Salivary glands of primary Sjögren's syndrome patients express factors vital for plasma cell survival

Lower labial minor salivary gland biopsies from 21 consecutive patients fulfilling the American-European consensus group criteria (AECC) of pSS [31] and 10 individuals not fulfilling the pSS criteria were used in this study. All biopsies had been performed between 2004 and 2007 at the Department of Otolaryngology/Head and Neck Surgery at Haukeland University Hospital in Bergen. All tissue sections had been previously evaluated by an oral pathologist and the focus score (FS) (the number of focal mononuclear cell infiltrates with >50 mononuclear cells per 4 mm²) was determined. All patients had FS ≥1 and were divided in two groups; patients with given FS equal to 1 (FS = 1) and patients with FS equal to or more than 2 (FS ≥2). Hematoxylin and eosin (H&E) stained formalin fixed and paraffin embedded minor salivary gland tissue sections were further screened for the presence of germinal center (GC) like structures. Individuals that were evaluated for pSS, but did not fulfill the AECC criteria served as non-pSS tissue controls and were divided into chronically inflamed (CIG) (small focal infiltrates with less than 50 cells; chronically inflamed tissue not sufficient for SS) and normal glands (NG) (no inflammation in the salivary gland tissue). Medical records were obtained from patients' charts at the Department of Rheumatology, Haukeland University Hospital and information was collected regarding routine laboratory assessments including RF detection, antinuclear antibodies (ANA), anti-Ro/SSA, anti-La/SSB, and serum immunoglobulin levels (IgG, IgA and IgM). The characteristics of patients and subjects included in the study are shown in Table 1. The study was approved by the Committee of Ethics at the University of Bergen (145/96-44.96 and 242.06). All studied subjects gave their informed consent.

The following primary anti-human antibodies were used in this study: mouse monoclonal CD138 (1:00) (clone MI15, Dako A/S, Glostrup, Denmark), rabbit polyclonal IgA (1:50000) (Dako A/S), rabbit polyclonal IgG (1:30000) (Dako A/S), mouse monoclonal Ki-67 (1:150) (clone MIB-1, Dako A/S), rabbit monoclonal Bcl-2 (1:100) (clone E17, Abcam plc., Cambridge, UK), mouse monoclonal CXCL12 (1:250) (clone 79014.111, R&D Systems, Minneapolis, MN USA), rabbit polyclonal IL-6 (1:500) (Abcam plc) and rat anti-PNAd carbohydrated epitope (clone MECA-79, BD Bioscience, Trondheim, Norway). Isotype and concentration matched antibodies (mouse IgG1 and rabbit Ig fraction, Dako A/S) were used as negative controls.

Statistical analyses were performed using SPSS version 15.0 software (IBM, Chicago, IL, USA). Normal distributions of the results were tested by Kolmogorov-Smirnov. All data were normally distributed and thus statistic significance was tested by Student t-test and presented as mean ± standard error of mean (SEM). Differences were considered significant when P < 0.05.

The pSS patients used in this study were divided into two groups according to FS; one group consisted of patients with FS = 1 and one group consisted of patients with FS ≥2. By morphology, we detected GC-like structures (GC+) in salivary glands of 7/21 (33%) patients with pSS. The majority of the GC+ patients were in the FS ≥2 group; mean FS in the GC+ was 4.4 as compared to 2 in the GC- patients (Table 1). Subjects not fulfilling the criteria of pSS were also divided into two groups; 1) subjects with chronic inflammation which presented as diffuse non-focal infiltration of cellular aggregates with less than 50 cells in the minor salivary gland tissue and 2) subjects with normal salivary gland histology; occasional cellular aggregates but not more than expected in normal tissue, and absence of focal inflammation. The mean unstimulated salivary secretion levels were lower than 1.5 ml/15 minutes in all study groups (Table 1). Eight pSS patients were positive for Ro/SSA or/and La/SSB. The mean antinuclear antibodies (ANA) levels in serum were 2.47 in pSS with FS = 1, 3.56 in pSS with FS ≥2, 0.35 in subjects with chronically inflamed glands and 0.77 for subjects with normal gland histology. Only one pSS patient had high amounts of rheumatoid factor in serum; however, five pSS patients were highly positive for cyclic citrullinated peptide antibody (CCP). There were no significant differences in IgG and IgA levels between the groups. The IgM levels were significantly higher in pSS with FS ≥2 compared to chronically inflamed (P < 0.05) and normal (P < 0.05) subjects (Table 1).

