Skip to main content
NCBI home page
Journal of Anatomy logoLink to Journal of Anatomy
. 2021 Jan 31;239(1):70–80. doi: 10.1111/joa.13402

Multiscale structural characterization of the vertebral endplate in animal models

Magdalena Wojtków 1,✉, Maciej Głowacki 2, Celina Pezowicz 1
PMCID: PMC8197957  PMID: 33521970

Abstract

Research in the field of spinal biomechanics, including analyses of the impact of implants on the stability of the spine, is conducted extensively in animal models. One of the basic problems in spinal implantation is the transfer and distribution of loads carried by the spine on the surfaces of the vertebral bodies. An important factor in proper cooperation of spinal implants with the vertebrae is the endplate (EP), which is why the EP in the animal model used for testing should be as similar as possible to the human EP. Therefore, this study involved multiscale structural and morphometric analyses of the animal models most commonly used in spinal biomechanics research, i.e. pig, ovine, and bovine tail. The tests were performed on 28 lumbar porcine, ovine, and bovine vertebrae. Both cranial and caudal EPs were analysed in three selected areas: anterior, middle, and posterior EPs. The conducted tests included a morphometric analysis of the trabecular bone (TB) layer of the EP as well as microscopic analysis at the mesoscale (total thickness) and microscale (thickness of the individual EP layers). The porcine EP had a characteristic increased circumferential thickness (~3 mm) with a significant narrowing in the central region (50%–60%). The convex cranial ovine EP had a constant thickness throughout the cross‐section and the concave caudal EP showed ~35% narrowing in the central region. The thickest EPs were observed in the bovine tail model with negligibly small narrowing in the central region (~5%). The thickness of the cartilaginous layer in the porcine and bovine models reached up to 1 mm in the peripheral regions and decreased in the central part. The growth plate layer had a similar thickness in all the models. On the other hand, the narrowing of the total thickness of the EPs in the central region was mainly due to a decrease in the VEP thickness. In the ovine and bovine models, the central region of the EP was characterized by large isotropy and trabeculae of mixed or rod‐like shape. By contrast, in the pig, this region had plate‐like trabeculae of anisotropic nature. The porcine model was identified as best reflecting the shape and structure of the human EP and as the best surrogate model for the human EP model. This choice is particularly important in the context of biomechanical research.

Keywords: animal model, anisotropy, endplate, geometrical parameters, microscopic analysis

Based on the conducted research, the authors of this study obtained data on the shape of EPs of the lumbar spines of the porcine, ovine, and bovine (tail) models at different scale levels. The central region of the endplate was characterised by large isotropy and trabeculae of mixed or rod‐like shape in the ovine and bovine models. By contrast, in the pig, this region had plate‐like trabeculae of anisotropic nature. The porcine model was identified as best reflecting the shape and structure of the human EP and as the best surrogate model for the human EP model.

graphic file with name JOA-239-70-g007.jpg

1. INTRODUCTION

One of the elements determining the proper cooperation of the intervertebral disc (IVD)–vertebra complex is the endplate (EP) the separates elastic tissues of the IVD from hard, rigid tissues of the vertebral bodies. Despite the mismatch of the mechanical properties between these structures, they successfully transfer complex loads acting on the spine.

Furthermore, what is important, the EP seems to play a key role in spine reconstruction procedures using interbody implants. One of the basic problems in spinal implantation is the transfer and distribution of loads carried by the spine on the surfaces of the vertebral bodies. Disorders or problems with obtaining the correct distribution of transferred loads most often result from the selection of a geometrically inappropriate implant. The use of too small implants makes it impossible to support them on the epiphyseal rim of the EP, which is characterized by the greatest thickness and high strength parameters (Grant et al., 2001). The most common consequence of such unsuccessful implantation is subsidence of the implant into one or both vertebral bodies and destruction of the osseous structure of the vertebral bodies, resulting in secondary instability of the implanted area, nerve damage and pain (Grant et al., 2002). On the other hand, implantation of an oversized implant may lead to damage or irritation of adjacent structures, especially the spinal cord (Zhou et al., 2000).

Research in the field of spinal biomechanics, including analyses of the impact of implants on the stability of the spine, is carried out extensively in animal spine models such as the pig, sheep, goat, and bovine tail. In the literature, numerous works are analysing various geometrical structural, mechanical, and biochemical parameters of both the vertebra and the IVD, aimed at, among others, determining the optimal model of the spine that best reflects the human spine (Alini et al., 2008; Busscher et al., 2010; Casaroli et al., 2018; Smit, 2002; Szotek et al., 2004; Wilke et al., 1997). As already mentioned, the EP is an important factor in the proper cooperation of the IVD‐vertebra complex and spinal implants with the vertebrae. Hence, an EP in the animal model should be as similar as possible to the human EP.

It must be pointed out that animal and human EPs have different structures. In general, in quadrupeds, the EP consists of a cartilaginous endplate (CEP), bony vertebral endplate (VEP), and the growth plate (GP). The CEP layer becomes more calcified closer to the VEP, and its progression of calcification increases with age. The VEP is composed of subchondral bone (SB), which is a thin layer of dense cortical bone, and trabecular bone (TB). Below the VEP, one can find the growth plate, which is a tiny cartilage membrane between the vertebra and EP structure (Broom & Thambyah, 2018; Moore, 2000; Wojtków et al., 2019). In the case of humans, the EP takes the form of the epiphyseal ring, covering the peripheral regions of the vertebral body in the transverse plane. It contains both the VEP and CEP, although in the human spine the CEP extends only to the epiphyseal ring, whereas in quadruped spines it extends right to the perimeter of the disc. Moreover, animal EPs are separated from the vertebra by the GP, which does not exist in the human spine in this form (Broom & Thambyah, 2018; Wade, 2018). Human spine growth is a very complex process in which vertebral growth occurs by endochondral ossification. The human EP also varies spatially in both its morphology and composition.

