ARTHRITIS & RHEUMATISM Vol. 60, No. 2, February 2009, pp 460469 DOI 10.1002/art.

24248 2009, American College of Rheumatology

Altered Integrin Mechanotransduction in Human Nucleus Pulposus Cells Derived From Degenerated Discs
Christine Lyn Le Maitre,1 Jennie Frain,2 Jane Millward-Sadler,2 Andrew P. Fotheringham,2 Anthony John Freemont,2 and Judith Alison Hoyland2
Objective. Several studies have demonstrated biologic responses of intervertebral disc (IVD) cells to loading, although the mechanotransduction pathways have not been elucidated. In articular chondrocytes, which have a phenotype similar to that of IVD cells, a number of mechanoreceptors have been identified, with 51 integrin acting as a predominant mechanoreceptor. The purpose of this study was to investigate the role of integrin signaling in IVD cells during mechanical stimulation and to determine whether RGD integrins are involved. Methods. Human nucleus pulposus (NP) cells derived from nondegenerated and degenerated discs were subjected to dynamic compressive loading in the presence of an RGD inhibitory peptide. Expression of the 51 heterodimer in IVD tissue was examined by immunohistochemistry and possible alternative mechanoreceptors by real-time quantitative polymerase chain reaction. Results. Aggrecan gene expression was decreased following loading of NP cells from nondegenerated and degenerated discs. This response was inhibited by treatment with an RGD peptide in cells from nondegenerated, but not degenerated, IVDs. Immunohistochemistry demonstrated that expression of the 5 1 heterodimer was unaltered in degenerated IVD tissue as compared with normal IVD tissue. Conclusion. Our results indicate that the mech1 Christine Lyn Le Maitre, PhD: Sheffield Hallam University, Sheffield, UK; 2Jennie Frain, PhD, Jane Millward-Sadler, PhD, Andrew P. Fotheringham, MSc, Anthony John Freemont, MD, FRCP, FRC Path, Judith Alison Hoyland, PhD: University of Manchester, Manchester, UK. Address correspondence and reprint requests to Judith Alison Hoyland, PhD, Tissue Injury and Repair Group, School of Clinical and Laboratory Sciences, Faculty of Medical and Human Sciences, The University of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT, UK. E-mail: Judith.hoyland@manchester.ac.uk. Submitted for publication April 10, 2008; accepted in revised form October 13, 2008.

anotransduction pathways are altered in cells from degenerated IVDs. Mechanosensing in NP cells from nondegenerated discs occurs via RGD integrins, possibly via the 51 integrin, while cells from degenerated discs show a different signaling pathway that does not appear to involve RGD integrins. The intervertebral disc (IVD) is an important load-bearing structure, with daily activities exposing the IVD to dynamic oscillatory loads (1). These forces arise from a combination of gravity and muscle tension during movement and result in various types of pressure being experienced within the IVD. Abdominal and back muscles stabilize the spine and resist gravitational forces to allow spinal movement, but by doing so, they exert additional compressive forces on the lumbar discs (2). These combined forces result in a wide variety of mechanical stimuli, including compression, hydrostatic pressure, shear stress, torsion and flexion of the spine, electrokinetic effects, and volumetric changes (38). The biologic response to these stimuli varies according to the type of cells in the IVD, the nature of loading experienced, and the duration of the load. Several studies have demonstrated biologic responses to a number of different types of load in IVD cells, with many focusing on the effects of compressive and hydrostatic loading (3,4,914). However, the mechanotransduction pathways have not been elucidated. Many cell types detect movement or forces via a transmembrane receptor, which triggers subsequent remodeling of the cytoskeleton, leading to changes in cell metabolism via numerous downstream events, such as posttranslational modifications (15). In articular chondrocytes, which share phenotype similar to that of IVD cells, a number of transmembrane receptors have been identified that can fulfill this role, including CD44, anchorin II, and integrin receptors (16,17). Integrins are a family of transmembrane glycosy460

INTEGRIN MECHANOTRANSDUCTION IN IVD CELLS

461

lated proteins that mediate dynamic interactions between the cell surface and the immediate surrounding environment (18). Structurally, they are a large family of obligate heterodimers comprising 2 noncovalently associated subunits, referred to as the and subunits, which combine to form 24 distinct integrin receptors. The integrins can be grouped into subfamilies based on ligand interactions. One such family is the RGD integrins, where RGD (arginine-glycine-aspartic acid) is the ligand motif that these integrins recognize and interact with in extracellular matrix (ECM) molecules such as fibronectin. The RGD integrins are those containing the 5, 8, v, and IIb subunits (18) and include the classic fibronectin receptor 51. Integrins (particularly, the 51 heterodimer) have been identified in articular chondrocytes (19) and are responsible for the initial mechanotransduction response within these cells (17,2022). Stimulation of human articular chondrocytes in the presence of an inhibitory RGD peptide has been shown to block the usual molecular response observed following loading, in that there was an increase in cell proliferation and proteoglycan synthesis (23), membrane hyperpolarization (24), and increased aggrecan and decreased matrix metalloproteinase 3 (MMP-3) gene expression (25). Interestingly, the 51 integrinassociated pathway has been shown to activate an altered intracellular signaling pathway in osteoarthritic (OA) chondrocytes following mechanical stimulation (24), which raises the possibility of altered mechanotransduction mechanisms between healthy and diseased cells. Consistent with this is our recent study showing that the application of hydrostatic load to IVD cells resulted in an anabolic response in cells derived from nondegenerated discs, with no response seen in cells derived from degenerated discs (14), which suggests that mechanotransduction pathways may also be altered in degenerated IVDs. Few studies have investigated integrin expression in the IVD. Yu et al (26) demonstrated that expression of the 5 integrin subunit decreased following application of cyclic hydrostatic pressure to porcine IVD cells, suggesting this may be part of a mechanotransduction pathway. However, no studies have investigated the expression of the 51 heterodimer or its functionality in human IVDs, and few studies have investigated the expression of the individual integrin subunits in human IVDs. Nettles et al (27) demonstrated the production of 5 and 1 individual subunits in degenerated human anulus fibrosus (AF) tissue (albeit only 3 samples) and in porcine nucleus pulposus (NP) and AF. More recently, Gilchrist et al (28) investigated porcine IVD cell attach-

