Involvement of CD4+,CD57+ T cells in the disease activity of rheumatoid arthritis

Authors


Abstract

Objective

To evaluate the relationship between the frequency of peripheral CD57+ T cells and the physical status of rheumatoid arthritis (RA) patients, and to perform cytokine analysis of these CD57+ T cells.

Methods

Four-color fluorescence-activated cell sorter analysis was performed to detect both cell surface antigens and intracellular cytokines in peripheral blood leukocytes, using monoclonal antibodies against CD3, CD4, CD8, CD57, interferon-γ (IFNγ), and interleukin-4 (IL-4). RA patients were clinically evaluated with a modified Health Assessment Questionnaire (M-HAQ), joint score, face scale, and visual analog scale (VAS) assessing pain and disease activity.

Results

There was a significant correlation between the frequency of CD4+,CD57+ T cells and erythrocyte sedimentation rate (ESR), whereas a correlation was not found between the frequency of CD8+,CD57+ T cells and ESR. The frequency of CD4+,CD57+ T cells also showed a significant correlation with the mHAQ score, VAS, and face scale. Again, there was no significant correlation between the above-mentioned clinical scores and the frequency of CD8+,CD57+ T cells. Flow cytometric analysis of intracellular cytokines revealed that 14.5% of the CD57+ T cells produced IFNγ, whereas only 2.8% of the CD57+ T cells produced IL-4 in RA patients.

Conclusion

Evidence showing that the frequency of CD4+,CD57+ T cells among CD3+ cells of RA patients had a significant correlation not only with ESR but also with the physical status of the patients, and that a large proportion of the CD4+,CD57+ T cells had the capacity to produce IFNγ, strongly suggests that these CD4+,CD57+ T cells are involved in the immunopathogenesis of RA.

The CD57 (HNK-1 or Leu-7) antigen was initially described as a 110-kd glycoprotein presented by natural killer (NK) cells, which are morphologically known to be large, granular lymphocytes (LGLs) (1). At present, it is known that a small subset of T cells also expresses the CD57 antigen and that these CD57+ T cells are different from true NK (CD3−,CD57+) cells in that the former have low NK activity and are not necessarily LGLs morphologically (1). CD57+ T cells can also be distinguished from conventional NK T cells by the absence of any bias toward using the T cell receptor BV11 and AV24 segments (2). CD57+ T cells frequently lack the CD28 antigen but still possess a strong capacity to proliferate and produce cytokines, possibly through costimulatory pathways other than CD28−,CD80/86 interactions (3, 4). As for killer activity, lectin-dependent cellular cytotoxicity and anti–CD3-induced cytotoxicity are reported to be preferentially mediated by CD57+ T cells (5). Thus, CD57+ T cells have unique characteristics in their function compared with classic CD57− T cells; however, the in vivo role of CD57+ T cells is still unclear.

The notion that there may be some association of CD57+ T cells with the pathophysiology of rheumatoid arthritis (RA) comes from the fact that the number of CD57+ T cells is often increased in the peripheral blood, joint fluid, and joint-adjacent bone marrow of RA patients (6, 7). Moreover, the lymphoproliferative disorder of LGLs expressing both CD3 and CD57 antigens (T-LGL leukemia) is known to be complicated by the symptoms of RA (1). A number of studies have demonstrated oligoclonality of these CD57+ T cells, indicating that the high frequency of the CD57+ T cells in RA patients reflects the in vivo clonal expansion of these cells in response to still-unknown antigens (6). Although these facts suggest that CD57+ T cells are involved in the pathophysiology of RA, there is little information as to how the circulating CD57+ T cells contribute to disease manifestations in RA, and, to our knowledge, there has been no report evaluating the cytokine profile of CD57+ T cells in RA patients. In this study, we demonstrated that the proportion of circulating CD4+,CD57+ T cells, but not CD8+, CD57+ T cells, was correlated positively with the clinical features of RA patients. In addition, a large proportion of CD4+,CD57+ T cells was shown to have the capacity to produce interferon-γ (IFNγ) in RA patients.