The macroscopic structure of the EP, i.e. its width, thickness, and topography, have been described by only a few authors (van der Houwen et al., 2010; Panjabi et al., 2001; Pitzen et al., 2004; Wang et al., 2011; Zhao et al., 2009). The conducted research has shown a characteristic narrowing of the thickness of the human EP in its central area, occurring regardless of the analysed segment of the spine or location of the EP relative to the IVD. Based on an analysis of the geometrical parameters of the EPs, the researchers hypothesized that cranial EPs (located on the superior surface of the vertebral body) are more susceptible to damage because they are thinner and are supported by lower density cancellous tissue (Roberts et al., 1997; Zhao et al., 2009). However, the issue of structural analysis of the EP on a microscopic scale, taking into account its layers, has not yet been addressed by researchers.

Therefore, this study aims to provide a detailed description of the structure and morphometrical parameters of the animal EPs, most commonly used in studies of spinal biomechanics, i.e. the pig, ovine, and bovine tail.

2. MATERIAL AND METHODS

2.1. Research material

The tests were carried out on 28 vertebrae of the following animal models: porcine (lumbar spine‐ L1, L3, L5; n = 9), ovine (lumbar spine‐ L1, L3, L5; n = 9), and bovine (tail segment‐ Co1÷Co4; n = 10). Three porcine and three ovine spines were collected from immature animals aged between 6 to 8 months, and the immature/ young adults’ bovine tails (extracted from three spines) were from animals aged approximately 18 months. All spines were obtained from a local meat‐processing slaughterhouse. Immediately after collection, the research material was placed in double plastic packages and stored at –20°C until the test date. Before starting the tests, the spines were thawed and individual vertebrae were selected for testing, then they were cleaned of adhering soft tissues and the IVDs were removed. The procedure was carried out with the utmost caution so as not to damage the CEP layer. All the tests were carried out on both the cranial and caudal EPs of each of the analysed vertebrae. A total of 56 EPs were tested, within which three areas were analysed: anterior, middle, and posterior (in the sagittal plane). The middle region was defined as the thinner location or concavity of the EP if present in a given animal model. Otherwise, the geometrical centre in the analysed cross‐section was selected.

The conducted tests describing the EP structure included morphometric analysis of the TB layer of the EP as well as microscopic analysis of the EP at the mesoscale (total thickness) and microscale (thickness of the individual EP layers).

2.2. Microscopic analysis

A 3 mm slice was excised from the middle part of each of the vertebrae along the sagittal plane. For this purpose, a metallographic cutter was used with a diamond grinding wheel, working under constant water flow (Accutom‐5, Struers, Denmark). Microscopic analyses included such parameters as the total EP thickness, the thickness of the CEP layer, the SB layer as well as the TB layer forming the bony endplate VEP, and the GP layer (Figure 1). Using the term endplate (EP), we refer to this macro‐scale structure in animal models, as one whole tissue, formed by the CEP, VEP, and GP layers.

FIGURE 1.

FIGURE 1

Open in a new tab

Microscopic image showing the layered structure of the endplate (EP) of the porcine model with indication of the analysed areas (CEP, cartilage endplate layer; SB, subchondral bone layer; TB, trabecular bone structure; GP, growth plate). CEP layer is located between the tidemark line (i.e. the line separating the disc tissue and CEP layer, marked as first dotted white line from the top of image) and cement line (tissue mineralization line, second dotted white line from the top). SB layer is between cement line and line within VEP, where trabecular bone structure (and porosity) begins to appear. This area is also characterized by the darker colour of the tissue in the photos. Below TB layer can be indicated which lower boundary was identified as the boundary between VEP and GP, i.e. bone and cartilage tissue of the endplate

An analysis of the selected structures was carried out in the anterior, middle, and posterior areas of the EP using a stereomicroscope (Zeiss Stereo Discovery V20, Zeiss). All the measurements were carried out 10 times using specialized software (AxioVision, Zeiss). Due to the endplate structure irregularity, the microscopic analyses were carried out manually by the examiner each time. To ensure the high repeatability of the layers’ height measurements, reached by similar identification of boundaries between layers, all of the measurements were carried out by one examiner.

2.3. Microtomography examination

The microstructural properties of the TB layer of the bony EP were determined with the use of a microtomography (SkyScan 1172, Bruker). The EPs were scanned using an Al‐Cu filter with a resolution of 9 µm, 180‐degree scanning range, lamp intensity of 153 µA, voltage of 65 kV, and exposure time of 1,130 ms. The entire vertebrae were scanned; therefore, the analysis was performed before microscopic examinations. After the image reconstruction (NRecon, Bruker), three cylindrical volumes of interest (VOIs) were created within each EP, which were located in the analysed regions (anterior, middle, posterior). The created VOIs had a diameter of 3 mm, while their height was adjusted to the thickness of the TB layer in the examined region (Figure 2). To avoid artifacts and misanalysis, the selected VOIs contained only the TB structure (located between the SB layer and the GP).

FIGURE 2.

FIGURE 2

Open in a new tab

Visualization of the areas analysed during the study: (a) view of the reconstructed vertebral body, (b) selected cranial EP, and (c) an EP slice together with the analysed cylindrical volumes of interest presenting differentiation in the thickness of the TB layer of the EP (marked red) depending on the area of the analysis. White dotted lines indicate the cutting lines of the model. Presented based on the porcine model

Analysis of the microstructural parameters (CtAn, Bruker, Belgium) was carried using three‐dimensional parameters describing the TB microarchitecture, such as bone volume to total volume ratio in the VOI (BV/TV [%]), total porosity (Po(tot) [%]), trabecular number (Tb.N [mm−1]), trabecular thickness (Tb.Th [mm]), trabecular spacing (Tb.Sp [mm]), degree of anisotropy (DA [‐]), and structural model index specifying the type of bone tissue (SMI [‐]).

2.4. Statistical analysis

Statistical analysis was performed using the Prism 7 software (GraphPad). The normality of the analysed parameter distribution was verified using the Shapiro–Wilk normality test. The Brown–Forsythe test was carried out to compare the variances of the groups. Statistical differences between the analysed parameters for the cranial and caudal EPs were examined using the Mann–Whitney U test. Analysis of the statistical significance of the studied parameter concerning various animal models was carried out using the Kruskal–Wallis test together with the post hoc Dunn test. All the analyses were performed at a significance level of p < 0.05. The results were presented in the form of means with standard deviations.

