MMP Expression in Rheumatoid Inflammation

ORIGINAL ARTICLE MMP expression in rheumatoid inﬂammation: the rs11568818 polymorphism is associated with MMP-7 expression at an extra-articular site

MG Kazantseva1, NA Hung2, J Highton3 and PA Hessian1

Matrix metalloproteinases (MMPs) contribute to the joint damage in rheumatoid arthritis (RA). Less is known of the involvement of MMPs at extra-articular sites of rheumatoid inflammation. We assessed the relative contribution from MMP-1, MMP-3, MMP-7 and MMP-12 to joint and extra-articular tissue destruction and inflammation by comparing gene expression in joint synovia and subcutaneous rheumatoid nodules from RA patients. Expression of MMP-1 and MMP-3 predominated in synovia, whereas MMP-12 expression was significantly higher in rheumatoid nodules. Markedly higher MMP-7 expression distinguished a subgroup of nodules that featured infiltrating monocyte/macrophage-producing MMP-7 protein. The high MMP-7 expression in nodules was associated with the single-nucleotide polymorphism (SNP) rs11568818 ( À 181A4G, MMP-7 promoter) and more active inflammation within the nodule lesions. Patients with such nodules had significantly earlier age of RA onset. Our findings indicate that the expression of MMP-1 and MMP-3 occurs relatively independent of the tissue microenvironment with substantial expression also at extra-articular sites. MMP-12 expression reflects the involvement of monocyte/macrophages in rheumatoid inflammation. Evidence for the association between the rs11568818 SNP and increased MMP-7 expression is restricted to nodules, which indicates that consequences of the MMP-7 polymorphism are likely to manifest within aspects of immune/inflammatory activity that are monocyte/macrophage-mediated.

Genes and Immunity (2013) 14, 162–169; doi:10.1038/gene.2012.65; published online 24 January 2013 Keywords: matrixmetalloproteinase; inﬂammation; gene expression; polymorphism; MMP-7

INTRODUCTION and MMP-3 that, respectively, target collagenous and non- 11–13 Rheumatoid arthritis (RA) is a prototypic chronic, inflammatory collagen components of the extracellular matrix. Much less disease. The principal clinical feature is progressive destruction of is known of the tissue-destructive process in rheumatoid nodules joint tissue components, loss of function and disability.1 or the contribution MMPs make towards inflammation in this Destruction of extra-articular tissues may also be observed, such extra-articular lesion. The expression of MMP-2, MMP-3 and as the development of subcutaneous rheumatoid nodules. MMP-9 proteins has been noted previously in rheumatoid 14 Nodules occur in 25–30% of cases and are generally associated nodules, as part of an active granulomatous inflammation. In with more severe disease.2 other granulomas, MMP-9 and MMP-12 emerge as mediators of The subcutaneous rheumatoid nodule is a necrotising the extensive connective tissue destruction and/or tissue 15–17 granuloma that has a highly organised and uniform structure.3 remodelling. In a comparison of gene expression between The centre of the nodule comprises necrotic tissue that includes inflamed RA synovium and subcutaneous nodules, our preliminary deposits of necrotic collagen and reticulin fibres admixed data highlighted two further members of the MMP enzyme family, with fibrinoid material, cellular debris and substantial amounts of MMP-7 (matrilysin-1; PUMP-1) and MMP-12 (macrophage S100 proteins.2,4 This necrotic centre is surrounded by a ‘palisade’ elastase), both with functions that may contribute to the of monocyte/macrophages infiltrating from the vascular peri- granulomatous inflammation and tissue destruction seen in phery.3 T cells and putative dendritic cells colocalise immediately rheumatoid nodules. outside the ‘palisade’ layer and in the perivascular area.5–7 While MMP-7 is distinct among the MMP family as it lacks the plasma cells may be present, there is no significant B-lymphocyte carboxyl-terminal haemopexin-like domain common to all other component to the inflammatory infiltrate in subcutaneous members that would normally function to restrict substrate 18 nodules.8 specificity. Consequently, MMP-7 possesses potent and broad Tissue destruction is a feature of both the rheumatoid nodule proteolytic activity. MMP-7 is produced by a variety of normal and joint synovial lesions in RA.8 The irreversible destruction of cells, including glandular and mucosal epithelial cells, 18 cartilage, tendon and bone that comprise the synovial joints is keratinocytes, fibroblasts and also by monocyte/macrophages 19–21 well documented. Inflammatory cytokines, including tumour at various stages of their differentiation. 22,23 necrosis factor (TNF)-a and interleukin (IL)-1b,9,10 stimulate the MMP-12 is expressed predominantly by macrophages production of matrix metalloproteinases (MMPs), including MMP-1 and is an important modulator of macrophage trafficking during

1Department of Physiology, Otago School of Medical Sciences, University of Otago, Dunedin, New Zealand; 2Department of Pathology, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand and 3Department of Medicine, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand. Correspondence: Dr PA Hessian, Department of Physiology, Otago School of Medical Sciences, University of Otago, Dunedin PO Box 913, Dunedin 9054, New Zealand. E-mail: [email protected] Received 15 June 2012; revised 22 November 2012; accepted 10 December 2012; published online 24 January 2013 Metalloproteinases in inﬂammation MG Kazantseva et al 163

