A novel mouse model of intervertebral disc degeneration shows altered cell fate and matrix homeostasis

Abstract

Intervertebral disc degeneration and associated low back and neck pain is a ubiquitous health condition that affects millions of people world-wide, and causes high incidence of disability and enormous medical/societal costs. However, lack of appropriate small animal models with spontaneous disease onset has impeded our ability to understand the pathogenetic mechanisms that characterize and drive the degenerative process. We report, for the first time, early onset spontaneous disc degeneration in SM/J mice known for their poor regenerative capacities compared to “super-healer” LG/J mice. In SM/J mice, degenerative process was marked by decreased nucleus pulposus (NP) cellularity and changes in matrix composition at P7, 4, and 8 weeks with increased severity by 17 weeks. Distinctions between NP and annulus fibrosus (AF) or endplate cartilage were lost, and NP and AF of SM/J mice showed higher histological grades. There was increased NP cell death in SM/J mice with decreased phenotypic marker expression. Polarized microscopy and FTIR spectroscopy demonstrated replacement of glycosaminoglycan-rich NP matrix with collagenous fibrous tissue. The levels of ARGxx were increased in, indicating higher aggrecan turnover. Furthermore, an aberrant expression of collagen X and MMP13 was observed in the NP of SM/J mice, along with elevated expression of Col10a1, Ctgf, and Runx2, markers of chondrocyte hypertrophy. Likewise, expression of Enpp1 as well as Alpl was higher, suggesting NP cells of SM/J mice promote dystrophic mineralization. There was also a decrease in several pathways necessary for NP cell survival and function including Wnt and VEGF signaling. Importantly, SM/J discs were stiffer, had decreased height, and poor vertebral bone quality, suggesting compromised motion segment mechanical functionality. Taken together, our results clearly demonstrate that SM/J mouse strain recapitulates many salient features of human disc degeneration, and serves as a novel small animal model.

Introduction

Low back pain is a ubiquitous health condition that affects millions of people world-wide with a life-time prevalence as high as 84% of the population [1]. Recent studies that measured the burden of diseases in the US and global population ranked low back pain as the 1^st, and neck pain as the 4^th condition for years lived with disability [2]. The medical and societal costs of treating this complex pathology was estimated to be $85.9-624.8 billion [2]. Chronic low back pain also has a strong correlation with psychiatric disorders such as depression [3]. In addition, opioids were the most commonly prescribed class of drug for treating low back pain, according to insurance claims data from 2004 to 2006, despite the lack of their long-term efficacy [4]. With increasing drug abuse and overdose deaths associated with opioid analgesics, it is critical to tackle the chronic low back/neck pain as a major health problem [5].

Importantly, intervertebral disc degeneration is the major risk factor associated with low back pain [6]. The intervertebral disc is composed of inner-most, glycosaminoglycan (GAG)-rich, notochord-derived, nucleus pulposus (NP), outer concentric layers of sclerotome-derived, fibrocartilaginous annulus fibrosus (AF), and superior and inferior, cartilaginous end-plates. The most abundant proteoglycan in NP is aggrecan, but other proteoglycans including versican, biglycan, decorin, and fibromodulin are expressed at low levels [7–14]. Elevated levels of aggregating proteoglycans in turn keeps water within the NP and give the tissue its distinct water-binding ability. Intervertebral disc degeneration is often chronic with multiple etiologies driving complex pathological processes. With aging and degeneration, the NP loses its water binding matrix, becomes fibrotic, and its ability to absorb mechanical loads is compromised. This structural degeneration is accompanied by phenotypic changes in cells [7].

There is an incomplete understanding of the pathogenesis of the disease process partially due to lack of an appropriate animal model. Various large and small animal models that involve genetic manipulations or traumatic injuries have been used [15,16]. However, injury-induced disc degeneration is acute and predominantly driven by inflammatory processes, thus differing from most human cases of disc degeneration [17]. Likewise, genetic modification only allows investigating roles of specific genes or proteins in the disease process, which does not capture multifactorial and possibly polygenic etiology of the human condition. Nonetheless, a few animal models such as chondrodystrophic dogs, and sand rats (Psammomys obesus), a desert gerbil, have been reported to show early-onset spontaneous disc degeneration [18,19]. However, several issues, including lack of tools for molecular analysis, limited availability, high cost of animals and housing, and time involved in aging, limit their wide-spread use in research. For these reasons, there is a great need for small animal models that are not only economical and convenient, but also a better representative of human disc pathology.

MRL/MpJ has been considered as a “super healer” mouse strain due to its high cartilage tissue regenerative capacity in response to full-thickness ear wound [20]. LG/J, one of the parent strains of MRL/MpJ, has also shown a remarkable ability to heal ear as well as articular cartilage injury [21]. On the other hand, SM/J strain is at the opposite end of the cartilage regeneration spectrum, showing little to no regeneration of ear and articular cartilage injuries [21]. However, despite the known differences in cartilage regenerative capacity, there are no published reports on the cartilage homeostasis in naïve animals. Both SM/J and LG/J are inbred strains of mice that were selected for their small and large body size respectively, however these strains have a similar life-span [22,23]. Genetic studies of these strains as well as their recombinant inbred lines have revealed that the tissue repair phenotypes of ear and articular cartilage is heritable and correlated to each other [21]. Specifically, genes that are associated with DNA repair and Wnt signaling, including Xrcc2, Pcna, Axin2 and Wnt16 were found to have strong correlation with cartilage healing [24]. However, there are no studies that have investigated the differences in disc health, especially in the absence of overt injury in these strains of mice. We hypothesized that SM/J mice with diminished regenerative capacity have compromised intervertebral health. We show for the first time that SM/J mice exhibited spontaneous disc degeneration at a very early age, evidenced by increased NP cell death, changes in matrix composition and cell phenotype. Importantly, the altered cell phenotype was underscored by loss of NP cell markers and acquisition of hypertrophic chondrocyte identity. Our results indicate that SM/J mice represent a novel animal model and a valuable tool to study mechanisms of spontaneous disc degeneration.

