Key points

Introduction

pSS can be considered a systemic autoimmune disease with a strong organ bias. The involvement of the exocrine glands is prevalent and drives the pathognomonic manifestations of dryness that define the sicca syndrome. Although the involvement of the ocular glands and the mucosal exocrine glands in the respiratory and gastrointestinal systems is less studied, the accessibility of minor and major salivary gland tissue has dramatically contributed to the understanding of pSS pathogenesis. Differently from other diseases, where target organs embody disease casualties (eg, the kidneys in lupus or the lungs in vasculitides), there is clear evidence that the salivary glands represent both the pathogenic hub of pSS and its end target. Salivary glands provide the local source of autoantigens, ectopic signals for lymphocyte organization and survival, and the microenvironment that supports autoantibody production.

Compelling evidence suggests that several biological processes can be involved in pSS and contribute to the establishment of salivary gland pathology. Those encompass pathways of activation belonging both to the innate and acquired immune systems, such as interferon (IFN) activation, defective T-regulatory activity, increased number and function of helper T cell (T _H ) 17, and excessive costimulation. Lymphoneogenesis with germinal center (GC) formation and clonal expansion of malignant B cells is ascribed to the aberrant B-cell activation that characterized a subset of patients with high serum level of autoantibodies, systemic manifestations, and high risk of lymphoma development.

Whether those pathogenic processes coexist in the same patient and manifest in different disease phases or whether different pathogenic processes occur in different subjects still represents an area of debate.

Introduction

pSS can be considered a systemic autoimmune disease with a strong organ bias. The involvement of the exocrine glands is prevalent and drives the pathognomonic manifestations of dryness that define the sicca syndrome. Although the involvement of the ocular glands and the mucosal exocrine glands in the respiratory and gastrointestinal systems is less studied, the accessibility of minor and major salivary gland tissue has dramatically contributed to the understanding of pSS pathogenesis. Differently from other diseases, where target organs embody disease casualties (eg, the kidneys in lupus or the lungs in vasculitides), there is clear evidence that the salivary glands represent both the pathogenic hub of pSS and its end target. Salivary glands provide the local source of autoantigens, ectopic signals for lymphocyte organization and survival, and the microenvironment that supports autoantibody production.

Compelling evidence suggests that several biological processes can be involved in pSS and contribute to the establishment of salivary gland pathology. Those encompass pathways of activation belonging both to the innate and acquired immune systems, such as interferon (IFN) activation, defective T-regulatory activity, increased number and function of helper T cell (T _H ) 17, and excessive costimulation. Lymphoneogenesis with germinal center (GC) formation and clonal expansion of malignant B cells is ascribed to the aberrant B-cell activation that characterized a subset of patients with high serum level of autoantibodies, systemic manifestations, and high risk of lymphoma development.

Whether those pathogenic processes coexist in the same patient and manifest in different disease phases or whether different pathogenic processes occur in different subjects still represents an area of debate.

Salivary gland epithelial cells and activation of the interferon type I pathway

The cross talk between salivary gland epithelial cells (SGECs) and the recruited lymphocytes plays a central role in salivary gland pathology. Some of the viruses and bacteria that display specific epithelial tropism have been implicated in SS pathogenesis. Although a single responsible agent has not been identified, there are reports of viral involvement in support of the persistent antigenic exposure to locally recruited lymphocytes. Evidence of Epstein-Barr virus, cytomegalovirus, human herpesvirus type 6, human T-lymphotropic virus type I, human herpesvirus type 8, and coxsackie virus infection have been provided in pSS.