3. RESULTS

3.1. Total thickness of the EP at the mesoscale

Statistical analysis showed significant statistical differences in the total thickness of the caudal and cranial EPs in all the animal models. Therefore, these two structures were subjected to separate analyses (Figure 3).

FIGURE 3.

FIGURE 3

Open in a new tab

Graphs of the total thickness of cranial (a, b) and caudal (c, d) EPs. The boxes on the graphs A and C show 25th and 75th percentiles (distribution of the data); the line inside each box represent the median, while the whiskers indicate the minimum and maximum values. The graphs B and D show profiles of both cranial and caudal EPs

The total thickness of the cranial EP in the anterior region was highest in the bovine tail model and averaged 4.20 ± 0.61 mm compared to 3.54 ± 0.62 mm in the pig (p < 0.0001) and 2.58 ± 0.34 mm in the sheep (p < 0.0001). A similar relationship was noticed in the posterior EP region, where the bovine tail EP was also the thickest.

However, the largest differences in the shape and thickness of the cranial EP were observed in the middle region. In the porcine model, there was a significant decrease in thickness of about 60% compared to the anterior region (p < 0.0001) and 55% compared to the posterior region (p < 0.0001). In the ovine model, the narrowing did not exceed 10% compared to the peripheral regions, whereas in the case of the bovine model it decreased by 15% compared to the anterior region (p = 0.0036) and was 10% thicker compared to the posterior region (p = 0.0036).

A similar relationship in the total EP thickness between the animal models was observed in the case of caudal EPs (Figure 3c,d). The thickest of the analysed models was the bovine EP with an average thickness of about 3.69 ± 0.64 mm in the peripheral regions (p < 0.0001) and 3.60 ± 0.28 mm in the middle region (p < 0.0001). By contrast, the pig had the thinnest EP with the thickness of 3.29 ± 0.32 mm in the anterior region, 1.65 ± 0.21 mm in the central region, and 3.03 ± 0.26 mm in the posterior region. The EP narrowing in this animal in the central region was about 50% (p < 0.0001 to the peripheral regions) compared to the 5% narrowing in the bovine and 35% in the sheep (p < 0.0001 and p = 0.0015 to, respectively, the anterior and posterior regions).

3.2. Thickness of the EP layers at the microscale

The thickness of the CEP layer (Table 1) was most diverged in terms of the analysed regions in the porcine model. The thickest CEP layer in this animal model was in the posterior EP region and amounted to, respectively, 0.81 ± 0.33 mm for the cranial EP and 0.75 ± 0.23 mm for the caudal EP. The CEP in the middle region was thinner (relative to peripheral regions) by 83% in the cranial EP (p < 0.0001) and by 76% in the caudal EP (p < 0.0001). Both in the ovine and bovine tail models, the CEP was thickest in the caudal EP and its thickness was slightly lower in the middle region. In all the analysed animal models, the CEP thickness was similar in the peripheral regions.

TABLE 1.

Thickness values of individual EP layers of the analysed animal models in the anterior, middle, and posterior regions

Endplate layers thickness (mm)
Cranial Caudal
CEP SB TB GP CEP SB TB GP
Pig
Anterior 0.77 ± 0.17 a 0.36 ± 0.11 a 2.66 ± 0.57 a 0.14 ± 0.03 a , b 0.73 ± 0.23 a 0.40 ± 0.12 a 2.34 ± 0.32 a 0.14 ± 0.03 b
Middle 0.14 ± 0.04 c 0.22 ± 0.11 c 1.15 ± 0.13 c 0.10 ± 0.01 0.18 ± 0.04 c 0.21 ± 0.11 1.37 ± 0.28 c 0.12 ± 0.02
Posterior 0.81 ± 0.33 0.36 ± 0.12 2.26 ± 0.39 0.10 ± 0.02 0.75 ± 0.23 0.28 ± 0.11 2.16 ± 0.22 0.12 ± 0.04
Bovine
Anterior 1.05 ± 0.13 a 0.86 ± 0.31 a , b 2.36 ± 0.53 a 0.26 ± 0.11 1.06 ± 0.16 a 0.58 ± 0.14 a , b 2.41 ± 0.43 0.25 ± 0.05 a
Middle 0.76 ± 0.17 c 0.28 ± 0.11 2.77 ± 0.48 c 0.19 ± 0.05 0.95 ± 0.20 c 0.28 ± 0.12 2.61 ± 0.30 0.19 ± 0.04 c
Posterior 1.02 ± 0.16 0.34 ± 0.14 2.36 ± 0.52 0.21 ± 0.07 1.09 ± 0.18 0.34 ± 0.12 2.49 ± 0.44 0.32 ± 0.10
Ovine
Anterior 0.51 ± 0.01 a 0.40 ± 0.12 a 2.06 ± 0.34 0.13 ± 0.02 0.57 ± 0.09 0.69 ± 0.14 a 2.14 ± 0.79 a 0.12 ± 0.04
Middle 0.39 ± 0.10 0.21 ± 0.11 c 1.95 ± 0.21 c 0.14 ± 0.02 0.52 ± 0.08 0.16 ± 0.09 c 1.48 ± 0.22 c 0.11 ± 0.03
Posterior 0.42 ± 0.15 0.42 ± 0.12 2.34 ± 0.44 0.13 ± 0.02 0.56 ± 0.07 0.40 ± 0.12 1.52 ± 0.26 0.12 ± 0.02
Human
Anterior — — — x — — — x
Middle 0.25 ± 0.03 (Palepu et al., 2019) — — 0.30 ± 0.09 (Palepu et al., 2019) — —
Posterior — — — — — —
Open in a new tab

The values are presented as means ± SD. The differences between the analysed EP regions were examined using the Friedman test (Dunn post hoc).

a

Statistical significance between the anterior and middle regions.

b

statistical significance between the anterior and posterior regions.

c

Statistical significance between the middle and posterior regions, x‐ do not exist in the human spine.

The thickness of the SB layer (Table 1) is similar between analysed models; however, in the porcine model was the lowest. The thickest SB layer in all analysed models was anterior EP region. Again, the thinner and practically constant in all analysed animal models was the middle region. The cranial SB layer height decrease in this area (in relation to the peripheral regions) was about, respectively, 39% (p = 0.0058), 54% (average, p < 0.0001), and 49% (p = 0.0003 for anterior vs. middle, and p < 0.0001 for posterior vs. middle region) in porcine, bovine, and ovine model. Presented dependence was analogical for caudal SB layers of the endplate.

