Intercritical circulating levels of neo-epitopes reflecting matrixmetalloprotease-driven degradation as markers of gout and frequent gout attacks

Transcription

1 RHEUMATOLOGY Concise report Rheumatology 2016;55: doi: /rheumatology/kew235 Advance Access publication 2 June 2016 Intercritical circulating levels of neo-epitopes reflecting matrixmetalloprotease-driven degradation as markers of gout and frequent gout attacks Ana M. Valdes 1, Tina Manon-Jensen 2, Abhishek Abhishek 1, Wendy Jenkins 1, Anne Sofie Siebuhr 2, Morten A. Karsdal 2, Sally Doherty 1, Weiya Zhang 1, Helen Richardson 1, Michael Doherty 1 and Anne-Christine Bay-Jensen 2 CLINICAL SCIENCE Abstract Objective. Recurrent flares constitute the main clinical burden of gout. Our aim was to assess whether biomarkers measuring MMP tissue degradation could be used as markers of frequent gout flares. Methods. Fasting plasma samples from 112 men with gout and 170 controls, along with serum samples from 447 men with gout collected at baseline from an ongoing clinical trial, were analysed by ELISA for neo-epitopes from MMP degradation of collagens type I (C1M) and type III (C3M). The log 10 levels of both markers were compared between cases and controls and between gout patients with three or more gout attacks in the past year and those with two or less attacks. Results. The circulating levels of C1M and C3M correlated with gout status in the case control study. Levels of both markers were associated with frequent gout flares (53 attacks in the past year) in both cohorts (odds ratio, OR = 3.1; 95% CI: 1.4, 6.8; P = for log 10 C1M, and OR = 6.7; 95% CI: 2.3, 19.3; P = for log 10 C3M). The area under the curve in a receiver operating characteristic analysis of frequent flares increased from 0.68 to 0.74 in one cohort and from 0.60 to 0.66 in the other when log 10 C1M and log 10 C3M were added to clinical variables of the model. Conclusion. C1M and C3M, reflective of interstitial matrix destruction, are associated with gout status and with frequent gout flares in men, suggesting that increased MMP activity may contribute to gout flares. Further research is needed to find out whether this is independent of dietary and lifestyle risk factors for acute gout. Key words: gout, matrix metalloprotease, interstitial matrix, flares, collagen neoepitopes, biochemical markers, tissue degradation, inflammation Rheumatology key messages. Markers reflecting matrix metalloprotease degradation of connective tissue correlate with frequent gout attacks in men.. The current study offers a novel insight into the molecular mechanisms underlying frequent gout flares.. Biomarkers can be used to identify patients at risk of frequent attacks of gout. Introduction 1 Academic Rheumatology, Clinical Sciences Building, University of Nottingham, Nottingham City Hospital, Nottingham, UK and 2 Rheumatology, Nordic Bioscience Biomarkers and Research, Herlev, Denmark Submitted 25 August 2015; revised version accepted 28 April 2016 Correspondence to: Ana M. Valdes, Academic Rheumatology, Clinical Sciences Building, University of Nottingham, Nottingham City Hospital, Hucknall Road, Nottingham, NG5 1PB, UK. Ana.Valdes@nottingham.ac.uk Chronic gout can cause joint damage and tophi, but recurrent acute attacks are its most common clinical manifestation. There is considerable interindividual variation in the frequency of acute attacks [1, 2]. Although gout attacks are precipitated by excess intake of purine rich food, alcohol, rapid urate lowering due to the institution of potent urate lowering treatment, acute illness, joint injury and! The Author Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please journals.permissions@oup.com

2 Biochemical markers of frequent gout attacks in men dehydration, many patients report that their gout attacks are spontaneous, with no obvious triggers [3]. Additionally, the evidence for the risk factors that are believed to trigger acute gout attacks is sparse, and there are no biomarkers that can identify people at risk of recurrent acute gout. Thus, there is need for understanding of the molecular mechanisms that predispose to frequent gout flares and for identifying patients at risk of frequent gout attacks in order to enable a targeted treatment strategy. Acute gout is associated with severe inflammation, which results in increased endopeptidase activity, such as that of MMPs, thus leading to elevated degradation of extracellular matrix proteins [4] of the connective tissue [5 7]. The interstitial matrix is one of the main local areas affected by inflammation, and the main components affected are types I and III collagen. MMP-2 and - 9 have been implicated in the degradation of extracellular matrix components in gout [5, 8]. Collagens type I (C1M) and C3M are neo-epitopes of types I and III collagen, both generated by MMP-2 and 9, and they have been