Dr Y.-M. Huang, Division of Neuroimmunology, Karolinska Institute, Alfred Nobels Allé 10, 141 83 Stockholm, Sweden. E-mail: email@example.com
Studies in experimental animal models of human autoimmune diseases have revealed that CD4+CD25+ T regulatory (Tr) cells are of thymic origin and have potentials in preventing auto-aggressive immunity. Myasthenia gravis (MG) is the best-characterized autoimmune disease. Changes in the thymus are found in a majority of patients with MG. Thymectomy has beneficial effects on the disease severity and course in a substantial proportion of MG patients. But the occurrence and characteristics of Tr cells have not yet been defined in MG. We determined the frequencies and properties of circulating CD4+CD25+ versus CD4+CD25– cells in MG patients and healthy controls (HCs), with special focus on the effect of thymectomy on CD4+CD25+ cells. CD4+CD25high cells comprise only about 2% of blood lymphocytes in both MG patients and HCs. Frequencies of CD4+CD25high cells were similar in MG patients irrespective of treatment with thymectomy. CD4+CD25+ cells in both MG patients and HCs are mainly memory T cells and are activated to a greater extent than CD4+CD25– cells, as reflected by high levels of CD45RO and human leucocyte antigen (HLA)-DR-positive cells. In both MG patients and HCs, CD4+CD25+ cells also contained a high proportion of CD95-expressing cells as possible evidence of apoptosis-proneness. Upon stimulation with anti-CD3/CD28 monoclonal antibodies, CD4+CD25+ cells responded more vigorously than CD4+CD25– cells in MG, irrespective of treatment with thymectomy, as well as in HCs. Although CD4+CD25– cells are mainly naïve T cells, in non-thymectomized MG patients, they are activated to a greater extent as reflected by higher expression of HLA-DR and CD95 on the surface compared to HCs. The data thus show that there is no deficiency of CD4+CD25+ cells in MG, nor is the proportion of CD4+CD25+ cells influenced by thymectomy.
In the murine immune system, CD4+CD25+ T regulatory (Tr) cells have been shown to act as key controllers of self-reactive T cells and to contribute to the maintenance of immunological self-tolerance [1, 2]. Transfer of CD4+CD25+ cells reduces the pathology in experimental autoimmune diseases such as insulin-dependent diabetes mellitus and in colitis, whereas depletion of CD4+CD25+ cells gives rise to systemic autoimmune diseases [3–5]. In the murine immune system, a majority of mature CD4+CD25+ cells are generated in the thymus in the first few days after birth. Abrogation of the migration of these cells to the periphery from the beginning of their ontogeny results in the development of organ-specific autoimmune diseases [6, 7]. Recent data implicate that, in the human, CD4+CD25+ cells from blood [8, 9] and thymus [10, 11] exhibit in vitro functions similar to murine CD4+CD25+ cells, inhibiting both Th1 and Th2 responses. But significance of CD4+CD25+ in the control of immune responses in humans remains to be established .
Myasthenia gravis (MG) is an autoantibody-mediated autoimmune disorder leading to a reduction in the number of striated muscle acetylcholine receptors (AChRs) that initiate muscle contraction . MG is a prototype autoimmune disease because the target for aggressive autoimmunity, the AChR, is well characterized, which is in contrast to most other autoimmune diseases where the target autoantigen(s) have remained enigmatic. Destruction of AChR results in the abnormal muscle fatigability and weakness that is characteristic of MG. MG is also of particular interest as it is associated in a majority of the patients with changes in the thymus in the form of either hyperplasia or thymoma. The disease responds to treatment with general immuno-suppression as well as to thymectomy in a substantial number of patients [14–16].
Considering the thymic origin and intrinsic anergic and suppressive properties of CD4+CD25+ Tr cells , it is important to determine the frequency and characteristics of this subset in patients with MG. Here, we report on the frequencies of CD4+CD25+ Tr cells among blood lymphocytes in MG patients and healthy controls (HCs). As it was reported that the suppressive activity of CD4+CD25+ Tr cells is mainly through cell-surface contact pathway [18, 19], we analysed the expression of the functionally important cell-surface molecules of CD4+CD25high Tr cells in comparison with CD4+CD25– T cells with special focus on thymectomy therapy.