Figure 1. Comparison of MMP gene expression in rheumatoid synovial and nodule tissues. (a) Comparisons of MMP-1, MMP-3, MMP-7 and MMP-12 gene expression in rheumatoid synovia (K; n ¼ 11) and rheumatoid nodule tissues (’; n ¼ 24). Gene expression was analysed using quantitative RT-PCR. All mRNA levels are expressed relative to GAPDH. Individual sample values are shown with the group mean±s.d. (horizontal bars); NS, not significant (P40.05); ***Po0.001 by Mann–Whitney, two-sided U-test. (b) Comparisons of MMP-7 gene expression in rheumatoid synovia (n ¼ 11) and nodule tissues with low MMP-7 (L) (n ¼ 17) or high MMP-7 (H) expression (n ¼ 7). High or low MMP-7 expression in the nodule tissue was distinguished by a nominal threshold, defined by the nodule group mean±1 s.e. NS, not significant (P40.05); **Po0.01; ***Po0.001 by Kruskal–Wallis Dunn’s multiple comparison test. Data for individual samples are shown as detailed above. inflammation, through its effect upon several matrix components, Table 1. Patient demographics including elastin and basement membrane proteins.23 Owing to the functional redundancy among MMP family Tissue specimen members, the biological relevance of an individual MMP is 24 primarily determined by its particular pattern of expression. In Nodule Synovium tissues affected by inflammation, multiple signalling pathways that involve glucocorticoids, transforming growth factor-b, TNF-a, Patients 24a 10b IL-1b, IL-6 or interferon (IFN)-g have been shown to regulate MMP Gender 20 Females 8 Females ± ± gene expression.25 Sequence variations within gene pro- Age (years) 58.8 11 57.8 14.4 Dis. duration (years) 17.2±10.8 15.0±9.4 moter regions are also associated with altered MMP gene 26–28 Nodules 24 (100%) 8 (80%) expression. For the MMP-7 promoter, both an A to G Erosions 23 (95.8%) 9 (90%) transition at the À 181-bp position ( À 181A4G; rs11568818), Rheumatoid factor 21 (87.5%) 9 (90%) and a C to T substitution at position À 153 ( À 153C4T; ACPA 14/15 (93.3%)c 7/8 (87.5%) rs11568819) have been shown to increase MMP-7 gene ESR 29.5±19.3 31.9±24.3 expression in vitro.26 Consequently, the activity of the À 181G Abbreviations: ACPA, anti-citrullinated peptide antibody; ESR, erythrocyte allele promoter region is about two- to threefold higher a 26 sedimentation rates. Unless indicated, all values are mean±s.d. Three with an associated increase in protein expression. In genetic paired synovial and nodule samples were included, obtained at the same association studies, the MMP-7 À 181A4G polymorphism time. bTwo synovial samples were from the same patient, taken B5 29–31 has been associated with increased risk of several cancers, months apart. cACPA values ware not available for all patients. coronary artery disease26 and, recently, with the degree of disability in RA.32 This study compared expression of the MMP-1, MMP-3, MMP-7 further study, we distinguished high or low MMP-7 expression and MMP-12 genes in RA synovial tissue with that in extra- as levels of gene expression above a nominal threshold (1.20 ng articular nodule lesions. Increased MMP-7 gene expression was of gene-specific RNA), defined by the nodule group mean±1s.e. found in a subgroup of rheumatoid nodules that was associated By this criterion, 7 of 24 nodules (29%) showed significantly with the À 181A4G polymorphism in the MMP-7 promoter region higher MMP-7 expression (2.73±1.73 ng RNA; mean±s.d.) when and manifest solely in nodules. This is first genotype–phenotype compared with remaining nodules with low MMP-7 expression study, which links the À 181A4G polymorphism and increased (0.14±0.19 ng RNA; Po0.001) or to levels of MMP-7 expression in MMP-7 expression in the tissue-destructive inflammatory lesions all synovial tissues (0.11±0.08; Po0.01; Figure 1b). of RA.

MMP-7 ( À 181A4G) promoter polymorphism is associated with increased MMP-7 expression in rheumatoid nodules RESULTS We further investigated whether a genetic predisposition under- MMP expression in rheumatoid synovial and nodule tissues lies the increased MMP-7 expression observed in a subgroup of Expression levels of the genes MMP-1, MMP-3, MMP-7 and MMP-12 rheumatoid nodule tissues. Two functional single-nucleotide were compared in rheumatoid synovial and nodule tissues polymorphisms (SNPs), À 181A4G and À 153C4T, that occur (Table 1) using real-time polymerase chain reaction (RT–PCR) in the promoter region of the MMP-7 gene were considered. analysis. Both MMP-1 and MMP-3 expressions were significantly Among the RA patient cohort studied, the À 153TT genotype was higher in synovia compared with nodule tissues (Po0.0001; not detected, regardless of MMP-7 expression levels. Overall, the Figure 1a). In contrast, MMP-12 expression was significantly higher analysis showed a different frequency of distribution for the in nodule tissues when compared with the synovial tissue MMP-7 À 181AG and À 181GG genotypes, with the latter (Po0.001; Figure 1a). Although the analysis showed no difference significantly more frequent among RA patients with high MMP-7 overall in MMP-7 gene expression between synovial and nodule expression in nodule tissue (P ¼ 0.015; Table 2). In contrast, tissues (P40.05; Figure 1a), we did observe noticeably higher MMP- there was no difference in the frequency of distribution for the 7 expression in a subgroup of individual nodule tissues. To facilitate MMP-7 À 153CT genotype (P40.05; Table 2).

& 2013 Macmillan Publishers Limited Genes and Immunity (2013) 162 – 169 Metalloproteinases in inﬂammation MG Kazantseva et al 164 Table 2. Frequency of MMP-7 SNPs in patients with RA donating nodules associated with high and low MMP-7 expression

Genotype Patient tissue contribution P-valuea

MMP-7 (H) MMP-7 (L) nodules nodules

MMP-7 À 181A/G À 181A/A 0 (0%)b 6 (35%) À 181A/G 2 (29%) 10 (59%) 0.52 À 181G/G 5 (71%) 1 (6%) 0.015

MMP-7 À 153C/T À 153C/C 5 (71%) 15 (88%) À 153C/T 2 (29%) 2 (12%) 0.55 À 153T/T 0 (0%) 0 (0%) UD Abbreviations: H, high; L, low; MMP, matrix metalloproteinase; RA, rheumatoid arthritis; SNP, single-nucleotide polymorphism; UD, undeter- mined. aP-value by Fisher’s exact test. bData are number of patients with the percentage of total indicated.