Results

SM/J mouse intervertebral discs show accelerated structural degeneration

Previous studies that compared connective tissues of LG/J and SM/J mice primarily focused on their regenerative capacities in response to tissue injuries [20,25]. However, injury-driven progression of degeneration differs substantially from age-related degenerative changes. Moreover, there are no reports comparing intervertebral disc phenotype of these strains. We therefore examined the time-dependent changes in disc health in LG/J and SM/J mice in the absence of an overt injury. Safranin O/Fast green staining of 8-week- and 17-week-old mouse intervertebral discs revealed that LG/J mice had healthy discs, characterized by GAG-rich NP with normal cellularity and clear demarcation between NP and AF tissues with intact lamellar AF structure (Figure 1A, B). On the other hand, discs of SM/J mice showed degenerative phenotypes at 8 weeks including moderate cell loss, altered cell distribution and morphology (black arrow heads), and changes in NP matrix composition and integrity (white arrow heads). The level of degeneration became much more severe by 17 weeks (Figure 1A, B). Notably, NP of 17-week-old SM/J mice showed pronounced cell loss, presence of non-vacuolated cells (black arrow heads), and replacement of its bulging, GAG-rich matrix with dense fibrous tissue (white arrow heads). The clear distinction between NP and AF tissue compartments was lost, and some inner AF layers showed inward buckling (black arrow). Tissue demarcation between NP and the endplates was indistinct (black arrow). Interestingly, discs of SM/J mice showed similar levels of degeneration at 4 weeks and signs of milder degeneration at P7 (Supplementary Figure 1A).

Figure 1. SM/J mice show severe intervertebral disc degeneration.

(A, B) Safranin O/Fast Green staining of coronal section of 8- and 17-week-old LG/J and SM/J mouse intervertebral discs showed healthy discs of LG/J, and degenerated discs of SM/J (top row, scale bars=200 μm). High magnification view of NP tissue showed LG/J mice with vacuolated cells and a normal cellularity, while SM/J mice showed loss of vacuolated cells and presence of chondrocyte-like cells (black arrowheads) sparsely distributed within condensed and fibrotic matrix (white arrowheads, second row, scale bars=50 μm). In SM/J mice, tissue boundaries between NP and AF (third row, scale bars=50 μm), and between NP and endplate (bottom row, scale bars=50 μm) were blurred at 17 weeks (black arrow). (C) Histological grading using modified Thompson scale was performed. There was a significant difference between two mouse strains in the proportion of degenerated NP at both age groups, and in the proportion of degenerated AF at 17 weeks. Data were collected from 3 discs per mouse (n=6 mice per group). χ² test. NS=not significant; ****, p≤0.0001.

We scored histological changes based on the modified Thompson scale (Supplementary Table 1). The grade distributions for NP and AF between LG/J and SM/J mice were significantly different (Figure 1C, Supplementary Figure 1B). Importantly, from P7 onwards, greater proportion of SM/J discs had NP with higher grades, indicating an increased incidence of disc degeneration. Similarly, more SM/J discs showed higher AF grades at P7, 4, and 17 weeks. Level-by-level average Thompson grades were also quantified (Figure 2A, B). At 17 weeks, the average grades of NP in SM/J mice were significantly higher than that of LG/J mice for all three caudal levels (C6-C9) (Figure 2A). Similarly, AF of SM/J at both C7/8 and C8/9 had higher average grades (Figure 2B). Furthermore, both at P7 and 4 weeks, the average grades of NP and AF of SM/J mice were significantly higher compared to LG/J mice (Supplementary Figure 1C). Interestingly, the width-to-height aspect ratio of the NP compartment differed in 8-week-old SM/J mice, suggesting that SM/J mice had altered NP shape (Figure 2C). Taken together, these results clearly indicated that intervertebral discs of SM/J mice spontaneously degenerate at an early age.

Figure 2. SM/J mouse intervertebral discs have higher average Thompson grades.

(A, B) Average modified Thompson score of NP and AF for each caudal disc level was shown as scatter plots. All three levels of NP and two levels of AF in SM/J discs at 17 weeks had significantly higher average Thompson grades, indicating increased severity of degeneration. There was no significant difference between these mice at 8 weeks. (C) Measurements of NP aspect ratio, determined by width divided by height of the tissue, showed significantly lower ratio at 8 weeks. Data were collected from 3 discs per mouse (n=6 mice per group). t-test. NS=not significant; *, p≤0.05; **, p≤0.01; ****, p≤0.0001.

NP tissue of SM/J mice show increased apoptosis and loss of cellular phenotype

We then investigated if decreased cellularity in degenerating NP of SM/J mice was due to increased apoptosis. Use of the TUNEL assay clearly demonstrated increased cell death in the NP compartment of 8-week-old SM/J compared to LG/J mice (Figure 3A). At 17 weeks, this increase in TUNEL positive cells was not as pronounced as 8 weeks (Figure 3A). Since one of the hallmarks of disc degeneration is the change in cellular phenotype, we determined if tissue degeneration of SM/J mouse NP was accompanied by cellular phenotypic alterations. Immunofluorescence staining showed that expression of NP phenotypic markers, CA3, GLUT-1 (Slc2a1), and KRT19 [26], was robust and highly specific to the NP compartment (Figure 3B). Specifically, CA3 was expressed at similar levels in both 8-week-old LG/J and SM/J mice. However, GLUT-1 and KRT19 levels were already lower in SM/J mice at this age. Importantly, in the degenerated SM/J discs at 17 weeks, there was loss of expression of all three NP markers. In addition SM/J mice showed lower expression of SDC4, a cell surface proteoglycan that contributes to maintenance of NP phenotype [27], suggesting that these cells are phenotypically different from the original NP cells.

Figure 3. NP cells in SM/J mice undergo apoptosis and lose their notochordal phenotype.

