ORIGINAL ARTICLE
Neutrophil Gelatinase B–associated Lipocalin and Matrix Metalloproteinase-9 Complex as a Surrogate Serum Marker of Mucosal Healing in Ulcerative Colitis Magali de Bruyn, MSc,* ,† Ingrid Arijs, PhD,† Willem-Jan Wollants, Dipl Ing,† Kathleen Machiels, MSc,† Kristel Van Steen, PhD,‡ Gert Van Assche, MD, PhD,† Marc Ferrante, MD, PhD,† Paul Rutgeerts, MD, PhD,† Séverine Vermeire, MD, PhD,† and Ghislain Opdenakker, MD, PhD*
Background: The current standard for the assessment of mucosal healing after therapy in inflammatory bowel diseases is endoscopy. However, a high need exists for noninvasive, accurate surrogate markers.
Methods: In 2 independent cohorts, levels of serum neutrophil gelatinase B–associated lipocalin and matrix metalloproteinase-9 complex (NGAL– MMP-9) from patients with active ulcerative colitis (UC) before and after first treatment with infliximab and from healthy controls (HC) were determined with zymography and sandwich enzyme-linked immunosorbent assay. The response to infliximab was defined as complete mucosal healing (Mayo endoscopic subscore 0–1) at control endoscopy. Data were analyzed with SPSS, and P values ,0.05 were considered significant.
Results: In cohort 1 (n ¼ 66; median age, 30 yr; 38% female), serum NGAL–MMP-9 levels significantly increased at baseline in UC patients versus HC
(103.8 versus 42.4 ng/mL; P , 0.0001), whereas 55% of the patients had normal C-reactive protein levels. NGAL–MMP-9 levels significantly decreased after therapy in UC responders (from 116.3 ng/mL to 32.0 ng/mL; P , 0.0001) and in nonresponders (from 94.7 ng/mL to 54.1 ng/mL; P ¼ 0.047). In cohort 2 (n ¼ 132; median age, 39 yr; 53% female), NGAL–MMP-9 levels increased at baseline in active UC patients versus HC (86.5 versus 60.4 ng/ mL; P ¼ 0.10), whereas 45% of the patients had normal C-reactive protein levels. NGAL–MMP-9 levels significantly decreased after therapy in responders (from 87.5 ng/mL to 16.3 ng/mL; P , 0.0001) but not in nonresponders (from 82.7 ng/mL to 57.8 ng/mL; P ¼ 0.19). After pooling the data, a cutoff value of 97.7 ng/mL for NGAL–MMP-9 complex was determined to predict complete mucosal healing with high specificity (91%).
Conclusions: Serum NGAL–MMP-9 is suggested as a new surrogate marker for the assessment of mucosal healing in UC patients treated with infliximab. (Inflamm Bowel Dis 2014;20:1198–1207) Key Words: NGAL–MMP-9 complex, ulcerative colitis, mucosal healing, surrogate serum marker, neutrophil
O
ne of the challenges in inflammatory bowel disease (IBD) research is to have accurate biomarkers with predictive
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.ibdjournal.org). Received for publication November 5, 2013; Accepted April 3, 2014. From the *Department of Microbiology and Immunology, Rega Institute for Medical Research, KU Leuven, Leuven, Belgium; †Department of Clinical and Experimental Medicine, Translational Research Center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium; and ‡Montefiore Institute, Electrical Engineering and Computer Science, University of Liège, Liège, Belgium. I. Arijs and M. Ferrante are postdoctoral fellows and S. Vermeire is a Senior Clinical Investigator of the Fund for Scientific Research of Flanders (FWOVlaanderen). M. de Bruyn is a PhD student funded by a fellowship from the Agency for Innovation by Science and Technology (IWT). P. Rutgeerts, S. Vermeire, G. Van Assche and M. Ferrante report following conflicts of interest: grant support, lecture fees and consulting fees from Centocor and Schering-Plough. Supported by FWO-Vlaanderen and “Geconcerteerde Onderzoeksacties GOA 2013-014”. Reprints: Ghislain Opdenakker, MD, PhD, Department of Microbiology and Immunology, Rega Institute for Medical Research, Minderbroedersstraat 10, blok x-bus 1030, Leuven 3000, Belgium (e-mail: [email protected]). Copyright © 2014 Crohn’s & Colitis Foundation of America, Inc. DOI 10.1097/MIB.0000000000000068 Published online 27 May 2014.
1198
| www.ibdjournal.org
value, correlating with disease activity, responsive to the effect of treatment, and preferentially playing a defined pathophysiological role in the disease. At present, mainly C-reactive protein (CRP) and erythrocyte sedimentation rate are used as blood markers for IBD, whereas calprotectin and lactoferrin are useful fecal markers. Nonetheless, no single biomarker has been proven to be ideal for all proposed indications.1 Ulcerative colitis (UC) is mostly characterized by influx of neutrophils to the active site of inflammation.2 Being specialized leukocytes, neutrophils have an end-stage differentiation and contain preformed granules.3 Secondary and tertiary neutrophil granules contain, respectively, neutrophil gelatinase B–associated lipocalin ([NGAL], also called lipocalin-2) and matrix metalloproteinase-9 ([MMP-9], also known as gelatinase B). Furthermore, MMP-9 can form a covalent complex with NGAL. However, the functionality of this complex is still debated. One hypothesis is that by formation of this complex, NGAL protects MMP-9 from autodegradation,4 whereas it was also proposed that NGAL supports the allosteric activation of pro-MMP-9.5 In contrast, the functions of both individual proteins are well studied. Inflamm Bowel Dis Volume 20, Number 7, July 2014
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL6,7 is expressed in response to the activation of Tolllike receptors during infections,8,9 and it has been shown to inhibit bacterial growth by sequestering iron-laden siderophores.10 NGAL protein or messenger RNA expression levels have been correlated with parameters of active IBD.11–16 MMP-9, as a member of the MMP family, is a zinc-dependent endopeptidase involved in many developmental processes, including angiogenesis, wound healing, and extracellular matrix degradation. Despite the involvement in many normal physiological processes, MMP-9 has been associated with abnormal disease conditions17,18 and is considered a tuner and amplifier of inflammatory reactions.17,19,20 MMP-9 levels have been shown to be elevated in the feces of UC patients and to correlate well with disease activity.21 Recently Sela-Passwell et al22 have shown that neutralizing antibodies with tissue inhibitor of MMPslike mechanisms against MMP-2 and MMP-9 can attenuate the development of colitis in IBD mouse models. Studies investigating the decrease of NGAL or MMP-9 after treatment with infliximab were mostly performed in Crohn’s disease (CD) patients and not in UC patients.23,24 Moreover, the role of NGAL–MMP-9 as a complex has only been investigated in 1 study, indicating elevated levels of NGAL–MMP-9 in the urine of pediatric IBD patients.11 Because frequent endoscopic assessments are costly and uncomfortable for the patient, a high need exists for noninvasive, accurate markers. In contrast with CD patients, in whom CRP is an accurate marker, CRP is normal in more than 50% of UC patients. With this study, we aimed to investigate for the first time levels of NGAL–MMP-9 complex in the serum of patients with active UC before and after first treatment with infliximab and to compare the accuracy of NGAL–MMP-9 complex with CRP in the assessment of mucosal healing.
MATERIALS AND METHODS Patient Sampling In a retrospective study, performed at the Leuven University Hospital, sera from 2 independent cohorts of active UC patients before and after first administration of infliximab (Remicade; Centocor, Horsham, PA) and from healthy controls (HC) were investigated. Cohort 1 included 66 infliximab-naive patients with active UC and 40 HC (median age, 30.4 yr; 38% female). In cohort 2, 132 active UC patients who received a first dose of infliximab and 129 HC (median age, 38.8 yr; 53% female) were included. The baseline UC patient characteristics of the 2 cohorts are shown in Tables 1 and 2. The patients underwent endoscopy within a week before the first intravenous infliximab infusion of 5 mg/kg body weight, and a second endoscopy was performed after 8 to 14 weeks. The response to treatment with infliximab was determined at the time of the second endoscopy and defined as complete mucosal healing when the Mayo endoscopic subscore was 1 or 0. Patients were defined as nonresponders when they did not reach this complete healing, although some of them showed endoscopic improvement. Histological activity was determined by a panel of expert pathologists on biopsies
NGAL–MMP-9 as a Surrogate Serum Marker in UC
taken at the time of endoscopy before and after treatment with infliximab. Histological activity was scored according to the pathologic assessment reports as severely active or active inflammation (¼2), mildly active inflammation (¼1), and no inflammation (¼0). Histological improvement was defined as no inflammation (score 0) after the treatment. From all UC patients, CRP levels and neutrophil counts were measured in the serum before and after infliximab therapy, at time points corresponding to the time of endoscopy. From all individuals, a written informed consent was obtained, and the study was approved by the University Hospital Ethics Committee (Vlaams erfelijkheidsonderzoek Crohn en colitis ulcerosa S-53684).
Gelatin Zymography Analysis To define alterations in the different molecular forms of gelatinases, sera from UC patients and HC from cohort 1 were first analyzed with gelatin zymography analysis.25 Briefly, 100 mL of serum was prepurified using gelatin Sepharose beads (GE Healthcare, Buckinghamshire, United Kingdom) and mini spin columns (Bio-Rad Laboratories, Hercules, CA). The bound gelatinases were eluted from the column with 20 mL Tris/glycine/ sodium dodecyl sulfate nonreducing loading buffer (Invitrogen, Carlsbad, CA), and the prepurified samples were then separated in 7.5% acrylamide gels copolymerized with 1 mg/mL porcine gelatin (Sigma-Aldrich, St Louis, MO).26 The gels were then washed with 2.5% Triton X-100 (Sigma-Aldrich) for 40 minutes and incubated overnight at 378C in 50 mM Tris-HCl (pH 7.5) supplemented with 10 mM CaCl2 (Sigma-Aldrich). The gels were stained with 0.25% Coomassie Brilliant Blue-R (Sigma-Aldrich) and scanned using standard settings. Band densities of all individual gelatinase forms (MMP-2, MMP-9 multimers, NGAL– MMP-9 complexes, and MMP-9 monomers) were analyzed and quantified with Image J 1.48 software (NIH Windows version), and the obtained densities of different gelatinase forms were plotted as a ratio versus pro-MMP-2 levels.
Sandwich Enzyme-linked Immunosorbent Assay All serum NGAL–MMP-9 complex levels were determined with the use of a commercial anti-human NGAL–MMP-9 complex enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Abingdon, UK) and assayed according to the manufacturer’s guidelines. This ELISA incorporates 2 antibodies with different antigen specificities: one against MMP-9 and the other against NGAL. Hence, with this ELISA only the NGAL–MMP-9 complexes are measured. The absorbance was measured at 450 nm with a spectrophotometer. NGAL–MMP-9 complex levels were quantified with the use of a calibration curve of human NGAL–MMP-9 complex as a standard.
Statistical Analysis Data were analyzed with SPSS Statistics 20.0 software (SPSS, Inc, Chicago, IL) with the use of the nonparametric Mann–Whitney U test for unpaired samples (significance indicated by *) and Wilcoxon’s signed rank test for paired samples www.ibdjournal.org |
1199
Inflamm Bowel Dis Volume 20, Number 7, July 2014
de Bruyn et al
TABLE 1. Cohort 1 Patient Characteristics and Laboratory Markers at the Start of First Treatment with Infliximab for UC Responders and Nonresponders Baseline Patient Characteristics of Cohort 1
UC Responders (n ¼ 28)
UC Nonresponders (n ¼ 38)
Pa
15/13 (54/46) 38.7 (30.4–51.1) 6.1 (1.3–16.7)
25/13 (66/34) 42.4 (30.6–58.2) 6.9 (2.0–12.1)
0.44b 0.44c 0.94c
NA 10/28 (36) 18/28 (64)
NA 14/38 (37) 24/38 (63)
0/21 8/21 13/21 16/28 3/28 4.9 16/28 5.2
1/32 8/32 23/32 23/38 5/38 4.9 20/38 5.5
Male/female (%) Median (IQR) age at first infliximab (yr) Median (IQR) duration of disease before first infliximab (yr) Disease extent before starting infliximab therapy (%) E1 (proctitis) E2 (left-sided colitis) E3 (pancolitis) Histological activity score at first infliximab (%) 0 (inactive) 1 (mildly active) 2 (active) Mayo endoscopic subscore 3 before first infliximab (%) Active smoking at first infliximab (%) Median (IQR) CRP at first infliximab (mg/L) Normal CRP (,5 mg/L) levels at baseline (%) Median (IQR) amount of neutrophils (109/L) Concomitant medication at first infliximab (%) 5-ASA Intravenous/systemic oral corticosteroids Immunosuppressant (AZA or 6-MP)
(0) (38) (61) (57) (11) (1.0–14.2) (57) (3.4–7.3)
20/28 (71) 10/28 (36) 16/28 (57)
(3) (25) (72) (61) (13) (1.9–11.8) (53) (3.9–8.2)
35/38 (92) 12/38 (32) 12/38 (32)
NA 1b 1b 1b 0.37b 0.55b 0.81b 1b 0.58c 0.80b 0.61c 0.04a,b 0.79b 0.05a,b
P , 0.05 was considered significant. The statistical difference was analyzed by Fisher’s exact test. c The statistical difference was analyzed by Mann–Whitney U test. IQR, interquartile range; 5-ASA, 5-aminosalicylates; AZA, azathioprine; 6-MP, mercaptopurine. a
b
(significance indicated by ^). Spearman correlation analysis (correlation coefficient ¼ r), Kendall’s tau rank correlation (correlation coefficient ¼ t), Fisher’s exact test, and receiver operating characteristic (ROC) analysis were also performed in SPSS. Influences of demographic and clinical variables on NGAL–MMP-9 levels were first tested using univariate (Mann–Whitney U test) and bivariate analyses (Spearman’s rho or Kendall’s tau). Second, independent associations and influences of confounding variables were assessed using a linear regression model applying a forward selection approach. Multicolinearity was explored by determining the variance inflation factor (VIF) and tolerance value. A VIF value greater than 5 and a tolerance value smaller than 0.20 implied colinearity. P values of ,0.05 were considered significant.
