Meta-analysis of myeloperoxidase G-463/A polymorphism in anti-neutrophil cytoplasmic autoantibody-positive vasculitis

Abstract

Wegener's granulomatosis, microscopic polyangiitis and Churg Strauss syndrome are small-vessel vasculitides associated with anti-neutrophil cytoplasmic antibodies (ANCA) directed against proteinase 3 (PR3) and myeloperoxidase (MPO). A G to A polymorphism at position 463 in the promoter region of the MPO gene, which leads to the loss of a SP1 transcription binding site in an Alu hormone responsive element, reduces MPO expression. We hypothesized that MPO alleles may play a role in determining disease susceptibility or severity in ANCA-associated vasculitis (AASV). MPO genotypes were determined by restriction fragment length polymorphism polymerase chain reaction (RFLP/PCR) in 134 Caucasian patients (Wegener's granulomatosis, n = 69; microscopic polyangiitis, n = 65; PR3–ANCA n = 91; MPO–ANCA, n = 43) and 150 matched healthy controls. There was no difference in survival to renal failure or death in patients with the different MPO alleles (χ² = 0·904, P = 0·6362) or in presenting serum creatinine concentration based on MPO genotype (χ² = 0·389, P = 0·8232). There was no significant difference in genotype frequencies between controls (13AA, 102GG, 35GA) and patients (14AA, 97GG, 23GA: χ² = 1·75, P = 0·417), patients with Wegener's granulomatosis (5AA, 53GG, 11GA: χ² = 1·864, P = 0·3938) or patients with microscopic polyangiitis (9AA, 44GG, 12GA: χ² = 1·682, P = 0·4317). A meta-analysis of our study and two previous studies showed that there was no association between the myeloperoxidase G-463/A polymorphism and the risk of developing ANCA-associated vasculitis; GG versus GA plus AA (odds ratio 1·14; 95% confidence interval 0·86–1·50). The MPO G-463/A polymorphism is not a risk factor for the development or severity of AASV.

Introduction

Autoantibodies to myeloperoxidase (MPO) are found in microscopic polyangiitis and Churg Strauss syndrome and less commonly in Wegener's granulomatosis. Myeloperoxidase (MPO) is found in polymorphonuclear leucocytes, monocytes and some macrophages [1], and catalyses the interaction between hydrogen peroxide and chloride to generate hypochlorous acid and other anti-microbial compounds. ANCA-mediated activation of neutrophils and monocytes leads to the release of MPO and the generation of reactive oxygen and nitrogen products which are thought to mediate endothelial injury in systemic vasculitis [2–4]. Several single nucleotide polymorphisms have been identified in the MPO locus, one of which is located in the promoter region (−463G to A) and leads to the loss of an SP1 transcription binding site in an Alu hormone-responsive element [5].The G allele was reported to be associated with a threefold increase in transcriptional activity of the gene in a transient transfection assay [5]. We hypothesized that if MPO played a role in the pathogenesis of endothelial injury in ANCA-associated vasculitis (AASV) then the MPO G-463/A polymorphism would influence disease severity and prognosis. An increased incidence of the MPO A allele was associated with disease relapse in a patient cohort with AASV. Furthermore, the GG genotype was found to be associated with an increased risk for disease susceptibility in females with MPO–ANCA associated vasculitis but not males [6]. A subsequent study, however, did not show an association between this polymorphism and the risk of developing vasculitis [7]. These two previous studies yielded inconclusive results. It also seemed likely that our study would not have sufficient power to be certain that there was no association between the MPO G-463/A polymorphism and susceptibility to AASV. To address this we carried out a meta-analysis of the data from our study and the two previous studies on the impact of the MPO G-463/A polymorphism on the risk of developing AASV.

Patients and methods

Subjects for MPO allotyping

We studied 150 disease-free Caucasian individuals and 134 Caucasian patients with AASV. Patients were put into disease categories using the Chapel Hill Consensus conference definitions [8]. In addition to a positive ANCA status, all patients had biopsy evidence of vasculitis.

ANCA testing

ANCA activity of samples was determined by indirect immunofluorescence on ethanol-fixed neutrophils using standard techniques [9] and by antigen-specific enzyme-linked immunosorbent assay (ELISA), as described previously [10].

