Three-dimensional morphology of the pericellular matrix of intervertebral disc cells in the rat

Abstract

Intervertebral disc cells are surrounded by a pericellular matrix that is biochemically and morphologically distinct from other extracellular matrix regions. Although the function of the pericellular matrix is not fully understood, prior studies of pericellular matrix-chondrocyte regions in articular cartilage (termed ‘chondrons’) suggest that the size, shape, and mechanical properties of the pericellular matrix significantly influence the micromechanical environment of the contained cells. A first step in understanding the role of the pericellular matrix in the intervertebral disc is to quantify the three-dimensional morphology and zonal variations of these regions across the disc. In this study, three-dimensional reconstructions and morphometric measurements of pericellular matrix-cell regions were obtained in situ using fluorescence confocal microscopy of en bloc sections of nucleus pulposus and anulus fibrosus of the rat disc immunolabeled for type VI collagen. The morphology of the pericellular matrix and cells varied significantly across regions, with distinct pericellular matrix aspect ratios (largest/smallest diameter) showing shapes that were generally large and rounded in the nucleus pulposus (average of 1.9), and ellipsoidal and discoidal in the inner (2.4) and outer anulus fibrosus (2.8). The average pericellular matrix volume per cell was found to be significantly larger in the nucleus (6424 µm³) than that of inner (1903 µm³) and outer (1433 µm³) anulus. Pericellular matrix regions containing 1 or 2 cells were the dominant subgroup in the rat intervertebral disc at both 1 and 12 months of age. Multicellular pericellular matrix regions were present more often in the younger nucleus pulposus and outer anulus fibrosus. The orientation of the pericellular matrix regions further varied significantly across the disc, reflecting local collagen matrix architecture. These studies provide new information on the organization and shape of intervertebral disc cells and their surrounding pericellular matrix, which may provide new insights into the mechanisms that regulate cell-matrix interactions.

Introduction

The intervertebral disc (IVD) is a heterogeneous soft tissue that contributes to flexibility and load support in the spine. The IVD consists of three distinct anatomical regions, the nucleus pulposus (NP), and the inner and outer anulus fibrosus (AF) that exhibit biochemical and biomechanical differences during development, aging and degeneration of the disc (Rufai et al. 1995; Antoniou et al. 1996; Nerlich et al. 1997; Hayes et al. 2001; Roughley, 2004). The NP is an amorphous and proteoglycan-rich material that confers a high swelling pressure that contributes to fluid pressurization for compressive load support (Urban & Maroudas, 1981). The AF contains a highly organized and collagen-rich lamellar structure that acts to resist the high tensile loads generated by the contained NP (Galante, 1967; Nachemson & Elfstrom, 1970; Setton & Chen, 2004). Disc cells also have demonstrated regional variations in cell morphology and cytoskeletal composition across the entire tissue (Trout et al. 1982a; Postacchini et al. 1984; Errington et al. 1998; Ishii et al. 1991; Guilak et al. 1999b; Baer & Setton, 2000; Hastreiter et al. 2001; Bruehlmann et al. 2002; Johnson & Roberts, 2003). AF cells may be elongated and align with the principal collagen fiber direction, and may extend cell processes into extracellular matrix regions (Errington et al. 1998; Bruehlmann et al. 2002; Baer et al. 2003; Johnson & Roberts, 2003). Cells of the immature NP may be more rounded and exist in ‘clusters’ with significant changes in morphology observed with maturity and aging of the IVD (Trout et al. 1982a,b; Ishii et al. 1991; Hastreiter et al. 2001).