Plasma cells in the salivary gland tissue identified by the expression of CD138 showed both plasma blast and mature plasma cell morphology. Cells with mature plasma cell morphology had characteristic eccentrically placed nucleus, prominent cytoplasm and oval shape, whereas cells with plasma blast morphology were smaller, spherical and with higher nucleus-to-cytoplasm ratio.

CD138+ plasma cells were detected in the salivary gland tissue of all investigated patients. In patients with FS = 1, CD138 expressing cells were detected interstitially as small cell clusters close to ductal and acinar epithelium and in the periphery of larger infiltrates. In the biopsy samples from patients with FS ≥2, CD138 expression was identified interstitially and both in the periphery and within larger focal infiltrates (Figure 1A).

Further, we detected CD138+ plasma cells within some GC-like structures as described in more detail later.

Significantly higher numbers of CD138+ interstitial cells were detected in pSS patients with FS ≥2 compared to the other groups tested (Figure 1B). There were no significant differences in the amount of CD138+ interstitial cells between the pSS group with FS = 1, chronic inflammation and normal glands (Figure 1B). When looking at focal infiltrates, we found no significant differences between the percentage of CD138+ cells in pSS patients with FS = 1 and pSS patients with FS ≥2 (Figure 1C). In our material, the presence of GC-like structures had no impact on the numbers of detected CD138 plasma cells.

An analysis of immunoglobulin phenotype showed the presence of both IgA and IgG expressing plasma cells in the salivary gland tissue of all patients and subjects tested (Figure 2). IgA expressing plasma cells were detected in small cell clusters in close proximity to ductal and acinar epithelium in all groups tested and in the periphery of smaller and larger focal infiltrates in pSS salivary glands. IgA+ cells clearly resembled plasma cell morphology (Figure 2A). IgG expressing plasma cells were found both inside and in the periphery of large focal infiltrates in pSS patients with FS = 1 and FS ≥2, whereas interstitial accumulation of IgG expressing plasma cells was observed mostly in the pSS patients with high FS. Similarly to IgA, IgG+ cells clearly resembled plasma cell morphology (Figure 2B).

Approximately the same frequencies of IgA expressing plasma cells were noted in the interstitial areas of all three groups tested (Figure 2C). However, not significantly, a reduction in IgA+ plasma cells was noted in the focal infiltrates of pSS patients with FS ≥2 (Figure 2D).

Both pSS groups (FS = 1 and FS ≥2) and subjects with chronically inflamed salivary glands had significantly higher numbers of interstitial IgG plasma cells compared with normal tissue. However, pSS patients with high FS demonstrated the most prominent increase in the interstitial IgG plasma cells (Figure 2E). The number of IgG expressing plasma cells in the focal infiltrates was higher in the pSS with FS ≥2 but due to great variation within the groups, the differences were not significant (Figure 2F).

Interestingly, the focal infiltrates of pSS patients with FS = 1 contained similar frequencies of IgA and IgG expressing plasma cells, whereas in the focal infiltrates of pSS patients with high FS, the numbers of IgG expressing plasma cells were three times higher than numbers of IgA positive plasma cells (Figure 2D, F). The presence of GC-like structures appeared to have no impact on the amounts of IgA or IgG expressing plasma cells. No correlation was detected when we compared the increased number of IgG versus IgA plasma cells in the salivary glands and serum IgG and IgA levels in patients with pSS.

