The response of human anulus fibrosus cells to cyclic tensile strain is frequency-dependent and altered with disc degeneration.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 62, No. 11, November 2010, pp 3385–3394 DOI 10.1002/art.27643 © 2010, American College of Rheumatology The Response of Human Anulus Fibrosus Cells to Cyclic Tensile Strain Is Frequency-Dependent and Altered with Disc Degeneration Hamish T. J. Gilbert, Judith A. Hoyland, and Sarah J. Millward-Sadler Results. The expression of catabolic genes (MMP-3 and ADAMTS-4) in AF cells derived from nondegenerated tissue decreased in response to 1.0 Hz of CTS, whereas changing the frequency to 0.33 Hz resulted in a shift toward matrix catabolism. Application of 1.0 Hz of CTS reduced anabolic gene expression (aggrecan and type I collagen) in AF cells derived from degenerated tissue, with 0.33 Hz of CTS resulting in increased catabolic gene expression. Conclusion. The response of human AF cells to CTS is frequency-dependent and is altered by degeneration. Objective. Mechanical loads are important for homeostasis of the intervertebral disc (IVD) cell matrix, with physiologic and nonphysiologic loads leading to matrix anabolism and catabolism, respectively. Previous investigations into the effects of load on disc cells have predominantly used animal models, with the limited number of human studies focusing primarily on nucleus pulposus cells. The aim of this study was to examine the effect of cyclic tensile strain (CTS) on human anulus fibrosus (AF) cells to ascertain whether the response was frequency-dependent and to compare AF cells derived from nondegenerated and degenerated tissue samples. Methods. AF cells were isolated from nondegenerated and degenerated human IVDs, expanded in monolayer, and cyclically strained for 20 minutes, applying 10% strain at a frequency of 1.0 Hz or 0.33 Hz with the use of a Flexcell strain device. Total RNA was extracted from the cells at baseline (control) and at 1, 3, and 24 hours following application of CTS. Real-time quantitative polymerase chain reaction was used to analyze gene expression of matrix proteins (aggrecan, type I collagen, and type II collagen) and enzymes (matrix metalloproteinases [MMPs] 3, 9, 13, and ADAMTS-4). The intervertebral disc (IVD) can be divided into 2 main regions, the central gelatinous nucleus pulposus (NP) and the outer anulus fibrosis (AF). The forces that act on these regions and their component cells are multiple and complex, with cells in the NP exposed to hydrostatic pressures and cells within the AF exposed primarily to tensile strain (1,2). Internal disc pressures are predicted to range from 0.1 MPa (while lying down) to 2.5 MPa (while lifting a heavy weight) (3), with tissue strains reaching 25% during physiologic compression (4). Mechanical load is known to be important for the maintenance and integrity of the IVD cell matrix, with physiologic mechanical stimuli leading to matrix anabolism (5–11). Hutton et al (12) found that in canine NP cells, 1 MPa of hydrostatic pressure led to increased collagen and proteoglycan synthesis as compared with atmospheric pressure, while Matsumoto et al (11) reported increased collagen synthesis in rabbit NP cells exposed to 20% cyclic tensile strain (CTS). Furthermore, exposure of the IVD to nonphysiologic mechanical load (e.g., overloading or immobilization) results in altered cellular metabolism, which leads to degenerative changes (9,13–16). Wang et al (9) showed that the anabolic response of rabbit IVD explants to dynamic Supported by Arthritis Research UK (grant 17850). The Intervertebral Disc Research Group at the University of Manchester School of Biomedicine is supported by the Manchester Academic Health Sciences Centre and the National Institute of Health Research, Manchester Biomedical Research Centre. Hamish T. J. Gilbert, BSc (Hons), Judith A. Hoyland, PhD, Sarah J. Millward-Sadler, PhD: University of Manchester, Manchester, UK. Address correspondence and reprint requests to Sarah J. Millward-Sadler, PhD, School of Biomedicine, Faculty of Medical and Human Sciences, The University of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT, UK. E-mail: Jane.Sadler@ manchester.ac.uk. Submitted for publication January 29, 2010; accepted in revised form June 29, 2010. 3385 3386 compression was inhibited when the compression was changed to a static force and that static in vivo bending of murine IVDs resulted in increased cell death and decreased type II collagen gene expression (14). The type, magnitude, frequency, and duration of force have all been identified as important factors in determining IVD cell metabolism in response to load (9,10,13,15,17–21). Maclean et al (7) found that NP cells in an in vivo rat tail model responded in a frequency-dependent manner, with low-frequency (0.01 Hz) and highfrequency (1 Hz) loading having anabolic and catabolic effects, respectively, whereas the response of AF cells was mainly magnitude-dependent. A frequencydependent response was also shown in hydrostatically loaded porcine NP and AF cells, where a frequency of 5 Hz resulted in poor collagen synthesis and increased degradation, whereas frequencies above and below this value had minimal effects (18). Animal models have been predominantly used in mechanical loading studies, but differences in solute transport and matrix biochemistry due to variations in disc size and age, respectively, prevent their use as an accurate model for