Transforming and epidermal growth factors in degenerated intervertebral discs Y. T. Konttinen, P. Kemppinen, T. F. Li, E. Waris, H. Pihlajamäki, T. Sorsa, M. Takagi, S. Santavirta, G. S. Schultz, M. G. Humphreys-Beher From the University of Helsinki and Helsinki University Central Hospital, Finland, Tohoku University School of Medicine, Sendai, Japan and the University of Florida, Gainesville, USA

e studied the presence of anabolic growth factors in human herniated intervertebral discs (IVD) using a reverse transcriptase-polymerase chain reaction (RT-PCR) and immunohistochemistry. Messenger RNA (mRNA) was isolated from the nucleus pulposus using oligo (dT)25 superparamagnetic beads and probing with gene-specific primers in RT-PCR. mRNA coding for TGF-
(3/10), EGF (0/10), TGF-1 (0/10) and TGF-3 (2/10) or the EGF receptor (EGF-R; 0/10) and TGF- type-II receptor (0/10) was found only occasionally. Beta-actin was always present and positive sample controls confirmed the validity of the RT-PCR assay. These RT-PCR findings were confirmed using immunohistochemical staining of EGF and TFG-, whereas TGF-
protein was always found associated with discocytes. We conclude that the nucleus pulposus of the herniated IVD is vulnerable to proteolytic degradation and depletion of proteoglycans due to the lack and/or

W

Y. T. Konttinen, MD, PhD, Senior Researcher of the Finnish Academy and Consultant Rheumatologist P. Kemppinen, MSc E. Waris, BSc Department of Anatomy, Institute of Biomedicine, University of Helsinki, Siltavuorenpenger 20A, 00014 Helsinki, Finland. T. F. Li, MD H. Pihlajamäki, MD, PhD S. Santavirta, MD, PhD, Professor and Head Department of Orthopaedics and Traumatology, Helsinki University Central Hospital, Topeliuksenkatu 5, 00260 Helsinki, Finland. Y. T. Konttinen, MD, PhD T. Sorsa, DDS, PhD, Professor Institute of Dentistry, Mannerheimintie 172, University of Helsinki, Helsinki, Finland. M. Takagi, MD, PhD, Associate Professor Department of Orthopaedic Surgery, Yamagata University School of Medicine, 2-2-2 Iida-Nishi, Yamagata 990-9585 Japan. G. S. Schultz, MD, PhD, Professor Wound Healing Institute, Department of Gynaecology M. G. Humphreys-Beher, PhD, Associate Professor Department of Oral Biology College of Dentistry, University of Florida, Gainesville, Florida 32610-0424, USA. Correspondence should be sent to Dr Y. T. Konttinen. ©1999 British Editorial Society of Bone and Joint Surgery 0301-620X/99/69321 $2.00 1058

low production of anabolic growth factors/receptors which could increase the local synthesis of the extracellular matrix. J Bone Joint Surg [Br] 1999;81-B:1058-63. Received 23 July 1998; Accepted after revision 6 January 1999

The functional unit of the lumbar spine consists of two vertebral bodies combined by a fibrocartilaginous intervertebral disc (IVD). This synchondrosis, or uncovertebral, 1 joint allows movement in flexion and extension during which it functions as a shock absorber and as a dynamic 2 cushion. It is composed of a proteoglycan (PG) and waterrich nucleus (nucleus pulposus) surrounded by lamellae of 3,4 collagen organised to a fibrous annulus. This arrangement is to a large extent responsible for the biomechanical 5 properties of a healthy IVD. In a degenerated IVD this local homeostasis has not been maintained because of a decreased content of PG and water in the nucleus pulpo6-8 sus. The altered properties of the degenerating IVD 9 increase the mechanical stress of the lumbar spine. Research has so far focused on the different catabolic factors and proteases which can degrade PG core proteins 10-13 (proteoglycanases). Less attention has been paid to the regenerative capacity of the nucleus. An adult nucleus pulposus is an avascular tissue. Systemic regulatory factors and nutrients are delivered to the IVD cells by diffusion through the extracellular matrix, which also regulates the behaviour of discocytes by matrix14,15 cell interactions. Another potential source of regulatory factors could be the local discocytes. The members of the transforming growth factor (TGF) super family have been shown to play an important role in the anabolic metabolism of the IVD. In mature canine IVDs, TGF- and epidermal growth factor (EGF) elicited greater proliferative responses than other factors such as fibroblast growth factor (FGF) 16 and insulin-like growth factor (IGF). In addition, TGF can upregulate the expression of the tissue inhibitors of 17 some metalloproteinases. Our aim was to assess the anabolic potential of degenerated/herniated IVDs by determining the local expression of TGF and EGF. The responsiveness of the discocytes to these soluble factors was assessed based on the demonstration of synthesis of their respective receptors. THE JOURNAL OF BONE AND JOINT SURGERY