ment on various matrices in the presence of integrin subunitblocking antibodies. They showed inhibition of fibronectin binding by blockade of the 1 subunit and suggested that this was due to blockade of the 51 heterodimer (the classic fibronectin receptor). However, no studies have investigated the role of integrin signaling in IVD cells during mechanical stimulation, and to date, it is unclear whether RGD integrins are involved in the IVD response to mechanical stimulation. In order to assess this, we subjected human NP cells derived from nondegenerated and degenerated discs to dynamic compressive loading in the presence of an RGD inhibitory peptide and then investigated the expression of 51 heterodimer and integrin subunits by immunohistochemistry, together with the possible expression of alternative mechanoreceptors by polymerase chain reaction (PCR). PATIENTS AND METHODS
Tissue selection and grading of IVDs. Human IVD tissue was obtained either at the time of surgery or at postmortem examination. Informed consent of the patient or the relatives of the deceased was obtained. The study was approved by the local research ethics committee. Postmortem IVD tissue. Nine IVDs were recovered from 6 subjects within 18 hours of death (Table 1). These consisted of full-thickness wedges of IVD of 120 of arc, which were removed anteriorly, allowing well-orientated blocks of tissue for histologic assessment. Degenerated IVD tissue. Nine patients (n 15 samples) were selected based on the presence of magnetic resonance imagingdiagnosed degeneration and progression to anterior resection, either for spinal fusion or for IVD replacement surgery because of chronic low back pain. Some patients underwent fusion at more than one level because of instability. Patients with a history of sciatica sufficient to warrant their seeking medical opinion or patients experiencing classic sciatica were excluded. Tissue preparation. A block of tissue, incorporating the AF and the NP in continuity, was fixed in 10% neutral buffered formalin, decalcified in EDTA, and processed to paraffin wax. Hematoxylin and eosinstained sections were analyzed for the degree of morphologic degeneration (29). A score of 03 indicates a histologically normal (nondegenerated) IVD, a score of 47 indicates evidence of moderate degeneration, and a score of 812 indicates severe degeneration. For purposes of the loading and gene expression studies, samples were classified as nondegenerated or degenerated, but these 3 categories were used for immunohistochemical analysis. Creation of IVD constructs. Normal human cadaveric IVD tissue (n 3; predetermined using power calculations) was obtained from the spines of cadavers: one from an 18-year-old man with L3/4 disc grade 1, another from a

462

LE MAITRE ET AL

Table 1. Characteristics of the paraffin-embedded intervertebral disc samples used for integrin immunohistochemistry Sample 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Rib control Sample source Postmortem Postmortem Postmortem Postmortem Postmortem Postmortem Postmortem Postmortem Surgical Postmortem Surgical Surgical Surgical Surgical Surgical Surgical Surgical Surgical Surgical Surgical Surgical Surgical Surgical Surgical Postmortem Age/sex of subject 37/M 47/M 47/M 61/M 47/M 37/M 37/M 37/M / 61/M 75/M / 75/F 46/M 58/M /F 58/M / 58/M 22/M 73/F 46/M /F /M 47/M Histology grade* 1 1 1 2 2 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 10 11 12 12 NA

* Histologic scoring was performed as previously described (29), based on the degree of degeneration: grades 03 normal (nondegenerated), grades 47 moderate degeneration, and grades 812 severe degeneration. NA not applicable.

47-year-old man with L4/5 disc grade 2, and the third from a 79-year-old man with L5/S1 disc grade 2. Degenerated human IVD tissue (n 3) was obtained from patients undergoing spinal surgery: one from a 75-year-old man with L5/S1 disc grade 7, another from a 35-year-old woman with L5/S1 disc grade 10, and the third from a 43-year-old man with L4/5 disc grade 10. NP tissue was finely minced and digested with 300350 PUK/ml of Pronase (Calbiochem, Nottingham, UK) in Dulbeccos modified Eagles medium (DMEM)Hams F-12 for 1 hour at 37C and then washed in DMEMHams F-12. NP cells were isolated in 0.25% type II collagenase (Invitrogen, Paisley, UK) and 0.01% hyaluronidase (Sigma, Poole, UK) for 4 hours at 37C. Cells were then expanded in monolayer and used at early passage (2). Cells were trypsinized and resuspended in 1.2% medium-viscosity sodium alginate (Sigma) in 0.15M NaCl at a density of 1 106 cells/ml. Sixty microliters of cell/alginate suspension was added to the inner (5 mm diameter) foam ring of BioFlex 6-well compression plates (Flexcell International, from Dunn Labortechnik, Asbach, Germany). Constructs were then formed by polymerization with 200 mM CaCl2 to yield 3.1-mm high samples. CaCl2 was removed, constructs were washed twice with 0.15M NaCl and medium, 2 ml of complete medium was added, and the constructs were incubated for 2 weeks at 37C in a humidified atmosphere containing 5% CO2 to allow cell redifferentiation (30) prior to loading.