investigating human IVD mechanobiology (22). Thus, to ascertain the effects of load on human IVD, in vitro mechanical loading systems capable of directing specific and quantifiable mechanical stimuli to cultured cells have been used to study magnitudedependent responses. Handa et al (23) found that hydrostatic pressure influenced proteoglycan synthesis and matrix metalloproteinase (MMP) gene expression in human NP explants, with low hydrostatic pressure promoting matrix anabolism and high hydrostatic pressure leading to matrix catabolism, a response also confirmed by studies conducted by Neidlinger-Wilke et al (15). Although these human studies show the importance of mechanical load in IVD cell matrix regulation, the majority of studies have focused on the response of hydrostatic pressure–stimulated NP cells, with only a few studies investigating the response of AF cells to load. Using human and bovine tissue explants, Ishihara et al (21) showed that hydrostatic pressure influenced proteoglycan synthesis by NP cells and inner AF cells, while outer AF cells remained unaffected by this type of load. Using canine cells, Hutton et al (12) found that proteoglycan and collagen synthesis were stimulated in NP cells but inhibited in AF cells following hydrostatic pressure. In studies of human cell cultures, Le Maitre et al (24) observed that the response of NP cells exposed to hydrostatic pressure was absent in AF cells exposed to an identical stimulus. These findings support speculation GILBERT ET AL that hydrostatic pressure is not the predominant load experienced by AF cells in vivo and suggest that investigating a more physiologically relevant load would be beneficial in progressing our understanding of AF cell biomechanics. To date, there are only limited studies of AF cells treated with CTS, the more physiologically relevant load. Wuertz et al (25) found that 6% CTS had only a modest effect on matrix gene expression of human AF cells in 3-dimensional culture as compared with the influence of different osmotic potentials. Sowa and Agarwal (26) found that CTS exhibited a protective effect against inflammatory cytokine–stimulated catabolism in monolayer cultures of rat AF cells, whereas Rannou et al (27) found that CTS initiated a catabolic response in rabbit AF cells. Although the effect of mechanical load on nondegenerated IVDs has been documented, investigations into the implications of degeneration during loading remain limited. However, a recent study showed that NP cells derived from degenerated IVDs failed to respond to hydrostatic pressure (24). This inability to respond to physiologic loads in the normal anabolic way could result in altered matrix regulation and lead to the progression of degenerative disc disease. As such, if the response to load is altered in cells derived from degenerated IVDs then elucidation of the altered mechanotransduction pathway could lead to future therapeutic targets for the prevention or treatment of degenerative disc disease. To our knowledge, there are currently no studies investigating the effect of frequency on cyclically strained human AF cells, nor any studies comparing the response of AF cells derived from nondegenerated or degenerated IVDs to load. For the present study, an in vitro loading system (FX-4000 Flexcell Tension system) capable of directing CTS to cells in monolayer was used to assess the effect of a physiologically relevant mechanical stimulus on human AF cells. The aims were to investigate the effect of CTS on AF cells isolated from human IVDs to ascertain whether the response was frequency-dependent and to investigate whether the response of AF cells derived from nondegenerated and degenerated IVDs differed. MATERIALS AND METHODS IVD tissue samples. Human IVD tissue was collected at the time of surgery or at postmortem examination (within 18 hours of death), with the consent of the patient or the relatives of the decedent and with the approval of Central Manchester, Bury, Rochdale, Salford, and Trafford Research Ethics Committees. A section of formalin-fixed, paraffin-embedded tissue RESPONSE OF HUMAN AF CELLS TO CYCLIC TENSILE STRAIN Table 1. Characteristics of the intervertebral disc tissue specimens examined Sample Source Grade* Age, years Disc level 1 2 3 4 5 6 Postmortem Postmortem Postmortem Postmortem Surgical Surgical 1 (nondegenerated) 1 (nondegenerated) 1 (nondegenerated) 7 (mildly degenerated) 9 (degenerated) 9 (degenerated) 57 46 37 57 29 66 L4/5 L5/S1 L5/S1 L2/3 L4/5 L4/5 * Histologic changes in the tissues were graded on a scale of 0–12, where 0–3 ⫽ nondegenerated, 4–7 ⫽ mildly degenerated, and 8–12 ⫽ severely degenerated. derived from the entire IVD containing intact NP and AF was graded for histologic changes as previously reported (28). Tissue changes were scored on a scale of 0–12, where 0–3 ⫽ nondegenerated tissue, 4–7 ⫽ mildly degenerated, and 8–12 ⫽ severely degenerated. Characteristics of the tissue samples used in these studies are detailed in Table 1. Isolation and culture of AF cells. Within 24 hours of death or surgical removal, AF tissue (including the inner and outer anuli) was separated from the IVD, finely minced, digested with 300–350 PUK/ml of Pronase (Calbiochem) for 1 hour, and then with 0.25% type II collagenase (Invitrogen) and 0.01% hyaluronidase (Sigma) for 4 hours at 37°C with constant agitation in Dulbecco’s modified Eagle’s medium. AF cells were cultured in standard medium, which consisted of Dulbecco’s modified