ARTHRITIS & RHEUMATISMVol. 44, No. 9, September 2001, pp 2024–2028© 2001, American College of RheumatologyPublished by Wiley-Liss, Inc.

Matrix Metalloproteinase 9, Apoptosis, andVascular Morphology in Early Arthritis

Alexander Fraser, Ursula Fearon, Richard Reece, Paul Emery, and Douglas J. Veale

Objective. To examine matrix metalloproteinase 9(MMP-9) in the synovial fluid (SF) and synovial mem-brane (SM) in relation to vascular endothelial cell (EC)apoptosis, vascular endothelial growth factor (VEGF),and SM vascular pattern.

Methods. Thirty-four patients underwent needlearthroscopy of the knee joint; 12 had early rheumatoidarthritis (RA), 12 had early psoriatic arthritis (PsA),and 10 had osteoarthritis (OA). The early RA and earlyPsA patients were matched for disease activity. SF levelsof MMP-9 and VEGF were measured by an enzyme-linked immunosorbent assay, and EC apoptosis wasmeasured by TUNEL assay. MMP-9 expression wasexamined in SM by immunohistochemistry. Synovialtissue explants were stimulated with VEGF, and MMP-9levels were measured in the supernatants. The synovialvascular pattern was recorded.

Results. SF MMP-9 levels were significantlyhigher in early PsA patients than in early RA patients;OA patients had minimal levels. MMP-9 levels corre-lated with blood vessel morphology and SF VEGF levels.MMP-9 expression was greater in early PsA SM than inearly RA SM, but the difference was not significant. Incontrast however, EC apoptosis was greater in early RASM than in early PsA SM. MMP-9 levels increased2-fold and 9-fold, respectively, in SM explant culturesupernatants on day 7 in response to stimulation with25 ng/ml and 50 ng/ml of VEGF.

Conclusion. SF MMP-9 levels correlate with thepattern of SM neovascularization and SF VEGF levelsin early inflammatory arthritis, and VEGF increasesMMP-9 production by SM. Endothelial cell apoptosis,

however, appears to be more prevalent in early RA. Thiscombination of factors may explain the pattern ofdifferential angiogenesis in these arthritides.

Rheumatoid arthritis (RA) and psoriatic arthritis(PsA) are the 2 most common forms of chronic, dis-abling inflammatory arthritis. Distinct patterns of syno-vial membrane (SM) vascular morphology have recentlybeen described in RA and the seronegative arthritides (1).

Angiogenesis, or new blood vessel growth, is aphysiologic process central to the instigation, propaga-tion, and maintenance of tumor growth and chronicinflammatory conditions, including psoriasis and RA(2,3). An invading tumor and the eroding synovialpannus are both dependent on the support of a nour-ishing capillary bed. Recently, there have been manyexciting developments in the study of angiogenesis, andthe role of angiogenic inhibitors in the treatment ofcertain neoplasms is under investigation (4).

Matrix metalloproteinases (MMPs) play a signif-icant role in regulating angiogenesis, and MMP-9 (gela-tinase B) in particular has been shown to dissolveextracellular matrix, initiating and promoting new vesselformation (5). Vascular endothelial growth factor(VEGF) is also central to the angiogenic process (6) andhas been isolated from the joints of RA patients. RAendothelial cells (EC) express the VEGF receptors flt-1and kinase insert domain receptor (7), and SM expres-sion of VEGF is significantly different in early RA andearly PsA (8). In vitro studies suggest that VEGFup-regulates MMP-9 expression in vascular smooth mus-cle cells (9).

Dysregulated cellular apoptosis is also implicatedin the etiology of malignancy and RA (10), and apopto-sis of the EC may provide another important mechanismfor angiogenic control (11). Human microvascular ECapoptosis is inhibited by VEGF in vitro (11).

In this in vivo and ex vivo study, we examinedboth the proangiogenic factor MMP-9 and EC apoptosis

Alexander Fraser, MB, BCh, MRCPI, Ursula Fearon, PhD,Richard Reece, MB, BCh, MRCP, Paul Emery, MA, MD, FRCP,Douglas J. Veale, MD, FRCP, FRCPI: University of Leeds, Leeds, UK.

