1 Introduction

Articular cartilage is a transparent connective tissue with smooth surface, which can relieve vibration and withstand appropriate mechanical stress. Besides, cartilage can be carved easily to fit the desired profile, and thus, it is widely used as a transplant material in surgery [1, 2]. Although most cartilages are composed of similar chemical components including water, collagen, and proteoglycan (PG), the cartilages at different organs [such as nasal septal cartilage and articular cartilage (AC)] have different properties [3,4,5].

According to prophase research, the internal structure of bovine nasal cartilage (BNC) was isotropic [6] rather than anisotropic of AC. However, the opposite result was obtained by magnetic resonance imaging (MRI) [1]. The authors found that the long axis of collagen fiber is paralleled to the axial direction of the medial plane, and the collagen fiber in vertical and caudocephalic planes is obvious anisotropy, while they are little anisotropy in medial plane.

Currently, the main research methods on cartilage are MRI [1], biomechanical analysis [1], biochemical analysis [7, 8], and other imaging technologies [9]. Biochemistry and biomechanics can investigate large pieces of tissues or thicker parallel sections, but they are limited by resolution. Similarly, MRI and other forms of microscopic technology are hard to synchronously determine the content and structure of PG and collagen at micro-resolution. Compared to the methods above, Fourier transform infrared imaging (FTIRI) can synchronously measure the infrared spectra and images of cartilage samples with fine spatial and spectral resolution [10,11,12,13]. By combining with chemometric methods [10, 14, 15] and polarization analysis [13, 16], FTIRI could be used to measure the spatial distribution and the orientation of main components in the cartilage matrix.

In mid-infrared spectral analysis of cartilage, the useful characteristic bands are concentrated in the spectral region of 800–1800 cm−1 [17]. In detail, amide I and amide II (1590–1500 cm−1) bands are more suitable to characterize collagen, as well as PG to sugar band [18]. Amide III band is neither recommended to characterize collagen nor PG in cartilage [17, 18]. In addition, the axis of orientation of the collagen fibrils is parallel to the transition moments of amide II, which will be helpful to get the orientation information of collagen fiber in cartilage [18, 19]. Simultaneously, there are potential linkages between the mechanical anisotropy with the anisotropies of collagen of BNC [1]. Thus, research on collagen fibers will be more representative in the anisotropy study of BNC.

To obtain and understand the optical properties of cartilage might be conducive to the diagnosis in cartilage reshaping study. Simultaneously, it might be significant for the development of noninvasive optical diagnosis to minimize the nonspecific thermal damage in laser-assisted cartilage reshaping procedures [20,21,22]. In this paper, therefore, the orientation of collagen fiber in BNC would be investigated by analyzing the absorbance of amide II band after FTIRI on BNC section with polarization. This study could be used to further explicate the anisotropic results of BNC and help us to understand the structure and distribution of collagen in BNC at molecule level, which would be beneficial for BNC to be a better transplant material in cartilage reshaping studies.

2 Materials and methods

2.1 BNC samples

The BNC tissue (2-year bull) was obtained from Nanjing Mofeng Slaughterhouse (Jiangsu Province, China), which was approved by the institutional review committees and was performed in compliance with the relevant laws and institutional guidelines in accordance with the EU Directive 2010/63/EU. After skin was removed, the tissue was washed with physiological saline and cut into BNC blocks. A BNC block (5 mm × 5 mm × 7.5 mm) was taken from the top of the whole nasal septum, and marked with a black rectangle in Fig. 1a. It was rapidly cooled by liquid nitrogen and then sliced into 10-μm-thickness sections in plane XY, YZ, and XZ using Leica CM1900 cryotome. Figure 1b–e shows the 3D spatial coordinate diagram of the BNC block. The X-axis direction is consistent with the forward direction of the BNC. Planes XY, YZ, and XZ are paralleled to the bottom, anterior, and lateral sections, respectively. These sections were picked up with Low-E slides (MirrIR slides, Kevley Technologies) and dried in air for 2 h.

Fig. 1
figure 1

The anterior view of bovine nasal septum oriented in anatomical position (a), the schematic diagram of 3D spatial coordinate (b), and three planes of XY (c), YZ (d), and XZ (e) of bovine nasal cartilage (BNC) block. The sampling location of the BNC block is marked with black rectangle in the anterior view (a). The f, g, and h are visible images (upper row) and IR chemimaps of amide II (bottom row) of BNC sections in plane XY, plane YZ, and plane XZ, respectively. The Fourier transform infrared (FTIR) images come from the region marked with red rectangle in visible images, and the red arrows indicate the corresponding locations (A—anterior, M—medial, RE—rear, L—left, R—right, B—bottom, and T—top). The maximum value of color bar is 1.2 in absorbance

Full size image

2.2 FTIRI method

PerkinElmer Spotlight-400 system was used for FTIRI. The system included an FTIR spectrometer, an infrared microscope, a 16-element liquid N2 cooled Mercuric–Cadmium–Telluride (MCT, 400 μm × 25 μm) array detector, and a single-element MCT (100 μm × 100 μm) detector. FTIRI experiments were performed in reflective imaging mode with 8 cm−1 spectral step and 25 μm pixel size when BNC sections were placed on the removable stage. The LED white light and the incident infrared light were focused on the position of sample for locating and scanning visible image (upper row in Fig. 1f–h) and FTIR image (bottom row in Fig. 1f–h), respectively. A polarizer was inserted into the incident light path and rotated at different polarization angles for polarized FTIRI on BNC specimens. Furthermore, the anisotropy of the BNC was investigated by analyzing the absorbance of characterized band, amide II. The schematic representation of the polarized FTIRI setup is shown in Fig. 2a.

