Journal of Neuroimmunology, 42 (1993) 215-222 215 ~, 1993 Elsevier Science Publishers B.V. All rights reserved 0165-5728/93/$06.00

JNI 02301

Human muscle acetylcholine receptor reactive T and B lymphocytes in the peripheral blood of patients with myasthenia gravis

Qing Y i a, Ritva Pirskanen b and Ann Kari Lefvert a

Immunological Research Laboratory and Department of Medicine, Karolinska Hospital, Stockholm, Sweden, and h Department of Neurolo~', South Hospital, Stockholm, Sweden

(Received 9 July 1992) (Revision received 8 September 1992)

(Accepted 8 September 1992)

Ke3' words: Human acetylcholine receptor; T lymphocyte; B lymphocyte; Myasthenia gravls; ELISPOT assay

S u m m a r y

The prevalence of T and B cells reactive with the acetylcholine receptor (AChR) of human skeletal muscle was studied in 33 patients with myasthenia gravis (MG), 18 patients with other neurological diseases (OND) or autoimmune disorders (AD) and 27 age- and sex-matched healthy controls. T cell stimulation was estimated by enumerating cells secreting interferon (IFN)-y and interleukin (IL)-2 in response to the AChR, whereas B cell reactivity was estimated by enumerating cells secreting IgG antibodies binding to the AChR. AChR-reactive T cells were increased in the peripheral blood of patients with MG as compared to patients with OND, AD and healthy individuals. Of the patients with MG, 29/33 (87.7%) had numbers of IFN-y secreting cells higher than the mean _+ 2 SD of the mean of controls as compared to 4,/18 (22.2%) of patients with OND or AD and 2/27 (7.4%) of the controls. The mean value of the numbers of AChR-reactive T cells in the patients with MG was 19.6/105 PBMC, corresponding to 1/5100 PBMC. Comparable results were obtained also for IL-2-secreting cells. Anti-AChR IgG antibody-secreting cells were detected in the blood of 30/33 (91%) of the patients with MG, 3/18 (16.7%) of the patients with OND or AD and 2/25 (8%) of the controls. The mean value of the antibody-secreting cells in MG was 11.7 cells/106 PBMC corresponding to 1/70400 PBMC in the patients with MG, compared to a mean value of antibody-secreting cells in the patients with OND or AD of 0.33 and controls of 0.16 cells/106pBMC.

Introduct ion

The disease myasthenia gravis (MG) is characterized by increased fatigability and weakness of striated mus- cles, caused by a reduction of the amount of functional acetylcholine receptors (AChR) on the postsynaptic endplate of the neuromuscular junction (Drachman, 1981). Antibodies against the AChR are present in > 90% of patients with MG (Lefvert et al., 1978). These antibodies lead to loss of functional receptors by direct blockade of transmitter binding, by cross-linking and consequently increased internalization of receptors (Drachman, 1981) and by complement-mediated lysis

Correspondence to: A.K. Lefvert, Department of Medicine, Karolin- ska Hospital, S-104 01 Stockholm, Sweden.

(Engel, 1980). There is, however, a discrepancy be- tween the presence and the amount of such antibodies and neuromuscular symptoms, both in patients with MG and in other conditions (Lefvert et al., 1978), suggesting that analysis of serum antibodies gives an incomplete picture of the specific repertoire.

The AChR is a pentamer comprising a2, 8 , Y and 6 subunits. About 60% of the AChR antibodies in MG are directed against extracellular sites on the a-sub- units, termed the main immunogenic region (Lind- strom, 1986; Tzartos et al., 1988). The synthesis of anti-AChR antibodies is regulated by AChR-specific T cells (Hohlfeld et al., 1984, 1986). T cells from patients with MG proliferate in response to AChR from Tor- pedo, to recombinant human AChR and to synthetic peptides from the a-subunit (Newsom-Davis et al., 1989a). These T cells react predominantly with the

216

c~-subunit (Hohlfeld et aI., 1987, 19881. AChR-specific T cell clones have been isolated from the thymus and peripheral blood of MG patients (Hohlfeld et al., 1984; Melms et aI., 1988). Such T cells are also present in healthy individuals, although in lower frequency (Sommer et al., 19911.

In this study, PBMC from patients with MG and control groups were analyzed on the single-cell level with regard to reactivity with affinity-purified human skeletal muscle AChR. The results show that T and B cells reactive with human receptor were present in a high percentage of the patients with MG.