RESULTS Serum NGAL–MMP-9 Complex Levels Are Increased at Baseline in Active UC Patients and Decrease After Treatment With Infliximab mers,
1200
To visualize all gelatinase forms (MMP-2, MMP-9 monoMMP-9 multimers, and NGAL–MMP-9 complex), | www.ibdjournal.org
zymography analysis was performed on prepurified serum samples from 66 UC patients (28 responders and 38 nonresponders) before and after therapy with infliximab and from 40 HC (cohort 1) (Fig., Supplemental Digital Content 1, http://links.lww.com/IBD/A481). A template of zymography analysis is shown in Figure 1, whereby different gelatinase forms can be distinguished. In fact, this test measures the levels of 4 analytes at once, and with appropriate standardization, it yields a richer information content than any single ELISA.25 At baseline, NGAL–MMP-9 serum levels from patients of cohort 1 with active UC were significantly elevated in comparison with HC levels (P , 0.01). After treatment with infliximab, NGAL–MMP-9 complex levels significantly decreased in both UC responders (P , 0.0001) and nonresponders (P ¼ 0.011), although the decrease was more profound for UC responders (Fig. 2A). Pro-MMP-9 multimer and monomer levels were also significantly decreased after treatment with infliximab, whereas activated forms of MMP-9 and MMP-2 were rarely detectable (Table 3). Zymography is a useful technique to visualize various molecular forms and to measure relative gelatinase levels; however, it is a laborious technique. Therefore, we validated our analyses of the NGAL–MMP-9 complex with sandwich ELISA
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 as a Surrogate Serum Marker in UC
TABLE 2. Cohort 2 Patient Characteristics and Laboratory Markers at the Start of First Treatment with Infliximab for UC Responders and Nonresponders Baseline Patient Characteristics of Cohort 2
UC Responders (n ¼ 88)
UC Nonresponders (n ¼ 44)
48/40 (55/45) 39.7 (28.4–50.7) 5.5 (2.3–10.6)
28/16 (64/36) 43.4 (32.9–53.0) 7.0 (1.4–12.6)
0.35b 0.39c 0.82c
2/88 (2) 31/88 (35) 55/88 (65)
NA 15/44 (34) 29/44 (66)
NA 1b 0.85b
2/72 44/72 26/72 44/88 9/88 6.0 39/88 5.2
0/34 26/34 8/34 24/44 5/44 6.4 17/44 5.7
1b 0.52b 0.40b 0.71b 1b 0.12c 0.58b 0.43c
Male/female (%) Median (IQR) age at first infliximab (yr) Median (IQR) duration of disease before first infliximab (yr) Disease extent before starting infliximab therapy (%) E1 (proctitis) E2 (left-sided colitis) E3 (pancolitis) Histological score at the start of infliximab therapy (%) 0 (inactive) 1 (mildly active) 2 (active) Mayo endoscopic subscore 3 before first infliximab (%) Active smoking at first infliximab (%) Median (IQR) CRP at first infliximab (mg/L) Normal CRP (,5 mg/L) levels at baseline (%) Median (IQR) amount of neutrophils (109/L) Concomitant medication at first infliximab (%) 5-ASA Intravenous/systemic oral corticosteroids Immunosuppressant (AZA or 6-MP)
(3) (61) (36) (50) (10) (1.6–16.6) (44) (3.9–7.3)
59/88 (67) 33/88 (38) 50/88 (57)
(0) (77) (23) (55) (11) (3.7–27.9) (39) (4.0–8.1)
Pa
0.01a,b 0.71b 0.002a,b
39/44 (89) 18/44 (41) 12/44 (27)
P , 0.05 was considered significant. The statistical difference was analyzed by Fisher’s exact test. c The statistical difference was analyzed by Mann–Whitney U test. IQR, interquartile range; 5-ASA, 5-aminosalicylates; AZA, azathioprine; 6-MP, mercaptopurine. a
b
(Fig. 2B) and thus confirmed that serum NGAL–MMP-9 complex levels were elevated at baseline in active UC patients as compared with HC (P , 0.0001). After treatment with infliximab, NGAL– MMP-9 complex serum levels were significantly decreased in UC responders (median range from 116.3 to 32.0 ng/mL; P , 0.0001), whereby only 3 responders had unchanged levels of NGAL–MMP-9 after therapy. Overall, serum NGAL–MMP-9 complex levels decreased in 90% of the UC responders. For UC nonresponders, only a borderline significant decrease was observed (median range from 94.7 to 54.1 ng/mL; P ¼ 0.0469), which was significantly less pronounced in comparison with UC responders (P ¼ 0.0074). For almost 40% of the UC nonresponders, NGAL–MMP-9 complex serum levels increased instead of decrease after treatment. Importantly, the NGAL–MMP-9 complex serum levels of UC responders decreased to a significantly lower level after treatment than the levels of UC nonresponders (P ¼ 0.03). Finally, a strong correlation was found between zymography analysis of the NGAL–MMP-9 complex and sandwich ELISA (r ¼ 0.835; P , 0.0001) (Fig. 3). To corroborate the data from UC patient in cohort 1, we retrospectively selected 132 UC patients (88 responders and 44 nonresponders) for an independent cohort 2 and performed
sandwich ELISA on serum samples collected before and after treatment with infliximab. As shown in Figure 2C, an increase in NGAL–MMP-9 complex levels was found at baseline for active UC patients versus HC (86.5 versus 60.4 ng/mL; P ¼ 0.10). After treatment, a significant decrease of NGAL–MMP-9 levels was found for UC responders (median range from 87.5 to 16.3 ng/mL; P , 0.0001). More specifically, 83% of the UC responders had decreased NGAL–MMP-9 levels after treatment. However, NGAL–MMP-9 levels did not significantly decrease after treatment for UC nonresponders (median range from 82.7 to 57.7 ng/mL; P ¼ 0.19), whereby 41% of the UC nonresponders had increased or similar NGAL–MMP-9 levels after treatment. Moreover, NGAL–MMP-9 levels of UC responders decreased to a significantly lower level after treatment than the levels of UC nonresponders after treatment (P , 0.0001). Because cohort 2 confirmed and validated the results from cohort 1, the data from both cohorts were pooled. As shown in Figure 2D, pooled data from 129 HC, 116 UC responders, and 82 UC nonresponders confirm that serum NGAL–MMP-9 complex levels significantly increased at baseline in active UC patients versus HC (87.3 ng/mL versus 60.4 ng/mL; P ¼ 0.012). Moreover, NGAL–MMP-9 levels significantly decreased after treatment www.ibdjournal.org |
1201
de Bruyn et al
Inflamm Bowel Dis Volume 20, Number 7, July 2014
patients who had active disease at baseline, which was associated with higher levels of NGAL–MMP-9, were treated effectively with corticosteroids. Importantly, these confounding characteristics were not different between patients who responded to infliximab and those who did not respond to infliximab (Tables 1 and 2).
Serum NGAL–MMP-9 Complex Levels Correlate With Neutrophil Counts and Complement CRP as an Inflammatory Marker FIGURE 1. Template of zymography analysis showing different gelatinase forms in prepurified samples. (Pro) MMP-9 multimers are detected at the highest molecular weights (in kilodaltons), followed by NGAL–MMP-9 complex, pro-MMP-9 monomers, MMP-9 monomers, pro-MMP-2, and MMP-2. This template illustrates gelatinase levels of a HC and a UC responder and nonresponder patient from cohort 1. Serum samples from UC patients before and after infliximab therapy were loaded in adjacent lanes. R1, serum sample of a UC responder patient before infliximab therapy; R2, serum sample of the same UC responder patient after infliximab therapy; NR1, serum sample of a UC nonresponder before infliximab therapy; NR2, serum sample of the same UC nonresponder after infliximab therapy.
with infliximab for UC responders (median range from 90.8 to 20.7 ng/mL; P , 0.0001) and to a lesser extent for UC nonresponders (median range from 86.1 to 56.2 ng/mL; P ¼ 0.02). NGAL–MMP-9 levels decreased to significantly lower levels after treatment for UC responders compared with UC nonresponders (P , 0.0001). Furthermore, the influence of several demographic and clinical variables on serum NGAL–MMP-9 levels was determined. We found that the concomitant use of corticosteroids (intravenous or oral systemic) and disease activity (as was assessed by the Mayo endoscopic subscore) showed a correlation with serum NGAL–MMP-9 levels at the start of infliximab therapy (t ¼ 0.215; P , 0.0001 and t ¼ 0.255; P , 0.0001, respectively). No significant differences were found for disease extent (P ¼ 0.281), histological activity (P ¼ 0.601), concomitant treatment with 5-aminosalicylates (P ¼ 0.295), or immunosuppressants (P ¼ 0.077). Univariate linear regression analysis confirmed our findings for association of the use of corticosteroids (P ¼ 0.001) and disease activity (P ¼ 0.003) with NGAL–MMP-9 levels. Furthermore, also gender (female) (P ¼ 0.06) and concomitant treatment with immunosuppressants (P ¼ 0.095) were moderately associated with NGAL–MMP-9 levels at baseline. In the multiple linear regression model, we included gender, disease activity at the start of infliximab therapy, and concomitant treatment with corticosteroids or immunosuppressants. The analysis indicated that the concomitant use of corticosteroids (b ¼ 0.201; P ¼ 0.004) and disease activity at baseline (b ¼ 0.177; P ¼ 0.012) had an influence on serum NGAL–MMP-9 levels at baseline. VIF and tolerance showed that there was no colinearity between both parameters (VIF ¼ 1.03 and tolerance ¼ 0.96). These data underline the fact that
1202
| www.ibdjournal.org
To assess its role as a serum marker of inflammation and mucosal healing, we correlated NGAL–MMP-9 complex serum levels with currently used blood markers in IBD. Correlations were made with neutrophil counts and CRP levels. In cohort 1, the results indicated that NGAL–MMP-9 complex levels correlated well (r ¼ 0.430; P , 0.0001) with neutrophil counts, in line with the fact that NGAL–MMP-9 is a product of neutrophils. A more moderate correlation was found with CRP levels (r ¼ 0.266; P ¼ 0.002). More than half (57%) of the UC responders had no elevated CRP levels (,5 mg/L) at baseline; of whom, only 1 patient also had normal (,42.2 ng/mL) NGAL–MMP-9 complex levels at the start of anti–tumor necrosis factor (TNF) treatment. In UC nonresponders, 53% had normal CRP levels at baseline; of whom, only 9 patients (24%) also had normal levels of NGALMMP-9. These results indicated that, especially for UC responders, serum NGAL–MMP-9 complex is more representative of the disease activity than CRP. In cohort 2, NGAL–MMP-9 levels significantly correlated with the amount of neutrophils (r ¼ 0.463; P , 0.01) and CRP levels (r ¼ 0.477; P , 0.01). In 45% of the UC responder and nonresponder patients, CRP levels were not elevated at baseline (,5 mg/L). Half of the patients with normal CRP levels also had NGAL–MMP-9 levels below the median HC level (60.4 ng/mL). After pooling the data, correlation factors of NGAL–MMP-9 with CRP (r ¼ 0.429; P , 0.01) and neutrophil counts (r ¼ 0.414; P , 0.01) were determined (Fig. A and B, Supplemental Digital Content 2, http://links.lww.com/ IBD/A482). Overall, in 48% of the active UC patients, CRP was not elevated (,5 mg/L) at baseline; of whom, 36% of the patients also had normal (,60.4 ng/mL) NGAL–MMP-9 levels.
Correlation Analysis With Mayo Endoscopic Subscores Indicates That Serum NGAL–MMP-9 Complex Levels Constitute a Surrogate Marker of Mucosal Healing For patients from cohort 1, serum NGAL–MMP-9 complex levels significantly correlated with Mayo endoscopic subscores (t ¼ 0.317; P , 0.01), and the highest NGAL–MMP-9 complex levels were observed in patients with Mayo endoscopic subscores of 2 and 3. After therapy with infliximab, NGAL–MMP-9 levels decreased rapidly and correlated with restoration of mucosal integrity. Interestingly, the mean NGAL–MMP-9 complex levels did not differ between Mayo endoscopic subscores 0 and 1. When pointwise Mayo endoscopic subscore gradation was related to
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 as a Surrogate Serum Marker in UC
FIGURE 2. Decrease of serum NGAL–MMP-9 complex levels after therapy with infliximab as measured by zymography analysis (A), whereby NGAL–MMP-9 levels are represented as a ratio versus pro-MMP-2 levels, and sandwich enzyme-linked immunosorbent assay analysis (B–D). A, B, Serum NGAL–MMP-9 levels from patients of cohort 1. C, Serum NGAL–MMP-9 levels from patients of cohort 2. D, The pooled data from cohorts 1 and 2. Asterisks (*) indicate significance by nonparametric Mann-Whitney U test for unpaired samples. Arrowheads (^) indicate significance by Wilcoxon’s signed rank test for paired samples. ^P , 0.05 or *P , 0.05; ^^P , 0.01; ** P , 0.01; ^^^P , 0.001; ***P , 0.001.
CRP levels, a more moderate correlation was obtained (t ¼ 0.299; P , 0.01). Whereas we observed a gradual increase in NGAL– MMP-9 complex levels with increasing Mayo endoscopic subscores, CRP was only markedly increased in patients who had a Mayo endoscopic subscore of 3. A similar analysis was performed identifying the correlation of neutrophil counts with Mayo endoscopic subscores (t ¼ 0.233; P , 0.01). Finally, ROC analysis was performed whereby NGAL– MMP-9 complex values corresponding to Mayo endoscopic subscores 0 or 1 were compared with NGAL–MMP-9 complex levels corresponding to Mayo endoscopic subscores 2 or 3. The
analysis showed that NGAL–MMP-9 complex levels lower than 97.7 ng/mL were predictive of mucosal healing with a specificity and sensitivity of 93% and 43%, respectively. The area under the curve (AUC) was 0.75, and a positive predictive value (PPV) of 96% and a negative predictive value (NPV) of 31% were determined. ROC analysis with CRP levels corresponding to Mayo endoscopic subscores 0 to 1 and 2 to 3, respectively, was also performed. The obtained ROC curve was comparable to the curve plotted for NGAL–MMP-9 serum levels (AUC difference of 0.70 versus 0.75; P ¼ 0.43). However, the clinical cutoff value of 5 mg/L for CRP corresponded to www.ibdjournal.org |
1203
Inflamm Bowel Dis Volume 20, Number 7, July 2014
de Bruyn et al
TABLE 3. Cohort 1 Serum Levels of Gelatinase Forms as Determined by Zymography, for HC and UC Responders or Nonresponders, Before and After Treatment with Infliximab Ratio of Gelatinase Form Versus Pro-MMP-2
HC (n ¼ 40), Median (IQR)
Pro-MMP-9 multimers 0.53 (0.38–0.83) NGAL–MMP-9 complex 0.34 (0.15–0.68) Pro-MMP-9 monomers 1.67 (1.24–2.63) MMP-9 monomers ND MMP-2 ND
UC Nonresponders (n ¼ 38), Median (IQR)
UC Responders (n ¼ 28), Median (IQR) a
Before Infliximab
After Infliximab
P
Before Infliximab
After Infliximab
Pa
0.80 (0.58–1.12) 0.77 (0.50–1.36) 2.08 (1.54–2.77) ND ND
0.52 (0.38–0.72) 0.25 (0.14–0.52) 1.39 (1.12–1.90) ND ND
0.0004b ,0.0001b ,0.0001b — —
0.77 (0.49–1.27) 0.59 (0.42–1.19) 2.09 (1.44–3.24) ND ND
0.66 (0.46–0.93) 0.41 (0.22–0.73) 1.73 (1.30–2.40) ND ND
0.023b 0.017b 0.013b — —
P , 0.05 was considered significant. The statistical difference was analyzed by Wilcoxon’s signed rank test. ND, not detectable; IQR, interquartile range.
a
b
a lower specificity (82%) for the prediction of mucosal healing than NGAL–MMP-9 (93%). In cohort 2, previous correlations were validated because serum NGAL–MMP-9 complex levels (t ¼ 0.382; P , 0.01) and CRP levels (t ¼ 0.408; P , 0.01) significantly correlated with Mayo endoscopic subscores. A similar analysis was performed
FIGURE 3. Correlation plot of serum NGAL–MMP-9 complex levels as measured from patients of cohort 1 by sandwich ELISA and zymography analysis. NGAL–MMP-9 serum levels of UC responders and nonresponders before and after therapy with infliximab are shown in red and blue, respectively. HC samples are indicated as black diamonds, nonresponders as circles, and responders as triangles. R1, serum sample of a UC responder patient before therapy; R2, serum sample of a UC responder patient after therapy; NR1, serum sample of a UC nonresponder before therapy; NR2, serum sample of a UC nonresponder after therapy.
1204
| www.ibdjournal.org
identifying the correlation of neutrophil counts with Mayo endoscopic subscores (t ¼ 0.368; P , 0.01). Finally, ROC analysis was performed whereby NGAL–MMP-9 complex levels lower than 93.2 ng/mL were predictive of mucosal healing with a specificity and sensitivity of 91% and 44%, respectively. The AUC was 0.78, and a PPV of 41% and NPV of 70% were determined. ROC analysis for CRP levels corresponding to Mayo endoscopic subscores 0 to 1 and 2 to 3, determined a cutoff value of 7.2 mg/L, predicting mucosal healing with 44% sensitivity, 93% specificity, 46% PPV, and 72% NPV, with an AUC of 0.80. However, the clinical cutoff value of 5 mg/L for CRP corresponded to a lower specificity (86%) for the prediction of mucosal healing than NGAL–MMP-9 (91%). Pooling the data from cohorts 1 and 2, serum NGAL– MMP-9 complex levels (t ¼ 0.370; P , 0.01) and CRP levels (t ¼ 0.372; P , 0.01) significantly correlated with Mayo endoscopic subscores (Fig. 4A, B). A similar analysis was performed identifying the correlation of neutrophil counts with Mayo endoscopic subscores (t ¼ 0.308; P , 0.01) (data not shown). ROC analysis of the pooled data determined an AUC of 0.78 and a cutoff value of 97.7 ng/mL for NGAL–MMP-9 complex, predicting mucosal healing with 43% sensitivity, 91% specificity, 92% PPV, and 40% NPV (Fig. 4C). For CRP, a similar AUC of 0.77 was determined, and a cutoff value of 7.3 mg/L was determined, predicting mucosal healing with 39% sensitivity, 91% specificity, 91% PPV, and 38% NPV (Fig. 4D). Finally, when considering the clinical cutoff value of 5 mg/L for CRP, a lower specificity (85%) was found for the prediction of mucosal healing than for NGAL–MMP-9 (91%).
Correlation Analysis With Histological Activity Scores Indicates That Serum NGAL–MMP-9 Complex Levels Constitute a Surrogate Marker of Histological Improvement In a subset of patients (n ¼ 103), information on histological activity before and after treatment with infliximab was also
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 as a Surrogate Serum Marker in UC
FIGURE 4. A, B, Correlations of NGAL–MMP-9 complex levels and CRP levels with Mayo endoscopic subscores. A, Correlation of NGAL–MMP-9 complex values from all UC patients (pooled cohorts 1 and 2) with the corresponding Mayo endoscopic subscores, regardless of their response to therapy or time point before or after the treatment. The dotted line represents the mean NGAL–MMP-9 complex level of HC subjects (60.4 ng/mL). B, Correlation of CRP levels from all UC patients with Mayo endoscopic subscores. The dotted line represents the clinical cutoff value of 5 mg/L. Mean values along with the standard deviation of the mean are represented in function of the corresponding Mayo endoscopic subscores. C, D, ROC analysis of serum NGAL–MMP-9 complex values and CRP levels corresponding to Mayo endoscopic subscores 0 to 1 versus Mayo endoscopic subscores 2 to 3. C, A cutoff value of 97.7 ng/mL for NGAL–MMP-9 was determined to predict mucosal healing with 43% sensitivity, 91% specificity, 92% PPV, and 40% NPV. D, For CRP levels, a cutoff value of 7.3 mg/L was determined to predict mucosal healing with 39% sensitivity, 91% specificity, 91% PPV, and 38% NPV (D).
analyzed. Histological activity scores (severely active or active ¼ 2, mildly active ¼ 1, inactive ¼ 0) were determined by a panel of expert pathologists on biopsies taken at the time of endoscopy at the start of infliximab therapy and after treatment. Correlation analysis of histological activity with NGAL–MMP-9 levels showed a significant correlation (t ¼ 0.218; P , 0.0001), and NGAL–MMP-9 levels gradually increased for every score point (linear trend test, P , 0.0001) (Fig. A, Supplemental Digital Content 3, http://links.lww.com/IBD/A483). Moreover, a significant correlation was found between histological improvement and mucosal healing (t ¼ 0.465; P , 0.0001) and between the endoscopic Mayo score and histological score (t ¼ 0.505; P , 0.0001). Furthermore, ROC analysis was performed to evaluate the performance of NGAL–MMP-9 to discriminate between histological activity (active and mildly active) (n ¼ 182) and no activity (n ¼ 24). The AUC (95% confidence interval) for
NGAL–MMP-9 levels was 0.713 (0.627–0.799). A cutoff value of 65 ng/mL for NGAL–MMP-9 was able to predict histological healing with 52% sensitivity and 91% specificity (Fig. B, Supplemental Digital Content 3, http://links.lww.com/IBD/A483).
DISCUSSION Biomarkers are crucial to aid clinicians in IBD diagnosis and management. Currently available blood and fecal biomarkers provide reproducible and quantitative tools to distinguish IBD from noninflammatory diarrhea and assess disease activity.27 However, the discovery of novel biomarkers that correlate with mucosal healing or predict long-term disease course would improve current clinical assessment greatly. With the aim to critically assess serum gelatinase levels in UC patients, we first used the labor-intensive gelatin zymography www.ibdjournal.org |
1205
de Bruyn et al
analysis25 to measure multiple analytes simultaneously (MMP-2, MMP-9 monomers, MMP-9 multimers, and NGAL–MMP-9 complex). Hence, the information content is much richer than that of a single parameter. Zymography may also reveal—individually— the activation forms and degradation products of the studied enzymes, whereas with ELISA, all these molecules are measured as a mixture, without information about single molecular species. Zymography analysis, performed on serum samples before and after treatment with infliximab from cohort 1, indicated that the NGAL–MMP-9 complex was intact and that the levels were significantly higher in active UC patients at baseline compared with HC. This indicated that NGAL–MMP-9 is an inflammatory marker related to disease activity. Moreover, the covalently formed complex of MMP-9 with NGAL significantly decreased after treatment with infliximab in UC responders and to a lesser extent in nonresponders. It needs to be stressed that the NGAL–MMP-9 is a readout of not 1, but 2 different molecules, assembled primarily in neutrophils. To validate zymography analysis, we performed sandwich ELISAs, whereby only the NGAL–MMP-9 complex was detected. The zymography data were indeed validated and also confirmed by sandwich ELISA analysis on samples from an independent UC patient cohort. To investigate whether and how NGAL–MMP-9 complex competes with currently used serum biomarkers to assess mucosal healing, we compared NGAL–MMP-9 with CRP. Hepatocytes rapidly increase the production of CRP under the influence TNF-a, interleukin-1b, and interleukin-6.1 Importantly, such elevations are not specific to IBD as increased levels are also found in various other disease settings.28 Although CRP is now used as the standard inflammatory marker in the clinic, 52% of UC patients do not present elevated CRP levels at baseline.28 In our study, we confirmed this finding and found that 48% of active UC patients had normal CRP levels at baseline, whereas only 29% of UC responders and 33% of UC nonresponders had normal NGAL–MMP-9 levels at baseline. Good biomarkers should not only have the characteristic to indicate disease activity (as determined by degree of inflammation) but also should indicate response to treatment and subsequent mucosal healing, as these are important outcomes as well. Therefore, NGAL–MMP-9 complex levels in sera correlated with endoscopic activity, as measured by the Mayo score. We found a good correlation with a gradual increase in NGAL–MMP-9 complex levels with every point of the score. Furthermore, ROC analysis was performed for pooled data of both cohorts, identifying an NGAL–MMP-9 complex cutoff value of 97.7 ng/mL corresponding to complete mucosal healing with a specificity of 91%. In this respect, the cutoff of 7.3 mg/L for CRP was less sensitive (39% versus 43%) in predicting mucosal healing than NGAL–MMP-9 with a specificity of 91%. In addition, we also studied the predictive value of serum NGAL–MMP-9 complex levels for histological improvement. In a subset of patients, we were able to define histological activity before and after treatment with infliximab and found that NGAL–MMP-9 levels correlated with histological activity scores. Interestingly, histological improvement correlated nicely
1206
| www.ibdjournal.org
Inflamm Bowel Dis Volume 20, Number 7, July 2014
with mucosal healing, as defined by endoscopic evaluation. NGAL–MMP-9 levels lower than 65 ng/mL were found to be predictive of histological improvement. Prospective studies measuring NGAL–MMP-9 complex levels in patients undergoing anti-TNF therapy are underway to confirm our data and are needed to determine whether this novel serum marker could reduce the amount of endoscopies. It will also be interesting to investigate whether NGAL–MMP-9 complex levels decrease in patients who receive other types of treatment. This will be necessary to validate further that NGAL–MMP-9 is a reliable surrogate marker. As for the functionality of the NGAL– MMP-9 complex in relation to response to the treatment, it is difficult to state whether it plays an active role or is rather a consequence due to less circulating neutrophils as a result of control of inflammation. A study by the group of MacDonald29 suggested that MMPs could play an active role because these are able to degrade anti-TNF antibodies. In this particular study, it was found that MMP-3 and MMP-12 were able to degrade infliximab antibodies, thereby releasing Fc monomers and F(ab)2. These cleaved forms were still functional; however, a reduction was seen in their ability to neutralize soluble TNF-a after coincubation with IBD proteins. This study warrants further research into the potential cleavage of infliximab antibodies by MMP-9 or the NGAL– MMP-9 complex and its subsequent effect on response to treatment. Because the NGAL–MMP-9 complex is a neutrophil product and UC is mostly characterized by influx of neutrophils to the active site of inflammation,2 it may not be surprising that serum NGAL–MMP-9 complex levels correlated strongly with neutrophil counts. However, correlation analysis of neutrophil counts with Mayo endoscopic subscores indicated a weaker correlation (t ¼ 0.308) than that was found for NGAL–MMP-9 levels (t ¼ 0.370). A limitation of the current retrospective study is the lack of information on fecal calprotectin levels. Fecal calprotectin is a sensitive marker of intestinal inflammation and correlates well with the degree of endoscopic activity.30 A recent article by Kolho et al31 showed that fecal MMP-9 had an almost comparable performance to fecal calprotectin in UC patients, suggesting its applicability as a surrogate marker of inflammation. Moreover, the EMBARK study showed that combinations of fecal calprotectin and serum MMP-9 for UC and combinations of fecal calprotectin, serum MMP-9, and serum interleukin-22 in CD demonstrated the strongest association with imaging/endoscopy-defined inflammation.32 These recent studies highlight the current relevance of our study, and future studies should therefore investigate the performance (or accuracy) of serum and/or fecal NGAL–MMP-9 complex levels compared with fecal calprotectin. Nonetheless, the advantages of a serum marker over a fecal marker are evident and are related to patient and caregiver convenience. The present study forms a strong stimulus to obtain more functional information about NGAL–MMP-9 as a complex. So far, only 1 study has investigated its relevance in IBD. Manfredi et al11 showed that urinary NGAL–MMP-9 level was an independent predictor of pediatric IBD. In cancer research, however,
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 complex is already a well-known biomarker.33– Moreover, an important issue relates to the specificity of NGAL–MMP-9 complex as a biomarker for UC or CD. We trust that the present data will be a stimulus to study this biomarker in large CD patient cohorts. To conclude, we found that serum NGAL–MMP-9 complex levels were increased in patients with active UC compared with HC. NGAL–MMP-9 complex levels significantly decreased after treatment with infliximab and correlated well with mucosal healing. Importantly, NGAL–MMP-9 complex complemented CRP in predicting disease activity and mucosal healing. Therefore, we propagate that NGAL–MMP-9 complex can be used as a surrogate serum marker in UC patients, thereby potentially decreasing the need for endoscopy.
35
Acknowledgments The authors thank Jennifer Vandooren and Erik Martens for their help with zymography analysis and Vera Ballet for managing the VLECC (Vlaams erfelijkheidsonderzoek Crohn en colitis ulcerosa) database of IBD patients at the Leuven University Hospital.
REFERENCES 1. Vermeire S, Van Assche G, Rutgeerts P. Laboratory markers in IBD: useful, magic, or unnecessary toys? Gut. 2006;55:426–431. 2. Qin X. Etiology of inflammatory bowel disease: a unified hypothesis. World J Gastroenterol. 2012;18:1708–1722. 3. Borregaard N. Neutrophils, from marrow to microbes. Immunity. 2010;33: 657–670. 4. Kubben FJ, Sier CF, Hawinkels LJ, et al. Clinical evidence for a protective role of lipocalin-2 against MMP-9 autodegradation and the impact for gastric cancer. Eur J Cancer. 2007;43:1869–1876. 5. Tschesche H, Zolzer V, Triebel S, et al. The human neutrophil lipocalin supports the allosteric activation of matrix metalloproteinases. Eur J Biochem. 2001;268:1918–1928. 6. Kjeldsen L, Johnsen AH, Sengelov H, et al. Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J Biol Chem. 1993;268:10425–10432. 7. Kjeldsen L, Bainton DF, Sengelov H, et al. Identification of neutrophil gelatinase-associated lipocalin as a novel matrix protein of specific granules in human neutrophils. Blood. 1994;83:799–807. 8. Flo TH, Smith KD, Sato S, et al. Lipocalin 2 mediates an innate immune response to bacterial infection by sequestrating iron. Nature. 2004;432: 917–921. 9. Chan YR, Liu JS, Pociask DA, et al. Lipocalin 2 is required for pulmonary host defense against Klebsiella infection. J Immunol. 2009;182: 4947–4956. 10. Nielsen BS, Borregaard N, Bundgaard JR, et al. Induction of NGAL synthesis in epithelial cells of human colorectal neoplasia and inflammatory bowel diseases. Gut. 1996;38:414–420. 11. Manfredi MA, Zurakowski D, Rufo PA, et al. Increased incidence of urinary matrix metalloproteinases as predictors of disease in pediatric patients with inflammatory bowel disease. Inflamm Bowel Dis. 2008;14: 1091–1096. 12. Bolignano D, Della Torre A, Lacquaniti A, et al. Neutrophil gelatinaseassociated lipocalin levels in patients with Crohn disease undergoing treatment with infliximab. J Investig Med. 2010;58:569–571. 13. Oikonomou KA, Kapsoritakis AN, Theodoridou C, et al. Neutrophil gelatinase-associated lipocalin (NGAL) in inflammatory bowel disease: association with pathophysiology of inflammation, established markers, and disease activity. J Gastroenterol. 2012;47:519–530.
NGAL–MMP-9 as a Surrogate Serum Marker in UC
14. Yesil A, Gonen C, Senates E, et al. Relationship between neutrophil gelatinase-associated lipocalin (NGAL) levels and inflammatory bowel disease type and activity. Dig Dis Sci. 2013;58:2587–2593. 15. Nielsen OH, Gionchetti P, Ainsworth M, et al. Rectal dialysate and fecal concentrations of neutrophil gelatinase-associated lipocalin, interleukin-8, and tumor necrosis factor-alpha in ulcerative colitis. Am J Gastroenterol. 1999;94:2923–2928. 16. Arijs I, De Hertogh G, Lemaire K, et al. Mucosal gene expression of antimicrobial peptides in inflammatory bowel disease. PLoS One. 2009; 4:e7984. 17. Vandooren J, Van den Steen PE, Opdenakker G. Biochemistry and molecular biology of gelatinase B or matrix metalloproteinase-9 (MMP-9): the next decade. Crit Rev Biochem Mol Biol. 2013;48:222–272. 18. Baugh MD, Perry MJ, Hollander AP, et al. Matrix metalloproteinase levels are elevated in inflammatory bowel disease. Gastroenterology. 1999;117:814–822. 19. Opdenakker G, Van den Steen PE, Van Damme J. Gelatinase B: a tuner and amplifier of immune functions. Trends Immunol. 2001;22:571–579. 20. Opdenakker G, Van den Steen PE, Dubois B, et al. Gelatinase B functions as regulator and effector in leukocyte biology. J Leukoc Biol. 2001;69: 851–859. 21. Annahazi A, Molnar T, Farkas K, et al. Fecal MMP-9: a new noninvasive differential diagnostic and activity marker in ulcerative colitis. Inflamm Bowel Dis. 2013;19:316–320. 22. Sela-Passwell N, Kikkeri R, Dym O, et al. Antibodies targeting the catalytic zinc complex of activated matrix metalloproteinases show therapeutic potential. Nat Med. 2011;18:143–147. 23. Geboes K, Rutgeerts P, Opdenakker G, et al. Endoscopic and histologic evidence of persistent mucosal healing and correlation with clinical improvement following sustained infliximab treatment for Crohn’s disease. Curr Med Res Opin. 2005;21:1741–1754. 24. Gao Q, Meijer MJ, Schluter UG, et al. Infliximab treatment influences the serological expression of matrix metalloproteinase (MMP)-2 and -9 in Crohn’s disease. Inflamm Bowel Dis. 2007;13:693–702. 25. Vandooren J, Geurts N, Martens E, et al. Zymography methods for visualizing hydrolytic enzymes. Nat Methods. 2013;10:211–220. 26. Descamps FJ, Martens E, Opdenakker G. Analysis of gelatinases in complex biological fluids and tissue extracts. Lab Invest. 2002;82: 1607–1608. 27. Iskandar HN, Ciorba MA. Biomarkers in inflammatory bowel disease: current practices and recent advances. Transl Res. 2012;159:313–325. 28. Tall AR. C-reactive protein reassessed. N Engl J Med. 2004;350: 1450–1452. 29. Biancheri A, Di Sabatino A, Joe-Njoku I, et al. Differential Cleavage of Anti-tumor Necrosis Factor-alpha Agents by Matrix Metalloproteinase (MMP)-10 and MMP-12 in Inflammatory Bowel Disease. Dublin, Ireland: ECCO; 2011. 30. D’Haens G, Ferrante M, Vermeire S, et al. Fecal calprotectin is a surrogate marker for endoscopic lesions in inflammatory bowel disease. Inflamm Bowel Dis. 2012;18:2218–2224. 31. Kolho KL, Sipponen T, Valtonen E, et al. Fecal calprotectin, MMP-9, and human beta-defensin-2 levels in pediatric inflammatory bowel disease. Int J Colorectal Dis. 2014;29:43–50. 32. Faubion WA Jr, Fletcher JG, O’Byrne S, et al. EMerging BiomARKers in Inflammatory Bowel Disease (EMBARK) study identifies fecal calprotectin, serum MMP9, and serum IL-22 as a novel combination of biomarkers for Crohn’s disease activity: role of cross-sectional imaging. Am J Gastroenterol. 2013;108:1891–1900. 33. Yan L, Borregaard N, Kjeldsen L, et al. The high molecular weight urinary matrix metalloproteinase (MMP) activity is a complex of gelatinase B/MMP-9 and neutrophil gelatinase-associated lipocalin (NGAL). Modulation of MMP-9 activity by NGAL. J Biol Chem. 2001;276: 37258–37265. 34. Bolignano D, Donato V, Lacquaniti A, et al. Neutrophil gelatinaseassociated lipocalin (NGAL) in human neoplasias: a new protein enters the scene. Cancer Lett. 2010;288:10–16. 35. Mir SU, Jin L, Craven RJ. Neutrophil gelatinase-associated lipocalin (NGAL) expression is dependent on the tumor-associated sigma-2 receptor S2RPgrmc1. J Biol Chem. 2012;287:14494–14501.
www.ibdjournal.org |
1207
Neutrophil Gelatinase B–associated Lipocalin and Matrix Metalloproteinase-9 Complex as a Surrogate Serum Marker of Mucosal Healing in Ulcerative Colitis Magali de Bruyn, MSc,* ,† Ingrid Arijs, PhD,† Willem-Jan Wollants, Dipl Ing,† Kathleen Machiels, MSc,† Kristel Van Steen, PhD,‡ Gert Van Assche, MD, PhD,† Marc Ferrante, MD, PhD,† Paul Rutgeerts, MD, PhD,† Séverine Vermeire, MD, PhD,† and Ghislain Opdenakker, MD, PhD*
Background: The current standard for the assessment of mucosal healing after therapy in inflammatory bowel diseases is endoscopy. However, a high need exists for noninvasive, accurate surrogate markers.
Methods: In 2 independent cohorts, levels of serum neutrophil gelatinase B–associated lipocalin and matrix metalloproteinase-9 complex (NGAL– MMP-9) from patients with active ulcerative colitis (UC) before and after first treatment with infliximab and from healthy controls (HC) were determined with zymography and sandwich enzyme-linked immunosorbent assay. The response to infliximab was defined as complete mucosal healing (Mayo endoscopic subscore 0–1) at control endoscopy. Data were analyzed with SPSS, and P values ,0.05 were considered significant.
Results: In cohort 1 (n ¼ 66; median age, 30 yr; 38% female), serum NGAL–MMP-9 levels significantly increased at baseline in UC patients versus HC
(103.8 versus 42.4 ng/mL; P , 0.0001), whereas 55% of the patients had normal C-reactive protein levels. NGAL–MMP-9 levels significantly decreased after therapy in UC responders (from 116.3 ng/mL to 32.0 ng/mL; P , 0.0001) and in nonresponders (from 94.7 ng/mL to 54.1 ng/mL; P ¼ 0.047). In cohort 2 (n ¼ 132; median age, 39 yr; 53% female), NGAL–MMP-9 levels increased at baseline in active UC patients versus HC (86.5 versus 60.4 ng/ mL; P ¼ 0.10), whereas 45% of the patients had normal C-reactive protein levels. NGAL–MMP-9 levels significantly decreased after therapy in responders (from 87.5 ng/mL to 16.3 ng/mL; P , 0.0001) but not in nonresponders (from 82.7 ng/mL to 57.8 ng/mL; P ¼ 0.19). After pooling the data, a cutoff value of 97.7 ng/mL for NGAL–MMP-9 complex was determined to predict complete mucosal healing with high specificity (91%).
Conclusions: Serum NGAL–MMP-9 is suggested as a new surrogate marker for the assessment of mucosal healing in UC patients treated with infliximab. (Inflamm Bowel Dis 2014;20:1198–1207) Key Words: NGAL–MMP-9 complex, ulcerative colitis, mucosal healing, surrogate serum marker, neutrophil
O
ne of the challenges in inflammatory bowel disease (IBD) research is to have accurate biomarkers with predictive
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.ibdjournal.org). Received for publication November 5, 2013; Accepted April 3, 2014. From the *Department of Microbiology and Immunology, Rega Institute for Medical Research, KU Leuven, Leuven, Belgium; †Department of Clinical and Experimental Medicine, Translational Research Center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium; and ‡Montefiore Institute, Electrical Engineering and Computer Science, University of Liège, Liège, Belgium. I. Arijs and M. Ferrante are postdoctoral fellows and S. Vermeire is a Senior Clinical Investigator of the Fund for Scientific Research of Flanders (FWOVlaanderen). M. de Bruyn is a PhD student funded by a fellowship from the Agency for Innovation by Science and Technology (IWT). P. Rutgeerts, S. Vermeire, G. Van Assche and M. Ferrante report following conflicts of interest: grant support, lecture fees and consulting fees from Centocor and Schering-Plough. Supported by FWO-Vlaanderen and “Geconcerteerde Onderzoeksacties GOA 2013-014”. Reprints: Ghislain Opdenakker, MD, PhD, Department of Microbiology and Immunology, Rega Institute for Medical Research, Minderbroedersstraat 10, blok x-bus 1030, Leuven 3000, Belgium (e-mail: [email protected]). Copyright © 2014 Crohn’s & Colitis Foundation of America, Inc. DOI 10.1097/MIB.0000000000000068 Published online 27 May 2014.
1198
| www.ibdjournal.org
value, correlating with disease activity, responsive to the effect of treatment, and preferentially playing a defined pathophysiological role in the disease. At present, mainly C-reactive protein (CRP) and erythrocyte sedimentation rate are used as blood markers for IBD, whereas calprotectin and lactoferrin are useful fecal markers. Nonetheless, no single biomarker has been proven to be ideal for all proposed indications.1 Ulcerative colitis (UC) is mostly characterized by influx of neutrophils to the active site of inflammation.2 Being specialized leukocytes, neutrophils have an end-stage differentiation and contain preformed granules.3 Secondary and tertiary neutrophil granules contain, respectively, neutrophil gelatinase B–associated lipocalin ([NGAL], also called lipocalin-2) and matrix metalloproteinase-9 ([MMP-9], also known as gelatinase B). Furthermore, MMP-9 can form a covalent complex with NGAL. However, the functionality of this complex is still debated. One hypothesis is that by formation of this complex, NGAL protects MMP-9 from autodegradation,4 whereas it was also proposed that NGAL supports the allosteric activation of pro-MMP-9.5 In contrast, the functions of both individual proteins are well studied. Inflamm Bowel Dis Volume 20, Number 7, July 2014
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL6,7 is expressed in response to the activation of Tolllike receptors during infections,8,9 and it has been shown to inhibit bacterial growth by sequestering iron-laden siderophores.10 NGAL protein or messenger RNA expression levels have been correlated with parameters of active IBD.11–16 MMP-9, as a member of the MMP family, is a zinc-dependent endopeptidase involved in many developmental processes, including angiogenesis, wound healing, and extracellular matrix degradation. Despite the involvement in many normal physiological processes, MMP-9 has been associated with abnormal disease conditions17,18 and is considered a tuner and amplifier of inflammatory reactions.17,19,20 MMP-9 levels have been shown to be elevated in the feces of UC patients and to correlate well with disease activity.21 Recently Sela-Passwell et al22 have shown that neutralizing antibodies with tissue inhibitor of MMPslike mechanisms against MMP-2 and MMP-9 can attenuate the development of colitis in IBD mouse models. Studies investigating the decrease of NGAL or MMP-9 after treatment with infliximab were mostly performed in Crohn’s disease (CD) patients and not in UC patients.23,24 Moreover, the role of NGAL–MMP-9 as a complex has only been investigated in 1 study, indicating elevated levels of NGAL–MMP-9 in the urine of pediatric IBD patients.11 Because frequent endoscopic assessments are costly and uncomfortable for the patient, a high need exists for noninvasive, accurate markers. In contrast with CD patients, in whom CRP is an accurate marker, CRP is normal in more than 50% of UC patients. With this study, we aimed to investigate for the first time levels of NGAL–MMP-9 complex in the serum of patients with active UC before and after first treatment with infliximab and to compare the accuracy of NGAL–MMP-9 complex with CRP in the assessment of mucosal healing.
MATERIALS AND METHODS Patient Sampling In a retrospective study, performed at the Leuven University Hospital, sera from 2 independent cohorts of active UC patients before and after first administration of infliximab (Remicade; Centocor, Horsham, PA) and from healthy controls (HC) were investigated. Cohort 1 included 66 infliximab-naive patients with active UC and 40 HC (median age, 30.4 yr; 38% female). In cohort 2, 132 active UC patients who received a first dose of infliximab and 129 HC (median age, 38.8 yr; 53% female) were included. The baseline UC patient characteristics of the 2 cohorts are shown in Tables 1 and 2. The patients underwent endoscopy within a week before the first intravenous infliximab infusion of 5 mg/kg body weight, and a second endoscopy was performed after 8 to 14 weeks. The response to treatment with infliximab was determined at the time of the second endoscopy and defined as complete mucosal healing when the Mayo endoscopic subscore was 1 or 0. Patients were defined as nonresponders when they did not reach this complete healing, although some of them showed endoscopic improvement. Histological activity was determined by a panel of expert pathologists on biopsies
NGAL–MMP-9 as a Surrogate Serum Marker in UC
taken at the time of endoscopy before and after treatment with infliximab. Histological activity was scored according to the pathologic assessment reports as severely active or active inflammation (¼2), mildly active inflammation (¼1), and no inflammation (¼0). Histological improvement was defined as no inflammation (score 0) after the treatment. From all UC patients, CRP levels and neutrophil counts were measured in the serum before and after infliximab therapy, at time points corresponding to the time of endoscopy. From all individuals, a written informed consent was obtained, and the study was approved by the University Hospital Ethics Committee (Vlaams erfelijkheidsonderzoek Crohn en colitis ulcerosa S-53684).
Gelatin Zymography Analysis To define alterations in the different molecular forms of gelatinases, sera from UC patients and HC from cohort 1 were first analyzed with gelatin zymography analysis.25 Briefly, 100 mL of serum was prepurified using gelatin Sepharose beads (GE Healthcare, Buckinghamshire, United Kingdom) and mini spin columns (Bio-Rad Laboratories, Hercules, CA). The bound gelatinases were eluted from the column with 20 mL Tris/glycine/ sodium dodecyl sulfate nonreducing loading buffer (Invitrogen, Carlsbad, CA), and the prepurified samples were then separated in 7.5% acrylamide gels copolymerized with 1 mg/mL porcine gelatin (Sigma-Aldrich, St Louis, MO).26 The gels were then washed with 2.5% Triton X-100 (Sigma-Aldrich) for 40 minutes and incubated overnight at 378C in 50 mM Tris-HCl (pH 7.5) supplemented with 10 mM CaCl2 (Sigma-Aldrich). The gels were stained with 0.25% Coomassie Brilliant Blue-R (Sigma-Aldrich) and scanned using standard settings. Band densities of all individual gelatinase forms (MMP-2, MMP-9 multimers, NGAL– MMP-9 complexes, and MMP-9 monomers) were analyzed and quantified with Image J 1.48 software (NIH Windows version), and the obtained densities of different gelatinase forms were plotted as a ratio versus pro-MMP-2 levels.
Sandwich Enzyme-linked Immunosorbent Assay All serum NGAL–MMP-9 complex levels were determined with the use of a commercial anti-human NGAL–MMP-9 complex enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Abingdon, UK) and assayed according to the manufacturer’s guidelines. This ELISA incorporates 2 antibodies with different antigen specificities: one against MMP-9 and the other against NGAL. Hence, with this ELISA only the NGAL–MMP-9 complexes are measured. The absorbance was measured at 450 nm with a spectrophotometer. NGAL–MMP-9 complex levels were quantified with the use of a calibration curve of human NGAL–MMP-9 complex as a standard.
Statistical Analysis Data were analyzed with SPSS Statistics 20.0 software (SPSS, Inc, Chicago, IL) with the use of the nonparametric Mann–Whitney U test for unpaired samples (significance indicated by *) and Wilcoxon’s signed rank test for paired samples www.ibdjournal.org |
1199
Inflamm Bowel Dis Volume 20, Number 7, July 2014
de Bruyn et al
TABLE 1. Cohort 1 Patient Characteristics and Laboratory Markers at the Start of First Treatment with Infliximab for UC Responders and Nonresponders Baseline Patient Characteristics of Cohort 1
UC Responders (n ¼ 28)
UC Nonresponders (n ¼ 38)
Pa
15/13 (54/46) 38.7 (30.4–51.1) 6.1 (1.3–16.7)
25/13 (66/34) 42.4 (30.6–58.2) 6.9 (2.0–12.1)
0.44b 0.44c 0.94c
NA 10/28 (36) 18/28 (64)
NA 14/38 (37) 24/38 (63)
0/21 8/21 13/21 16/28 3/28 4.9 16/28 5.2
1/32 8/32 23/32 23/38 5/38 4.9 20/38 5.5
Male/female (%) Median (IQR) age at first infliximab (yr) Median (IQR) duration of disease before first infliximab (yr) Disease extent before starting infliximab therapy (%) E1 (proctitis) E2 (left-sided colitis) E3 (pancolitis) Histological activity score at first infliximab (%) 0 (inactive) 1 (mildly active) 2 (active) Mayo endoscopic subscore 3 before first infliximab (%) Active smoking at first infliximab (%) Median (IQR) CRP at first infliximab (mg/L) Normal CRP (,5 mg/L) levels at baseline (%) Median (IQR) amount of neutrophils (109/L) Concomitant medication at first infliximab (%) 5-ASA Intravenous/systemic oral corticosteroids Immunosuppressant (AZA or 6-MP)
(0) (38) (61) (57) (11) (1.0–14.2) (57) (3.4–7.3)
20/28 (71) 10/28 (36) 16/28 (57)
(3) (25) (72) (61) (13) (1.9–11.8) (53) (3.9–8.2)
35/38 (92) 12/38 (32) 12/38 (32)
NA 1b 1b 1b 0.37b 0.55b 0.81b 1b 0.58c 0.80b 0.61c 0.04a,b 0.79b 0.05a,b
P , 0.05 was considered significant. The statistical difference was analyzed by Fisher’s exact test. c The statistical difference was analyzed by Mann–Whitney U test. IQR, interquartile range; 5-ASA, 5-aminosalicylates; AZA, azathioprine; 6-MP, mercaptopurine. a
b
(significance indicated by ^). Spearman correlation analysis (correlation coefficient ¼ r), Kendall’s tau rank correlation (correlation coefficient ¼ t), Fisher’s exact test, and receiver operating characteristic (ROC) analysis were also performed in SPSS. Influences of demographic and clinical variables on NGAL–MMP-9 levels were first tested using univariate (Mann–Whitney U test) and bivariate analyses (Spearman’s rho or Kendall’s tau). Second, independent associations and influences of confounding variables were assessed using a linear regression model applying a forward selection approach. Multicolinearity was explored by determining the variance inflation factor (VIF) and tolerance value. A VIF value greater than 5 and a tolerance value smaller than 0.20 implied colinearity. P values of ,0.05 were considered significant.
RESULTS Serum NGAL–MMP-9 Complex Levels Are Increased at Baseline in Active UC Patients and Decrease After Treatment With Infliximab mers,
1200
To visualize all gelatinase forms (MMP-2, MMP-9 monoMMP-9 multimers, and NGAL–MMP-9 complex), | www.ibdjournal.org
zymography analysis was performed on prepurified serum samples from 66 UC patients (28 responders and 38 nonresponders) before and after therapy with infliximab and from 40 HC (cohort 1) (Fig., Supplemental Digital Content 1, http://links.lww.com/IBD/A481). A template of zymography analysis is shown in Figure 1, whereby different gelatinase forms can be distinguished. In fact, this test measures the levels of 4 analytes at once, and with appropriate standardization, it yields a richer information content than any single ELISA.25 At baseline, NGAL–MMP-9 serum levels from patients of cohort 1 with active UC were significantly elevated in comparison with HC levels (P , 0.01). After treatment with infliximab, NGAL–MMP-9 complex levels significantly decreased in both UC responders (P , 0.0001) and nonresponders (P ¼ 0.011), although the decrease was more profound for UC responders (Fig. 2A). Pro-MMP-9 multimer and monomer levels were also significantly decreased after treatment with infliximab, whereas activated forms of MMP-9 and MMP-2 were rarely detectable (Table 3). Zymography is a useful technique to visualize various molecular forms and to measure relative gelatinase levels; however, it is a laborious technique. Therefore, we validated our analyses of the NGAL–MMP-9 complex with sandwich ELISA
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 as a Surrogate Serum Marker in UC
TABLE 2. Cohort 2 Patient Characteristics and Laboratory Markers at the Start of First Treatment with Infliximab for UC Responders and Nonresponders Baseline Patient Characteristics of Cohort 2
UC Responders (n ¼ 88)
UC Nonresponders (n ¼ 44)
48/40 (55/45) 39.7 (28.4–50.7) 5.5 (2.3–10.6)
28/16 (64/36) 43.4 (32.9–53.0) 7.0 (1.4–12.6)
0.35b 0.39c 0.82c
2/88 (2) 31/88 (35) 55/88 (65)
NA 15/44 (34) 29/44 (66)
NA 1b 0.85b
2/72 44/72 26/72 44/88 9/88 6.0 39/88 5.2
0/34 26/34 8/34 24/44 5/44 6.4 17/44 5.7
1b 0.52b 0.40b 0.71b 1b 0.12c 0.58b 0.43c
Male/female (%) Median (IQR) age at first infliximab (yr) Median (IQR) duration of disease before first infliximab (yr) Disease extent before starting infliximab therapy (%) E1 (proctitis) E2 (left-sided colitis) E3 (pancolitis) Histological score at the start of infliximab therapy (%) 0 (inactive) 1 (mildly active) 2 (active) Mayo endoscopic subscore 3 before first infliximab (%) Active smoking at first infliximab (%) Median (IQR) CRP at first infliximab (mg/L) Normal CRP (,5 mg/L) levels at baseline (%) Median (IQR) amount of neutrophils (109/L) Concomitant medication at first infliximab (%) 5-ASA Intravenous/systemic oral corticosteroids Immunosuppressant (AZA or 6-MP)
(3) (61) (36) (50) (10) (1.6–16.6) (44) (3.9–7.3)
59/88 (67) 33/88 (38) 50/88 (57)
(0) (77) (23) (55) (11) (3.7–27.9) (39) (4.0–8.1)
Pa
0.01a,b 0.71b 0.002a,b
39/44 (89) 18/44 (41) 12/44 (27)
P , 0.05 was considered significant. The statistical difference was analyzed by Fisher’s exact test. c The statistical difference was analyzed by Mann–Whitney U test. IQR, interquartile range; 5-ASA, 5-aminosalicylates; AZA, azathioprine; 6-MP, mercaptopurine. a
b
(Fig. 2B) and thus confirmed that serum NGAL–MMP-9 complex levels were elevated at baseline in active UC patients as compared with HC (P , 0.0001). After treatment with infliximab, NGAL– MMP-9 complex serum levels were significantly decreased in UC responders (median range from 116.3 to 32.0 ng/mL; P , 0.0001), whereby only 3 responders had unchanged levels of NGAL–MMP-9 after therapy. Overall, serum NGAL–MMP-9 complex levels decreased in 90% of the UC responders. For UC nonresponders, only a borderline significant decrease was observed (median range from 94.7 to 54.1 ng/mL; P ¼ 0.0469), which was significantly less pronounced in comparison with UC responders (P ¼ 0.0074). For almost 40% of the UC nonresponders, NGAL–MMP-9 complex serum levels increased instead of decrease after treatment. Importantly, the NGAL–MMP-9 complex serum levels of UC responders decreased to a significantly lower level after treatment than the levels of UC nonresponders (P ¼ 0.03). Finally, a strong correlation was found between zymography analysis of the NGAL–MMP-9 complex and sandwich ELISA (r ¼ 0.835; P , 0.0001) (Fig. 3). To corroborate the data from UC patient in cohort 1, we retrospectively selected 132 UC patients (88 responders and 44 nonresponders) for an independent cohort 2 and performed
sandwich ELISA on serum samples collected before and after treatment with infliximab. As shown in Figure 2C, an increase in NGAL–MMP-9 complex levels was found at baseline for active UC patients versus HC (86.5 versus 60.4 ng/mL; P ¼ 0.10). After treatment, a significant decrease of NGAL–MMP-9 levels was found for UC responders (median range from 87.5 to 16.3 ng/mL; P , 0.0001). More specifically, 83% of the UC responders had decreased NGAL–MMP-9 levels after treatment. However, NGAL–MMP-9 levels did not significantly decrease after treatment for UC nonresponders (median range from 82.7 to 57.7 ng/mL; P ¼ 0.19), whereby 41% of the UC nonresponders had increased or similar NGAL–MMP-9 levels after treatment. Moreover, NGAL–MMP-9 levels of UC responders decreased to a significantly lower level after treatment than the levels of UC nonresponders after treatment (P , 0.0001). Because cohort 2 confirmed and validated the results from cohort 1, the data from both cohorts were pooled. As shown in Figure 2D, pooled data from 129 HC, 116 UC responders, and 82 UC nonresponders confirm that serum NGAL–MMP-9 complex levels significantly increased at baseline in active UC patients versus HC (87.3 ng/mL versus 60.4 ng/mL; P ¼ 0.012). Moreover, NGAL–MMP-9 levels significantly decreased after treatment www.ibdjournal.org |
1201
de Bruyn et al
Inflamm Bowel Dis Volume 20, Number 7, July 2014
patients who had active disease at baseline, which was associated with higher levels of NGAL–MMP-9, were treated effectively with corticosteroids. Importantly, these confounding characteristics were not different between patients who responded to infliximab and those who did not respond to infliximab (Tables 1 and 2).
Serum NGAL–MMP-9 Complex Levels Correlate With Neutrophil Counts and Complement CRP as an Inflammatory Marker FIGURE 1. Template of zymography analysis showing different gelatinase forms in prepurified samples. (Pro) MMP-9 multimers are detected at the highest molecular weights (in kilodaltons), followed by NGAL–MMP-9 complex, pro-MMP-9 monomers, MMP-9 monomers, pro-MMP-2, and MMP-2. This template illustrates gelatinase levels of a HC and a UC responder and nonresponder patient from cohort 1. Serum samples from UC patients before and after infliximab therapy were loaded in adjacent lanes. R1, serum sample of a UC responder patient before infliximab therapy; R2, serum sample of the same UC responder patient after infliximab therapy; NR1, serum sample of a UC nonresponder before infliximab therapy; NR2, serum sample of the same UC nonresponder after infliximab therapy.
with infliximab for UC responders (median range from 90.8 to 20.7 ng/mL; P , 0.0001) and to a lesser extent for UC nonresponders (median range from 86.1 to 56.2 ng/mL; P ¼ 0.02). NGAL–MMP-9 levels decreased to significantly lower levels after treatment for UC responders compared with UC nonresponders (P , 0.0001). Furthermore, the influence of several demographic and clinical variables on serum NGAL–MMP-9 levels was determined. We found that the concomitant use of corticosteroids (intravenous or oral systemic) and disease activity (as was assessed by the Mayo endoscopic subscore) showed a correlation with serum NGAL–MMP-9 levels at the start of infliximab therapy (t ¼ 0.215; P , 0.0001 and t ¼ 0.255; P , 0.0001, respectively). No significant differences were found for disease extent (P ¼ 0.281), histological activity (P ¼ 0.601), concomitant treatment with 5-aminosalicylates (P ¼ 0.295), or immunosuppressants (P ¼ 0.077). Univariate linear regression analysis confirmed our findings for association of the use of corticosteroids (P ¼ 0.001) and disease activity (P ¼ 0.003) with NGAL–MMP-9 levels. Furthermore, also gender (female) (P ¼ 0.06) and concomitant treatment with immunosuppressants (P ¼ 0.095) were moderately associated with NGAL–MMP-9 levels at baseline. In the multiple linear regression model, we included gender, disease activity at the start of infliximab therapy, and concomitant treatment with corticosteroids or immunosuppressants. The analysis indicated that the concomitant use of corticosteroids (b ¼ 0.201; P ¼ 0.004) and disease activity at baseline (b ¼ 0.177; P ¼ 0.012) had an influence on serum NGAL–MMP-9 levels at baseline. VIF and tolerance showed that there was no colinearity between both parameters (VIF ¼ 1.03 and tolerance ¼ 0.96). These data underline the fact that
1202
| www.ibdjournal.org
To assess its role as a serum marker of inflammation and mucosal healing, we correlated NGAL–MMP-9 complex serum levels with currently used blood markers in IBD. Correlations were made with neutrophil counts and CRP levels. In cohort 1, the results indicated that NGAL–MMP-9 complex levels correlated well (r ¼ 0.430; P , 0.0001) with neutrophil counts, in line with the fact that NGAL–MMP-9 is a product of neutrophils. A more moderate correlation was found with CRP levels (r ¼ 0.266; P ¼ 0.002). More than half (57%) of the UC responders had no elevated CRP levels (,5 mg/L) at baseline; of whom, only 1 patient also had normal (,42.2 ng/mL) NGAL–MMP-9 complex levels at the start of anti–tumor necrosis factor (TNF) treatment. In UC nonresponders, 53% had normal CRP levels at baseline; of whom, only 9 patients (24%) also had normal levels of NGALMMP-9. These results indicated that, especially for UC responders, serum NGAL–MMP-9 complex is more representative of the disease activity than CRP. In cohort 2, NGAL–MMP-9 levels significantly correlated with the amount of neutrophils (r ¼ 0.463; P , 0.01) and CRP levels (r ¼ 0.477; P , 0.01). In 45% of the UC responder and nonresponder patients, CRP levels were not elevated at baseline (,5 mg/L). Half of the patients with normal CRP levels also had NGAL–MMP-9 levels below the median HC level (60.4 ng/mL). After pooling the data, correlation factors of NGAL–MMP-9 with CRP (r ¼ 0.429; P , 0.01) and neutrophil counts (r ¼ 0.414; P , 0.01) were determined (Fig. A and B, Supplemental Digital Content 2, http://links.lww.com/ IBD/A482). Overall, in 48% of the active UC patients, CRP was not elevated (,5 mg/L) at baseline; of whom, 36% of the patients also had normal (,60.4 ng/mL) NGAL–MMP-9 levels.
Correlation Analysis With Mayo Endoscopic Subscores Indicates That Serum NGAL–MMP-9 Complex Levels Constitute a Surrogate Marker of Mucosal Healing For patients from cohort 1, serum NGAL–MMP-9 complex levels significantly correlated with Mayo endoscopic subscores (t ¼ 0.317; P , 0.01), and the highest NGAL–MMP-9 complex levels were observed in patients with Mayo endoscopic subscores of 2 and 3. After therapy with infliximab, NGAL–MMP-9 levels decreased rapidly and correlated with restoration of mucosal integrity. Interestingly, the mean NGAL–MMP-9 complex levels did not differ between Mayo endoscopic subscores 0 and 1. When pointwise Mayo endoscopic subscore gradation was related to
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 as a Surrogate Serum Marker in UC
FIGURE 2. Decrease of serum NGAL–MMP-9 complex levels after therapy with infliximab as measured by zymography analysis (A), whereby NGAL–MMP-9 levels are represented as a ratio versus pro-MMP-2 levels, and sandwich enzyme-linked immunosorbent assay analysis (B–D). A, B, Serum NGAL–MMP-9 levels from patients of cohort 1. C, Serum NGAL–MMP-9 levels from patients of cohort 2. D, The pooled data from cohorts 1 and 2. Asterisks (*) indicate significance by nonparametric Mann-Whitney U test for unpaired samples. Arrowheads (^) indicate significance by Wilcoxon’s signed rank test for paired samples. ^P , 0.05 or *P , 0.05; ^^P , 0.01; ** P , 0.01; ^^^P , 0.001; ***P , 0.001.
CRP levels, a more moderate correlation was obtained (t ¼ 0.299; P , 0.01). Whereas we observed a gradual increase in NGAL– MMP-9 complex levels with increasing Mayo endoscopic subscores, CRP was only markedly increased in patients who had a Mayo endoscopic subscore of 3. A similar analysis was performed identifying the correlation of neutrophil counts with Mayo endoscopic subscores (t ¼ 0.233; P , 0.01). Finally, ROC analysis was performed whereby NGAL– MMP-9 complex values corresponding to Mayo endoscopic subscores 0 or 1 were compared with NGAL–MMP-9 complex levels corresponding to Mayo endoscopic subscores 2 or 3. The
analysis showed that NGAL–MMP-9 complex levels lower than 97.7 ng/mL were predictive of mucosal healing with a specificity and sensitivity of 93% and 43%, respectively. The area under the curve (AUC) was 0.75, and a positive predictive value (PPV) of 96% and a negative predictive value (NPV) of 31% were determined. ROC analysis with CRP levels corresponding to Mayo endoscopic subscores 0 to 1 and 2 to 3, respectively, was also performed. The obtained ROC curve was comparable to the curve plotted for NGAL–MMP-9 serum levels (AUC difference of 0.70 versus 0.75; P ¼ 0.43). However, the clinical cutoff value of 5 mg/L for CRP corresponded to www.ibdjournal.org |
1203
Inflamm Bowel Dis Volume 20, Number 7, July 2014
de Bruyn et al
TABLE 3. Cohort 1 Serum Levels of Gelatinase Forms as Determined by Zymography, for HC and UC Responders or Nonresponders, Before and After Treatment with Infliximab Ratio of Gelatinase Form Versus Pro-MMP-2
HC (n ¼ 40), Median (IQR)
Pro-MMP-9 multimers 0.53 (0.38–0.83) NGAL–MMP-9 complex 0.34 (0.15–0.68) Pro-MMP-9 monomers 1.67 (1.24–2.63) MMP-9 monomers ND MMP-2 ND
UC Nonresponders (n ¼ 38), Median (IQR)
UC Responders (n ¼ 28), Median (IQR) a
Before Infliximab
After Infliximab
P
Before Infliximab
After Infliximab
Pa
0.80 (0.58–1.12) 0.77 (0.50–1.36) 2.08 (1.54–2.77) ND ND
0.52 (0.38–0.72) 0.25 (0.14–0.52) 1.39 (1.12–1.90) ND ND
0.0004b ,0.0001b ,0.0001b — —
0.77 (0.49–1.27) 0.59 (0.42–1.19) 2.09 (1.44–3.24) ND ND
0.66 (0.46–0.93) 0.41 (0.22–0.73) 1.73 (1.30–2.40) ND ND
0.023b 0.017b 0.013b — —
P , 0.05 was considered significant. The statistical difference was analyzed by Wilcoxon’s signed rank test. ND, not detectable; IQR, interquartile range.
a
b
a lower specificity (82%) for the prediction of mucosal healing than NGAL–MMP-9 (93%). In cohort 2, previous correlations were validated because serum NGAL–MMP-9 complex levels (t ¼ 0.382; P , 0.01) and CRP levels (t ¼ 0.408; P , 0.01) significantly correlated with Mayo endoscopic subscores. A similar analysis was performed
FIGURE 3. Correlation plot of serum NGAL–MMP-9 complex levels as measured from patients of cohort 1 by sandwich ELISA and zymography analysis. NGAL–MMP-9 serum levels of UC responders and nonresponders before and after therapy with infliximab are shown in red and blue, respectively. HC samples are indicated as black diamonds, nonresponders as circles, and responders as triangles. R1, serum sample of a UC responder patient before therapy; R2, serum sample of a UC responder patient after therapy; NR1, serum sample of a UC nonresponder before therapy; NR2, serum sample of a UC nonresponder after therapy.
1204
| www.ibdjournal.org
identifying the correlation of neutrophil counts with Mayo endoscopic subscores (t ¼ 0.368; P , 0.01). Finally, ROC analysis was performed whereby NGAL–MMP-9 complex levels lower than 93.2 ng/mL were predictive of mucosal healing with a specificity and sensitivity of 91% and 44%, respectively. The AUC was 0.78, and a PPV of 41% and NPV of 70% were determined. ROC analysis for CRP levels corresponding to Mayo endoscopic subscores 0 to 1 and 2 to 3, determined a cutoff value of 7.2 mg/L, predicting mucosal healing with 44% sensitivity, 93% specificity, 46% PPV, and 72% NPV, with an AUC of 0.80. However, the clinical cutoff value of 5 mg/L for CRP corresponded to a lower specificity (86%) for the prediction of mucosal healing than NGAL–MMP-9 (91%). Pooling the data from cohorts 1 and 2, serum NGAL– MMP-9 complex levels (t ¼ 0.370; P , 0.01) and CRP levels (t ¼ 0.372; P , 0.01) significantly correlated with Mayo endoscopic subscores (Fig. 4A, B). A similar analysis was performed identifying the correlation of neutrophil counts with Mayo endoscopic subscores (t ¼ 0.308; P , 0.01) (data not shown). ROC analysis of the pooled data determined an AUC of 0.78 and a cutoff value of 97.7 ng/mL for NGAL–MMP-9 complex, predicting mucosal healing with 43% sensitivity, 91% specificity, 92% PPV, and 40% NPV (Fig. 4C). For CRP, a similar AUC of 0.77 was determined, and a cutoff value of 7.3 mg/L was determined, predicting mucosal healing with 39% sensitivity, 91% specificity, 91% PPV, and 38% NPV (Fig. 4D). Finally, when considering the clinical cutoff value of 5 mg/L for CRP, a lower specificity (85%) was found for the prediction of mucosal healing than for NGAL–MMP-9 (91%).
Correlation Analysis With Histological Activity Scores Indicates That Serum NGAL–MMP-9 Complex Levels Constitute a Surrogate Marker of Histological Improvement In a subset of patients (n ¼ 103), information on histological activity before and after treatment with infliximab was also
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 as a Surrogate Serum Marker in UC
FIGURE 4. A, B, Correlations of NGAL–MMP-9 complex levels and CRP levels with Mayo endoscopic subscores. A, Correlation of NGAL–MMP-9 complex values from all UC patients (pooled cohorts 1 and 2) with the corresponding Mayo endoscopic subscores, regardless of their response to therapy or time point before or after the treatment. The dotted line represents the mean NGAL–MMP-9 complex level of HC subjects (60.4 ng/mL). B, Correlation of CRP levels from all UC patients with Mayo endoscopic subscores. The dotted line represents the clinical cutoff value of 5 mg/L. Mean values along with the standard deviation of the mean are represented in function of the corresponding Mayo endoscopic subscores. C, D, ROC analysis of serum NGAL–MMP-9 complex values and CRP levels corresponding to Mayo endoscopic subscores 0 to 1 versus Mayo endoscopic subscores 2 to 3. C, A cutoff value of 97.7 ng/mL for NGAL–MMP-9 was determined to predict mucosal healing with 43% sensitivity, 91% specificity, 92% PPV, and 40% NPV. D, For CRP levels, a cutoff value of 7.3 mg/L was determined to predict mucosal healing with 39% sensitivity, 91% specificity, 91% PPV, and 38% NPV (D).
analyzed. Histological activity scores (severely active or active ¼ 2, mildly active ¼ 1, inactive ¼ 0) were determined by a panel of expert pathologists on biopsies taken at the time of endoscopy at the start of infliximab therapy and after treatment. Correlation analysis of histological activity with NGAL–MMP-9 levels showed a significant correlation (t ¼ 0.218; P , 0.0001), and NGAL–MMP-9 levels gradually increased for every score point (linear trend test, P , 0.0001) (Fig. A, Supplemental Digital Content 3, http://links.lww.com/IBD/A483). Moreover, a significant correlation was found between histological improvement and mucosal healing (t ¼ 0.465; P , 0.0001) and between the endoscopic Mayo score and histological score (t ¼ 0.505; P , 0.0001). Furthermore, ROC analysis was performed to evaluate the performance of NGAL–MMP-9 to discriminate between histological activity (active and mildly active) (n ¼ 182) and no activity (n ¼ 24). The AUC (95% confidence interval) for
NGAL–MMP-9 levels was 0.713 (0.627–0.799). A cutoff value of 65 ng/mL for NGAL–MMP-9 was able to predict histological healing with 52% sensitivity and 91% specificity (Fig. B, Supplemental Digital Content 3, http://links.lww.com/IBD/A483).
