Among the possible etiologic and triggering factors involved in Sjögren's syndrome, the discussion about a relationship between viral infections causing development of autoimmune reactions began some decades ago. The putative role of different viruses in Sjögren's syndrome can be viewed in the light that salivary glands are a site of latent viral infections. Potential viral triggers include a number of viruses including Epstein–Barr virus (EBV), widely studied in relation to Sjögren's syndrome (James, Harley and Scofield, 2001).
Hepatitis C virus (HCV) infection has in some populations been frequently detected in patients with primary Sjögren's syndrome (Garsia-Carrasco et al, 1997). Analysis of the association between chronic lymphocytic sialadenitis and chronic HCV liver disease showed that histological features of Sjögren's syndrome were significantly more common in HCV-infected patients (57%) compared with controls (5%) (Haddad et al, 1992).
Lymphotropic viruses have the potential to trigger the autoimmune process. Some of the immunoreactive regions within the La/SS-B protein have been found to have sequence similarities with proteins of EBV, HHV-6 and HIV-I (Haaheim et al, 1996). It seems reasonable that these viruses can promote autoantibody (particularly anti-La/SS-B) production through molecular mimicry or exposure of La/SS-B homologue sequences on cellular surfaces after translocation of cryptic self-determinants.
Furthermore, a possible relationship between Sjögren's syndrome and Helicobacter pylori infection has been suspected. In both these conditions there are increased risks of developing mucosa-associated lymphoid tissue lymphoma (Isaacson and Spencer 1987; Parsonnet et al, 1994). It has been suggested that infection with H. pylori might trigger a widespread clonal B-cell disorder in Sjögren's syndrome (De Vita et al, 1996). Studies on antibodies against H. pylori in Sjögren's syndrome though have given conflicting results as to whether seroreactivity is elevated or not (Showji et al, 1996; Aragona et al, 1999; Theander et al, 2001). However, improvement of sicca symptoms has been indicated after eradication treatment of H. pylori (Figura et al, 1994)
Immunopathology and autoimmunity
Immunohistologic analysis of lymphoid cell infiltration in exocrine glands in Sjögren's syndrome shows a predominance of T cells with fewer B cells and macrophages (Jonsson et al, 2001). Adhesion molecules and activated lymphocyte function-associated antigen type 1 (LFA-1) promote homing and occasionally characteristic cell clustering similar to follicular structures of lymph nodes. Expression of the mucosal lymphocyte integrin αEβ7 and its ligand E-cadherin suggest a mucosal origin of a subpopulation of the infiltrating cells (Kroneld et al, 1998). There is an increased expression of HLA-DR/DP/DQ molecules on acinar and ductal epithelial cells (Jonsson et al, 1987) presumably because of local production of IFN-γ by activated T cells. The majority of T cells in the lympocytic infiltrates are CD4+ T-helper cells with a CD4/CD8 ratio well over two. Most of these T cells bear the memory phenotype CD45RO+ and express the α/β T cell receptor and LFA-1, and may contribute significantly to B cell hyperactivity. There is indication of oligoclonal expansion of certain TCR Vβ family expressing lymphocytes (Sumida et al, 1992). The findings in peripheral blood in Sjögren's syndrome have yielded findings similar to those in salivary glands, although a difference in magnitude of immune activation is often evident.
The B cells make up roughly 20% of the infiltrating cell population in affected glands. The B cells produce immunoglobulins with autoantibody activity for IgG (rheumatoid factor), Ro/SSA and La/SSB (Halse et al, 1999a). A substantial number of the B cells are of CD5+ phenotype (B-1 cells) (Dauphinee, Tovar and Talal, 1988). Production of IgG predominates in Sjögren's syndrome whereas synthesis of IgA is more abundant in normal salivary glands.
A large number of autoantibodies have been reported in both primary and secondary Sjögren's syndrome, reflecting both B cell activation and a loss of immune tolerance in the B cell compartment (MacSween, Govidie and Anderson, 1967; Feltkamp and Van Rossum, 1968; Manthorpe, Permin and Tage-Jensen, 1979; Ben-Chetrit et al, 1988; Atkinson et al, 1989; Inagak et al, 1991; Hauschild et al, 1993; Markusse, Otten and Vroom, 1993; Kausman et al, 1994; Tzioufas and Moutsopoulus, 1994; Boire et al, 1995; Bacman et al, 1996, 1998; Kino-Ohsaki et al, 1996; Haneji et al, 1997; Font et al, 1998; Elagib et al, 1999; Freist et al, 1999; Ono et al, 1999; Watanabe et al, 1999). In some cases, the presence of these antibodies is related to the extent and severity of disease in Sjögren's syndrome. The non-organ-specific autoantibodies anti-Ro/SSA and anti-La/SSB are the diagnostically most important and the best characterized autoantibodies in primary Sjögren's syndrome (Jonsson et al, 2001). The majority of anti-Ro-positive sera also react with the denatured form of a 52-kDa protein termed Ro52, which is structurally distinct from Ro60 and probably does not directly associate with the Ro ribonucleoprotein particle (Ben-Chetrit et al, 1988; Boire et al, 1995). However, the two Ro proteins colocalize to surface membrane blebs on apoptotic cells where they may become targets of an autoimmune response (Ohlsson et al, 2002).