Mean MMP-7 expression level was significantly higher in rheumatoid nodule tissue from patients with RA homozygous for the G allele at position À 181 within the MMP-7 promoter when compared with the wild type, homozygous for the A allele (2.18±2.1 vs 0.06±0.06 ng RNA, respectively; Po0.01; Figure 2a). A low prevalence of the À 153C4T SNPs precluded the analysis of MMP-7 expression in relation to haplotypes. Figure 2. An association between the À 181A/G polymorphism and increased MMP-7 expression in rheumatoid nodule lesions. (a) MMP-7 The availability of paired synovial and nodule tissue samples gene expression (mean±s.d.) in nodules from 24 patients carrying allowed for further examination of any genotype–phenotype the AA (n ¼ 6), AG (n ¼ 12) and GG (n ¼ 6) genotype at position association. Analysis of paired synovial and nodule samples À 181. Gene expression was analysed using quantitative RT–PCR. All from three separate patients showed that increased MMP-7 mRNA levels are expressed relative to GAPDH. **Po0.01 using expression is restricted to nodule tissue in those patients Kruskal–Wallis test, with post-test analysis by Dunn’s multiple with a À 181GG genotype. These data suggest that the increased comparisons. (b, top panel) Mean MMP-7 expression in paired MMP-7 expression is dictated by a combination of the nodule synovial (S) and nodule tissues (N) from three separate patients, tissue microenvironment and MMP-7 ( À 181GG) promoter poly- labelled P1, P2 and P3. (Bottom panel) Restriction fragment length morphism (Figure 2b). polymorphism (RFLP) analysis of DNA extracted from the same paired tissue samples (see ‘Top panel’ above) associates patients’ MMP-7 genotype with gene expression in nodule tissue. Lane M: Inflammation features associated with MMP-7 expression 100-bp molecular marker. The proinflammatory cytokines, TNF-a and IFN-g, make important contributions to the inflammatory environment within the necrosis was similar in most tissues and appeared unaffected by rheumatoid nodule.8 However, we found no significant associa- the levels of MMP-7 expression. Accordingly, there was no tion between the expression of either the MMP-7 or MMP-12 genes significant difference in histological scores for granuloma forma- and the TNF or IFNG gene expression within rheumatoid nodules tion (P ¼ 0.2; Figure 3e). (data not shown). This was the case for nodules regardless of high With indications of more active inflammation within nodules or low MMP-7 expression. with high MMP-7 expression, we considered additional demo- To gain further insight into differences between nodules with graphics for the patients subgrouped by nodule MMP-7 expression high or low MMP-7 expression, nodule sections were assessed for (Table 3). Age at ‘biopsy’ between patients contributing nodules collagen content, neutrophil influx and the presence of granu- with high or low MMP-7 expression was the same. However, loma. Representative histology and the associated histological patients contributing nodules with high MMP-7 expression were scores are shown in Figure 3 (see also Supplementary Figures 1 significantly younger when diagnosed with RA and consequently and 2). The collagen in nodules with high MMP-7 expression was had significantly longer disease duration at the time of ‘biopsy’, more fibrinoid and ‘altered’ in appearance (Figure 3a and b) when indicating an earlier onset of RA. Measures of serum C-reactive compared with the collagen in nodules with low MMP-7 protein and erythrocyte sedimentation rates were similar between expression (Figure 3c and d). The latter had prominent areas of the two groups. There was no effect of current or former smoking intact and dense collagen fibres typical of mature collagen. on MMP-7 expression nor any association with methotrexate Corresponding histological scores reflect this appearance showing therapy (Table 3). significantly less amounts of intact and mature collagen in nodule tissues with high MMP-7 expression (arbitrary score 1.7±0.57 vs 2.5±0.55; mean±s.d.; P ¼ 0.04; Figure 3e). Nodules with high Different patterns of MMP-12 and MMP-7 protein expression in MMP-7 expression also showed a prominent neutrophil influx nodule tissue. (Figure 3b), observed throughout the tissue including perivascular Monocyte/macrophages are the principal inflammatory cells infil- and necrotic areas. Histological scores for neutrophil influx were trating the nodule.6,33 In nodules with high MMP-7 expression, significantly higher in nodules with high MMP-7 expression increased neutrophils mean a more varied myeloid cell infiltration. (arbitrary score 1.7±0.91 vs 0.58±0.58; P ¼ 0.05; Figure 3e). As These myeloid cells produce the S100A8 and S100A9 proteins,4,34 expected, the degree of granuloma development and associated a characteristic that was used to establish any links between

Genes and Immunity (2013) 162 – 169 & 2013 Macmillan Publishers Limited Metalloproteinases in inﬂammation MG Kazantseva et al 165

Figure 3. Histological features of nodule tissue associated with high and low MMP-7 expression. (a) Viewed at lower magnification ( Â 10), shows collagen fibrinoid in appearance and extensive infiltration of inflammatory cells in nodule tissue with high MMP-7 expression. A boxed area is further magnified ( Â 40) in (b) with infiltrating neutrophils indicated by arrowheads. (c) Viewed at lower magnification ( Â 10), shows prominent areas of intact and dense collagen fibres and diffuse inflammatory cell infiltrate in nodule tissue with low MMP-7 expression. A boxed area is further magnified ( Â 40) in (d). Scale bars 100 mm(a and c) and 20 mm(b and d). (e) Haematoxylin and eosin (H&E) sections of nodule tissue with high MMP-7 (H) (shaded; n ¼ 5) and low MMP-7 (L) (unshaded; n ¼ 6) expression were assessed for collagen content, neutrophil influx and granuloma with necrosis formation. Shown are the mean±s.d.; NS, not significant (P40.05); *Po0.05 by Kruskal–Wallis rank test.

MMP-7 or MMP-12 expression and the myeloid cell infiltration. Inflammatory cells producing MMP-7 or MMP-12 protein were Comparisons of gene expression in rheumatoid nodules found no identified using double immunofluorescence with a combination association between MMP-7 and either CAGA (encoding S100A8) of antibodies that specifically identified each MMP while or CAGB (S100A9) expression (data not shown). However, MMP-12 distinguishing CD14 þ monocyte/macrophages. Distinctly differ- expression was positively and significantly associated with S100A8 ent patterns of MMP-7 and MMP-12 protein expression were gene expression (Spearman’s correlation, r ¼ 0.49; P ¼ 0.015; evident in nodules. MMP-7 was produced by a subset of CD14 þ Supplementary Figure 3). monocyte/macrophages infiltrating the nodule (Figure 4). These

& 2013 Macmillan Publishers Limited Genes and Immunity (2013) 162 – 169 Metalloproteinases in inflammation MG Kazantseva et al 166 CD14 þ /MMP-7 þ monocyte/macrophages were located in the but when identified, these isolated individual cells did not innermost regions of the palisade layer (Figure 4). produce MMP-7. We also considered whether other inflammatory cells in the MMP-12 protein was restricted to CD14 þ monocyte/macro- nodule are responsible for MMP-7 production. CD3 þ T cells were phages in perivascular regions. MMP-12 was neither detected in generally restricted to the perivascular regions in the CD14 þ cells that had infiltrated the tissue beyond the perivascular outer vascular zone of the nodule and did not produce MMP-7 regions nor in any cells that lacked CD14 expression (Supple- protein. Similarly, CD20 þ B cells were rare in the nodule tissue, mentary Figure 3).