Materials and methods
Patients and controls. Twenty-three MG patients (12 females) were enrolled in the study. The age range is 18–82 (50 ± 20) years. Fourteen patients were under 40 years of age at MG onset. The diagnosis of MG was based on a typical history, verified by testing muscular fatigability, positive response to cholinesterase inhibitors and a decremental response following repetitive motor nerve stimulation. Two patients had severe generalized MG (stage IIB) and the remaining 21 patients had mild generalized MG (IIA) based on the Ossermann–Oosterhuis scale . Fourteen patients had been thymectomized. Seven of these patients had thymic hyperplasia, six had a thymoma, and one had thymic atrophy. At the time of the study, 10 patients were treated with cholinesterase inhibitors only, nine with azathioprine (100–150 mg/day), one with only cyclosporine, two with azathioprine + cyclosporine (250 mg/day), and one patient received azathioprine + prednisone (7.5 mg/day).
Twenty-one HCs (nine males) aged 24–64 (44 ± 14) years were blood donors or staff from the Department.
Culture medium and antibodies. RPMI 1640 supplemented with 10% heat-inactivated fetal calf serum (FCS), 50 IU/ml penicillin, 50 µg/ml streptomycin and 2 mm glutamine was used for cell culture (all from Gibco, Paisley, Scotland). For activation of T cells, anti-CD3 (UCHT1) and anti-CD28 were purchased from R & D (Minneapolis, MN, USA). For immunostaining, phycoerythrin (PE)-, fluorescein isothiocyanate (FITC)- and peridinin chlorophyll protein (PerCP)-conjugated monoclonal antibodies (MoAbs) against CD4, CD25, CD45RO, CD69, CD80, CD86, CD95, CD152, CD154 and human leucocyte antigen (HLA)-DR were purchased from PharMingen (San Diego, CA, USA).
Cell isolation and culture. Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll density gradient centrifugation. For stimulation, PBMCs at 106/ml were re-suspended in culture medium and cultured in U-bottom-96-well plates in the presence of MoAb against CD3 (10 µg/ml) and CD28 (2 µg/ml) at 37 °C with 5% CO2 for 16 h.
Flow cytometry. Three-colour staining was used to analyse cell-surface molecule expression and intracellular cytokine secretion. Cells were washed twice with 0.5% bovine serum albumin (BSA) in phosphate-buffered saline (PBS). 105 cells/per test were incubated at 4 °C for 20 min with FITC-, PE- or PerCP-conjugated MoAb (5 µg/ml). After washing twice with cold 0.5% BSA in PBS, flow cytometry was performed in a FACS-Calibur instrument using cellquest software (Becton-Dickinson, Stockholm, Sweden). FITC-, PE- and PerCP-conjugated isotype control MoAbs were used.
Statistical analysis. Paired t-test was used to compare continuous variables of CD4+CD25+ and CD4+CD25– subsets. Unpaired t-test was used to compare continuous variables of MG patients and HCs. For multiple comparisons of different MG patient groups and HCs, based on Bartlett's test for homogeneity of variances, anova or nonparametric anova test was chosen. A two-sided P-value <0.05 was considered statistically significant.
Proportions of CD4+CD25+ cells were similar in blood in myasthenia gravis and healthy controls
Less than 10% of lymphocytes detectable in blood were CD4+CD25+, with similar frequencies in MG patients (9 ± 1%) and HCs (8 ± 1%). Blood lymphocytes did not stain strongly for CD25, the interleukin-2 receptor α-chain. Unlike what is seen in the mouse, the CD4+CD25+-cell population in both MG patients and HCs was not highly distinct, but contained different intensities of CD25+ cells. In freshly prepared lymphocytes, (Fig. 1A), a majority of CD4+ cells were CD25–. A large proportion of CD4+CD25+ cells expressed CD25 at low intensities (CD25low), and a very small proportion of CD4+CD25+ cells expressed CD25 at very high intensities (CD25high). Roughly 2% of all blood lymphocytes expressed CD25high, with similar frequencies in MG patients (2 ± 0.4%) and HCs (2.1 ± 0.7%).
Upon stimulation with anti-CD3 and anti-CD28 MoAbs, expression of CD25 was increased among CD4+ cells (Fig. 1B). As shown in Fig. 1(B), CD4+CD25+ cells with low and high intensities were increased, while CD4+CD25– cells were decreased. Activation of lymphocytes by anti-CD3 + anti-CD28 MoAbs resulted in increased proportion of CD4+CD25low cells (16 ± 2% in MG patients, 15 ± 2% in HCs) and of CD4+CD25high (5 ± 2% in MG patients, 4 ± 1% in HCs). There were thus no differences for these variables between MG patients and HCs.