Inhibitory peptide treatments. One hour prior to loading, constructs were left untreated or were treated with either 50 g/ml of GRGDSP (Calbiochem), a peptide that competes for integrin ligand RGD binding sites or 50 g/ml of GRADSP (Calbiochem), a control peptide for GRGDSP. All treatments were performed in triplicate. Application of compressive load. BioFlex compression plates containing the alginate/NP cell constructs were loaded onto a Flexercell FX-4000C compression loading system (Flexcell International), 300 l of medium was applied per sample, which ensured no leakage or drying out of samples during loading, and compressive load was applied for 2 hours at 0.350.95 MPa, where the applied stress was oscillated sinusoidally between these 2 limits at a frequency of 1 Hz in the presence or absence of peptide inhibitors. The Flexercell FX-4000C functions via the application of compressed gas to the base of the BioFlex compression plates (which have a flexible membrane), resulting in compression of the sample (contained within the foam circle) against the fixed platen at controlled rates of pressure (31). Control plates were placed in the incubator at 37C without application of load. Samples were harvested 1 hour after loading to assess cell viability and aggrecan gene expression. Cell viability. Cell viability was assessed using carboxyfluorescein diacetate succinimidyl ester (CFSE-DA) and propidium iodide (PI) staining, where CFSE-DA stains viable cells green, and PI stains nonviable cells red. Alginate constructs were incubated at 37C for 10 minutes in 1 ml of phosphate buffered saline containing 10 l of 6 mM PI and 2 l of 5 mM CFSE-DA (both from Sigma). Cells were visualized with an inverted microscope equipped with a dual-wavelength bypass filter (450 nm for CFSE-DA/520 nm for PI). Emission of green fluorescence and red fluorescence indicated viable cells and nonviable cells, respectively. Total cell numbers were counted manually and tallied as viable and nonviable cells to allow for calculation of the percentage of viability. RNA extraction, complementary DNA (cDNA) synthesis, and real-time quantitative PCR for aggrecan gene expression. Prior to TRIzol extraction, alginate constructs were washed in 0.15M NaCl and dissolved in dissolving buffer (55 mM sodium citrate, 30 mM EDTA, 0.15M NaCl, pH 6) at 37C for 30 minutes. RNA was then extracted from triplicate constructs using TRIzol reagent (Gibco, Paisley, UK) according to the manufacturers instructions. Reverse transcription was performed as described previously (30), and real-time quantitative PCR was used to investigate the effects of dynamic compressive loading on aggrecan gene expression. Realtime quantitative PCR was performed as described previously (30), using 5-TCGAGGACAGCGAGGCC-3 as the forward primer, 5 -ATGGAACACGATGCCTTTCACCACGA-3 (with FAM at the 5 end and minor groove binder at the 3 end) as the probe, and 5-TCGAGGGTGTAGCGTGTAGAGA-3 as the reverse primer (Applied Biosystems, Warrington, UK). Data were analyzed using the 2Ct method to calculate the difference in cycle thresholds and using the housekeeping gene ribosomal subunit 18S (Pre-Developed Assay Reagents; Applied Biosystems) and unloaded controls for normalization (30). Localization of 5 and 1 integrin subunits and 51 heterodimer in human IVDs. Tissue samples from 24 IVDs were selected for immunohistochemical analysis. These consisted of 6 nondegenerated (grades 03), 8 moderately degen-

INTEGRIN MECHANOTRANSDUCTION IN IVD CELLS

463

Table 2. Primer sequences, conditions, and product sizes for standard real-time polymerase chain reaction (PCR) Primer sequences Gene GAPDH Forward (533) CCCATCACCATCTTCCAGG GGCTGTTGGTGAATTCAGTG GTGCCTGCAGAAGAATATGG GGCTGTTGGTGAATTCAGTG GAGCAGCAAGGACTTTGGG CCTTATGGACCTGTCTTACTC TCCCAGTATGACACACATATTGC Reverse (335) GGCCATCCACAGTCTTCTG GCTGCAGGTAGACGTAGAC GGTAGCCTACATCGCAGGC GTCTACGTCTACCTGCAGC GGGTACACTTCAAGACCAGC GCTGAAATTCTTCAGTAACTGC CACCTTCTTCGACTGTTGAC PCR conditions Annealing at 35 cycles Annealing at 36 cycles Annealing at 42 cycles Annealing at 36 cycles Annealing at 36 cycles Annealing at 38 cycles Annealing at 38 cycles 57.4C; 57.4C; 57.4C; 57.4C; 57.4C; 57.4C; 55.1C; Product size 354 bp 300 bp 475 bp 286 bp 619 bp 589 bp 549 bp Accession number M33197 NM_181501 NM_002203 NM_002205 NM_002210 NM_002211 NM_000610