The thickest EP layer was the TB layer (Table 1), i.e. the TB structure of the bony EP. Again, the largest decrease in the thickness of this layer in the central region relative to the peripheral regions was observed in the porcine model and amounted to no more than 57% (p < 0.0001). An interesting correlation was observed in the ovine model, where the TB structure was thickest in the central region and amounted to, respectively, 2.77 ± 0.48 mm and 2.61 ± 0.30 mm in the cranial and caudal EP (p < 0.0001).

The thickness of the GP layer in the ovine model was practically constant in all the analysed regions and averaged 0.13 ± 0.02 mm for the cranial EP and 0.12 ± 0.03 mm for the caudal EP. The thickest height of this layer, reaching up to about 0.30 mm, was observed in the bovine model, where there was a significant narrowing of this layer in the middle region, in relation to peripheral regions (by less than 30% in the cranial EP and 40% in the caudal EP).

3.3. TB morphometric analysis of the VEPs

Statistical analysis found no statistically significant differences between the TB morphometric parameters in the caudal and cranial EPs, which is why the data were analysed together. The means of the analysed morphometric parameters are presented in Table 2.

TABLE 2.

The analysed microstructural parameters of the TB layer of the EP of the porcine, bovine, and ovine models with differentiation of the analysed EP areas: anterior, middle, and posterior

BV/TV (%) Tb.N (1/mm) Tb.Th (mm) Tb.Sp (mm) Po(tot) (%)
Pig
Anterior 61.67 ± 14.97 a 1.918 ± 0.410 a 0.3383 ± 0.1169 0.3006 ± 0.0690 a 38.33 ± 14.70 a
Middle 75.73 ± 13.84 2.346 ± 0.487 c 0.3358 ± 0.0934c 0.2001 ± 0.0758 c 24.27 ± 13.84
Posterior 70.98 ± 16.25 1.792 ± 0.352 0.4064 ± 0.1170 0.2707 ± 0.0635 29.02 ± 16.25
Bovine
Anterior 30.92 ± 10.78 a 2.082 ± 0.224 a 0.1474 ± 0.0409 0.2511 ± 0.0362 69.08 ± 10.78 a
Middle 20.01 ± 4.88 1.502 ± 0.153 c 0.1578 ± 0.07448 0.2936 ± 0.07696 79.99 ± 4.88
Posterior 28.22 ± 9.24 1.836 ± 0.399 0.1616 ± 0.0553 0.2930 ± 0.0783 71.78 ± 9.24
Ovine
Anterior 30.07 ± 2.77 b 2.124 ± 0.157 0.1433 ± 0.0217 b 0.2875 ± 0.0243 69.93 ± 2.77 b
Middle 29.83 ± 1.94 c 2.101 ± 0.302 0.1509 ± 0.0084 0.2819 ± 0.0441 70.17 ± 1.94 c
Posterior 38.06 ± 2.72 2.169 ± 0.127 0.1664 ± 0.0131 0.2862 ± 0.0223 63.10 ± 2.46
Human
Anterior 21.2 (Rutges et al., 2011) —

0.1507 (Rutges et al., 2011)

0.18 ± 0.05 (Rodriguez et al., 2012)

 — 44 ± 14 (Rodriguez et al., 2012)
Middle 20.7 (Rutges et al., 2011) — —
Posterior 18.9 (Rutges et al., 2011) — —
Open in a new tab

The values are presented as means ± SD. The differences between the analysed EP regions were examined using the Friedman test (Dunn post hoc).

a

Statistical significance between anterior and middle regions.

b

Statistical significance between anterior and posterior regions.

c

statistical significance between middle and posterior regions.

Analysis of the microarchitecture of the trabeculae of the TB layer showed that in the porcine model it was much denser in the entire EP volume compared to the other tested animal models (Figure 4). A significantly higher bone volume was observed in the analysed regions (BV/TV, p < 0.0001 for all regions between the porcine and bovine models) and a decrease in the porosity of the structure (Po(tot)) by about half (p < 0.0001 for all regions between the ovine and bovine models). Moreover, trabeculae in the porcine model were significantly, at least twice, thicker (Tb.Th) in all the analysed regions (p < 0.0001) than in other animal models. There was a variation in the number of trabeculae (Tb.N) depending on the analysed region. In the anterior region, it was similar in all the models; in the middle region, it was lowest in the case of the bovine (p < 0.0001 compared to the pig and p = 0.0019 compared to the sheep); while in the posterior region, the greatest number of trabeculae was observed in the sheep (p = 0.0171 compared to the pig and p = 0.0433 compared to the bovine). The distances between individual trabeculae (Tb.Sp) were similar between animal models in the anterior and posterior regions of the EP; however, in the central region in the pig, they were significantly smaller (p = 0.0018 compared to the bovine tail and p = 0.0394 compared to the sheep).

FIGURE 4.

FIGURE 4

Open in a new tab

Reconstruction of the cranial and caudal EPs for the analysed animal models of the pig, sheep, and bovine (tail), indicating differences in the shape of the EP

The TB layer in the porcine model was denser in the middle region of the EP, resulting in an increase in Tb.N and a decrease in Tb.Sp relative to the peripheral regions. Due to the greater number of trabeculae and smaller distances between them in the central region, the porosity in this region was reduced by 37% relative to the anterior EP (p = 0.0016) and by 16% relative to the posterior EP. The least dense TB structure occurred in the anterior region, where trabeculae with an average thickness (Tb.Th) of 0.3383 ± 0.1169 mm were the thinnest and the distances between them were the largest (Tb.Sp 0.3006 ± 0.0690 mm on average).

An inverse relationship was observed in the bovine model, where the TB structure of the EP in the central region was least dense and its porosity was about 80%. The Tb.N parameter with an average value of 1.502 ± 0.153 [‐] in this region was lower by 27% and 18%, respectively, relative to the anterior (p = 0.0002) and posterior (p = 0.0320) regions. The values of Tb.Th and Tb.Sp were similar in all the analysed regions.