Anti-thyroid microsomal and antigastric parietal cell antibodies occur in about one-third of patients with both primary and secondary Sjögren's syndrome but other organ-specific antibodies are infrequent (Morrow et al, 1999). Antibodies to salivary duct antigens were described over 30 years ago but they have remained poorly characterized and their clinical significance is uncertain (MacSween et al, 1967; Feltkamp and Van Rossum, 1968). However, a recent study has demonstrated that this reaction is because of cross-reactivity with blood group antigens (Goldblatt et al, 2000).
Several other autoantibodies have been reported to be frequently present in the sera of patients with primary Sjögren's syndrome including antibodies directed against carbonic anhydrase (Inagak et al, 1991; Kino-Ohsaki et al, 1996), proteasomal subunits (Freist et al, 1999) and α-fodrin (Haneji et al, 1997). These findings are intriguing but await independent confirmation in larger cohorts of Sjögren's syndrome patients. The finding of serum autoantibodies directed against the muscarinic M3 receptor (expressed in salivary and lacrimal glands) in the majority of patients is an important advance in understanding the pathogenesis of impaired glandular function in Sjögren's syndrome (Bacman et al, 1996, 1998). This is of high interest as a recent study has shown that M3-muscarinic receptors are up-regulated in glandular acini (Beroukas et al, 2002).
Studies in the non-obese diabetic (NOD) mouse have indicated that muscarinic receptor autoantibodies are directed against the agonist-binding site of the molecule on the cell surface and interfere with secretory function of exocrine tissues in Sjögren's syndrome (Robinson et al, 1998a). Inhibitory effects of these autoantibodies on parasympathetic neurotransmission in Sjögren's syndrome have recently been experimentally shown (Waterman, Gordon and Rischmueller, 2000). However, the clinical significance of these antibodies in Sjögren's syndrome remains to be elucidated (Humphreys-Beher et al, 1999).
Rheumatoid factor is detected in the serum and saliva of 60–80% of primary Sjögren's syndrome patients (Atkinson et al, ; Markusseet al, 1993). There appears to be little role for somatic hypermutation in their generation in contrast to rheumatoid factor in rheumatoid arthritis (Elagib et al, 1999). A significant number of patients with primary Sjögren's syndrome have mixed oligoclonal cryoglobulins, many of them having IgM rheumatoid factor activity (Tzioufas et al, 1986). The latter frequently possess cross-reactive idiotypes notably the 17.109 (V kappa III b related) and G-6 (VH1 related) idiotypes that may serve as markers for lymphoma development in primary Sjögren's syndrome (Fox et al, 1986a; Tzioufas et al, 1996).
Oligoclonal or monoclonal B cell expansion, arising mainly from salivary glands but also from visceral organs and lymph nodes, has been reported to occur in 14–100% of Sjögren's syndrome patients (Anaya et al, 1996). In this respect, Sjögren's syndrome appears to be a crossroad between autoimmunity and malignancy and it is suggested that patients with evidence of clonal expansions of B cells in their salivary glands are at high risk of developing malignant lymphoma (Hyjek, Smith and Isaacson, 1988; Diss et al, 1995; Jordan et al, 1995). Various studies have reported that between 25 and 80% of salivary lymphoid infiltrates have morphologic and/or immunophenotypic evidence of low-grade lymphomas (Harris, 1999). However, there is no absolute correlation between clonality and the development of lymphoma. Although a high proportion of lymphoid cells may show evidence of immunoglobulin gene rearrangements, clonality does not necessarily predict progression to clinically overt lymphoma. The clinical benefit of immunogenotypic analysis in the diagnosis of salivary gland lymphoma in Sjögren's syndrome remains to be defined (His et al, 1996; Quintana, Kapadia and Bahler, 1997). A recent study reported that a history of swollen salivary glands, lymphadenopathy and leg ulcers predicted lymphoma development in patients with primary Sjögren's syndrome (Sutcliffe et al, 1998).
Immune-mediated tissue destruction
Highly up-regulated expression of HLA molecules, and the more recently demonstrated B-7 co-stimulatory molecules (Manoussakis et al, 1999), by salivary gland epithelium in Sjögren's syndrome is a potentially effective local antigen-presenting mechanism whereby HLA antigens could be involved in exocrine glandular destruction mediated directly or indirectly by CD4+ T cells. Such interaction may lead to further production of cytokines and stimulation of B cell proliferation and differentiation. Indeed, high levels of three tissue destructive cytokines, interleukin (IL)-1βT, IL-6 and tumour necrosis factor-α (TNF-α), are produced by epithelial cells. Infiltrating T cells mainly produce IL-10 and IFN-γ, while IL-6 and IL-10 are also elevated in peripheral blood (Halse et al, 1999b). A low capacity to produce IL-2 in Sjögren's syndrome might be because of absence of the T cell costimulatory signals resulting in the induction of anergy in the responding T cell population, but other explanations are also possible.