DISCUSSION The proteolytic activities of MMPs regulate tissue remodelling Table 3. Patient demographics associated with MMP-7 expression in (that is, repair and destruction) and influence the pro- and anti- nodules inflammatory signals generated during inflammation. In RA, Tissue specimen synovial tissue destruction is a well-documented consequence of MMP action and there is increasing recognition of roles for these MMP-7 (H) MMP-7 (L) P- proteases in mediating the actions of various chemokines and nodules nodules valuea cytokines.35,36 Much less is known of the contribution of MMPs to the inflammation and tissue necrosis that is a feature of the Patients 7 17 rheumatoid nodule. The aim of this study was to identify any Females 7 13 0.22 differences in the inflammatory environment between joint and Age at biopsy (years) 58.9±11.8b 58.8±11.1 0.986 Age at diagnosis (years) 33.1±15.4 47.9±14.1 0.036 extra-articular nodule inflammation in RA, concentrating Dis. duration at biopsy 27.1±10.5 12.6±7.5 0.001 particularly on the expression of the MMP-1, MMP-3, MMP-7 and (years) MMP-12 genes. The results revealed a differential contribution CRP 19.8±10.3 17.4±14.1 0.736 from MMPs towards joint and extra-articular inflammation in RA. ESR 38.8±22.0 25.9±17.9 0.216 In particular, a pattern of high MMP-7 gene expression was Current/former/non- 16.7/50/33.3c 23.5/35.3/ NSd identified, which was restricted to nodule tissue and associated smokers (%) 41.2 d with the À 181 A4G polymorphism (rs11568818). MTX (%) 57% 82% NS Our data show substantial expression of both MMP-1 and MMP-3 Abbreviations: CRP, C-reactive protein; ESR, erythrocyte sedimentation in synovial and nodule tissues. Several closely linked polymorph- rates; H, high; L, low; MMP, matrix metalloproteinase; MTX, methotrexate. isms across the MMP-1 and MMP-3 genetic loci are known to aP-value by unpaired Student’s t-test; bUnless indicated, all values are influence the expression of these two MMPs.28 Ultimately, the mean±s.d. cSmoking data were available on six patients contributing circulating levels of these enzymes in RA28 and, for MMP-3, the hi d 2 MMP-7 nodules. Significance determined by w test for combinations of level of disease activity in RA are affected.37 These associations are current smokers vs former smokers plus non-smokers or of current plus thought to reflect the local expression and activity of the MMPs former smokers vs non-smokers, and separately for patients receiving MTX within the joint, dictated by a tissue-specific environment, therapy within each nodule group. including inflammation. Although we did not assess MMP-1 or

Figure 4. Monocyte/macrophages produce MMP-7 protein in rheumatoid nodule tissue. (a) Representative haematoxylin and eosin (H&E) staining of a nodule tissue revealing a typical microstructure with a necrotic centre (NC) surrounded by a palisade layer (PL). Scale bar 200 mm. A boxed area is magnified further ( Â 40) in (b); scale bar 50 mm. (b) Double immunofluorescence staining of MMP-7 þ (green) and CD14 þ cells (red) with blue nuclei from Hoechst 33342 uptake. A boxed area targeting the area with MMP-7 þ and CD14 þ cells is magnified further ( Â 100) in ( Â and d); scale bars 20 mm. (c) MMP-7 þ cells stained green; and (d) double fluorescence staining (yellow) from CD14 þ monocyte/ macrophages (red) expressing MMP-7 þ (green).