The ovine model showed the densification of TB structure in the posterior region. Po(tot) decreased by 10% relative to the other regions (p = 0.0035 and p = 0.0179 relative to, respectively, the anterior and middle regions). Microstructural examinations also showed that the trabeculae in the posterior region were thicker (Tb.Th increased by 14%, p = 0.0179) relative to the anterior EP.

3.4. Anisotropy of the TB structure of the EP in the central region

Microscopic analysis of the EP structure showed varying orientation of trabeculae (Figure 4) in the TB layer between the analysed animal models in the middle EP region. In the porcine model, the trabeculae were arranged horizontally, parallel to the surfaces of the vertebral bodies, while their arrangement in the bovine and sheep was vertical, in accordance with the arrangement of the main trabeculae in the vertebral body structure (Figure 5). To identify the observed changes, an in‐depth analysis was carried out using micro‐computed tomography (micro‐CT).

FIGURE 5.

FIGURE 5

Open in a new tab

Varying orientation of trabeculae in the TB layer in the middle EP region shown on the example of the caudal EP of the pig and sheep obtained by (a) micro‐computed tomography (micro‐CT), (b) microscopy, and (c) presented schematically

Analysis of EP anisotropy (Figure 6) demonstrated that both the bovine tail and ovine animal models showed isotropic tendencies in all the analysed regions. The average values of the DA parameter were 0.3548 ± 0.1040, 0.3869 ± 0.0452, and 0.3151 ± 0.0851 in, respectively, the anterior, middle, and posterior regions of the bovine EP. In the case of the ovine model, the average values of the DA were, respectively, 0.4689 ± 0.1141, 0.4105 ± 0.1036, and 0.3613 ± 0.0675, with this model showing slightly lower isotropy than the bovine tail model.

FIGURE 6.

FIGURE 6

Open in a new tab

The values of the DA parameter for different animal models and analysed areas. The boxes show 25th and 75th percentiles (distribution of the data); the lines inside the boxes represent the median, while the whiskers indicate the minimum and maximum values

By contrast, in the porcine model, the anterior and posterior regions of the EP showed isotropic properties, but the central region was characterized by anisotropic properties (p < 0.0001). The change in the isotropic properties of the TB structure of the EP in the central region was consistent with the change in the arrangement of trabeculae observed during the microscopic analysis. The obtained DA values for this model were 0.3106 ± 0.1048 and 0.4348 ± 0.0960, respectively, for the anterior and posterior regions (p = 0.0429) and 0.7863 ± 0.0587 for the central region (p < 0.0001 with respect to the anterior region and p = 0.0429 with respect to the posterior region).

Also, it is worth noting that in the central region of the porcine model, the shape of the trabeculae also changed, which showed a plate tendency (mean SMI of −0.3690 ± 0.8750) compared to the mixed plate/rod structure in the anterior and posterior (p = 0.0038) regions (SMIs of, respectively, −0.8414 ± 1.112 and −1.498 ± 1.317, Figure 7). In the case of the microstructure of the trabeculae of the bovine EP, in the entire analysed EP structure they had rod shape with a cylindrical cross‐section, with the mean SMIs of 2.777 ± 0.9025, 3.396 ± 0.5103, and 0.2601 ± 0.7326, respectively, for the anterior, middle, and posterior regions. The ovine model showed a mixed TB structure, which was more plate‐like in the posterior region than in the anterior (p = 0.0373) and middle (p = 0.0014) regions.

FIGURE 7.

FIGURE 7

Open in a new tab

The values of the SMI parameter characterizing the shape of the trabeculae in relation to various animal models and analysed areas. The boxes show 25th and 75th percentiles (distribution of the data); the lines inside the boxes represent the median, while the whiskers indicate the minimum and maximum values. The presented images of trabeculae were obtained using micro‐CT

4. DISCUSSION

Based on the conducted research, the authors of this study obtained data on the shape of EPs of the lumbar spines of the porcine, ovine, and bovine (tail) models at different scale levels. The mesoscale analysis allowed us to determine the total thickness of caudal and cranial EPs. The microscopic analysis at the microscale allowed the specification of the EPs with differentiation of their individual layers. Additionally, the use of micro‐CT allowed for the specification of the TB microarchitecture of the EP layer.

The human EP exists in the form of the epiphyseal ring, which forms on the periphery of the cranial and caudal surfaces of the vertebral body (Boos & Aebi, 2008), while in the middle part there is only a thin layer of mineralized subchondral tissue. In the region adjacent to the nucleus, the thickness of calcification is the lowest, but it increases progressively toward the outer annulus (Coventry, 1969; Nosikova et al., 2012; Roberts et al., 1989; Zhao et al., 2009). In addition, the EP shows a characteristic narrowing in its central‐posterior region, occurring regardless of the analysed spinal segment and location of the EP in relation to the IVD (cranial /caudal) (Broom & Thambyah, 2018; van der Houwen et al., 2010; Lou et al., 2016; Zhao et al., 2009). In the lumbar spine, there are large visible differences in thickness between the cranial and caudal EPs, with an average thickness of 1.03 ± 0.24 mm for the cranial EP, 0.78 ± 0.16 mm for the caudal EP, and the greatest thickness in the posterior region (Wang et al., 2011; Zhao et al., 2009). An important feature describing the EP geometry is the so‐called concavity. This knowledge is particularly important for stabilizing the spine with intervertebral implants (including artificial IVDs) and designing spinal stabilization systems. The cranial and caudal surfaces of vertebral bodies may have one of three types of shape, i.e. concave, flat, or irregular. The caudal EPs of the lumbar region are most often concave in contrast with the cranial EPs, which are more flattened (Duran et al., 2017).

This study identified and described the shape of the EP of each of the tested animal models. The porcine EP had a characteristic increased circumferential thickness (~3 mm), with a significant narrowing observed in the central region both on the caudal and cranial EPs (between 50%–60%). However, the narrowing of the EP by itself was not synonymous with the occurrence of the concavity of the EP surface, it only indicated a change in the thickness in the sagittal plane. The thickness of the caudal EP in the pig was the smallest of the analysed models and its shape (slightly concave both in the cranial and caudal EPs) most closely resembled the shape of the human EP. The cranial ovine EP had a constant thickness throughout the cross‐section and the caudal EP showed ~35% narrowing in the central region. Moreover, the cranial EP is convex (Sheng et al., 2010), while the caudal EP is concave. The cranial EP of the sheep was the thinnest of all the analysed EPs, and its thickness in the peripheral regions was 2.58 ± 0.34 mm. On the other hand, the thickest EPs were observed in the bovine tail model, where the cranial EP had the lowest values in the posterior region, while the caudal EP had a negligible narrowing (~5%) in the central region. In all analysed animal models, the significant narrowing in the SB layer middle region was also observed.