Address correspondence and reprint requests to Douglas J.Veale, MD, FRCP, FRCPI, Rheumatology Rehabilitation ResearchUnit, 36 Clarendon Road, Leeds LS3 9NZ, UK.

Submitted for publication November 9, 2000; accepted inrevised form April 11, 2001.

2024

in early RA and PsA, the effects of VEGF on MMP-9production by SM, and the vascular morphology ofthe SM.

PATIENTS AND METHODS

Patients. Twenty-four patients with early (,12months’ duration), untreated inflammatory arthritis and clini-cally active knee involvement, defined as current pain, swell-ing, and/or heat (12 with early RA; 12 with early PsA) and 10osteoarthritis (OA) patient controls were recruited from theclinics in Leeds. The RA patients satisfied the AmericanCollege of Rheumatology (formerly, the American Rheuma-tism Association) criteria for diagnosis (12). Full, written, andinformed consent was obtained, and the study was approved bythe local ethics committee.

Arthroscopy. All patients underwent needle arthros-copy of a symptomatic knee joint at the Rheumatology Unit,Chapel Allerton Hospital (Leeds, UK). Under local anesthe-sia, a 2.7 mm–diameter needle arthroscope was inserted intothe knee joint, synovial fluid (SF) was aspirated, and SMbiopsy samples were obtained under full visualization. Thejoint was fully examined in each case, and the vascular patternwas recorded as previously described (1). SF samples werecentrifuged at 3,000 revolutions per minute for 10 minutes, andthe supernatants were frozen at 270°C. SM biopsies were snapfrozen in liquid nitrogen, foil wrapped, and stored at 270°C.

Immunohistology. Cryostat sections (4–7 mm thick)were cut, dried overnight at 37°C, fixed in acetone for 10minutes, and stored at 220°C. A routine 3-stage immuno-peroxidase staining technique incorporating avidin–biotin–immunoperoxidase complex was used. SM sections werestained with a specific polyclonal antibody to MMP-9 (SantaCruz, CA). For control sections, Tris buffer or IgG was usedinstead of the primary antibody. Color was developed in asolution containing diaminobenzidine tetrahydrochloride (Sig-ma, Poole, UK), 0.5% H2O2 in Tris buffer without saline (pH7.6). Slides were counterstained with hematoxylin andmounted.

Expression was quantified according to an establishedsemiquantitative scoring method using a 0–4 scale, where 0 5no stained cells, 1 5 1–25% stained cells, 2 5 25–50% stainedcells, 3 5 50–75% stained cells, and 4 5 75–100% stainedcells (13).

Enzyme-linked immunosorbent assay (ELISA).MMP-9 and VEGF were measured in synovial fluids by com-mercially available ELISA kits (R&D Systems, Oxford, UK).

TUNEL assay. Assessment of apoptosis was carriedout using an in situ cell death fluorescein detection kit fromBoehringer Mannheim (Lewes, UK). Cryostat sections (4–8mm thick) were cut, dried overnight at 37°C, fixed in acetonefor 10 minutes, and stored in foil at 270°C. SM sections werepostfixed in paraformaldehyde, pH 7.4, for 30 minutes andwashed twice in Tris buffer. Sections were incubated with 0.1%Triton X-100 in 0.1% sodium citrate for 2 minutes on ice,washed in Tris, and then incubated with the TUNEL reagents(terminal deoxynucleotidyl transferase and fluorescein-labelednucleotide mixture) for 1 hour at 37°C. For positive controls,sections were incubated for 10 minutes with DNase beforeTUNEL reagents. Apoptosis was scored in the endothelial

region using the semiquantitative scoring method describedabove (0–4 scale) (13).