TRANSFORMING AND EPIDERMAL GROWTH FACTORS IN DEGENERATED INTERVERTEBRAL DISCS

Materials and Methods Samples of nucleus pulposus from herniated IVDs were collected from ten patients, five men and five women with a mean age of 37.5 years (21 to 50). The clinical diagnosis had been confirmed before operation by MRI. The herniation was located between L4 and L5 in five and between L5 and S1 in five patients. The operation had been performed from the right or left depending on the side of the herniation. Fibrous tissue of the capsule of the hip, obtained at a total hip replacement performed for primary osteoarthritis, was used as a positive sample control. All the samples were frozen in dry ice precooled with isopentane immediately after removal. For each mRNA extraction 80 to 90 tissue sections 15 µm thick were used. Aerosol-resistant pipette tips (Biohit, Helsinki, Finland) were used for pipetting hot DNA-containing solutions. The mRNA was extracted as previously descri18 bed. Briefly, the procedure used oligo (dT)25 covalently attached to superparamagnetic polystyrene microbeads according to the manufacturer’s protocol (Dynal, Oslo, Norway). For each sample, 30 µl of beads (binding capacity + 2 ng of poly(A) mRNA/µl beads) were used. They were washed with lysis/binding buffer (100 mM Tris-HC1, 500 mM LiCl, 10 mM EDTA, 1% LiDS, 5 mM DTT, pH 8.0) using the magnetic particle collector supplied. The buffer (100 µl) was then added to the tissue sections and genomic DNA was sheared by passing the sample suspension 10 to 15 times through a 1 ml syringe with a 22G needle. The supernatant was incubated with the (dT)25 beads for 3 minutes at room temperature, washed twice with 200 µl of washing buffer with LiDS (10 mM Tris-HC1, 0.15 M LiCl, 1 mM EDTA, 0.1% LiDS, pH 8.0), twice with washing buffer without LiDS and finally once with 0.5 RT buffer (45 mM KC1, 5 mM Tris-HC1, pH 8.3). The rTth reverse transcription (Perkin Elmer, Branchburg, New Jersey) reaction was performed using 5 U of the enzyme in a total volume of 20 µl (90 mM KC1, 10 mM Tris-HC1, pH 8.3, 1 mM MnC12, 200M of dATP, dCTP, dGTP and dTTP) in thin-walled PCR tubes (Plastic Trade, Helsinki, Finland) topped with 50 µl of mineral oil (Sigma, St Louis, Missouri) using a thermal cycler (Pharmacia, Sollentuna, Sweden). The reaction was run for 1 minute at room temperature, 1 minute at +37°C, 5 minutes at +55°C and 10 minutes at +70°C. After the synthesis of the first strand of cDNA the beads were collected, resuspended in

