Synovial fluid mononuclear cells
CD25–CD4+ Responder T cells from peripheral blood
CD25–CD4+ Responder T cells from synovial fluid
In the homeostasis of the immune system regulatory cells play a major role. Removal of one group of regulatory cells, the CD25+CD4+ T cells, leads to autoimmune manifestations in experimental animal models, and reintroduction of this population prevents disease. This study addresses the role of such regulatory T cells in humans with an autoimmune disease, where we demonstrate the presence of CD25brightCD4+ T cells in the target organ of patients with active rheumatoid arthritis. The patients displayed an enrichment of CD25brightCD4+ T cells in synovial fluid as compared to peripheral blood. These cells are functional regulatory cells, as they were able to suppress in vitro proliferation of autologous T cells, bothfrom synovial and peripheral blood origin. Although the frequency of CD25brightCD4+ T cells varied between patients, it was found to be constant over time in any one joint during each relapse. Numbers were also comparable in two inflamed knee joints of one and the same patient, emphasizing the symmetry of the disease. In summary, it is striking that in addition to all activated, potentially pathological T cells the synovial fluid from RA patients also contains CD25-expressing CD4+ T cells with a regulatory capacity.
Rheumatoid arthritis (RA) is a chronic inflammatory disease primarily affecting peripheral joints. The local inflammation is characterized by a cellular infiltrate, and is often accompanied byan increased volume of synovial fluid. An essential role in this autoimmune process has been ascribed to the local T cells, all of which are memory cells and many carry an activated phenotype 1, 2. One commonly used marker for recently activated T cells is CD25, the IL-2Rα chain. However, expression of CD25 also delineates a population of regulatory CD4+ T cells isolated directly ex vivo without any additional activation 3. In rodents, these CD25+CD4+ regulatory T cells have been found to suppress the in vitro proliferation of T cells (reviewed in 4), and recently the first descriptions of such suppressor cells in healthy humans have been published 5, 6. Experimental in vivo studies have demonstrated that absence of regulatory T cells allows autoimmune manifestations to occur, while the addition of this T cell population can prevent or delay these diseases 7. For example, neonatal thymectomy of mice leads to the development of a wide spectrum of organ-specific autoimmune manifestations such as gastritis, oophoritis, orchitis, and thyroiditis. The transfer of small numbers of CD4+ T cells from normal mice efficiently inhibits all these manifestations 8, 9. Similar autoimmune diseases can be induced by reconstituting immunodeficient nude mice with CD4+ T cells, which have been depleted of CD25+CD4+ T cells 3.
In this study, we analyzed whether functional regulatory CD25+CD4+ T cells exist in the target organ of a human autoimmune disease. We have focused on the question if the synovial fluid of patients with RA contains a significant population of CD25+CD4+ T cells with the potential to suppress the proliferation of autologous CD4+ T cells.
2 Results and discussion
2.1 CD25+CD4+ T cells are enriched in inflamed joints of patients with rheumatoid arthritis
Paired samples of peripheral blood and synovial fluid mononuclear cells from 27 patients with RA were screened by flow cytometry for the presence of CD25+CD4+ T cells. Our results revealed similar frequencies of CD25+CD4+ T cells in peripheral blood of RA patients (median 0.7%, range 0.1–1.8% of all CD4+ T cells) and healthy subjects (median 1.1%, range 0.9–2.3) (Fig. 1A). Statistical analyses indicated no relation between age of subject and frequency of CD25+CD4+ T cells in either of the two study groups. In synovial fluid a significant enrichment of CD4+ T cells expressing CD25 was found as compared to peripheral blood of the same patients (median 7.1%, range 1.8–20) (Fig. 1A). Also, the CD25 staining on the synovial fluid cells was brighter than on cells from peripheral blood, as reflected by a higher mean fluorescence intensity (Fig. 1B). For functional studies, we subdivided the synovial CD25+CD4+ T cells into bright and intermediate populations (CD25bright and CD25int) (Fig. 1C). This is in accordance with the work by Baecher-Allan et al. 10, who demonstrated that the regulatory CD4+ T cells in peripheral blood of healthy human subjects preferentially reside within the CD25highCD4+ T cell population. In addition, Kuniyasu et al. 11 have demonstrated that regulatory T cells have the capacity to express higher levels of CD25 than activated cells. Therefore, in this study the gate for CD25 was deliberately set high for isolation of regulatory T cells as previous studies have clearly demonstrated the presence of activated T cells expressing CD25 in the joint fluid. When a lower gate was used, our measured frequencies of CD25+CD4+ T cells in peripheral blood of both RA patients and healthy subjects were very similar to what has been published by others, 4–14% (data not shown) as compared to 5–15% 12.