1 integrin 2 integrin 5 integrin v integrin 1 integrin

CD44

erated (grades 47), and 10 severely degenerated (grades 812) discs (Table 1). Immunohistochemistry was also used to localize the 51 heterodimer in the NP cells cultured in the alginate constructs. The immunohistochemistry protocol we followed has previously been described (30). Antibodies used were mouse monoclonal primary antibodies against human integrin 5 subunit (MAB1956; 1:100 dilution), human integrin 1 subunit (MAB1965; 1:400 dilution), and human 51 heterodimer (MAB1969; 1:600 dilution) (all from Chemicon, Southampton, UK). Negative controls in which mouse IgG (Dako, Cambridge, UK) replaced the primary antibody (at an equal protein concentration) were included. Image analysis. All slides were examined under a Leica RMDB microscope (Leica, Cambridge, UK), and images were captured using a digital camera and a Bioquant Nova image analysis system (Bioquant Image Analysis, Nashville, TN). Each section was divided into the NP, the inner AF, and the outer AF, and each was analyzed separately. Within each area, 200 cells were counted, and the number of immunopositive cells was expressed as a proportion of this. Alginate/NP constructs were analyzed for cells displaying 51 immunopositivity. Gene expression of mechanoreceptors. Tissues from 19 lumbar IVD samples were used for gene expression analysis. These consisted of 10 nondegenerated (grades 03; age range 3079 years) and 9 degenerated (grades 412; age range 2974 years) discs. RNA was extracted using TRIzol reagent, and cDNA was synthesized using BioScript RNase H reverse transcriptase (Bioline, London, UK) and random hexamers (Roche, Lewes, UK). Real-time PCR was performed to analyze expression of the integrin subunits 1, 2, 5, v, and 1 together with CD44 and the housekeeping gene GAPDH. Primers were designed using Amplify 1.2 software (available at http://engels.genetics.wisc.edu/amplify), and gene specificity was confirmed by BLAST searches of GenBank database sequences (Table 2). Twenty-five microliters of the PCR reaction product consisted of 5 l of cDNA (50 ng/l), 0.125 l of HotStar Taq DNA polymerase (Qiagen, West Sussex, UK), 0.25 l of forward primer (100 M), 0.25 l of reverse primer (100 M), 0.5 l of 40 mM dNTPs (Roche), and 2.5 l of buffer, which was brought to 25 l with sterile distilled H2O.

PCR was then performed on a PTC-200 Peltier thermal cycler (Labsystems, Epsom, UK) with an initial Taq activation step at 95C for 10 minutes, followed by 3542 cycles consisting of denaturation at 94C for 30 seconds, annealing at 55.157.4C for 30 seconds, and extension at 72C for 1 minute (Table 2). A final incubation at 72C for 10 minutes was then performed. PCR products were visualized on a 1% Tris borateEDTA weight/volume agarose gel using an ultraviolet transilluminator and the UV Grab program GeneSnap from Syngene (SLS, Manchester, UK). Statistical analysis. Loading study data. Data were determined to be nonparametric using the Shapiro-Wilke test. The Kruskal-Wallis test with Mann-Whitney U post hoc tests were used to assess the effect of dynamic compressive load on cell viability and aggrecan gene expression on nondegenerated and degenerated samples in the 3 treatment groups. Immunohistochemistry data. Data were nonparametric, and hence, Kruskal-Wallis tests with Mann-Whitney U post hoc tests were performed to compare the numbers of immunopositive cells in degenerated groups versus nondegenerated IVDs (scores 03) for each area of the IVD. In addition, Wilcoxon paired sample tests were used to compare the proportions of immunopositive cells in the different regions of the IVDs.

RESULTS Application of mechanical load to human disc cells and inhibition of the RGD integrins. All alginate constructs maintained cell viability at 80%, with no difference observed between samples subjected to dynamic compressive loading and unloaded controls and no difference with any treatment in cells derived from nondegenerated discs. A small decrease in cell viability was observed in loaded disc cells derived from degenerated IVDs compared with unloaded disc cells (P 0.05), with no difference seen between treatments (Figure 1). Application of compressive loading at 0.350.95

464

LE MAITRE ET AL

Figure 1. Percentage of viable cells following 2 hours of compressive loading (0.350.95 MPa at a frequency of 1 Hz) in untreated, control peptidetreated, and inhibitory peptidetreated cells from intervertebral discs (IVDs) obtained at surgery from patients undergoing spinal fusion or IVD replacement (degenerated) and at postmortem examination from normal subjects (nondegenerated). Values are the mean SEM. P 0.05 versus unloaded cells from degenerated discs.

MPa and a frequency of 1 Hz to NP cells derived from nondegenerated or degenerated discs significantly decreased aggrecan gene expression (P 0.05) (Figure 2). No difference in the response to loading was observed in NP cells that had been pretreated with the control peptide. However, in NP cells from nondegenerated discs pretreated with the RGD peptide, no change in aggrecan gene expression was observed following application of dynamic compressive load. In contrast, the response to dynamic compressive loading in NP cells from degenerated discs pretreated with the RGD peptide was not altered from that seen in untreated cells or

cells treated with control peptide; that is, a significant decrease in aggrecan gene expression in NP cells from degenerated discs was observed in all 3 treatment groups (P 0.05) (Figure 2). Localization of 5 and 1 integrin subunits and the 51 heterodimer in human IVDs. Immunoreactivity for 5 and 1 subunits and 51 heterodimer was observed in nondegenerated and degenerated IVDs. Positive control tissue (human rib cartilage) stained for each integrin and the heterodimer. IgG controls were negative (Figure 3). Staining was particularly prominent in the chondrocyte-like cells of the NP and the inner AF (Figure 3), with significantly lower numbers of immunopositive cells in the outer AF (P 0.05). No significant difference was observed in the number of immunopositive cells for the 5 subunit, the 1 subunit, or the 51 heterodimer identified in the NP or inner AF of nondegenerated discs as compared with IVDs with moderate or severe histologic degeneration (P 0.05 for all comparisons). Immunopositive staining for 51 heterodimer was observed in cells cultured in alginate (Figure 3). Gene expression of mechanoreceptors. Although gene expression was detected for CD44 and the integrin subunits in both nondegenerated and degenerated discs, a differential expression pattern was observed. Integrin subunits 1 and 5 and CD44 were strongly expressed in all samples, while v was moderately expressed in the

