Imbalanced Protein Expression Patterns of Anabolic, Catabolic, Anti-Catabolic and Inflammatory Cytokines in Degenerative Cervical Disc Cells: New Indications for Gene Therapeutic Treatments of Cervical Disc Diseases

Degenerative disc disease (DDD) of the cervical spine is common after middle age and can cause loss of disc height with painful nerve impingement, bone and joint inflammation. Despite the clinical importance of these problems, in current publications the pathology of cervical disc degeneration has been studied merely from a morphologic view point using magnetic resonance imaging (MRI), without addressing the issue of biological treatment approaches. So far a wide range of endogenously expressed bioactive factors in degenerative cervical disc cells has not yet been investigated, despite its importance for gene therapeutic approaches. Although degenerative lumbar disc cells have been targeted by different biological treatment approaches, the quantities of disc cells and the concentrations of gene therapeutic factors used in animal models differ extremely. These indicate lack of experimentally acquired data regarding disc cell proliferation and levels of target proteins. Therefore, we analysed proliferation and endogenous expression levels of anabolic, catabolic, ant-catabolic, inflammatory cytokines and matrix proteins of degenerative cervical disc cells in three-dimensional cultures. Preoperative MRI grading of cervical discs was used, then grade III and IV nucleus pulposus (NP) tissues were isolated from 15 patients, operated due to cervical disc herniation. NP cells were cultured for four weeks with low-glucose in collagen I scaffold. Their proliferation rates were analysed using 3-(4, 5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide. Their protein expression levels of 28 therapeutic targets were analysed using enzyme-linked immunosorbent assay. During progressive grades of degeneration NP cell proliferation rates were similar. Significantly decreased aggrecan and collagen II expressions (P<0.0001) were accompanied by accumulations of selective catabolic and inflammatory cytokines (disintegrin and metalloproteinase with thrombospondin motifs 4 and 5, matrix metalloproteinase 3, interleukin-1β, interleukin-1 receptor) combined with low expression of anti-catabolic factor (metalloproteinase inhibitor 3) (P<0.0001). This study might contribute to inhibit inflammatory catabolism of cervical discs.


Introduction
After middle age many people experience pain symptoms of cervical disc degeneration. Pain symptoms may get worse over time, which are accompanied by loss of disc height, painful nerve impingement, bone and joint inflammation [1][2][3][4][5][6]. The degenerative process can cause radiating pain and loss of mobility that have a major impact on the quality of life. Previous publications have analysed the pathology of cervical disc degeneration only from a morphologic view point using magnetic resonance imaging (MRI), which does not address the issue of biological treatment approaches. Until now the expression patterns of extracellular matrix (ECM) associated proteins in cervical nucleus pulposus cells are not published. Our current work is the first investigation concerning the endogenous expression patterns of ECM-associated proteins in degenerative cervical disc cells. Considerable anatomical differences between cervical and lumbar discs have been previously presented [7][8]. Furthermore, Mechanical properties in cervical discs have shown specific features and demonstrated some differences from lumbar discs [9][10]. The anatomical and mechanical differences might lead to functional modifications in cervical disc cells. These reasons suggest that biomolecular results from lumbar disc cells should not be directly projected onto cervical disc cells without any similar investigations. Numerous investigations have been made in lumbar discs to understand how bioactive factors combine to promote painful disc degeneration [11][12]. However, previous publications have not yet displayed the biomolecular differences or similarities between lumbar and cervical disc. Thus, the data of the current study address for the first time the biomolecular issue of cervical disc degeneration and might contribute valuably to gene therapeutic approaches of painful intervertebral disc degeneration.
Degenerative lumbar intervertebral discs (IVDs) have been targeted by different biological treatment approaches. Nucleus pulposus (NP) cells have been shown to play a central role in the maintenance of lumbar IVDs by organizing the expression of anabolic, catabolic, anti-catabolic and inflammatory cytokines that affect the synthesis and degradation of the IVD matrix. IVD degeneration is shown to be associated with imbalances of these factors combined with the declined cell density in adult IVDs [11][12][13][14][15][16][17][18][19][20][21][22][23]. Nevertheless, the amounts of lumbar NP cells and the concentrations of gene therapeutic factors used for regeneration of IVD tissues in animal models differ extremely [11][12][13][14][15][16]. These exhibit lack of experimentally acquired data regarding proliferation rates of NP cells and their endogenous expression levels of therapeutic target proteins. Recently we have reported about proliferation rates and imbalances of anabolic and catabolic factors regarding adult lumbar NP cells, and suggested potentially useful gene therapeutic targets [24].
