Relationship between serum levels of IL-18 and IgG1 in patients with primary Sjögren's syndrome, rheumatoid arthritis and healthy controls

Abstract

Primary Sjögren's syndrome (SS) is characterized by inflammation in salivary and lachrymal glands, with a local predominance of Th1-like cytokines, as well as the pleiotropic cytokine interleukin (IL) 18. High serum levels of polyclonal IgG are common, with a subclass imbalance in which IgG1 is increased and IgG2 is normal or low. IL-18 is also of pathogenetic importance in rheumatoid arthritis. In the present study we looked for any relationship between serum IL-18 as well as transforming growth factor (TGF) β1 versus IgA, IgM, and IgG subclass levels in SS (n = 16), rheumatoid arthritis (RA) (n = 15), and healthy controls (n = 15). SS was defined by the revised American-European classification criteria. IL-18 and TGF-β1 were analyzed with enzyme immunoassays (EIA), and IgG1, IgG2 and IgG3 by single radial immunodiffusion. In the composite group of RA, SS and normal controls, IgG1 and IL-18 were related (R = 0·52, P = 0·0005). No relation was found neither between IL-18 versus IgG2, IgG3 or IgA, nor between serum TGF-β1 versus any of the immunoglobulins. Since serum levels of IL-18 are related to serum IgG1, IL-18 may be of importance for IgG1 switch and/or release.

INTRODUCTION

Owing to its ability to stimulate the production of interferon gamma (IFN-γ) in synergy with interleukin (IL-)12, IL-18 was originally designated ‘IFN-γ-inducing factor’ and therefore regarded as an inducer of T-helper cell (Th) type1 immune responses [1]. By this mechanism, reduced IL-4-mediated production of murine IgE as well as IgM and IgG1 has been demonstrated [2,3]. The designation of the four human IgG subclasses (IgG1-IgG4) must not be confused with murine IgG subclasses (IgG1, IgG2a, IgG2b and IgG3), neither regarding biochemical or biological properties (such as complement fixation, Fc-receptor binding, and opsonization), nor regarding their dependency of Th type1 and type2-like cytokine responses. In a murine model of atherosclerosis, IL-18 deficient mice produced increased amounts of IgG1 antibodies to low-density-lipoproteins, increasing the IgG1/IgG2a antibody ratio by 50%, which implies that loss of IL-18 signalling can cause a type1 → type2 shift in mice [4]. However, in addition to its type1-driving property, IL-18 can act as a type2-promoting cytokine (with and without synergy with other cytokines, e.g. IL-2) inducing the production of IgE [5,6]. In synergy with IL-1α and IL-12, IL-18 has also been shown to induce mucosal IgA production [7]. Transforming growth factor beta-1 (TGF-β1) is generally regarded as an anti-inflammatory cytokine and is associated with the induction of Th type3-mediated immune regulation, including mucosal production of secretory IgA antibodies [8].

Primary Sjögren's syndrome (SS) is characterized by inflammation in the salivary and lachrymal glands. In salivary gland biopsies, type1 cytokines, including IL-18, predominate [9,10]. Reduced salivary gland expression of TGF-β1 has been reported in SS [11,12], and in TGF-β1 deficient mice, an SS-like lymphoproliferative disease develops [13]. Several studies have revealed raised serum levels of IL-6 and IL-10 in SS [14–18]. In rheumatoid arthritis (RA), a type1 inflammatory response including IL-18 is considered to be of pathogenetic importance [19,20].

High serum levels of polyclonal IgG are common in SS, and subclass analysis usually shows an imbalance with selective increase of IgG1, whereas IgG2 levels are normal or decreased [21]. Total serum IgG1 parallels the levels of antibodies of IgG1 subclass to Ro/SS-A [22] and La/SS-B [23]. These two autoantibodies are typical of SS [24]. Raised serum levels of IgG1 in SS have been associated with raised IL-10 levels [18]. In a recent study on primary Sjögren's syndrome, we found increased levels of circulating IL-18 compared to patients with rheumatoid arthritis and healthy subjects [25]. The present study was done as an extension to investigate possible relations between the serum levels of IL-18 and TGF-β1, respectively, versus the total serum levels of IgG-subclasses, IgM, and IgA in a group of patients and normal individuals with a wide range of IgG subclass and IgA levels.