The TB layer was the largest component (by volume) of the EP, with an average thickness of 2.14 mm. It was this structure that shows a characteristic narrowing in the central area of the EP in both the porcine and ovine models. By contrast, in the middle part of the EP in the bovine tail, there was a slight thickening in relation to the anterior and posterior parts, which makes this model the least favorable in the case of research on interbody implants.

The EP is covered with hyaline cartilage composed of collagen fibers, arranged parallel to the mineralized subchondral layer (Inoue, 1981; Moore, 2006). The average thickness of the cartilaginous layer in the human lumbar spine is 0.77 ±0.24 mm. This value changes in the anteroposterior direction of the vertebra, where there are significant differences in thickness. The cartilaginous layer has the minimum thickness in the central part of the EP (0.54 ± 0.12), which is 44% lower than in the anterior and posterior regions of the EP (Moon et al., 2013; Roberts et al., 1989). A similar character of the thickness distribution is shown by the CEP layer of the porcine and bovine models, where thickness reaches up to 1 mm in the anterior and posterior EPs, while in the middle part there is a decrease in thickness that also affects the characteristic concave shape of the EP. This is especially visible in the porcine model. At the same time, the porcine model also shows a drastic decrease in the thickness of the hyaline in relation to the anterior and posterior parts, even to the value of 0.14 mm. The thickness of the CEP layer plays a very important role in research on, among others, nutritional processes occurring on the border of the vertebra and the IVD or vertebral EP permeability. Hence, the choice of the appropriate animal model is crucial to the correctness of the conducted analyses.

Due to the different processes of growth and development of human and animal vertebrae, the results of the analysis of the GP of the animal EP are difficult to apply to the human spine (Casaroli et al., 2018; Hasler et al., 2010). However, in the current study, the GP layer had a practically constant thickness and was similar in all the tested models. This is important because it indicates that the analysed preparations were of a similar age. In animals, the GP undergoes mineralization as the aging process progresses (Casaroli et al., 2018; Rodrigues et al., 2017).

The presented geometrical and structural properties of individual layers of the EP allow us to treat this structure in mechanical terms as a multilayer composite, such as a sandwich structure (Broom & Thambyah, 2018). The increase in the core thickness of such composites significantly increases the strength parameters of the structure (Campbell, 2004). In our case, the inner layer is the TB, which is largely responsible for the total EP thickness. According to the authors, the observed much higher thickness of the EP in quadruped animals compared to the human model is due to the need to achieve a stronger and more durable structure connecting the vertebral bodies with the IVDs. Stabilization of a horizontally oriented spine requires greater muscular forces and passive tension, which result in greater loads than in the case of vertically oriented human spines (Alini et al., 2008; Casaroli et al., 2018).

However, the most important and interesting discovery is the anisotropy of the TB structure in the VEP layer of the porcine model compared to the ovine and bovine models. In the pig spine, the trabeculae in the middle of the EP are parallel to the vertebra. By comparison, in both ovine and bovine models, the trabeculae are oriented perpendicularly to the vertebral area. A change in the orientation of the trabeculae affects the EP strength. The observed parallel arrangement of the trabeculae in the middle part of the TB layer of the porcine model corresponds to some extent to the structure of the human EP. According to research by Dall’Ara (Dall'Ara et al., 2013), mineralized collagen fibers of bone tissue lie in the plane of the EP, so its mechanical properties correspond (in terms of the obtained values) to the properties of the external compact bone tissue of the vertebral body, in which collagen fibers are arranged vertically (along the long axis of the spine).

In the ovine and bovine models, the central region of the EP was characterized by large isotropy and trabeculae of mixed or rod‐like shape. On the other hand, the structure of the trabeculae observed in the pig in this region consisted of plate‐like trabeculae of anisotropic nature, which was also confirmed by Laffosse et al. (Laffosse et al., 2010). This observation suggests that the above animals may have different mechanisms of vertebral EP formation or different EP operating conditions during functional spinal movements. The trabeculae of the TB layer of the human EP are also plate‐like in structure, which changes with age into a predominantly rod‐like structure (Paietta et al., 2013).

An analysis of the microarchitecture of the TB layer showed that porcine EPs were characterized by a structure with the lowest porosity (~30%) compared to ~70% porosity obtained for the ovine and bovine tail models. The epiphyseal ring in human EPs shows a similar porosity (Malandrino et al., 2014) or a slightly greater porosity of ~40%–50% (Zehra et al., 2015). However, porcine trabeculae are twice as thick as in other animal models and ~30% thicker than in human EPs (Rodriguez et al., 2012; Zehra et al., 2018). At the same time, the number of trabeculae and their spacing in the peripheral regions is similar between all the analysed animal models.

The structural analysis of the individual EP layers of the animal models presented in this study is an important contribution to the existing state of knowledge in this field.

Based on the conducted structural and morphometric analyses of the examined animal models, the authors of this study identified the porcine model as best reflecting the shape and structure of the human EP and consider it to be the best surrogate model for the human EP model. This choice is particularly important in the context of biomechanical research related to spinal implants, in which the shape of the EP (including the existing narrowing and concavity in the central‐posterior part) and its structure are critical to achieving proper fixation and ensuring proper transfer of loads.

One of the limitations of this study is the fact that the research covered only the EPs of the lumbar spine. In different sections of the spine, both the shape and size of the vertebrae differ from each other, which also affects the geometrical parameters and the shape of the EP. However, there is a similar structural pattern between all the vertebrae (Galbusera, 2018). Therefore, similar structural and morphometric relationships of the EP should occur in the remaining sections of the spine. Moreover, this study did not present an analysis of the mechanical parameters characterizing the EPs of the examined animal models. This aspect of the research is also crucial in terms of selecting the appropriate animal model for planned biomechanical research. The aforementioned scope of research, including analysis of the mechanical properties at the macro and nano scales, will be the subject of future research by the authors. On the other hand, the data on structural and morphometric parameters and mechanical properties of the EP of the ovine model were presented in a study by Wojtków and Pezowicz (Wojtków et al., 2019).

