Sjögren’s syndrome, should we sign?

Type I interferons (IFNs) play a significant role in the pathogenesis of primary Sjögren’s syndrome (pSS). Derailed IFN production may contribute to overactivation of the innate and adaptive immune system, finally leading to B-cell hyperactivity, autoimmunity, and dysfunctioning of the glandular tissue. In pSS patients many IFN-stimulated genes and proteins are overexpressed, resulting in a so-called IFN signature. Although initially mainly type I inducible genes were believed to be overexpressed in pSS patients, more recent evidence indicates that also type II IFN genes can be overrepresented. Not all pSS patients seem to exhibit an IFN signature and the key question that remains to be answered is whether the presence of a type I and/or type II IFN signature reflects merely a more severe variant or more active phase of the disease or that it identifies a distinct subgroup of patients with a different etiopathogenesis.

Type I IFNs have been implicated in the pathogenesis of several systemic autoimmune diseases, including pSS.[1] pSS is characterized by a chronic inflammatory process of the salivary glands and lacrimal glands, leading to the dominant complaints of dry mouth and dry eyes. The disease is, however, certainly not restricted to exocrine glands as reflected by complaints of fatigue and pain and a wide range of extraglandular manifestations, explaining the diverse clinical presentation of pSS patients.

Type I IFN comprises a large group of subtypes of IFNs of which IFNα and IFNβ are the most studied proteins. These subtypes share similar functions. Type I IFNs are prevalently induced by the binding of (viral) DNA/RNA to transmembrane pattern recognition receptors (Toll-like receptors (TLRs) in particular TLR3, TLR7, and TLR9) or to cytosolic sensors (e.g. RIG-1, MDA-5). Most body cells, including epithelial cells of exocrine glands, can produce IFNβ, whereas plasmacytoid dendritic cells (pDCs) are specialized to produce large amounts of IFNα, but also produce other subtypes; pSS patients express higher serum levels of type I IFN (IFNα) compared to healthy controls.[2] Other studies, however, detected serum IFNα only in a small proportion of pSS patients.[3] IFNα producing pDCs are found in the glandular infiltrates from pSS patients.[3,4] Not only microbial-derived nucleic acids, but also autologous nucleic acids and immune complexes composed of autoantibodies directed to RNA binding proteins are strong stimulants of IFNα production by pDCs in pSS patients.[3] These autoantigens (mostly SSA/SSB) are possibly derived from the intrinsically activated epithelial cells.[5] In turn, IFNs can upregulate the expression of these autoantigens by the epithelial cells. Furthermore, B-cells from patients with pSS can stimulate directly the secretion of IFNα by pDCs.[6] IFN production is regulated by various IFN regulatory factors (IRFs) and type I IFN production by pDCs is dependent on the transcription factors IRF5 and IRF7.[7] The importance of IRF5 in pSS is illustrated by genome-wide association studies that identified a polymorphism in IRF5 associated with pSS.[8] Another pSS-associated gene polymorphism is stat4, a transcription factor involved in the signaling of type I IFN, after binding to its receptor the type I IFN receptor (IFNAR). This receptor is widely expressed by many different cell types and binds all isoforms of type I IFNs.[9] IFNAR signaling results in the transcription of several genes that renders the cells with a kind of antiviral state, in order to protect the body from viral infections. Many other genes that are stimulated by type I IFN modulate both innate and adaptive immune responses, for example by increased antigen presentation, cytokine and chemokine production, and stimulation of effector T- and B-cell responses. Indeed, most evidence for a crucial role of (type I) IFN in pSS comes from studies analyzing mRNA transcripts and protein expression levels of genes stimulated by IFN. Measuring the expression of IFN-inducible genes also overcomes the problem that there are several type I IFN subtypes, as a result of which assessment of a single subtype I IFN does not provide clues about the overall type I IFN activity.[10] Initial microarray analysis of mRNA extracted from minor salivary gland tissue of pSS patients revealed an upregulated expression of many IFN-stimulated genes, the so-called IFN signature, compared to non-pSS sicca patients.[4,11] The finding that the increased expression of many (type I) IFN-stimulated genes can also been observed in circulating monocytes, peripheral blood mononuclear cells, and whole blood from pSS patients [12–14] indicates that IFN activity is clearly not only restricted to the target tissue itself, but is a more systemic feature of the disease. Possibly, continued activation of type I IFN production by triggers such as virus, autoantigen, and immune complexes in combination with a special genetic background and lack of proper control and regulation may be responsible for the observed overproduction of type I IFN in pSS patients.[15]