Materials and Methods

Patients Thirty-three patients, 22 women and 11 men, with

MG were included in the study. Their age varied between 15-88 years (mean 48 years). The duration of the disease was 1-28 years (mean 9 years). The diagno- sis of MG was based on a typical case history, a positive response to edrophonium, an increased decre- mental response following repetitive nerve stimulation and the presence of antibodies against the AChR of human skeletal muscle measured by radioimmunoassay as described (Lefvert et al., 1978). The clinical evalua- tion was done using the Osserman-Oosterhuis classifi- cation (Oosterhuis, 1964). Four of the 33 patients had ocular myasthenia (stage I), 20 had mild generalized disease (stage IIA), and nine had severe chronic dis- ease (stage IIB). Eighteen patients were thymec- tomized. Examination of thymic tissue showed hyper- plastic changes in 11, normal structure in three and thymoma in four cases. At the time of study, 26 of the patients were treated with cholinesterase inhibitors, 11 with azathioprine and four with corticosteroids.

Eighteen patients (12 female, 6 male) with other neurological diseases (OND) or autoimmune disorders (AD) were included in this study. Their age varied between 21 and 68 years (mean 50 years). Two of the patients had cerebrovascular disease, two multiple scle- rosis, three epilepsy, one cerebrovascular infarction, one headache of unknown origin, two Guillain-Barr~ syndrome, two Wegeners granulomatosis, three sys- temic lupus erythematosus, one autoimmune hemolysis and one Crohn's disease.

Twenty-seven sex- and age-matched healthy individ- uals, 18 women and nine men, were used as healthy controls. Their age varied between 24 and 77 years (mean 47 years).

Preparation of human skeletal muscle AChR Affinity-purified native human AChR was used in

this study. Crude human muscle receptor extract from denervated amputated legs was prepared as described

(Lcfvert et al., 19781. AChR was purified b~ it~ binding to c~-bungarotoxin (c~-BTX) (Sigma T-31119k immobi- lized on the culture plate and used as tile antigen. 1"o determine the specificity of the binding of AChR to the immobilized a-BTX, plates ~ei'c coated with ~'- BTX, washed and incubated with or without 10% of human serum to block the remaining piotein binding sites. After washing, 100 /xl of different dilutions of human muscle extract or muscle extract pre-satured with cold ~-BTX (10 /xg c~-BTX to 1 ml muscle ex- tract, incubated at 37°C for 30 rain) was added and incubated for 1 h at 37°C, followed by incubation with 100 /,tl of [~sI]e~-BTX (NEX-126, Dupont) for 1 h. Radioactivity (CPM) was measured by y-counter (1270 Rackgamma, LKB).

Control samples consisted of toxin without muscle extract, muscle extract without toxin and empty wells without toxin or muscle extract were examined in par- alle[. All tests were done in triplicate, and the tests were repeated three times.

L~olation of PBMC PBMC were isolated by Ficoll-Hypaque density gra-

dient centrifugation of heparinized blood from patients and healthy controls. The cells were washed twice and resuspended in culture medium (RPMI 1640) supple- mented with L-glutamine (4 mM), 1% penicill in-strep- tomycin and 10% heat-inactivated (56°C, 30 min) serum from individuals with blood group ABRh +.

Enumeration of human muscle AChR reactil'e T cells A solid-phase enzyme-linked immunospot (ELI-

SPOT) assay (Czerkinsky et al., 1988) for detection of IFN-y or IL-2 secretion by stimulated T cells at single-cell level was used for enumerat ion of the T cells stimulated by AChR.

Microtiter plastic plates (Nunclon, Nunc, Denmarkl were coated with 100 p.l 0.5 p ,g /ml a -BTX in PBS (pH 7.4), overnight at 4°C, followed by incubation with 10% of human serum for 1 h at 37°C. After washing, 100 ~1 of the human muscle receptor preparat ion diluted 1/20 was added to the wells and incubated for 1 h at 37°C. After washing, 200 p,l of freshly prepared PBMC (2 x l0 s cells/well) were added and incubated for 20 h at 37°C in 5% CO, .

