T-Lymphocyte Interferon-γ Receptor Binding in Patients With Myasthenia Gravis

Objective  To investigate some aspects of T-lymphocyte–dependent immune function in patients with myasthenia gravis (MG).

Design  Assay interferon-γ binding on T lymphocytes from patients with MG compared with healthy controls.

Setting  The study was performed on ambulatory patients in a tertiary care center, where MG was diagnosed according to Osserman criteria.

Patients  Thirty-six patients with MG (19 women and 17 men; mean±SD age, 50.2±17.6 years) were selected consecutively. They were assigned to groups 1, 2A, and 2B. Ten patients were treated with pyridostigmine bromide alone, 18 were treated with pyridostigmine and corticosteroids, and 8 were not yet treated. Thirty-six age- and sex-matched healthy nonsmoking subjects formed the control group.

Results  A significant (P<.001) decrease of T-lymphocyte interferon-γ binding was found in patients with MG compared with healthy controls (483±14 vs 734±13 receptors [SEM] per cell), whereas the ligand-receptor affinity values [SEM] were similar in the 2 groups (0.9±0.05 and 1.0±0.07 nmol/L).

Conclusion  These data indicate a persistent activation of the immune system in patients with MG, since reduced cell surface interferon-γ receptors seem to be related with enhanced T-lymphocyte immune function.

MYASTHENIA gravis (MG) is a T-lymphocyte–dependent and antibody-mediated autoimmune disease manifested by weakness and fatigue of voluntary muscles.¹ The basic defect observed in patients with MG is a reduction of nicotinic acetylcholine receptors (AChRs) brought about by an autoimmune antibody-mediated attack against the AChRs on the postsynaptic endplate.² Anti-AChR antibodies can be detected in serum samples from 85% to 90% of patients with MG, but the antibody titer does not correlate with disease severity. The mechanisms by which the antibodies may reduce available AChRs include accelerated degradation of AChRs,³ blockade of the ligand-binding site of the AChR molecule,⁴ and complement-mediated damage to the neuromuscular junction.⁵

Cytokines exert an important role in the pathogenesis and pathophysiology of MG by modulating autoantibody production and interfering with cell-mediated immunity.⁶ Cytokines can stimulate or suppress each other, thus producing a complicated network, as well as idiotypic–anti-idiotypic antibodies.

Among cytokines, interferons (IFNs) have a particular relevance. They are a family of proteins classically defined as having antiviral activity; in addition, the IFNs are known to be potent antiproliferative and immunomodulatory agents. Two IFN types have been identified and characterized: type 1 or viral IFNs, including IFN-α (leukocyte-derived) and IFN-β (fibroblast-derived), produced during viral or bacterial infection; and type 2 or immune IFN (IFN-γ), which is produced primarily by T lymphocytes on mitogen or antigen stimulation.

The actions of IFN-γ in the immune system include (1) promoting T-lymphocyte and B-cell proliferation; (2) inducing high-affinity interleukin (IL) 2 receptor expression; (3) generating cytotoxic T-lymphocyte activity; (4) modulating antibody, IL-1, and tumor necrosis factor production; (5) activating macrophages and driving most of their metabolic, secretory, and effector functions; and (6) increasing expression of major histocompatibility complex antigens on the surface of many cell types.⁷

Kott et al⁸ found in subjects with MG high levels of circulating IFN and a spontaneous in vitro IFN production by peripheral blood mononuclear cells (32% of patients), and a markedly deficient natural killer cell activity (73% of patients). These data would reflect a specific stimulation that might contribute to the pathogenesis of the autoimmune response. On the contrary, Confalonieri et al⁹ did not detect IFN-γ in the serum samples of patients with MG and healthy subjects. High numbers of IFN-γ messenger mRNA–expressing blood mononuclear cells,^10,11 further augmented on culture in the presence of AChR^10,12 or anti-AChR and anti-idiotypic antibodies,¹³ were detected in patients with MG compared with subjects with noninflammatory neurologic diseases and healthy controls. Such results parallel those achieved in experimental autoimmune MG, an animal model for human MG,^14,15 suggesting that IFN-γ may be involved in the development of MG and experimental autoimmune MG.

The IFN-γ receptor, the encoding gene of which is located on chromosome 6 and has already been cloned,¹⁶ is a transmembrane glycoprotein of 472 amino acids with an apparent molecular mass of 90 kd, distinct from the receptor protein for type 1 IFNs. The IFN-γ binds a single class of high-affinity receptors with a dissociation constant in the picomolar to nanomolar range, expressed in various human tissues, on the membranes of numerous human cell lines and peripheral blood lymphocytes (PBLs) or monocytes.^17-19 Freshly isolated normal human T lymphocytes activated in vitro with phytohemagglutinin, concanavalin A, or phorbol myristate acetate have been reported to express high- and lower-affinity IFN-γ receptors.²⁰

To our knowledge, no studies have been carried out on PBL IFN-γ receptors in patients with MG, in whom possible changes of lymphocyte biochemistry might be important for immunological reasons. During our research, we assayed T-lymphocyte IFN-γ binding in patients with MG and healthy controls to note some differences between these subject groups, which might be linked with a different immune status.

