To prospectively evaluate histopathologic, blood cellular, serologic, and clinical changes in response to abatacept treatment in patients with primary Sjögren's syndrome (SS).
To prospectively evaluate histopathologic, blood cellular, serologic, and clinical changes in response to abatacept treatment in patients with primary Sjögren's syndrome (SS).
Blood, saliva, and minor salivary gland biopsy samples were obtained before and after the last of 8 doses of abatacept in 11 primary SS patients. The histologic data evaluated the numbers of lymphocytic foci and B and T cell subtypes (CD20+, CD3+, CD4+, and CD8+). The numbers of FoxP3+ regulatory T cells were measured and the FoxP3:CD3 ratio was calculated. Histologic data were compared with results from peripheral blood and with changes in saliva secretion.
The numbers of lymphocytic foci decreased significantly (P = 0.041). Numbers of local FoxP3+ T cells decreased significantly in percentage of total lymphocytic infiltrates (P = 0.037). In the peripheral blood, B cells increased (P = 0.038). This was due to an expansion of the naive B cell pool (P = 0.034). When adjusting for disease duration, an increase was also noted for total lymphocytes (P = 0.044) and for CD4 cells (P = 0.009). Gamma globulins decreased significantly (P = 0.005), but IgG reduction did not reach significance. Adjusted for disease duration, saliva production increased significantly (P = 0.029).
CTLA-4Ig treatment significantly reduces glandular inflammation in primary SS, induces several cellular changes, and increases saliva production. Remarkably, this increase in saliva production is significantly influenced by disease duration.
Primary Sjögren's syndrome (SS) is defined as an autoimmune connective tissue disease typically affecting salivary and lacrimal glands with consecutively reduced saliva and tear production, as well as affecting extraglandular organs (). Patients with primary SS have an increased risk for the development of B cell lymphoma and cryoglobulinemic vasculitis ().
From the pathophysiologic perspective, T as well as B lymphocytes have been shown to be involved in the generation and perpetuation of inflammatory infiltrates in salivary glands (). Targeting the T lymphocyte part with, e.g., leflunomide or systemic application of cyclosporin A, showed histologic changes of the salivary glands, yet no marked clinical benefit ([4, 5]). Treatment strategies targeting B lymphocytes have recently been addressed in more detail (). The reduction of CD20 expressing B cells was of partial effect in selected patients (). Recently, the hypothesis has been put forward that it is not the lymphocyte subtype itself, but the interaction between B and T lymphocytes that is important in the pathophysiology of primary SS (). In a mouse model, the local application of CTLA-4Ig, a biologic agent that blocks the costimulation of T lymphocytes, showed a reduction of inflammation resulting in an increase in salivary gland function (). A therapeutic modulation of the costimulatory interaction between B and T cells in humans, however, has not yet been studied. In this study, we evaluated the effect of CTLA-4Ig on the salivary gland inflammation in primary SS patients. We measured local as well as systemic cellular changes and the effect on saliva secretion.
This prospective study was approved by the local ethics committee. All patients gave written informed consent prior to participation.
Patients ages >18 years with a diagnosis of primary SS according to the revised criteria of the American–European Consensus Group, including a positive histology from labial salivary glands, were eligible for inclusion prior to study enrollment (). Women of childbearing age were asked to take safe contraception.
Patients with secondary sicca syndrome, overlap syndrome, or immunosuppression with biologic or nonbiologic disease-modifying antirheumatic drugs, including cyclophosphamide, cyclosporine, azathioprine, or glucocorticoids, within the last 6 months were excluded. Further exclusion criteria were active infection, immunization within 4 weeks prior to enrollment or planned active immunization within the study period, pregnancy or breastfeeding, malignancy, serum creatinine level >1.5-fold above normal, aspartate aminotransferase or alanine aminotransferase >2.5-fold above normal, hemoglobin <9.5 gm/dl, platelets <50,000/μl, neutrophilic granulocytes <0.5/μl, and hypogammaglobulinemia (<50% below normal).
Labial salivary gland biopsy samples were obtained in a minimally invasive manner prior to study enrollment and 4 weeks after the last abatacept dose ().
The dosing regimen approved for the treatment of rheumatoid arthritis was used (). A total of 8 weight-adapted doses of abatacept (500 mg below 60 kg body weight or 750 mg above 60 kg body weight per infusion) at weeks 0, 2, 4, 8, 12, 16, 20, and 24 were infused in a standardized manner according to the manufacturer's guidelines.