Fig. 2
figure 2

Schematic representation of the polarized FTIRI (a) and schematic diagram of spectral processing (b)

Full size image

2.3 Spectral processing

Figure 2b shows the schematic diagram of spectral processing. The target infrared (IR) spectra were extracted from each region of interest (ROI) of the polarization FTIR images using the “Co-add Spectra” function of Spotlight software. The specific extractions were summarized in Table 1. For example, in plane XY, 13 spectra were extracted from anterior to rear (along X direction) with 100 μm step, while 24 spectra were extracted along Y direction (from left to right) with the interval of 200 μm width. The same method of spectral extraction was used in each polarization FTIR image of the corresponding slice. Software of Spotlight (PerkinElmer) and Origin (OriginLab) was used for image processing and spectral analysis, respectively.

Table 1 IR spectral extraction from ROI in each plane
Full size table

The absorbance of amide II band versus the polarization angles was calculated from the infrared absorption images at different polarization and plotted after spectral extraction from the relative ROIs. These plotted data were fitted according to \(A(r,\theta ) = \pm A(r)\cos^{2} (\theta + \theta_{0} (r)) + A_{0} (r)\) [18] to analyze the anisotropy of collagen in BNC. In the equation, r is the tissue depth in BNC slice and θ is the polarization angle. A(r, θ) is a dependent variable on depth and angle. A(r) and A0(r) are two depth-dependent scaling parameters. θ0(r) is a depth-dependent angle parameter, corresponding to the angular offset.

3 Results

The visible images and the chemimaps of amide II (1590–1500 cm−1) of all ROIs of the BNC sections are shown in Fig. 1f–h. The upper row represents the visible images of three sections cut from plane XY, YZ, and XZ, respectively; while the corresponding amide II chemimaps of the ROIs (red rectangles) are arranged in the bottom row. The red in the color bar means the maximum absorbance. The three chemimaps of amide II suggests that collagen was evenly distributed in each BNC section, except of that in plane XY showing more collagen content with red color at both ends where the BNC section overlapped slightly. On the other hand, collagen in the plane YZ shows the most concentrated distribution in three planes. It might indicate that the collagen distribution in plane YZ had a specific structure. The polarization analysis based on FTIR images of three planes would further discover the fine structure or anisotropy of collagen in bovine nasal septum.

Figure 3 shows the polarization dependences of absorbance of amide II band at different locations (corresponding to the red arrow position in Fig. 1) along the different directions (X, Y, and Z) in corresponding section planes. The mean R2 values of all the fitted curves are in the range of 0.93–0.99, which indicates the goodness of fit with excellent reliability.

Fig. 3
figure 3

Polarization dependence of absorbance of amide II band at different locations along different directions (X, Y, and Z) in corresponding section planes (a XY and b XZ, c XZ and d YZ, and e XY and f YZ), respectively. The data marked with (■), (●), and (▲) in (a–f) are corresponding to the anterior (bottom/top, left), and medial and rear (top/bottom, right) locations in respective slice. The mean R2 values of all fitted curves are in the range of 0.93–0.99

Full size image

In Fig. 3a and b, the data marked with (■), (●) and (▲) are corresponding to the anterior, medial, and rear locations in the slice along direction X, respectively; as well as corresponding to the bottom/top, medial, and top/bottom locations in the slice along direction Y in Fig. 3c, d; and the left, medial, and right locations in the slice along direction Y in Fig. 3e and f, respectively. It is clear that the peaks (troughs) of the fitted curves in Fig. 3a–e shift slightly with the location changing from anterior to the rear but no inversion, indicating the orientation of collagen fibers slightly waved without anisotropy. However, the fitted absorbance curves of amide II (Fig. 3f) corresponding to left, medial, and right locations in plane YZ are inversed from the medial region to the left (right) end, suggesting the ca.90° phase difference of these fitted curves or anisotropy between the medial and both ends.

For the further investigation on the anisotropy of amide II (collagen) along direction Y in plane YZ, the IR spectra of residual pixel columns in the regions of 200–800 μm, 1000–2600 μm, and 2800–3600 μm (Y-axis) were extracted from the polarized FTIR images, respectively. Then, the polarization dependence of the absorbance of amide II was drawn by fitting the absorbance obtained at different polarization angles with the previous equation, which are shown in Fig. 4.