shown to be elevated in OA, RA and spondyloarthritis [9]. The objective of this study was to investigate whether quantification of MMP-mediated destruction of interstitial matrix components, as assessed by the two neo-epitope biomarkers C1M and C3M, was associated with gout and with frequency of gout flares (53 vs <3 attacks per year). Methods Study participants from two separate studies were included. The current study did not require ethical approval because patients from the randomized control trial (RCT) and case control study gave consent for the re-use of samples for additional biomarker analyses. We carried out a case control study of 112 men with gout and 170 controls without gout or OA that were recruited from general practice surgeries in Nottinghamshire, UK. Gout patients were also identified from rheumatology clinics. Inclusion criteria were gout ascertained by rheumatologist clinical assessment. All the patients met ACR classification criteria [10]. Exclusion criteria were RA, lupus, PsA, renal failure, terminal illness, mental health problems, previous history of large joint replacement or inability to give informed consent. Fasting plasma samples were collected and stored at 80 C. The research protocol for the case control study was approved by the Nottinghamshire Research Ethics Committee 1 (REC reference 08H0403/36). We also undertook a community-based RCT involving a total of 447 men recruited consecutively from general practice surgeries in Nottinghamshire, Derbyshire and South Yorkshire, UK, as part of an ongoing clinical trial. Inclusion criteria were: gout ascertained by clinical assessment that met ACR classification criteria, and having suffered at least one gout attack in the past year. Exclusion criteria were the same as for the case control study (vide supra). Before randomization, individuals gave informed consent and then underwent a structured gout assessment (for signs of peripheral joint inflammation and damage, and subcutaneous tophi) and gave a blood sample. Serum samples were collected and stored at 80 C. The research protocol for the communitybased RCT was approved by the NRES Committee East Midlands Nottingham 1 (Rec Ref 12/EM/0044 Clinical Trials reference ). Informed consent given for the two cohorts was also valid for this study. Biomarker assessments Two different collagen degradation biomarkers were assessed, using solid phase competitive ELISAs developed and produced by Nordic Biosciences [11]. The assays measure MMP-generated neo-epitope fragments of C1M and type III (C3M). In brief, a 96-well streptavidin plate was coated with the appropriate biotinylated synthetic peptide dissolved in assay buffer and incubated for 30 min at 20 C. Peptide calibrator or sample was added to appropriate wells, and then 100 ml of a conjugated mab against the target sequence was added to wells and incubated. The mix of primary antibody and samples was incubated for 2 h at 20 C or 20 h at 4 C. After thorough washing of the microtitre plate, 100 ml tetramethylbenzinidine (KemEn-Tec cat.438oh) was added, and the plate was incubated for 15 min at 20 C in the dark. All of the above incubation steps included shaking at 300 rpm. After each incubation step, the plate was washed five times in washing buffer (20 mm Tris, 50 mm NaCl, ph 7.2). The tetramethylbenzinidine reaction was stopped by adding 100 ml of stopping solution (1% HCl), and absorbance was measured at 450 nm, with 650 nm as the reference. The samples were assessed in duplicates, including five quality control samples on each microtitre plate. The intra- and inter-assay coefficients of variation were below 15%. Statistical analysis The biomarker data were log 10 -transformed to achieve a normal distribution. Logistic regression analyses were carried out on gout status and on occurrence of three or more attacks in the past year without adjustment. The cut-off of three was chosen because it has been shown to relate to health-care utilization [12] and is the median annual number of attacks in the study population. Regressions were adjusted for age and BMI (multivariate analysis), additional confounders (age at first attack, presence of tophi and number of sites affected by tophi available in both cohorts) and confounders available only in the RCT (serum urate levels). In order to assess the discriminating power of the biomarkers, receiver operating characteristic curves were generated using prevalence of three or more gout attacks in the past year as an outcome for cases from both cohorts. All statistical analysis were performed using the statistical package R ( version 3.1.2, and P < 0.05 was considered statistically significant. Results The descriptive characteristics of the two study cohorts are presented in Table 1. Mean age and BMI were similar