Figure 2. Aggrecan gene expression in human nucleus pulposus cells derived from intervertebral discs (IVDs) that were left untreated or were treated with control peptide or integrin inhibitory peptide and then subjected to compressive loading for 2 hours at 0.350.95 MPa at a frequency of 1 Hz. RNA was extracted 1 hour after loading. IVDs were obtained at surgery from patients undergoing spinal fusion or IVD replacement (degenerated) and at postmortem examination from normal subjects (nondegenerated). Values are the mean SEM. P 0.05 versus unloaded cells.

INTEGRIN MECHANOTRANSDUCTION IN IVD CELLS

465

Figure 3. Photomicrographs demonstrating immunohistochemical localization of integrin subunits 5 and 1 and the heterodimer 51 in the nucleus pulposus from intervertebral discs (IVDs) obtained at surgery from patients undergoing spinal fusion or IVD replacement (degenerated) and at postmortem examination from normal subjects (nondegenerated). Cells extracted from nondegenerated and degenerated IVDs cultured in alginate for 2 weeks showed immunopositivity for 51. IgG negative controls showed no staining. Bars 380 m.

majority of samples. Integrin subunits 2 and 1 were only weakly expressed in nondegenerated and degenerated NP samples, with little expression detected in AF samples (Figure 4).

DISCUSSION The intervertebral disc is an important loadbearing structure in the human body that experiences a

Figure 4. Expression of GAPDH, 1 integrin, 2 integrin, 5 integrin, v integrin, 1 integrin, and CD44 genes in nucleus pulposus (NP) and anulus fibrosus (AF) cells derived from intervertebral discs (IVDs) obtained at surgery from patients undergoing spinal fusion or IVD replacement (degenerated; n 9) and at postmortem examination from normal subjects (nondegenerated; n 10), as determined by real-time polymerase chain reaction analysis. Red X indicates unavailable sample.

466

LE MAITRE ET AL

number of mechanical forces during daily activities. Cells within the IVD respond to these stresses with a multitude of anabolic and catabolic responses. As such, understanding the mechanisms of mechanosensing within the human IVD is of paramount importance to understanding the regulation of IVD matrix homeostasis. In this study, we investigated the role of integrin signaling in mechanotransduction pathways in the IVD. Application of dynamic compressive loading resulted in a significant decrease in aggrecan gene expression in both nondegenerated and degenerated discs. Interestingly, this response was inhibited by the application of an RGD peptide in nondegenerated, but not degenerated, discs. Having determined that RGD-binding integrins play a role in the mechanotransduction response in nondegenerated IVDs, we then investigated the expression and localization of the potential mechanoreceptor 51 integrin and its individual subunits (5 and 1) within both nondegenerated and degenerated human IVD samples. Our results showed that human NP and AF cells express both the 5 and 1 integrin subunits and the 51 heterodimer and that this expression was not altered with degree of degeneration. Additionally, we showed that both degenerated and nondegenerated IVDs express the 1 integrin subunit and CD44, as determined by PCR. Neither 1 integrins nor CD44 acts through an RGD motif: 1 integrins interact with collagen, whereas CD44 is a hyaluronan receptor, but either could act as potential alternative mechanoreceptors in degenerated IVDs. In this study, cells were subjected to dynamic compressive loading for 2 hours at 0.350.95 MPa at a frequency of 1 Hz, which Wilke et al (7) demonstrated to be the physiologic loading pattern experienced during jogging. In vivo, the IVD is routinely subjected to compressive loads that can exceed 2 MPa and are affected by posture and muscle activity (32). The directional compressive forces that are the result of spinal movement will also lead to fluid flow and changes in hydrostatic pressures within the IVD, all of which will have an influence on IVD cellular metabolism. With increasing degeneration, there is a decrease in hydration of the IVD, which is likely to result in a change in the type and magnitude of forces that the individual cells experience. We previously investigated the effect of hydrostatic loading on NP cells isolated from human IVDs (14). Hydrostatic loading is a major force experienced in the IVD, particularly in the nondegenerated disc, which behaves mechanically like a fluid. In the present