Wells of nitrocellulose bot tomed microtiter plates (Millititer-HAM, Millipore Company, Bedford, MA) were coated with 100 p,l aliquots of mouse monoclonal anti-human IFN-7 or anti-human IL-2 antibody (Genzyme Corporation, Boston, MA) at 200 U / m l (IFN-7) or 10 p ,g /ml (IL-2) at 4°C overnight. Coating solution was removed by suction through a Millititer vacuum filtration holder (Millipore). The plates were washed twice with PBS before the cells, which had been preincubated as described above, were trans- ferred to the coated wells and incubated for another 24

h at 37°C and 5% CO 2. To visualize spots correspond- ing to IFN-y- or IL-2-secreting cells, plates were washed with PBS and incubated with 100 /~I of rabbit poly- clonal anti-human IFN-y (1/400) or anti-human IL-2 (1/100) (Genzyme Corporation) at 37°C for 2 h. After washing, biotinylated anti-rabbit IgG diluted 1/1000 (Vector Lab., Burlingame, CA) was added followed by avidin-biot in-peroxidase complex diluted 1/200 (ABC Vectastain-Elite kit, Vector Lab) for 50 min. After peroxidase staining, using the substrate 3-amino-9-ethyl carbazol (Sigma), spots corresponding to cells that had secreted IFN-7 or IL-2 were enumerated by dissection microscope.

Control samples consisted of wells with toxin with- out addition of muscle ~c,.eptor; muscle receptor with- out toxin; cells incubated in complete medium only, together with purified protein derivate (PPD) (2.5 g g / m l ) and concanavalin A (ConA; 20 g g / m l ) were tested in parallel.

The data were expressed as numbers of spots per 105 PBMC. The coefficient of variation between tripli- cate values was 8.5%.

To determine whether AChR-induced increase in numbers of IFN-y spots was dependent on antigen- presenting cells, PBMC were depleted of monocytes / macrophages by iron-uptake depletion, using 80 mg of iron powder (Sigma) and 10 × 106 PBMC for 30 min at 37°C. Using this method, the depletion of CD14 + cells was more than 99%. PBMC with and without mono- cy tes /macrophages from two MG patients were then tested for the ability of the human receptor to increase the numbers of IFN-y-secreting cells.

Enumeration of anti-human muscle AChR lgG anti- body-secreting cells

An ELISPOT assay (Czerkinsky et al., 1983; M611er et al., 1985) using as coating antigen human muscle receptor bound to a -BTX as described above was used for enumerat ion of the antibody-secreting cells. Wells of nitrocellulose microtiter plates were coated with 0.5 g g / m l a-BTX, blocked with 10% of human serum and followed by 100/xl of muscle receptor diluted 1/15 in PBS (pH 7.4). Blocking with 10% human serum did not interfere with the specific antibody-binding to a -BTX- receptor complex but decreased the non-specific ab- sorption of muscle proteins to the wells. After washing, 200 ~1 of PBMC containing 0.5 x 10 6 cells was added to individual wells and incubated overnight at 37°C. After washing, 100 /xl of diluted high-affinity-purified biotinylated goat anti-human IgG antiserum (Sigma) was added. Washing, incubation with avidin-biot in- peroxidase complex, staining, and counting of spots was performed as described above. Control samples consisting of toxin without addition of muscle receptor and muscle receptor extract without toxin were tested in parallel.

CPM 2000- (A)

, ~1 . . . . . . . . . . . . . . . . . . . . ii . . . . 1/5 1/10 1/20 1/30 1140 1/50 1/60 1170 1180 l/q0 11100 112OO 11300

Dilution of Human Muscle Receptor Extract

217

CPM 2~] (B)

1000

Toxin Toxin only Muscle Exlract No Toxin Pre-~¢al~i Muscle Extract only No Muscle Ex~act Muscle Extract

Control Experiments

Fig. 1. (A) Identification of the binding of AChR from human muscle extract in different dilutions to immobilized a-bungarotoxin visual- ized by [125 I]-a-bungarotoxin. (B) Control experiments comparing the binding of AChR to toxin and non-AChR components from human muscle extract to plates and muscle extract pretreated with cold a-bungarotoxin with (shaded bars) or without blocking (open bars)

with 10% human serum.

Values were given as numbers of spots per 106 PBMC. Variation between quadruplicate cultures was < 10%.

Statistical analysis The Mann-Whitney U-test was used for statistical

comparison of spot numbers obtained from patients and controls, and ordinary analysis of variance for analysis of relation of numbers between IFN-y, IL-2- secreting cells and ant i-AChR IgG antibody-secreting cells in the patient groups.