Prior to the first and after the last abatacept infusion, we quantified stimulated saliva secretion in a standardized manner using the Saxon test (). At the same time points, concentrations of the immunoglobulin isotypes IgG, IgA, and IgM were determined by nephelometry (Dade Behring nephelometer) and lymphocyte subpopulations were analyzed by flow cytometry. For analysis of lymphocytes, EDTA blood samples were transferred into Trucount tubes (BD Biosciences) for fixation of leukocytes and lysis of erythrocytes. Cells were stained with the following BD Multitest reagents (BD Biosciences): CD3/CD8/CD45/CD4, CD3/CD16+CD56/CD45/CD19, and phycoerythrin-conjugated anti-CD27 (clone L128), as well as fluorescein isothiocyanate–conjugated anti-IgD (clone IA6-2). All antibodies were purchased from BD Biosciences. Data were acquired using the FACSCanto II flow cytometry system and FACSDiva software. Analysis was performed after gating on anti-CD45+–stained lymphocytes. The numbers and percentages of naive, nonswitched memory B cells and switched memory B cells were determined according to their CD27 and IgD surface expression. All measurements were performed according to the standard operating procedure guidelines of our accredited laboratory for flow cytometry.
Biopsy specimens from labial minor salivary glands were fixed in 4% formalin for 24 hours, embedded in paraffin, cut at thickness 2–3 μm, and stained with hematoxylin and eosin. In each biopsy sample, the area of total glandular parenchyma was measured in mm2. The inflammatory infiltrate was characterized as follows: the numbers of lymphocytic foci with >50 lymphocytic cells, of lymphoepithelial lesions, and of foci with active gland destruction by infiltrating lymphocytes per biopsy sample were counted and expressed per mm2 of glandular tissue. The numbers of lymphocytic foci were independently counted by 2 different pathologists in a blinded manner. In case of discrepant results, consensus was reached at a double-head microscope. Lobules of atrophy as well as focal lymphocytic infiltrates showing crush artifacts were excluded from the focus score. Percentages of glandular parenchyma affected by atrophy and interstitial fibrosis were assessed, where 0 = no atrophy/fibrosis, 1 = 1–10% of glandular area affected by atrophy/fibrosis, 2 = 11–20% of glandular area affected by atrophy/fibrosis, etc.
Serial tissue sections of salivary and gland biopsy specimens were immunohistochemically stained for expression of CD20, CD3, CD4, CD8, FoxP3, IgM, IgG, and IgA. The following antibodies were used for immunohistochemistry: CD3 (clone ON19) and CD4 (clone 4B12; both Leica Biosystems); CD8 (clone C8/144B) and CD20 (clone L26; both Dako); FoxP3 (clone 236A/E7; Abcam); and polyclonal IgG, IgA, and IgM (Dako). Prior to incubation with the primary antibody, 2–3-μm thick paraffin-embedded tissue sections were dewaxed, rehydrated, and pretreated either by boiling in 10 mM citrate buffer, pH 6.0, or in 10 mM Tris–1 mM EDTA, pH 9.0, in a microwave oven, or were digested with 1 mg/ml of trypsin 1:250 (Difco). Following pretreatment, slides were incubated for 5 minutes in 3% hydrogen peroxide (H2O2) with 0.1% sodium azide to block endogenous peroxidase activity. Following all subsequent steps, sections were then washed in Tris buffered saline (TBS) and incubated with the primary antibody diluted in TBS with 0.5% casein and 5% normal goat serum for 60 minutes at room temperature. In negative controls, the primary antibody was replaced with antibody dilution buffer in order to exclude nonspecific staining of endogenous peroxidase. Next, a polymer-based visualization system with horseradish peroxidase (EnVision Plus, Dako) was applied for 30 minutes. Finally, sections were developed in 0.02% 3,3′-diaminobenzidine (Sigma) with 0.01% H2O2, counterstained with hematoxylin, and mounted. Known positive controls were stained in parallel with each series. Double staining for CD3 and FoxP3 was performed on an automated immunostainer (Bond III, Leica Biosystems) after EDTA-mediated antigen retrieval, using a peroxidase-based visualization system (Bond Refine, Leica Biosystems) to demonstrate FoxP3 and an alkaline phosphatase–based polymer (Bond Refine Red, Leica Biosystems) for CD3.