Genes and Immunity (2013) 162 – 169 & 2013 Macmillan Publishers Limited Metalloproteinases in inflammation MG Kazantseva et al 167 MMP-3 polymorphisms, our expression data are consistent with mature,33 but not in the vascular region of nodules, indicates inflamed synovium as a site of significant MMP-1 and MMP-3 gene that these are not newly arrived cells. Consistent with expression in RA. Moreover, despite less expression in nodules, the this, we found no association between MMP-7 and S100A8 or actual amount of MMP-1 and MMP-3 transcript is substantial, S100A9 gene expression that would be expected from newly indicating that extra-articular sites of inflammation are a further arrived monocytes.40 Selective expression of MMP-7 in certain source of these two MMPs. It is also worth noting that we found a monocyte/macrophage subsets has been described previously, positive and significant association between MMP-1 and MMP-3 related to the site of inflammation and also to the degree gene expression within both synovial and nodule tissues. of associated inflammatory activity.19–21 In an analogous manner, Functioning MMP-3 activates a number of other pro-MMPs, and it appears that rheumatoid nodules may also represent a is critical for the processing of pro-MMP-1 to a fully active unique site of inflammation. Such a conclusion is reinforced enzyme.38,39 Although the function was not addressed, the by observations that MMP-7 expression within synovial association between MMP-1 and MMP-3 gene expression was tissues was consistently low, including in synovia where paired consistent with the concurrent activation of these enzymes nodules with high MMP-7 expression were available from the required for function. same patient. In contrast to MMP-1 and MMP-3 expression, we showed To assess the inflammatory activity within nodules, we significantly increased MMP-12 expression in nodules compared compared the expression of the genes encoding the proinflam- with RA synovial tissues, and overall that MMP-7 gene expression matory cytokines, TNF-a and IFN-g. These two cytokines are not was no different between the synovial and nodule groups. expected to influence MMP-7 expression,19 but they are both However, a critical observation was that a subgroup of nodules established as important mediators of the inflammatory expressed high levels of MMP-7 transcript, suggesting a role for environment within nodules and synovia.9,43 However, there was both MMP-7 and MMP-12 in nodule inflammation and tissue no difference in the expression of these genes between nodules remodelling. Interestingly, we found no association between with high MMP-7 expression and those nodules or synovium MMP-7 and MMP-12 gene expression in nodule tissue, suggesting where MMP-7 expression was lower. Differences in other that these enzymes are induced independently and have separate inflammatory indices, including a greater influx of neutrophils roles. Macrophages are an important source of MMP-1223 and we and change in collagen appearance, suggest that there is found that MMP-12 expression in nodules was associated with the heightened inflammatory activity in nodules with high MMP-7 expression of S100A8, suggesting that MMP-12 is derived from expression. All nodules included in this study were long-standing newly infiltrating monocyte/macrophages.40 Consistent with this lesions, removed for clinical reasons. Thus, it is more likely that the association, MMP-12 protein was restricted to CD14 þ monocyte/ differences in the inflammation between nodules with high or low macrophages in perivascular areas of the nodule tissue. This may MMP-7 expression reflect dynamic changes associated with underlie a direct role for MMP-12 in mediating macrophage ongoing or renewed inflammatory activity. Our results indicate trafficking into nodules similar to that shown for macrophage that within the context of chronic nodule inflammation, those infiltration of synovial tissue and the consequent development of nodules with high MMP-7 expression are more ‘active’. MMP-7 is joint inflammatory disease.41 An additional role for MMP-12 in known to promote neutrophil transmigration and has been nodule inflammation that involves chemokine cleavage cannot be proposed to modulate fibrosis signalling in other inflammatory discounted.42 settings.44,45 However, heightened inflammatory activity per se We sought evidence for why high MMP-7 gene expression was was not directly associated with the MMP-7 À 181A4G polymor- restricted to a subgroup of nodules and considered the phism, suggesting that high MMP-7 expression is a consequence importance of an underlying genetic component. Two functional rather than a cause of increased inflammation. Similarly, the polymorphisms, À 181A4G and À 153C4T, in the promoter infiltration of monocyte/macrophages producing MMP-7 protein is region of the MMP-7 gene have been shown to increase MMP-7 also a likely consequence of heightened inflammatory activity. expression in vitro.26 Consistent with this, we found that the Because MMP-7 displays strong proteolytic activity against a wide frequency of the G allele at position À 181 was significantly higher variety of extracellular matrix components and also non- in patients with high MMP-7 expression in nodules, and extracellular matrix bioactive molecules, like pro-TNF-a, some correspondingly that patients with the variant À 181GG geno- effect of MMP-7 on nodule pathology is likely. We found that the type show significantly higher MMP-7 expression in nodules more ‘active’ inflammation in nodules is independent of systemic than those with the wild-type AA genotype. This information measures of disease activity. Furthermore, there is no evidence identifies the presence of the À 181A4G polymorphism as a that smoking, which worsens established disease or that critical factor that underlies elevated MMP-7 gene expression in methotrexate therapy, which can predispose to nodule develop- rheumatoid nodules. No such association was found with variation ment, has any influence on MMP-7 expression in nodules. Rather, in genotype determined by the À 153C4T polymorphism. the onset of RA occurred much earlier in patients with high MMP-7 However, conclusive evidence for any effect of the variant expression in nodules. Thus, our data suggest a possible À 153TT genotype awaits further investigation as our patient association between the presence of the rs11568818 SNP and cohort did not contain any individual with this genotype. This is earlier onset of RA. This possibility also requires confirmation in a consistent with the low incidence of this genotype in the general larger patient cohort. population.26 Consequently, we were also unable to determine In summary, gene expression data suggest more prominent any combined contribution of the À 181A4G and À 153C4T roles for MMP-12 and MMP-7 in the pathogenesis of rheumatoid polymorphisms to haplotype that might influence levels of MMP-7 nodules than in the joint synovium. MMP-7 and MMP-12 gene expression. Despite the statistical significance of these expression reflects that monocyte/macrophages are the primary genetic data, the small sample size within genotype groups cellular component of nodule inflammation. We further show that requires that any conclusions are moderated. Clearly, a larger high and low MMP-7 gene expression provides a rare basis for sample is needed to gain more confidence in the association subdivision of the extra-articular nodule lesions in rheumatoid between the MMP-7 À 181A4G polymorphism and high MMP-7 inflammation. High MMP-7 gene expression in nodules is expression in nodules. predisposed by the À 181A4G polymorphism and associated We identified that a subset of CD14 þ monocytes/macrophages with heightened inflammatory activity in the lesion. Our data infiltrating the nodules produce MMP-7 protein. The location of indicate that consequences of the MMP-7 polymorphism are likely these CD14 þ /MMP-7 þ monocyte/macrophages in the innermost to manifest within aspects of immune/inflammatory activity that regions of the palisade layer, where they are expected to be more are monocyte/macrophage-mediated.

& 2013 Macmillan Publishers Limited Genes and Immunity (2013) 162 – 169 Metalloproteinases in inflammation MG Kazantseva et al 168 PATIENTS AND METHODS In interpreting the PCR-restriction fragment length polymorphism Patients and tissue samples analysis, the –181A and –153C homozygous wild-type alleles are represented as DNA bands of 150 and 160 bp, respectively; the –181G In all, 11 synovial membrane samples and 24 rheumatoid nodules were 46 and –153T polymorphic alleles are represented by DNA bands of 120 and obtained from 31 patients, all fulfilling the ACR revised criteria for RA. Patient demographics are summarised in Table 1. Additional demo- 130 bp, respectively. Correspondingly, heterozygous individuals display graphics are presented in Table 3 for those patients contributing nodules combinations of 150 and 120 bp for the –181 A4G polymorphism, and 160 with high or low MMP-7 expression. Synovial tissue was obtained during and 130 bp for the –153C4T polymorphism. The results of PCR-restriction joint replacement surgery and nodule tissue following elective surgical fragment length polymorphism analysis were confirmed by DNA sequen- cing of PCR products (Allan Wilson Centre, Massey University, Palmerston removal. Paired synovial and nodule samples were obtained at the same North, New Zealand; data not shown). time from three separate patients and two synovial samples were obtained from one other patient, approximately 5 months apart. Human tonsil RNA was used as reference. Ethics approval was obtained from the NZ Multi- Double immunofluorescence labelling Region Ethics Committee. Sections (7 mm) of frozen nodule tissues were cut and fixed in acetone for 10 min at 4 1C. After blocking with 5% normal donkey serum (Sigma, Castle Quantitation of gene expression by real-time PCR Hill, NSW, Australia) for 30 min at room temperature, the sections were incubated overnight at 4 1C with a goat polyclonal antibody against pro- and Total RNA was purified from RA synovial and nodule tissues using an active MMP-7 (diluted 1:50; R&D Systems, Minneapolis, MN, USA), followed by RNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s a 1.5 h incubation with AlexaFluor488-conjugated donkey anti-goat antibody instructions. Total RNA (1 mg) was reverse transcribed into complementary (diluted 1:1500; Invitrogen) at 41 C. The sections were then blocked with DNA using Superscript III (Invitrogen, Carlsbad, CA, USA) primed with normal sheep serum (Sigma) for 30 min at room temperature, and incubated oligo(dT)20. separately with cell-specific mouse monoclonal anti-CD3 (1:100; Dako, Expression of MMP-7, MMP-12, S100A8 and S100A9 genes was assessed Carpenteria, CA, USA), anti-CD20 (1:100; Dako) or anti-CD14 (clone FMC-17; by real-time PCR using SYBR GreenER gene expression assays (Invitrogen) undiluted culture supernatant) for 1 h at room temperature, followed by on an Applied Biosystems ABI 7300 Sequence Detecting System (Applied incubation with an AlexaFluor568-conjugated goat anti-mouse antibody Biosystems, Foster City, CA, USA). The SYBR GreenER assay primers were— (diluted 1:2000; Invitrogen) for 1.5 h at 41 C. For detection of pro-MMP-12 MMP-7 (sense): 50-GGAGGAGATGCTCACTTC GAT-30, MMP-7 (antisense): 0 0 0 antibody (R&D Systems) was diluted 1:10 and detected using combinations 5 -AGGAATGTCCCATACCCAAAGA-3 ; MMP-12 (sense): 5 -ACACCTGACATG of Alexa568-conjugated anti-goat and Alexa488-conjugated anti-mouse AACCGTGAG-30, MMP-12 (antisense): 50-ATGCCTGTGTTAATCTTGCTGA-30; 0 0 antibodies as described above. Cell nuclei were identified by counterstaining S100A8 (sense): 5 -CATGCCGTCTACAGGGATGA-3 , S100A8 (antisense): À 1 0 0 0 with Hoechst 33342 (2.5 mgml ). Negative controls omitted primary 5 -GACGTCTGCACCCTTTTTCC-3 ; and S100A9 (sense): 5 -GTGCGAAAAG antibodies. Staining was visualised by epifluorescence microscopy. ATCTGCAAAATTT-30, S100A9 (antisense): 50-GGTCCTCCATGATGTGTTC TATGA-30. Expression of MMP-1 and MMP-3 was assessed using either SYBR GreenER gene expression assays or commercially available TaqMan Differential tissue histology associated with MMP-7 expression in assays: MMP-1 (Hs00899658_m1) and MMP-3 (Hs00968305_m1) (Applied nodules Biosystems, Mulgrave, Australia). For the SYBR GreenER assays, the gene- Sections from rheumatoid nodules with high (n ¼ 5) or low (n ¼ 6) MMP-7 specific primers were—MMP-1 (sense): 50-ACTCTGGAGTAATGTCACACCT-30, expression were stained with a routine haematoxylin and eosin procedure, MMP-1 (antisense): 50-GTTGGTCCACCTTTCATCTTCA-30; and MMP-3 (sense): and then assessed in a ‘blinded’ manner by two investigators (NAH and 50-AGCAAGGACCTCGTTTTCATT-30, MMP-3 (antisense): 50-GTCAATCCCTGG Dr Tania Slatter, Department of Pathology, Dunedin, New Zealand). Each AAAGTCTTCA-30, Commercially available TaqMan assays were also used to section was scored for collagen content, the extent of neutrophil influx and measure TNF (Hs00174128_m1) and IFN-g (Hs00174143_m1) expression. the presence of granuloma with necrosis formation (Supplementary Test sample threshold cycle (Ct) values were extrapolated to standard Figures 1 and 2.) For collagen content, the scoring was on a scale of curves obtained from human tonsil RNA, to calculate the mean amount of 1 þ , where fibrinoid material predominated, through to 3 þ where gene-specific RNA in each sample. Using the appropriate assay, results for collagen fibrils were intact, aligned and mature. Tissues that intermediate individual genes were normalised to expression levels of the constitutively between these extremes were scored 2 þ . The extent of neutrophil influx expressed gene, glyceraldehede-3-phosphate dehydrogenase (GAPDH), for was scored 1 þ when only occasional sparsely distributed neutrophils individual samples. were present, 2 þ when neutrophils showed a patchy distribution and 3 þ when neutrophils were uniformly scattered throughout the tissue. Granuloma formation with accompanying necrosis was scored as 0 when Genotyping absent or 1 when present. Genomic DNA was isolated either from 100 mm cryosections of subcutaneous nodules tissue using a ChargeSwitch genomic DNA MicroTissue Kit (Invitrogen) or from peripheral blood using a QIAamp DNA blood mini Statistical analysis kit (Qiagen). Both methods followed the manufacturer’s instructions. The Mann–Whitney, two-sided U-test and Kruskal–Wallis test with post-test Genotyping for the MMP-7 À 181A4G and À 153C4T SNPs was analysis by Dunn’s multiple comparisons were used to analyse the results. performed using PCR-based restriction fragment length polymorphism Correlation coefficients were determined by rank correlation using non- analysis as described previously.26 Briefly, the MMP-7 promoter region, parametric Spearman’s test. Gender distribution and genotype frequencies from –300 to –22 bp, containing potential MMP-7 –181A4G and/or were compared using a non-parametric, two-sided Fisher’s exact test. –153C4T polymorphisms was amplified from genomic DNA and the PCR Histological scores were averaged and then analysed using a non- products were digested with restriction enzymes EcoRI and NlaIII, parametric Kruskal–Wallis rank test. respectively. For this analysis, the MMP-7 –181A4G primer sets were as All analyses were carried out using Prism 4 for Windows v.4.03 follows—MMP-7 –181 A4G (sense): 50-TGGTACCATAATGTCCTGAATG-30, (GraphPad Software Inc., La Jolla, CA, USA) and STATA11 (StataCorp, MMP-7 –181 A4G (antisense, mismatch primer): 50-TCGTTATTGGCAGG College Station, TX, USA) statistical software. A probability level of 5% was AAGCACACAATGAATT-30. This antisense primer was mutated from T to A at considered significant for all statistical analyses. the equivalent of position –177 bp (underlined) to create an EcoRI recognition site when the –181G allele exists. For analysis of the – 153C4T polymorphism, the sense mismatch primer was: 50-ACGAATACA ACKNOWLEDGEMENTS TTGTGTGCTTCCTGCCAATCA-30. This sense primer was mutated from A to C This work was supported by Arthritis New Zealand and the Health Research Council at the equivalent of –155 bp (underlined) to create an NlaIII recognition site of New Zealand. when the –153T allele exists. This was paired with the antisense primer: 50- TTTATATAGCTTCTCAGCCTCG-30. A16ml aliquot of PCR reaction was subjected to digestion at 37 1C for 3 h in a 20 ml reaction containing 10 U of EcoRI (Roche Applied Science, REFERENCES Mannheim, Germany) or NlaIII (New England Biolabs, Ipswich, MA, USA) as 1 Wolfe F, Michaud K, Gefeller O, Choi HK. Predicting mortality in patients with appropriate. Digestion products were separated on a 3% agarose gel and rheumatoid arthritis. Arthritis Rheum 2003; 48: 1530–1542. stained with Gel Red (Biotium, Hayward, CA, USA). 2 Ziff M. The rheumatoid nodule. Arthritis Rheum 1990; 33: 761–767.

Genes and Immunity (2013) 162 – 169 & 2013 Macmillan Publishers Limited Metalloproteinases in inflammation MG Kazantseva et al 169 3 Highton J, Hung N, Hessian P, Wilsher M. Pulmonary rheumatoid nodules 25 Parks WC, Wilson CL, Lopez-Boado YS. Matrix metalloproteinases as modulators of demonstrating features usually associated with rheumatoid synovial membrane. inflammation and innate immunity. Nat Rev Immunol 2004; 4: 617–629. Rheumatology 2007; 46: 811–814. 26 Jormsjo S, Whatling C, Walter DH, Zeiher AM, Hamsten A, Eriksson P. Allele- 4 Palmer DG, Hogg N, Allen CA, Highton J, Hessian PA. A mononuclear phagocyte specific regulation of matrix metalloproteinase-7 promoter activity is associated subset associated with cell necrosis in rheumatoid nodules: identification with with coronary artery luminal dimensions among hypercholesterolemic patients. monoclonal antibody 5.5. Clin Immunol Immunopathol 1987; 45: 17–28. Arterioscler Thromb Vasc Biol 2001; 21: 1834–1839. 5 Highton J, Kean A, Hessian PA, Thomson J, Rietveld J, Hart DN. Cells expressing 27 Jormsjo S, Ye S, Moritz J, Walter DH, Dimmeler S, Zeiher AM et al. Allele-specific dendritic cell markers are present in the rheumatoid nodule. J Rheumatol 2000; regulation of matrix metalloproteinase-12 gene activity is associated with cor- 27: 339–346. onary artery luminal dimensions in diabetic patients with manifest coronary 6 Duke OL, Hobbs S, Panayi GS, Poulter LW, Rasker JJ, Janossy G. A combined artery disease. Circ Res 2000; 86: 998–1003. immunohistological and histochemical analysis of lymphocyte and macrophage 28 Chen Y, Nixon NB, Dawes PT, Mattey DL. Influence of variations across the MMP-1 subpopulations in the rheumatoid nodule. Clin Exp Immunol 1984; 56: 239–246. and -3 genes on the serum levels of MMP-1 and -3 and disease activity in 7 Mellbye OJ, Forre O, Mollnes TE, Kvarnes L. Immunopathology of subcutaneous rheumatoid arthritis. Genes Immun 2012; 13: 29–37. rheumatoid nodules. Ann Rheum Dis 1991; 50: 909–912. 29 Li Y, Jin X, Kang S, Wang Y, Du H, Zhang J et al. Polymorphisms in the promoter 8 Highton J, Hessian PA, Stamp L. The rheumatoid nodule: peripheral or central to regions of the matrix metalloproteinases-1, -3, -7, and -9 and the risk of epithelial rheumatoid arthritis? Rheumatology 2007; 46: 1385–1387. ovarian cancer in China. Gynecol Oncol 2006; 101: 92–96. 9 Hessian PA, Highton J, Kean A, Sun CK, Chin M. Cytokine profile of the rheumatoid 30 Zhang J, Jin X, Fang S, Wang R, Li Y, Wang N et al. The functional polymorphism in nodule suggests that it is a Th1 granuloma. Arthritis Rheum 2003; 48: 334–338. the matrix metalloproteinase-7 promoter increases susceptibility to esophageal 10 Wikaningrum R, Highton J, Parker A, Coleman M, Hessian PA, Roberts-Thompson squamous cell carcinoma, gastric cardiac adenocarcinoma and non-small cell PJ et al. Pathogenic mechanisms in the rheumatoid nodule: comparison of lung carcinoma. Carcinogenesis 2005; 26: 1748–1753. proinflammatory cytokine production and cell adhesion molecule expression in 31 Peng B, Cao L, Ma X, Wang W, Wang D, Yu L. Meta-analysis of association between rheumatoid nodules and synovial membranes from the same patient. Arthritis matrix metalloproteinases 2, 7 and 9 promoter polymorphisms and cancer risk. Rheum 1998; 41: 1783–1797. Mutagenesis 2010; 25: 371–379. 11 Vincenti MP, Clark IM, Brinckerhoff CE. Using inhibitors of metalloproteinases to 32 Ye S, Patodi N, Walker-Bone K, Reading I, Cooper C, Dennison E. Variation in the treat arthritis. Easier said than done? Arthritis Rheum 1994; 37: 1115–1126. matrix metalloproteinase-3, -7, -12 and -13 genes is associated with functional 12 Ahrens D, Koch AE, Pope RM, Stein-Picarella M, Niedbala MJ. Expression of matrix status in rheumatoid arthritis. Int J Immunogenet 2007; 34: 81–85. metalloproteinase 9 (96-kd gelatinase B) in human rheumatoid arthritis. Arthritis 33 Palmer DG, Hogg N, Highton J, Hessian PA, Denholm I. Macrophage migration Rheum 1996; 39: 1576–1587. and maturation within rheumatoid nodules. Arthritis Rheum 1987; 30: 728–736. 13 Pakozdi A, Amin MA, Haas CS, Martinez RJ, Haines III GK, Santos LL et al. Mac- 34 Hessian PA, Edgeworth J, Hogg N. MRP-8 and MRP-14, two abundant Ca(2 þ )- rophage migration inhibitory factor: a mediator of matrix metalloproteinase-2 binding proteins of neutrophils and monocytes. J Leukoc Biol 1993; 53: 197–204. production in rheumatoid arthritis. Arthritis Res Ther 2006; 8: R132. 35 Brinckerhoff CE. Joint destruction in arthritis: metalloproteinases in the spotlight. 