Eagle’s medium, with 4.5 gm/liter glucose, Glutamax and pyruvate (both from Gibco) containing 50 g/ml of ascorbic acid, 250 ng/ml of amphotericin, 100 units/ml of penicillin, 100 g/ml of streptomycin (Invitrogen), and 10% fetal calf serum (Invitrogen). Cells were expanded in monolayer, with medium changes every 2–3 days. Cells with passage numbers of ⱕ6 were trypsinized and seeded onto untreated silicone membrane BioFlex culture plates 3387 (Flexcell International) at a density of 1 ⫻ 105 cells/ml in 2 ml of standard medium and were allowed to adhere for 48 hours (passage number ⬎6 has been found to influence cell behavior [Hoyland JA: unpublished observations]). Medium was changed to serum-free medium at 15–17 hours prior to application of CTS. Application of CTS using a FX-4000 Flexcell tension system. BioFlex culture plates containing AF cells in serumfree medium were placed into a vacuum manifold inside an incubator set to a temperature of 37°C in an atmosphere of 5% CO2. A precise amount of negative pressure was directed to the base of the culture plates, causing deformation of the silicone membranes over the loading posts. The use of 25-mm loading posts with BioFlex culture plates allows computercontrolled equibiaxial strain of the silicone membranes located directly above the posts. This strain is directly translated to the adherent cells, resulting in equibiaxial cellular deformation. The membrane regions and adherent cells located away from the loading posts experience biaxial strain with greater elongation in the radial direction as compared with the circumferential direction (29). For this study, a FX-4000 Flexcell tension system (Flexcell International) was used to apply CTS of 10% (directed by the computer), resulting in a 10% membrane strain at the central region and a strain of 14% radial and 5% circumferential at the membrane periphery (29,30). The AF is predicted to experience compressive tissue strains of 1–25% and radial tensile strains of up to 19% during multiaxial bending motions when compressed with a load that is physiologically similar to that of walking (3,4). Two different loading regimens were used: either 10% strain at a frequency of 0.33 Hz (1.5 seconds of extension, with 1.5 seconds of relaxation, at a rate of 20 cycles/minute) for 20 minutes or 10% strain at a frequency of 1.0 Hz (0.5 seconds of extension, with 0.5 seconds of relaxation, at a rate of 60 cycles/minute) for 20 minutes. Frequencies of 1.0 Hz and 0.33 Hz were chosen as being representative of physiologic locomotion and less than physiologic locomotion, respectively (31). Cells adhering to un- Table 2. Primer and probe sequences used in the real-time polymerase chain reaction analysis Target gene Forward primer Reverse primer GAPDH GGT-GAA-CGG-GAA-GCT-CAC-T AGG-TCA-GGT-CCA-CCA-CTG-A Aggrecan CCG-TGT-GTC-CAA-GGA-GAA-GG GGG-TAG-TTG-GGC-AGT-GAG-AC Type I collagen Type II collagen MMP-3* AGA-ACA-GCG-TGG-CCT-ACA-TG GCG-CGG-ATC-TCG-ATC-TCG ATG-GAG-ACT-GGC-GAG-ACT-TG GCT-GCT-CCA-CCA-GTT-CTT-CTT TGA-AGA-GTC-TTC-CAA-TCCTAC-TGT-TG CCC-GGA-GTG-AGT-TGA-ACC-A CTA-GAT-ATT-TCT-GAA-CAA-GGTTCA-TGC-A CAG-GAC-GGG-AGC-CCT-AGT-C MMP-9 MMP-13 CCC-CAG-GCA-TCA-CCA-TTCAAG ADAMTS-4 ACT-GGT-GGT-GGC-AGA-TGA-CA * MMP-3 ⫽ matrix metalloproteinase 3. Probe CCC-CAG-TGC-CAACGT-G CTG-ATA-GGC-ACTGTT-GAC CAG-CAG-ACT-GGCAAC CCC-AAT-CCA-GCAAAC-G TTT-GCT-CAG-CCT-ATCCAT TAC-GTG-ACC-TATGAC-ATC CTG-CCT-TCC-TCT-TC GAC-AAA-TCA-TCT-TCA-TCA-CCACCA-C TCA-CTG-TTA-GCA-GGT-AGC-GCT-TT ATG-GCC-GCA-TTC-C GenBank accession no. NM_002046 NM_001135 NM_000088 NM_001844 NM_002422 NM_004994 NM_002427 NM_005099 3388 GILBERT ET AL stimulated BioFlex culture plates served as controls. Stimulated and unstimulated AF cells were incubated at 37°C for 0 (baseline control), 1, 3, or 24 hours postload in serum-free medium, and total RNA was extracted. Cell viability. Cell viability was assessed before and after cyclic stretching using a trypan blue exclusion assay. Briefly, 100 l of trypan blue (0.4%; Sigma) was added to each well of a BioFlex culture plate containing unstimulated or stimulated AF cells. Stained AF cells were visualized using a light microscope (Leitz), and the percentage of viable cells (cells with unstained nuclei) was calculated by counting 100 cells in 5 random fields of view. Real-time quantitative polymerase chain reaction (PCR) analysis. Total RNA was extracted from each BioFlex culture plate well using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions, and samples were treated with DNase I (Ambion). RNA quality and quantity were determined using a NanoDrop ND-1000 spectrophotometer and 500 ng of RNA that had been reverse transcribed using a high-capacity reverse transcription kit (Applied Biosystems). Real-time quantitative PCR was performed in triplicate using TaqMan Universal PCR Master Mix (Applied Biosystems) with primers and probes at 900 nM and 250 nM, respectively, for GAPDH, aggrecan, types I and II collagen, MMPs 3, 9, and 13, and ADAMTS-4 (Table 2). Data were analyzed using the 2–⌬⌬Ct method (32,33), and the results were normalized to the endogenous control gene GAPDH and unloaded baseline controls. Primers were designed using Primer Express Software (Applied Biosystems) (Table 2). Statistical analysis. Nonparametric data, as determined by the Shapiro-Wilk test, were analyzed by MannWhitney U test. RESULTS Cell viability and baseline expression. AF cells isolated from nondegenerated and degenerated IVDs remained viable (⬎90%) throughout