In each biopsy specimen, the cellular composition of up to 3 different focal lymphocytic infiltrates was characterized. In each focus, a fixed area of 0.2 mm2 was evaluated that usually covered the entire infiltrate focus. When infiltrate foci were >0.2 mm2, a representative cross-section of the focus containing both central/periductal and peripheral zones was assessed. Criteria for selecting foci for counting lymphocytic subpopulations were presence of representative cross-sections throughout all consecutive tissue sections and well-preserved tissue morphology without crush artifacts. In these foci, the absolute numbers of CD20+, CD3+, CD4+, CD8+, and FoxP3+ lymphocytes were counted on consecutive tissue sections. The sum of CD20+ B cells and CD3+ T cells was used as an approximate value of total infiltrating lymphocytes. The composition of the infiltrates with respect to lymphocytic subpopulations was expressed relative to the total of CD20+ B cells and CD3+ T cells. All values were expressed as median values of the foci. The diffuse plasmacytic infiltrate was typed by counting the absolute numbers of IgG+, IgA+, and IgM+ cells in the entire biopsy specimen. Values were expressed per mm2 of biopsy area.
Wilcoxon's test was applied for a first analysis of changes over time. In a second step, we used a repeated-measures analysis of variance to adjust analysis of time effects for possible confounding by age and disease duration. IBM SPSS statistics for Windows, version 19.0, was used for analysis. P less than 0.05 was considered statistically significant.
Between July 2008 and October 2010, 11 patients (all women, median age 47 years, range 25–75 years) with a median disease duration of 6 years (range 0.3–48 years) were included and all completed the study. The subjective reason for participation was intolerable dryness of the mouth. None of the patients experienced extraglandular disease activity at study entry or during the study.
An overview of patient characteristics, including saliva and tear production, is shown in Table 1. Table 2 shows histopathologic, serologic, and peripheral blood parameters, with Figure 1 displaying the respective significant values.
|Age, years||Disease duration, years||RF at recruitment, IU/ml||SSA antibodies at recruitment, units||SSB antibodies at recruitment, units||Saliva before treatment, ml/2 mins||Saliva after treatment, ml/2 mins||Δ in saliva, ml||Tear production before treatment, mm/5 mins||Tear production after treatment, mm/5 mins|
|N||Before treatment||Range||After treatment||Range||Normal value||P|
|Lymphocytic foci in total||4.4||0–11||2.1||0–7||0.041a|
|CD20+ B cells + CD3+ T cells, mm2||558||143–1,991||332||79–1,549||NS|
|CD20+ B cellsb||48.31||8.55–78.66||40.61||13.62–95.71||NS|
|CD3+ T cellsb||51.69||23.14–91.45||59.39||4.29–86.38||NS|
|CD4+ T cellsb||36.12||4.13–74.72||42.70||12.08–64.51||NS|
|CD8+ T cellsb||22.38||7.63–86.96||30.30||9.48–91.43||NS|
|FoxP3+ regulatory T cellsb||5.25||0.78–32.09||2.26||0.00–35.57||0.037a|
|Gamma globulins, gm/liter||19.4||14.1–26.8||16.1||11.2–27.2||7.7–14.6||0.005a|
|Peripheral blood cells, per μl|
|CD3+ T cells||9||971||141–1,925||1,132||173–1,989||690–2,540||NS|
|CD4+ T cells||9||550||120–1,221||716||141–1,062||410–1,590||NS|
|CD8+ T cells||9||360||15–926||270||22–830||190–1,140||NS|
|Naive B cells||8||166||74–434||215||140–519||0.034a|
|Memory B cells||8||10.44||1.5–60.3||13.9||0.5–128.3||NS|
|Switched memory B cells||8||7.23||1.1–45.6||9.6||0.2–100.7||NS|
|Nonswitched memory B cells||8||3.64||0.4–26.5||5.5||0.2–27.6||NS|
|Adjusted for disease duration, gm/2 mins||11||1.61||0.51–2.33||1.74||0.49–2.65||≥2.75||0.029a|
An analysis of histopathologic parameters was performed in 10 patients (patient 2 had to be excluded because the posttreatment biopsy specimen did not contain sufficient amounts of glandular tissue). The total number of lymphocytic foci decreased significantly from 4.4 to 2.1 (P = 0.041). Evaluation was performed by 2 independent pathologists counting a decrease of 4.3 to 1.6 by the first observer and from 4.6 to 2.5 by the second observer (intraclass correlation coefficient before abatacept 0.95, after abatacept 0.76, and for the respective delta values 0.88). The median number of lymphocytic foci per mm2 decreased from 1.2 to 0.51 after therapy (not significant).