Since it has been proposed that the long-lived plasma cells survive without detectable proliferation [21, 32], a Ki-67/CD138 double-staining was performed.

Ki-67 was generally expressed on mononuclear cells in the focal infiltrates of pSS patients with both FS = 1 and FS ≥2 and in the interstitial cells of all groups included in this study (Figure 3). Frequencies of interstitially Ki-67+ cells were low in the normal tissue, but increased significantly in chronically inflamed and in pSS tissue with FS = 1. A remarkable increase of Ki-67+ interstitial cells was observed in the salivary glands of pSS patients with FS ≥2 (Figure 3A). Only a slight difference was detected in the expression of Ki-67 by focally infiltrating cells in pSS patients with FS = 1 and in pSS patients with FS ≥2 (Figure 3B). Moreover, GC positivity did not seem to associate with the numbers of focally infiltrating Ki-67+ cells.

Most of the CD138+ plasma cells did not express Ki-67. For the pSS groups with FS = 1 and FS ≥2 we detected on average four double-positive cells per gland. For the group with chronic inflammation we detected one double-positive cell per gland, whereas no double positive cells were detected in normal salivary gland tissue. Ki-67+ plasma cells were mainly observed in the periphery of larger infiltrates with high proliferating activity and in some small plasma cell aggregates often in close proximity to vascular structures (Figure 3C).

In order to determine whether the accumulating plasma cells in the minor salivary glands of pSS patients and controls are primed to survive for long periods of time [33, 34] we looked at the expression of anti-apoptotic Bcl-2 protein.

Bcl-2 was expressed in the cytoplasm of mononuclear cells both in the focal infiltrates of pSS patients and interstitially in all study groups (Figure 4A). We observed significantly higher expression of Bcl-2 in the interstitial cells of pSS patients with FS ≥2 compared with the three other groups (Figure 4B). Furthermore, significantly higher numbers of Bcl-2+ interstitial cells were found in the pSS tissue with FS = 1 comparing with normal, but not with chronically inflamed tissue (Figure 4B). When evaluating focal infiltrates, a dramatic increase in the levels of Bcl-2 expressing cells was observed in the pSS with high FS comparing with pSS with FS = 1 (Figure 4C). GC positivity did not seem to influence the numbers of Bcl-2 positive cells in the salivary gland tissue included in this study.

Notably, no Bcl-2 negative plasma cells could be detected in the salivary glands of pSS patients (both FS = 1 and FS ≥2) (Figure 5A-D). In the chronically inflamed and normal tissue both Bcl-2+ and Bcl-2- plasma cells were observed (Figure 5E-H)). Even though it is difficult to evaluate the intensity of the Bcl-2 expression, a more distinct cytoplasmic Bcl-2 staining was noted in plasma cells from pSS salivary glands compared to plasma cells from subjects with chronic inflammation and normal salivary gland tissue (Figure 5).

CXCL12 (also known as stromal derived factor 1; SDF-1) plays a crucial role in the migration and survival of plasma cells [22, 23, 27]. To investigate the possibility of CXCL12 to facilitate plasma cell survival in the inflamed salivary gland tissue, a double-staining with anti-CXCL12 and anti-CD138 was performed.

CXCL12 staining was detected on acinar and ductal epithelial cells in pSS patients and to a lesser extent in chronic inflammation and normal subjects (Figure 6A). CXCL12 was also expressed on the mononuclear cells interstitially in pSS and chronic inflammation subjects and in focal infiltrates of all pSS patients. No expression of CXCL12 was detected on the interstitial cells in normal salivary gland tissue (Figure 6A). Additionally, we observed expression of CXCL12 by resident adipocytes present in some areas of salivary gland tissue of pSS patients (Figure 6B).