study, however, we focused on the application of compressive load to cells derived from nondegenerated and degenerated NP IVDs. Compressive loading is a force that is regularly experienced by the IVD. In addition, there is a growing number of studies investigating the effect of compression on the different regions of the IVD and showing that the duration and amount of loading have an effect on proteoglycan/aggrecan expression (10,11,3335). The present study showed that the application of dynamic compressive loading to NP cells cultured in alginate resulted in decreased aggrecan gene expression in cells from both nondegenerated and degenerated discs. Previous studies investigating the application of compressive loading to animal IVD cells and tissues have found both increases and decreases in aggrecan gene expression, depending on the method of application, the magnitude, frequency, duration, and the period of rest following loading prior to assessment (4,10,11,13,34,3638). The present study is the first to apply compressive loading to human NP cells in a 3-dimensional matrix. In contrast to our findings, compressive loading of animal IVD tissue (both in vivo and in vitro) using pressures similar to those used in our study has been shown to result in increased aggrecan gene expression (10,11,34,3638). However, a number of studies to date have shown differential responses, depending on the frequency (1,10,39,40) and duration (41) of the applied load as well as on the time of sample harvest following loading (37), and this may explain the different responses to a similar magnitude of loading that have been observed between these studies. Indeed, Maclean et al (10) showed with rat tail compression studies that compressive loading at 1 MPa at a frequency of 1 Hz resulted in a greater up-regulation of catabolic gene expression than matrix gene expression, while a lower frequency resulted in a greater anabolic gene response. Additionally, the majority of studies investigating compressive loading have been in vivo studies, where the forces actually perceived by the cells, particularly in the healthy NP, may be different, rather than in vitro studies using isolated cells cultured in alginate. It is therefore possible that the deformation experienced by the cells in our study was higher than that observed in cells protected by the extensive ECM in vivo, thus giving rise to the catabolic response. Furthermore, since these cells were embedded in a 3-dimensional matrix, they would not have experienced any hydrostatic pressure, whereas they would have experienced hydrostatic pressure in vivo. It is also of note that our study is the first to

INTEGRIN MECHANOTRANSDUCTION IN IVD CELLS

467

investigate the effects of compressive loading on human cells, which could also explain the observed differences. Interestingly, the response of aggrecan gene expression following compressive loading of NP cells in alginate was inhibited in NP cells from nondegenerated discs by pretreatment with the inhibitory peptide (GRGDSP) but not the control peptide (GRADSP). Integrins containing 5, v, IIb, and 8 subunits are inhibited by RGD inhibitory peptides, which suggests that in NP cells derived from nondegenerated discs, the RGD integrins act as mechanoreceptors when cells embedded in alginate are subjected to compressive loading. RGD peptide inhibitors have previously been used to investigate the role of integrins in mechanotransduction pathways in articular chondrocytes (24,25), where the stimulation of aggrecan gene expression and the inhibition of MMP-3 gene expression observed following application of cyclic stretch at 0.1 MPa at a frequency of 0.33 Hz for 20 minutes was inhibited by pretreatment with RGD peptides (25). In articular chondrocytes, the RGD integrin receptor proposed to be the main mechanoreceptor is the 51 heterodimer (17,20 24). The expression of 51 by NP cells in vivo and in vitro following culture in alginate for 2 weeks would suggest that this could possibly be the integrin involved in the mechanotransduction response in cells from nondegenerated IVDs. A number of studies have investigated the effect of alginate culture of chondrocytes on matrix deposition and have shown that within 2 weeks, cells have synthesized components of both pericellular and territorial matrix (42,43). Thus, it is likely that the 51 integrin is operating via interaction with the newly synthesized pericellular and territorial matrix produced in this culture system. It is noteworthy that in NP cells derived from degenerated discs, pretreatment with the RGD peptide failed to inhibit their response to compressive loading, and a significant decrease in aggrecan gene expression was still observed following application of load. This suggests that in degenerated discs, an alternative mechanotransduction pathway may be involved that replaces the RGD mechanoreceptor signaling seen in nondegenerated discs. In cells derived from OA cartilage, the intracellular signaling that is activated following mechanical stimulation is altered as compared with normal pathways, but the cells still sense the mechanical signal through the 51 integrin (24). Our findings suggest that in the degenerated IVD, this is not the case. We have shown that although degenerated disc cells express the 51 integrin, it does not appear to be involved in the mechanotransduction pathway in these cells. The

51 integrin is also involved in the extracellular matrix interaction with fibronectin; thus, the expression of the 51 integrin in degenerated discs may play a greater role in ECM interactions than mechanotransduction. Gilchrist et al (28) demonstrated that binding to fibronectin in porcine IVDs was blocked by inhibition of the 1 integrin subunit, suggesting expression of the fibronectin receptor (51 integrin) in these disc cells. Fibronectin expression has been shown to be increased during disc degeneration (44) and, thus, may affect the availability of the 51 integrin for mechanosensing and/or may affect integrin-mediated signaling, as has been shown in OA chondrocytes with other integrin ECM relationships (45). Our findings indicate that IVD cells in degenerated discs may be using a different receptor to sense and respond to mechanical signaling, since these disc cells retained their response to mechanical stimuli in the presence of the RGD peptide. This mechanosensing could occur through a non-RGD integrin, such as 1, or a nonintegrin transmembrane molecule, such as CD44, both of which we have shown to be strongly expressed by human IVD cells. In OA, the increased production of mediators of inflammation, including interleukin-1, tumor necrosis factor , and nitric oxide, have been postulated to affect integrin-mediated signaling, either by modulating chondrocyte integrin expression (46) or by regulating intracellular signaling (47). In the degenerated IVD, an increase in these mediators of inflammation, particularly interleukin-1, is also seen (30,48), and mechanical loading has been shown to initiate signaling pathways that involve calcium (49) and nitric oxide (50). It is possible that these mediators could be involved in the altered mechanotransduction seen in our study, as has previously been reported for OA chondrocytes (24). Interestingly, we have recently shown that the application of 0.81.7 MPa of hydrostatic pressure at a frequency of 0.5 Hz for 2 hours resulted in a differential response in NP cells derived from nondegenerated versus degenerated discs, where degenerated discs failed to respond to this loading regimen (14). Together with the data presented herein, this suggests that different mechanotransduction pathways may be involved in sensing different types of mechanical load and that these pathways are altered during disc degeneration. In conclusion, we have demonstrated that when subjected to compressive loading, cells isolated from nondegenerated and degenerated human discs recognize mechanical stimulation through different mechanoreceptors. The inhibition of the aggrecan gene response to mechanical loading by RGD peptides in cells from