the culture period, with no change in viability following the application of CTS. Unloaded controls showed no significant change in gene expression for any of the genes investigated at any of the time points analyzed (data not shown). There was no significant difference in the baseline levels of relative gene expression for type II collagen, MMP-3, MMP-9, MMP-13, or ADAMTS-4 between AF cells from nondegenerated or degenerated IVDs. Baseline relative gene expression of aggrecan and type I collagen was significantly increased (P ⬍ 0.01) and decreased (P ⬍ 0.05), respectively, in AF cells derived from degenerated IVDs as compared with those from nondegenerated IVDs. Effects of CTS at a frequency of 1.0 Hz. Application of CTS at 10% strain and a frequency of 1.0 Hz for 20 minutes to AF cells from nondegenerated IVDs resulted in no significant change in the relative expression of any of the matrix genes investigated. Exposure of Figure 1. Effect of cyclic tensile strain on matrix gene expression in anulus fibrosus cells from nondegenerated and degenerated intervertebral discs (IVDs). Cells derived from nondegenerated or degenerated IVDs were mechanically stimulated (M/S) at 10% strain and a frequency of 1.0 Hz for 20 minutes and then incubated for up to 24 hours prior to analysis. Real-time quantitative polymerase chain reaction analysis was used to determine gene expression of matrix proteins. Levels of aggrecan (A) and type I collagen (B) were determined in degenerated and nondegenerated tissue relative to the housekeeping gene GAPDH and were normalized to the unstimulated baseline control. Values are the mean ⫾ SEM of samples from 3 subjects per group. ⴱ ⫽ P ⱕ 0.05 versus unstimulated baseline control. AF cells from degenerated IVDs to an identical stimulus resulted in a significant down-regulation of aggrecan and type I collagen genes at 24 hours following mechanical stimulation (5-fold [P ⬍ 0.01] and 6-fold [P ⬍ 0.05], respectively) (Figures 1A and B), but no change in the relative expression of the type II collagen gene (data not shown). Analysis of matrix-degrading enzymes in AF cells from nondegenerated IVDs subjected to a 1.0-Hz stimulus showed a significant decrease in the expression of the MMP-3 gene at 1 hour and at 24 hours poststimulation (6-fold [P ⬍ 0.05] and 7-fold [P ⬍ 0.01], respec- RESPONSE OF HUMAN AF CELLS TO CYCLIC TENSILE STRAIN 3389 Figure 2. Effect of cyclic tensile strain on matrix-degrading enzyme gene expression in anulus fibrosus cells from nondegenerated and degenerated intervertebral discs (IVDs). Cells derived from nondegenerated or degenerated IVDs were mechanically stimulated at 10% strain and a frequency of 1.0 Hz for 20 minutes and then incubated for up to 24 hours prior to analysis. Real-time quantitative polymerase chain reaction analysis was used to determine gene expression of matrix-degrading enzymes. Levels of matrix metalloproteinase 3 (MMP-3) (A) and ADAMTS-4 (B) were determined in degenerated and nondegenerated tissue relative to the housekeeping gene GAPDH and were normalized to the unstimulated baseline control. Values are the mean ⫾ SEM of samples from 3 subjects per group. ⴱ ⫽ P ⱕ 0.05 versus unstimulated baseline control. tively), but the decrease at 3 hours did not achieve significance (P ⫽ 0.07). Relative expression of the ADAMTS-4 gene was also significantly decreased at 1 hour poststimulation (7-fold; P ⬍ 0.05) (Figure 2B). Relative expression of the MMP-9 and MMP-13 genes remained unchanged (data not shown). AF cells from degenerated IVDs showed no change in the relative expression of the matrix-degrading enzyme genes at any time point following 1.0 Hz of stimulation. Effects of CTS at a frequency of 0.33 Hz. Mechanical stimulation of AF cells from nondegenerated Figure 3. Effect of cyclic tensile strain on matrix gene expression in anulus fibrosus cells from nondegenerated and degenerated intervertebral discs (IVDs). Cells derived from nondegenerated or degenerated IVDs were mechanically stimulated at 10% strain and a frequency of 0.33 Hz for 20 minutes and then incubated for up to 24 hours prior to analysis. Real-time quantitative polymerase chain reaction analysis was used to determine gene expression of matrix proteins. Levels of aggrecan (A), type I collagen (B), and type II collagen (C) were determined in degenerated and nondegenerated tissue relative to the housekeeping gene GAPDH and were normalized to the unstimulated baseline control. Values are the mean ⫾ SEM of samples from 3 subjects per group. ⴱ ⫽ P ⱕ 0.05 versus unstimulated baseline control. 3390 IVDs with 10% strain at a frequency of 0.33 Hz for 20 minutes resulted in decreased relative expression of type I collagen (9-fold decrease at all 3 time points; P ⬍ 0.05) and type II collagen (10-fold decrease at 1 hour [P ⬍ 0.01], 9-fold decrease at 3 hours [P ⬍ 0.05], and 8-fold decrease at 24 hours [P ⬍ 0.01]) (Figures 3B and C). AF cells from nondegenerated IVDs exposed to 0.33 Hz of CTS showed no difference in the relative gene expression of aggrecan between stimulated tissues and unstimulated controls (Figure 3A). The expression of types I and II collagen genes in AF cells from degenerated IVDs following exposure to 0.33 Hz of CTS was similar to that observed in AF cells from nondegenerated IVDs exposed to the same stimulus. The relative gene expression of type I collagen was decreased at all 3 time points (⬃8-fold; P ⬍ 0.01), whereas the relative gene expression of type II collagen was