So far a wide range of endogenously expressed bioactive factors, which are vital for designing objective gene therapeutic approaches, has not yet been investigated in degenerative cervical disc cells. Thus, we analysed proliferation rates of degenerative cervical NP cells and their endogenous expression levels of therapeutic target proteins in a three-dimensional collagen I scaffold. Since spinal disc herniation in adults predominantly occurs in discs of degeneration grade III and IV, we analysed cervical NP cells from those patients of disc degeneration grade III and IV, operated due to cervical disc herniation. Patient inclusion criteria for surgery were radiographically determined cervical disc herniation with nerve root compression on MRI, which correlated to primary symptoms that remained unresponsive to nonoperative treatment for six weeks or demonstrated progressive neurological deterioration in the face of conservative treatment.
Progressive grade of cervical disc degeneration is significantly associated with accumulation or low expression levels of selective bioactive factors, which might cause unfavourable phenotypic alternations that might impair IVDs regeneration. The results of this study might contribute to design objective gene therapeutic treatment approaches and help to inhibit the inflammatory catabolism of intervertebral disc tissue.

Cervical IVD Specimens
Cervical nucleus pulposus tissues were acquired from patients with informed consents. Patients were operated due to cervical disc herniation. Participants provided their written informed consent to participate in this study. Experimental studies of human cervical IVD specimens were approved by the local research ethics committee (Heidelberg University, University Medical Center Mannheim: project 2009-217N-MA). The Miyazaki MRI scoring system [25] was used to determine the degeneration grades of cervical IVDs. 15 Patients (15 cervical IVDs of degeneration grade III and IV) with a mean age of 55 years (range 35-75 years) were involved (table 1). For isolation of NP cells, residual NP tissues from cervical disc space were recruited and brought immediately to the lab in sterile phosphate buffered saline solution (PBS) (Sigma-Aldrich).

Isolation and Monolayer Culture of Cervical NP Cell
Isolation and monolayer expansion of cervical NP cells were performed as described in our previous publication [24]. Briefly, NP specimens were washed in PBS and carefully separated from AF tissues. NP tissues were then minced into small fragments of approximately 2 mm 3 and sequentially digested with pronase, collagenase II and hyaluronidase. After filtration of the samples through nylon mesh filters (75 gm), supernatants were centrifuged and pellets were suspended in 10 ml Dulbecco's Modified Eagle's Medium (DMEM) containing 1% v/v penicillin/streptomycin, 1% w/v glucose and 10% v/v FCS. By changing the culture medium every two day, NP cells were cultured for 2 weeks in 75 cm 2 tissue culture flask. Monolayer expanded cervical NP cells were then cryopreserved at 2196uC in culture medium containing 30% v/v FCS and 15% v/v dimethyl sulfoxide (DMSO).
For control two-dimensional (2D) culture of cervical NP cells in tissue culture dishes (100620 mm, Sigma-Aldrich), 4610 5 NP cells were seeded in 10 ml DMEM containing 1% v/v penicillin/ streptomycin, 1% w/v glucose and 10% v/v FCS. Cells were cultured for four weeks (37uC, 5% CO2) by changing the culture medium every two days.

Three-dimensional Culture of Cervical NP Cells
The three-dimensional (3D) culture of cervical NP cells in collagen I based cell carrier (CCC) was carried out as previously described [24]. For control 3D culture of cervical NP cells in agarose gels, 6-well plates were coated with a thin layer of 1% agar (Sigma-Aldrich). 2% agarose of high electroendosmosis (Sigma-Aldrich) was autoclaved and equilibrated to 37uC. The agarose was mixed with equal volume of DMEM containing 1% v/v penicillin/streptomycin, 1% w/v glucose and 10% v/v FCS. Then the solution was mixed with one volume part of cell suspension to yield a final density of 4610 5 cells per milliliter and 1 ml was added to the pre-coated 6-well plates. Following gelation (4uC, 10 min) the embedded cells were overlaid with culture medium. Cells were cultured for four weeks (37uC, 5% CO2) by changing the culture medium every two days. The NP cells were then processed for control quantification of collagen I expression.