IVD cells have been shown, in some studies, to be surrounded by a lacunae or ‘nest’ of extracellular matrix (ECM) that is distinct from other matrix regions (Trout et al. 1982a; Postacchini et al. 1984; Roberts et al. 1991b). This pericellular matrix (PCM) together with the enclosed cell(s) consists of a distinct cell-matrix unit (CMU) that has been termed ‘chondron’ for chondrocytes in articular cartilage (Poole et al. 1991; Poole, 1992; Guilak et al. 1999a; Choi et al. 2007). The PCM of IVD cells contains many of the same molecular constituents as other ECM regions, including collagen types I (mainly AF), II (mainly NP) (Roberts et al. 1991b), III (Roberts et al. 1991c; Aulisa et al. 1998), IX and XI (Roberts et al. 1991b; Eyre et al. 2002), aggrecan (Antoniou et al. 1996; Sztrolovics et al. 1997), fibronectin and laminin (Oegema et al. 2000; Hayes et al. 2001), but is generally defined by the unique presence of type VI collagen in the IVD (Wu et al. 1987; Roberts et al. 1991a,b), as for cartilage (Poole et al. 1988). Although the precise function of the PCM is not fully understood, prior studies have quantified mechanical properties of the PCM in chondrons and found them to be functionally distinct from that of the cell and the adjoining ECM (Alexopoulos et al. 2003, 2005a,b; Guilak et al. 2006; Choi et al. 2007). This mechanically functional layer may play important roles in transducing mechanical loading to the cell (Alexopoulos et al. 2005a), as well as regulating physical interactions of IVD cells with their ECM. Given evidence of substantial differences in IVD cell morphology and ECM composition across regions, it is likely that the PCM surrounding IVD cells varies in its morphology as well as physical function across the IVD.

In prior studies, three-dimensional reconstructions of IVD cells imaged via confocal microscopy were obtained using solid modeling algorithms, from which parameters of cell morphology were obtained (Baer & Setton, 2000). While three-dimensional cell morphology has been documented for IVD cells, no quantitative information is available on the presence or morphology of the PCM in the IVD. The objective of this study was to quantify the three-dimensional morphology of the in situ PCM using confocal imaging of en bloc sections of the IVD that were fluorescently-labeled for type VI collagen as a marker of the PCM. Studies were performed on rat IVD tissue from skeletally immature and mature animals to evaluate differences in PCM morphology among regions of the AF and the NP, as well as their changes with age.

Materials and methods

Specimen preparation

Intervertebral discs (IVD) were harvested from the lumbar spines of 1 month and 12 month old rats (Fischer 344, Harlan Inc., Chicago, IL, USA) immediately after sacrifice. The discs were fixed overnight in 4% paraformaldehyde (pH = 7.4) and thick frozen sections (60 µm) were taken from each IVD along the axial direction.

Immunohistochemistry for type VI collagen

A modified immunohistochemical staining protocol was adopted for en bloc type VI collagen immunostaining (Youn et al. 2006; Choi et al. 2007). In brief, the frozen sections were rinsed in PBS, incubated with 0.1 m trypsin-EDTA (37 °C for 2 hrs), and blocked with 10% normal donkey serum (25 °C for 3 h). Specimens were then incubated with primary antibody (1:50, rabbit anti-human collagen type VI, RDI-600-401-108, Fitzgerald, Concord, MA, USA) and followed by incubation with a fluorescent secondary antibody (1:8, FITC-conjugated anti-rabbit IgG, RDI-711095152, Fitzgerald). Specimens were rinsed again in TBS and mounted for imaging via confocal laser scanning microscopy (LSM 510, Zeiss, Jena, Germany).

Sampling and imaging of the PCM

Cells were often observed to be located within one contiguous PCM region. These cell-matrix units (CMU) were divided into three sub-groups depending on the number of cells enclosed in one contiguous PCM region (1 cell, 2 cells, or 3 or more cells within one PCM). In order to investigate the distribution of CMU subgroups over the entire disc, four separate fields of view were scanned across anterior, posterior, and lateral regions in each IVD (Fig. 1). The distribution of cells belonging to each of the three CMU subgroups was tabulated.

Fig. 1.

Photographs of intact IVD from rat lumbar spine and immunohistological sections. (i) IVD showed gelatinous central region and surrounding concentric fiber-reinforced lamellae region. NP: nucleus pulposus, AF: anulus fibrosus. (ii) The entire rat disc section (thickness: 60 µm) immunolabelled for type VI collagen showed PCM alignment with the fiber orientation in the AF. Inner AF (IAF) region was defined as the 20% of AF from the inner edge. Outer AF (OAF) region was defined as 20%∼80% of the AF tissue. P: posterior, A: anterior. (iii–v) Two-dimensional PCM morphology from the confocal microscopy varied across NP (iii), inner (iv) and outer AF regions (v) in the rat IVD. (Green – Type VI collagen; Red – cell nuclei; Bar = 20 µm).