CD4+CD25+ cells comprise of CD45RO, CD95 and HLA-DR-expressing cells
The CD4+CD25– and CD4+CD25+ cells were analysed for expression of the functionally important surface molecules CD45RO, CD69, CD95 and HLA-DR [Fig. 2], in order to gain insight into their action mechanism and to characterize these populations in a prototype autoimmune disease represented by MG. The proportions of cells expressing CD45RO, CD69, CD95 and HLA-DR were compared in the CD25– versus CD25+ subpopulations among CD4+ cells.
The level of cells expressing CD45RO, which is a molecule to identify memory T cells capable of responding to recall antigens, was higher among CD4+CD25+ cells than among CD4+CD25– cells both in MG patients (82 ± 3% versus 54 ± 4%; P < 0.0001) and HCs (82 ± 3% versus 52 ± 4%; P < 0.0001). Most CD25+ cells, in contrast to CD4+CD25– cells, also expressed CD95 (Fas, APO-1), a prototypical death receptor. This was true in both MG patients (75 ± 3% versus 48 ± 4%; P < 0.0001) and HCs (69 ± 3% versus 36 ± 3%; P < 0.0001). Low proportions of HLA-DR-expressing CD25+ cells were observed in both MG patients and HCs. Importantly, more CD25+ cells than CD25– cells expressed HLA-DR in both MG patients (7 ± 1% versus 5 ± 1%; P < 0.01) and HCs (6 ± 1% versus 3 ± 1%; P < 0.001). In both MG patients and HCs, resting CD4+CD25+ and CD4+CD25– cells showed similarly low proportions of cells expressing CD69 that is an early activation marker for T cells. Levels were 5 ± 1% versus 4 ± 1% in MG patients and 6 ± 2% versus 4 ± 1% in HCs.
In summary, CD4+CD25+ cells among blood lymphocytes from MG patients as well as HCs mainly consist of activated and memory T cells with apoptosis-proneness, as reflected by high proportions of cells expressing HLA-DR, CD45RO and CD95 [Fig. 2].
CD4+CD25+ but not CD4+CD25 cells express CTLA-4 on the surface
No clearly detectable cells expressing CD80, CD86 or CD40L were encountered among resting CD4+CD25+ and CD4+CD25– cells in both MG patients and HCs. Low proportions of CD4+CD25+ cells expressing CD152 (CTLA-4), a negative regulator for T-cell costimulation, were found in both MG patients (2 ± 2%) and HCs (2 ± 3%). In contrast, CD152 expression was not detectable on CD4+CD25– cells. This was the case in our MG patients irrespective of treatment with thymectomy, and in the HCs as well.
Upon activation with anti-CD3 plus anti-CD28 MoAbs, CD4+CD25+ but not CD4+CD25– cells were found to contain CD80+ and CD86+ cells (Fig. 3A). This was the case in both MG patients (4 ± 1% for CD80, 6 ± 2% for CD86) and HCs (3 ± 1% for CD80, 4 ± 1% for CD86). Compared to activated CD4+CD25– cells, activated CD4+CD25+ cells also contained higher proportions of cells expressing CD40L (12 ± 3% versus 4 ± 1% in MG patients and 6 ± 6% versus 3 ± 1% in HCs; P < 0.01 for both comparisons) (Fig. 3A), CD69 (63 ± 5% versus 40 ± 4% in MG patients and 61 ± 6% versus 43 ± 5% in HCs; P < 0.001 for both comparisons), CD152 (10 ± 2% versus 4 ± 1% in MG patients and 11 ± 2% versus 4 ± 1% in HCs; P < 0.002 for both comparisons) and HLA-DR (11 ± 3% versus 5 ± 1% in MG patients and 10 ± 2% versus 3 ± 1% in HCs; P < 0.01 for both comparisons) (Fig. 3B). Thus, CD4+CD25+ cells consist of a more substantial number of cells expressing cell-surface molecules that are associated with costimulation and activation upon stimulation. In contrast, CD4+CD25– cells mainly remain naïve as judged by lack of or the occurrence of only a very low proportion of cells expressing costimulatory molecules and HLA-DR.
Levels of CD4+CD25+ cells are similar in MG patients and HCs and not affected by thymectomy.
In mice and rats, it has been established that CD4+CD25+ cells are generated in the thymus . There is some evidence that CD4+CD25+ cells in humans are also generated in the thymus . The aetiology of MG patients is associated with thymic hyperplasia as well as with thymoma. Thymectomy benefits at least younger MG patients with thymic hyperplasia .