In pSS patients and chronic inflammation subjects, CD138+ interstitial plasma cells were mainly observed in close proximity to the CXCL12 expressing ductal and acinar epithelium and CXCL12 expressing mononuclear cells (Figure 6A). In contrast, in normal tissue clusters of CD138+ plasma cells were mainly observed in close proximity to CXCL12 negative acini. Interestingly, small aggregates of CD138+ plasma cells were in close contact to CXCL12 expressing adipocytes in pSS salivary gland tissue.

Significantly higher numbers of CD138+ interstitial plasma cells were in contact with CXCL12 expressing cells in pSS patients with FS ≥2 compared to chronic inflammation and normal subjects (Figure 6C). However, there were no statistical differences in the numbers of interstitial plasma cells in contact with CXCL12 between pSS with FS = 1 and chronic inflamed tissue. In normal glands, only few plasma cells were in contact with CXCL12 expressing epithelium (Figure 6C).

Plasma cells residing in the focal infiltrates of pSS patients were in contact with CXCL12 expressing cells (Figure 6A). The frequency of interactions between the CD138+ plasma cells and CXCL12 were significantly increased in the focal infiltrates of pSS with FS ≥2 compared with interactions seen in the pSS tissue with FS = 1 (Figure 6D).

IL-6 cytokine has proven to be vital for plasma cell survival [22]. Thus, we performed a double-staining for IL-6 and CD138, to investigate if IL-6 can support the survival of plasma cells in the salivary gland tissue.

IL-6 was expressed primarily on the salivary gland interstitial cells in all groups studied and the majority of salivary gland epithelial cells were negative for this cytokine (Figure 7). Only scattered IL-6+ acinar epithelial cells were detected in some of the pSS patients (data not shown). IL-6 was demonstrated within the majority of the focal infiltrates; however, the number of cells expressing IL-6 was generally higher in the pSS patients with FS ≥2 (Figure 7A).

Salivary glands from pSS patients had significantly higher numbers of CD138+ plasma cells in close contact with IL-6 producing cells than plasma cells in chronically inflamed and normal tissue (Figure 7B). Moreover, the number of plasma cells that interacted with IL-6, increased drastically from pSS patients with FS = 1 to pSS with FS ≥2 (Figure 7B). Since only few IL-6 expressing cells were detected in chronically inflamed and normal subjects, the contact between these cells and CD138+ plasma cells was rare (Figure 7B).

When investigating focal infiltrates, we detected higher numbers of CD138+ plasma cells in close proximity to the IL-6 producing cells in the pSS with high FS compared with pSS with FS = 1. However, due to variation in the presence of IL-6 producing cells, these differences were not significant (Figure 7C).

Formation of specific blood vessels called high endothelial venules (HEVs) has been detected in the inflamed tissue of autoimmune diseases such as the synovium in rheumatoid arthritis and in salivary gland tissue of SS patients [35‐37]. However, little is known about the phenotype and the function of these structures at the site of inflammation. To further characterize the distribution pattern of HEVs in the salivary glands of pSS and in particular their possible contribution to 'survival niches' for plasma cells, we performed for the first time a IHC triple staining with PNAd, CXCL12 and CD138. In addition, consecutive salivary gland tissue sections were single stained with PNAd and IL-6.

HEVs were detected interstitially in all groups tested. In pSS patients, HEVs were often localized in close proximity to and within focal infiltrates. Occurrence of HEVs increased with increasing inflammation and thus high numbers were detected in pSS patients with FS ≥2 (Figure 8A, B). Using the triple staining, only few of the HEV endothelial cells in pSS salivary glands were expressing CXCL12. CD138+ plasma cells were often observed in an extravascular location, sometimes tightly adherent to both CXCL12 positive and CXCL12 negative HEVs (Figure 8C, D). No CXCL12 positive HEVs were recognized in chronically inflamed or in normal salivary gland tissue.

To gain further insight into the mode of chemokine binding to HEVs, we next investigated the possible relationship between IL-6 and HEVs in the salivary glands. Indeed most of HEVs in our selected pSS material expressed IL-6 (Figure 9).