When subdividing the CD25 population, we were able to demonstrate that the CD25brightCD4+ T cells did not proliferate in response to anti-CD3 stimulation, while CD25intCD4+ T cells and CD25–CD4+ T cells did (Fig. 1D). The addition of IL-2 increased the proliferation of all three cell populations tested, indicating that the lack of proliferation by CD25brightCD4+ T cells after anti-CD3 stimulation alone was due to the fact that the cells were anergic rather than dying (Fig. 1D). These findings correlate well with previous studies in healthy subjects 12.
2.2 CD25brightCD4+ T cells from inflamed joints are phenotypically similar to CD25brightCD4+ T cells from peripheral blood
A panel of surface markers was selected to compare the phenotype of CD25brightCD4+ T cells from rheumatoid arthritis synovial fluid to corresponding cells from peripheral blood, both from healthy donors and RA patients. The majority of the CD25brightCD4+ T cells were activated memory cells expressing CD45RO, HLA-DR, CD122, CD58 and CD71 (Fig. 2C). This expression pattern resembles that of CD25brightCD4+ T cells from peripheral blood, both from RA patients and healthy donors (Fig. 2C). Differences observed were a higher frequency of CTLA-4-expressing cells, a higher intensity of the CD71 staining, and fewer cells expressing CD62L within the synovial CD25brightCD4+ T cell population. In summary, our data are consistent with previous publications on peripheral blood of healthy subjects 10, 13, 14.
2.3 CD25brightCD4+ T cells from inflamed joints have an immunoregulatory capacity
To investigate whether synovial CD25+CD4+ T cells had a suppressive capacity, six RA patients with a high frequency of synovial CD25brightCD4+ T cells were selected for functional studies (Fig. 3 and Table 1). CD25bright, CD25int and CD25–CD4+ T cells from synovial fluid of each patient were sorted according to the gates in Fig. 3A. A regulatory capacity of CD25+CD4+ T cells is commonly measured in co-cultures as a suppression of proliferation of responder CD25–CD4+ T cells. All patients tested had functional regulatory T cells. In four out of six patients the synovial CD25brightCD4+ T cells suppressed the proliferation of autologous synovial CD4+ responder T cells significantly at a 1:1 ratio of the two cell populations (Fig. 3B). In the other two patients, 50% suppression was achieved when a higher number of CD25brightCD4+ T cells were added to the cultures (Fig. 4A). The degree of suppression could not be related to the percentage of CD25+CD4+ T cells measured directly ex vivo, patient age, disease duration, or treatment (Fig. 3 and Table 1).
When synovial CD25intCD4+ T cells were used as potential regulatory T cells, enhanced instead of suppressed proliferation was measured in the co-cultures from four out of five patients (Fig. 3C). Phenotypically, the CD25intCD4+ T cells were similar to the CD25brightCD4+ T cells, exceptions being lower frequencies of CD62L and HLA-DR-expressing cells (Fig. 2B, C). However, the function differed dramatically. We suggest that the CD25brightCD4+ population is enriched with regulatory T cells in all patients analyzed, while the CD25intCD4+ T cell population holds a majority of "normal" activated T cells with a variable "contamination" of regulatory T cells in the different patients. Only in patient 1 was a suppressive effect of the synovial CD25intCD4+ T cells seen, which suggests a positive correlation between the frequency of CD25bright T cells and the frequency of regulatory T cells in the CD25intCD4+ T cell population. In summary, this emphasizes the importance to differentiate between activated and regulatory CD25+CD4+ T cells on the basis of their intensity of CD25 expression until specific regulatory T cell markers have been identified 10, 11.