Thirty-six outpatients with MG at the Department of Neurosciences, Section of Neurology, University of Pisa, Pisa, Italy, were selected consecutively according to Osserman²¹ criteria for MG. There were 17 men and 19 women (mean±SD age, 50.2±17.6 years; range, 21-82 years). Onset ages ranged from 17 to 48 years, and disease duration from 2 to 38 years. Patients with MG were assigned to group 1 (n=7), 2A (n=12), and 2B (n=17) according to the Osserman classification. Illness severity was scored by a quantitative clinical scale.²²

Single-fiber electromyographic results and response to intravenous edrophonium chloride were positive in all patients tested. Nine of 36 patients were negative for antibodies against human AChR by conventional radioimmunoassay (Table 1).

Serum IL-2 and soluble IL-2 receptor levels were used as markers for circulating activated T lymphocytes: for IL-2 and soluble IL-2 receptor immunoenzymatic assays, commercial kits were used (Benfer-Scheller, Milan, Italy).

At the time of observation and blood sampling, 18 patients were receiving pyridostigmine bromide and corticosteroids with (n=6) or without (n=12) immunosuppressive drugs, and 10 were receiving pyridostigmine alone. Eight subjects were "de novo" patients.

The control population consisted of 36 age- and sex-matched healthy subjects selected among blood donors, laboratory personnel, or volunteers aged 18 to 80 years (mean±SD, 49.6±19.3 years).

Lymphocytes were obtained from venous blood of all subjects, as previously described.²³ To get pure T lymphocytes, PBLs were incubated in Dulbecco-modified minimal Eagle medium (Sigma-Aldrich Corporation, Milan) with 10% fetal calf serum (Bio-Whittaker, PBI, Milan) at 4°C for 2 hours in mouse–anti-human IgG-coated Petri dishes. Such a panning procedure was repeated 3 times, so that 98% pure T-lymphocyte suspensions were obtained: T lymphocytes were identified morphologically and as CD3⁺ cells by flow cytometry (Figure 1).

To study IFN-γ binding on T-helper and T-cytotoxic lymphocytes, in a subset of our patients (n=12) and controls (n=10) we incubated lymphocytes freshly separated from blood in Dulbecco-modified minimal Eagle medium with 10% fetal calf serum at 4°C for 2 hours in mouse–anti-human CD4⁺- or anti-human CD8⁺- (both from Sigma-Aldrich Corporation) coated Petri dishes. Through a so-called negative selection, by using anti-CD4 antibodies, we obtained CD8⁺ T lymphocytes; conversely, by means of anti-CD8 antibodies, we obtained CD4⁺ lymphocytes (Figure 1).

Recombinant human IFN-γ was purchased from Sigma-Aldrich Corporation (specific antiviral activity, 1×10⁶ U/mg of protein) and radioiodinated with iodine 125 (¹²⁵I) Bolton-Hunter reagent, as previously described.²³ The initial specific activity of the ¹²⁵I–IFN-γ of different preparations ranged between 0.74 and 1.11 MBq/µg (20-30 µCi/µg), defined as bindable counts per minute per microgram of biologically active IFN-γ in an antiviral assay.²⁴

In standard binding assays, 6×10⁶ T lymphocytes from each subject were incubated in duplicate at 4°C in 750 µL of Dulbecco-modified minimal Eagle medium with HEPES buffer (10 mmol/L, pH 7.4) and 10% fetal calf serum for 2 hours with different amounts of ¹²⁵I–IFN-γ (0.1-0.5 ng).

In competitive binding experiments, increasing amounts of unlabeled IFN-γ were added to standard binding assays. Nonspecific binding was determined by adding in duplicate a 100-fold excess of unlabeled IFN-γ. At the end of incubation, samples were carefully layered over 300 µL of a di-n-butylphthalate–dinonylphthalate (2:1) mixture in microfuge tubes, and spun for 4 minutes at 15,000g at 4°C. Supernatants (containing free ¹²⁵I–IFN-γ) were discarded, and cell pellet radioactivity was counted in a gamma counter (Beckman Instruments, Fullerton, Calif) with 50% efficiency.

Final results about binding parameters were achieved through the Scatchard equation, using the McPherson²⁵ LIGAND program, which determines the maximal receptor number (B_max) and the dissociation constant. A molecular weight of 34 kd, corresponding to the dimeric form of IFN-γ, was used for calculations. Statistical evaluation was performed using the analysis of variance and the Pearson r correlation coefficient. Data are expressed as mean±SEM, unless otherwise indicated.