3 Ana M. Valdes et al. TABLE 1 Descriptive characteristics of the study cohorts and association between connective tissue degradation neoepitopes and gout and frequent gout flares Case control cohort Randomized clinical trial Trait Controls Gout patients Controls Gout patients N Gender, male % Age, mean (S.D.), years 66.2 (9.3) 63.1 (10.7) 62.8 (11.4) BMI, mean (S.D.), kg/m (4.5) 29.8 (4.8) 30.4 (5.3) Had gout attacks last year, % N/A attacks in past year, N (%) N/A 31 (27.7) 262 (58.6) Presence of tophi, % N/A C1M, mean (S.D.), ng/ml 47.5 (39.5) a 56.9 (38.7) a 31.7 (25.4) b C3M, mean (S.D.), ng/ml 24.5 (7.1) a 28.4 (7.8) a 24.3 (10.9) b Association with Association with log 10 C1M Gout, OR (95% CI), P-values 4.10 (1.51, 11.1), < gout flares unadjusted, 5.19 (0.99, 27.1), OR (95%CI), P-values < gout flares adjusted c, 4.96 (0.86, 28.8), OR (95% CI), P-values < gout flares adjusted d, OR (95% CI), P-values log 10 C3M Association with log 10 C1M Association with log 10 C3M 82.8 (9.9, 695) < (1.26, 2564), 2.88 (1.22, 6.80), 6.51 (2.22, 19.1), < <0.037 < (0.65, 1803), 2.83 (1.17, 6.81), 5.21 (2.05, 18.7), < <0.080 < (1.71, 15.87), < (1.11, 6.51), <0.027 a Levels measures in fasting serum samples. b Levels measured in plasma. c Adjusted for age, BMI, age at first attack, sites with tophi. d Adjusted for age, BMI, age at first attack, serum uric acid levels, tophi, sites with tophi. N: number. across groups investigated, although the gout patients were on average more obese than the controls. In the case control study, 57% had experienced an attack during the last year, while approximately half of those had 3 or more attacks within that period, with an average of 2.2 attacks. In the RCT, 59% had experienced 3 or more attacks during the past year, with an average of 4.1 attacks. Plasma levels of C1M were significantly higher than serum C1M levels, whereas the difference between plasma and serum C3M was not statistically significant. The distribution of the log 10 C1M and log 10 C3M values in gout cases and controls is shown in Supplementary Fig. S1 A and B, available at Rheumatology online. Association between the biomarkers, gout and flares In the case control study, both C1M and C3M levels were associated with gout (Table 1). Next, we investigated their association with frequency of gout attacks. On univariate analysis, the levels of both C1M and C3M were significantly associated with frequent gout flares in both studies (Table 1). The significance was attenuated after adjustment for age, BMI and age at first attack. A meta-analysis of results from both cohorts yielded an odds ratio (OR) of 3.11 (95% CI: 1.42, 6.81; P = ) for log 10 C1M and OR = 6.70 (95% CI: 2.33, 19.3; P = ) for log 10 C3M. We therefore investigated the potential additional role of C1M and C3M in identifying patients with frequent flare (Table 2). In both cohorts, the inclusion of C3M and C1M improved the area under the curve considerably, suggesting that biomarkers of connective tissue turnover may on the one hand provide important clues to the cause of frequent gout flares, and on the other hand may be used for diagnostic purposes in identifying those at a high risk of recurrent acute gout attacks. The distribution of the log 10 C1M and log 10 C3M levels in gout cases in relation to the number of attacks suffered in the past year is presented in Supplementary Fig. S1 C F, available at Rheumatology Online, for both cohorts. The receiver operating characteristics curves for these analyses are presented in Supplementary Fig. S2, available at Rheumatology Online. Finally, we investigated whether C1M and C3M are merely reflecting circulating levels of serum uric acid (SUA) or a relationship to the presence of tophi. We found no significant correlation between SUA and log 10 C1M (r = , P < 0.43) or log 10 C3M (r = , P < 0.98). There was also no correlation between presence of tophi and log 10 C1M (r = , P < 0.80) or log 10 C3M (r = , P < 0.72). Discussion There have been no tools available for serological quantification of local tissue damage in acute gout. The data in this study suggest that inflammation of the interstitial matrix, leading to increased MMP activity (and consequent release of the two main proteins of the interstitial matrix, types I and III collagen), could be associated with novel biomarkers of local tissue damage. This is the first study to report molecular markers not correlated with SUA that are strongly associated with frequent gout flares in