METHODS

The SS and RA patients were recruited at random after informed consent at the rheumatology unit, Linköping University Hospital. All patients and healthy controls were women. The median age was 62 years (range 37–74) in the 16 SS patients, 58 years (range 35–74) in the 15 RA patients, and 47 years (range 41–69) in the 14 healthy controls. The healthy controls were significantly younger than the RA and SS patients tested by Mann–Whitney's U-test (P = 0·04 for both), whereas the age difference between the two patient groups was not significant (P = 0·72). No patient or healthy control had any symptoms hinting infectious disease at the blood sampling occasion. In the SS group, 3 patients used hydroxychloroquine (200 mg/day), 2 used prednisolone (≤ 7·5 mg/day), and 2 used nonsteroidal anti-inflammatory drugs (NSAIDs) or coxibs. The medications in the RA group were: 8 oral methotrexate (7·5–15 mg once weekly), 8 prednisolone (≤ 10 mg/day), 2 sulfasalazine (2 g daily), 2 intramuscular gold sodium thiomalate, 1 leflunomide, and 9 NSAID/coxibs. Intra-articular corticosteroids were not allowed within 1 month prior to the study.

All SS patients fulfilled the revised version of the European criteria proposed by the American-European Consensus Group [26]. The presence of antibodies to SS-A was not a requirement for inclusion, but all 16 SS patients had precipitating anti-SS-A antibodies (Immunoconcepts, Sacramento, CA, USA), 11 (69%) had anti-SS-B antibodies, and 13 of 16 had antinuclear antibodies (ANA) detectable by immunofluorescence microscopy (HEp-2 cells, Immunoconcepts) at a serum dilution of at least 1 : 100. All RA patients met the requirements of the 1987 ACR classification criteria [27], and all except two (87%) were seropositive for agglutinating rheumatoid factor (RF). The disease activity in RA was estimated by a disease activity score (DAS-28) based upon a 28-joint count of swollen and tender joints, patient's global assessment of general health and erythrocyte sedimentation rate [28], with a mean DAS-28 of 4·4 (SD 1·3).

Serum samples were stored at −70°C until analysed. Enzyme-immuno assay (EIA) was used to analyse IL-18 (MBL, Nagoya, Japan) and TGF-β1 (BioSource, Europe S.A., Nivelles, Belgium). The analyses were performed in duplicates and according to the manufacturers instructions. IgG subclass analyses were performed by single radial immunodiffusion in 1·4% agarose containing mouse monoclonal antibodies against the corresponding IgG subclass (Binding Site Ltd, Birmingham, UK). Analyses of RF, IgA, IgM, and C-reactive protein (CRP) were done by nephelometry.

Statistics

Mann–Whitney's U-test and Spearman rank correlation test were used for statistical analysis.

Ethics

The study protocol was approved by the local ethics committee.

RESULTS

The median serum values of IgG1, IgG2, IgG3, IgM, IgA, IL-18 and TGF-β1 are given in Table 1. The patients with SS had increased levels of IgG1 (P < 0·0001), IgA (P < 0·05), and IL-18 (P < 0·01), whereas the levels of TGF-β1 were decreased (P < 0·05) compared with healthy controls. RA patients had increased levels of IgG3 (P < 0·05) and increased levels of IL-18 (P < 0·05) compared with controls.

Table 1.

Serum levels (and range) of IgG1, IgG2, IgG3, IgA, IgM, IL-18, and TGF-β1 in the groups of primary Sjögren's syndrome, RA, and healthy controls. Median values and range are given

	IgG1 (g/l)	IgG2 (g/l)	IgG3 (g/l)	IgA (g/l)	IgM (g/l)	IL-18 (pg/ml)	TGF-β1 (ng/ml)
RA	6·8(3·3–15·8)	3·2(0·9–6·0)	0·80*(0·2–1·8)	2·6(0·97–4·6)	1·3(0·63–3·0)	230*(150–660)	65(39·5–83·2)
SS	16·1***†(4·7–44·3)	3·0(1·1–6·2)	0·75(0·2–4·5)	3·4*(0·0–8·4)	1·55(0·7–2·4)	312**(145–755)	52*(27–72·8)
Controls	6·0(3·8–8·6)	3·3(2·3–5·8)	0·55(0·2–0·9)	2·1(1·1–3·6)	1·2(0·54–1·9)	150(100–405)	65(46·8–109·2)

^*P < 0·05

^**P < 0·01

^***P < 0·001 compared with healthy controls;

^†P < 0·001 compared with RA.

In the composite group of RA, SS and healthy controls, serum IL-18 and serum IgG1 were highly correlated (Rho = 0·52, P = 0·0005) (Fig. 1). When the SS patients were analysed separately, R was 0·504 (P = 0·06). In the composite group we also found a significant correlation between the serum levels of IgM and IL-18 (Rho = 0·36, p = 0·015), but no correlation between the IgM and TGF-β levels (P = 0·23). No relation was found between serum IL-18 versus IgG2, IgG3 or IgA (R−0·213, R 0·117, R 0·160, respectively), nor between serum TGF-β1 versus IgG1, IgG2, IgG3 or IgA (R−0·107, R−0·019, R−0·071, R 0·04, respectively).