Explant cultures. Control culture medium was Eagle’sminimum essential medium (EMEM) containing 10% fetalcalf serum supplemented with streptomycin (100 mg/ml) andpenicillin (100 IU/ml). Synovial tissue was obtained duringjoint replacement surgery from patients with established RA(n 5 3). Synovial tissue was trimmed of adipose tissue andwashed in culture medium. Fragments (2–3 mm) were placedin culture medium in 24-well plates for 24 hours at 37°C priorto the addition of test reagents. After 24 hours, the supernatantwas removed from each well (time 0), then EMEM or VEGFwas added to basal cultures to give final concentrations of 25ng/ml and 50 ng/ml for each sample. Explants were maintainedfor 7 days when supernatants were removed, stored at 270°C,and analyzed for proMMP-9.

Statistical analysis. Data were stored and analyzedusing SPSS for Windows version 9.0 software (SPSS, Chicago,IL). Correlations were examined by Spearman’s rank correla-tion test, and the Kruskal-Wallis and Mann-Whitney tests wereused to assess statistical differences between groups. P valuesof less than 0.05 were considered significant for differences andcorrelations, and all differences and correlations discussed aresignificant at this level or better.

RESULTS

The 12 patients with early RA (5 male [42%])had a mean age of 55 years (range 27–79 years) and amean disease duration of 7.2 months (range 1–12months). The 12 patients with early PsA (9 male [75%])had a mean age of 34 years (range 20–72 years) and amean disease duration of 7.3 months (range 1–12months). The early RA and early PsA patients werematched for disease activity (Table 1). The mean age ofthe 10 OA patients was 72 years (range 56–93 years).Macroscopically, a tortuous vascular pattern was notedin the SM of all 12 patients with early PsA but in noneof the SM samples from the 12 patients with early RA.

SF levels of MMP-9 were significantly greater inearly PsA than in early RA (mean 6 SEM 14.62 6 9.80ng/ml versus 5.02 6 3.98 ng/ml; P , 0.01), and in OA,MMP-9 levels were mostly undetectable (1.28 6 1.6

Table 1. Clinical parameters of disease activity in patients with earlyRA and early PsA*

Early RA(n 5 12)

Early PsA(n 5 12)

HAQ, range 0–24 8.55 6 2.22 7.57 6 1.63C-reactive protein, mg/liter 14.21 6 3.64 12.14 6 2.936Plasma viscosity, mPa 1.69 6 0.36 1.77 6 0.39Swollen joint count, range 0–66 5.28 6 3.21 4.42 6 2.1

* Values are the mean 6 SEM. RA 5 rheumatoid arthritis; PsA 5psoriatic arthritis; HAQ 5 Health Assessment Questionnaire.

MMP-9, APOPTOSIS, AND VASCULAR MORPHOLOGY IN EARLY ARTHRITIS 2025

ng/ml) (Figure 1). SF MMP-9 levels correlated signifi-cantly with the SM blood vessel pattern (r 5 0.57, P ,0.05) (Figure 2) and with the SF VEGF levels (r 5 0.452,P , 0.05).

SM expression of MMP-9 in early PsA (n 5 9)was higher than that in early RA (n 5 8) (mean 6 SEM1.75 6 0.31 versus 1.12 6 0.12); however, this differencedid not reach statistical significance. SM VEGF expres-sion, which has been reported previously (8), demon-strated higher expression in the early PsA SM comparedwith the early RA SM.

Endothelial cell apoptosis was increased in earlyRA SM compared with early PsA (mean 6 SEM 1.75 60.37 versus 0.78 6 0.32; P 5 0.057). SM apoptosis wasprimarily perivascular and stromal, with minimal levelsfound in the lining layer of early RA and early PsApatients (Figure 3).

Synovial membrane explants stimulated withVEGF at a concentration of 25 ng/ml for 7 days in-creased supernatant MMP-9 levels by 2-fold from base-line. A VEGF concentration of 50 ng/ml increasedMMP-9 levels by 9-fold (Figure 4).