PATIENTS AND METHODS

Patient population. Fifty-seven patients with RA (average age 58 years; range 37–81 years) as defined by the American College of Rheumatology (formerly, the American Rheumatism Association) criteria (8) and 15 healthy controls (average age 55 years; range 38–77 years) were evaluated in this study. Measurements of each patient's physical status included the modified Health Assessment Questionnaire (M-HAQ) (9), joint score, patient's assessment of pain on a visual analog scale (VAS; 0–100), patient's global assessment of disease activity on a VAS (0–100), and face scale (1–30).

Flow cytometry and monoclonal antibodies (mAb). Peripheral blood samples obtained from patients with RA and from healthy controls were subjected to hemolysis using the Whole Blood Lysing Reagent Kit (Beckman Coulter, Fullerton, CA). Cell surface antigens of peripheral blood leukocytes (PBLs) were detected using a FACSCalibur flow cytometer with the CELLQuest program (Becton Dickinson, Mountain View, CA). The mAb used were fluorescein isothiocyanate–anti-CD3 mAb, phycoerythrin (PE)–anti-CD4 mAb, peridinin chlorophyll protein–anti-CD8 mAb, and allophycocyanin–anti-CD57 mAb. For intracellular cytokine analysis, PE–anti-IFNγ mAb or PE–anti–interleukin-4 (anti–IL-4) mAb was used. All of these mAb were purchased from Becton Dickinson.

Cell stimulation and staining for intracellular cytokine analysis. We performed intracellular cytokine analysis according to the manufacturer's instructions (PharMingen, San Diego, CA). Briefly, PBLs were stimulated with phorbol myristate acetate and calcium ionophore for 4 hours. For the last 3 hours of stimulation, the intracellular transport inhibitor, nomensin (GolgiStop) was added for accumulation of intracellular cytokines. Afterward, cell surface antigens were stained with fluorochrome-labeled mAb specific for CD3, CD8, and CD57. After Cytofix/Cytoperm solution was added to fix and permeabilize the cells, intracellular cytokines were stained with either PE–anti-IFNγ mAb or PE–anti–IL-4 mAb and 4-color fluorescence-activated cell sorter (FACS) analysis was performed.

Statistical analysis. Statistical correlations were examined using Pearson's correlation analysis. The nonparametric Wilcoxon test was used to compare the frequencies of cytokine-producing cells between CD57+ and CD57− subsets.

RESULTS

Relationship between the frequency of CD4+,CD57+ T cells and erythrocyte sedimentation rate (ESR). Four-color FACS analysis was performed to detect cell surface antigens in PBLs from RA patients and healthy controls, using mAb against CD3, CD4, CD8, and CD57. As has been previously reported (6, 7), the frequencies of CD57+ T cells both in the CD4+ subset and in the CD8+ subset were significantly increased in RA patients (6.0% in CD4 subset, 32.6% in CD8 subset) compared with those in the controls (3.3% in CD4 subset, 20.9% in CD8 subset). Among the laboratory parameters studied, ESR had a significant correlation with the frequency of CD4+,CD57+ T cells (Figure 1). However, the frequency of CD8+,CD57+ T cells had no correlation with ESR.

Figure 1.

Correlation between the erythrocyte sedimentation rate (ESR) and the percent of CD57+,CD4+ cells (left) or CD57+,CD8+ cells (right). The correlation between ESR and CD57+,CD4+ cells was significant.

Relationship between the frequency of CD4+,CD57+ T cells and physical status of RA patients. As shown in Figure 2A, there was a significant correlation between the frequency of CD4+,CD57+ T cells and M-HAQ scores. This correlation was not found with the CD8 subset. Furthermore, scores from either VAS or face scale also had a significant correlation with the frequency of CD4+,CD57+ T cells (Figure 2B). Again, these correlations were not found with the CD8 subset (data not shown). The joint score had no relationship with the frequency of CD57+ T cells either in the CD4 or the CD8 subset (joint score versus CD4+,CD57+ cells r = 0.201, P = 0.135; joint score versus CD8+,CD57+ cells r = 0.111, P = 0.412).