4 Biochemical markers of frequent gout attacks in men TABLE 2 Receiver operating characteristics analysis results using prevalence of three or more gout attacks in the past year as the outcome in both studies Model Cases from case control RCT cohort Clinical variables only a AUC 0.679(95% CI: 0.565, 0.793) AUC 0.596(95% CI: 0.544, 0.649) Clinical variables + log 10 C1M AUC 0.734(95% CI: 0.642, 0.847) AUC 0.649(95% CI: 0.598, 0.699) Clinical variables + log 10 C3M AUC 0.712(95% CI: 0.602, 0.822) AUC 0.651(95% CI: 0.6, 0.701) Clinical variables + log 10 C1M + log 10 C3M AUC 0.736(95% CI: 0.631, 0.84) AUC 0.663(95% CI: 0.613, 0.713) The AUC and 95% CIs are shown for all four models tested. a Clinical variables are: age, BMI, age at first attack, presence of tophi and number of sites with tophi for the cases in the case control study; age, BMI, age at first attack, presence of tophi and serum uric acid levels for the gout patients from the RCT. AUC: area under curve. two independent cohorts. Circulating levels of these two neo-epitopes of connective tissue degradation by MMPs were also associated with gout status in the case control study. The inflammatory reaction typical of an acute gout attack is initiated by MSU crystals [13 15]. The activation of resident cells by MSU induces the synthesis of several inflammatory mediators, and induces cells of the vasculature to express potent chemokines (e.g. IL-8 [14]) and adhesion molecules responsible for the massive recruitment of neutrophils to the joint [15]. The recruitment of a large number of neutrophils to the affected joint during the effector phase is the pathological hallmark of gouty arthritis [13]. A recent report has shown that circulating levels of IL-8 are increased during both the acute and intercritical phases of gout [14]. IL-8 regulates the release of MMP9 from neutrophils [15]. The increased levels of C1M and C3M, reflecting the degradation of collagen types 1 and 3 by MMPs in gout patients compared with controls, is therefore consistent with the increased levels of IL-8 in gout patients reported by others. A possible explanation for the correlation seen between circulating levels of C1M and C3M during the intercritical phase with frequent gout flares, but not with tophi or SUA levels, is that these are markers of increased neutrophil activation by IL-8, and of a generally higher inflammatory state that is not directly dependent on SUA levels. Our study offers important insight into the molecular mechanisms that accompany frequent gout attacks, and holds promise for the development of additional biomarkers for identifying gout patients at a high risk of developing acute attacks of gout. We have identified markers linked to inflammation but not to the presence of tophi. However, we did not test for cartilage or bone damage; therefore, we cannot exclude the possibility that the two neo-epitopes tested may be reflecting joint damage. Inflammation-driven MMP activity results in tissue destruction, with the subsequent release of small protein fragments such as types I and III collagen by MMPs, measured by C1M and C3M. The biomarkers studied here are not specific to gout and have been shown to be both diagnostic and prognostic in an array of diseases with local tissue inflammation and destruction (such as RA, OA and AS [16]). For example, C1M was found to be correlated with disease activity (i.e. DAS28), as well as predictive for structural progression in RA [6, 7]. Serum C3M levels are also elevated in a subgroup of patients with inflammatory OA [9, 17] and in patients with AS (compared with controls) [7, 11, 16, 17]. In the diseases mentioned above (except OA, which is not autoimmune), the autoimmune response results in inflammation of the interstitial matrix. We note some study limitations. Both cohorts included only men, and it is possible, though unlikely, that results may be different in women. The majority of gout patients however are men, and even if results were different between genders, the value of these markers would remain. We see strong differences between the mean values of plasma and serum samples. Although serum and heparinized plasma specimens are considered equivalent for many assays, differences in results between these sample types have been reported for many chemistry analytes, including albumin, alkaline phosphatase, calcium, carbon dioxide, chloride, creatine kinase, glucose, lactate dehydrogenase, inorganic phosphorus, potassium and total protein [18]. Nonetheless, in spite of the differences seen between serum and plasma, the association with frequent gout flares is seen in both types of specimens, suggesting that these biomarkers can be used in both cases but separate reference intervals for plasma samples need to be established. The study is not prospective, so we do not know the true prognostic value of these biomarkers, and the frequent gout attacks were self-reported. Because of the retrospective nature of the data analysed, we cannot distinguish whether increased MMP activity was contributing to the onset of a gout flare or whether increased MMP activity was a result of gout flares. Another limitation is that not all participants had crystal confirmation, though all fulfilled ACR criteria for gout. On the other hand, our study shows that technically the results should be generalizable because we have included both fasting plasma and non-fasting sera and the results were consistent. Further prospective studies are required to find out whether these biomarkers are associated with the development of gout in people with hyperuricaemia, and with frequent acute attacks of gout or with structural joint damage in those with gout. Quantification of tissue damage holds promise for the development of additional