Fig. 4
figure 4

The polarization dependences of amide II absorbance along direction Y in plane YZ at the different regions of a 200–800 μm (■, ●, ▲), b 1000–2600 μm (◄, ♦, ►), and c 2800–3600 μm (□, ○, ∆)

Full size image

As shown in Fig. 4, the polarization dependence of amide II absorbance is flat and closed to a line at the 4th and 14th pixel column (800 and 2800 μm Y coordinate), suggesting that the collagen orientation (anisotropy) remarkably changes from direction Z to Y around 800 μm and from direction Y to Z around 2800 μm in the 2D plane YZ.

The amide II absorption profile of the FTIR image of the BNC (YZ) section along direction Y was obtained under 0° and 90° polarizing angles and is shown in Fig. 5. It can be found that the locations in the range of 600–930 μm and 2680–2980 μm should be both transitional regions of anisotropy of collagen, which is consistent with Fig. 4. As is well known, the absorbance of the characteristic band was highest when the direction of the molecular vibration transition moment was consistent with the polarization direction of polarized light [23]. Therefore, anisotropy or orientation changing of collagen could be clearly showed out on both sides of the transitional regions.

Fig. 5
figure 5

The amide II band absorption profiles of the Fourier transform infrared (FTIR) image of BNC section under 0° and 90° polarization, which are along Y coordinate in plane YZ. The black and red lines corresponded to the 0° and 90° polarization. The location of interesting is from 250 μm to 3600 μm

Full size image

4 Discussion

According to Fig. 3a–e, a common phenomenon is the slight angular deviations (phase shift) shown among the fitting curves, which might be due that the direction of collagen fiber in BNC structure is oscillating and irregular linear type. Thus, the slight deviation is included in little anisotropy. Furthermore, it is understandable that the phase difference of approaching 90° for the fitted curve at different locations can be regarded as strong anisotropy. Therefore, amide II along direction X and direction Z similarly shows little anisotropy. Amide II just shows strong anisotropy along direction Y in plane YZ.

Based on the analysis above, collagen fiber shows little anisotropy in planes XY, XZ, and along the direction Z in plane YZ, while collagen fiber shows strong anisotropy along the direction Y in plane YZ. It implies that the long axis of collagen fiber is along the direction Y, the transverse axis direction of BNC. Therefore, it is significant that one needs to pay attention to the variation along the transverse axis (Y) direction of BNC when BNC is used as the transplant material.

Compared to the reported results [1] that the BNC shows strong anisotropy in 2 planes of YZ and XY (vertical and caudocephalic), the FTIRI investigation provides a consistency, that is strong anisotropy along direction Y in plane YZ. It does not show strong anisotropy along direction Y in plane XY, which is the sole difference. The essential reason might be the BNC block coming from the top end of the nasal septum where is in very weak anisotropy rather than the center position. Thus, the anisotropic performance in different parts of BNC might be different. By comparative analysis, the long (Y) axis direction of collagen fibers in the whole BNC should be generally consistent with the transverse axis direction of BNC [1], while the distribution of collagen fiber in other directions of BNC is related to the block’s position in the bovine nasal septum.

Furthermore, Figs. 3, 4, and 5 of Ref. [1] disclosed the direction dependences of optical and physical properties at defined points, called anisotropy in this reference. They also reflect the changes in orientation and optical property of collagen with the location and polarized angle in the whole plane, which are named anisotropy too and in agreement with the results of this FTIRI paper. The FTIRI investigation has further revealed the anisotropy of the spatial arrangement of collagen fibers along two orthogonal axes direction of the same plane.

Based on the results mentioned above, the distributions of collagen fiber in the top of BNC are deduced and summarized in Fig. 6. The distributions of collagen fiber in top of BNC along direction Y in plane YZ are deduced as either arched (curves 1–3 in Fig. 6b) or mildly wavy (curve 4–7 in Fig. 6c. Actually, the anisotropic structure of collagen in top of BNC might be the mixture of both above (not shown here). The distribution of collagen fiber shows little anisotropy in other planes. The orientation of collagen fiber changes along direction Y (transverse axis) in plane YZ at both transitional regions from 600 to 930 μm and from 2680 to 2980 μm, which will be helpful for the biomechanical research on BNC. The possible anisotropy of collagen fiber can be confirmed by combining actual biomechanical characteristics [1] with quantitative images of BNC by polarized light microscopy [1].

Fig. 6
figure 6

Spatial coordinates (a) and schematic diagram of anisotropy of collagen fiber in top of BNC (b, c). The first, second, and third curves represent the possible arched structures, and the fourth, fifth, sixth, and seventh curves represent the possible wavy structures

Full size image

5 Conclusion

FTIRI combined with polarized light technology was used to investigate the collagen anisotropy in the top end of BNC. Collagen fiber is less anisotropic along directions X and Z. However, it is strong anisotropy along direction Y in plane YZ, which is the long axis of the collagen fiber consistent with the transverse axis direction of BNC. Arched and wavy structures of collagen fiber even their mixture in BNC are proposed. The strength of anisotropy is different in different parts of the BNC means that each piece of BNC used for clinical might be distinctive, so that it is necessary to carry out anisotropic analysis in detail. These findings have important clinical significance in cartilage reshaping procedures and other medical/industrial applications. A better understanding of the structure of BNC (the transplant material) will be helpful to improve the success rate of transplant surgery and accelerate the postoperative recovery process of the patient.