(A) TUNEL assay (green) of LG/J and SM/J intervertebral disc sections showed increased cell death in NP of SM/J mice, especially at 8 weeks (top row; scale bars=300 μm). Corresponding high magnification images of the area within the box are shown (second and fourth column; Bars=150 μm). (B) Immunofluorescence staining of NP markers including CA3, GLUT-1, and KRT19 shows that all three markers were lost in the NP of 17-week-old SM/J mice. GLUT-1 and KRT19 levels were also decreased in SM/J mice at 8 weeks. SDC4 levels were low in SM/J mice at both age groups (scale bars=200 µm). Dotted lines were drawn to demarcate different tissue compartments within the disc. All staining was performed using at least 3 animals per group.

Matrix composition of NP is altered in degenerated SM/J discs

To better characterize the histopathology of degenerated SM/J discs, we investigated changes in NP matrix composition. Total collagen content and fiber orientation was visualized by Picrosirius red staining followed by polarized light microscopy. As expected, the NP matrix of healthy LG/J and SM/J mice showed little to no Picrosirius red staining as well as absence of collagen fibers when observed under polarized light. On the other hand, AF and vertebral bone showed strong staining with highly organized collagen fibers. In the severely degenerated disc of SM/J mice, NP tissue compartments stained intensely with Picrosirius red. When observed under polarized light, these degenerated discs evidenced disorganized collagen fibers within the NP compartments, some of which appeared to be a continuation of the collagen fibers of the buckled inner AF (Figure 4).

Figure 4. NP matrix of SM/J mice have increased collagen content.

Lack of Picrosirius red staining in NP matrix of LG/J mice at both 8 and 17 weeks. SM/J mice showed little to no staining at 8 weeks, while discs of 17-week-old animals showed strong signal, indicating increased collagen content (top row). Collagen fibers visualized under polarized microscope (bottom row) supported drastic upregulation of collagen content in NP compartment in 17-week-old SM/J mice. Scale bars=200 pm. Staining was performed using 6 animals per group.

To obtain further insights into composition of the intervertebral disc matrix in SM/J mice, we used Fourier transform infrared (FTIR) spectroscopy to analyze discs of 17-week-old mice. Cluster analysis of infrared imaging spectra of LG/J and SM/J disc sections revealed anatomic regions qualitatively similar to the corresponding histology images (Figure 5A). Average second derivative spectra from the NP, AF, and CGP clusters were generated for further analyses (Figure 5B). In agreement with increased Picrosirius red staining, collagen content represented by a peak at 1338 cm⁻¹ in the NP of SM/J mice was significantly higher than that of LG/J mice (Figure 5C). On the other hand, similar levels of collagen were observed in the AF of LG/J and SM/J mice (Figure 5C). When the inner and outer AF were analyzed separately, both compartments showed no significant difference in their collagen content (Figure 5D). In addition, there was a substantial decrease in proteoglycan levels represented by a peak at 1156 cm⁻¹ in the NP of SM/J mice, while no changes were seen in the AF compartment (Figure 5E). Both inner and outer AF showed similar levels of proteoglycan between SM/J and LG/J mice (Figure 5F). Total protein content represented by a peak at 1660 cm⁻¹ was significantly higher in the NP of SM/J mice (Figure 5G). In contrast, AF of SM/J mice had slightly lower total protein content, and this difference was more pronounced in outer than inner AF (Figure 5G, H). It should be noted that during the cluster analysis, NP and inner AF of some SM/J mice were automatically grouped as same cluster. This suggested that the compositional distinction between these tissue compartments was reduced in SM/J animals. In addition, the levels of proteoglycan and total protein contents between the NP and AF of SM/J mice were more or less comparable (Figure 5E, G), indicating that the NP matrix composition was becoming similar to that of AF. Taken together, these results indicated that the NP matrix underwent significant transformation in the degenerated discs of SM/J mice.

Figure 5. FTIR analysis show altered disc matrix composition in SM/J mice.

(A) Safranin O/Fast Green staining of representative LG/J and SM/J discs at 17 weeks (left column), and corresponding spectral cluster analysis images (right column). Scale bars= 200 µm. Cluster analysis of infrared imaging spectra of the discs showed anatomic regions qualitatively similar to the histology images. (B) A representative of average inverted second derivative spectra generated from NP, AF, and CGP clusters of LG/J and SM/J mice. Mean absorbance values of peak at 1660 cm⁻¹, 1338 cm⁻¹, and 1156 cm⁻¹ were used form total protein, collagen, and proteoglycan content respectively. (C, D) Average second derivative peak values for collagen content in different tissue compartments were represented with scatter plots. NP of SM/J mice had a significant increase in collagen content, while no difference was seen in AF, inner AF (IAF), and outer AF (OAF) between LG/J and SM/J mice. (E, F). Average second derivative peak values for proteoglycan content were significantly lower in NP of SM/J mice. No difference was seen in AF, IAF, or OAF between the two mouse strains. (G, H) Total protein content in NP of SM/J mice was significantly higher, while that in AF was statistically significantly but minimally lower. Data for all FTIR analysis were collected from n=6 mice per strain with 3 sections/disc/animal. Scatter plots shown as mean ± SEM. t-test. NS=not significant; *, p≤0.05; ***, p≤0.001; ****, p≤0.0001.

We then investigated the expression of different matrix proteins in SM/J discs. Despite the decrease in proteoglycan levels seen in FTIR analysis, NP of both strains at 8 and 17 weeks showed high expression of aggrecan (Figure 6A). However, there was a prominent increase of aggrecan ARGxx neoepitope in the NP of 17-week-old SM/J mice (Figure 6B). Interestingly, both collagen I and II immunostaining showed no difference in the level and pattern of expression between LG/J and SM/J mice (Figure 6C, D).

Figure 6. Aggrecan degradation is increased in the NP of SM/J mice.

(A-D) Immunofluorescence staining of aggrecan, ARGxx, collagen I, and collagen II (red). (A) Aggrecan showed strong expression in NP and some expression in inner AF in both LG/J and SM/J mice at both age groups. (B) ARGxx, aggrecan neoepitope generated by ADAMTS-dependent degradation, was significantly increased in SM/J mice at 17 weeks. (C, D) At both ages, collagen I and II were mainly expressed in AF of both mouse strains. There was no discernable difference in the expression of these proteins between SM/J and LG/J mice. Dotted lines were drawn to demarcate different tissue compartments within the disc. Scale bar=200 µm. All staining was performed using at least 3 animals per group.