For three-dimensional (3D) geometric reconstructions, a total of 198 CMUs were randomly selected for analysis from the NP, inner and outer AF regions (Fig. 1, Table 1). For each selected CMU, a 3D volume was obtained from images of 30–60 serial sections (512 × 512 pixels), taken at intervals of 0.5 µm (AF) or 1 µm (NP) using a 63x objective (water-immersion lens, 1.3NA, Zeiss). To evaluate differences in cell density in situ across the regions and between two age groups, cell number in the reconstructed 3D volume was counted. The in-situ cell density was calculated from the total cell number included within the entire tissue volume of each thick section examined.

Table 1.

Summary of PCM volume and sample numbers in the rat IVD

	NP			IAF			OAF
PCM Volume (µm3)	1 cell†	2 cells	3+ cells	1 cell†	2 cells	3+ cells	1 cell†	2 cells	3+ cells
1 month	6021 ± 1770	13792 ± 5368‡	23675 ± 9205	2214 ± 931	3396 ± 1224	4656 ± 1707	1706 ± 492	2528 ± 836	3672 ± 1653
PCM	(12)	(12)	(6)	(17)	(13)	(6)	(12)	(12)	(6)
Cell	(12)	(24)	(21)	(17)	(26)	(19)	(12)	(24)	(18)
12 months	7740 ± 2184§	10391 ± 3081	15315 ± 8016	2982 ± 1158§	5069± 1968‡'§	7885 ± 2232§	2210 ± 1352	4818 ± 2151§	5567 ± 2393
PCM	(12)	(12)	(6)	(21)	(15)	(6)	(12)	(12)	(6)
Cell	(12)	(24)	(18)	(21)	(30)	(20)	(12)	(24)	(21)

The PCM volume (mean ± s.d.) varies significantly in different CMU subgroups across three anatomical regions of the disc for 1 month and 12-month-old rat. Sample number (n): n-value in parentheses indicates numbers of individual PCM or cell evaluated for this study.

^†P < 0.05, as compared to 2 and 3+ cell CMU subgroups in the same region (ANOVA and post hoc Tukey test).

^‡P < 0.05, significant different from the 3+ cell CMU subgroup (ANOVA).

^§P < 0.05, significantly different from the corresponding subgroups at 1 month old (ANOVA).

Three-dimensional quantitative measurements of the PCM and cell

3D reconstruction and morphological measurement of CMUs were also performed using a custom program and graphic user interface, as described previously (Youn et al. 2006). Briefly, serial images were smoothed using a median filter, and a modified marching cubes algorithm (Guilak, 1994) was used to render the volumes corresponding to CMU, and to cell, as well as their corresponding surfaces based on the 3D coordinates of the acquired images (IsoSurface function, MATLAB, The MathWorks, Natick, MA, USA). Based on a pre-determined optimal iso-intensity threshold (Youn et al. 2006), the surface contours were formed by triangular surface patches containing three vertices that fell on the border between PCM and ECM for the CMUs or between PCM and Cell for the cells. Volume was reconstructed from tetrahedral elements with one random vertex inside the volume. Calibration studies have shown these methods for 3D volume imaging and reconstruction to be accurate within 2% (when estimated using calibration objects, Guilak, 1994; Youn et al. 2006).

To analyze unique features of the PCM and cell across regions and between two age groups, morphological parameters including volume, surface area, principal direction of the CMU, and principal diameters from shape fitting, were extracted from the rendered 3D volumes and surface areas (Fig. 2). The surface area of the PCM was calculated as the sum of areas of each triangular surface patch defining the outer border between the PCM and ECM. Similarly, cell surface area was calculated from the inner surface of PCM. The CMU volume was defined as the total volume enclosed by the reconstructed surface, which was determined as the sum of the volume of multiple tetrahedrons; cell volume was calculated in the same manner. PCM volume was obtained by subtracting the cell volume from the CMU volume. To determine the principal angle of CMU orientation, all the vertices forming the iso-intensity surface contour were screened to find the two vertices with the maximum distance, forming the principal axis of the CMU. The orientation of the principal direction was calculated by its direction cosine with respect to the unit vector along the axial direction of the spine.