Determination of MPO allotypes by polymerase chain reaction (PCR) amplification

Genomic DNA was isolated using a nucleic acid extraction kit (Fisher Scientific, Loughborough, UK) from peripheral blood obtained from subjects using a Perkin Elmer DNA Thermocycler. PCR products were generated using 250 ng of genomic DNA as a template and the forward primer 5′-CCG TAT AGG CAC ACA ATG GTG AG-3′ and reverse primer 5′-GCA ATG GTT CAA GCG ATT CTT C-3′. The PCR reaction was carried out using a 30-µl reaction with a final concentration of 50 µm 2′-deoxyribonucleoside 5′-triphosphate (dNTPs), 1·5 mM MgCl, 0·1 µM for each primer and 1 unit of Taq polymerase. After denaturation at 95°C for 5 min, the cycling conditions were as follows: 35 cycles at 94°C for 1 min, 56°C for 1 min and 72°C for 1 min with a final extension at 72°C for 7 min. The PCR products were then digested with 5 units of AciI overnight at 37°C and separated on a 2% agarose gel. The G to A substitution at position 463 leads to the loss of a AciI restriction site which yields a 61-base pairs (bp) fragment that serves as an internal control (Fig. 1). The PCR products were then analysed through a 1% agarose gel in TBE buffer [90 mM borate acid, 2·5 mM ethylenediamine tetraacetic acid (EDTA)] containing 1 µg/ml of ethidium bromide and visualized by UV transillumination.

Fig. 1.

Agarose gel electrophoresis of amplified DNA by restriction fragment length polymorphism polymerase chain reaction using allele specific primers for myeloperoxidase.

Neutrophil isolation

Neutrophils were isolated using a method adapted from Toothill et al. [11]. Peripheral blood was collected from healthy donors into tubes containing acid citrate dextrose (9 : 1 dilution). The blood cells were first sedimented across Hespan (2·5% hydroxyethyl starch), then neutrophils were isolated by density gradient centrifugation at 500 g using isotonic Percoll. Neutrophils were 99% viable by trypan blue exclusion and were 98–99% pure when stained with haematoxylin.

Myeloperoxidase assay

MPO activity was analysed in neutrophils from individuals of known MPO allotypes (3A/A, 3G/A and 3G/G). MPO activity was assayed spectrophotometrically by determining the decomposition of hydrogen peroxide using o-dianisidine as the hydrogen donor [12]. Thus, 5 × 10⁶ neutrophils were lysed with ice-cold 0·5% hexadecyltrimethyl–ammonium bromide in 50 mmol/l phosphate buffer (pH 6·0) (50 µl) for 30 min at room temperature. The cell lysate was then centrifuged for 15 min at 12 500 g. MPO activity was determined by the addition of 0·1 ml of the supernatant to 2·9 ml of 50 mmol/l phosphate buffer containing 0·167 mg/ml o-dianisidine dihydrochloride (Sigma, Gillingham, UK) and 0·0005% hydrogen peroxide in a microwell plate. Purified MPO (Calbiochem, Nottingham, UK) was used at concentrations of 79 pg/ml to 250 pg/ml, diluted in 50 mmol/l phosphate buffer, to produce a standard curve in order to determine that the spectrophotometric assay using o-dianisidine as a substrate provided accurate measurement of MPO activity. Neutrophil numbers per well were also assessed using a standard curve prepared using 3·5 × 10⁶−5 × 10⁶ neutrophils per well in order to determine that a linear dose relationship existed between neutrophil number and MPO activity. For subsequent experiments the MPO supernatant fraction from 5 × 10⁶ neutrophils were assayed per sample. All assays were performed in triplicate.

Absorbances at 450 nm were measured immediately after addition of the chromogen, o-dianisidine dihydrochloride using a Titertek spectrophotometer and the change in absorbance at 460 nM over a 5-min period was measured at 25°C and the linear portion of the tracing was used for analysis. The molar absorbency of oxidized o-dianisidine (1·13 × 10⁴/min) was used to calculate the moles of hydrogen peroxide decomposed, one unit of MPO activity being the quantity of enzyme being able to convert 1 µmol of hydrogen peroxide to water at 25°C. Therefore, 1 µmol (i.e. 1 unit) of hydrogen peroxide gives a change in absorbance of 1·13 × 10²/min. Furthermore, the unit of MPO per well (5 × 10⁶ neutrophils) was determined by the maximal change in absorbance (A) divided by the time period (5 min), the sum of which is multiplied by the molar absorbency coefficient 1·13 × 10²/min. All assays were performed in triplicate.