14 Matsushita I, Uzuki M, Matsuno H, Sugiyama E, Kimura T. Rheumatoid Arthritis Rheum 1991; 34: 1073–1075. nodulosis during methotrexate therapy in a patient with rheumatoid arthritis. 36 Mohammed FF, Smookler DS, Khokha R. Metalloproteinases, inflammation, and Mod Rheumatol 2006; 16: 401–403. rheumatoid arthritis. Ann Rheum Dis 2003; 62(Suppl 2): ii43–ii47. 15 Crouser ED, Culver DA, Knox KS, Julian MW, Shao G, Abraham S et al. Gene 37 Mattey DL, Nixon NB, Dawes PT, Ollier WE, Hajeer AH. Association of matrix expression profiling identifies MMP-12 and ADAMDEC1 as potential metalloproteinase 3 promoter genotype with disease outcome in rheumatoid pathogenic mediators of pulmonary sarcoidosis. Am J Respir Crit Care Med 2009; arthritis. Genes Immun 2004; 5: 147–149. 179: 929–938. 38 Suzuki K, Enghild JJ, Morodomi T, Salvesen G, Nagase H. Mechanisms of activation 16 Sheen P, O’Kane CM, Chaudhary K, Tovar M, Santillan C, Sosa J et al. High MMP-9 of tissue procollagenase by matrix metalloproteinase 3 (stromelysin). Biochemistry activity characterises pleural tuberculosis correlating with granuloma formation. 1990; 29: 10261–10270. Eur Respir J 2009; 33: 134–141. 39 Klein T, Bischoff R. Physiology and pathophysiology of matrix metalloproteases. 17 Vaalamo M, Kariniemi AL, Shapiro SD, Saarialho-Kere U. Enhanced expression of Amino Acids 2010; 41: 271–290. human metalloelastase (MMP-12) in cutaneous granulomas and macrophage 40 Ancuta P, Liu KY, Misra V, Wacleche VS, Gosselin A, Zhou X et al. Transcriptional migration. J Invest Dermatol 1999; 112: 499–505. profiling reveals developmental relationship and distinct biological functions of 18 Burke B. The role of matrix metalloproteinase 7 in innate immunity. Immuno- CD16 þ and CD16 À monocyte subsets. BMC Genomics 2009; 10:403. biology 2004; 209: 51–56. 41 Wang X, Liang J, Koike T, Sun H, Ichikawa T, Kitajima S et al. Overexpression of 19 Busiek DF, Baragi V, Nehring LC, Parks WC, Welgus HG. Matrilysin expression by human matrix metalloproteinase-12 enhances the development of inflammatory human mononuclear phagocytes and its regulation by cytokines and hormones. arthritis in transgenic rabbits. Am J Pathol 2004; 165: 1375–1383. J Immunol 1995; 154: 6484–6491. 42 Dean RA, Cox JH, Bellac CL, Doucet A, Starr AE, Overall CM. Macrophage-specific 20 Busiek DF, Ross FP, McDonnell S, Murphy G, Matrisian LM, Welgus HG. The metalloelastase (MMP-12) truncates and inactivates ELR þ CXC chemokines and matrix metalloprotease matrilysin (PUMP) is expressed in developing human generates CCL2, -7, -8, and -13 antagonists: potential role of the macrophage in mononuclear phagocytes. J Biol Chem 1992; 267: 9087–9092. terminating polymorphonuclear leukocyte influx. Blood 2008; 112: 3455–3464. 21 Halpert I, Sires UI, Roby JD, Potter-Perigo S, Wight TN, Shapiro SD et al. Matrilysin 43 Feldmann M, Maini RN. Lasker Clinical Medical Research Award. TNF defined as a is expressed by lipid-laden macrophages at sites of potential rupture in athero- therapeutic target for rheumatoid arthritis and other autoimmune diseases. Nat sclerotic lesions and localizes to areas of versican deposition, a proteoglycan Med 2003; 9: 1245–1250. substrate for the enzyme. Proc Natl Acad Sci USA 1996; 93: 9748–9753. 44 Li Q, Park PW, Wilson CL, Parks WC. Matrilysin shedding of syndecan-1 regulates 22 Shapiro SD, Kobayashi DK, Ley TJ. Cloning and characterization of a unique chemokine mobilization and transepithelial efflux of neutrophils in acute lung elastolytic metalloproteinase produced by human alveolar macrophages. J Biol injury. Cell 2002; 111: 635–646. Chem 1993; 268: 23824–23829. 45 Kelly MM, Leigh R, Gilpin SE, Cheng E, Martin GE, Radford K et al. Cell-specific gene 23 Shipley JM, Wesselschmidt RL, Kobayashi DK, Ley TJ, Shapiro SD. Metalloelastase expression in patients with usual interstitial pneumonia. Am J Respir Crit Care Med is required for macrophage-mediated proteolysis and matrix invasion in mice. 2006; 174: 557–565. Proc Natl Acad Sci USA 1996; 93: 3942–3946. 46 Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS et al. The 24 Sternlicht MD, Werb Z. How matrix metalloproteinases regulate cell behavior. American Rheumatism Association 1987 revised criteria for the classification of Annu Rev Cell Dev Biol 2001; 17: 463–516. rheumatoid arthritis. Arthritis Rheum 1988; 31: 315–324.

Supplementary Information accompanies this paper on Genes and Immunity website (http://www.nature.com/gene)

& 2013 Macmillan Publishers Limited Genes and Immunity (2013) 162 – 169