1059

50 µl of DEPC (Sigma)-treated H2O and heated for 1 minute at +95°C to melt the cDNA:RNA hybrid. The beads containing the first strand of cDNA were immediately collected. The pooled gene-specific sense and antisense primers were donated by Dr Humphreys-Beher (Table I). The beads containing the first strand of cDNA were washed once with 100 µl of PCR buffer (10 mM Tris-HC1, 1.5 mM MgC12, 50 mM KC1, 0.1% Triton X-100, pH 8.8) and synthesis of the second strand of cDNA was done with gene-specific sense and antisense primers (0.8M/PCR reaction) and 200M of dATP, dCTP, dGTP and dTTP in PCR buffer in thin-walled PCR tubes topped with 50 µl of mineral oil. Hot start at +80°C was used when adding 1.0 U of the thermostable DNA polymerase (Finnzymes, Espoo, Finland) in a final reaction volume of 30 µl. Samples were denatured for 2 minutes at +95°C and annealed for 1 minute at +61°C followed by extension for 5 minutes at +72°C. After melting the strands for 2 minutes at +95°C the supernatant containing the second strand of cDNA was transferred to a new PCR tube in which were performed 35 (for -actin) or 45 cycles (TGF-1, TGF-3, TGF-II-R, TFG-
, EGF and EGF receptor (EGF-R)) of PCR reactions consisting of 1 minute of denaturation at +95°C, 1 minute of annealing at a primer-specific temperature of +60°C for TGF-1, TGF-3, EGF and EGF-R, +61°C for -actin, +63°C for TGF-II-R and TGF-
and 1 minute of extension at +72°C. For the last cycle 5 minutes of extension at +72°C were used. The beads containing the first strand of cDNA of each sample were used for all of the genes studied. Amplified DNA was run on a 1.5% modified NuSieve agarose gel (FMC BioProducts, Rockland, Maine) for verification of size. PCR fragments were extracted from the gel using silica-gel-membrane-based QIAquick columns according to the manufacturer’s protocol (Qiagen Inc, Chatsworth, California) which were quantified, sequenced (20 to 75 ng/PCR fragment) using fluorescein-labelled dye terminator kits (PE Applied Biosystems, Foster City, California) and analysed on an automatic sequencer 373 A (PE Applied Biosystems). Frozen tissue samples were cut into sections 8 µm thick, mounted on gelatin-coated glass slides and fixed in acetone for 5 minutes at +4°C. Intrinsic peroxidase activity was abolished by pretreating the tissue sections in 0.1% H2O2 in methanol for 30 minutes. Sections were rinsed in 0.3% Triton-X in 0.1M PBS before staining using the avidin-

Table I. Details of the nucleotide sequences of the sense and antisense primers, the PCR amplification product and the primer concentrations Gene

Sense primer

Antisense primer

PCR product (bp)

Primer concentration* (M)

-actin TGF-1 TGF-3 TGF-II-R TGF-
EGF EGF-R

GTGGGGCGCCCCAGGCACCA CAAGCAGAGTACACACAGCA ATTACCTCCAAGGTTTTCCG TGTGTTCCTGTAGCTCTGATG ATGGTCCCCTCGGCTGGCAG TATGTCTGCCGGTGCTCAGAA GTGACCGTTTGGGAGTTGATG

CTCAATGTCACGCACGATTTC GATGCTGGGCCCTCTCCAGC GCCCGCTTCTTCCTCTGACC GATCTTGACTGCCTCTGTCTC CCTCTGGGCTCTTCAGACCTC AGCGTGGCGCAGTTCCCACCA AAACCAGTCTGTGGGTCTAAG