5 Ana M. Valdes et al. biomarkers for identifying gout patients at a high risk of developing frequent acute attacks of gout. Funding: This work was supported by the European Commission Framework Programme 7 programmes D- BOARD and EurHEALTHAgeing, and by Arthritis Research UK grant number Disclosure statement: A.-S.S. is a full-time employee at Nordic Bioscience. T.M.-J. is an employee of Nordic Bioscience, a privately owned biotechnology company involved in the development of biomarkers. M.D. has received honoraria for participation in advisory boards for gout and osteoarthritis from AstraZeneca, Nordic Biosciences and Novartis and Academic Rheumatology Nottingham receives an Investigator-led research grant on gout from AstraZeneca. A.-C.B.-J. is a full-time employee and share-holder of Nordic Bioscience, a privately owned biotech company. M.A.K. is a full time employee of and holds shares in Nordic Bioscience. All other authors have declared no conflicts of interest. Supplementary data Supplementary data are available at Rheumatology Online. References 1 Doherty M. New insights into the epidemiology of gout. Rheumatology 2009;48 (Suppl 2):ii Rothenbacher D, Primatesta P, Ferreira A, Cea-Soriano L, Rodriguez LA. Frequency and risk factors of gout flares in a large population-based cohort of incident gout. Rheumatology 2011;50: Neogi T, Chen C, Niu J et al. Alcohol quantity and type on risk of recurrent gout attacks: an internet-based casecrossover study. Am J Med 2014;127: Zhen EY, Brittain IJ, Laska DA et al. Characterization of metalloprotease cleavage products of human articular cartilage. Arthritis Rheum 2008;58: Chu SC, Yang SF, Lue KH et al. The clinical significance of gelatinase B in gouty arthritis of the knee. Clin Chim Acta 2004;339: Siebuhr AS, Wang J, Karsdal M et al. Matrix metalloproteinase-dependent turnover of cartilage, synovial membrane, and connective tissue is elevated in rats with collagen induced arthritis. J Transl Med 2012;10: Bay-Jensen AC, Platt A, Byrjalsen I et al. Effect of tocilizumab combined with methotrexate on circulating biomarkers of synovium, cartilage, and bone in the LITHE study. Semin Arthritis Rheum 2014;43: Karsdal MA, Bay-Jensen AC, Leeming DJ, Henriksen K, Christiansen C. Quantification of end products of tissue destruction in inflammation may reflect convergence of cytokine and signaling pathways implications for modern clinical chemistry. Biomarkers 2013;18: Siebuhr AS, Petersen KK, Arendt-Nielsen L et al. Identification and characterisation of osteoarthritis patients with inflammation derived tissue turnover. Osteoarthritis Cartilage 2014;22: Singh JA, Solomon DH, Dougados M et al. Development of classification and response criteria for rheumatic diseases. Arthritis Rheum 2006;55: Bay-Jensen AC, Leeming DJ, Kleyer A et al. Ankylosing spondylitis is characterized by an increased turnover of several different metalloproteinase-derived collagen species: a cross-sectional study. Rheumatol Int 2012;32: Saseen JJ, Agashivala N, Allen RR et al. Comparison of patient characteristics and gout-related health-care resource utilization and costs in patients with frequent versus infrequent gouty arthritis attacks. Rheumatology 2012;51: Gagné V, Marois L, Levesque JM et al. Modulation of monosodium urate crystal induced responses in neutrophils by the myeloid inhibitory C-type lectin-like receptor: potential therapeutic implications. Arthritis Res Ther 2013;15:R Kienhorst LB, van Lochem E, Kievit W et al. Gout is a chronic inflammatory disease in which high levels of interleukin 8 (CXCL8), myeloid-related protein 8/myeloidrelated protein 14 complex, and an altered proteome are associated with diabetes mellitus and cardiovascular disease. Arthritis Rheumatol 2015;67: Chakrabarti S, Patel KD. Regulation of matrix metalloproteinase-9 release from IL-8 stimulated human neutrophils. J Leukoc Biol 2005;78: Bay-Jensen AC, Byrjalsen I, Siebuhr AS et al. Serological biomarkers of joint tissue turnover predict tocilizumab response at baseline. J Clin Rheumatol 2014;20: Bay-Jensen AC, Wichuk S, Byrjalsen I et al. Circulating protein fragments of cartilage and connective tissue degradation are diagnostic and prognostic markers of rheumatoid arthritis and ankylosing spondylitis. PLoS One 2013;8:e Miles RR, Roberts RF, Putnam AR, Roberts WL. Comparison of serum and heparinized plasma samples for measurement of chemistry analytes. Clin Chem 2004;50: