Enhanced systemic matrix metalloproteinase response in Helicobacter pylori gastritis

Abstract

Background. Helicobacter pylori causes chronic gastritis, peptic ulcer disease, and is the most important risk factor for non-cardia gastric cancer, and has been shown to upregulate matrix metalloproteinases (MMPs) in infected gastric mucosa. MMPs are proteolytic enzymes regulated by tissue inhibitors of metalloproteinases (TIMPs).

Aims. We set up this study to find out whether H. pylori gastritis induces systemic MMP response.

Methods. Serum samples were collected from patients undergoing gastroscopy; 26 patients had H. pylori gastritis and 18 were H. pylori-negative controls with normal gastric mucosa. Serum MMP levels were analysed by enzyme-linked immunosorbent assay.

Results. Significantly elevated serum levels of collagenase-2 (MMP-8), gelatinase B (MMP-9), neutrophil elastase (NE), and myeloperoxidase (MPO), and reduced serum levels of gelatinase A (MMP-2) and TIMP-1 were demonstrated in patients with H. pylori gastritis as compared to H. pylori-negative controls. No significant differences were shown in serum matrilysin-1 (MMP-7) levels.

Conclusions. For the first time, we show enhanced MMP-8 response in H. pylori infection together with other neutrophil degranulation products (MMP-9, MPO, NE). Elevated circulating neutrophil degranulation product levels in serum of H. pylori-positive patients reflect accelerated proteolysis and oxidative stress, and may contribute to extraintestinal sequelae, such as cardiovascular diseases.

• Gastritis
• Helicobacter pylori
• infection
• matrix metalloproteinase

Introduction

Helicobacter pylori causes gastritis, and, although the infection is asymptomatic in most cases, some infected persons develop severe gastric diseases and may be at risk for some extragastric disorders as well. H. pylori is considered the most important single risk factor in peptic ulcer disease and non-cardiac gastric cancer 1. Despite the activated immune response of the host, as shown for example by abundant polymorphonuclear and mononuclear leukocyte infiltration in the gastric mucosa as well as high levels of specific H. pylori antibodies 2, the chronic infection persists for decades unless actively eradicated 3. Neither the reasons for chronicity nor those for the development of serious sequelae of the infection are completely clarified.

Matrix metalloproteinases (MMPs) are genetically distinct but structurally related zinc-dependent metalloendopeptidases, which can be classified based on their primary structures and substrate specificities into different groups, such as collagenases (MMP-1, -8, -13), gelatinases (MMP-2, -9), stromelysins (MMP-3, -10, -11, -19), matrilysins (MMP-7, -26), membrane-type MMPs (MT-MMPs) (MMP-14, -15, -16, -17, -24, -25), and other MMPs 4, 5. MMPs can collectively degrade almost all components of extracellular matrix and basement membrane, as well as process serpins, growth factors, pro- and anti-inflammatory cytokines and chemokines, as well as apoptotic signals to modulate immune response 4–7. They have also been shown to be important in malignant pathological processes 8. The excess activity of MMPs leads to host damage by causing connective tissue break-down 5, 7. In host defence against bacteria, leukocytes migrate to the site of inflammation and, due to infectious or inflammatory stimuli, MMP-8 and MMP-9 are released and locally activated 5, 7, 9. The cleavages of non-matrix cytokine and chemokine substrates of MMP-8 and MMP-9 can be decisive and subsequently can direct the pro- and anti-inflammatory actions of the neutrophil-derived MMPs 5, 7, 9. In this regard, MMP-8 and MMP-9 have recently been shown to modify immune response by modulating chemokinesis and to exert defensive and anti-inflammatory properties 5, 6. Recently the role of MMPs in immune response has been under keen investigation in various infectious diseases 7. The activity of MMPs is regulated by antiproteases called tissue inhibitors of metalloproteinases (TIMPs) 10 and recently some specific TIMPs, including TIMP-1, were demonstrated to be locally upregulated in the gastric mucosa of H. pylori patients 11. Neutrophil elastase (NE) is a serine protease which is degranulated from azurophilic granules of neutrophils when exposed to inflammatory stimuli and microbial virulence factors. NE can degrade almost all extracellular matrix and plasma proteins as well as activate proMMPs and inactivate TIMP-1 5, 12, 13. Degranulation of neutrophils releases also antimicrobial substances such as myeloperoxidase (MPO). MPO, by generating oxidants such as hypochlorous acid, is able not only to oxidatively inactivate pathogenic microbes but also to inactivate TIMP-1 as well as to activate latent proMMPs 5, 12, 14. Thus, MPO oxidatively and NE proteolytically can potentiate the destructive MMP cascades.

Key messages

• For the first time, we show enhanced MMP-8 response in Helicobacter pylori infection together with other neutrophil degranulation products (MMP-9, myeloperoxidase, neutrophil elastase).

• Elevated circulating neutrophil degranulation product levels in serum of H. pylori-positive patients reflect accelerated proteolysis and oxidative stress, and may contribute to extraintestinal sequelae, such as cardiovascular diseases.

The role of MMPs in gastrointestinal tract diseases, such as gastric ulcer 15 and possibly cancer 16, 17, has been suggested. As gastric ulcer and cancer are important late gastric sequelae of a chronic H. pylori infection, the association of MMPs with H. pylori infection has recently gained a lot of interest 7. In fact, regarding H. pylori infection, increased expression of MMP-2 18–22, MMP-7 19, 20, 23, 24, and MMP-9 18–22, 25, 26 has been demonstrated by cell and gastric biopsy experiments whereas the role of MMP-8 is practically unknown 11. The present study was set up to find out whether H. pylori gastritis induces an enhanced MMP-8 response and if the local gastric infection is also reflected in a systemic MMP-2, -7, -9, TIMP-1, NE, and MPO response by measuring circulating levels in serum of these patients.

Methods

Patients

A total of 26 Caucasian patients (age range 42–82, median 64 years; 18 females) with histologically proven H. pylori gastritis and 18 H. pylori-negative Caucasian patients (age range 62–85, median 72 years; 15 females) with macroscopically and histologically proven normal gastric mucosa were included. Both the patients and the controls were gastroscopied at Herttoniemi Hospital in 2004–2005. Table I shows the underlying diseases of the patients. The indications for and macroscopical and histological findings of the H. pylori-positive patients are shown in Table II. The gastroscopy indications for the controls were the following: reflux (seven patients), gastric pain (five patients), anaemia (three patients), weight loss (one patient), and suspected coeliac disease (two patients). The H. pylori-negative controls were older than the H. pylori-positive patients (P=0.0241, Mann-Whitney test).

Table I.  Underlying diseases of the Helicobacter pylori-positive patients and the H. pylori-negative controls. No statistically significant differences in the presence of any of the underlying diseases were found between the two groups.

	Number of patients with underlying diseases
Underlying disease	H. pylori-positive patients (n=26)	H. pylori-negative controls (n=18)
Hypertension	7	8
Coronary artery disease	4	3
Atrial fibrillation	1	1
Hypercholesterolaemia	4	3
Diabetes type I	0	1
Diabetes type II, no complications	0	2
Diabetes type II, nephropathy	0	1
Bronchial asthma	3	4
Epilepsy	1	0
Sarcoidosis	1	0
Osteoarthritis	2	3
Rheumatoid arthritis	0	1
Osteoporosis	3	4
Migraine	1	0
Hiatal hernia, Barrett's oesophagus	0	1
Crohn's disease	0	1
Diverticular disease and irritable colon	2	2
Prostatic hyperplasia	2	1

Table II.  Characteristics of Helicobacter pylori-positive gastroscopied patients.

No	Sex	Age (y)	Indication for endoscopy	Erosions and ulcers	Chronic^a inflammation antrum/corpus	Acute^b inflammation antrum/corpus
1	F	56	coeliac disease?		3/2	1/1
2	M	68	dyspepsia		3/3	2/2
3	F	43	coeliac disease?		2/2	1/1
4	F	43	reflux		2/1	1/1
5	F	49	H. pylori resistance		3/2	1/1
6	F	79	anaemia		2/1	3/1
7	F	65	dyspepsia		3/2	2/2
8	F	64	reflux		2/2	1/1
9	F	80	reflux		3/2	1/1
10	F	57	dyspepsia	gastric ulcer	2/2	2/2
11	M	42	dyspepsia		2/1	1/1
12	M	77	weight loss		2/2	1/1
13	F	58	dyspepsia	gastric ulcer	2/1	1/1
14	M	71	reflux		2/1	1/0
15	M	82	anaemia	few^c erosions	2/2	0/2
16	F	58	dyspepsia		2/2	1/1
17	F	70	dyspepsia		2/3	0/2
18	F	70	H. pylori resistance		1/1	0/0
19	M	54	reflux		2/1	2/2
20	F	54	dyspepsia	few erosions	1/0	1/0
21	F	50	reflux		2/1	3/1
22	F	62	dyspepsia		2/1	1/1
23	F	73	dyspepsia		2/2	1/1
24	M	71	reflux		2/2	1/1
25	M	78	dyspepsia		3/1	2/2
26	F	66	dyspepsia	gastric ulcer	3/2	1/1

^aChronic (plasma cell/lymphocytic) inflammation was scored (0=none, 1=mild, 2=moderate, 3=severe) separately for antrum and corpus.

^bAcute (granulocytic) inflammation was scored (0=none, 1=mild, 2=moderate, 3=severe) separately for antrum and corpus.

^c1–5 macroscopical minimal erosions in the gastric mucosa.

The Ethics Committee of the Hospital District of Helsinki and Uusimaa approved the study, and all the participants gave their written informed consent.

Gastric histology

The gastric biopsies (two from antrum and two from corpus) were stained with haematoxylin-eosin, Alcian Blue (pH 2.5)-periodic acid-Schiff, and modified Giemsa stains. Grade of gastritis (Table I) was assessed by one experienced pathologist according to the updated Sydney system 27.

Serum

Serum samples were collected from patients in context with the endoscopy and were stored at −20°C until analysed. MMP-2, -7, -8, -9, NE, MPO, and TIMP-1 analyses were carried out by enzyme-linked immunosorbent assay as earlier described 28, 29. MMP-2, MMP-7, MMP-8, MMP-9, and TIMP-1 concentrations were determined using commercially available enzyme-linked immunosorbent assay (ELISA) kits. Biotrak ELISA systems (Amersham Biosciences UK Ltd, Buckinghamshire, UK) were used for MMP-2, MMP-8, and MMP-9 according to the manufacturer's protocol. DuoSet ELISA development Systems (R&D Systems, Minneapolis, USA) for TIMP-1 and Quantakine MMP-7 (total) immunoassay for MMP-7 were used correspondingly. All samples were analysed in duplicate. The used MMP- and TIMP-1 ELISAs detect as informed by the manufacturer active, pro-, complexed and fragmented forms of the studied MMPs and TIMP-1. MPO and NE levels were measured according to manufacturer's instructions by commercial kits purchased from Immunodiagnostic AG (Bensheim, Germany) and Bender MedSystems mbH (Vienna, Austria), respectively. The secondary antibody in each kit was conjugated with horseradish peroxidase, and tetramethyl benzidine was used as a substrate. The absorbance was measured at 450 nm using Labsystems Multiskan RC (Thermo Bioanalysis Corporation, Santa FE, USA). The levels of MMPs, TIMP-1, MPO, and NE were expressed as ng per mL, and for calculation of MMP/TIMP-1 ratios the levels were converted to mol per L.

Serum C-reactive protein (CRP) levels (≥0.3 mg/L) were measured by routine methods.

Statistical analysis

Data were analysed by using GraphPad Prism version 4.0c (GraphPad Inc., San Diego, California, USA). Data of two groups were compared by the Mann-Whitney test. Data are presented as median (25%–75% percentiles). Correlations were studied using Spearman's rank correlation test. A P-value less than 0.05 was considered statistically significant.

Results

The serum levels of sensitive CRP did not significantly differ (P=0.5828) between the H. pylori-positive gastritis patients (median 1.0, range, <0.3–7.0 mg/L) and the H. pylori-negative controls with normal gastric mucosa (median 0.95, range <0.3–12.0 mg/L). However, MMP-8, MMP-9, NE, and MPO serum levels of patients with H. pylori gastritis were significantly elevated when compared to the serum levels of H. pylori-negative control patients (Figure 1). The levels were 27.6 (21.2–39.6) versus 11.3 (8.05–14.5) ng of MMP-8 per mL, P<0.0001 (Figure 1A), and 170 (132–251) versus 112 (73.5–201) ng of MMP-9 per mL, P=0.0290 (Figure 1B). For NE and MPO, the corresponding levels were 169.5 (120.5–225.0) versus 89.0 (69.50–133.5) ng of NE per mL, P=0.0004 (Figure 1C), and 81.50 (60.50–109.5) versus 46.50 (35.00–68.00) ng of MPO per mL, P=0.0075 (Figure 1D), respectively. The serum MMP-2 levels of H. pylori gastritis patients were significantly decreased when compared to those of H. pylori-negative control patients (1921 (1616–2369) versus 2787 (2040–3420) ng of MMP-2 per mL, P=0.004) (Figure 2A). Furthermore, serum levels of TIMP-1 were significantly lower in H. pylori-positive patients 61.50 (52.00–76.00) than in H. pylori-negative controls 78.00 (61.00–85.50) ng per mL, P=0.0463 (Figure 2B). In addition, the MMP-8/TIMP-1 and MMP-9/TIMP-1 ratios for H. pylori-positive patients were significantly higher than those for H. pylori-negative controls (P<0.0001 and P=0.0047, respectively). On the contrary, serum levels of MMP-7 between H. pylori-positive and -negative subjects did not significantly differ (P=0.242). MMP-7 levels correlated with acute inflammation (r=0.47, P=0.016, scores for antrum and corpus together), but the levels of other biomarkers tested did not correlate either with acute or chronic or total inflammation scored either separately for antrum and corpus or both together. Among the H. pylori-negative control patients, the levels of MMP-7 (r=0.67, P=0.0022), MMP-9 (r=0.47, P=0.0480), and TIMP-1 (r=0.55, P=0.0185) correlated with increasing age.

Figure 1.  Serum levels of neutrophilic degranulation markers matrix metalloproteinase (MMP)-8 (A), MMP-9 (B), neutrophil elastase (NE) (C), and myeloperoxidase (MPO) (D) of 26 Helicobacter pylori gastritis patients and 18 H. pylori-negative controls with normal gastric mucosa. Data of two groups compared by the Mann-Whitney test. The horizontal line indicates median.

Figure 2.  Serum matrix metalloproteinase (MMP)-2 (A) and tissue inhibitor of metalloproteinase (TIMP)-1 (B) levels of 26 Helicobacter pylori gastritis patients and 18 H. pylori-negative controls with normal gastric mucosa. Data of two groups compared by the Mann-Whitney test. The horizontal line indicates median.

Discussion

We found significantly elevated serum levels of neutrophilic degranulation markers MMP-8, MMP-9, NE, and MPO in patients with H. pylori gastritis as compared to H. pylori-negative controls with macroscopically and histologically normal gastric mucosa. In addition, significantly decreased levels of serum MMP-2 and TIMP-1 were detected in H. pylori gastritis patients when compared to H. pylori-negative controls. These findings indicate, for the first time, that H. pylori infection is associated with an enhanced MMP-8 response. The finding strongly suggests enhanced neutrophil degranulation further supported by elevated serum levels of NE and MPO in H. pylori-positive patients as compared to H. pylori-negative controls. Elevated MPO may indeed potentiate the oxidative activation of both MMP-8 and MMP-9 as well as oxidatively inactivate TIMP-1 12, 14. Enhanced levels of MMP-8 and MPO indicate increased degranulation of specific granules of neutrophils. NE, a serine protease, can also accelerate MMP-cascades by activating latent proMMPs and inactivating TIMP-1 5, 12. The lower serum levels of TIMP-1 and higher MMP-8/TIMP-1 ratios in H. pylori gastritis patients as compared to H. pylori-negative individuals suggest that the proteolytic activity of MMP-8 may be relatively unopposed by TIMP-1 in chronic H. pylori gastritis or that the higher TIMP-1 levels in controls were due to the older age.

Earlier studies have shown an increased production of MMP-9, another MMP from C-particles or tertiary granules of neutrophils 5, 12, in H. pylori-infected gastric and other cells 18, 20–22, 25, 26, 30 and in gastric biopsies from H. pylori-positive patients 18, 25. In the present study, we could confirm and further extend the association of H. pylori infection with elevated MMP-9 production, and we could show, for the first time, also a systemic elevation of MMP-9 in serum of patients with chronic H. pylori gastritis. In addition to MMP-9, we also assayed the serum levels of MMP-2, both of which can degrade type IV collagen abundant in basement membranes. Earlier studies have demonstrated in H. pylori infection that local expression of MMP-2 in gastric biopsies or in infected cell lines is upregulated 18–22 or there is no effect at all 25, 31. In the present study, the serum MMP-2 levels were found to be lower in H. pylori-positive patients as compared to H. pylori-negative subjects with normal gastric mucosa. Recently McQuibban et al. 32 showed that MMP-2 cleaved monocyte chemoattractant protein-3, which resulted in chemokine antagonism and reduced inflammation. In our study, the systemic ‘downregulation’ of MMP-2 in H. pylori infection may reflect, at least to some extent, enhanced inflammation against H. pylori infection.

MMP-7 expression levels have been shown to be enhanced in H. pylori-infected cell lines or in gastric biopsies 19, 20, 23, 24, and MMP-7 has been found to be capable of activating MMP-8 33. However, in the present study, the serum levels of MMP-7 did not significantly differ between H. pylori-positive patients and controls with normal gastric mucosa.

Although it is generally considered that H. pylori has no proven role in extraintestinal sequelae, except probably for unexplained iron deficiency anaemia and idiopathic thrombocytopenic purpura 1, a recent meta-analysis indicated an association between more virulent CagA-positive (positive for protein coded by cytotoxin-associated gene) H. pylori infection and ischaemic heart disease 34. H. pylori 18, 19, 30 and especially CagA-positive strains 20, 22, 25 have been shown to upregulate MMP-9 production. Furthermore, according to a recent finding, successful therapy, in contrast to unsuccessful treatment, of H. pylori infection results in downregulation of MMP-9 in gastric biopsies taken from patients after therapy 35. There is increasing evidence supporting an association of MMPs, especially MMP-9, with the pathogenesis of cardiovascular diseases 36. Recent studies have shown that circulating MMP-8 and MMP-9 concentrations together with an elevated MMP-8/TIMP-1 ratio were significantly associated with cardiovascular disease deaths, indicating that elevated circulating MMP-8, MMP-9 and MMP-8/TIMP-1 can be regarded as important prognostic factors in patients with cardiovascular diseases 37–39. We found in this study that chronic H. pylori infection, lasting for decades unless eradicated, induces a systemic host response by increasing the serum levels of two functionally distinct MMPs, a collagenase MMP-8 and a gelatinase MMP-9, both of which may have a role in cardiovascular diseases 37–39. In addition, serum MMP-8/TIMP-1 and MMP-9/TIMP-1 ratios were significantly elevated. Overall, this suggests that, by increasing MMP-8 and MMP-9 (together with other tissue-destructive neutrophil degranulation products, MPO and NE) in systemic circulation where the activity is only moderately opposed by low levels of TIMP-1, chronic H. pylori infection may predispose to cardiovascular diseases. This may be especially possible in combination with elevated levels of sensitive serum CRP. Nevertheless, in our study, there was no significant difference in the serum CRP levels between the H. pylori-positive and -negative subjects.