The density of infiltrating lymphocytes within the foci diminished, reflected by a reduction of the total of CD20+ B cells as well as CD3+ T cells. This decrease affected all lymphocytic subpopulations, including CD4+, CD8+, and FoxP3+ T cells. Notably, the relative composition of lymphocytic subpopulations in the lymphocytic foci changed: while the fraction of CD20+ cells decreased, there was a relative increase of CD3+ T cells, including CD4+ and CD8+ subpopulations. Of particular interest was the relative reduction of FoxP3 regulatory cells (P = 0.037). A representative case is shown in Figure 2 (patient 4).
The diffuse plasmacytic infiltrate consisted of IgG+, IgA+, and IgM+ elements, with IgA+ cells being most prevalent and IgM+ cells comprising the smallest fraction. After therapy, the plasmacytic infiltrate and its relative composition remained unchanged (data not shown).
Active gland destruction characterized by a lymphocytic infiltrate within glandular structures was apparent in 3 biopsy samples before initiation of therapy. This low number did not allow statistical analysis; however, it is notable that after therapy, active destruction was not seen in any of the biopsy specimens. Atrophy was found in up to 10% of glandular tissue in 9 patients; in 1 patient, atrophy was present in 60–70% of glandular tissue. The atrophy score slightly improved by 1 point in 3 patients and deteriorated by 1 point in 2 patients, whereas it did not change in 5 patients after abatacept treatment.
Interstitial fibrosis was more variable before therapy, affecting between 10% and more than 90% of biopsy areas. After therapy with abatacept, there was an unchanged fibrosis score in 4 patients, an increase by 1 point in 4 patients, a decrease by 1 point in 1 patient, and an important decrease by 9 points in 1 patient. Changes in atrophy and fibrosis scores were not statistically significant.
Gamma globulin serum levels decreased significantly (P = 0.005), as well as when adjusting for disease duration (P = 0.009), from a median of 19.4 to 16.1 gm/liter. IgG decreased only slightly from a median of 15.5 to 14.4 gm/liter; there were no concordant nor significant changes in IgA and IgM serum levels.
When adjusting for disease duration, the increase in B and T lymphocytes reached statistical significance (P = 0.044), as did the increase in CD4 cells (P = 0.009). Increases in CD3 cells did not quite reach significance (P = 0.052), while CD8+ T cells showed a decrease (not significant).
There was a significant increase in B cells from a median of 174 to a median of 242 (P = 0.038). Regarding subpopulations, naive B cells increased from a median of 166 to a median of 215 (P = 0.034), and all other B cell subsets such as memory B cells, switched memory B cells, and nonswitched memory B cells did not show significant changes.
Saliva secretion increased from a median of 1.61 to a median of 1.74 gm/2 minutes (not significant). Since disease duration might have an influence on saliva production, data were adjusted for this possible confounder. Indeed, when adjusting for disease duration, the increase in saliva secretion reached significance (P = 0.029).
Patient 1 spontaneously reported alleviation of swallowing and increasing tear production during the study and thereafter, yet this was not reflected in our measurements. After study completion, she developed arthritis for the first time and was sufficiently retreated with abatacept. Patient 3 reported absence of former recurrent parotitis during the study and until 19 months after completion of the study. Patient 5 reported amelioration of severe fatigue leading to an increase in working capacity. Patient 6 reported an overall ameliorated well-being specifically regarding fatigue. Patient 8 felt no benefit during the study period despite a marked increase in salivary flow. Remarkably, 6 months after the study, she asked to resume abatacept treatment since, retrospectively, she judged the effect to be clinically relevant. Patient 11 was free of prior severe Raynaud's phenomenon. In 2 of 11 patients, the self-assessment correlated with changes in saliva production and the number of lymphocytic foci (Table 3). Due to the small number of patients, we did not plan to perform a global assessment using the Short Form 36 questionnaire.
|Patient 1||Patient 2||Patient 3||Patient 4||Patient 5||Patient 6||Patient 7||Patient 8||Patient 9||Patient 10||Patient 11|
|Self-assessment||↑ amelioration||Unchanged||↑ amelioration||Unchanged||↑ amelioration||↑ amelioration||Unchanged||Unchanged||Unchanged||Unchanged||↑ amelioration|
No serious adverse events and no infusion reactions were recorded. Patient 2 experienced a transient increase in liver enzymes caused by concomitant rifampin medication for latent tuberculosis. Patient 6 developed diverticulitis that was successfully treated with antibiotic combination therapy. Patient 11 demonstrated lupus-like skin lesions after the third abatacept infusion. These were successfully treated with topical glucocorticoids and a represcription of hydroxychloroquine that had been stopped prior to study enrollment.