468

LE MAITRE ET AL

nondegenerated discs suggests that these cells recognize the mechanical stimulus through RGD integrins in a manner similar to that of articular chondrocytes. Expression of the 51 heterodimer would suggest that this is a likely candidate as a mechanoreceptor in these NP cells. The 51 integrin is also expressed in degenerated IVDs, with no change in expression observed during disc degeneration. However, the presence of RGD peptides did not inhibit the response of these cells to compressive loading, suggesting that cells in the degenerated discs do not use the 51 integrin or other RGD integrins as mechanoreceptors during this loading stimulus. Thus, we have demonstrated differences in the mechanotransduction pathways of nondegenerated and degenerated IVDs. Further elucidation of the signaling events activated by mechanical stimuli in IVD cells derived from nondegenerated and degenerated discs will lead to a better understanding of how IVD matrix homeostasis is maintained by mechanical stimuli in health and disease.
AUTHOR CONTRIBUTIONS
Professor Hoyland 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 design. Le Maitre, Frain, Hoyland. Acquisition of data. Le Maitre, Frain, Fotheringham. Analysis and interpretation of data. Le Maitre, Frain, MillwardSadler, Fotheringham, Freemont, Hoyland. Manuscript preparation. Le Maitre, Millward-Sadler, Freemont, Hoyland. Statistical analysis. Le Maitre, Hoyland.

10.

11.

12.

13.

14.

15.

16. 17.

18. 19.

20.

REFERENCES
1. 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;21:597603. 2. Adams M, Bogduk N, Burton K, Dolan P. The biomechanics of back pain. London: Churchill Livingstone; 2002. 3. Iatridis JC, Mente PL, Stokes IA, Aronsson DD, Alini M. Compression-induced changes in intervertebral disc properties in a rat tail model. Spine 1999;24:9961002. 4. Lotz JC, Colliou OK, Chin JR, Duncan NA, Liebenberg E. Compression-induced degeneration of the intervertebral disc: an in vivo mouse model and finite-element study. Spine 1998;23: 2493506. 5. Lotz JC, Chin JR. Intervertebral disc cell death is dependent on the magnitude and duration of spinal loading. Spine 2000;25: 147783. 6. Li H, Wang Z. Intervertebral disc biomechanical analysis using the finite element modeling based on medical images. Comput Med Imaging Graph 2006;30:36370. 7. 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;24:75562. 8. Sato K, Kikuchi S, Yonezawa T. In vivo intradiscal pressure measurement in healthy individuals and in patients with ongoing back problems. Spine 1999;24:246874. 9. Hsieh AH, Lotz JC. Prolonged spinal loading induces matrix

21.

22.

23. 24.

25.

26.

27.

metalloproteinase-2 activation in intervertebral discs. Spine 2003; 28:17818. Maclean JJ, Lee CR, Alini M, Iatridis JC. Anabolic and catabolic mRNA levels of the intervertebral disc vary with the magnitude and frequency of in vivo dynamic compression. J Orthop Res 2004;22:1193200. Maclean JJ, Lee CR, Alini M, Iatridis JC. The effects of shortterm load duration on anabolic and catabolic gene expression in the rat tail intervertebral disc. J Orthop Res 2005;23:11207. Iatridis JC, Maclean JJ, Roughley PJ, Alini M. Effects of mechanical loading on intervertebral disc metabolism in vivo. J Bone Joint Surg Am 2006;88 Suppl 2:416. Hutton WC, Toribatake Y, Elmer WA, Ganey TM, Tomita K, Whitesides TE. The effect of compressive force applied to the intervertebral disc in vivo: a study of proteoglycans and collagen. Spine 1998;23:252437. 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;45:56375. Setton LA, Chen J. Mechanobiology of the intervertebral disc and relevance to disc degeneration. J Bone Joint Surg Am 2006;88 Suppl 2:527. Loeser RF. Chondrocyte integrin expression and function. Biorheology 2000;37:10916. Millward-Sadler SJ, Salter DM. Integrin-dependent signal cascades in chondrocyte mechanotransduction. Ann Biomed Eng 2004;32:43546. Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell 2002;110:67387. Loeser RF, Sadiev S, Tan L, Goldring MB. Integrin expression by primary and immortalized human chondrocytes: evidence of a differential role for 11 and 21 integrins in mediating chondrocyte adhesion to types II and VI collagen. Osteoarthritis Cartilage 2000;8:96105. Millward-Sadler SJ, Wright MO, Lee H, Nishida K, Caldwell H, Nuki G, et al. Integrin-regulated secretion of interleukin 4: a novel pathway of mechanotransduction in human articular chondrocytes. J Cell Biol 1999;145:1839. Millward-Sadler SJ, Khan NS, Bracher MG, Wright MO, Salter DM. Roles for the interleukin-4 receptor and associated JAK/ STAT proteins in human articular chondrocyte mechanotransduction. Osteoarthritis Cartilage 2006;14:9911001. Chowdhury TT, Salter DM, Bader DL, Lee DA. Integrin-mediated mechanotransduction processes in TGF-stimulated monolayer-expanded chondrocytes. Biochem Biophys Res Commun 2004;318:87381. Chowdhury TT, Appleby RN, Salter DM, Bader DA, Lee DA. Integrin-mediated mechanotransduction in IL-1 stimulated chondrocytes. Biomech Model Mechanobiol 2006;5:192201. Millward-Sadler SJ, Wright MO, Lee H, Caldwell H, Nuki G, Salter DM. Altered electrophysiological responses to mechanical stimulation and abnormal signalling through 51 integrin in chondrocytes from osteoarthritic cartilage. Osteoarthritis Cartilage 2000;8:2728. Millward-Sadler SJ, Wright MO, Davies LW, Nuki G, Salter DM. Mechanotransduction via integrins and interleukin-4 results in altered aggrecan and matrix metalloproteinase 3 gene expression in normal, but not osteoarthritic, human articular chondrocytes. Arthritis Rheum 2000;43:20919. Yu SJ, Qiu GX, Burton Y, Sandra R, Cari W, Albert Y. Expression of integrin 5 and actin in the cells of intervertebral disc under cyclic hydrostatic pressure in vitro. Zhonghua Wai Ke Za Zhi 2005;43:16058. In Chinese. Nettles DL, Richardson WJ, Setton LA. Integrin expression in cells of the intervertebral disc. J Anat 2004;204:51520.