DISCUSSION

Different patterns of synovial vascularity de-scribed at arthroscopic examination have reliably beenshown to relate to either RA or seronegative arthritis(1). The finding of elongated, tortuous, and “bushy”

capillaries in PsA may indicate an abnormal or “imma-ture” angiogenesis with important pathogenic implica-tions. New blood vessel formation is central to allfibrovascular disorders, including tumor growth andpsoriasis (2–4), and increasingly, the EC is recognized asplaying a pivotal role in inflammation and immunologicmodulation (3). The process of angiogenesis is con-trolled by a variety of angiogenic inducers and inhibitors,and it is the balance of these factors which determinesthe quality and quantity of new vessel growth (2).

Apoptosis, or programmed cell death, is a tightlyregulated process that controls cell turnover in tissues.Consequently, apoptosis may control the morphofunc-tional characteristics of a given tissue and is a normalpart of the physiologic processes of tissue expansion andinvolution. Disturbed apoptosis of specific cell types hasbeen implicated in the development and persistence ofcancer and autoimmune conditions.

MMP-9 (gelatinase B or 92-kd gelatinase) de-grades components of the extracellular matrix and playsa critical role in the tumor model of angiogenesis (5,14).Experimental inhibitors of MMP-9 have been shown tosuppress tumor growth by restricting the expansion ofEC (4). VEGF is also known to be a key regulator ofangiogenesis. It stimulates the proliferation and migra-tion of EC and enhances their proteolytic activity (11).In vitro studies show that VEGF up-regulates MMP-9production by perivascular smooth muscle cells and that

Figure 1. Levels of matrix metalloproteinase 9 (MMP-9) in thesynovial fluid of patients with early psoriatic arthritis (PsA), earlyrheumatoid arthritis (RA), and osteoarthritis (OA). Data are shown asbox-plots. Each box represents the 25th to the 75th percentiles. Linesoutside the boxes represent the 10th and the 90th percentiles. Linesinside the boxes represent the median.

Figure 2. Scattergram showing the relationship between synovial fluidmatrix metalloproteinase 9 (MMP-9) levels and the presence orabsence of a tortuous synovial membrane vessel pattern in patientswith early psoriatic arthritis (PsA) and early rheumatoid arthritis(RA).

2026 FRASER ET AL

this action is mediated by the flt-1 high-affinity receptor.VEGF treatment also accelerates smooth muscle cellmigration through synthetic barriers, and this response isblocked by MMP inhibition, suggesting that this action ismediated via up-regulation of MMPs (9). Both factorshave been shown to enhance EC survival and inhibitapoptosis, and this may further contribute, in part, totheir proangiogenic activity (11,15).

In this study, we demonstrated that high levels ofMMP-9 and low levels of EC apoptosis are closelyrelated to SM vascular patterns in early PsA comparedwith early RA. Furthermore, the MMP-9 levels corre-lated with SF VEGF levels and with the pattern ofvascularity found in the SM tissue. SM expression ofVEGF and MMP-9 was also higher in early PsA; how-ever, this difference did not reach statistical significance.These 2 patient groups were matched for clinical corre-lates of disease severity (Health Assessment Question-

Figure 3. Synovial membrane (SM) biopsy samples from patients withA, early rheumatoid arthritis (RA), B, early psoriatic arthritis (PsA),and C, osteoarthritis (OA) obtained at arthroscopy and stained usinga cell death fluorescein detection kit (TUNEL assay). Early RA SMdemonstrates increased levels of endothelial cell apoptosis (arrows),which is undetectable in OA and early PsA SM.

Figure 4. Levels of matrix metalloproteinase 9 (MMP-9) detected inthe supernatant of synovial explant cultures on day 7, when stimulatedwith either 0 (basal), 25, or 50 ng/ml of vascular endothelial growthfactor (VEGF).

MMP-9, APOPTOSIS, AND VASCULAR MORPHOLOGY IN EARLY ARTHRITIS 2027

naire score, C-reactive protein level, plasma viscosity,and swollen joint count), and SM levels of the keyproinflammatory cytokine transcription factor, nuclearfactor kB, has previously been shown to be higher in theearly RA cohort (8). Consequently, the quantitative andqualitative vascular differences identified are not merelya reflection of inflammatory severity. The RA SMexplant cultures confirm that ex vivo, VEGF stimulatesthe SM to secrete MMP-9, and therefore, the proangio-genic properties of these factors are interdependent.