5. CONCLUSIONS

This study shows that in terms of structure and morphometry, porcine EPs appear to be the best surrogate for the human EP. These EPs were characterized by increased thickness occurring circumferentially and a significant narrowing in the central region (50–60%), with a simultaneous occurrence of the concavity of the surface. Such relationships, also occurring in human EPs, were not observed in the ovine or bovine tail models. The above‐mentioned narrowing of the EP height/ thickness was most visible in the TB layer of the EP, but also in the SB and CEP layers. In all the models, the GP layer had a similar thickness, but this element was not present in the human EPs.

The microarchitecture of the TB layer of the EP also most closely resembled human tissue in the porcine model. Both cases showed about 30%–40% porosity and parallel arrangement of the trabeculae to the vertebral body in the central area of the EP (in the case of human tissues, this function is performed by collagen fibers in the CEP). It is worth noting that in the porcine model the trabeculae were arranged differently depending on the EP region: the middle region showed an anisotropic structure, the nature of which changed to isotropic structure toward the peripheral regions. This relationship was not observed in other animal models.

AUTHOR CONTRIBUTIONS

Magdalena Wojtków: concept/design, acquisition of data, data analysis/interpretation, drafting the manuscript, approval of the article. Maciej Głowacki: concept/design, critical revision of the manuscript, approval of the article. Celina Pezowicz: concept/design, data interpretation, critical revision of the manuscript, approval of the article.