Cells in the NP compartment of SM/J show evidence of hypertrophy

While collagen I and II levels were similar in NP of SM/J and LG/J mice, there was a dramatic increase in collagen X expression in the NP compartment of SM/J mice at 17 weeks (Figure 7A). The induction of collagen X expression and age-dependent changes in cellular morphology as well as decreased expression of NP markers suggested that the cells in the NP compartment of SM/J mice were undergoing profound phenotypic changes. Indeed, immunofluorescence staining showed that MMP13, a classical marker of hypertrophy, was upregulated in the NP as well as AF compartments of SM/J mice at 17 weeks (Figure 7B).

Figure 7. NP of SM/J mice have increased collagen X and MMP13 expression.

(A) Immunofluorescence staining of collagen X showed minimal expression in NP and AF of both LG/J and SM/J at 8 weeks (left column; scale bar =300 µm). Hypertrophic zone of growth plate stained strongly for collagen X. At 17 weeks, there was a dramatic increase in collagen X within the NP tissue of SM/J mice (middle column; scale bar=300 µm). Images in the third column show the area within the box from the middle column in higher magnification (Scale bar =150 pm). (B) Immunofluorescence staining of LG/J and SM/J discs showed increased expression of MMP13, a marker of hypertrophic chondrocytes, in the NP and AF compartments of 17-week-old SM/J mice (Scale bar=200 µm). Dotted lines were drawn to demarcate different tissue compartments within the disc. All staining was performed using at least 3 animals per group.

To gain insight into the underlying mechanism driving the altered phenotype, we investigated changes in gene expression that preceded structural degeneration using tissue RNA isolated from the NP of 7-week-old mice. SM/J mice expressed significantly higher levels of Col10a1, although this increase was not apparent by immunostaining at 8 weeks (Figure 8A). While there was only a trend of increase in Mmp13 expression, a significantly higher proportion of SM/J mice showed the presence of Mmp13 (Figure 8B). Ctgf and Runx2, two other markers of hypertrophic chondrocytes, were also upregulated in SM/J mice, strongly suggesting that the molecular phenotype of NP cells was altered at 7 weeks (Figure 8C, D). Despite the increased expression of several genes linked to chondrocyte hypertrophy, there were no significant differences in Adamts4, Fmod, and Il6 expression (Figure 8E-G). Interestingly, expression of Col11a1, which has been shown to be highly expressed in cartilaginous tissues [28], was significantly higher in SM/J mice (Figure 8H). Levels of Vegfa, a pro-survival factor for NP cells [29], were significantly lower in SM/J mice (Figure 8I). In contrast to cartilage [24], NP of SM/J mice had lower expression of Cdkn1a, a cell-cycle inhibitor (Figure 8J). In addition, SM/J mice expressed significantly less Axin2, a target of Wnt signaling pathway (Figure 8K). On the other hand, Levels of Bmp2 and its target, Msx2, were similar between SM/J and LG/J mice (Figure 8L, M).

Figure 8. NP cells of SM/J mice show evidence of hypertrophy.

(A-Q) Real time RT-PCR analysis of NP tissue from 7-week-old LG/J and SM/J mice represented by scatter plots (mean ± SEM). (A-G) mRNA expression of markers of chondrocyte hypertrophy, including Col10a1, Ctgf, and Runx2, were significantly upregulated in SM/J mice. Although Mmp13 levels were not different, significantly higher percentage of SM/J mice expressed Mmp13. (E-G) Levels of Adamts4, Fmod, and Il6 were not different between LG/J and SM/J mice. (H) Expression of Col11a1 was significantly higher in SM/J mice. (I, J) Levels of Vegfa and Cdkn1a, pro-survival factors, were lower in SM/J mice. (K) Expression of a Wnt/B-catenin signaling target, Axin2, was lower in SM/J mice. (L, M) Expression of Bmp2, and its target, Msx2, was not different between LG/J and SM/J mice. (N, O) Levels of Ankh as well as Spp1 that are involved in modulating mineralization were not different between two mouse strains. (P, Q) Enpp1 level and the number of discs that expressed Alpl were significantly higher in SM/J mice. n=10 samples per group (10 mice for LG/J and 15 mice for SM/J). Significance of proportionality was tested using χ² test (Mmp13, Adamts4, and Alpl). T test for normally distributed data, Mann-Whitney test for non-normally distributed data, Shapiro-Wilk normality test was done to check the distribution. NS=not significant; *, p≤0.05; **, p≤0.01.

Since a major function of hypertrophic chondrocytes is to produce proteins involved in controlling matrix mineralization, we measured levels of Ankh, a pyrophosphate (PPi) transmembrane channel, and Spp1/Opn, and did not see any difference (Figure 8N, O). However, levels of Enpp1, an enzyme that generates PPi, and the percentage of discs expressing Alpl, was significantly higher in SM/J mice (Figure 8P, Q). Taken together, these results clearly demonstrated that NP cells of SM/J mice have altered gene expression that is indicative of hypertrophic phenotype.

SM/J discs have altered biomechanical properties

With degradation of matrix molecules and changes in matrix composition, degenerated discs have compromised mechanical function [30,31]. Using an electromechanical testing system, 20 cycles of compression-tension as well as continuous compression (creep test) force was applied to mouse caudal bone-disc-bone motion segments to calculate compressive stiffness and range of motion (ROM), and neutral zone (NZ) stiffness and ROM (Figure 9A). The discs of SM/J mice had significantly higher compressive stiffness (Figure 9B), without changes in compressive ROM (Figure 9C). SM/J discs also showed substantially higher NZ stiffness (Figure 9D), lower NZ ROM (Figure 9E), and lower creep displacement (Figure 9F). These results clearly demonstrated that SM/J mouse intervertebral discs were stiffer.

Figure 9. Discs of SM/J mice have compromised biomechanical properties.