$$\begin{array}{c}\text{Unit MPO activity per sample}(5\times {10}^6\text{neutrophils})=\\ (\text{A}/5\text{min})\times 1·13\times {10}^2/\text{min.}\end{array}$$

Statistical analysis

Differences in allotype and allele frequencies between ANCA-positive vasculitis patients and controls were analysed by applying the χ² test. To reject the null hypothesis, a probability of 0·05 (two-tailed) was used. Differences in normally distributed continuous variables were compared using Student's t-test and for non-normally distributed data by the Mann–Whitney test. For comparisons between more than two groups, the repeated-measures analysis of variance or Kruskal–Wallis tests were performed. Two-sided P-values less than 0·05 were considered significant. Survival to end-stage renal failure or death by MPO G-463/A allotype was examined by the Kaplan–Meier product limit method and equality of survival between groups was tested by the log-rank test (Stata Statistical Software, Release 5·0 College Station, TX, USA).

Meta-analysis

Two investigators extracted the data individually from the three studies. The meta-analysis examined the effect of the at risk allele GG which increases myeloperoxidase transcription against the GA and AA alleles. In the study by Fiebeler et al. [7], allele frequencies of the MPO G-463/A polymorphism in controls was derived from a reference population in northern Germany [13]. In an initial analysis we evaluated GG versus GA and AA, GG and GA versus AA, and GG versus AA in patients with AASV compared with controls. A further analysis was carried out comparing GG versus GA and AA in patients with anti-MPO antibodies or with anti-PR3 antibodies, with controls. We analysed the data using the fixed effects model and generated odds ratios with 95% confidence intervals (CI) using Review Manager 4·2. We estimated between study heterogeneity using the χ²-based Cochrane Q statistic.

Results

Patient demographics

Of the 134 patients with AASV, 69 had Wegener's granulomatosis and 65 had microscopic polyangiitis. Ninety-one patients had PR3–ANCA and 43 had MPO–ANCA. All patients and controls were genotyped successfully.

Kaplain–Meier survival analysis

There was no difference in survival to renal failure or death in patients with the different MPO alleles (χ² = 0·904, P = 0·6362) (Fig. 2).

Fig. 2.

Kaplain–Meier survival estimates by myeloperoxidase 463G/A polymorphism.

MPO genotype and renal function

There was no significant difference in serum creatinine concentration at presentation in patients with AASV based on MPO genotype (χ² = 0·389, P = 0·8232) (Fig. 3).

Fig. 3.

Serum creatinine (Creat) analysis by myeloperoxidase 463G/A polymorphism.

Overall distribution of MPO genotypes

There was no difference in the overall distribution of MPO genotypes in controls and vasculitis patients (χ² = 1·75, P = 0·417) (Table 1). No skewing was observed in gene frequency between controls and vasculitis patients (χ² = 0·032, P = 0·8554).

Table 1.

Overall distribution of myeloperoxidase genotypes and allele frequencies in controls and vasculitis patients.

	Controls n = 150 (%)	Vasculitis patients n = 134 (%)
Genotype
AA	13 (8·7)	14 (10·4)
GA	35 (23·3)	23 (17·2)
GG	102 (68)	97 (72·4)
		χ2 = 1·75
		P = 0·417
Gene frequency (%)
A	(20·3)	(19)
G	(79·7)	(81)

Overall distribution of MPO genotypes divided by disease category

There was no difference in the distribution of MPO genotypes when controls were compared with Wegener's patients (χ² = 1·864, P = 0·3938) or microscopic polyangiitis patients (χ² = 1·682, P = 0·4317). No skewing was observed in gene frequency between controls and Wegener's patients (χ² = 0·866, P = 0·3521) or patients with microscopic polyangiitis (χ² = 0·2676, P = 0·6056) (Table 2).

Table 2.

Overall distribution of myeloperoxidase genotypes and allele frequencies in controls and patients with Wegener's granulomatosis (WG) and microscopic polyangiitis (MPA).