Recently conducted studies on chemokine patterns have pointed further to the role of epithelial cells in the pathogenesis of Sjögren's syndrome (Amft and Bowman 2001; Xanthou et al, 2001; Salomonsson et al, 2002). This offers new insight into the mechanisms of leukocyte attraction and formation of secondary lymphoid tissue structures.
Even though the mechanism(s) behind the characteristic destruction of salivary glands in Sjögren's syndrome remain obscure, immunopathological findings demonstrate that infiltrating cytotoxic T cells (CTL) could play a role in this event. Upon recognition of a proper MHC–antigen complex presented by a target cell, CTLs induce cell death through one of its two main and independent pathways, the perforin-mediated or the Fas-mediated pathway. Interestingly, expression of Fas has also been detected among infiltrating mononuclear cells in salivary glands of MRL/lpr mice, a murine model displaying similar features as human systemic lupus erythematosus and Sjögren's syndrome (Skarstein et al, 1997).
Expression of Fas, Fas-L, Bcl-2 and other apoptosis associated genes/proteins has been detected by RT-PCR and immunohistochemical staining of minor salivary glands in patients with Sjögren's syndrome (Kong et al, 1997; Nakamura et al, 1998). In particular, ductal and acinar epithelial cells but to some degree also infiltrating mononuclear cells express abnormal levels of Fas and FasL, especially in cases with heavy mononuclear cell infiltration. Ductal epithelial cells expressing Fas were usually situated inside or adjacent to a dense focus (Ohlsson et al, 2002). Most in situ studies have clearly shown a low grade or even absence of apoptosis among infiltrating mononuclear cells (Kong et al, 1997; Nakamura et al, 1998). The presence of granzyme A in Sjögren's glands (Alpert et al, 1994) suggests that rather than apoptosis the perforin pathway of CTL mediated killing may be involved in destruction of salivary glands.
Among the salivary gland infiltrating T cells some express activation markers such as CD25, proto-oncogene products and HLA-DR, but few T cells proliferate as determined by cell cycle studies. Also, it seems difficult to stimulate the T lymphocytes in Sjögren's syndrome with the autoantigens Ro/SSA and La/SSB (Halse et al, 1996) although a recent study has indicated no T cell responses, at least to La/SSB (Davies et al, 2002). These findings suggest that many cells are of memory T cell phenotype; either few of them are autoantigen specific or alternatively many of them are in a state of anergy. In both cases lack of stimulation of T cells will also hamper the apoptotic signals.
As already alluded to there are genetic associations which may predispose individuals to Sjögren's syndrome, in particular the genes encoding products of the MHC and immune receptors, but also other genes. It is thus natural to seek more knowledge using genetically well-characterized, inbred animal models that are available for study (Jonsson and Skarstein 2001). In particular, the current challenge is to find links between a particular genetic background and phenotypic expression(s) of this disease.
Any proposed animal model should fulfill certain criteria and features found in the human disease. Moreover, the clinical symptoms of Sjögren's syndrome in humans usually appear relatively late in life thus making examination of early events difficult. An animal model of the disease would make it possible to study earlier events and to identify potentially important immune reactions in the pathogenesis of this disease. Finally, both immune manipulation and the effects of drug therapy can be studied in animals (Jonsson and Skarstein 2001).
The earlier attempts to induce Sjögren's syndrome in animals by injection with salivary gland extracts with or without adjuvants and/or other supplements would give rise to a transient inflammation which was self-limiting and did not mirror the human disease in either the temporal course of events or in the serological profile. The better models of Sjögren's syndrome are the mice with spontaneous autoimmune disease with long-lasting and progressive exocrinopathy, but even in these cases the disorder has at best represented only secondary Sjögren's syndrome (Jonsson and Skarstein 2001). Both anti-Ro (Wahren et al, 1994) and anti-La (St. Clair et al, 1991) have been detected in murine models of spontaneous Sjögren's syndrome. As these autoantibodies are the dominant serological marker in patients with primary Sjögren's syndrome, this finding is an important starting point for future work.
NOD.B10.H2b mice have been found to exhibit exocrine gland lymphocytic infiltration typical of Sjögren's syndrome-like disease and dysfunction observed in NOD mice, but without the insulitis and diabetes (Robinson et al, 1998b). These findings indicate that murine sicca syndrome occurs independently of autoimmune diabetes and that the congenic NOD.B10.H2b mouse represents a novel murine model of primary Sjögren's syndrome.
More recently, gene segregation experiments on a (NOD.QxB10.Q) F2 cross and genetic mapping have revealed one locus associated with sialadenitis on chromosome 4 (LOD score 4.7) (Johansson et al, 2002). In this study it was shown that the genetic control of sialadenitis seemed to be unique in comparison with diabetes and also arthritis, as no loci associated with these diseases have been identified at the same location.