Fig. 2.

A schematic of the geometric parameters measured from 3D images. (left) The principal angle of the CMU (α) relative to the axial axis of the spine (z) was determined based on the principal axis (PA) of the CMU and cells. (right) The aspect ratio was defined by the ratio between three diameters (D₃ > D₂ > D₁) of the fitted ellipsoid of the CMU and cells.

The geometries of the CMU and cell were separately described by a best-fit ellipsoid of their respective 3D surface (Fig. 2) (Baer & Setton, 2000). Three diameters were used to characterize the size of the CMU and cell, where D₃ corresponds to the long axis of the CMU (D₃ > D₂ > D₁). The aspect ratios between the major (D₃) and minor diameters (D₂ or D₁) were used to characterize the shape of the CMU and cell. The average thickness of the PCM along the transverse plane (defined by D₂, D₁) was calculated by subtracting the radius of the corresponding fitted cell from that of the fitted CMU.

Regional and subgroup variations in the morphometric parameters of the PCM and cells were tested by one-factor or two-factor analysis of variance (ANOVA and post hoc Tukey test; S-PLUS. Insightful Corp., Seattle, WA, USA) with statistical significance reported at the 95% confidence level.

Results

Virtually all cell nuclei were associated with a PCM region that was positively labeled with type VI collagen (Fig. 1). Cell density, as well as the incidence and morphology of PCM-cell regions varied significantly among different anatomical regions and CMU subgroups, as shown for 3D reconstructions from serial images (Fig. 3). The in situ cell density in the inner and outer AF was significantly higher than that of the NP at 1 and 12 months of age in the rat (Fig. 4, P < 0.00001, ANOVA and post hoc Tukey test), but decreased at all regions by an average of ∼30% at 12 months of age (P < 0.05, ANOVA). CMUs in the NP region exhibited a more rounded shape than other regions. CMUs in the inner AF were often ellipsoidal, however, and those in the outer AF were more elongated and discoidal in shape. Multiple cells were enclosed within one continuous PCM region in the NP. These CMUs were organized in a spheroidal shape with no preferable spatial orientation. CMUs in the AF also contained multiple cells but were more highly oriented with average principal angles of 65.2 ± 17.9° (mean ± SD, n = 138) relative to the axial axis of the spine, in alignment with the principal collagen fiber direction of the AF (Cassidy et al. 1989; Marchand & Ahmed, 1990). Disc cell morphology, defined as the internal surface of the PCM, was also observed to vary among different regions (Figs 1, 3).

Fig. 3.

Three-dimensional reconstructions of 1 cell (A1-F1), 2 cell (A2-F2) and 3+ cell CMUs (A3-F3). A and D: 2D confocal sections of the PCM; B and E: reconstructed 3D images of the PCM and enclosed cells; C and F: 3D fitted ellipsoids (denoted by yellow-red grid surfaces). A–C are from the NP and D–F are from the AF.

Fig. 4.

The in situ cell density (upper) and the average PCM thickness on the transverse plane of the CMU long axis (lower) in the rat IVD at 1 and 12 months old. *P < 0.0001, as compared to inner and outer AF (ANOVA and post hoc Tukey test). ^P < 0.05, ^^P < 0.0001, as compared to 12 months old (ANOVA).

In NP tissues of the 1-month-old rat, a majority of cells (> 80%) resided in CMUs containing 2 cells or 3+ cells (Fig. 5), reflecting prior observations of ‘cell clustering’. The high incidence of CMUs containing 2, 3 or more cells was associated with a numerical average of four cells per CMU across all NP. In 12-month-old NP, however, the number of cells residing in 1 cell CMUs increased (∼40%), while that in 2 cell CMUs decreased (3+ cell CMUs remained unchanged). In inner AF regions, a majority of cells were found in 1 or 2 cell CMUs (more than ∼90%) at both 1 and 12 months of age with a numerical average of three cells per CMU. In the immature outer AF regions, cells were present in 3+ cell CMUs (∼45%) in nearly equal numbers with 1 and 2 cell CMUs (∼55%), with an average of four cells per CMU. This trend reversed towards more cells being present in 1 and 2 cell CMUs at 12 months of age (∼80%).

