Impaired salivary gland function in NOD miceAssociation with changes in cytokine profile but not with histopathologic changes in the salivary gland.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 54, No. 7, July 2006, pp 2300–2305 DOI 10.1002/art.21945 © 2006, American College of Rheumatology Impaired Salivary Gland Function in NOD Mice Association With Changes in Cytokine Profile but Not With Histopathologic Changes in the Salivary Gland Malin V. Jonsson, Nicolas Delaleu, Karl A. Brokstad, Ellen Berggreen, and Kathrine Skarstein Objective. To characterize the chronologic disease course and possible interrelationships between salivary gland inflammation, hyposalivation, and cytokine levels in NOD mice, a model for Sjögren’s syndrome (SS). Methods. NOD mice of different ages were used to mimic different disease stages of SS. Histopathologic findings and rates of salivary secretion were compared between 8-week-old, 17-week-old, and 24-week-old female mice. In addition, 10 cytokines were analyzed in serum and saliva obtained from NOD and BALB/c mice. Results. In NOD mice, the salivary flow rate did not change between 8 weeks and 17 weeks of age, while a significant decrease in the salivary flow rate occurred between 17 weeks and 24 weeks of age (P < 0.001). In contrast, significant histopathologic changes in the salivary glands occurred before 17 weeks of age. Chronic inflammatory cell infiltrates were characterized by T and B cell infiltration. Interestingly, in one-third of the mice, proliferating cells were observed in the focal infiltrates. Significant changes in the levels of interleukin-2 (IL-2), IL-5, and granulocyte–macrophage colony-stimulating factor in serum, and in the levels of IL-4 and tumor necrosis factor ␣ (TNF␣) in saliva occurred contemporarily with the decrease in salivary flow. Correlation analyses revealed a negative association between salivary secretion and the levels of IL-4, interferon-␥, and TNF␣ in saliva obtained from NOD mice, while the correlation with inflammatory changes in the glands was consistently weak. Conclusion. Consistent with previous findings, our results indicate at least 2 phases of SS-like disease in NOD mice. Hyposalivation was preceded by inflammatory changes in the salivary glands, whereas abrupt changes in secretion occurred without significant progression of inflammation. Changes in cytokine levels are an indication of the mechanisms involved in the adaptive immune response in the transition from early to overt disease. Sjögren’s syndrome (SS) is an autoimmune disease characterized by oral and ocular dryness. Histopathologically, the disease is manifested by focal lymphocytic infiltrates within the target organ, but a direct association between the degree of lymphoid infiltration and exocrine dysfunction is not always obvious. Inhibition of acinar cell enervation by autoantibodies to the muscarinic M3 receptor (M3R) may inhibit secretion of saliva (1). The NOD mouse is a recognized animal model for the study of SS. The first signs of SS-like disease— mononuclear cell infiltration in the salivary glands— occur at 8 weeks of age and are accompanied by a loss of salivary secretion later in life (2–4). Genetically modified NOD mice such as interleukin-4 knockout (IL-4⫺/⫺) (5), interferon-␥ knockout (IFN␥⫺/⫺), and IFN␥ receptor knockout (IFN␥R⫺/⫺ (6) mice have been characterized in relation to SS-like disease. Despite having glandular infiltration similar to that in NOD/LtJ or NOD.B10-H2b mice, Supported by the Faculty of Odontology, University of Bergen, Bergen, Norway, the L. Meltzer Foundation, the Research Council of Norway, the Norsk Dental Depot Fund for Dental Research, the Broegelmann Foundation, Helse Vest, and the Strategic Research Program at Helse Bergen. Malin V. Jonsson, DMD, Nicolas Delaleu, BSc, Karl A. Brokstad, PhD, Ellen Berggreen, DMD, PhD, Kathrine Skarstein, DMD, PhD: University of Bergen, Bergen, Norway. Address correspondence and reprint requests to Malin V. Jonsson, DMD, Department of Oral Sciences, Oral Pathology and Forensic Odontology, Haukeland University Hospital, N-5021 Bergen, Norway. E-mail: firstname.lastname@example.org. Submitted for publication December 22, 2005; accepted in revised form March 24, 2006. 