	Controls n = 150 (%)	WG patients n = 69 (%)	MPA patients n = 65 (%)
Genotype
AA	13 (8·7)	5 (7·2)	9 (13·8)
GA	35 (23·3)	11 (16)	12 (18·5)
GG	102 (68)	53 (76·8)	44 (67·7)
		χ2 = 1·864	χ2 = 1·682
		P = 0·3938	P = 0·4317
Gene frequency (%)
A	(20·3)	(15·2)	(23·1)
G	(79·7)	(84·8)	(76·9)

Overall distribution of MPO genotypes by ANCA specificity

There was no significant difference in MPO genotypes between controls and patients with MPO–ANCA (χ² = 3·527, P = 0·7154) or PR3–ANCA (χ² = 3·783, P = 0·1508) (Table 3). No skewing was observed in gene frequency between controls and MPO–ANCA-positive patients (χ² = 0·8666, P = 0·3521) or PR3–ANCA-positive patients (χ² = 2·667, P = 0·1025).

Table 3.

Overall distribution of myeloperoxidase (MPO) genotypes divided by anti-neutrophil cytoplasmic antibodies (ANCA) specificity.

	Controls n = 150 (%)	PR3–ANCA n = 91 (%)	MPO–ANCA n = 43 (%)
Genotype
AA	13 (8·7)	7 (7·7)	7 (16·3)
GA	35 (23·3)	17 (14·3)	10 (23·3)
GG	102 (68)	71 (78·0)	26 (60·4)
		χ2 = 3·783	χ2 = 3·527
		P = 0·1508	P = 0·7154
Gene frequency (%)
A	(20·3)	(15·4)	(30·2)
G	(79·7)	(84·6)	(69·8)

Overall distribution of MPO genotypes by gender

No statistically significant difference was noted when comparing controls against patients based on gender (males χ² = 0·491, P = 0·782; females χ² = 2·565, P = 0·2774).

MPO assay

MPO unit activity/5 × 10⁶ neutrophils were significantly higher in neutrophils from MPO 463G/G donors (0·7123 ± 0·1982), compared with MPO 463A/A donors (0·1321 ± 0·1691) (P < 0·05) or MPO G/A donors (0·4019 ± 0·1522) (P < 0·05).

Meta-analysis

We found no association between the GG genotype and the risk of developing AASV (Table 4). For all patients with AASV, the odds ratio (OR) for the risk of developing vasculitis was 1·14 (95% CI 0·86–1·50) when GG homozygous patients were compared with the other genotypes combined. None of the other comparisons of genotypes were significant. Similarly, there was no association between the GG genotype compared with other genotypes combined in patients with MPO–ANCA vasculitis (OR 1·21; 95% CI 0·82–1·81) or PR3–ANCA vasculitis (OR 1·09; 95% CI 0·79–1·51). There was no significant heterogeneity between the trials for any of the outcomes.

Table 4.

Summary odds ratios (OR) for the association of anti-neutrophil cytoplasmic antibodies (ANCA)-associated vasculitis with myeloperoxidase (MPO) G-463/A polymorphism.

Comparison group	Studies	Participants	OR (fixed), 95% CI
ANCA-associated vasculitis versus controls
GG versus GA + AA	3	944	1·14 (0·86–1·50)
GG + GA versus AA	3	944	0·77 (0·44–1·36)
GG versus GA	3	888	1·23 (0·91–1·65)
GA versus AA	3	327	0·64 (0·35–1·18)
GG versus AA	3	673	0·83 (0·47–1·47)
MPO–ANCA-associated vasculitis versus controls	3	690	1·21 (0·82–1·81)
PR3–ANCA-associated vasculitis versus controls	3	803	1·09 (0·79–1·51)

Test for heterogeneity P > 0·1 for all comparisons. CI: confidence interval.

Discussion

ANCA has been shown to ligate neutrophil cell surface expressed target antigens to Fcγ receptors and this results in the generation of reactive oxygen production [2,14] and nitric oxide [3], as well as degranulation with the releaseof granule constituents, MPO, PR3 and elastase [15]. MPO is a haem enzyme found in the azurophilic granules of neutrophils and monocytes that is released after ANCA activation of neutrophils. MPO catalyses a reaction between hydrogen peroxide and chlorine to generate the potent oxidant hypochlorous acid [16]. In addition, MPO-dependent oxidation of tyrosine generates tyrosine radicals and a further interaction with nitric oxide leads to the oxidant nitrotyrosine [17].