Figure 2.

Correlation between the Modified Health Assessment Questionnaire (M-HAQ) score and the percent of CD4+,CD57+ cells (left) or CD8+,CD57+ cells (right) among CD3+ T cells. The correlation between the M-HAQ score and CD4+,CD57+ cells was significant. B, Correlation between the pain score by visual analog scale (VAS) (left), the patient's global assessment of disease activity by VAS (middle), or the face scale score (right) and the percent of CD4+,CD57+ cells among CD3+ T cells. All 3 correlations were significant. The percent of CD8+,CD57+ cells among CD3+ T cells had no significant relationship with any of these scores (data not shown).

Flow cytometric analysis of intracellular IFNγ and IL-4 production. PBLs from 7 RA patients and 5 healthy controls were examined by flow cytometry for their capacity to produce IFNγ and IL-4. As shown in Figure 3A, a considerable proportion of the CD3+,CD57+ cells (average 14.5%) in the RA patients produced IFNγ, whereas only 2.8% of the CD3+,CD57+ cells produced IL-4. Similar results were obtained in the healthy controls. Thus, circulating CD57+ T cells were considered to belong to the Th1 subset, both in the RA patients and in the healthy controls. IFNγ production by CD57+ T cells in RA patients was further analyzed using gating on either the CD4 or the CD8 subset (Figure 3B). These experiments revealed that, in the CD4 subset, the frequency of IFNγ-positive cells was much higher in the CD57+ T cell subpopulation than in the conventional CD57− T cell subpopulation. In contrast, this difference in the frequency of IFNγ-producing cells between the CD57+ and CD57− subpopulations was not detected in the CD8 subset.

Figure 3.

Substantial production of interferon-γ (IFNγ) but not of interleukin-4 (IL-4) in CD57+ T cells both in patients with rheumatoid arthritis (RA) and in age-matched controls. Peripheral blood leukocytes were stimulated with phorbol myristate acetate and calcium ionophore in the presence of nomensin. After staining of cell surface antigens with fluorochrome-labeled monoclonal antibodies (mAb) specific for CD3, CD8, and CD57, intracellular cytokines were stained with anti-IFNγ or anti–IL-4 mAb. X-axis represents proportions of IFNγ- or IL-4–positive cells among CD3+,CD57+ cells. B, Production of IFNγ in the CD57+ and CD57− T cell subpopulations among the CD4 and CD8 subsets in RA patients. Analysis gate was set on CD8+ or CD8− subset (almost equivalent to the CD4+ subset because double-negative T cells were negligibly low; data not shown), and the proportion of IFNγ-positive cells in each subset is shown. IFNγ production in the CD4+,CD57+ subpopulation was significantly increased compared with that in the CD4+,CD57− subpopulation. Values are the mean and SD.

DISCUSSION

Recent reports that CD57+ T cells were persistently increased in number in RA patients and that RA frequently occurs in patients with T-LGL (phenotypically, CD3+,CD57+) leukemia suggest the involvement of these CD57+ T cells in the pathophysiology of RA (1, 6, 7). However, the role of CD57+ T cells in RA has been controversial. Some authors believe they play a role in activating RA (3, 10, 11), whereas Arai et al have suggested a suppressive role of CD57+ T cells in RA (7). In the present study, we clearly demonstrated that the proportion of CD4+,CD57+ T cells in PBLs was positively correlated with scores assessing the patients' physical status, which include M-HAQ, VAS, and face scale. Together with the data showing a significant correlation between ESR and the proportion of CD4+,CD57+ T cells, this strongly suggested that CD4+,CD57+ T cells are positively associated with the disease activity of RA. Since joint score has no relationship with the proportion of CD4+,CD57+ T cells in PBLs, it is possible that the CD4+,CD57+ T cells contribute to disease activity while having little direct influence on the synovitis itself. In support of this possibility, Martens et al demonstrated a significant correlation between the frequency of CD4+,CD28− T cells, most of which have been reported to have the CD57 surface antigen, and the extraarticular progression of RA (10). It has also been shown that a number of synovial CD4+ T cells have a negative correlation with scores for knee pain (12).