R e s u l t s

Determination of toxin-AChR binding Figure 1A shows a representative curve of the bind-

ing of human AChR to a-BTX. The effects of blocking with 10% human serum on the unspecific binding of muscle extract to plates is shown in Fig. lB. These results demonstrate that human AChR is bound and

218

J" ,,r,

.2

~ 4O

~r~ 20

Z l 5 I'l~l I .'_~¢, I .'Z(I 1.411 I'3U [.'6(I l.l(lO

Dilution of Human Muscle Receptor Extract

Fig. 2. Numbers of IFN-y-secreting cells/"105 PBMC after incuba- tion with different concentrations of human muscle AC hR extract from a patient ~ith MG (shaded bars) and a healthy control (open

bars).

concentrated by binding to ce-BTX. Control experi- ments using muscle extract pretreated with cold c~-BTX showed no binding of [IeSI]a-BTX (Fig. 1B)

Selection of concentration of ce-BTX-AChR for T cell stimulation

Figure 2 shows the influence of different concentra- tions of the muscle receptor preparation on the num- bers of IFN-T secreting cells in a healthy individual and in a patient with MG. The maximum of IFN-y- secreting cells were obtained using a dilution of muscle receptor preparation of 1 / 1 0 - 1 / 4 0 .

Depletion of monocytes/macrophages from PBMC reduced the stimulatory response to human AChR measured as numbers of IFN-y-secreting cells (Fig. 3).

Spontaneous production of IFN-y attd IL-2 The numbers of spontaneously IFN-y- and IL-2-

secreting cells were determined in cultures with com- plete medium alone. Patients with MG had lower prevalence of both IFN-y- and IL-2-secreting cells than patients with O N D / A D or healthy individuals (Table 1).

-~ 4o

~a

;~. 20

z ~ 10

[ [1

Patients

Fig. 3. Stimulated numbers of IFN-y secreting cells.,' 105 PBMC with (open bar) or without (shaded bar) depletion of m(mocytes/ 'macro-

phages in response to the human AChR in two patients with MG.

Effect of AChR and a-BTX on T cell activation The stimulated numbers of cells, obtained by sub-

traction with the numbers of spots secreted sponta- neously without antigen, are shown in Table 2. There were differences in the stimulated numbers of AChR- reactive cells between the patients with MG and with O N D / A D or controls (P < 0.001) (Fig. 4).

Using a cut-off level of the mean _+ 2 SD (4.0 _+ 6.4) of mean values obtained with cells from the healthy controls, 29/33 (87.8%) of the MG patients, 4 /18 (22.2%) of the O N D / A D patients and 2 /27 (7.4%) of the controls responded with T cell activation to the c~-BTX-receptor complex. The differences between the groups were significant (P < 0.001). The mean value of stimulated AChR-reactive IFN-y-secreting T cells in the patients with MG was 19.6/105 PBMC, corre- sponding to 1/5100 mononuclear cells.

The prevalence of individuals whose ceils responded with IL-2 secretion was 25//31 (81%) in the MG pa- tients as compared to 3 /18 (16.7%) in the O N D / A D patients and 2 /26 (7.7%) in the controls. The mean value of stimulated IL-2-secreting T cells in the patients with MG was 15.3/105 PBMC. corresponding to 1/6500 mononuclear cells.

~1~ 4 0 I

"..a 3 o ,

L

20 ' >% ~J

..~ 10'

z

1 FN-y

I I I I I I

MG

8

_L

Controls OND/AD

,50, IL-2

40,

e

10 ' +

MG Controls OND/AD

Fig. 4. Stimulated numbers of IFN-y- and IL-2-secreting cells,/105 PBMC after incubation with 1/20 dilution of human muscle AChR in patients with myasthenia gravis (MG), healthy controls (controls) and patients with other neurological diseases (OND) / ' au to inmmne disorders (AD).

TABLE 1

Spontaneous IFN-y- and IL-2-secreting cells/105 PBMC from pa- tients with myasthenia gravis (MG), other neurological diseases (OND)/autoimmune disorders (AD) and controls.