Isolation of 3D Cultured Cervical NP Cells
Isolation of three-dimensional cultured cervical NP cells from collagen I scaffold was performed as formerly described [24]. To isolate three-dimensional cultured cervical NP cells from control agarose gels, the AgarACE Agarose-Digesting Enzyme (Promega) was used. Each 200 mg of agarose gel slice was transferred to a 1.5 ml micro-centrifuge tube and melted for 10 min at 65uC. After whirling for 2 seconds, the tube was transferred to a 42uC heating block and the gel was digested for 15 min with AgarACE Agarose-Digesting Enzyme (2U). After digestion of the gels, samples were filtered through nylon mesh filter (75 gm), supernatants were centrifuged for 2 min (10006g) and cell pellets were washed twice in PBS for 2 min (10006g). The NP cells were then processed for control quantification of collagen I expression.
Proliferation Assay of Cervical NP Cells As previously described [24] the MTT assay was applied to determine the proliferation of cervical NP cells. Briefly, NP cells were suspended in 0.5 ml culture medium and 100 ml duplicates of cell suspension were plated into flat-bottomed 96 well plates in addition to duplicate of blank control wells of medium alone. After 24 h of incubation, MTT reagent (10 ml) was added to each well. Following 3 h of incubation, 100 ml SDS-HCl solution was added for additional incubation of 4 h. The average absorbance value (570 nm) of the blank duplicate readings was subtracted from the average values of the sample duplicate readings and cell concentration was calculated from the standard curve. Cervical NP cell proliferation data represent the mean of at least three individual experiments.
For the isolation and quantification of target proteins from cervical NP cells, 4610 5 cells were cultured for four weeks in collagen I scaffold, TC dish or agarose gel. NP cell pellets were harvested and pellets were washed twice for 5 min in cold PBS (25006g). Proteins were isolated by using the radio-immunoprecipitation assay (RIPA) buffer as formerly described [24]. Protein concentrations in samples were determined according to the

Statistical Data Analysis
Landis and Koch [25][26][27] based interpretations with k statistics and agreement percentage among two observers (interobserver reliability) were applied to estimate the reliability of the MRI evaluations. Frequency of disagreement was calculated for each degeneration grade. The software IBM SPSS Statistics 20, Armonk New York USA was applied for statistical analysis. 1way ANOVA and pairwise comparisons were used to analyze cell proliferation rates and protein expression levels as a function of degeneration grade, age and gender. Significance in all cases was set at P,0.05.

Interobserver Reliability of MRI-grading
The interobserver agreement was excellent (k = 0.885) and the calculated frequency of agreement was 92.31%.
Degenerative Cervical NP Cell Proliferation Rates 4610 5 degenerative cervical NP cells were cultured in collagen I scaffold for four weeks. Equivalent cell proliferation rates were determined from all samples of degeneration grade III and IV.
The mean values of proliferation rates were 1.086610 6 (613167) and 1.058610 6 (618661) cells for grade III and IV respectively (P,0.0061). The confirmed cell proliferation rates between degeneration grade III and IV differed with about 2.6% (table 2 and figure 1a). Age and gender do not seem to play distinct role in influencing proliferation rate of degenerative cervical NP cells (data not shown).