We examined whether there is a deficiency of CD4+CD25+ cells in MG patients and whether age of onset, immunosuppressive therapy and, in particular, thymectomy could affect this cell population regarding frequencies and levels of cells expressing functionally important molecules. Upon subgrouping the MG patients according to age at onset of MG, the patients with an onset <40 and those with an onset >40 years of age showed similar frequencies of CD25low (8 ± 4% versus 9 ± 5%) and CD25high (1.9 ± 0.1% versus 2 ± 0.3%) cells among lymphocytes, as well as similar levels of CD4+CD25+ cells expressing CD45RO, CD69, CD95 and HLA-DR. Similarly, patients treated with cholinesterase inhibitors only and patients treated with immunosuppressive drugs showed no differences regarding frequencies of CD25low (9 ± 3% versus 8 ± 4%) and CD25high (2 ± 0.3% versus 1.8 ± 0.2%) cells, and of cells expressing CD45RO, CD69, CD95 or HLA-DR. Furthermore, upon subgrouping regarding treatment with thymectomy, the MG patients showed no differences for CD25low (9 ± 2% versus 8 ± 1%) and for CD25high (2 ± 0.2% versus 1.8 ± 0.3%), nor for levels of cells expressing CD45RO, CD69, CD95 or HLA-DR (data not shown).
The frequencies of CD4+CD25+ cells observed after stimulation with anti-CD3 + anti-CD28 MoAbs, and the levels of CD4+CD25+ cells expressing CD80, CD86, CD40L, CD69, CD152 and HLA-DR were also analysed upon subgrouping the MG patients according to age at onset of MG, immunosuppressive drug therapy and thymectomy. No significant differences were found (data not shown).
CD4+CD25– T cells are activated to a larger extent and more apoptosis-prone in myasthenia gravis
As mentioned above, CD4+CD25– cells mainly remained naïve with low CD95 expression compared to CD4+CD25+ cells, as reflected by low levels of HLA-DR and CD95+ cells. This was observed in both MG patients and HCs. Interestingly, levels of HLA-DR-positive resting CD4+CD25– cells were higher in MG patients compared to HCs (5 ± 0.7% versus 3 ± 0.4%; P = 0.01) (Fig. 3A). Upon subgrouping the MG patients based on age at onset of MG, immunosuppressive drug therapy and thymectomy, non-thymectomized MG patients were found to have higher levels of CD4+CD25– cells expressing HLA-DR compared to thymectomized patients and HCs (7 ± 2% versus 4 ± 1% and 3 ± 0.4%; P = 0.01) (Fig. 3B). Levels of cells expressing CD95 among resting CD4+CD25– cells differed in MG patients (48 ± 4%) versus HCs (36 ± 3%; P = 0.03). Upon subgrouping the MG patients according to age at onset of MG, immunosuppressive drug therapy and thymectomy, no differences in levels of CD95+ resting CD4+CD25– cells were observed versus HCs [Fig. 4]. These findings suggest that CD4+CD25– cells in MG patients may have enhanced susceptibility to CD95-mediated activation-induced cell death.
It has been well established in mice that CD25+ cells of the CD4+-cell population in normal naïve animals bear the ability to prevent autoimmune diseases in vivo, and that these cells in vitro suppress the activation and proliferation of CD4+CD25– T cells upon antigenic simulation . This functionally distinct subpopulation of Tr cells is generated in the normal thymus. It is provided with naturally occurring anergic and suppressive properties . It plays critical roles not only in preventing autoimmune diseases  but also in controlling tumour immunity [22, 23] and transplantation tolerance [24–26]. It remains to be investigated whether a deficiency or abnormality of this population is responsible for or involved in a human autoimmune disease like MG.
In this study, we found that blood lymphocytes from MG patients and HCs have similar frequencies of CD4+CD25+ T cells, constituting <10% of lymphocytes. Upon activation of lymphocytes, CD4+CD25+ T cells were upregulated to a similar extent in the MG patients and the HCs further supporting that there is no deficiency of CD4+CD25+ T cells in MG patients. But the use of CD25 as a marker for Tr cells in humans is complicated by the fact that this molecule also represents an activation marker of effector T cells . This contrasts to the situation in mice in which essentially all CD4+CD25+ cells are Tr cells . It has recently been demonstrated that the bulk of CD4+ T cells among human blood lymphocytes expresses CD25. Of the CD25+ T cells, only a very small population possesses anergic and suppressive properties. This unique population expresses brightest CD25 (CD25high). It comprises 1–2% of human blood lymphocytes [17, 28]. In line with these data, we show that CD4+CD25+ T cells do not constitute a homogeneous population neither in MG patients nor in HCs. A large proportion of CD4+CD25+ T cells shows low intensity CD25 expression, and only around 2% of the lymphocytes show high intensity CD25 expression. Upon activation of lymphocytes, certain CD25– T cells convert into CD25+ cells with low as well as high intensity, with similar frequencies among MG patients and in HCs. Although it is supposed that Tr cells are generated from the thymus, thymectomy in MG patients showed no obvious influence on the frequency of CD25high among CD4+ T cells.