540 442 541 430 495 360 1157

0.8 0.8 0.8 0.8 0.8 0.8 0.8

* per PCR reaction VOL. 81-B, NO. 6, NOVEMBER 1999

1060

Y. T. KONTTINEN,

P. KEMPPINEN,

Fig. 1

T. F. LI,

ET AL

Fig. 2

Figure 1 – Panels A to E – ß-actin mRNA amplification products (540 bp) after 35 cycles in RT-PCR in which panel A represents the first amplification round with the oligo (dT)25 beads and panel E the last, respectively. Amplification products were electrophoresed on 1.5% modified agarose gel using ethidium bromide both in the gel and in the electrophoresis buffer and photographed under ultraviolet light. Lanes marked L contain a 100 bp ladder starting from 200 bp upwards. Lane 2 is the amplification product from THR as a positive amplification control and lanes 3 to 12 are amplification products of ß-actin from sections of discs obtained from ten operated patients. Lanes 13 and 14 are negative controls without template and without template and primers, respectively, followed by a 100 bp ladder. Panel B – TGF-ß1 mRNA amplification product (442 bp). Only the THR-positive control, but not any of the patient samples, contained TGF-ß1. Lanes are arranged as in panels A and E. Panel C – TGF-ß mRNA amplification product (541 bp). Two of the patient samples, lanes 8 and 11, contain TGF-ß3 II-R. Lanes are arranged as in panels A and E. Panel D – TGF-ßII-R mRNA amplification product (430 bp). Only the THR-positive control, but not any of the patient samples, contain TGF-ß. Lanes are arranged as in panels A and E. Figure 2 – Panels A to E – ß-actin mRNA amplification products (540 bp) after 35 cycles in RT-PCR in which panel A represents the first amplification round with the oligo (dT)25 beads and panel E the last, respectively. Otherwise the size standards, positive sample control, patient samples and negative RT-PCR controls are as arranged as in Figure 1. Panel B – TGF-
mRNA amplification product (495 bp). Three of the patient samples, lanes 4, 8 and 12, contain TGF-
. Lanes are arranged as in panels A and E. Panel C – Epidermal growth factor (EGF) mRNA (360 bp). Only the THR-positive control, but not any of the patient samples, contains EGF. Lanes are arranged as in panels A and E. Panel D – Epidermal growth factor receptor (EGF-R) mRNA (1157 bp). Only the THR positive control, but not any of the patient samples, contains EGF-R. Lanes are arranged as in panels A and E. 19

biotin-peroxidase complex (ABC) method as described 20,21 elsewhere. Briefly, serial sections were sequentially incubated with: 1) normal rabbit or horse serum (1:20; Vector Laboratory, Burlingame, California) depending on the primary antibody used for 20 minutes at +22°C; 2) primary antibodies consisting of a) polyclonal goat anti-human EGF IgG 10 µg/ml (R&D Systems, Abingdon, UK) (no cross-reactivity detected by Western blot or direct ELISA for rhTGF-
, rhIL-
, rhIL-1, rhIL-2, rhIL-3, rhIL4, hPDGF); b) monoclonal mouse anti-human TGF-
antibodies, diluted 1:400 (Chemicon, Temecula, Canada) (no cross-reactivity detected with EGF or TFG-) or c) monoclonal mouse anti-human TGF-1, 2, 3, IgG1, 20 µg/ml

(Genzyme Diagnostics, Cambridge, Massachusetts) (primary antibodies were incubated overnight in a moist chamber at +4°C); 3) biotinylated rabbit anti-goat IgG or biotinylated horse anti-mouse IgG, respectively, for 1 hour at +22°C; 4) ABC complexes in 0.1M PBS (1:100; Vector Laboratory) for 1 hour at +22°C; and 5) 0.06% 3,3-diaminobenzidine tetrahydrochloride (DAB; Sigma Chemicals) and 0.006% H2O2. All sections were dehydrated in a graded ethanol series, cleared in xylene and mounted in Diatex. The specificity of the reactions was tested by the omission of the primary antibody, which was replaced with the corresponding normal goat serum or monoclonal mouse IgG control antibody THE JOURNAL OF BONE AND JOINT SURGERY

TRANSFORMING AND EPIDERMAL GROWTH FACTORS IN DEGENERATED INTERVERTEBRAL DISCS

of the same subtype and concentration, but with irrelevant specificity (Dakopatts a/s, Glostrup, Denmark).

Results All samples of the herniated lumbar IVD and of the positive control fibrous capsule contained beta-actin demonstrating the successful extraction of mRNA (Figs 1 and 2). The fibrous capsule control contained TGF-
, EGF, TGF-1, TGF-3, TGF-
, EGF-R and TGF- type II receptor showing that the processing of the sample, probing and amplification protocol worked to full satisfaction. The identity of the PCR amplification products was confirmed, not only by verification of size, but also by nucleotide sequencing. All the experiments were run twice and gave consistent results. Histological examination of the semiserial sections confirmed that the samples contained chondrocyte-like discocytes embedded in amorphous extracellular matrix, whereas spindle-shaped, fibroblast-like

1061

cells embedded in fibrillar collagenous matrix were practically absent from all samples. Regarding the herniated IVD samples, TGF-
was found in 3 out of 10 samples (Fig. 2, panel B). In addition, TGF3 was found in 2 out of 10 samples (Fig. 1, panel C) which, however, were not the same as those containing TGF-
mRNA (Fig. 1, panel C, Fig. 2, panel B). EGF (Fig. 2, panel C), TGF-1 (Fig. 1, panel B), EGF-R (a receptor shared by EGF and TGF-
; Fig. 2, panel D) and TGF- type-II receptor (Fig. 1, panel D) were not found in any of the IVD samples studied. Immunohistochemical ABC staining of TGF-
, TGF- (1, 2, 3) and EGF (Fig. 3) in the positive sample control, i.e., fibrous capsule of the hip, gave positive staining results with moderate to strong staining of the fibroblast-like cells of the capsule (Figs 3C and 3D). TGF-
was also found in the discocytes (Fig. 4), but TGF- and TGF were not found in situ in the nucleus pulposus of human IVDs (Figs 3A and 3B).

Fig. 3 Immunohistochemical avidin-biotin-peroxidase complex (ABC) staining of TGF-ß and EGF. Discocytes do not contain any TGF-ß. Figure 3A – Staining with monoclonal mouse anti-human TGF-ß antibodies which recognise the ß1, ß2 and ß3 isoforms of TGF-ß (
500). EGF staining gave similar results (not shown). Figure 3B – A semiserial section showing that the field in panel A contains a chondron and individual discocytes embedded in the extracellular matrix (
500) (haematoxyin counterstain). Figs 3C and 3D – Positive sample controls stained for TGF-ß and EGF, respectively. Fibrous tissue of the capsule of the hip, obtained at a total hip replacement operation performed for a primary osteoarthritis, was used as a positive sample control (
625).

VOL. 81-B, NO. 6, NOVEMBER 1999

1062

Y. T. KONTTINEN,

P. KEMPPINEN,

Fig. 4 Immunohistochemical ABC staining of TGF-
. There is strong labelling of the discocytes in the chondron (
625). Negative staining control performed with a monoclonal mouse IgG of the same subtype and concentration instead of primary mouse anti-human TGF-
antibodies was negative.

Discussion Low back pain is the most common physical complaint of the middle-aged. During skeletal maturation, the clear gelatinous nucleus pulposus of childhood becomes a firm fibrous plate. In early adult life, fissures and cracks appear in the periphery and then extend to the central region of the IVD. Histological studies have shown that the number of viable cells in all regions of the IVD declines sharply, especially in the central regions. A decrease in PG and water content and an increase in the concentration of noncollagenous proteins accompany these morphological changes. These age-related changes increase the possibility of degeneration/herniation of the IVD and contribute to the 22 loss of mobility and strength of spine. Herniated IVD specimens have been shown spontaneously to produce increased amounts of nitric oxide, inter23 leukins, prostaglandins and different MMPs. By contrast, there have been no studies on the eventual local production of molecules able to stimulate a compensatory increase in 16 the synthesis of PG, namely TGF and EGF. We now report that such locally active auto and paracrine factors are only occasionally produced in the nucleus pulposus of a herniated IVD. We conclude that these two events together, namely depletion of inflammation-induced ‘proteoglycanases’/PG and a low capacity of the nucleus pulposus of herniated IVD to compensate for the PG loss, make the nucleus pulposus vulnerable. These observations, however, do not allow any conclusions with regard to healthy human disc tissue because the exposure of the nucleus pulposus to rupture may significantly influence the expression of mRNAs and the presence of growth factors in herniated IVD tissue. It is also of interest that TGF- suppresses some MMP synthesis, but increases the synthesis of their 24 endogenous inhibitors. Furthermore, TGF- inhibits the 25 production of urokinase-type plasminogen activator and

T. F. LI,

ET AL

stimulates that of the plasminogen activator inhibitor type 26 1, which exerts a protective influence against degradation of PG. This combined effect on MMP and plasminogen systems would be particularly detrimental, because plasmin is able to activate proteolytically the proMMPs to the 27-29 corresponding active enzyme species. In the nucleus pulposus the lack of an anabolic effect seems to be combined with a lack of an anticatabolic effect. All discocytes stained for TGF-
. There is thus an apparent discrepancy between the local expression of TGF
mRNA (3/10) and its translation product (10/10). This might indicate a short half-life for the TGF-
mRNA and/or a long half-life for the corresponding protein. Alternatively, it is possible that exogenous TGF-
is deposited in the 30 lamina of the discocyte. Because the discocytes did not contain mRNA coding for the TGF-
receptor (=EGF-R), they would seem to be inherently non-responsive to TGF-
. TGF-
and EGF share a common receptor, EGR-R, which was not locally produced, nor was TGF- type-II receptor. Thus, the situation prevailing in the nucleus pulposus could be analogous to that in degenerating osteoarthritic cartilage in which catabolic cytokines like IL-1, TNF-
and IL-6, are produced in excess compared with anabolic cytokines, e.g., TGF-, bFGF and IGF. Lack or low level of expression of anabolic growth factors and their respective receptors is probably a significant risk factor predisposing to degeneration of the IVD and thus herniation. This is further augmented by the low cell-to-extracellular-matrix ratio 31 prevailing in the nucleus pulposus, which contains only 3 approximately 4000 to 5000 cells/mm. Routine histological examination of the samples showed the typical morphological features of the nucleus pulposus. More or less cell-rich chondrons were found embedded in an amorphous extracellular matrix. Usually no spindleshaped, fibroblast-like cells or fibrillar collagen were seen. This finding confirms that the samples were indeed from the nucleus pulposus and not from the annulus fibrosus. To obtain such a clear-cut dissection of these two elements of the IVDs, we used an isolation protocol for mRNA based on the use of tissue sections and superparamagnetic oligo (dT) beads An internal standard with -actin and a positive control using capsular tissue from the hip confirmed the validity of our methods. If there had been any production of growth factor and receptor in situ, it would have been detected by the current protocol for isolation, probing and amplification of mRNA. To a large extent, parallel results were obtained using immunohistochemistry. This work was supported by the Yrjö Jahnsson Foundation and by the Helsinki University Central Hospital vo-grant. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

References 1. Soini J, Antii-Poika I, Tallroth K, et al. Disc degeneration and angular movement of the lumbar spine: comparative study using plain and flexion-extension radiography and discography. J Spinal Disord 1991;4:183-7. THE JOURNAL OF BONE AND JOINT SURGERY

TRANSFORMING AND EPIDERMAL GROWTH FACTORS IN DEGENERATED INTERVERTEBRAL DISCS

2. Humzah MD, Soames RW. Human intervertebral disc: structure and function. Anat Rec 1988;220:337-56. 3. Urban J. Disc biochemistry in relation to function. In: Weinstein JN, Wiesel SW, eds. The lumbar spine. The International Society for the Study of the Lumbar Spine. Philadelphia: WB Saunders Co, 1990: 231-44. 4. Chelberg MK, Banks GM, Geiger DF, Oegema TR. Identification of heterogeneous cell populations in normal human intervertebral disc. J Anat 1995;186:43-53. 5. Holm SH. Nutrition of the intervertebral disc. In: Weinstein JN, Wiesel SW, eds. The lumbar spine. The International Society for the Study of the Lumbar Spine. Philadelphia: WB Saunders Co, 1990: 244-60. 6. Buckwalter JA. Aging and degeneration of the human intervertebral disc. Spine 1995;20:1307-14. 7. Pritzker KP. Aging and degeneration in the lumbar intervertebral disc. Orthop Clin North Am 1977;8:66-77. 8. Krajickova J, Polakova R, Smetana K, Vytasek R. Age-dependent changes in proteoglycan biosynthesis in human intervertebral discs. Folia Biol Praha 1995;41:41-51. 9. Santavirta S, Konttinen YT. Lumbar spine: studies of pathobiology, clinical examination and surgical treatment. Curr Opin Orthop 1991; 2:206-11. 10. Cole TC, Melrose J, Ghosh P. Cysteine proteinase inhibitors of the canine intervertebral disc. Biochim Biophys Acta 1988;952:201-7. 11. Liu J, Roughley PJ, Mort JS. Identification of human intervertebral disc stromelysin and its involvement in matrix degradation. J Orthop Res 1991;9:568-75. 12. Ng SC, Weiss JB, Quennel R, Jayson MI. Abnormal connective tissue degrading enzyme patterns in prolapsed intervertebral discs. Spine 1986;11:695-701. 13. Sedowofia KA, Tomlinson IW, Weiss JB, Hilton RC, Jayson MI. Collagenolytic enzyme systems in human intervertebral disc: their control, mechanism, and their possible role in the initiation of biomechanical failure. Spine 1982;7:213-22. 14. Meredith JE, Fazeli B, Schwartz MA. The extracellular matrix as a survival factor. Mol Biol Cell 1993;4:953-61. 15. Adams JC, Watt FM. Regulation of development and differentiation by the extracellular matrix. Development 1993;117:1183-98. 16. Thompson JP, Oegema TR Jr, Bradford DS. Stimulation of mature canine intervertebral disc by growth factors. Spine 1991;16:253-60. 17. Su S, DiBattista JA, Sun Y, Li WQ, Zafarullah M. Up-regulation of tissue inhibitor of metalloproteinases-3 gene expression by TGF-beta in articular chondrocytes is mediated by serine/threonine and tyrosine kinases. J Cell Biochem 1998;70:517-27.