Endogenous Expression Levels of Catabolic, Anticatabolic and Inflammatory Cytokines in Degenerative Cervical NP Cell 4610 5 degenerative cervical NP cell were cultured for four weeks in collagen I scaffold. High and increasing expression levels of the catabolic factor ADAMTS-4 with mean expression values of 1718663.7 pg/ml and 33086123 pg/ml were verified for degeneration grades III and IV respectively (P,0.0001), which corresponds to a 1.9 fold increase in mean expression values. Moreover, higher but equivalent expression levels of ADAMTS-5 with mean expression values 41366191 pg/ml and 42156160 Table 2. Proliferation rates of degenerative cervical NP cells in 3D culture. The inflammatory cytokines IL-1b, IL-1 R, TNF-a and TNF-a R1 were detected relatively at very low expression levels. In degeneration grade III and IV the expression levels of IL-1b and IL-1 R were higher than that of TNF-a and TNF-a R1. Their respective mean expression values were 11260.728 pg/ml and 12262.55 pg/ml (P,0.0001) for IL-1b, 12262.36 pg/ml and Cervical NP specimens of degenerative grade III and IV were isolated from 15 patients. From each specimen 4610 5 cells were grown in collagen I scaffold for four weeks to determine the endogenous expression levels (ELISA) of MMPs on the basis of disc degeneration grade (DDG). For each sample 100 mg total protein extracts were applied per experiment. Whiskers min to max of box plots show MMP-1 expression levels (Fig. 2a), MMP-2 expression levels (Fig. 2b), MMP-3 expression levels (Fig. 2c), MMP-7 expression levels (Fig. 2d) and MMP-13 expression levels (Fig. 2e). doi:10.1371/journal.pone.0096870.g002 Table 3. Levels of endogenous protein expression for catabolic and anti-catabolic cytokines in degenerative cervical NP cells.

Levels of Endogenous Anabolic and Matrix Proteins in Degenerative Cervical NP Cells
Decreasing expression levels of matrix proteins aggrecan and collagen II were determined in degenerative cervical NP cells of degeneration grade III and IV. The respective mean expression values of aggrecan were 2668261861 pg/ml and 145016936 pg/ ml (P,0.0001) and of collagen II were 95676384 pg/ml and 60456198 pg/ml (P,0.0001). The mean expression values of aggrecan and collage II in degeneration grade III were about 1.8 and 1.6 fold higher than that in degeneration grade IV (table 4  and figure 5a-b). The expression level of collagen I in cervical NP cells remained below the minimum detectable dose (MDD) of our detection system (MDD,217 pg/ml). Similarly in control 3D culture (agarose gel) and control 2D culture (TC dishes) the expression level of collagen I in cervical NP cells remained below the minimum detectable dose (data not shown). Moreover, the expression levels of the analyzed anabolic factors remained below the minimum detectable dose (MDD) of our detection system (MDD,1-25 pg/ml). They were expressed either at extremely low level or not expressed at all. The calculated significance of changes in expression levels as a function of age by decade showed decreasing expression levels of aggrecan (P,0.0001) and collagen II (P,0.0001). Gender does not appear to play any role in influencing the expression levels of the anabolic and matrix proteins (data not shown).

Discussion
In previous publications the issues of biological treatment approaches in degenerative cervical disc diseases have not been reported, even though cervical disc degeneration, which could be associated with pain and loss of mobility, is clinically important. So far, the pathology of cervical disc degeneration has been studied only from a morphologic view point using MRI [1][2][3][4][5][6]. Different biological treatment approaches, which were carried out in degenerative lumbar discs, have shown the importance of NP cells for the maintenance of IVD matrix. They coordinate the expression of anabolic, catabolic, anti-catabolic and inflammatory cytokines that affect the synthesis or degradation the IVD matrix.
IVD degeneration is associated with imbalances of these factors combined with the declined cell density in adult IVDs [11][12][13][14][15][16][17][18][19][20][21][22][23]. Therefore, accurate knowledge regarding the quantity and quality of these factors is vital for designing of rational gene therapeutic approaches. So far proliferation rates of degenerative cervical NP To determine the endogenous expression levels of TIMPs in cervical NP cells, 15 NP specimens of degenerative grade III and IV of herniated discs were used. In collagen I scaffold 4610 5 NP cells from each specimen were grown for four weeks, and on the basis of disc degeneration grade (DDG) the protein concentration of TIMPs were defined (ELISA) from 100 mg total protein extracts of each sample. TIMP-1 expression levels (Fig. 3a), TIMP-2 expression levels (Fig. 3b), TIMP-3 expression levels (Fig.  3c) and TIMP-4 expression levels (Fig. 3d)  Therefore, we analyzed the proliferation rates of degenerative cervical NP cells and their endogenous protein expression levels of 28 anabolic, catabolic, anti-catabolic and inflammatory cytokines as well as matrix proteins. Cervical IVD specimens of degeneration grade III and IV from 15 adult patients, operated due to cervical disc herniation, were used.