This is the first study reporting on treatment with abatacept in patients with primary SS. In addition to the reduction of inflammation within the salivary glands, the data show an increase in saliva production when adjusting for disease duration. This latter finding could indicate that primary SS treatment starting early in the disease course might be of greater benefit. Efforts aiming at early diagnosis should therefore be supported. The abatacept-induced reduction of lymphocytic foci together with locally reduced CD20+ B cells and T cell subsets supports the hypothesis of an important role of T cell costimulation and of T cell–B cell interaction in the pathogenesis of primary SS.
Interestingly, our histologic analysis showed a significant reduction in FoxP3+ Treg cells upon treatment. Large amounts of Treg cells can be found in sites of inflamed tissue such as in rheumatoid synovitis (), thyroiditis (), and salivary glands of primary SS patients ([15, 16]). The increased numbers of Treg cells in autoimmune inflammation appear to represent a biologic effort to counterbalance proinflammatory and potentially destructive disease processes. This interpretation is in line with findings of a positive correlation of Treg cell numbers and severity of cell infiltration in salivary glands ([15, 16]). Recent studies have shown that upon abatacept treatment, B7 molecules on antigen-presenting cells are blocked. Therefore, the CD28-mediated activation and expansion of effector T cells and of Treg cells are hampered; however, the function of Treg cells is enhanced ([17, 18]). The reduced frequency of Treg cells in the salivary glands of our primary SS patients is in accordance with these reports and suggests better immunologic control over disease mechanisms. Based on a recent study, this might even indicate a beneficial effect on potential lymphoma formation ().
At the blood level, we found an unexpected increase in B cells. The total number of B cells increased significantly, and this was caused by an increase in naive B cells. Memory B cells and switched as well as nonswitched B cells did not change in numbers. We recently demonstrated that switched memory B cells but not naive B cells correlate with clinical response to treatment with rituximab in rheumatoid arthritis (). Because abatacept indirectly leads to a reduction of cytokines such as interleukin-6, the mechanism of recruitment of naive B cells as seen in our study remains an open but interesting question.
Immunoglobulin levels slightly decreased, but changes did not reach significance. Respecting the indirect effect of abatacept on B cells and the half-life of immunoglobulins, it is conceivable that the treatment period of 6 months was too short to show an effect.
Despite the significant histologic and cellular effects, the clinical benefit regarding salivary secretion remained minimal in our patients. Interestingly, the increase in saliva secretion was comparable to the data reported for rituximab therapy (). Considering the effective dosing regimen in animals, the dose in our patients probably would have to be increased in order to demonstrate a clearly measurable and clinically relevant effect (). Based on the disease dependency demonstrated in our patient cohort, further studies should aim at a very early treatment start.
As in many other unblinded studies, the self-reported benefit, saliva production, and histologic changes did not show a correlation (). This may be explained by the composition of saliva, but also by a placebo effect. The latter is demonstrated by patient 1, who reported a substantial clinical benefit but showed a marked reduction in saliva secretion.
A weakness of our study is the small number of patients. This was primarily due to the invasive character of our study protocol, with biopsy samples taken before and after treatment. Furthermore, the primary focus of our pilot study was the characterization of histologic and cellular changes and not the detailed measurement of clinical effects. A second issue to be considered is the measurement of stimulated salivary flow by the Saxon test. More sensitive tests and the characterization of saliva composition might have depicted more pronounced differences before and after treatment.
In summary, this is the first study to show that abatacept to treat primary SS leads to a reduction of local glandular inflammation and to a significant increase in saliva secretion when adjusted for disease duration, and that treatment early in the disease course is more effective than at a later stage.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Villiger had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Adler, Caversaccio, Villiger.
Acquisition of data. Adler, Körner, Caversaccio, Villiger.
Analysis and interpretation of data. Adler, Körner, Förger, Huscher, Caversaccio, Villiger.
The authors would like to thank Dr. Michael von Gunten, Pathologie Länggasse, Bern, for participation in evaluating the focus score, and Ms Isabelle Estella, Institute of Pathology, University of Bern, for technical assistance.