Group IFN-y IL-2

MG Range 5-58 5-30 (n = 33) Mean±SD 17.4_+9.8 13.1 _+5.2

P < 0.01/0.01 0.01/0.01

OND/AD Range 15-60 9-50 (n = 18) Mean_+ SD 30.+ 12.0 21.3_+ 10.8

Controls Range 12-55 8-40 (n = 27) Mean_+ SD 29.5_+ 10 .9 19.0_+9.0

P-values refer to difference between MG and OND/AD, and be- tween MG and controls.

219

0

z

3O'

20,

10,

i .L

MG Controls

| I l l

OND/AD

Fig. 5. Numbers of anti-human AChR IgG antibody-producing cells/106 PBMC in patients with myasthenia gravis (MG), healthy controls (controls) and patients with other neurological diseases

(OND)/autoimmune disorders (AD).

In cul tures exposed to a - B T X or muscle extract a lone, the s t imula ted numbers of IFN- 7- or IL-2- secre t ing cells did not differ be tween pa t ien t s with M G and pa t ien t s with O N D / A D or cont ro ls (Table 2).

T h e r e was no cor re la t ion be tween the numbers of cells secre t ing IFN- 7 and IL-2 spon taneous ly in pa t ien t groups or in hea l thy individuals . The re was, however , a cor re la t ion be tween the s t imula ted numbers of IFN-y - and IL-2-secre t ing ceils in pa t ien ts with M G in re- sponse to the r ecep to r ( r = 0.68, P < 0.001).

No cor re la t ion was found be tween clinical var iables or t r e a t m e n t and numbers of cy tok ine-secre t ing cells f rom the pa t ien t s with MG.

Anti-AChR IgG antibody-secreting cells A n t i - A C h R IgG an t ibody-secre t ing cells were de-

t ec ted in pe r i phe ra l b lood from 3 0 / 3 3 (91%) M G pa t ien t s (Fig. 5). The mean value of the ant ibody- secre t ing cells in the M G pa t i en t s ' pe r iphe ra l b lood was 11.7 ce l l s /106 P B M C (range 0-30) , co r r e spond ing to 1 / 7 0 4 2 2 m o n o n u c l e a r ceils. A n t i - A C h R I g G anti-

body-secre t ing cells were also d e t e c t e d in 3 of 18 (16.7%) pa t ien ts with O N D / A D and 2 of 25 (8%) controls , with a mean value of 0.33 cells ( range 0 - 3 ) and 0.16 ce l l s /106 ( range 0 - 2 ) PBMC, respect ively. The d i f ferences be tween the groups were signif icant ( P < 0.001).

No IgG an t ibody-secre t ing B cells were found in cont ro l wells with toxin or muscle extract a lone.

T h e r e was no cor re la t ion be tween cl inical var iables or t r e a tmen t and the n u m b e r of the an t ibody-secre t ing cells. A cor re la t ion was found, however, be tween the numbers of A C h R - s t i m u l a t e d IFN -y - se c r e t i ng cells and the numbers of a n t i - A C h R IgG an t ibody-secre t ing cells in pa t ien t s with M G ( r = 0.726, P = 0.0001, Fig. 6). A cor re la t ion also exis ted be tween numbers of A C h R - s t imula ted IL-2-secre t ing cells and the numbers of a n t i - A C h R IgG an t ibody-secre t ing cells in the pa t i en t group ( r = 0.442, P = 0.003). Lymphocytes from M G pa t ien ts and also E B V - t r a n s f o r m e d B cell l ines that secrete a n t i - A C h R an t ibodies (Lefver t and Holm, 1987)

TABLE 2 Toxin or muscle extract and toxin (T)-AChR-induced IFN-y- and IL-2-secreting cells/105 PBMC from patients with myasthenia gravis (MG). other neurological diseases (OND)/autoimmune disorders (AD) and controls.

Group IFN-y-producing cells IL-2-producing cells

Toxin Muscle T-ACh R Toxin Muscle T-ACh R

MG Range 0-10 0-11 (n = 33) Mean ± SD 2.9 + 2.6 3 + 2.7

P ( < ) ns/ns ns/ns

OND/AD Range 0-9 0-8 (n = 18) Mean + SD 2.3 + 2.2 2.5 ± 2.8

Controls Range 0-10 0-9 (n = 27) Mean ± SD 2.7 + 2.3 2.9 +_ 3.5

4-46 0-6 0-7 4-33 19.3 + 9.8 2.3 + 1.5 3 -+ 8.1 15.3 -+ 6.5

0.001/0.001 ns/ns ns/ns 0.001/0.001

0-15 0-7 0-8 0-12 5.0 _+ 4.6 2.5 _+ 2.3 2.7 + 2.6 3.9 ± 3.8

0-18 0-8 0-7 1-15 4.0+3.2 2.8+2.1 3.1 +3.1 3.8±2.8

P-values refer to difference between MG and OND/AD, and between MG and controls.