VOL. 81-B, NO. 6, NOVEMBER 1999

1063

18. Konttinen YT, T¨ornwall J, Kemppinen P, et al. Neutral endopeptidase (EC 3.4.24.11) in labial salivary glands in healthy controls and in patients with Sj¨ogren’s syndrome. Ann Rheum Dis 1996;55:513-9. 19. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex (ABC) inimmunoperoxidase techniques: a comparison between ABC and unlabelled antibody (PAP) procedures. J Histochem Cytochem 1981;29:577-80. 20. Konttinen YT, Gr¨onblad M, Korkala O, Tolvanen E, Polak JM. Immunohistochemical demonstration of subclasses of inflammatory cells and active, collagen-producing fibroblasts in the synovial plicae of lumbar facet joints. Spine 1990;15:387-90. 21. Koski H, Konttinen YT, Gu XH, Hietanen J, Malmstr¨om M. Transforming growth factor-2 in labial salivary glands in Sj¨ogren’s sydrome. Ann Rheum Dis 1995;54:744-7. 22. Buckwalter JA, Woo SL-Y, Goldberg VM, et al. Soft tissue aging and musculoskeletal function. J Bone Joint Surg [Am] 1993;75-A: 1533-48. 23. Kang JD, Stefanovic-Racic M, McIntyre LA, Georgescu HI, Evans CH. Toward a biochemical understanding of human intervertebral disc degeneration and herniation: contributions of nitric oxide, interleukins, prostaglandin E2 and matrix metallo-proteinases. Spine 1997;22: 1065-73. 24. Edwards DR, Leco KJ, Beaudry PP, et al. Differential effects of transforming growth factor-beta 1 on the expression of matrix metalloproteinases and tissue inhibitors of metalloproteinases in young and old human fibroblasts. Exp Gerontol 1996;31:207-23. 25. Flaumenhaft R, Abe M, Mignatti P, Rifkin DB. Basic fibroblast growth factor-induced activation of latent transforming growth factor beta in endothelial cells: regulation of plasminogen activator activity. J Cell Biol 1992;118:901-9. 26. Cao HJ, Hogg MG, Martino LJ, Smith TJ. Transforming growth factor-beta induces plasminogen activator inhibitor type-1 in cultured human orbital fibroblasts. Invest Ophthalmol Vis Sci 1995;36:1411-9. 27. He CS, Wilhelm SM, Pentland AP, et al. Tissue cooperation in a proteolytic cascade activating human interstitial collagenase. Proc Natl Acad Sci USA 1989;86:2632-6. 28. Salo T, Liotta LA, Keski-Oja J, Turpeenniemi-Hujanen T, Tryggvason K. Secretion of basement membrane collagen degrading enzyme and plasminogen activator by transformed cells: role in metastasis. Int J Cancer 1982;30:669-63. 29. Werb Z, Mainardi CL, Vater CA, Harris ED. Endogenous activation of latent collagenase by rheumatoid synovial cells: evidence for a role of plasminogen activator. N Engl J Med 1977;296:1017-23. 30. Flaumenhaft R, Rifkin DB. The extracellular regulation of growth factor action. Mol Biol Cell 1992;3:1057-65. 31. Maroudas A, Stockwell R, Nachemson A, Urban J. Factors involved in the nutrition of human lumbar intervertebral disc: cellularity and diffusion of glucose in vitro. J Anat 1975;120:113-30.