2300 CYTOKINE PROFILE AND SS-LIKE DISEASE PROGRESSION NOD.B10-H2b.IL-4⫺/⫺ and NOD IL-4⫺/⫺ mice had salivary secretion comparable with that in 4-week-old mice of the same strain (5). NOD IFN␥⫺/⫺ and NOD IFN␥R⫺/⫺ mice did not display SS-like disease features in the salivary glands and retained a normal rate of salivary secretion (6). Various cytokines have been studied in plasma (7), saliva, and salivary gland tissue obtained from patients with SS (8). The aim of this study was to investigate the relationship between lymphoid infiltration and exocrine dysfunction in NOD mice. Ten cytokines were analyzed in saliva and serum to further characterize local and systemic immune reactions. MATERIALS AND METHODS Animals. Female NOD and BALB/c mice were purchased from Taconic (Bomholtgård, Denmark). The animals were maintained under standard animal-housing conditions. Serum glucose levels were measured using the Reflotron Plus glucose test kit (Roche Diagnostics, Laval, Quebec, Canada) and did not differ significantly between or within age groups in NOD or BALB/c mice. The weight of NOD mice did not decrease between 17 weeks and 24 weeks of age. The experimental protocol was approved by the Committee for Research on Animals/Forsøksdyrutvalget (79-04/BBB). Measurement of stimulated salivary flow rate. Prior to stimulation, mice were fasted for a minimum of 5 hours (with water ad libitum) and anesthetized using 0.10 ml of ketamine per 10 grams of body weight. After stimulation of secretion by pilocarpine (Sigma, St. Louis, MO) in saline (0.5 g/gm body weight) via the femoral artery to ensure reliable uptake, saliva was collected with capillary tubes for 10 minutes, and the volume was determined. The samples were stored at ⫺80°C until analyzed. Measurement of cytokines in serum and saliva. Serum and saliva samples obtained from the mice were analyzed using a mouse cytokine Ten-Plex assay kit (catalog no. LMC0001; BioSource, Nivelles, Belgium), as recommended by the manufacturer, measured on a Luminex 100 system (Luminex, Austin, TX), and analyzed using StarStation software (Applied Cytometry Systems, Dinnington, Sheffield, UK). Mean cytokine levels were compared within and between the different age groups. Evaluation of salivary gland inflammation. Submandibular and sublingual salivary glands were surgically removed, snap-frozen in isopentane by liquid nitrogen, and stored at ⫺80°C. Five-m–thick sections were obtained using a cryostat (Leica Instruments, Nussloch, Germany) and placed onto SuperFrost Plus glass slides (Menzel, Braunschweig, Germany). Hematoxylin and eosin (H&E) staining was performed to determine the degree of inflammation. Salivary gland tissue samples were evaluated and morphometrically analyzed using a Leica DMLB light microscope connected to a ColorView III camera and AnalySIS software (Soft Imaging System, Munster, Germany), to determine the focus score (i.e., the number of foci comprising ⱖ50 mononuclear cells/mm2 of glandular 2301 tissue) (9) and the ratio index (i.e., the ratio of the area of inflammation to the total area of glandular tissue) (10). At least 2 tissue sections from both glands were examined, first by H&E staining and then consecutively by immunohistochemical analysis. In the majority of cases, histomorphologic features and the degree of inflammation were similar. In case of inconsistencies, the tissue block was cut down further, and new H&E-stained sections were evaluated. Immunohistochemical analysis. Immunostaining was performed by the avidin–biotin complex method, as described previously (10), using the following antibodies: for T cells (CD4), rat IgG2B,, clone GK1.5 (R&D Systems, Abingdon, UK); for B cells (B220), rat IgG2B, clone RA3-6B2 (R&D Systems); for proliferating cells (Ki-67), rat IgG2a, clone TEC-3 (Dako, Glostrup, Denmark); and for follicular dendritic cells (FDCs), rat IgG2c, clone FDC-M1 (BD Biosciences, San Jose, CA). Briefly, following fixation in cold acetone, endogenous peroxidase (Blocking Kit) and biotin were blocked (Avidin/Biotin Blocking Solution; Vector Laboratories, Burlingame, CA). Nonspecific binding was inhibited by normal rabbit serum. Diaminobenzidine was used as chromogen. Sections were counterstained with hematoxylin. Unless specified otherwise, all reagents were purchased from Dako (Glostrup, Denmark). Statistical analysis. Data were analyzed using one-way analysis of variance followed by the Bonferroni posttest for selected groups (modified unpaired Student’s 2-tailed t-test for multiple group comparisons). To normalize skewed distributions and improve the homogeneity of variance, cytokine data were log-transformed prior to all statistical analyses. To determine the linear relationship between 2 variables, values were compared using Pearson’s correlation test (2-tailed). Correlation analyses were restricted to NOD mice and data sets in which a scientific reason for a causal connection was given. P values less than 0.05 were considered significant. Statistical analyses were performed using GraphPad Prism 4.0 software (San Diego, CA). RESULTS Impaired salivary secretion and histopathologic changes in the salivary glands. The mean salivary flow rate in 24-week-old NOD mice was reduced by ⬃70% compared with that in the 8-week-old NOD mice (P ⬍ 0.001), the 17-week-old NOD mice (P ⬍ 0.001), and the age-matched BALB/c mice (P ⬍ 0.001) (Figure 1). Significant changes were not detected in any of the other groups. Periductal inflammatory cell foci were observed in the submandibular glands in 2 of 6 8-week-old NOD mice and in all of the 17-week-old and 24-week-old NOD mice (Figures 2A–C). Inflammation was assessed by the focus score and the ratio index, and the mean focus scores were 0.1 in the 8-week-old NOD mice, 0.5 in the 17-week-old NOD mice, and 0.7 in the 24-week-old NOD mice (Figure 1). The mean ratio indexes were 0.003 in the 8-week-old NOD mice, 0.019 in the 17- 2302 week-old NOD mice, and 0.039 in the 24-week-old NOD mice (data not shown). An association was observed between the results obtained with both methods (r ⫽ 0.9497, P ⬍ 0.001). Foci comprising ⬍50 mononuclear cells/mm2 of glandular tissue were observed in 3 of the remaining 8-week-old NOD mice. Scattered, nonfocal infiltration was observed in samples obtained from the BALB/c mice and was considered to represent a normal morphologic appearance (Figure 2D). In contrast to the decrease in the rate of salivary secretion that occurred between 17 and 24 weeks of age (P ⬍ 0.001), a significant increase in inflammation was detected when comparing 8-week-old and 17-week-old NOD mice (P ⬍ 0.05) (Figures 1 and 2A and B). A reduction in the salivary flow rate showed only a weak association with inflammation (r ⫽ ⫺0.4688, P ⫽ 0.0497 and r ⫽ ⫺0.5710, P ⫽ 0.0133 for focus score and ratio index, respectively). The most pronounced change in inflammation (focus score and ratio index) was observed between 8-week-old and 24-week-old NOD mice (P ⬍ 0.01), but neither the focus score nor the ratio index increased significantly when comparing 17-week-old and 24-week-old NOD mice (P ⬎ 0.05) (Figure 1). Cellular composition and lymphoid organization. Infiltration by T cells and B cells was observed in focal periductal and perivascular infiltrates. Scattered clusters of T cells and B cells were also observed in close relationship to (and sometimes infiltrating into) ductal and acinar epithelium. Interestingly, in 2 of 6 17-weekold NOD mice and in 2 of 6 24-week-old NOD mice, inflammatory cell infiltrates contained distinct areas of T cells and B cells (Figures 2E and F, respectively), prolifer- Figure 1. Salivary secretion in NOD mice ages 8 weeks, 17 weeks, and 24 weeks, and age-matched BALB/c mice. The salivary flow rate is expressed as microliters of saliva secreted per minute per gram of body weight. The focus score (in NOD mice) represents the number of foci comprising ⱖ50 mononuclear cells/mm2 of glandular tissue. Bars show the mean and SEM for each group. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001. JONSSON ET AL Figure 2. A–D, Hematoxylin and eosin–stained frozen sections of submandibular salivary gland tissue, showing focal inflammation in the 3 groups of NOD mice and lack of focal inflammation in age-matched BALB/c mice. A, Representative section from 2 of 6 8-week-old NOD mice that presented with focal mononuclear cell infiltrates in the submandibular salivary glands. B, Representative section from 17week-old NOD mice, all of which presented with moderate to large chronic inflammatory cell foci within the submandibular salivary glands. C, Representative section from 24-week-old NOD mice, all of which showed infiltration similar to that in 17-week-old NOD mice. D, Focal mononuclear cell infiltration was not observed in frozen submandibular salivary gland tissue sections obtained from BALB/c mice, as represented here by a specimen from a 24-week-old BALB/c mouse. E–H, Immunohistochemistry showing T cell/CD4 cell infiltration (E), B cell/B220 cell infiltration (F), Ki-67⫹ proliferating cells (G), and follicular dendritic cells (H) in tissue from a 17-week-old NOD mouse. Incubations with antibody diluent were performed as negative controls. Salivary gland tissue from the age-matched BALB/c mice served as negative tissue controls, and associated cervical lymph nodes served as positive tissue controls. Controls are not shown. Bars ⫽ 0.1 mm. CYTOKINE PROFILE AND SS-LIKE DISEASE PROGRESSION ating cells (Figure 2G), and FDCs (Figure 2H). The focus score was significantly higher in these mice compared with that in mice that did not present with proliferating cells and FDCs in the salivary glands (P ⬍ 0.05). Changes in cytokine levels in serum and saliva. Mean IL-2 levels were significantly increased in the serum of 24-week-old BALB/c mice compared with 24-week-old NOD mice (P ⬍ 0.05) (Figure 3A). IL-4 was detected in all serum and saliva samples investigated. In serum, the level of IL-4 was similar within age groups and strains of mice. In contrast, a significant increase in the level of IL-4 was detected in saliva from the 24-week-old NOD mice compared with 8-week-old NOD mice (P ⬍ 0.05) (Figure 3B). Correlation analyses showed a negative association between the salivary flow rate and the level of IL-4 in saliva (r ⫽ ⫺0.6899, P ⫽ 0.0015) (Figure 3C), but not with salivary gland inflammation (r ⫽ 0.5957, P ⫽ 0.0091 and r ⫽ 0.5526, P ⫽ 0.0174 for focus score and ratio index, respectively). IL-5 was also detected in all serum and saliva samples investigated. In the NOD mice, serum levels of IL-5 were similar at 8 weeks and 17 weeks of age but decreased significantly between 17 and 24 weeks of age (P ⬍ 0.001) (Figure 3A). A reduction in the level of IL-5 in serum was also observed in 24-week-old BALB/c mice, but the level remained significantly higher than that in 24-week-old NOD mice (P ⬍ 0.001). The levels of granulocyte–macrophage colonystimulating factor (GM-CSF) were significantly increased in 24-week-old NOD mice (P ⬍ 0.001) and in 24-week-old BALB/c mice compared with the levels in 8-week-old mice of the respective strains (P ⬍ 0.05) (Figure 3A). Low levels of IFN␥ were detected in serum and saliva from both NOD and BALB/c mice at 8 and 17 weeks of age. Despite the fact that at 24 weeks of age, IFN␥ was no longer detectable in half of the samples from NOD and BALB/c mice, serum and saliva levels did not differ significantly. The salivary flow rate correlated with IFN␥ levels in saliva (r ⫽ ⫺0.7604, P ⫽ 0.0002) (Figure 3C), whereas an insignificant negative association was observed with increased inflammation (r ⫽ 0.4616, P ⫽ 0.0538 and r ⫽ 0.4662, P ⫽ 0.0511 for focus score and ratio index, respectively). Tumor necrosis factor ␣ (TNF␣) was present in all serum samples. The level of TNF␣ was significantly increased in saliva from 24-week-old NOD mice compared with 8-week-old NOD mice (P ⬍ 0.05) (Figure 3B) and was negatively correlated with the salivary flow rate (r ⫽ ⫺0.7471, P ⫽ 0.0004) (Figure 3C) but with neither the focus score (r ⫽ 0.5563, P ⫽ 0.0165) nor the ratio index (r ⫽ 0.5432, P ⫽ 0.0198). Finally, IL-6, IL-10, 2303 Figure 3. Cytokine levels in serum and saliva samples obtained from NOD and BALB/c mice. A and B, Levels of interleukin-2 (IL-2), IL-5, and granulocyte–macrophage