The greater levels of EC apoptosis found in bloodvessels of early RA SM compared with early PsA SMsuggest that reduced EC survival or increased turnovermay occur in RA. This may be an important controllingfactor for angiogenesis. The high levels of MMP-9 andVEGF found in PsA SF may contribute to the inhibitionof EC apoptosis.

These data suggest that differential production ofMMP-9 and VEGF by the SM and differential ECapoptosis may be contributing factors in the develop-ment of different angiogenic phenotypes in RA and PsAat early disease stages. These findings have importantimplications for pathogenesis and targeted therapy.

REFERENCES

1. Reece RJ, Canete JD, Parsons WJ, Emery P, Veale DJ. Distinctvascular patterns of early synovitis in psoriatic, reactive, andrheumatoid arthritis. Arthritis Rheum 1999;42:1481–4.

2. Koch AE. Angiogenesis: implications for rheumatoid arthritis.Arthritis Rheum 1998;41:951–62.

3. Lowe PM, Lee ML, Jackson CJ, To SS, Cooper AJ, Schreiber L.The endothelium in psoriasis. Br J Rheumatol 1995;132:497–505.

4. Shalinsky DR, Brekken J, Zou H, McDermott CD, Forsyth P,Edwards D, et al. Broad antitumor and antiangiogenic activities ofAG3340, a potent and selective MMP inhibitor undergoing ad-vanced oncology trials. Ann N Y Acad Sci 1999;878:236–70.

5. Sang QX. Complex role of matrix metalloproteinases in angiogen-esis [review]. Cell Res 1998;8:171–7.

6. Folkman J. Angiogenesis and angiogenesis inhibition: an over-view. EXS 1997;79:1–8.

7. Fava RA, Olsen NJ, Spencer-Green G, Yeo KT, Yeo TK, Berse B,et al. Vascular permeability factor/endothelial growth factor(VPF/VEGF): accumulation and expression in human synovialfluids and rheumatoid synovial tissue. J Exp Med 1994;180:341–6.

8. Fearon U, Reece R, Smith J, Emery P, Veale DJ. Synovialcytokine and growth factor regulation of MMPs/TIMPs: implica-tions for erosions and angiogenesis in early rheumatoid andpsoriatic arthritis patients. Ann N Y Acad Sci 1999;878:619–21.

9. Wang H, Keiser JA. Vascular endothelial growth factor upregu-lates the expression of matrix metalloproteinases in vascularsmooth muscle cells: role of flt-1. Circ Res 1998;83:832–40.

10. Vaishnaw AK, McNally JD, Elkon KB. Apoptosis in the rheumaticdiseases. Arthritis Rheum 1997;40:1917–27.

11. Nor JE, Christensen J, Mooney DJ, Polverini PJ. Vascular endo-thelial growth factor (VEGF)-mediated angiogenesis is associatedwith enhanced endothelial cell survival and induction of Bcl-2expression. Am J Pathol 1999;154:375–84.

12. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,Cooper NS, et al. The American Rheumatism Association 1987revised criteria for the classification of rheumatoid arthritis.Arthritis Rheum 1988;31:315–24.

13. Cunnane G, Fitzgerald O, Hummel KM, Gay RE, Gay S, Bresni-han B. Collagenase, cathepsin B and Cathepsin L gene expressionin the synovial membrane of patients with early inflammatoryarthritis. Rheumatology (Oxford) 1999;38:32–42.

14. Himelstein BP, Canete-Soler R, Bernhard EJ, Dilks DW, MuschelRJ. Metalloproteinases in tumor progression: the contribution ofMMP-9. Invasion Metastasis 1994;14:246–58.

15. Vu TH, Shipley JM, Bergers G, Berger JE, Helms JA, Hanahan D,et al. MMP-9/gelatinase B is a key regulator of growth plateangiogenesis and apoptosis of hypertrophic chondrocytes. Cell1998;93:411–22.

2028 FRASER ET AL