REFERENCES

  1. Alini, M. , Eisenstein, S.M. , Ito, K. et al. (2008) Are animal models useful for studying human disc disorders/degeneration? European Spine Journal, 17, 2–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Boos, N. & Aebi, M. (2008) Spinal disorders ‐ Fundamentals of diagnosis and treatment. Berlin, Germany: Springer. [Google Scholar]
  3. Broom, N.D. & Thambyah, A. (2018) The soft‐hard tissue junction structure, mechanics and function. Cambridge, United Kingdom: Cambridge University Press. [Google Scholar]
  4. Busscher, I. , Ploegmakers, J.J. , Verkerke, G.J. & Veldhuizen, A.G. (2010) Comparative anatomical dimensions of the complete human and porcine spine. European Spine Journal, 19, 1104–1114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Campbell, F. (2004) Adhesive bonding and integrally cocured structure: A way to reduce assembly costs through parts integration. In: Campbell, F. (Ed.) Manufacturing processes for advanced composites. Oxford, United Kingdom: Elsevier, pp. 241–301. [Google Scholar]
  6. Casaroli, G.W. , Wade, K. , Villa, T. & Wilke, H.J. (2018) Animal models for spine biomechanics. In: Galbusera, F. & Wilke, H.J. (Eds.) Biomechanics of the spine basic concepts, spinal disorders and treatments. London, United Kingdom: Elsevier, pp. 279–296. [Google Scholar]
  7. Coventry, M.B. (1969) Anatomy of the intervertebral disk. Clinical Orthopaedics and Related Research, 67, 9–15. [PubMed] [Google Scholar]
  8. Dall'Ara, E. , Karl, C. , Mazza, G. et al. (2013) Tissue properties of the human vertebral body sub‐structures evaluated by means of microindentation. Journal of the Mechanical Behavior of Biomedical Materials, 25, 23–32. [DOI] [PubMed] [Google Scholar]
  9. Duran, S. , Cavusoglu, M. , Hatipoglu, H.G. , Sozmen Ciliz, D. & Sakman, B. (2017) Association between measures of vertebral endplate morphology and lumbar intervertebral disc degeneration. Canadian Association of Radiologists Journal, 68, 210–216. [DOI] [PubMed] [Google Scholar]
  10. Galbusera, F. (2018) The spine: Its evolution, function, and shape. In: Galbusera, F. & Wilke, H.J. (Eds.) Biomechanics of the spine basic concepts, spinal disorders and treatments. London, United Kingdom: Academic Press, pp. 3–9. [Google Scholar]
  11. Grant, J.P. , Oxland, T.R. & Dvorak, M.F. (2001) Mapping the structural properties of the lumbosacral vertebral endplates. Spine, 26(8), 889–896. [DOI] [PubMed] [Google Scholar]
  12. Grant, J.P. , Oxland, T.R. , Dvorak, M.F. & Fisher, C.G. (2002) The effects of bone density and disc degeneration on the structural property distributions in the lower lumbar vertebral endplates. Journal of Orthopaedic Research, 20, 1115–1120. [DOI] [PubMed] [Google Scholar]
  13. Hasler, C. , Sprecher, C.M. & Milz, S. (2010) Comparison of the immature sheep spine and the growing human spine. Spine, 35(23), E1262–E1272.. [DOI] [PubMed] [Google Scholar]
  14. Inoue, H. (1981) Three‐dimensional architecture of lumbar intervertebral discs. Spine, 6(2), 139–146. [DOI] [PubMed] [Google Scholar]
  15. Laffosse, J.M. , Odent, T. , Accadbled, F. et al. (2010) Micro‐computed tomography evaluation of vertebral end‐plate trabecular bone changes in a porcine asymmetric vertebral tether. Journal of Orthopaedic Research, 28, 232–240. [DOI] [PubMed] [Google Scholar]
  16. Lou, J. , Liu, H. , Rong, X. , Li, H. , Wang, B. & Gong, Q. (2016) Geometry of inferior endplates of the cervical spine. Clinical Neurology and Neurosurgery, 142, 132–136. [DOI] [PubMed] [Google Scholar]
  17. Malandrino, A. , Lacroix, D. , Hellmich, C. , Ito, K. , Ferguson, S.j. & Noailly, J. (2014) The role of endplate poromechanical properties on the nutrient availability in the intervertebral disc | Elsevier Enhanced Reader. Osteoarthritis and Cartilage, 22, 1053–1060. [DOI] [PubMed] [Google Scholar]
  18. Moon, S.M. , Yoder, J.H. , Wright, A.C. , Smith, L.J. , Vresilovic, E.J. & Elliott, D.M. (2013) Evaluation of intervertebral disc cartilaginous endplate structure using magnetic resonance imaging. European Spine Journal, 22, 1820–1828. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Moore, R.J. (2000) The vertebral end‐plate: What do we know? European Spine Journal, 9, 92–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Moore, R.J. (2006) The vertebral endplate: disc degeneration, disc regeneration. European Spine Journal, 15, 333–337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Nosikova, Y.S. , Santerre, J.P. , Grynpas, M. , Gibson, G. & Kandel, R.A. (2012) Characterization of the annulus fibrosus–vertebral body interface: identification of new structural features. Journal of Anatomy, 221, 577–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Paietta, R.C. , Burger, E.L. & Ferguson, V.L. (2013) Mineralization and collagen orientation throughout aging at the vertebral endplate in the human lumbar spine. Journal of Structural Biology, 184, 310–320. [DOI] [PubMed] [Google Scholar]
  23. Palepu, V. , Rayaprolu, S. & Nagaraja, S. (2019) Differences in trabecular bone, cortical shell, and endplate microstructure across the lumbar spine. International Journal of Spine Surgery, 13, 361–370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Panjabi, M.M. , Chen, N.C. , Shin, E.K. & Wang, J.L. (2001) The cortical shell architecture of human cervical vertebral bodies. Spine (Phila Pa 1976), 26, 2478–2484. [DOI] [PubMed] [Google Scholar]
  25. Pitzen, T. , Schmitz, B. , Georg, T. et al. (2004) Variation of endplate thickness in the cervical spine. European Spine Journal, 13, 235–240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Roberts, S. , McCall, I.W. , Menage, J. , Haddaway, M.J. & Eisenstein, S.M. (1997) Does the thickness of the vertebral subchondral bone reflect the composition of the intervertebral disc? European Spine Journal, 6, 385–389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Roberts, S. , Menage, J. & Urban, J.P. (1989) Biochemical and structural properties of the cartilage end‐plate and its relation to the intervertebral disc. Spine (Phila Pa 1976), 14, 166–174. [DOI] [PubMed] [Google Scholar]
  28. Rodrigues, S.A. , Thambyah, A. , Broom, N.D. (2017) How maturity influences annulus‐endplate integration in the ovine intervertebral disc: a micro‐ and ultra‐structural study. Journal of Anatomy, 230, 152–164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Rodriguez, A.G. , Rodriguez‐Soto, A.E. , Burghardt, A.J. , Berven, S. , Majumdar, S. & Lotz, J.C. (2012) Morphology of the human vertebral endplate. Journal of Orthopaedic Research, 30, 280–287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Rutges, J.P. , van der Jagt, O.P. , Oner, F.C. et al. (2011) Micro‐CT quantification of subchondral endplate changes in intervertebral disc degeneration. Osteoarthritis Cartilage, 19, 89–95. [DOI] [PubMed] [Google Scholar]
  31. Sheng, S.R. , Wang, X.Y. , Xu, H.Z. , Zhu, G.Q. & Zhou, Y.F. (2010) Anatomy of large animal spines and its comparison to the human spine: a systematic review. European Spine Journal, 19, 46–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Smit, T.H. (2002) The use of a quadruped as an in vivo model for the study of the spine ‐ biomechanical considerations. European Spine Journal, 11, 137–144. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Szotek, S. , Szust, A. , Pezowicz, C. , Majcher, P. & Bedzinski, R. (2004) Animal models in biomechanical spine investigations. Bulletin of the Veterinary Institute in Puławy, 48, 163–168. [Google Scholar]
  34. van der Houwen, E.B. , Baron, P. , Veldhuizen, A.G. , Burgerhof, J.G.M. , van Ooijen, P.M.A. & Verkerke, G.J. (2010) Geometry of the intervertebral volume and vertebral endplates of the human spine. Annals of Biomedical Engineering, 38, 33–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Wade, K. (2018) Chapter 8 ‐ vertebral endplates. In: Galbusera, F. & Wilke, H.‐J. (Eds.) Biomechanics of the spine basic concepts, spinal disorders and treatments. London, United Kingdom: Academic Press, pp. 125–140. [Google Scholar]
  36. Wang, Y. , Battie, M.C. , Boyd, S.K. & Videman, T. (2011) The osseous endplates in lumbar vertebrae: thickness, bone mineral density and their associations with age and disk degeneration. Bone, 48, 804–809. [DOI] [PubMed] [Google Scholar]
  37. Wilke, H.J. , Kettler, A. & Claes, L.E. (1997) Are sheep spines a valid biomechanical model for human spines? Spine, 22(20), 2365–2374. [DOI] [PubMed] [Google Scholar]
  38. Wojtkow, M. , Kiełbowicz, Z. , Bieżyński, J. & Pezowicz, C. (2019) Quantitative and qualitative assessment of the impact of osteoporosis on endplate layers. Biocybernetics and Biomedical Engineering, 39, 797–805. [Google Scholar]
  39. Zehra, U. , Robson‐Brown, K. , Adams, M.A. & Dolan, P. Porosity and thickness of the vertebral endplate depend on local mechanical loading. Spine, 40(15), 1173–1180. [DOI] [PubMed] [Google Scholar]
  40. Zehra, U. , Robson‐Brown, K. , Adams, M. & Dolan, P. (2018) Vertebral endplate porosity reflects nutritional and mechanical demands. Orthopaedic Proceeding, 96‐B. [Google Scholar]
  41. Zhao, F.D. , Pollintine, P. , Hole, B.D. , Adams, M.A. & Dolan, P. (2009) Vertebral fractures usually affect the cranial endplate because it is thinner and supported by less‐dense trabecular bone. Bone, 44, 372–379. [DOI] [PubMed] [Google Scholar]
  42. Zhou, S.H. , McCarthy, I.D. , McGregor, A.H. , Coombs, R.R. & Hughes, S.P. (2000) Geometrical dimensions of the lower lumbar vertebrae–analysis of data from digitised CT images. European Spine Journal, 9, 242–248. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Anatomy are provided here courtesy of Anatomical Society of Great Britain and Ireland

RESOURCES