Pathogen-induced SGEC damage is deemed responsible for the release of autoantigens and induction of intrinsic mechanisms of cellular responses. It is believed that critical modifications of the SGEC expression profiles occur on injury and that SGECs isolated from pSS patients can maintain those changes in vitro. Toll-like receptor (TLR) engagement, in particular TLR3 for SGECs and TLR7 and TLR8 for plasmacytoid dendritic cells (pDCs), is involved in the earliest phases of pSS pathology. This results in local up-regulation of the type I IFN genes in both SGECs and pDCs. The importance of type I INF activation is further supported by the association between IRF5 and STAT4 polymorphisms with disease susceptibility. Locally released IFN stimulates autocrine and paracrine release of the B-cell activating factor (BAFF) that, in turn, induces B-cell activation and supports autoantibody production and formation of immune complexes that stimulate recruited pDCs for the further release of IFN. The presence of the risk alleles for IRF5 correlates with a high level of IRF5 mRNA in both peripheral blood mononuclear cells and SGECs and with increased levels of IFN-induced gene transcripts. Similarly, STAT4 activation is responsible for the IL-12–dependent activation of natural killer (NK) cells, polarization of naïve CD4 ⁺ T cells to IFN-γ producing T _H 1 cells, and the IL-23–dependent expansion of T _H 17 cells, thus dramatically contributing to the autoimmune process.

The pathogenic cross talk established in the earliest phases of pSS between leukocytes and SGECs is maintained during the length of the disease. SGECs provide key signals that enable recruitment (CCL17, CCL19, CCL21, CCL22, CXCL10, CXCL12, and CXCL13 ) and adhesion molecules that facilitate the organization of the lymphocytic aggregates and are responsible for the classical periductal distribution of the foci.

To the recruited lymphocytes, SGECs provide survival factors and antigen presentation. SGECs release autoantigens in apoptotic blebs or in actively secreted vesicles. Direct presentation of nuclear extractable proteins, such as Ro52, which correlates with salivary inflammation, in association with the up-regulated MHC-II and in complex with CD40 has also been described.

SGECs are also responsible for the production of T-cell homeostatic cytokines, such as IL-7 and thymic stromal lymphopoietin (TSLP). IL-7 is a key mediator of salivary gland T-cell activation ; it is released on TLR activation and its levels correlate with the focus score (number of periductal foci in 4 mm ² ) and the total number of infiltrating T cells. An increased number of IL-7R cells is found in the salivary glands of pSS patients compared with sicca patients. Finally, ectopic IL-7 contributes to IFN-γ–mediated T _H 1 response and the formation of T _H 17. In contrast to IL-7, TSLP is decreased in pSS patients’ salivary glands and TSLP expression negatively correlates with T-cell infiltration and markers of inflammation. As such, TSLP seems to function as a mediator in tissue homeostasis. As discussed previously, SGECs express BAFF, which is a critical factor in B-cell activation and survival.

The ectopic production of lymphocyte survival factors supports the expansion and persistence within the glands of autoreactive clones, naturally deemed to die for the scarce expression of follicle-associated chemokine receptors (such as CXCR5) and BAFF receptor. Not surprisingly, the serum expression of those cytokines correlates significantly with local expansion and organization of the lymphocytic aggregates, with markers of disease activity and autoantibody production, thus providing a promising therapeutic target in pSS.

T lymphocytes

The composition of the lymphocytic infiltration of the salivary gland is dependent on the grade of infiltration and inflammation. Small lymphocytic foci, organized around the intralobular and interlobular ducts mainly consist of CD4 ⁺ T cells and dendritic cells. Progressive increase of the B-cell component is observed, during the disease course, which correlates with onset of systemic manifestations and increased focus score. Formation of germinal centers (GCs), areas of lymphocyte proliferation within the dense packed B-cell infiltrate, is observed in approximately 20% to 25% of SS patients. More common is the presence of organized foci, with separate T-cell or B-cell rich areas, in absence of fully formed GCs.

Although the potential viral etiology and the predominance of the type I IFN response suggest a predominance of CD8 ⁺ T cells, CD4 ⁺ cells embody the majority of infiltrating T lymphocytes. CD8 ⁺ T cells are present in pSS infiltrates but at relatively low numbers and are not considered to play a major part in pathogenesis.