INTEGRIN MECHANOTRANSDUCTION IN IVD CELLS

469

28. Gilchrist CL, Chen J, Richardson WJ, Loeser RF, Setton LA. Functional integrin subunits regulating cell-matrix interactions in the intervertebral disc. J Orthop Res 2007;25:82940. 29. 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;55:917. 30. 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:R73245. 31. Flexcell International. Flexcell culture systems. URL: http://www. flexcellint.com/gallery.htm. 32. Schultz A, Andersson G, Ortengren R, Haderspeck K, Nachemson A. Loads on the lumbar spine: validation of a biomechanical analysis by measurements of intradiscal pressures and myoelectric signals. J Bone Joint Surg Am 1982;64:71320. 33. Korecki CL, Maclean JJ, Iatridis JC. Dynamic compression effects on intervertebral disc mechanics and biology. Spine 2008;33: 14039. 34. Wang DL, Jiang SD, Dai LY. Biologic response of the intervertebral disc to static and dynamic compression in vitro. Spine 2007;32:25218. 35. Walsh AJ, Lotz JC. Biological response of the intervertebral disc to dynamic loading. J Biomech 2004;37:32937. 36. 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;30:11226. 37. Maclean JJ, Roughley PJ, Monsey RD, Alini M, Iatridis JC. In vivo intervertebral disc remodeling: kinetics of mRNA expression in response to a single loading event. J Orthop Res 2008;26: 57988. 38. 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;28:97381. 39. Gokorsch S, Nehring D, Grottke C, Czermak P. Hydrodynamic stimulation and long term cultivation of nucleus pulposus cells: a new bioreactor system to induce extracellular matrix synthesis by nucleus pulposus cells dependent on intermittent hydrostatic pressure. Int J Artif Organs 2004;27:96270. 40. Kasra M, Merryman WD, Loveless KN, Goel VK, Martin JD, Buckwalter JA. Frequency response of pig intervertebral disc cells

41.

42.

43.

44.

45.

46.

47.

48.

49. 50.

subjected to dynamic hydrostatic pressure. J Orthop Res 2006;24: 196773. Huang J, Ballou LR, Hasty KA. Cyclic equibiaxial tensile strain induces both anabolic and catabolic responses in articular chondrocytes. Gene 2007;404:1019. Hauselmann HJ, Masuda K, Hunziker EB, Neidhart M, Mok SS, Michel BA, et al. Adult human chondrocytes cultured in alginate form a matrix similar to native human articular cartilage. Am J Physiol 1996;271(3 Pt 1):C74252. Beekman B, Verzijl N, Bank RA, von der Mark K, TeKoppele JM. Synthesis of collagen by bovine chondrocytes cultured in alginate: posttranslational modifications and cellmatrix interaction. Exp Cell Res 1997;237:13541. Oegema TR, Johnson SL, Aguiar DJ, Ogilvie JW. Fibronectin and its fragments increase with degeneration in the human intervertebral disc. Spine 2000;25:27427. Probstmeier R, Pesheva P. Tenascin-C inhibits 1 integrin-dependent cell adhesion and neurite outgrowth on fibronectin by a disialoganglioside-mediated signaling mechanism. Glycobiology 1999;9:10114. Jobanputra P, Lin H, Jenkins K, Bavington C, Brennan FR, Nuki G, et al. Modulation of human chondrocyte integrins by inflammatory synovial fluid. Arthritis Rheum 1996;39:14302. Clancy RM, Rediske J, Tang X, Nijher N, Frenkel S, Philips M, et al. Outside-in signaling in the chondrocyte: nitric oxide disrupts fibronectin-induced assembly of a subplasmalemmal actin/rho A/focal adhesion kinase signaling complex. J Clin Invest 1997;100: 178996. Le Maitre CL, Hoyland JA, Freemont AJ. Catabolic cytokine expression in degenerate and herniated human intervertebral discs: IL-1 and TNF expression profile. Arthritis Res Ther 2007;9:R77. Pritchard S, Erickson GR, Guilak F. Hyperosmotically induced volume change and calcium signaling in intervertebral disk cells: the role of the actin cytoskeleton. Biophys J 2002;83:250210. Benallaoua M, Richette P, Francois M, Tsagris L, Revel M, Corvol M, et al. Modulation of proteoglycan production by cyclic tensile stretch in intervertebral disc cells through a post-translational mechanism. Biorheology 2006;43:30310.