It is possible that MPO generated oxidants play an important role in the vascular injury seen in AASV. Furthermore, enzymatically inactivated MPO can act as an immunoregulatory mediator by virtue of its capacity to induce the secretion of chemokines and cytokines by endothelial cells and hence may be an important factor for the induction of persistent inflammation [18]. In the microenvironment of the inflammatory response, approximately 40% of the MPO is inactivated rapidly by oxidation [12].

An Alu element preceding the MPO gene consists of four hexamer half-sites, related to the consensus recognition sequence for a nuclear hormone receptor binding site, AGGTCA, orientated as direct repeats spaced by 2, 4 and 2 bp (DR-2-4-2). Gel shift experiments and transfection assays demonstrate that these sequences include binding sites for retinoic acid and thyroid hormone receptors [5]. The first DR-2 elements of the series do not bind known receptors but do bind the SP1 transcription factor. The two alleles of the MPO gene differ at one position within this element, this first DR-2/G element increases transcription of a chloramphenicol acetyltransferase reporter gene by 25-fold in cellular transfection assays, whereas the corresponding DR-2 element from the 463A allele is a less effective transactivator [5]. This results in one allele with (463G) and one without (463A) a strong SP1 binding site. The 463G allele is within the SP1 binding site, while the 463A allele abolishes the site, creating a receptor for oestrogen instead [5]. The binding site for the SP1 transcription factor at 463G/A in the G allele is a plausible factor for the observed higher expression of MPO [5].

Our study examined the hypothesis that if MPO was involved in the pathogenesis of vascular injury then individuals with AASV who had the MPO 463G allele, which was associated with higher MPO levels, would have more severe disease than patients with the A allele. We show that the allelic expression directly reflects the level of expression of neutrophil MPO. The GG genotype was associated with 1·75-fold higher MPO expression than the GA genotype and 5·4-fold higher expression than the AA genotype, which is in keeping with an earlier study [19]. The higher expressing GG genotype may be associated with higher levels of MPO on the surface of primed monocytes and neutrophils, available for binding by circulating MPO–ANCA leading to monocyte and neutrophil activation. Further, the release of increased amounts of MPO might mediate endothelial injury.

The present study demonstrates that the MPO G-463/A promoter polymorphism does not increase disease severity as judged by the degree of renal impairment. Further, this polymorphism did not influence the rate of progression to renal failure or death. The polymorphism did not represent a genetic risk factor for the development of AASV. No association between the MPO 463G/A polymorphism was noted between controls and patients divided by disease category or ANCA specificity. Furthermore, no association was observed between the MPO G-463/A polymorphism and vasculitis in patients based on gender difference, given the evidence that females have up to twofold higher MPO activity levels than males and that this level fluctuates with serum oestradiol concentrations [20].

In a study of 142 patients with AASV and 129 ethnically matched controls, Reynolds et al. [6] reported an increased incidence of the GG genotype in females with MPO–ANCA vasculitis (86% GG; P = 0·045; OR 3·54; 95% CI: 1·10–11·49). No association was noted between the MPO 463G/A and disease risk in males (64% GG, P = 1·00, OR 0·91; 95% CI, 0·31–2·50). In patients with MPO–AASV there was an association between the GA/AA genotype with an earlier age of onset (P = 0·03) and an increased risk of disease relapse (χ² = 6·32, P = 0·012; relative risk 3·21; 95% CI: 1·43–18·2) in comparison with the GG genotype. However, the study of Fiebeler et al. [7] did not show any association between the MPO G-463A polymorphism and AASV.

This meta-analysis shows that the MPO G/A polymorphism is not a risk factor for susceptibility for AASV. Overall, the GG genotype or the combination of GG and GA genotypes did not confer susceptibility to disease and the GG genotype did not increase susceptibility in MPO–ANCA or PR3–ANCA associated vasculitis.

In conclusion, our study demonstrates that there is no association between the MPO G-463/A polymorphism and susceptibility or severity of AASV. However, we have also shown that the MPO 463G/A polymorphism impacts directly upon neutrophil MPO activity. This suggests that the MPO levels in leucocytes and the MPO G/A polymorphism does not play an important role in the pathophysiology of AASV. The possibility of a type II error cannot be excluded in this study. In order to detect a 20% difference with 80% power, 200 patients and 400 controls would be needed. However our meta-analysis goes some way to addressing this point, as with a reasonably large number of patients and controls no association between the MPO G/A polymorphism and susceptibility to AASV was seen.