(A) An example of a loading curve extrapolated at the final 20^th cycle was redrawn so that the tension and compression curves were made to intersect before various parameters were measured. (B, C) Compressive stiffness was higher in SM/J mice while compressive ROM was not significantly different. (D, E) SM/J mice had significantly higher NZ stiffness and lower NZ ROM. (F) Creep displacement was significantly lower in SM/J mice. All of the parameters indicated that SM/J discs were stiffer than that of LG/J. Data are represented as scatter plots (mean ± SEM). N=1-3 motion segments for 5 animals per group. NZ: neutral zone; ROM: range of motion; t-test. NS=not significant; *, p≤0.05; **, p≤0.01.

Vertebral bone characteristics of SM/J mice is compromised

SM/J mice exhibit relatively low bone mineral density compared to other inbred strains such as NZB/B1NJ [32,33], however, detailed comparison of vertebral bone parameters with LG/J mice has not been reported. In order to characterize and compare the vertebral bone parameters in LG/J and SM/J mice, we performed micro-computed tomography (µCT) analysis on three vertebrae (Ca5-7) and intervertebral discs (Ca5/6-Ca7/8) (Figure 10A). SM/J disc height was significantly lower than that of LG/J at both 8 weeks and 17 weeks of age (Figure 10B). However, as predicted by their smaller body size, the vertebral body of SM/J was shorter than that of age-matched LG/J mice (Figure 10C). Consequently, the disc height index (DHI) was higher in SM/J mice at 8 weeks; this difference was no longer significant at 17 weeks probably due to growth of the vertebrae and concomitant decrease in disc height as a result of degeneration (Figure 10D). In addition, SM/J mice showed significantly lower bone volume fraction (BV/TV) and trabecular number (Tb. N.), and higher trabecular spacing (Tb. Sp.) (Figure 10E-G). While trabecular thickness (Tb. Th.) was not different between LG/J and SM/J mice (Figure 10H), connective density (Conn. Dens.) was significantly lower, and structural model index (SMI) was significantly higher in SM/J mice (Figure 10I, J). Taken together, the µCT analysis confirmed the smaller bone size of SM/J mice, and also suggested that SM/J vertebrae are inferior in quality due to sparse distribution of trabecular bone.

Figure 10. SM/J mice demonstrate poor vertebral bone health.

(A) Representative µCT scans of caudal motion segment (Ca6-7) of LG/J and SM/J at 17 weeks, showing a coronal cut-plane through 3D reconstructions. Scale bars=1 mm. (B-D) Disc height, vertebral length, and disc height index (DHI) measurements of LG/J and SM/J at both ages were shown as scatter plots (mean ± SEM). SM/J mice had significantly smaller disc height as well as vertebral length, resulting in higher DHI at 8 weeks. (E-J) Measurements of various vertebral trabecular bone parameters demonstrated that, compared to LG/J, SM/J mice had lower bone volume fraction (BV/TV), trabecular number (Tb. N.), and connectivity density (Conn. Dens.), but higher trabecular separation (Tb. Sp.), and structure model index (SMI), suggesting that SM/J mice had compromised vertebral bone quality. Trabecular thickness (Tb. Th.) was not significantly different between two mouse strains. All analyses were done with n=6 mice per group with 3 consecutive vertebrae/animal. Scatter plots shown as mean ± SEM. t-test. NS=not significant; *, p≤0.05; **, p≤0.01; ***, p≤0.001; ****, p≤0.0001.

Discussion

Research on chronic disorders such as intervertebral disc degeneration and associated neck and back pain can greatly benefit from access to small animal models that accurately capture aspects of human disease progression. However, there has been no readily available small animal model that shows disc degeneration at an early age without genetic, dietary, or surgical manipulation. Due to their differences in many phenotypic traits, SM/J and LG/J mice have been used for a number of genetic studies [34-36]. While, previous studies investigated the injury related cartilage pathologies in these mice, in this study, we focused on differences in intervertebral disc health between these mouse strains without an overt injury. Our study clearly showed that SM/J mice develop an early disc degeneration characterized by extracellular matrix fibrosis, increased cell death, and changes in NP cell phenotype. Our results indicate that SM/J mice are valuable research tools to study pathogenetic mechanisms of disc degeneration.

Histopathological grading confirmed that SM/J mice show signs of degeneration as early as P7, which become prominent by 17 weeks. Previous reports on age-dependent disc degeneration in C57BL/6 mice showed that degeneration occur at 24 months [37]. At 17 weeks, SM/J mice have just attained skeletal maturity, suggesting that the disc degeneration is not due to aging. Although one of the risk factors for disc degeneration is age, degenerated discs are also observed in young population, implying that non-age-related factors are important in the disease pathogenesis [38,39]. Indeed, in a retrospective cohort study of twin population, Battié and colleagues reported that the genetics rather than aging was the most influential factor contributing to disc degeneration [40], highlighting the utility of using SM/J mice for studies of disc pathology. Interestingly, at 8 weeks, the NP of SM/J mice had altered width-to-height aspect ratio compared to LG/J, indicating difference in tissue shape. Changes in disc shape have been correlated with human disc degeneration [41]. The altered NP shape is also thought to compromise the normal tissue mechanics as well as the loads that cells experience [41,42], suggesting that this factor may provoke accelerated disc pathology in SM/J mice.

Increased collagen and total protein content, as well as decreased proteoglycan in SM/J mice measured by Picrosirius red and FTIR analyses recapitulated the changes seen in human degeneration [7,43]. The grouping of NP and inner AF into the same cluster following FTIR analysis suggested blurred distinction between these compartments in SM/J mice. Likewise, similar levels of proteoglycan and total protein between NP and AF of SM/J mice indicated that the matrix composition of NP was closer to fibrocartilage, which has inferior load-dampening capacity compared to healthy NP [44]. However, surprising lack of altered expression and localization of collagens I and II suggested that these collagens did not contribute to increased collagen content of the SM/J NP. Instead, elevated levels of collagen X, known to be expressed in degenerated human discs [45], and col11a1 were the likely source of collagen accumulation measured by Picrosirius red and FTIR analyses. While aggrecan immunostaining was not overtly different between the strains, a significant increase in the aggrecan neoepitope, ARGxx, clearly highlighted ADAMTS-mediated degradation of aggrecan in SM/J NP. It is also plausible that proteoglycans in SM/J mice are under-glycosylated, or the hyaluronan content is altered, both of which are risk factors for disc degeneration [46-49].