The suppressive action by CD4+CD25+ Tr cells is independent of soluble factors like inhibitory cytokines . Rather, the suppressive action by CD4+CD25+ Tr cells is dependent on cell–cell interactions between CD25+ cells and CD25– T cells, especially via T-cell costimulatory pathways [18, 19]. In the present study, we therefore focused on expression of functionally important surface molecules on CD4+CD25+ cells versus CD4+CD25– cells. We observed that resting CD4+CD25+ cells are mainly memory T cells as reflected by high proportions of cells expressing CD45RO, i.e. a molecule carried by memory T cells and capable of responding to recall antigens. Therefore, when stimulated with anti-CD3 + anti-CD28 MoAbs, resting CD4+CD25+ cells responded vigorously by expressing costimulatory molecules (CD40L, CD80, CD86 and CD152) as well as CD69, an early activation marker on lymphocytes. Furthermore, a majority of CD4+CD25+ cells express CD95, a member of the tumour necrosis factor (TNF) superfamily, which transduces apoptotic signals in a variety of cell types. High expression of CD95 may indicate cell apoptosis-proneness. The apoptosis-prone nature may meet the needs to protect against uncontrolled expansion of anergic and suppressive CD4+CD25+ cells, and to maintain the balance between initiation and downregulation of immune responses. CD4+CD25+ cells from MG patients and HCs showed a similar phenotype, and did not differ when considering age at onset of MG, or ongoing immunotherapy of the disease or, most interestingly, thymectomy.
In contrast to CD4+CD25+ cells, CD4+CD25– cells were mainly naïve T cells with a low proportion of cells expressing CD95. Upon stimulation, CD4+CD25– cells did not express CD80 and CD86, and expressed low CD40L, CD69, CD152 and HLA-DR. Interestingly, CD4+CD25– cells exhibited higher proportion of HLA-DR-positive cells in MG patients compared to HCs, in particular in the non-thymectomized patients. CD4+CD25– cells from MG patients also comprised of higher proportion of CD95+ cells compared to HCs. This may indicate that, in MG patients, CD4+CD25– cells are activated to a larger extent. It is known from a 3-year follow-up study before versus after thymectomy by Lefvert's group that both autoreactive T cells and B cells are present in the thymus in MG patients . T-cell but not B-cell reactivities against AChR decreased after thymectomy which tended to correlate with disease severity. It has recently been demonstrated that deletion of CD4+CD25+ Tr cells is not sufficient to induce autoimmune diseases but that an additional signal is required to activate autoreactive effector CD4+CD25– cells . There are also reports from other autoimmune diseases that frequencies of CD4+CD25+ Tr cells do not change, when comparing proteoglycan-induced arthritis and control naïve mice , and rheumatoid arthritis patients versus healthy subjects . Our recent study on CD4+CD25+ Tr cells in multiple sclerosis, i.e. a disease with proposed auto-aggressive immunity against central nervous system autoantigens, also revealed negative data regarding frequencies and activation of this cell population .
CD4+CD25+ cells may not represent a homogenous cell population. We are aware that up till now there has been a lack of specific key marker(s) for human Tr cells. Novel surface molecules related to the function of CD4+CD25+ have been continuously identified, such as the glucocorticoid-induced TNF receptor, the transcription factor Foxp3 and l-seletin (CD62L) [34, 35]. Subtypes of CD4+CD25+ Tr cells may share phenotype, functional features and mechanisms of action.
Taken together, it seems that CD4+CD25+ Tr cells may not play a critical role neither in controlling MG patients nor being affected by thymectomy, nor being critically involved in some of the other diseases with autoimmune background hitherto studied.
This work was supported by King Gustaf V Foundation, EU grants (QRLT-2000-01918 and QLK3-CT-2001-00225), the Swedish Research Council, Hjärnfonden and Åke Wiberg Foundation.