colony-stimulating factor (GM-CSF) in serum (A), and levels of IL-4 and tumor necrosis factor ␣ (TNF␣) in saliva (B). Values are the mean and SEM. C, A decreased salivary flow rate (microliters of saliva secreted per minute per gram of body weight) in NOD mice was associated with increased levels of IL-4, TNF␣, and interferon-␥ (IFN␥) in saliva. Data were analyzed using one-way analysis of variance followed by Bonferroni posttest for comparison of selected groups accounting for multiple group comparisons. N24 ⫽ NOD mice at 24 weeks; B24 ⫽ BALB/c mice at 24 weeks; N08 ⫽ NOD mice at 8 weeks; N17 ⫽ NOD mice at 17 weeks; B08 ⫽ BALB/c mice at 8 weeks; B17 ⫽ BALB/c mice at 17 weeks. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001. and IL-12 were detected in serum and saliva, but the levels of these cytokines did not differ significantly between age groups or strains, and IL-1␤ was detected 2304 JONSSON ET AL only in serum from one 17-week-old and one 24-weekold NOD mouse (data not shown). DISCUSSION Clinical symptoms of primary SS develop late in the disease course; thus, in most patients primary SS is most likely diagnosed at an advanced stage of disease (1). Consistent with results reported by other investigators (2), our results suggest that hyposalivation in NOD mice does not follow the occurrence of focal lymphoid infiltration and thus cannot be explained solely by the destruction or replacement of glandular tissue by inflamed cells (1,4). An alternative murine model for SS, the MRL/lpr mouse, has a normal salivary flow rate despite more pronounced inflammation and several other features of human SS (3,9,10). Results of adoptive transfer experiments indicated that SS is a T cell– mediated autoimmune disease (10), although B cells may contribute to hyposalivation (1,11). In our study, significant changes in the levels of IL-2, IL-5, and GM-CSF in serum and in the levels of IL-4 and TNF␣ in saliva occurred contemporarily with the decrease in the salivary flow rate. However, the NOD mouse strain must be considered a model of general as well as specific immune dysregulation, and the observed changes may, to some extent, represent immune processes unrelated to the SS-like disease (2– 4). In addition, environmental factors as well as other interactions may influence the incidence and extent of autoimmune manifestations of disease. Therefore, caution must be exerted when comparing NOD mice that are acquired from different suppliers and hosted in different environments. IL-2 plays a crucial role in the maintenance of natural immunologic self-tolerance by promoting growth and suppressor functions of regulatory T cells, and neutralization of circulating IL-2 led to the aggravation of diverse autoimmune manifestations (12). Although recent studies have indicated that the absence of IL-5 can induce a shift toward adaptive immune responses, the decrease in the level of IL-5 is difficult to interpret, because little is known about the role of IL-5 in SS. Increased levels of IL-4 in saliva further imply a function of the adaptive immune system. Assuming a connection between salivary glands and saliva, our findings of IL-4 and TNF␣ in saliva from both NOD and BALB/c mice are in accordance with results of previous studies (2). The negative association between IL-4 and salivary secretion is in conformity with findings in NOD.B10-H2b.IL-4⫺/⫺ mice, which experience focal salivary gland inflammation but no loss of salivary secretion, possibly related to the lack of M3R antibodies of an IgG1 isotype (5). GM-CSF is a major regulator of granulocytes and macrophages, and a recent report indicated a role of GM-CSF in inflammation and autoimmunity (13). Unfortunately, little is known about the role of GM-CSF in SS. TNF␣ is one of the cytokines capable of inducing GM-CSF production, and, interestingly, in our study the overt stage of disease in NOD mice was associated with increased levels of TNF␣ in saliva. TNF␣ is considered to be a key cytokine in the pathogenesis of rheumatic disease. Anti-TNF␣ therapies were successfully introduced in the treatment of rheumatoid arthritis, and initial results in