significantly decreased at 1 hour and at 24 hours post-CTS (14-fold [P ⬍ 0.01] and 9-fold [P ⬍ 0.05], respectively), although the decreased expression at the 3-hour time point was not significant as compared with control (Figures 3B and C). In contrast, AF cells from degenerated IVDs showed decreased expression of aggrecan 24 hours poststimulus (6-fold; P ⬍ 0.01) (Figure 3A) but showed no significant change at 1 hour or at 3 hours poststimulation. Exposure of AF cells from nondegenerated and degenerated IVDs to CTS at a frequency of 0.33 Hz resulted in an increase in the relative gene expression of MMP-9 at 3 hours poststimulation (4-fold; P ⬍ 0.05), with a return toward baseline values by 24 hours (Figure 4A). MMP-13 relative gene expression was decreased at all 3 time points in AF cells from nondegenerated IVDs, but the difference did not reach significance. AF cells from degenerated IVDs subjected to 0.33 Hz of CTS showed increased relative gene expression of MMP-13 at 3 hours poststimulation (2-fold; P ⬍ 0.05), with expression returning to baseline values by 24 hours (Figure 4B). ADAMTS-4 relative gene expression remained unchanged in AF cells from nondegenerated IVDs following 0.33 Hz of CTS but was decreased at 1 hour poststimulation in AF cells from degenerated IVDs exposed to the same stimulus (5-fold; P ⬍ 0.05) (Figure 4C). There was no change in gene expression of MMP-3 following cyclic stretch of AF cells from nondegenerated or degenerated IVDs at a frequency of 0.33 Hz (data not shown). DISCUSSION AF cells are exposed to CTS during normal physiologic movement of the spine, and this mechanical GILBERT ET AL Figure 4. Effect of cyclic tensile strain on matrix-degrading enzyme gene expression in anulus fibrosus cells from nondegenerated and degenerated intervertebral discs (IVDs). Cells derived from nondegenerated or degenerated IVDs were mechanically stimulated at 10% strain and a frequency of 0.33 Hz for 20 minutes and then incubated for up to 24 hours prior to analysis. Real-time quantitative polymerase chain reaction analysis was used to determine gene expression of matrix-degrading enzymes. Levels of matrix metalloproteinase 9 (MMP-9) (A), MMP-13 (B), and ADAMTS-4 (C) were determined in degenerated and nondegenerated tissue relative to the housekeeping gene GAPDH and were normalized to the unstimulated baseline control. Values are the mean ⫾ SEM of samples from 3 subjects per group. ⴱ ⫽ P ⱕ 0.05 versus unstimulated baseline control. RESPONSE OF HUMAN AF CELLS TO CYCLIC TENSILE STRAIN stimulation is important in regulating the AF cell matrix. This study is the first to compare the effects of CTS on human AF cells derived from both nondegenerated and degenerated IVDs, where the aims were to characterize the response in terms of anabolic and catabolic gene expression and to compare the effects of loading frequency. The AF cells derived from both nondegenerated and degenerated IVDs showed a frequency-dependent response to CTS, although the response was most dramatic in the nondegenerated AF cells. At a frequency of 1.0 Hz, which is similar to the frequency of normal physiologic locomotion in humans (31), AF cells derived from nondegenerated tissue exhibited decreased gene expression of MMP-3 and ADAMTS-4, suggesting a shift in matrix regulation toward reduced catabolism. In contrast, altering the frequency to 0.33 Hz resulted in a catabolic response of AF cells derived from nondegenerated IVDs, with decreased expression of types I and II collagen genes and increased expression of the MMP-9 gene. Frequency has previously been shown to be an important factor in determining cellular metabolic activity in response to load. MacLean et al (7) found a distinct frequency-dependent response in NP cells using a rat tail dynamic compression model, with anabolic gene expression elevated at a frequency of 0.01 Hz and catabolic gene expression elevated at a 1.0-Hz frequency. Although the AF cells in that study responded less dramatically than the NP cells, there was still an obvious frequency-dependent response to 0.2 MPa of compression, with anabolic gene expression elevated at the 1.0-Hz frequency only, suggesting an overall anabolic response at the tissue level (7). Similarly, the response of AF cells to CTS at a frequency of 1.0 Hz in our study suggests an overall anabolic effect as a consequence of reduced catabolic gene expression. The differences in the genes regulated between these studies could be due to species variation and differences in the type and magnitude of force experienced at the cellular level. The in vitro model system we used examines the effect of cell deformation due to CTS only, and although we are unable to exclude cell stimulation due to fluid flow, our system allowed the removal of all other types of mechanical stimulation present during in vivo mechanical loading. The biochemical processes that enable cells to respond to mechanical stimuli in a frequency-dependent manner remain largely unknown; however, a study investigating the effects of CTS on bone cells found that osteoblasts used different integrin receptors at different 3391 frequencies of strain, suggesting that there are different mechanotransduction pathways operating between frequencies (34). Other factors that could be involved in the frequency-dependent response observed in the present study are the differences in the strain rate and the number of cycles (400 cycles at the 0.33-Hz frequency and 1,200 cycles at the 1.0-Hz