We found that human T lymphocytes from patients with MG and healthy controls constitutively express high-affinity IFN-γ receptors. The binding of ¹²⁵I–IFN-γ was specific, because only unlabeled IFN-γ inhibited the binding (nearly by 87%), whereas the same amount (100 ng) of IFN-α or IFN-β was ineffective. A representative experimental set of competitive binding of ¹²⁵I–IFN-γ and unlabeled IFN-γ to T lymphocytes is shown in Figure 2. Scatchard analysis of the data yielded a linear plot, representing a single binding site model.²³

We found no significant differences in dissociation constant values between patients with MG and healthy controls (0.9±0.05 vs 1.0±0.07 nmol/L), but highly significant (P<.001) differences in B_max values between the 2 subject groups (Figure 3): 483±14 vs 734±13 receptors per cell. No significant differences in T-lymphocyte IFN-γ receptor density were noted between men and women in patient and control groups, and subject age did not affect B_max values.

By comparing subgroups of patients with MG with each other, we observed no significant differences in mean B_max values among patients treated with pyridostigmine and corticosteroids, those treated with pyridostigmine alone, and those de novo (Table 2), or among patients belonging to subgroups 1, 2A, and 2B. No correlation was found between B_max values and disease severity scores.

CD4⁺ lymphocytes were largely responsible for IFN-γ binding in patients and controls. Significantly lower amounts of IFN-γ receptors on CD4⁺T lymphocytes were assayed from patients with MG (irrespective of the treatment and the disease form or severity) than on CD4⁺ lymphocytes from controls (733±11 vs 1103±15) (Figure 4). CD8⁺ T lymphocytes seemed to have minimal (if any) amounts of IFN-γ receptors in patients and healthy subjects. As for total T-lymphocyte binding assays, CD4⁺T-lymphocyte IFN-γ dissociation constant values did not significantly differ between patients with MG and controls (1.0±0.06 vs 0.9±0.07 nmol/L).

Cytofluorometrically, no significant differences in the relative percentages of CD4⁺ and CD8⁺ lymphocytes were observed among patients in subgroups 1, 2A, and 2B (n=4 for each group) treated with pyridostigmine alone (n=4), pyridostigmine and corticosteroids (n=5), or untreated (n=3). Values ranged from 59% to 66% and from 34% to 40% for CD4⁺ and CD8⁺ cells, respectively.

We found significant differences between patients with MG and healthy controls in serum IL-2 and soluble IL-2 receptor levels (Figure 5 and Figure 6). In patients with MG, no significant differences were observed among the various subgroups of distinct disease forms and treatment types (Table 3).

Linear regression analysis performed on patients with MG showed significant negative correlations between IFN-γ receptor B_max values and serum IL-2 levels (r=−0.83, P<.01), and between ¹²⁵I–IFN-γ binding and soluble IL-2 receptor levels (r=−0.88, P<.01) (Figure 7 and Figure 8).

Interferon γ is a cytokine with pleiotropic effects²⁶; in particular, it is able to modulate the immune network in the central nervous system and systemically. The initial event in the action of IFN-γ is the binding to specific receptors found on different cell types, including PBLs.²⁰ Data from various laboratories suggest that the signals resulting from the binding of IFN-γ to its receptor play an obligate role in T-lymphocyte activation.

Using phytohemagglutinin (5 µg/mL) or concanavalin A (5 ng/mL) as immunostimulants, we found reduced IFN-γ binding in activated T lymphocytes from control subjects compared with resting T lymphocytes: 302±14 vs 716±21 receptors per cell.²³ Such results parallel those of Faltynek and Princler,²⁰ who assayed a 2-fold reduced number of IFN-γ receptors on activated T lymphocytes than on resting T lymphocytes.

We found that patients with MG have a significantly reduced number of T-lymphocyte IFN-γ receptors vs healthy controls. In particular, IFN-γ binding on T-helper CD4⁺ lymphocytes was decreased. Cytotoxic CD8⁺ T lymphocytes from patients and controls did not seem to bear IFN-γ receptors on their surface.

Patients undergoing therapy with corticosteroids did not show significantly different T-lymphocyte IFN-γ binding compared with those who were untreated or treated with pyridostigmine alone. Therefore, according to our results drug treatment does not seem to have any relevant effect on T-lymphocyte IFN-γ binding.

The present data of reduced T-lymphocyte IFN-γ binding in patients with MG parallel previous findings in patients with multiple sclerosis,²³ another neuroimmune disease in which T-lymphocyte activation and cytokine network derangement occur. Therefore, decreased T-lymphocyte IFN-γ binding is not a specific marker for MG.

In conclusion, our findings of decreased IFN-γ receptor B_max values on T-helper lymphocytes note an in vivo CD4⁺ T-lymphocyte activation in MG. It might be that the reduction of IFN-γ receptors is due, at least in part, to receptor down-regulation by endogenously produced IFN-γ, because monoclonal antibodies to IFN-γ have been reported to prevent a large portion of the phytohemagglutinin-induced decrease of IFN-γ B_max values.²⁰

Accepted for publication November 3, 1998.

Reprints: Paolo Bongioanni, MD, PhD, Department of Neurosciences, University of Pisa, 56126 Pisa, Italy (e-mail: bongioanni@sssup1.sssup.it).