Fig. 5.

Percentage of cells residing in different CMU subgroups in the rat IVD at 1 month (top) and 12 months (bottom). ^P < 0.01, *P < 0.0001, significantly different from each other (ANOVA and post hoc Tukey test).

Significant variations in PCM volume (Table 1) and surface area (data not shown) were found among different CMU subgroups and different regions. The PCM volume was 50%–150% larger in multiple cell CMUs, as compared to 1 cell CMUs. In a corresponding CMU subgroup, PCM volume is always largest in the NP, followed by the inner AF region. Overall, the PCM volume increased 30%–50% from 1 month to 12 months of age in the rat. Similarly, the average PCM volume associated with one cell varied significantly from NP to AF regions at both 1 month (Fig. 6) and 12 months (Fig. S1, supplementary material), with the highest PCM volume per cell in the NP and lowest in the outer AF. Interestingly, no difference for the PCM volume per cell was seen among different CMU subgroups. The ratio of the surface area of the CMUs over PCM volume (SA/V, µm⁻¹) remained relatively constant in all CMU subgroups and also across different anatomical regions (0.4 ± 0.1, mean ± s.d., n = 198). Average PCM thickness, measured on the transverse plane of the long axis (2–4 µm), also did not differ across anatomical regions (P > 0.05, ANOVA) but did increase significantly with age (Fig. 4, P < 0.05, ANOVA).

Fig. 6.

The volume of PCM associated with a single cell in different CMU subgroups in the rat IVD at 1-month old. *P < 0.0001, **P < 0.005, NP significantly different from corresponding CMU subgroup in the inner and outer AF (ANOVA and post hoc Tukey test). ^P < 0.05, significant difference detected between two subgroups in the AF.

The size of the CMUs in the NP was significantly larger than that in the AF regions (Fig. 7). The minor diameters (both D₂ and D₁ on the transverse plane) in the NP (> 20 µm) were twice that of values in the AF (< 20 µm, P < 0.05, ANOVA). The major diameter (D₃ along the long axis) was also significantly larger in the NP than AF, for 1 cell and 2 cell CMUs (P < 0.05, ANOVA and post hoc Tukey test), possibly related to the larger sizes of immature NP cells (Aguiar et al. 1999; Guilak et al. 1999b; Baer et al. 2003; Chen et al. 2006). Considering different CMU subgroups, D₃increased with the number of cells contained within a CMU (∼30, 45, and 60 µm, for 1, 2 and 3+ cell CMUs, respectively) for both NP and AF regions. However, no evidence of a difference was found for D₁ and D₂ among all subgroups (P > 0.05, ANOVA). The major aspect ratio (D₃/D₁) ranged from 1.2 to 3.0 in the NP for all CMU subgroups (Fig. 7). In the AF, for the single cell CMU, this ratio is ∼2.5, but varied to be as high as ∼8 in the 3+ cell CMU. The aspect ratio on the transverse plane (D₂/D₁) showed less difference among different CMU subgroups. It ranged from 1.2 to 2.2 in the NP and AF. All these quantitative measurements reflected the rounded CMU in the NP and elongated CMU morphology in the AF, as seen in the 3D images.

Fig. 7.

The size and shape of CMU in the rat IVD at 1-month old. (left) Three diameters from fitted ellipsoids of 3D reconstructed images of CMUs illustrating size of the PCM in the NP (upper) and AF (lower) regions. (right) The aspect ratio between the diameters (D₃/D₁ upper and D₂/D₁ lower) for all CMUs. Note that the magnitudes of the axis in these two panels are different. †P < 0.05, AF significantly lower than the corresponding group in the NP except D₃ in the 3+ cell CMU subgroup (ANOVA). *P < 0.05, **P < 0.005, significantly different between subgroups for the same diameter (ANOVA).