frequency). Both the strain rate and the cycle number could affect which mechanotransduction pathways are activated and, consequently, could alter the kinetics of such pathways. The response of AF cells derived from degenerated tissue was also found to be frequency-dependent, with 1.0 Hz of cyclic strain leading to reduced matrix anabolism (decreased aggrecan and type I collagen gene expression) and 0.33 Hz cyclic strain leading to matrix catabolism (decreased anabolic gene expression and increased catabolic gene expression). Altering the frequency of the strain led to a switch in gene regulation, from matrix-degrading enzyme genes to matrix genes for AF cells derived from nondegenerated tissue and from matrix genes to matrix-degrading enzyme genes for AF cells derived from degenerated tissue. Although an alteration in gene regulation occurred in both cell types, the frequency-dependent response was most prominent in the AF cells derived from nondegenerated tissue, where there was a shift in matrix regulation, from reduced catabolism at 1.0 Hz (suggesting matrix anabolism at the tissue level) to increased catabolism at 0.33 Hz. This change in matrix regulation, which was due to altered frequency, from an overall anabolic effect to a catabolic effect was not replicated in the AF cells derived from degenerated tissue, where both frequencies ultimately led to matrix catabolism. This appears to be primarily due to the altered response observed between AF cells derived from nondegenerated and degenerated tissue at a 1.0-Hz frequency of strain, suggesting that the response of AF cells to CTS is not only dependent on the frequency, but also on whether the cells are derived from nondegenerated or degenerated tissue. Our data suggest that AF cells derived from degenerated tissue have not only lost their ability to down-regulate matrix enzyme gene expression in response to a 1.0-Hz frequency of CTS, but are also responding by reducing their matrix protein gene expression (decreased aggrecan and type I collagen). It is also interesting that the differential response to CTS between AF cells derived from nondegenerated IVDs and those derived from degenerated IVDs occurred through the regulation of a subset of different genes, with nondegenerated AF cells altering their matrix-degrading enzyme gene expression and degener- 3392 ated AF cells altering their matrix protein gene expression. This lack of structural matrix gene regulation with CTS in AF cells derived from nondegenerated IVDs has been reported in 2 previous studies, where strains of 1% or 5% and strains of 6%, respectively, resulted in no change in the relative gene expression of aggrecan and biglycan and in the relative gene expression of aggrecan and types I and II collagen, respectively (26,27). Although the study by Rannou et al (27) found no changes in metalloproteinase (MMP-1 and MMP-3) expression, that study and the study by Sowa et al (26) suggested that CTS modulates matrix regulation in AF cells derived from nondegenerated tissue by altering the expression of matrix-degrading enzyme genes as opposed to structural matrix genes. Another difference in the cellular response of AF cells derived from nondegenerated and degenerated tissue to CTS was the time taken for a change in gene expression to occur following exposure to 1.0 Hz of mechanical stimulation. Matrix-degrading enzyme gene expression was reduced at 1 hour, and matrix protein gene expression was reduced at 24 hours, following application of CTS in AF cells derived from nondegenerated and degenerated tissues, respectively. This apparent delay in the CTS-induced change in matrix gene expression in AF cells derived from degenerated IVDs could be indicative of an altered or alternative mechanotransduction pathway being in operation. Both nondegenerated and degenerated AF cells responded catabolically to 0.33 Hz of CTS; however, the response of degenerated AF cells included the regulation of additional genes (up-regulation of MMP-13 and down-regulation of ADAMTS-4), again suggesting potential differences in mechanotransduction. Although there are no other studies comparing the effects of load on AF cells derived from nondegenerated and degenerated tissues, the effect of age on loaded AF cells has been investigated. Korecki et al (20) found that AF cells from mature (18–24 months) bovine caudal discs responded anabolically, with increased expression of types I and II collagen genes, while AF cells from young (4–6 months) bovine caudal discs responded catabolically, with increased expression of MMP-3 gene. Thus, the effect of age appears to be dissimilar to the effect of degeneration on the response of mechanically loaded AF cells. This suggests that age and degeneration are separate processes that have differential effects on the response of AF cells to load. Equally, the differences observed between the studies could be due to species variation or limitations in using relatively young cows as a model for age. GILBERT ET AL The difference observed in the cellular response to CTS at a frequency of 1.0 Hz between nondegenerated and degenerated AF cells suggests that these 2 cell types may use different mechanotransduction pathways. A previous study by our group (35) showed an altered integrin mechanotransduction pathway operating in hydrostatically loaded NP cells derived from degenerated IVDs. Together, these results suggest that the mechanotransduction pathways differ in disc cells derived from both the AF and the NP regions of nondegenerated and degenerated tissues. The glycosaminoglycan (GAG) content of the AF has been shown to be decreased in the IVDs of patients with low back pain (36), as well as with increasing age of the tissue (37), with loss of the negatively charged GAGs from the AF matrix potentially leading to a hypoosmotic cellular environment. Furthermore, changes in the osmotic pressure have been shown to influence mechanical load–regulated gene expression in AF and NP cells (25). The mechanotransduction pathway that is active in AF cells derived from degenerated IVDs may therefore be altered in order to maintain mechanical load–regulated gene expression during hypo-osmotic conditions, leading to the altered response observed between nondegenerated and degenerated AF cells when exposed to 1.0 Hz of CTS under physiologically nondegenerated osmotic conditions. The amount of strain acting on disc cells during physiologic loading is likely to differ with degeneration due to loss of matrix proteins, leading to diminished elasticity and increased stiffness of the disc (38,39). This difference between the native mechanical environment of cells derived from nondegenerated and degenerated IVDs could lead to differences in the threshold of mechanical force needed to elicit a cellular response. Interestingly, this process of altered mechanical threshold has been shown in tenocytes, where the amount of strain needed to cause a change in gene expression was increased when tenocytes were initially treated with tension deprivation (40). If the AF cells derived from degenerated IVDs had altered their mechanical force threshold due to loss of tension in their native environment, then 10% strain, a physiologic magnitude capable of reducing matrix catabolism in nondegenerated AF cells, could potentially be interpreted as a nonphysiologic understrain, leading to decreased matrix anabolism. In conclusion, our data are the first to show that the response of human AF cells to CTS, in terms of regulation of matrix protein and matrix-degrading enzyme gene expression, is frequency-dependent. This RESPONSE OF HUMAN AF CELLS TO CYCLIC TENSILE STRAIN finding supports the concept that physiologic mechanical loads are important for matrix homeostasis and promotion of matrix anabolism in healthy disc cells. The frequency of load has been shown to be an important factor in determining cell metabolic activity, and our data suggest that frequencies below physiologically “normal” levels could lead to degradation of the IVD matrix. Interestingly, AF cells derived from degenerated tissue respond differently to CTS as compared with those derived from nondegenerated tissue. Specifically, our data suggest that physiologic mechanical loads could be detrimental to disc cell matrix homeostasis in degenerated IVDs, potentially leading to the progression of degenerative disc disease. Although this study is limited by the in vitro nature of the loading culture system, removing the AF cells from their native environment excludes many uncharacterized extrinsic factors, allowing the effect of cell deformation to be considered in isolation. Such a system will allow further investigation of the response of IVD cells to CTS and will help elucidate the mechanotransduction pathways operating in cells derived from nondegenerated and degenerated IVDs. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Millward-Sadler had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Gilbert, Hoyland, Millward-Sadler. Acquisition of data. Gilbert. Analysis and interpretation of data. Gilbert, Hoyland, MillwardSadler. REFERENCES 1. Adams MA, McNally DS, Dolan P. ‘Stress’ distributions inside intervertebral discs: the effects of age and degeneration. J Bone Joint Surg Br 1996;78:965–72. 2. Shirazi-Adl SA, Shrivastava SC, Ahmed AM. Stress analysis of the lumbar disc-body unit in compression: a three-dimensional nonlinear finite element study. Spine 1984;9:120–34. 3. Wilke HJ, Neef P, Caimi M, Hoogland T, Claes LE. New in vivo measurements of pressures in the intervertebral disc in daily life. Spine 1999;24:755–62. 4. O’Connell GD, Johannessen W, Vresilovic EJ, Elliott DM. Human internal disc strains in axial compression measured noninvasively using magnetic resonance imaging. Spine 2007;32:2860–8. 5. Wenger KH, Woods JA, Holecek A, Eckstein EC, Robertson JT, Hasty KA. Matrix remodeling expression in anulus cells subjected to increased compressive load. Spine 2005;30:1122–6. 6. Kasra M, Goel V, Martin J, Wang ST, Choi W, Buckwalter J. Effect of dynamic hydrostatic pressure on rabbit intervertebral disc cells. J Orthop Res 2003;21:597–603. 7. MacLean JJ, Lee CR, Alini M, Iatridis JC. Anabolic and catabolic mRNA levels of the intervertebral disc vary with the magnitude 3393 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. and frequency of in vivo dynamic compression. J Orthop Res 2004;22:1193–200. Hutton WC, Elmer WA, Boden SD, Hyon S, Toribatake Y, Tomita K, et al. The effect of hydrostatic pressure on intervertebral disc metabolism. Spine 1999;24:1507–15. Wang DL, Jiang SD, Dai LY. Biologic response of the intervertebral disc to static and dynamic compression in vitro. Spine 2007;32:2521–8. Wuertz K, Godburn K, MacLean JJ, Barbir A, Stinnett Donnelly J, Roughley PJ, et al. In vivo remodeling of intervertebral discs in response to short- and long-term dynamic compression. J Orthop Res 2009;27:1235–42. Matsumoto T, Kawakami M, Kuribayashi K, Takenaka T, Tamaki T. Cyclic mechanical stretch stress increases the growth rate and collagen synthesis of nucleus pulposus cells in vitro. Spine 1999; 24:315–9. Hutton WC, Elmer WA, Boden SD, Hyon S, Toribatake Y, Tomita K, et al. The effect of hydrostatic pressure on intervertebral disc metabolism. Spine 1999;24:1507–15. MacLean JJ, Lee CR, Grad S, Ito K, Alini M, Iatridis JC. Effects of immobilization and dynamic compression on intervertebral disc cell gene expression in vivo. Spine 2003;28:973–81. Court C, Colliou OK, Chin JR, Liebenberg E, Bradford DS, Lotz JC. The effect of static in vivo bending on the murine intervertebral disc. Spine J 2001;1:239–45. Neidlinger-Wilke C, Wurtz K, Urban JP, Borm W, Arand M, Ignatius A, et al. Regulation of gene expression in intervertebral disc cells by low and high hydrostatic pressure. Eur Spine J 2006;15 Suppl 3:S372–8. Lotz JC, Chin JR. Intervertebral disc cell death is dependent on the magnitude and duration of spinal loading. Spine 2000;25: 1477–83. MacLean JJ, Lee CR, Alini M, Iatridis JC. The effects of short-term load duration on anabolic and catabolic gene expression in the rat tail intervertebral disc. J Orthop Res 2005;23: 1120–7. Kasra M, Merryman WD, Loveless KN, Goel VK, Martin JD, Buckwalter JA. Frequency response of pig intervertebral disc cells subjected to dynamic hydrostatic pressure. J Orthop Res 2006;24: 1967–73. Hutton WC, Elmer WA, Bryce LM, Kozlowska EE, Boden SD, Kozlowski M. Do the intervertebral disc cells respond to different levels of hydrostatic pressure? Clin Biomech (Bristol, Avon) 2001;16:728–34. Korecki CL, Kuo CK, Tuan RS, Iatridis JC. Intervertebral disc cell response to dynamic compression is age and frequency dependent. J Orthop Res 2009;27:800–6. Ishihara H, McNally DS, Urban JP, Hall AC. Effects of hydrostatic pressure on matrix synthesis in different regions of the intervertebral disk. J Appl Physiol 1996;80:839–46. Alini M, Eisenstein SM, Ito K, Little C, Kettler AA, Masuda K, et al. Are animal models useful for studying human disc disorders/ degeneration? Eur Spine J 2008;17:2–19. Handa T, Ishihara H, Ohshima H, Osada R, Tsuji H, Obata K. Effects of hydrostatic pressure on matrix synthesis and matrix metalloproteinase production in the human lumbar intervertebral disc. Spine 1997;22:1085–91. Le Maitre CL, Frain J, Fotheringham AP, Freemont AJ, Hoyland JA. Human cells derived from degenerate intervertebral discs respond differently to those derived from non-degenerate intervertebral discs following application of dynamic hydrostatic pressure. Biorheology 2008;45:563–75. Wuertz K, Urban JP, Klasen J, Ignatius A, Wilke HJ, Claes L, et al. Influence of extracellular osmolarity and mechanical stimulation on gene expression of intervertebral disc cells. J Orthop Res 2007;25:1513–22. 3394 26. Sowa G, Agarwal S. Cyclic tensile stress exerts a protective effect on intervertebral disc cells. Am J Phys Med Rehabil 2008;87:537–44. 27. Rannou F, Richette P, Benallaoua M, Francois M, Genries V, Korwin-Zmijowska C, et al. Cyclic tensile stretch modulates proteoglycan production by intervertebral disc annulus fibrosus cells through production of nitrite oxide. J Cell Biochem 2003;90: 148–57. 28. Sive JI, Baird P, Jeziorsk M, Watkins A, Hoyland JA, Freemont AJ. Expression of chondrocyte markers by cells of normal and degenerate intervertebral discs. Mol Pathol 2002;55:91–7. 29. Vande Geest JP, Di Martino ES, Vorp DA. An analysis of the complete strain field within Flexercell membranes. J Biomech 2004;37:1923–8. 30. Liu B, Qu MJ, Qin KR, Li H, Li ZK, Shen BR, et al. Role of cyclic strain frequency in regulating the alignment of vascular smooth muscle cells in vitro. Biophys J 2008;94:1497–507. 31. Winter DA. Biomechanical motor patterns in normal walking. J Mot Behav 1983;15:302–30. 32. Le Maitre CL, Freemont AJ, Hoyland JA. The role of interleukin-1 in the pathogenesis of human intervertebral disc degeneration. Arthritis Res Ther 2005;7:R732–45. 33. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2⫺⌬⌬CT method. Methods 2001;25:402–8. 34. Salter DM, Robb JE, Wright MO. Electrophysiological responses of human bone cells to mechanical stimulation: evidence for GILBERT ET AL 35. 36. 37. 38. 39. 40. specific integrin function in mechanotransduction. J Bone Miner Res 1997;12:1133–41. Le Maitre CL, Frain J, Millward-Sadler J, Fotheringham AP, Freemont AJ, Hoyland JA. Altered integrin mechanotransduction in human nucleus pulposus cells derived from degenerated discs. Arthritis Rheum 2009;60:460–9. Stevens RL, Ryvar R, Robertson WR, O’Brien JP, Beard HK. Biological changes in the annulus fibrosus in patients with lowback pain. Spine 1982;7:223–33. Scott JE, Bosworth TR, Cribb AM, Taylor JR. The chemical morphology of age-related changes in human intervertebral disc glycosaminoglycans from cervical, thoracic and lumbar nucleus pulposus and annulus fibrosus. J Anat 1994;184:73–82. Iatridis JC, Setton LA, Weidenbaum M, Mow VC. Alterations in the mechanical behavior of the human lumbar nucleus pulposus with degeneration and aging. J Orthop Res 1997;15: 318–22. Iatridis JC, Setton LA, Foster RJ, Rawlins BA, Weidenbaum M, Mow VC. Degeneration affects the anisotropic and nonlinear behaviors of human anulus fibrosus in compression. J Biomech 1998;31:535–44. Arnoczky SP, Lavagnino M, Egerbacher M, Caballero O, Gardner K, Shender MA. Loss of homeostatic strain alters mechanostat “set point” of tendon cells in vitro. Clin Orthop Relat Res 2008;466:1583–91.