Although the initial evidence accredited the importance of type I IFNs in pSS, more recent data indicate that not only type I IFN genes, but also type II IFN genes are involved. Hall et al. [16] observed that the majority of genes induced by IFNα (type I IFN) are also induced by IFNγ (type II IFN) and that only few genes are exclusively induced by IFNα. Taking this into account, it became apparent that both type I and type II IFN-stimulated genes can be upregulated in peripheral blood and salivary gland tissue of pSS patients.[16–18] Based on a small set of mRNA transcripts that are preferentially induced by either type I IFN or type II IFN, it was concluded that in the peripheral blood of pSS patients type I IFN-stimulated genes appear to dominate, whereas the opposite is the case in the glandular tissue where type II IFN-stimulated genes are the most pronounced.[17,18] Various approaches and definitions are being used to calculate an IFN score and to assign a type I or type II IFN signature to pSS patients. Despite these discrepancies is it clear that not all pSS patients upregulate type I and II IFN genes to the same extent and three groups of IFN profiles have been identified in pSS patients: type I IFN dominant, type II IFN dominant, or a mixed-type I/II IFN phenotype. Approximately 60% of the pSS patient have a high type I IFN score (signature) in blood based upon the expression of a few (2–5) type I IFN-stimulated genes or whole blood myxovirus-resistance protein A levels.[10,14,18] A significant number (slightly less than half of these patients) concomitantly also have a type II IFN signature; only relatively few pSS patients (7%) exhibit only a type II IFN signature.[18] Based upon the expression of a single IFN-induced protein, Hall et al. [17] observed that 17% of the patients demonstrated high type I IFN activity in the minor salivary glands, 21% high type II IFN activity, and 21% exhibited a mixed-type I/II IFN profile. Thus, there is significant heterogeneity in type I and type II IFN activities between blood versus salivary glands and amongst different individual patients.

Derailed type I IFN production, as seen in pSS, may contribute directly to the hyperactivity and autoreactivity of B cells characteristic for this disease.[19,20] Chronic expression of type I IFN also may also influence the function of glandular tissue, for example by promoting apoptosis of the epithelial cells or loss of integrity of the epithelium and consequently xerostomia. Irrespective of the fact whether the pSS patients exhibit high type I or type II IFN activity (i.e. high IFN score) and how and in which tissue this activity has been determined, the emerging picture is that pSS patients with such a signature have a more severe phenotype reflected not only serologically for example by higher serum IgG levels, more (and higher titers of) autoantibodies (rheumatoid factor, ANA, anti-SSA, and anti-SSB antibodies), and lower C3 levels, but also by lower salivary and tear flow rates, and higher disease activity as measured by the ESSDAI scores.[10,16–18] The key question therefore is whether the (operationally defined) type I and/or type II signature reflects merely a more severe variant or more active phase of the disease or identifies a distinct subgroup of patients with a different etiopathogenesis. The absence of an IFN signature in pSS patients does not necessarily imply that IFN (type I/II) does not play a role. Some patients may just produce lower amounts of IFN, due to less stimulation of pattern recognition receptors or their genetic makeup. Levels of IFN might in that case not reach the threshold levels required for the transcription of IFN-dependent genes. In addition, other (yet unknown) regulatory mechanisms may negatively influence transcription of these genes. Together, this may lead to an underestimation of in how many pSS patients IFN activity is involved in the pathogenesis. Furthermore, whether only IFN-positive patients would benefit from certain therapies, including IFN-targeted therapies, remains to be seen. Many factors regulate the production of IFN, including ligands for pattern recognition receptors and cytokines, which likely change during the disease process. Indeed, although in a small group of patients the IFN signature was reported to be longitudinally stable over time (3.6 ± 2.5 years), the signature seems to change in a number of patients during this period.[14] Long-term cohort studies with larger numbers of pSS patients are urgently needed to see whether the IFN signature changes over time, whether and how it is associated with various serological and clinical disease parameters, and whether treatment can affect the signature. To compare signatures from different studies a unifying definition of the signature is warranted. Establishing the activity of the IFN pathways has revealed the pivotal role of IFNs in the pathogenesis of pSS, at least in a subcategory of patients, and is very important for further unraveling of this complex disease. Whether this ‘signing’ of pSS patients is of additional value for the identification of different patient categories, that cannot be done with the existing serological and clinical parameters, awaits long-term analysis of a larger cohort of patients.

Financial & competing interests disclosure

F Kroese, K de Leeuw, and H Bootsma are supported by grants from the Dutch Arthritis Foundation. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.