Discussion

The findings of this study provide new quantitative data for the in situ three-dimensional PCM morphology of IVD cells, revealing significant differences in the shape, size, and cellular content of the PCM across tissue regions, as well as during maturation and aging. CMUs containing 1 or 2 cells are present in the greatest frequency in the rat IVD at both 1 and 12 months of age, with substantial distances of extracellular matrix separating these CMU regions. Across all regions and CMU types, similar values were observed for PCM thickness and PCM volume per cell, suggesting that the cell somehow regulates PCM morphology for optimal features that are preserved across all domains. While these features are not known, they may include characteristics such as a maximum diffusion distance from extracellular matrix regions to the cell, or a minimum thickness of a mechanically functional layer separating cell and extracellular matrix regions.

In both the immature and mature IVD, many NP cells were found closely packed together in multiple-cell-containing CMUs, consistent with the evidence of ‘cell clustering’ (Trout et al. 1982a; Ishii et al. 1991; Hastreiter et al. 2001; Nerlich et al. 2002; Hunter et al. 2004). In general, the percentage of cells present in pairs or in CMUs containing 3 or more cells was much higher than that reported for thin sections of human IVD samples (Hastreiter et al. 2001). These differences may relate to differences in species, age and degeneration between rat and human samples (average 63-years-old, Thompson Grades II-V), but likely relates to the use of 3D reconstruction of PCM morphology that enables visualization of large and contiguous PCM regions. Independent of the number of cells enclosed in one CMU group, however, the CMU aspect ratio remained similar, nearly 2, which may indicate that multiple individual PCM-cell units in the NP form a ball-like CMU in that region. A majority of cells were present in multiple cell CMUs in the outer AF of immature rats, but not mature rats nor inner AF, in which most cells were present in 1 and 2 cell CMUs. In contrast to the NP region, the CMU aspect ratio in the AF region increased uniformly with increasing cell numbers contained within a CMU, indicating cells in this region are arranged as a string-like CMU along the collagen fiber. With respect to age-related changes, decreases in cell density, increases in PCM volume, and a shift towards fewer cells in each CMU were observed with aging, changes that reflect a pattern for cell differentiation and senescence that demonstrate active regulation and reshaping of PCM regions. It is noteworthy that a prior study counting ‘pair density’ of cells in the human IVD reported a positive correlation between numbers of cells contained in pairs with increasing grade of degeneration (Hastreiter et al. 2001). This finding, in contrast to age-related decreases in cell numbers contained in CMUs for the older rats reported here, implies that cell clustering behaviors may be a key feature that distinguished normal aging from degeneration or pathological processes in the IVD (Hastreiter et al. 2001; Roberts et al. 2006). It is not clear from the data in the current study, however, if the observed decreases in cell density with age arise from growth or aging-associated increases in PCM thickness, as density was estimated from representative volumes of fixed size rather than the entire IVD volume.

It is of interest to compare key geometric parameters of PCM morphology in the IVD with those quantified for ‘chondrons’ in articular cartilage (Youn et al. 2006). The inner AF region presents many similar features in CMUs as for chondrons in cartilage, such as the PCM volume per cell (∼2000 µm³ in cartilage) and PCM thickness (∼2–4 µm in cartilage) that are comparable for cartilage. The PCM aspect ratio is also similar between two types of tissues, except for the 3+ cell CMU group in the inner AF, which is much larger than 3+ cell chondron in deep zones of cartilage. In general, the PCM distribution over the tissue is more organized in the IVD, likely due to highly oriented extracellular collagen matrix. Furthermore, the marked tail-like structures observed in cartilage chondrons (Poole et al. 1988) were not seen in the disc. It is worth noting that cell processes have been observed, although infrequently, in the rat IVD when labeled with stains recognizing cytoplasm (Baer et al. 2003). It is likely that the co-labeling of nucleus and PCM structures, but not cellular structures, obscured an ability to visualize cellular processes in the rat IVDs examined here.