NP of SM/J mice showed increased cell death and loss of phenotypic markers, CA3, GLUT-1 (Slc2a1), and KRT19, suggesting that the surviving cells have lost their molecular identity and transdifferentiated, and/or they were not of NP-origin. Both in vitro, and ex vivo organ culture studies have shown that large notochordal NP cells can differentiate into cells with various morphologies [50,51]. In addition, lower expression of SDC4, a heparan-sulphate proteoglycan, in SM/J would upregulate levels of SOX9, a master chondrogenic transcription factor, as well as affect Shh signaling, further contributing to loss of notochordal NP phenotype [27,52,53]. Decreased Axin2 also indicated diminished Shh-dependent canonical Wnt signaling, important for preservation of NP phenotype [53,54]. The phenotypic similarity of SM/J NP cells to hypertrophic chondrocytes was demonstrated by a robust expression of collagen X and MMP13 mRNA and protein, also seen during human disc degeneration [55,56]. Similarly, increased Runx2 and Ctgf in NP of SM/J supported the notion of altered cell phenotype. Noteworthy, while Ctgf expressed by healthy NP cells is critical in maintenance of matrix homeostasis, its level is elevated during human disc degeneration [57,58]. Interestingly, although the NP cells of SM/J resembled hypertrophic chondrocytes, they did not upregulate Fmod and Il6. Furthermore, high levels of Col11a1 suggested some characteristics of AF and endplate [28].

Vegfa is an important cell survival factor for both chondrocytes and NP cells [29,59]. Similarly, Cdkn1a regulates the cell cycle and regeneration, as well as plays a crucial role in blocking apoptosis [60–62]. From this perspective, the increase in NP cell death in SM/J mice could be due to decreased Vegfa and Cdkn1a. Moreover, it is possible that low GLUT-1 expression in SM/J subjected the cells to metabolic constraint and sensitized them to cell death as glycolysis is the main pathway of ATP generation in NP cells [63]. The support for the notion that NP cells are replaced with cells of different origin also comes from studies by Merceron et al. and Yang et al. [44,64]. Using a lineage tracing approach, we have shown that cells from the neighboring tissues populate the NP compartment following loss of the majority of NP cells [44]. Likewise, Yang et al., showed that the decrease in the number of notochordal NP cells in response to needle injury was followed by the appearance of chondrocytic cells [64]. The occurrence of dystrophic mineralization of intervertebral discs increases with aging and degeneration [65,66]. Both ANK and ENPP1 modulate levels of extracellular PPi, a potent inhibitor of hydroxyapatite crystal formation [67]. However, increased Enpp1 in combination with increased Alpl, an enzyme that utilizes PPi to generate Pi, suggest that hydroxyapatite crystal formation is promoted in SM/J.

The quality of NP matrix is extremely important for its proper biomechanical function. Increased collagen and decreased proteoglycan content within the NP, in combination with increased compressive and NZ stiffness, and decreased NZ ROM and creep displacement indicated stiffer and fibrocartilaginous property of the SM/J mouse discs. In addition, our µCT analysis of vertebral bone parameters of SM/J mice showed a positive correlation between bone and disc qualities, suggesting that the degenerative changes in the disc were not secondary to the changes in vertebral bone health. Our data thus underscore that with degeneration, SM/J mouse spinal motion segments become stiffer as suggested by a recent clinical report [31].

In summary, our study clearly demonstrates that SM/J mice show classical hallmarks of human disc degeneration that include NP cell loss, changes in NP cell phenotype, decreased proteoglycan and increased collagen content, and compromised biomechanical properties [68,69]. As such, this makes the SM/J mice an ideal experimental animal model to study pathogenetic mechanisms underlying disc degeneration.

Materials and Methods

Mice

All procedures regarding housing, breeding, and collection of animal tissues was performed as per approved protocols by Institutional Animal Care and Use Committee (IACUC) of the Thomas Jefferson University, in accordance with the IACUC’s relevant guidelines and regulations. Both SM/J and LG/J mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Animals were socially housed under barrier conditions using aseptic techniques. Mice were provided with Lab Diet 5010 Laboratory Autoclavable Rodent Diet ad libitum. Both male and female mice at ages between 7-days and 17-weeks were analyzed. These time points correspond to actively proliferating (P7) and stabilized (4 week) cell population in the NP compartment, early stage of matrix accumulation in NP (8 week) and skeletal maturity (17 week) of the animals.

Histological analysis

Caudal spine tissues from LG/J and SM/J mice at various ages were dissected and immediately fixed in 4% PFA in PBS at 4 °C for 48 h, decalcified with 20% EDTA at 4 °C for 15 days, and then embedded in paraffin. Coronal sections of 7 μm thickness were cut. Intervertebral disc tissue sections (Ca 6-9) were stained with Safranin O/Fast Green/Hematoxylin or Picrosirius red, then visualized using a light microscope (Axio Imager 2, Carl Zeiss) or a polarizing microscope (Eclipse LV100 POL, Nikon). Imaging of Safranin O/Fast Green stained tissues were performed using 5x/0.15 N-Achroplan (Carl Zeiss) or 10x/0,3 EC Plan-Neofluar (Carl Zeiss) objectives, Axiocam 105 color camera (Carl Zeiss), and Zen2™ software (Carl Zeiss). For Picrosirius red stained tissues, 10x/0.25 Pol/WD 7.0 objective (Nikon), Digital Sight DS-Fi2 camera (Nikon), and NIS Elements Viewer software (Nikon) were used. To evaluate degeneration of IVD, mid-coronal sections from three caudal disc levels per mouse were scored using a modified Thompson grading scale (Supplementary Table 1) by 6 blinded observers [70]. Histopathological scores were collected from n=6 mice per group with 3 discs per mouse (total 18 discs per group).