The reasons why only a small number of H. pylori-infected individuals develop serious gastric and extragastric sequelae are not known, but bacterial, host, and environmental factors are thought to be involved 1. MMPs are enzymes which influence the host immune response in infectious diseases 7 and contribute to the pathogenesis of gastric ulcers 15 and possibly cancer 16, 17, 40. Recently in a work by Hellmig and co-workers, patients with certain genetic variants of MMP-9 had a higher risk for gastric ulcers than others 15. Furthermore, elevated plasma levels of MMP-9 have been demonstrated in patients with gastric cancer 41. In the present study, none of the Helicobacter-positive gastritis patients had duodenal ulcer but three of them had gastric ulcer. The MMP-9 levels of these three gastric ulcer patients were clearly above the median of all patients, and close or above the 75% percentile value (data not shown). In addition to MMP-9, the efficient collagen-degrading protease MMP-8 has recently been suggested to have a role in gastrointestinal diseases such as gastric cancer 40. Thus, it is not surprising that elevated levels of MMP-8, mostly originating from neutrophils, were detected in H. pylori gastritis in the present study. MMP-8 expression and levels are mainly regulated by selective release of the contents of specific granules upon neutrophil degranulation 5, 7, 12. Prepacked and presynthesized MMP-8 is released by triggered neutrophils without significant de novo synthesis 12. This also concerns two other neutrophilic degranulation products studied, i.e. MPO and NE. Therefore, the systemic elevated level of MMP-8 (and that of MPO and NE) in H. pylori infection as shown in the present study is not eventually reflected by local gastric production of MMP-8 mRNA 11.

Our main interest was to see if local H. pylori gastritis could associate with a systemic MMP response as demonstrated by circulating serum MMP and neutrophilic degranulation product (MPO and NE) levels, and thus it was crucial to include a control group with endoscopically verified normal gastric mucosa. Our control patients were all referred to upper endoscopy for various reasons and could have other diseases or infectious processes than H. pylori that may increase the MMP levels. However, no significant differences were detected either in the presence of underlying diseases or in the serum CRP levels between the H. pylori-positive gastritis patients and the controls.

In conclusion, in H. pylori gastritis the serum levels of MMP-8, MMP-9, MMP-8/TIMP-1, MMP-9/TIMP-1, MPO, and NE are elevated, but those of MMP-2 and TIMP-1 decreased as compared to the serum levels in H. pylori-negative individuals with healthy gastric mucosa. Our results for the first time suggest the association of H. pylori infection with a systematically enhanced neutrophil-derived proteolytic and oxidative response which may contribute to extraintestinal sequelae, such as cardiovascular diseases. Our novel findings raise the question whether we should search and treat H. pylori infection more actively, not only because of abnormalities in the gastrointestinal tract but also considering extraintestinal sequelae.

Acknowledgements

This study was in part presented at the XXth International Workshop on Helicobacter and Related Bacteria in Chronic Digestive Inflammation, European Helicobacter Study Group, Istanbul, Turkey, 20–22 September 2007. This work has been supported by grants from the Academy of Finland and Helsinki University Central Hospital Research (EVO) Funds. Declaration of interest: Dr Pentti Sipponen has served as a speaker, a consultant, and an advisory board member for Biohit Plc (a company that produces laboratory pipettes and develops and markets laboratory tests), and as a speaker for meetings and congresses of many pharmaceutical companies. In the last 10 years, he has not received any funds from any companies or organizations, except the funding from the University Hospital. Dr Sipponen owns stocks and shares in Biohit Plc, including some other companies on the stock-market in Helsinki. Dr Sipponen does not own any personal patents. The other authors have no competing interests. The authors alone are responsible for the content and writing of the paper.