Salivary gland infiltrating CD4 ⁺ T cells display an activated phenotype and contribute to disease pathogenesis, both releasing pathogenic cytokines and providing support to recruited B cells. CD4 ⁺ cell-derived cytokines are mainly represented by IFN-γ and interleukin (IL)-17, which are associated with salivary gland tissue damage. IFN-γ is produced both by T cells and NK cells and is involved in epithelial cell activation, promoting antimicrobial protection, apoptosis, inflammation, and tissue damage by up-regulation of IFN II-induced genes. In the periphery, up-regulated IFN-inducible genes (IFN signature, both types I and II) are found in approximately 50% of SS patients, as reviewed elsewhere.

The presence of a type I IFN signature has been associated with a more severe disease phenotype, including a higher European League Against Rheumatism Sjögren’s Syndrome Disease Activity Index score, increased autoantibody production, and hypergammaglobulinemia. Nonetheless, a recent study demonstrated that in patients with high salivary gland expression of IFN-inducible genes, patients could be divided in groups with type I, type II, and mixed predominance, with the type II predominance associated with higher focus score and increased lymphoma risk. Stratifying the patients according to the type of IFN predominance might provide an interesting tool for disease prognosis and response to treatment.

The authors’ group has recently demonstrated that in an animal model of pSS, infiltrating T cells also produce IL-22, a mucosa-associated homeostatic cytokine involved in epithelial response to damage. Aberrant salivary gland expression of IL-22 enables the ectopic production of the lymphoid chemokines CXCL13 and CXCL12, respectively, by fibroblasts and epithelial cells, thus contributing to lymphocyte organization and GC formation. In pSS, T _H 17 and NKp44 ⁺ NK cells represent the major cellular source of IL-22.

Although increasing evidence shows involvement of T _H 2 activity in the salivary glands, the exact contribution of T _H 2 cytokines is unclear. A correlation between the degree of local inflammation and a bias toward a T _H 1-predominant infiltrate has been made. Nonetheless, significant increase in T _H 2 cytokines is observed in GC-positive patients. The authors’ group previously demonstrated the association between IL-18, a known cytokine involved in the T _H 1/T _H 2 switch, with the increased production of antibodies against ENA. Another cytokine, IL-21, involved in T _H 17 cell homeostasis, plasma cells differentiation, and GC formation, is found in pSS salivary glands. Its expression correlates with IgG1 levels, anti-Ro/SSA antibody titters, and the degree of lymphocytic infiltration. In pSS patients, an increase in number of both circulating and salivary gland IL-21 ⁺ T follicular helper cells correlates with the number of memory B cells and plasma cells.

CD4-CD8 double-negative T cells have also been reported to infiltrate the salivary glands, at later stages of disease; those are associated with the symptoms of dryness and presence of GC-like structures in the salivary glands.

B cells and lymphoneogenesis

Progressive B-cell aggregation in the salivary glands is mediated by the aberrant production of the B-cell chemoattractant CXCL13, derived from SGECs, IL-22–activated stromal cells, and macrophages. Alongside CXCL13, ectopic expression of CCL21, CCL19, and CXCL12 has been described in the salivary glands of pSS patients and animal models of the disease. The expression of these lymphoid chemokines, responsible for recruitment, positioning, and degree of organization of B and T compartments, can be detected in discrete areas within ectopic lymphoid structures, resembling the lymphoid organization observed in secondary lymphoid organs (SLOs). Increased gene expression of those chemokines and their receptors correlates with B-cell accumulation and presence of GC in pSS salivary glands.