To date, the cytokine profile of CD57+ T cells has been studied both in healthy patients and in patients with various diseases (13, 14). However, to our knowledge, there has been no study evaluating cytokine production by CD57+ T cells in patients with RA. Intracellular cytokine analysis in the present study demonstrated the Th1 cytokine profile of CD57+ T cells in both RA patients and healthy controls. Moreover, within the CD4 subset, the proportion of IFNγ-positive cells was much larger in the CD57+ T cell subpopulation than in the classic CD57− T cell subpopulation. Considering the recent notion that a predominance of Th1 cytokines seems to be of pathogenetic importance in RA (15), the high potential for IFNγ production by CD4+,CD57+ T cells supports the concept that these cells positively control the activity of RA.

However, some previous reports have provided data implying a suppressive role of CD57+ T cells against the inflammation of RA and other diseases, as well as against inflammation arising from normal immune responses (7, 16). In contrast to our findings, Arai et al found an inverse relationship between ESR and the proportion of CD57+ T cells in the joint fluid of RA patients (7). The discrepancy between the findings of Arai et al and our current findings could be attributable to the difference in T cell subsets (CD4/CD8) studied. Arai and colleagues evaluated the proportion of CD3+,CD57+ cells, whereas we subdivided these CD3+,CD57+ T cells into CD4/CD8 subsets and found that only a proportion of the CD4+,CD57+ cells had a positive correlation with ESR. It is possible that the inverse relationship to ESR seen in the CD3+,CD57+ T cell subset may reflect the relationship between the frequency of CD8+,CD57+ T cells and ESR. Although we did not detect such an inverse relationship in the CD8+,CD57+ subset in the PBLs, joint fluid may contain CD57+,CD8+ cells that have a more suppressive function compared with these cells in PBLs. Thus, the difference in the source of T cells studied could be another possible explanation for the observed discrepancy; CD57+ T cells in joint fluid may be different in character from those in PBLs. In fact, we could not detect any association between the proportion of CD4+,CD57+ T cells in PBLs and joint scores.

Other previous reports have suggested the suppressive role of CD8+,CD57+ T cells. CD8+,CD57+ T cells from healthy subjects inhibit pokeweed mitogen–driven B cell differentiation, proliferation, and immunoglobulin production (16, 17). Thus, the functions of CD57+ T cells seem to be divergent, and the difference in the CD4/CD8 subset studied might contribute, at least in part, to the conflicting results reported. CD4+,CD57+ T cells could have a role in the exacerbation of RA, whereas CD8+,CD57+ T cells might suppress the activity of RA. When considering the fact that inherited susceptibility to RA is associated with class II genes encoding certain HLA–DR molecules, it is worth noting that CD4+,CD57+, but not CD8+,CD57+, T cells are associated with both ESR and the clinical status of RA. It is tempting to speculate that a significant proportion of CD4+,CD57+ T cells recognize autoantigens presented with certain major histocompatibility complexes, giving rise to an exacerbation of RA. In support of this hypothesis, Schmidt et al demonstrated that CD4+,CD28− T cells, most of which also express the CD57 surface antigen, showed significant autoreactivity (3). In order to draw some firm conclusion regarding the role of CD4+,CD57+ T cells in RA, a more precise characterization of both the CD57 molecule and CD4+ T cells with this CD57 antigen is needed.

Ancillary