To analyze the efficacy of suppression, experiments were performed where variable numbers of CD25brightCD4+ T cells were added to a constant number of responder cells. A dose-dependent suppression was observed in all patients (Fig. 4A, B). From one patient enough synovial fluid cells were obtained to repeat the experiment. The results indicate a reproducible suppression at cell ratios of 1:1 or 1:3 and higher variability in suppression when fewer CD25brightCD4+ T cells were added (Fig. 4B). The dose-dependent suppression compares to those from peripheral blood in healthy subjects, both our own (Fig. 4C) and previously published results 5, 10, 13–15. Due to the low frequency of CD25brightCD4+ T cells in peripheral blood and the anemic state of RA patients, functional studies on peripheral blood were not performed in RA patients. Phenotypically, the responder cells from peripheral blood of healthy donors and RA patients were comparable (Fig. 2A), as were the CD25brightCD4+ T cells (Fig. 2C).
The synovial CD25brightCD4+ T cells were also able to suppress responder cells of peripheral blood origin (Fig. 5). CD25–CD4+ T cells from peripheral blood of two RA patients (patient 1 and 6) were sorted according to the gates shown in Fig. 5A and used as responder T cells. At a 1:1 ratio, autologous CD25brightCD4+ T cells from synovial fluid suppressed the proliferation of peripheral blood responder cells (Fig. 5B). The degree of suppression was similar to that of synovial responder cells in one patient, but lower in the other. These cells differed phenotypically from responder T cells of synovial fluid origin, with higher frequencies of CD62L and CD45RA-positive cells, but lower frequency and intensity of CD122 (Fig. 2A). In one of the patients an experiment was performed, where variable numbers of CD25brightCD4+ T cells were added to a constant number of peripheral blood responder cells (Fig. 5C). The pattern of increased suppressive capacity of CD25brightCD4+ T cells with increasing cell numbers was the same, irrespective if the responder cells were of synovial or peripheral blood origin (see Fig. 4A). This suggests that suppression by regulatory T cells isolated from the joints can affect not only local "target" T cells within the joint, but also T cells from other compartments exhibiting different cellular and cytokine environments. Indeed, CD25+CD4+ regulatory T cells have been shown to act without antigen specificity both in animal models of autoimmunity 4 and tumor rejection 16, as well as in in vitro experiments of healthy subjects 13, 14.
All culture supernatants were screened by ELISA for the cytokines IFN-γ, IL-10 and IL-13. IFN-γ was the only detectable cytokine, and only when IL-2 had been added to the culture (data not shown). This was possibly due to the low numbers of cells used in the cultures.
Comparing joint-derived and circulating cells in humans, it is interesting to note the many similarities despite the fact that the synovial cells are under the influence of a severe inflammation. Since RA is a disease characterized by relapses, it was of interest to investigate if the frequency of regulatory CD25brightCD4+ T cells fluctuated over time. For ethical reasons, the joints were only analyzed at time points when excess synovial fluid was removed for medical reasons. Four patients with RA visited the clinic multiple times during the study period (Table 1). At each time point the frequency of CD25brightCD4+ T cells was measured. Interestingly, these cells did not vary considerably over time in a single joint (Fig. 6). Thus, the affected joint kept its ratio of potentially regulatory cells to total number of CD4+ T cells in each relapse. Three out of four patients had both knee joints inflamed, which gave us the opportunity to compare the frequency of CD25brightCD4+ T cells from two sites of the same patient. To our surprise, the frequencies were comparable between the two knee joints. And from one of these patients, the same frequency of CD25brightCD4+ T cells was also measured in a third joint, an inflamed elbow (data not shown). However, different patients display different levels of CD25brightCD4+ T cells (Fig. 6).
Altogether, our data emphasize that the distinction between CD25+-activated and regulatory T cells must be established for each disease and cell compartment investigated.
