, 0 < ε < ε toe , was characterized by a second order formula [16] σ = A ε 2 + Bε + C, (6) where A , B and C are constants, and with an exponential formula [17] where D and F are constants.

Biochemical analyses
The sample surface was blot with a paper towel to remove surplus PBS, and the wet weight was taken as the average of three measurements. The samples were lyophilized (freeze-dried) for 24 h, and the dry weight was taken as the average of three measurements. Water content was subsequently determined by the difference between wet and dry weights. One part of the samples was used to determine the hydroxyproline and uronic acid contents. The samples were incubated in 150 mM sodium acetate including 5 mM l -cysteine, with 15 mM EDTA and a 1 mg/ml concentration of papain at pH of 5.8 and 60 °C for 16 h to digest the proteoglycans [18] . The samples were then boiled for 10 min for the deactivation of the enzyme. The hydroxyproline content was quantified following the procedure outlined in Brown et al. [19] . After papain digestion, the samples were hydrolyzed in 10 M HCl at 108 °C for 16 h. Following cooling, 1 M sodium hydroxide was used to neutralize the samples. An oxidizing solution with chloramine-T was added and mixed well, and allowed to react for 5 min. Ehrlich's reagent was then added to each sample and they were mixed using a vortex mixer. The samples were then incubated in a water bath at 60 °C for 45 min. A microplate reader was used to read the absorbance at 540 nm. The hydroxyproline content was determined three times and averaged, and the content was normalized by wet and dry weights.
The uronic acid content was quantified as outlined in Blumenkrantz and Asboe-Hansen [20] . First, ethanol precipitation was used to remove possible salts from the papain digested samples. Then, samples were cooled on crushed ice and sulfuric acid/sodium tetraborate solution was added. The samples were shaken and warmed in a water bath at 100 °C for 5 min. The samples were cooled, m-hydroxyphenyl reagent was added and after 5 min the absorbance was read at 540 nm. The uronic acid content was determined three times and averaged, and the content was normalized by wet and dry weights.
The elastin content was determined with Fastin Elastin Assay (Biocolor Ltd., Carrickfergus, County Antrim, United Kingdom) according to the manufacturer's instructions [21] , using the other part of the samples. First, the samples were boiled in 0.25 M oxalic acid for 1 h to extract the elastin. To ensure that all elastin was recovered, boiling was repeated three times, which was verified to be sufficient in our pilot measurements. Elastin was precipitated with a precipitation reagent from two aliquots of pooled extracts, and reacted with dye for 90 min. The dyed elastin, red-brown in color, was dissociated and the solution was placed in a microwell plate. The absorbance was read at 513 nm and the elastin content was determined by comparison to a standard curve. The content was normalized by wet and dry weights.

Ethics Statement
The tissues used in this study were obtained from an abattoir, as leftovers from meat production, therefore no further ethical approvals were required for this study.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships which have or could be perceived to have influenced the work reported in this article.