Inclusion criteria for surgery were radiographically determined intervertebral disc herniation with nerve root compression on MRI, which correlated to primary symptoms that remained unresponsive to non-operative treatment for six weeks or demonstrated progressive neurological deterioration in the face of conservative treatment. Our results may therefore be representative of degenerative cervical discs from such patients.
We determined comparable proliferation rates of cervical NP cells between degeneration grades III and IV (table 2, figure 1a). Age and gender do not seem to play a role in influencing proliferation rates of degenerative cervical NP cells. The data might suggest that progressive degenerative changes in adult IVDs may not be triggered by decreasing of NP cells. This might occur instead because of undesirable phenotypic changes in NP cells. Previous studies of lumbar NP cells have also indicated similar cell densities in adult age of different disc degeneration grades [24,28]. Conversely, some studies of lumbar NP cells reported a decline of NP cell densities with increasing degeneration grades [29][30][31]. Certainly, during adolescence IVDs have considerably reduced cell density than during childhood, for the reason that the cell environment within the disc severely changes during growth. During growth IVD size increases and blood supply as well as diffusion decrease, which result in cell death triggered by decreased concentration of glucose and oxygen [32][33][34][35][36][37].
High and increasing expression levels of catabolic factor ADAMTS-4 with about 1.9 fold upturn were recorded between degeneration grades III and IV (table 3, figure 1b-c). Additionally, high but comparable expression levels of ADAMTS-5 were recorded in both grades of degeneration. In contrast to the previous finding, which showed the age of patients is associated with the percentage of immunohistological ADAMTS-5 positive lumbar disc cells [38], we found here age and gender to be independent factors regarding the expression ADAMTS-4 and ADAMTS-5. These data, supporting the previous report on lumbar NP cells [24], indicate that ADAMTS-4 and ADAMTS-5 could represent attractive targets for biological treatment approaches of degenerative disc diseases. Moreover, they have been shown to cause aggrecan degradation in human osteoarthritic cartilage [39][40][41].
Compared to all other tested catabolic factors MMP-3 was expressed at highest levels, but its expression levels did not alter with increasing grades of degeneration. Its mean expression level was about 33 fold of MMP-1, 118 fold of MMP-2, 25 fold of MMP-7 and 20 fold of MMP-13 (Table 3 and figure 2a-e). MMP-8, MMP-9 and MMP-10 were expressed either at extremely low level or not expressed at all, as their expression levels remained below the minimum detectable dose (table 3).
In contrast to a previous report, where immunohistological staining was negative for TIMP-1 in MMP-3 positive stained surgical lumbar specimens [42], the high expression of MMP-3 in degenerative cervical NP cells was counteracted by even higher and increasing expression levels of TIMP-1 (1.9 fold of MMP-3) Figure 4. Protein expression levels of endogenous inflammatory cytokines in cervical NP cells. 15 degenerative grade III and IV herniated cervical discs were used to isolate NP specimens. Collagen I scaffold was used to culture 4610 5 NP cells from each specimen for four weeks. The endogenous expression levels of inflammatory cytokines were confirmed (ELISA) from 100 mg total protein extracts of each sample on the basis of disc degeneration grade (DDG). Box plots with whiskers min to max show IL-1b expression levels (Fig. 4a), IL-1 R expression levels (Fig. 4b), TNF-a expression levels (Fig. 4c) and TNF-a R1 expression levels (Fig. 4d) [20]. Moreover the most extensive immunohistochemical stainings were seen for MMP-1, MMP-2, MMP-3, and MMP-9 and much less for MMP-7 and MMP-8, and these up-regulations were paralleled by greater expression of TIMP-2 and not TIMP-1 [41]. Furthermore the number of immunopositive cells for MMP-1, MMP-3, MMP-13 and ADAMTS-4 increased with the severity of degeneration and this was accompanied by increased number of immunopositive cells for TIMP-1 and TIMP-2 but not for TIMP-3 [19]. Our data suggest that ectopic expression of TIMP-3, an inhibitor of ADAMTs, and repression of MMP-3 would be more interesting to improve the regeneration potential of degenerative cervical NP cells. On the other hand, as TIMP-1 and TIMP-2, inhibitors of MMP-3, are expressed at higher levels than MMP-3, their ectopic expression might not be potentially effective. It would be quite more interesting to focus on their mutational and posttranslational alterations.