Width-to-height aspect ratio measurement

Aspect ratio of NP was determined by width divided by height of the NP tissue measured from Safranin O/Fast Green staining images of coronal tissue sections using ImageJ software (http://rsb.info.nih.gov/ij/). Data were collected from 6 mice per group with 3 discs per mouse.

TUNEL assay

TUNEL assay was performed on disc tissue sections using “In situ cell death detection” Kit (Roche Diagnostic). Briefly, sections were de-paraffinized and permeabilized with Proteinase K (20 μg/mL) for 15 min at room temperature. Then TUNEL assay was carried out per manufacturer’s protocol. The sections were washed and mounted with ProLong® Gold Antifade Mountant with DAPI (Thermo Fisher Scientific, P36934), and were visualized with Axio Imager 2 (Carl Zeiss) using 5x/0.15 N-Achroplan (Carl Zeiss) or 10x/0,3 EC Plan-Neofluar (Carl Zeiss) objectives, AxioCam MRm camera (Carl Zeiss), and Zen2™ software (Carl Zeiss).

Immunofluorescence microscopy

Coronal disc tissue sections of 7 μm thickness were de-paraffinized and incubated either in microwaved citrate buffer for 20 min, with proteinase K for 10 min at room temperature, or with Chondroitinase ABC for 30 min at 37 °C for antigen retrieval. Then the sections were blocked in 5% normal serum (Thermo Fisher Scientific, 10000C) in PBS-T (0.4% Triton X-100 in PBS), and incubated with antibody against Aggrecan (1:50, Millipore, AB1031), Collagen I (1:100, Abcam, ab34710), Collagen II (1:400, Fitzgerald, 70R-CR008), Collagen X (1:500, Abcam, ab58632), CA3 (1:150, Santa Cruz, sc-50715), KRT19 (1:3, DSHB, TROMA-III/supernatant), Syndecan-4 (1:100, Millipore, ABT157), or MMP13 (1:200, Abcam, ab39012) in blocking buffer at 4 °C overnight. For GLUT-1 (1:200, Abcam, ab40084) and ARGxx (1:200, Abcam, ab3773) staining, Mouse on Mouse Kit (Vector laboratories, BMK-2202) was used for blocking and primary antibody incubation. Tissue sections were thoroughly washed and incubated with Alexa Fluor®-594 (Ex: 591 nm, Em: 614 nm) conjugated secondary antibody (Jackson ImmunoResearch Lab, Inc.), at a dilution of 1:700 for 1 h at room temperature in dark. The sections were washed again with PBS-T (0.4% Triton X-100 in PBS) and mounted with ProLong® Gold Antifade Mountant with DAPI (Thermo Fisher Scientific, P36934). All mounted slides were visualized with Axio Imager 2 (Carl Zeiss) using 5x/0.15 N-Achroplan (Carl Zeiss) or 10x/0,3 EC Plan-Neofluar (Carl Zeiss) objectives, X-Cite® 120Q Excitation Light Source (Excelitas), AxioCam MRm camera (Carl Zeiss), and Zen2™ software (Carl Zeiss).

Fourier transform infrared (FTIR) imaging spectroscopy data collection and processing

Infrared (IR) spectral imaging data were acquired from 5 μm-thick sections of decalcified mouse caudal disc tissues collected from 17-week-old LG/J and SM/J mice (n=6 mice per strain). Data were acquired in the mid-IR region, 4000-800 cm⁻¹ at 8 cm⁻¹ spectral resolution and 25 µm spatial resolution using a Spectrum Spotlight 400 FT-IR Imaging system (Perkin Elmer, Shelton, CT). Data collection time ranged from approximately 20 - 40 min, depending on the size of the tissue sample. Absorbances from collagen and proteoglycan dominated the spectra with peaks at multiple frequencies [71]. An extended multiplicative scatter correction (EMSC) was applied to the spectra to remove the scattering and offset effects [72]. Second derivative differentiation with a 9-point Savitzky–Golay smoothing was applied to spectral data to resolve peaks that underlie broad absorbances. The EMSC corrected second derivative spectra were multiplied by negative 1 to convert to positive peaks for visualization and quantification. Absorbances from proteins in the amide I region centered at 1660 cm⁻¹, and in the collagen side chain vibration region at 1338 cm⁻¹ were evaluated, as well as absorbances from the sugar ring of proteoglycans at 1156 cm⁻¹. For quantification, mean absorbance values of peaks of interest were taken across NP, AF, or cartilage growth plate (CGP) region as defined by spectral cluster analysis. Significant differences in parameters were assessed by Student t test with p≤0.05 considered significant.

Spectral Cluster analysis

The preprocessed spectra were used for K-means cluster analysis to define anatomical regions and tissue types within the tissue section spectral images. Cluster analysis of IR spectral imaging data has previously been performed to qualitatively differentiate tissue types located in different anatomical regions [73]. With this method, regions of IR images are separated into two or more classes, or “clusters”, with similar spectral properties. The K-means partitional clustering method starts with a random selection of K objects that are to be used as cluster targets (defined by the operator), where K is determined a priori. During each cycle of this clustering method, the remaining objects (pixels of spectral image) are assigned to one of these clusters based on the distance from each of the K targets. New cluster targets are then calculated as the means of the objects in each cluster, and the procedure is repeated until no objects are re-assigned after the updated mean calculations [74].

Tissue RNA isolation and Real Time RT-PCR Analysis

NP tissues were dissected from lumbar and caudal discs of 7-week-old LG/J and SM/J mice, and immediately placed in TRIzol® Reagent (Life Technologies). 10-15 mice per strain were sacrificed for tissue RNA isolation. For LG/J mice, NP tissues from one mouse were pooled together per one sample (total n=10 samples), and for SM/J mice, due availability of the material, NP tissues from 2 mice were pooled together per one sample for 5 samples and tissue from one mouse were pooled together per one sample for the addtional 5 samples (total n=10 samples). Isolated NP tissues collected in TRIzol® Reagent were homogenized with a Pellet Pestle Motor (Sigma Aldrich, Z359971) for 1 minute. Total RNA was extracted from the tissue lysates using Direct-zol® MiniPrep Plus kit (Zymo Research). The purified, DNA-free RNA was converted to cDNA using EcoDry™ Premix (Clontech). Template cDNA and gene-specific primers were added to Power SYBR Green master mix (Applied Biosystems) and mRNA expression was quantified using the Step One Plus Real-time PCR System (Applied Biosystems). GAPDH was used to normalize gene expression. Melting curves were analyzed to verify the specificity of the RT-PCR and the absence of primer dimer formation. Thermal cycle was programmed for 20 s at 95 °C as initial denaturation, followed by 40 cycles of 3 s at 95 °C, and 30 s at 60 °C, with final melt curve and extension for 15 s at 95 °C, 1 m at 60 °C, and 15 s at 95 °C.