|RA patient||Age (years)||Sex (M/F)||Disease duration (years)||RF||DMARD treatment||SR||CRP||% CD25bright of CD4+ T cells||% CD25+ of CD4+ T cells|
|Patient 1b)||35||F||12||–||Methotrexate + infliximab||41||51||11||20|
|Patient 2b, c)||39||F||11||–||Methotrexate||NA||8||3||4|
|Patient 4b)||47||F||8||+||Methotrexate + infliximab||41||78||3||9|
|Patient 5b)||55||F||11||+||Methotrexate + infliximab||71||174||3||7|
|Patient 9c)||23||F||0.5||–||Methotrexate + infliximab||45||63||1||2|
3 Concluding remarks
This is, to the best of our knowledge, the first study to demonstrate the presence and even enrichment of regulatory CD25+CD4+ T cells in the target organ of patients with an autoimmune disease. These cells had the capacity to suppress the proliferation of autologous CD25–CD4+ T cells in vitro. Although the contribution of this cell population to the in vivo situation is difficult to assess, there are several possible scenarios that can explain why, despite the presence of regulatory T cells, chronic inflammation and autoimmune reactions occur in the joint; (i) the presence of regulatory T cells delays or prevents the progression to an erosive inflammation, (ii) the regulatory T cells suppress the immune system, thereby preventing the inflammation to naturally resolve, and thus, perpetuate the chronic inflammation, or (iii) there are, despite an enrichment, not enough regulatory T cells to combat the ongoing inflammation. Extended longitudinal studies of functional changes of these CD25brightCD4+ T cells will help determining their role in the different stages of disease activity as well as in the long-term chronic inflammation. It is also of utmost interest to investigate whether these suppressor cells are also present in the cellular infiltrates in the synovial tissue, since the regulatory function is contact-dependent 17. Immunohistochemical stainings of synovial biopsies indicate that T cells in the infiltrates rarely express CD25 18, suggesting a much lower frequency of CD25brightCD4+ T cells in synovial tissue as compared to synovial fluid. Future studies, including analyses of the frequency as well as the suppressive capacity of synovial tissue CD25brightCD4+ T cells isolated from surgical specimen, will hopefully clarify this issue.
Before definite markers for regulatory CD25brightCD4+ T cells have been defined, it is difficult to compare animal and human studies. The consensus from animal work is that roughly 10% of the naive CD4+ T cell population has a regulatory capacity 3. In humans, the regulatory capacity is contained within the CD25highCD4+ T cell population 10. Despite a significant enrichment in the joint as compared to peripheral blood, such high numbers were only detected in a minority of our patients (Table 1 and Fig. 1). Green et al. 19 recently demonstrated that the local frequency of regulatory CD25+CD4+ T cells in the pancreasdropped to approximately half preceding the onset of diabetes in a transgenic mouse model. The corresponding analysis can not be performed in humans; healthy joints do not even contain T cells. Butthe Green study suggests, as does our own, that it is important to investigate the local target tissue, and one should not expect the frequency in peripheral blood to mirror that in the target tissue.
There are several possibilities to explain the enrichment of CD25+CD4+ regulatory T cells in the joint. Chemokines, preferentially derived from dendritic cells, have recently been described to attract for CD25+CD4+ T cells carrying the chemokine receptors CCR4 and CCR8 20. Interestingly, the presence of CCR4+ T cells has recently been described in the rheumatoid joint 21, 22. Also, mature dendritic cells exist in this compartment (our unpublished data and 23, 24). Secondly, it has been suggested that proinflammatory cytokines, which are abundant in the rheumatoid joint, can up-regulate CD25 expression 25. However, it has been shown by others that up-regulation of CD25 per se is not sufficient to create suppressor cells 11, 14. This indicates that the enrichment of regulatory T cells in the synovial fluid is not due solely to a cytokine rich environment. Thirdly, the notion that IL-10-producing T cells, another population of regulatory cells related to Tr1 cells 26, are best propagated in vitro in the presence of immunosuppressive drugs is interesting 27. Therefore, it is possible that the increased frequency of regulatory T cells in RA patients could depend on the corticosteroid treatment RA patients are receiving. The disease modifying anti-rheumatic drugs that all our patients were treated with could thus eitherdirectly increase the number of regulatory T cells by acting as a growth factor, or indirectly by down-regulating the inflammation. Such a scenario of parallel decrease in inflammation and increasein frequency of regulatory T cells was elegantly demonstrated in a TNF-induced experimental model of diabetes 19. In this model the frequency of pancreatic CD25+CD4+ regulatory T cells was markedly enhanced when the disease was delayed by TNF withdrawal.
Today, many researchers favor the concept that innate immune responses drive the pathology of RA, and that the role of T cells in initiating and perpetuating the chronic inflammation still hasto be substantiated. Our results, showing that regulatory T cells are enriched in the inflamed joint, strengthen the importance of T cells in RA. Possibly, the presence of these suppressor cells could in part account for the cellular hyporeactivity previously reported in RA patients 28, 29. It should be noted that the presence of regulatory T cells is not always beneficial: depletion of the CD25+CD4+ T cells in tumors enabled successful tumor rejection 16.