Fig. 1.

In the composite group of RA (•), SS (○) and healthy controls (+), serum IL-18 and serum IgG1 were correlated (R 0·524, P = 0·0005).

All healthy controls, and all except one SS patient, had CRP values below the detection level (<10 mg/l). In the RA group the median CRP was 26 mg/l (range: 10–68).

DISCUSSION

In a recent study comparing patients with SS, RA and healthy individuals, we found the highest serum IL-18 concentrations in SS, intermediate in RA, and lowest in healthy controls [25], and speculated that raised serum levels of IL-18 may reflect ongoing inflammation in the target organs. In the present study we analysed the serum levels of IL-18 and TGF-β1 in relation to total levels of the IgG subclasses and IgA in the composite group of patients with SS, RA and healthy controls. We found a highly significant correlation between IL-18 and IgG1 levels, but not between IL-18 and IgG2, IgG3 or IgA, respectively. IL-18 levels were also correlated to the serum levels of IgM. No significant correlations were found between serum levels of TGF-β1 and IgG subclasses, IgM or IgA. In contrast to Perrier et al. [18] we found decreased serum levels of TGF-β1 in SS both compared to healthy controls and to RA patients, but with no correlation to the levels of IgG subclasses or IgA. The positive correlation between IL-18 and IgG1 indicates that IL-18 may be important for the triggering of IgG1 production. It has previously been demonstrated that autoantibodies to SS-A [22] and SS-B [23] in Sjögren's syndrome are primarily of IgG1 subclass. Since IgG1 antibodies can promote inflammation, e.g. by complement activation and Fcγ-receptor-mediated phagocyte activation [29], and since local production of IL-18 [9,10] as well as autoantibodies [30] has been demonstrated in inflamed salivary glands, it is conceivable that IL-18-induced IgG1 autoantibody production is of aetio-pathogenetic importance in Sjögren's syndrome. Similarly, it is possible that the subnormal levels of TGF-β may reflect an insufficient anti-inflammatory control in this disease.

The mechanisms of human Ig class/subclass switches and release are not fully understood. In-vitro studies in mice show a relatively consistent pattern, even allowing IgG subclass patterns to be used as markers of type1/type2 cytokine patterns. IL-18, as well as TGF-β1, participates in type1/type2 regulation and both may contribute to immunoglobulin class/subclass switching in mice [5,31]. In contrast, in vitro studies in humans show inconsistent results concerning the roles of different cytokines for Ig production. The production of IgG1 may involve IL-2, IL-4 or IL-10 [29,31–31]. Induction of IgG2 has been suggested to involve IFN-γ and IL-12 [29,32,33], whereas IgG3 production possibly engages IL-2, IL-4 and IL-10 [29,31,32,34]. The production and/or release of IgG4 seems to be directed by IL-4 and IL-13 [29,31,33,35,36]. Different costimulating agents and B-cell lines were used in these in vitro studies, and one possible explanation for the contrasting results is differences in B cell activation stage in the used cell lines, which might affect the ability to up-regulate germline transcripts in response to CD40 triggering [37]. Studies of circulating antibody levels in humans have shown predominance of IgG1 and IgG3 in type1-associated diseases like Lyme Borreliosis [38], Plasmodium falciparum infection [39], and multiple sclerosis [40]. Considering the dual effects of IL-18 [5,6,41], however, it does not appear appropriate to discuss whether the IL-18-associated elevation of circulating IgG1, found in the present study, is the result of a type 1- or a type 2-dominated immune response. The same accounts for the indications of IL-10-mediated IgG1 production/release in vitro [34]and in vivo in Sjögren's syndrome [18].

Although subnormal IgG2 levels are common in SS [21], this was not the case in the present study, and contrary to expected, no correlation could be demonstrated between serum TGF-β1 and IgG2, IgA or any other immunoglobulin. This lack of correlation may be due to the limited study population. Another explanation could be that IgA is produced mainly in mucosal tissues and therefore cytokine levels in serum may not reflect the local mucosal cytokine regulation.

In summary, this study revealed a positive correlation between serum levels of IL-18 and IgG1 in a composite study population including healthy persons, patients with rheumatoid arthritis, and patients with primary Sjögren's syndrome. The highest levels were noted in primary Sjögren's syndrome, indicating that IL-18 may be involved in the triggering of IgG1 production and hypergammaglobulinaemia in primary Sjögren's syndrome. Further studies are needed to shed light on this hypothesis.