Mechanical testing and data analysis

Mechanical properties of caudal motion segments (C4/5-C8/9) from 17-week-old LG/J and SM/J mice (n ≥ 12 discs/group) were assessed with an electromechanical testing system (Instron, Norwood, MA) using a 5 N load cell. Caudal spines were dissected from mice and stored in PBS-soaked gauze at −80 °C until the day of testing. Before testing, spines were slowly thawed and scanned in micro-CT for accurate measurement of disc height and cross-sectional area. Subsequently, motion segments were prepared by cutting through superior and inferior unused discs and removing extraneous tissue to expose only clean vertebral bodies and intervertebral disc. Samples were secured to the superior, mobile fixture of the Instron with a small clamp, and to the bottom, immobile fixture via a potting technique using polymethylmethacrylate (PMMA). The mechanical testing protocol consisted of 20 cycles of compression/tension from −1.5 to +0.5 N at 0.5 Hz immediately followed by a 1 h creep load of −1.5 N [75]. All testing was conducted in 1X PBS bath at room temperature.

Compression data from the 20th test cycle were analyzed for compressive range of motion (ROM) and stiffness, and neutral zone (NZ) ROM and stiffness by fitting the data to a sigmoid function [76]. The maximum of the 1st derivative of the function represents the transition from tension to compression. Meanwhile, the extremum of the 2nd derivative represent the boundaries of the NZ, or the zone during which the samples experiences a change in extension without exogenous loading. Compressive ROM was defined as the displacement between the inflection point of the curve and the displacement at −1.5 N. Compressive stiffness was defined from the raw data as the slope of the line fit from −1.5 to −0.5 N. NZ ROM was defined as the displacement between the boundaries of the NZ, while NZ stiffness was defined as the slope of the line connecting these points. ROM measurements were normalized by dividing by the disc height, while stiffness measurements were normalized by multiplying with the disc height and dividing by the disc cross-sectional area. Creep displacement was measured from raw displacement of the disc from the beginning to the end of the 1 h test. Creep data were normalized by dividing by the disc height [75].

Micro-CT analysis

Micro-CT (µCT) scans (MicroCT40, SCANCO Medical, Switzerland) were performed on spines of LG/J and SM/J mice fixed with 4% PFA [77]. Caudal spine segments incorporating Ca5-7 were scanned with an energy of 70 kVp, a current of 114 mA, and a 200 ms integration time producing a resolution of 16 mm³ voxel size. 6 mice per strain per age group were used. The cross sectional scans were analyzed to quantify changes in trabecular bone microarchitecture by first drawing a region of interest (ROI) that contoured the outer boundary of the trabecular bone throughout the entire caudal vertebral body, excluding the cortical bone. The ROI were then compiled into 3D data sets using a Gaussian filter (σ = 0.8, support = 1) to reduce noise, and converted to binary images with a fixed grey-scale threshold of 200. The 3D data sets were assessed for the following parameters using software supplied by the system manufacturer: trabecular bone volume (BV), total volume (TV), bone volume fraction (BV/TV), trabecular thickness (Tb. Th), trabecular number (Tb. N), trabecular separation (Tb. Sp), connectivity density (Conn. Dens.), disc height between each vertebra, and vertebral length. Disc height was determined by measuring the ventral, central and dorsal lengths of the discs, and then averaging these values. Disc height index (DHI) was calculated by dividing average disc height by the length of adjacent vertebral bodies. All heights were calculated by multiplying the number of cross sectional slices spanning the structure of interest by the known thickness of each slice.

Statistics

At least 6 animals per strain (3 discs per animal) were used for all quantitative analysis. Quantitative data are presented as the mean ± SEM. Data distribution was checked with Shapiro-Wilk normality test, and Mann-Whitney test was used for non-normally distributed data. Differences between two groups were assessed by t test. Comparisons among more than two groups were done by ANOVA. For histopathological analysis showing percent-degenerated-discs, χ² test was used. All statistical analyses were done using Prism7 (GraphPad Software). P ≤ 0.05 was considered statistically significant.

Supplementary Material

1. Supplementary Figure 1. NP of SM/J mice show signs of degeneration at P7 and 4 weeks.

(A) Safranin O/Fast Green staining of intervertebral disc tissues from LG/J and SM/J mice at P7 and 4 weeks. SM/J mice already started showing signs of mild degeneration at P7 (Top row; scale bar=100 μm). By 4 weeks, the level of degeneration is similar to that of 8 and 17 weeks (Bottom row; scale bar=200 μm). (B) There was a significant difference between two mouse strains in the proportion of degenerated NP and AF at both age groups. χ² test. (C) Average modified Thompson grade was higher in SM/J mice for both NP and AF at P7 and 4 weeks. Scatter plots show mean ± SEM. Data were collected from at least 2 discs per mouse (n=6 mice per group). t-test. NS=not significant; *, p≤0.05; **, p≤0.01; ***, p***<0.001; ****, p≤0.0001.

Highlights.

• SM/J mice show early onset spontaneous intervertebral disc degeneration

• Disc degeneration in SM/J mice is characterized by decreased proteoglycan and increased collagen content

• NP cells of SM/J mice undergo accelerated cell death

• Cells in the NP compartment of SM/J mice show phenotype similar to hypertrophic chondrocytes

• Intervertebral discs of SM/J mice show compromised biomechanical properties