DISCUSSION Biomarkers are crucial to aid clinicians in IBD diagnosis and management. Currently available blood and fecal biomarkers provide reproducible and quantitative tools to distinguish IBD from noninflammatory diarrhea and assess disease activity.27 However, the discovery of novel biomarkers that correlate with mucosal healing or predict long-term disease course would improve current clinical assessment greatly. With the aim to critically assess serum gelatinase levels in UC patients, we first used the labor-intensive gelatin zymography www.ibdjournal.org |
1205
de Bruyn et al
analysis25 to measure multiple analytes simultaneously (MMP-2, MMP-9 monomers, MMP-9 multimers, and NGAL–MMP-9 complex). Hence, the information content is much richer than that of a single parameter. Zymography may also reveal—individually— the activation forms and degradation products of the studied enzymes, whereas with ELISA, all these molecules are measured as a mixture, without information about single molecular species. Zymography analysis, performed on serum samples before and after treatment with infliximab from cohort 1, indicated that the NGAL–MMP-9 complex was intact and that the levels were significantly higher in active UC patients at baseline compared with HC. This indicated that NGAL–MMP-9 is an inflammatory marker related to disease activity. Moreover, the covalently formed complex of MMP-9 with NGAL significantly decreased after treatment with infliximab in UC responders and to a lesser extent in nonresponders. It needs to be stressed that the NGAL–MMP-9 is a readout of not 1, but 2 different molecules, assembled primarily in neutrophils. To validate zymography analysis, we performed sandwich ELISAs, whereby only the NGAL–MMP-9 complex was detected. The zymography data were indeed validated and also confirmed by sandwich ELISA analysis on samples from an independent UC patient cohort. To investigate whether and how NGAL–MMP-9 complex competes with currently used serum biomarkers to assess mucosal healing, we compared NGAL–MMP-9 with CRP. Hepatocytes rapidly increase the production of CRP under the influence TNF-a, interleukin-1b, and interleukin-6.1 Importantly, such elevations are not specific to IBD as increased levels are also found in various other disease settings.28 Although CRP is now used as the standard inflammatory marker in the clinic, 52% of UC patients do not present elevated CRP levels at baseline.28 In our study, we confirmed this finding and found that 48% of active UC patients had normal CRP levels at baseline, whereas only 29% of UC responders and 33% of UC nonresponders had normal NGAL–MMP-9 levels at baseline. Good biomarkers should not only have the characteristic to indicate disease activity (as determined by degree of inflammation) but also should indicate response to treatment and subsequent mucosal healing, as these are important outcomes as well. Therefore, NGAL–MMP-9 complex levels in sera correlated with endoscopic activity, as measured by the Mayo score. We found a good correlation with a gradual increase in NGAL–MMP-9 complex levels with every point of the score. Furthermore, ROC analysis was performed for pooled data of both cohorts, identifying an NGAL–MMP-9 complex cutoff value of 97.7 ng/mL corresponding to complete mucosal healing with a specificity of 91%. In this respect, the cutoff of 7.3 mg/L for CRP was less sensitive (39% versus 43%) in predicting mucosal healing than NGAL–MMP-9 with a specificity of 91%. In addition, we also studied the predictive value of serum NGAL–MMP-9 complex levels for histological improvement. In a subset of patients, we were able to define histological activity before and after treatment with infliximab and found that NGAL–MMP-9 levels correlated with histological activity scores. Interestingly, histological improvement correlated nicely
1206
| www.ibdjournal.org
Inflamm Bowel Dis Volume 20, Number 7, July 2014
with mucosal healing, as defined by endoscopic evaluation. NGAL–MMP-9 levels lower than 65 ng/mL were found to be predictive of histological improvement. Prospective studies measuring NGAL–MMP-9 complex levels in patients undergoing anti-TNF therapy are underway to confirm our data and are needed to determine whether this novel serum marker could reduce the amount of endoscopies. It will also be interesting to investigate whether NGAL–MMP-9 complex levels decrease in patients who receive other types of treatment. This will be necessary to validate further that NGAL–MMP-9 is a reliable surrogate marker. As for the functionality of the NGAL– MMP-9 complex in relation to response to the treatment, it is difficult to state whether it plays an active role or is rather a consequence due to less circulating neutrophils as a result of control of inflammation. A study by the group of MacDonald29 suggested that MMPs could play an active role because these are able to degrade anti-TNF antibodies. In this particular study, it was found that MMP-3 and MMP-12 were able to degrade infliximab antibodies, thereby releasing Fc monomers and F(ab)2. These cleaved forms were still functional; however, a reduction was seen in their ability to neutralize soluble TNF-a after coincubation with IBD proteins. This study warrants further research into the potential cleavage of infliximab antibodies by MMP-9 or the NGAL– MMP-9 complex and its subsequent effect on response to treatment. Because the NGAL–MMP-9 complex is a neutrophil product and UC is mostly characterized by influx of neutrophils to the active site of inflammation,2 it may not be surprising that serum NGAL–MMP-9 complex levels correlated strongly with neutrophil counts. However, correlation analysis of neutrophil counts with Mayo endoscopic subscores indicated a weaker correlation (t ¼ 0.308) than that was found for NGAL–MMP-9 levels (t ¼ 0.370). A limitation of the current retrospective study is the lack of information on fecal calprotectin levels. Fecal calprotectin is a sensitive marker of intestinal inflammation and correlates well with the degree of endoscopic activity.30 A recent article by Kolho et al31 showed that fecal MMP-9 had an almost comparable performance to fecal calprotectin in UC patients, suggesting its applicability as a surrogate marker of inflammation. Moreover, the EMBARK study showed that combinations of fecal calprotectin and serum MMP-9 for UC and combinations of fecal calprotectin, serum MMP-9, and serum interleukin-22 in CD demonstrated the strongest association with imaging/endoscopy-defined inflammation.32 These recent studies highlight the current relevance of our study, and future studies should therefore investigate the performance (or accuracy) of serum and/or fecal NGAL–MMP-9 complex levels compared with fecal calprotectin. Nonetheless, the advantages of a serum marker over a fecal marker are evident and are related to patient and caregiver convenience. The present study forms a strong stimulus to obtain more functional information about NGAL–MMP-9 as a complex. So far, only 1 study has investigated its relevance in IBD. Manfredi et al11 showed that urinary NGAL–MMP-9 level was an independent predictor of pediatric IBD. In cancer research, however,
Inflamm Bowel Dis Volume 20, Number 7, July 2014
NGAL–MMP-9 complex is already a well-known biomarker.33– Moreover, an important issue relates to the specificity of NGAL–MMP-9 complex as a biomarker for UC or CD. We trust that the present data will be a stimulus to study this biomarker in large CD patient cohorts. To conclude, we found that serum NGAL–MMP-9 complex levels were increased in patients with active UC compared with HC. NGAL–MMP-9 complex levels significantly decreased after treatment with infliximab and correlated well with mucosal healing. Importantly, NGAL–MMP-9 complex complemented CRP in predicting disease activity and mucosal healing. Therefore, we propagate that NGAL–MMP-9 complex can be used as a surrogate serum marker in UC patients, thereby potentially decreasing the need for endoscopy.
35
Acknowledgments The authors thank Jennifer Vandooren and Erik Martens for their help with zymography analysis and Vera Ballet for managing the VLECC (Vlaams erfelijkheidsonderzoek Crohn en colitis ulcerosa) database of IBD patients at the Leuven University Hospital.
REFERENCES 1. Vermeire S, Van Assche G, Rutgeerts P. Laboratory markers in IBD: useful, magic, or unnecessary toys? Gut. 2006;55:426–431. 2. Qin X. Etiology of inflammatory bowel disease: a unified hypothesis. World J Gastroenterol. 2012;18:1708–1722. 3. Borregaard N. Neutrophils, from marrow to microbes. Immunity. 2010;33: 657–670. 4. Kubben FJ, Sier CF, Hawinkels LJ, et al. Clinical evidence for a protective role of lipocalin-2 against MMP-9 autodegradation and the impact for gastric cancer. Eur J Cancer. 2007;43:1869–1876. 5. Tschesche H, Zolzer V, Triebel S, et al. The human neutrophil lipocalin supports the allosteric activation of matrix metalloproteinases. Eur J Biochem. 2001;268:1918–1928. 6. Kjeldsen L, Johnsen AH, Sengelov H, et al. Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J Biol Chem. 1993;268:10425–10432. 7. Kjeldsen L, Bainton DF, Sengelov H, et al. Identification of neutrophil gelatinase-associated lipocalin as a novel matrix protein of specific granules in human neutrophils. Blood. 1994;83:799–807. 8. Flo TH, Smith KD, Sato S, et al. Lipocalin 2 mediates an innate immune response to bacterial infection by sequestrating iron. Nature. 2004;432: 917–921. 9. Chan YR, Liu JS, Pociask DA, et al. Lipocalin 2 is required for pulmonary host defense against Klebsiella infection. J Immunol. 2009;182: 4947–4956. 10. Nielsen BS, Borregaard N, Bundgaard JR, et al. Induction of NGAL synthesis in epithelial cells of human colorectal neoplasia and inflammatory bowel diseases. Gut. 1996;38:414–420. 11. Manfredi MA, Zurakowski D, Rufo PA, et al. Increased incidence of urinary matrix metalloproteinases as predictors of disease in pediatric patients with inflammatory bowel disease. Inflamm Bowel Dis. 2008;14: 1091–1096. 12. Bolignano D, Della Torre A, Lacquaniti A, et al. Neutrophil gelatinaseassociated lipocalin levels in patients with Crohn disease undergoing treatment with infliximab. J Investig Med. 2010;58:569–571. 13. Oikonomou KA, Kapsoritakis AN, Theodoridou C, et al. Neutrophil gelatinase-associated lipocalin (NGAL) in inflammatory bowel disease: association with pathophysiology of inflammation, established markers, and disease activity. J Gastroenterol. 2012;47:519–530.
NGAL–MMP-9 as a Surrogate Serum Marker in UC
14. Yesil A, Gonen C, Senates E, et al. Relationship between neutrophil gelatinase-associated lipocalin (NGAL) levels and inflammatory bowel disease type and activity. Dig Dis Sci. 2013;58:2587–2593. 15. Nielsen OH, Gionchetti P, Ainsworth M, et al. Rectal dialysate and fecal concentrations of neutrophil gelatinase-associated lipocalin, interleukin-8, and tumor necrosis factor-alpha in ulcerative colitis. Am J Gastroenterol. 1999;94:2923–2928. 16. Arijs I, De Hertogh G, Lemaire K, et al. Mucosal gene expression of antimicrobial peptides in inflammatory bowel disease. PLoS One. 2009; 4:e7984. 17. Vandooren J, Van den Steen PE, Opdenakker G. Biochemistry and molecular biology of gelatinase B or matrix metalloproteinase-9 (MMP-9): the next decade. Crit Rev Biochem Mol Biol. 2013;48:222–272. 18. Baugh MD, Perry MJ, Hollander AP, et al. Matrix metalloproteinase levels are elevated in inflammatory bowel disease. Gastroenterology. 1999;117:814–822. 19. Opdenakker G, Van den Steen PE, Van Damme J. Gelatinase B: a tuner and amplifier of immune functions. Trends Immunol. 2001;22:571–579. 20. Opdenakker G, Van den Steen PE, Dubois B, et al. Gelatinase B functions as regulator and effector in leukocyte biology. J Leukoc Biol. 2001;69: 851–859. 21. Annahazi A, Molnar T, Farkas K, et al. Fecal MMP-9: a new noninvasive differential diagnostic and activity marker in ulcerative colitis. Inflamm Bowel Dis. 2013;19:316–320. 22. Sela-Passwell N, Kikkeri R, Dym O, et al. Antibodies targeting the catalytic zinc complex of activated matrix metalloproteinases show therapeutic potential. Nat Med. 2011;18:143–147. 23. Geboes K, Rutgeerts P, Opdenakker G, et al. Endoscopic and histologic evidence of persistent mucosal healing and correlation with clinical improvement following sustained infliximab treatment for Crohn’s disease. Curr Med Res Opin. 2005;21:1741–1754. 24. Gao Q, Meijer MJ, Schluter UG, et al. Infliximab treatment influences the serological expression of matrix metalloproteinase (MMP)-2 and -9 in Crohn’s disease. Inflamm Bowel Dis. 2007;13:693–702. 25. Vandooren J, Geurts N, Martens E, et al. Zymography methods for visualizing hydrolytic enzymes. Nat Methods. 2013;10:211–220. 26. Descamps FJ, Martens E, Opdenakker G. Analysis of gelatinases in complex biological fluids and tissue extracts. Lab Invest. 2002;82: 1607–1608. 27. Iskandar HN, Ciorba MA. Biomarkers in inflammatory bowel disease: current practices and recent advances. Transl Res. 2012;159:313–325. 28. Tall AR. C-reactive protein reassessed. N Engl J Med. 2004;350: 1450–1452. 29. Biancheri A, Di Sabatino A, Joe-Njoku I, et al. Differential Cleavage of Anti-tumor Necrosis Factor-alpha Agents by Matrix Metalloproteinase (MMP)-10 and MMP-12 in Inflammatory Bowel Disease. Dublin, Ireland: ECCO; 2011. 30. D’Haens G, Ferrante M, Vermeire S, et al. Fecal calprotectin is a surrogate marker for endoscopic lesions in inflammatory bowel disease. Inflamm Bowel Dis. 2012;18:2218–2224. 31. Kolho KL, Sipponen T, Valtonen E, et al. Fecal calprotectin, MMP-9, and human beta-defensin-2 levels in pediatric inflammatory bowel disease. Int J Colorectal Dis. 2014;29:43–50. 32. Faubion WA Jr, Fletcher JG, O’Byrne S, et al. EMerging BiomARKers in Inflammatory Bowel Disease (EMBARK) study identifies fecal calprotectin, serum MMP9, and serum IL-22 as a novel combination of biomarkers for Crohn’s disease activity: role of cross-sectional imaging. Am J Gastroenterol. 2013;108:1891–1900. 33. Yan L, Borregaard N, Kjeldsen L, et al. The high molecular weight urinary matrix metalloproteinase (MMP) activity is a complex of gelatinase B/MMP-9 and neutrophil gelatinase-associated lipocalin (NGAL). Modulation of MMP-9 activity by NGAL. J Biol Chem. 2001;276: 37258–37265. 34. Bolignano D, Donato V, Lacquaniti A, et al. Neutrophil gelatinaseassociated lipocalin (NGAL) in human neoplasias: a new protein enters the scene. Cancer Lett. 2010;288:10–16. 35. Mir SU, Jin L, Craven RJ. Neutrophil gelatinase-associated lipocalin (NGAL) expression is dependent on the tumor-associated sigma-2 receptor S2RPgrmc1. J Biol Chem. 2012;287:14494–14501.
www.ibdjournal.org |
1207