The inflammatory cytokines IL-1b and TNF-a as well as their receptors IL-1 R and TNF-a R1 were expressed relatively at very low levels. However, higher expression levels of IL-1b and IL-1 R were detected in degeneration grade III and IV than that of TNFa and TNF-a R1. Moreover, increased mean expression values of about 10% were recorded between degeneration grade III and IV for IL-1b, IL-1 R and TNF-a, whereas the mean expression values of TNF-a R1 remained unaffected (table 4 and figure 4a-d). Based on our data, IL-1b, IL-1R as well as TNF-a might be involved in the pathogenesis of cervical disc degeneration, where IL-1b and IL-1 R might act as better therapeutic targets. Furthermore, treatment of lumbar disc cells with 10 ng/ml recombinant IL-1b has shown increased expression levels of ADAMTS-4, MMP-3 and MMP-13 with decreased expression levels of aggrecan, collagen II, collagen I and SOX6 [43]. Additionally, real time PCR and immunohistochemistry studies have previously shown the expression of IL-1b, IL-1 R, TNF-a and TNF-a R1 in lumbar NP cells of healthy and degenerative discs, and their expression levels increased with increasing degeneration grades [21,41].
We determined with increasing grades of degeneration decreased expression levels of matrix proteins aggrecan and collagen II in cervical NP cells. Between degeneration grade III and IV about 1.8 and 1.6 folds of decreased aggrecan and collagen II expressions were recorded respectively (table 4 and figure 5a-b). The expression level of collagen I remained below the minimum detectable dose of our detection system (table 4). Moreover, the expression level of collagen I in cervical NP cells remained also below the minimum detectable dose (data not shown) in control 3D culture (agarose gel) as well as in control 2D culture (TC dishes). The calculated significance of concentration changes as a function of age by decade showed decreasing expression levels of aggrecan (P,0.0001) and collagen II (P,0.0001). Gender does not appear to play any role in influencing the expression levels of matrix proteins (data not shown). Degeneration grade and age correlated changes were also shown in lumbar disc cells [24,[44][45].
Growth factors have been shown to be important biological components to stimulate matrix synthesis [11,46]. However, their endogenous expression levels in cervical IVD cells have never been studied. Using immunohistochemical analysis only a few number of publications have presented the expression of growth factors bFGF, TGF-b1, TGF-b2 and growth factor receptors TGFb RII, FGF R3 and BMP RI in lumbar IVDs [47,22]. Although we applied a large amount of total protein extract from cervical NP cells (100 mg), the protein expression levels of all tested growth factors remained in contrast below the minimum detectable doses. In addition, the minimum detectable doses of the tested growth factors are very low (table 4). Hence, our data emphasize imbalances between the expression levels of anabolic and catabolic proteins in degenerative cervical NP cells, which might result in a catabolic inflammatory metabolism of the disc matrix.
Therefore, the endogenous protein expression data of the anabolic, catabolic, anti-catabolic and inflammatory cytokines in degenerative cervical NP cells suggest that suppression of the catabolic factors (MMP-3, ADAMTS-4 and ADAMTS-5) along with the inflammatory cytokines (IL-1b and IL-1 R) might be a favorable gene therapeutic approach. This approach could be combined with ectopic expression of the anabolic factors and the anti-catabolic factor TIMP-3, an inhibitor of ADAMTS-4 and ADAMTS-5. The joint course might improve the regeneration potential of degenerative cervical disc cells. Figure 5. Endogenous expression levels of matrix proteins in degenerative cervical NP cells. From 15 herniated cervical discs of grade III and IV NP tissues were isolated and 4610 5 cells from each sample were cultured in collagen I scaffold for four weeks. On the basis of disc degeneration grade (DDG) the concentration of aggrecan, collagen I and collagen II were measured (ELISA) from 100 mg total protein extracts of each sample. Aggrecan expression levels (Fig. 5a) and collagen II expression levels (Fig. 5b) are shown using box plots with whiskers min to max. Collagen I expression level remained below minimum detectable dose of our detection system (table 4). doi:10.1371/journal.pone.0096870.g005