Finally, the idea that local immunosuppression by regulatory T cells affect disease activity and progression is appealing. It is tempting to speculate that with the enrichment of functional CD25brightCD4+ regulatory T cells in the inflamed joints, apparent in our study, the immune system tries to actively regulate autoimmune reactions and inflammation in the target organ. An increase in the number or function of the CD25brightCD4+ T cells and a parallel decrease in number or function of the pathogenic cells could possibly account for the phases of remission in this chronic disease.
4 Materials and methods
4.1 Sample material
Patients (n=27, age range 23–79 years, median 48 years) with RA (as defined by the American College of Rheumatology criteria) 30 were recruited from the Rheumatology Clinic at the Karolinska Hospital, Stockholm, Sweden. Synovial fluid and peripheral blood samples were obtained from the patients when excess synovial fluid was removed from swollen joints. Peripheral blood samples were also obtained from 7 healthy donors (age range 26–51 years, median 38 years). The investigation was performed after human subject approval from the Karolinska Hospital. Informed consent was obtained from all study subjects. In Table 1 a summary of clinical features of the patients included in the functional and longitudinal studies is provided.
4.2 Cell separation and flow cytometry
Mononuclear cells were prepared from peripheral blood (PB) and synovial fluid (SF) by Ficoll separation (Ficoll-Paque, Pharmacia, Sweden). For phenotypical characterization, cells were stainedwith anti-CD122-PE (PharMingen, USA), anti-CTLA-4-PE (Beckman Coulter, USA), anti-CD62L-FITC (Beckman Coulter), anti-CD45RO-PE (PharMingen), anti-CD45RA-FITC (Becton Dickinson BD, USA), anti-HLA-DR-FITC (BD), anti-CD58-FITC (Beckman Coulter), anti-CD71-PE (Beckman Coulter). As isotype controls IgG1-FITC, IgG2a-PE, IgG1-PE and IgG2a-FITC from BD were used. For frequency determinations cells were stained with anti-CD4-PerCP and anti-CD25-APC (both from BD). The stained cells were analyzed by flow cytometry on a FACSsort (BD). Synovial fluid mononuclear cells (SFMC) were sorted into threedifferent populations; CD25–CD4+ T cells, also referred to as responder cells, R, of SF origin (RSF), CD25intCD4+ T cells (intermediate) and CD25brightCD4+ T cells by using a FACS Vantage SE cell sorter (BD). We also sorted CD25–CD4+ responder T cells from peripheral blood mononuclear cells (PBMC) of patients 1 and 6 (RPB). Additionally, CD25brightCD4+ T cells and CD25–CD4+ T cells from PBMC of healthy donors were sorted. The purity of the sorted populations was determined by FACS reanalysis of an aliquot of sorted cells, and was 90% on average (Fig. 1C). Small, dying and large, activated T cells were excluded from both the sorting and analysis lymphocyte gates.
4.3 Proliferation assays and ELISA
Co-culture experiments were set up with 2×104 autologous APC, 5×103 CD25–CD4+ responder T cells (RPB or RSF), 5×103 CD25intCD4+ or CD25brightCD4+ cells, and, where indicated, rhIL-2 (50 U/ml) in plate-bound anti-CD3-coated wells (OKT-3 1 μg/ml). Autologous PBMC were used as APC after depletion of T cells (Dynabead separation with anti-CD3 beads, Dynal, Norway) followed by irradiation (25 Gy). Cells were incubated at 37°C for 6 days, the last 15–18 h in the presence of [3H]thymidine. Standard sandwich-ELISA were performed to determine the concentrations of IFN-γ, IL-10 and IL-13 in culture supernatants after 5 days of culture. The antibodies used were bought from MABTECH AB, Sweden (IFN-γ) and PharMingen BD, USA (IL-10 and IL-13).
4.4 Statistical analysis
The frequencies of CD25-expressing cells from peripheral blood and synovial fluid were compared by the Wilcoxon Rank test.
We are grateful to Birgitta Wester for excellent cell sorting and to the patients and personnel at the Rheumatology Clinic for providing samples. This work was funded by the Tore Nilsson, Börje Dahlin, and Nanna Svartz' Foundations, the Swedish Association against Rheumatism, the Swedish Medical Association, the King Gustaf the V's 80 year Foundation, and the Swedish Medical Research Council.