In a high proportion of pSS patients, the B cells present within salivary gland GCs lack the markers commonly found in typical GC B cells, such as CD10 and CD38. A small percentage of pSS patients (approximately 5%) presents with CD21 ⁺ CD38 ⁺ B cells in the lymphoid infiltrates, consistent with the phenotype observed in chronically activated tonsillar GC B cells. Still, most B cells found in the salivary glands of pSS patients display a phenotype consistent with Transitional type 2 B cells, a mature subset of transitional B cells, and marginal zone–type B cells (CD19 ⁺ IgD ⁺ CD38-IgM ⁺ CD21 ⁺ CD23 ⁺ and CD20 ⁺ IgD ⁺ CD38-CD21 ⁺ CD24 ⁺ ). Recently, CD138 ⁺ Bcl-2 ⁺ plasma cells have been detected in areas characterized by CXCL12 and IL-6 expression in pSS salivary glands with high focus score. Moreover, the presence of anti-Ro/SSA and anti-La/SSB (whose serum detection is an important diagnostic criteria)–producing cells has been described in the periphery of GC-like structures in labial salivary gland biopsies. Increasing interest has been raised on the ability of B cells to contribute to the development of autoimmunity, specifically the ability of B cells to regulate dendritic cells, activate T cells, act as antigen-presenting cells, and secrete cytokines.

The cross talk between T cells and B cells is of paramount importance for the generation of the GC reaction. GC-like structures can be identified in salivary gland histologic sections and their presence correlates with higher focus score and extensive B-cell infiltration. Histologically, GCs can be identified as areas of lighter staining (due to the presence of lymphoblasts and follicular dendritic cell networks) within darker confined and organized areas of active B-cell proliferation mainly populated by centroblasts. In minor salivary glands, segregation into dark and light zone compartments is not as easily identified as in major glands. Sensitivity in the detection can be enhanced by CD21, CD23, bcl-6, and CD35 stainings. GCs, both in minor and major salivary glands, host the machinery that enables the process of B-cell affinity maturation. The presence of the enzyme activation-induced cytidine deaminase (AID) has been described by the authors’ group within GC-positive pSS salivary glands.

The possible association between ectopic GC formation and the development of B-cell lymphoma (observed in 3.4%–7% of pSS patients ) had been long suspected on the basis of the biological risk related to clonal cell expansion at ectopic sites. In SLOs, the direction of travel of lymphocytes is established by stringent gradients of chemokine expression and resident stromal cells set the availability of locally released survival factors. Lymphocyte survival and proliferation in SLO is also controlled by the access to the follicles and limited by the competition between alloreactive and autoreactive clones (that present down-regulated CXCR5 and BAFF receptor expression). In the salivary glands, the ability to screen autoreactive clones is doubtful in GCs that lack a rigorous anatomic segregation. Although a process of progressive transformation of the resident salivary glands stroma toward lymphoid stroma has been described, this does not seem to reproduce the degree of organization present in SLO GCs. Expression of CXCL13 is not confined within the B-cell area; the expression of CCL21 is limited to the perivascular myofibroblasts and aberrant niches for B aggregation are provided by locally activated SGECs. This “confusing” microenvironment for lymphocyte migration is accompanied by aberrant and diffuse expression of BAFF and a proliferation-inducing ligand (APRIL) that enables survival of autoreactive, pathogenic clones that in SLOs would be classically excluded from the follicles. It is, therefore, not surprising that somehow GC formation is associated with lymphoma development. Theander and colleagues, however, could not establish a positive correlation between the 2 histologic entities (14%) but did define a strong negative correlation between the absence of GC detection and the nondevelopment of B-cell lymphoma, supporting the concept that uncontrolled B-cell clonal expansion represents a critical step in lymphomagenesis. Nonetheless, a more rigorous characterization of the additional stages involved in lymphomagenesis is needed to better understand this phenomenon.

In this context, it is important to highlight the largely unexplored pathogenic cross talk between the GC and the areas of lymphoepithelial proliferation within the lymphoepithelial sialoadenitis (LESA). Large areas of LESA are commonly considered to be premalignant lesions in inflamed glands. Within established samples of MALT lymphoma, malignant B cells that bear histological features of immunoblasts or centroblasts are preferentially distributed in sheets. These sheets either infiltrate the reactive follicles, assuming the characteristic marginal one distribution, or, more often, spread in the interfollicular area adjacent to the LESA. A clonal relationship between the polyclonal B-cell expansion occurring within the GC of pSS patients and the following lymphomatous transformation has been established; however, the role of epithelial cells and the pathogenic relevance of LESA/GC interaction during lymphomagenesis should be further exploited.