patients with SS were encouraging (14). However, subsequent clinical trials failed to confirm these findings (15). In conclusion, our findings indicate at least 2 phases of the SS-like disease manifested in NOD mice. Although major inflammatory cell infiltration is present in the gland, salivary secretion remains unchanged, whereas hyposalivation takes place without significant changes in histopathologic features. The observation of Ki-67⫹ cells in the inflammatory cell infiltrates indicates a role for local proliferation in addition to recruitment of inflamed cells in propagation of the disease. From this point of view, we hypothesize that qualitative changes in immunologic processes such as cytokine expression and lymphoid organization rather than quantitative inflammatory changes modify salivary secretion, and reactions within the chronic inflammatory cell infiltrates deserve special attention. ACKNOWLEDGMENTS We gratefully acknowledge excellent technical assistance by Gunnvor Øijordsbakken, Gudveig Fjell, and Åse Eriksen, and Joanna G. Stormark and the staff at the animal facilities for taking good care of the mice. REFERENCES 1. Jonsson R, Bowman SJ, Gordon TP. Sjogren’s syndrome. In: Koopman WJ, Moorland LW, editors. Arthritis and allied conditions. 15th ed. Philadelphia: Lippincott Williams & Wilkins; 2005. p. 1681–705. 2. Yamano S, Atkinson JC, Baum BJ, Fox PC. Salivary gland cytokine expression in NOD and normal BALB/c mice. Clin Immunol 1999;92:265–75. 3. Humphreys-Beher MG, Brayer J, Cha S, Nagashima H, Diggs S, Peck AB. Immunogenetics of autoimmune exocrinopathy in the NOD mouse: more than meets the eye. Adv Exp Med Biol 2002;506(Pt B):999–1007. 4. Rosignoli F, Roca V, Meiss R, Leceta J, Gomariz RP, Perez Leiros C. Defective signalling in salivary glands precedes the autoimmune CYTOKINE PROFILE AND SS-LIKE DISEASE PROGRESSION 5. 6. 7. 8. 9. 10. response in the non-obese diabetic mouse model of sialadenitis. Clin Exp Immunol 2005;142:411–8. Gao J, Killedar S, Cornelius JG, Nguyen C, Cha S, Peck AB. Sjogren’s syndrome in the NOD mouse model is an interleukin-4 time-dependent, antibody isotype-specific autoimmune disease. J Autoimmun 2006;26:90–103. Cha S, Brayer J, Gao J, Brown V, Killedar S, Yasunari U, et al. A dual role for interferon-␥ in the pathogenesis of Sjogren’s syndrome-like autoimmune exocrinopathy in the nonobese diabetic mouse. Scand J Immunol 2004;60:552–65. Szodoray P, Alex P, Brun JG, Centola M, Jonsson R. Circulating cytokines in primary Sjogren’s syndrome determined by a multiplex cytokine array system. Scand J Immunol 2004;59:592–9. Fox RI, Kang HI, Ando D, Abrams J, Pisa E. Cytokine mRNA expression in salivary gland biopsies of Sjogren’s syndrome. J Immunol 1994;152:5532–9. Jonsson R, Tarkowski A, Backman K, Holmdahl R, Klareskog L. Sialadenitis in the MRL-l mouse: morphological and immunohistochemical characterization of resident and infiltrating cells. Immunology 1987;60:611–6. Skarstein K, Johannessen AC, Holmdahl R, Jonsson R. Effects of 2305 11. 12. 13. 14. 15. sialadenitis after cellular transfer in autoimmune MRL/lpr mice. Clin Immunol Immunopathol 1997;84:177–84. Robinson CP, Brayer J, Yamachika S, Esch TR, Peck AB, Stewart CA, et al. Transfer of human serum IgG to nonobese diabetic Ig null mice reveals a role for autoantibodies in the loss of secretory function of exocrine tissues in Sjogren’s syndrome. Proc Natl Acad Sci U S A 1998;95:7538–43. Setoguchi R, Hori S, Takahashi T, Sakaguchi S. Homeostatic maintenance of natural Foxp3(⫹) CD25(⫹) CD4(⫹) regulatory T cells by interleukin (IL)-2 and induction of autoimmune disease by IL-2 neutralization. J Exp Med 2005;201:723–35. Hamilton JA. GM-CSF in inflammation and autoimmunity. Trends Immunol 2002;23:403–8. Steinfeld SD, Demols P, Appelboom T. Infliximab in primary Sjogren’s syndrome: one-year followup. Arthritis Rheum 2002;46: 3301–3. Mariette X, Ravaud P, Steinfeld S, Baron G, Goetz J, Hachulla E, et al. Inefficacy of infliximab in primary Sjogren’s syndrome: results of the randomized, controlled Trial of Remicade in Primary Sjogren’s Syndrome (TRIPSS). Arthritis Rheum 2004;50:1270–6.