Quantitative measurement of 3D geometry of the PCM and cells in situ as performed here has provided a reliable method to quantify changes in PCM features in the immature and mature rat disc. The rat was chosen for study here due to its interest as an animal model for IVD degeneration (Silberberg et al. 1979; Iatridis et al. 1999; MacLean et al. 2003; Elliott & Sarver, 2004; Singh et al. 2005), for screening biological interventions such as cell and growth factor supplementations (Nishimura & Mochida, 1998; Gruber & Hanley, 2002; Crevensten et al. 2004; Kawakami et al. 2005), for its potential for study in organ culture (Risbud et al. 2003; Lim et al. 2006) and for its documented presence of notochordal cells in the immature NP (Hunter et al. 2004). Nevertheless, the morphometric features of the IVD in this small animal model may be limited when extrapolated to the human IVD, particularly when evaluating aging-related changes in the cell and PCM geometric parameters. These rodents exhibit a prolonged period of growth in body mass that extends beyond 12 months and also contain notochordal cells even in the NP of the 12-month-old animal. Both of these features are markedly different from what has been shown for the human IVD, complicating any interpretation of comparable aging-related changes in the geometric parameters of the human. Future studies of the human IVD using these techniques will be of great interest. Also, this study made use of specialized immunohistochemical staining methods that may be subject to some limitations. Use of thick tissue sections with visualization via immunofluorescence was necessary to obtain reconstructions of the large geometric structures corresponding to CMUs in the IVD. Use of thick sections, however, relies upon enzymatic digestion to achieve antibody penetration that required strict control of the digestion conditions (duration, enzyme selection, etc.) to avoid loss of epitope reactivity through entire thickness. It is likely histological processing introduced some errors in quantitative estimates, however, due to factors such as tissue swelling. In preliminary work, tissue wet weight was compared immediately after tissue dissection to that after tissue processing and found to increase by 12% on average. The precise contributions of this swelling artifact to quantitative measures of cell and PCM geometry are difficult to assess as the measurement methods rely on these processing methods. While type VI collagen was chosen as the marker to identify the PCM border based on reports in articular cartilage (Youn et al. 2006), many other macromolecules exist in the PCM region (see Introduction) that could be visualized by immunolabeling allowing for identification of other morphometric features of the PCM. Other candidate markers could include collagen type IX, fibronectin, and laminins that have been shown to be abundant in the PCM regions of multiple cell types. Studies that co-label several of these candidate molecules, or compare geometric measures determined via two different labeling methods, would be instructive for evaluating the significance of type VI collagen in defining the PCM in the IVD and other tissues. In additional, other visualization methods such as electron microscopy can be used to reveal more detailed structure and geometry in PCM (Gruber & Hanley, 2002; Akhtar et al. 2005).

Previous computational models of cell-matrix interactions have suggested that cell aspect ratio (i.e., elongation) and orientation relative to collagen lamella are important determinants of the micromechanical stimuli experienced by IVD cells (Baer & Setton, 2000; Baer et al. 2003). However, these studies did not incorporate a PCM region, which has since been demonstrated to affect predicted cell strain, fluid flow and fluid pressure profiles for chondrocytes (Guilak, 2000; Alexopoulos et al. 2005a). The values for aspect ratio of the PCM and cells in the NP and AF regions measured here were higher than 1, indicating that the PCM regions and cell were fully three-dimensional ellipsoids rather than axisymmetric spheroids. This suggests that a fully 3D FEM analysis incorporated with 3D geometry, as well as material properties and degree of anisotropy of the PCM, will be required to accurately assess the micromechanical environment for IVD cells. In this regard, quantitative 3D geometric measurements of the PCM obtained with this approach are an important first step towards understanding a role for the PCM in governing cell micromechanics and mechanobiology in the IVD.

Supplementary Materials

The following supplementary material is available for this article:

Figure S1

The volume of PCM associated with a single cell in the rat IVD at 12-month old. *P < 0.05, significant difference detected between the NP and outer AF (ANOVA). **P < 0.0001, NP significantly different from corresponding CMU subgroups in the inner and outer AF (ANOVA and post hoc Tukey test).

Figure S2

The size and shape of CMU in the rat IVD at 12-month old. (left) Three diameters from fitted ellipsoids of 3D reconstructed images of CMUs illustrating size of the PCM in the NP (upper) and AF (lower) regions. (right) The aspect ratio between the diameters (D3/D1 upper and D2/D1 lower) for all CMUs. Note that the magnitudes of the axis in these two panels are different. *P < 0.05, ^P < 0.01, significantly different (ANOVA).

This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/j.1469-7580.2007.00784.x

(This link will take you to the article abstract).

Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.