__(}

"¢ 16-

=f, I 4-

12-

C~.~ 8-

.<

.&

q.-

i? g

O O

o °

z 0 " _~ " ~'0 " l ; :'0 _,'5 3'o 3', L , 4 ; 50

Number of IFN- T secreting cells Fig. 6. Relationship between the numbers of IFN-y-secreting cells induced by human AChR and the numbers of anti-AChR IgG

antibody-secreting cells in patients with MG.

did not secrete antibodies binding to toxin only (data not shown).

Discussion

This study describes the prevalence of T and B cells reactive with the AChR of human skeletal muscle in the peripheral blood of patients with MG. About 90% of the MG patients had both B cells secreting antibod- ies against human receptor and T cells that were stimulated by receptor, as compared to 17% of the patients with O N D / A D or 8% of the controls. The assays also allowed the calculation of the numbers of specific T and B cells. The requirement of antigen-pre- senting cells such as monocytes/macrophages for the T cell stimulation supports the specificity of this immune response.

The numbers of cells producing IFN-T and IL-2 spontaneously was significantly lower in the patients with MG as compared to the healthy controls. This is in agreement with our previously reported data (,~hl- berg et al., 1991; Yi et al., 1992). The reason for this observation is as yet unknown. Possibly, it might be related to the functional balance between different T cell subsets, e.g. TH1 and TH2 cells. The increased numbers of AChR-stimulated T cells in MG patients as compared to healthy controls, however, was not caused by the lower numbers of cells spontaneously producing cytokines in MG patients. Even when disregarding the difference in spontaneous production of cytokines be- tween the groups, MG patients had higher numbers of T cells stimulated by human AChR (P < 0.01; data not shown).

The use of human muscle AChR in the investigation of the relevant T and B cell reactivities in MG has been hampered by the difficulty of obtaining sufficient quantities of the purified receptor. Regarding T cell stimulation, a few studies have been performed in

which preparations of partially purified truman muscle AChR were used to stimulate lymphocytes from MG patients and controls as measured by [3H]thymidine incorporation (McQuillen ct al., 198.3; Hohlfeld et al., 1984). Most studies have used AChR from a species of the electric ray, Torpedo, or synthetic peptides covering part or the whole sequence of the a-subunit of human AChR (Richman et al., 1976; Conti-Tronconi et al., 1983; Link et al., 1991). However, Torpedo receptor, which shows approximately 75c'~ homology with the human AChR (Noda et al., 1983). has been shown to be unable to stinmlate all potentially reactive T cells (Newsom-Davis et al., 1989b). The cross-reactivity with human anti-receptor antibodies is also limited (Lind- strom et al., 1978; Newsom-Davis et al., 1989a). Thus, Torpedo receptor cannot be considered as an appropri- ate antigen in studies of human MG. Moreover, both T and B cell epitopes are present on subunits on the AChR other than the a-chain (Protti et al., 1991). To get a true picture of the T and B cell repertoire in MG, it was thus considered important to use native human receptor. The use of affinity-purified AChR from human muscle crude extract by its binding to immobi- lized a-BTX is widely employed in ELISA-systems for the detection of serum antibodies against AChR (Hin- man eta[ . , 1983: Jailkhani et al., 1986: Sundewal[ and Lefvert, 1990; Lu et al., 1991). In this study, this system was adopted for the ELISPOT assays of T and B cells. The results demonstrate that by using this system it is possible to obtain sufficient amounts of purified recep- tor to bind secreted antibodies and to stimulate T cells (Figs. 1, 2).

Our use of human receptor is also a likely explana- tion for the higher prevalence of both T and B cells reactive with the receptor than found in other studies (17-67%) employing Torpedo receptor or peptides from the a-subunit (Richman et al., 1976: Newsom-Davis et al., 1989b; Sommer et al., 1991). The low prevalence of T cells stimulated by AChR in healthy individuals confirms the results obtained by others and shows that the normal repertoire contains also T cells specific for the AChR (Richman et al., 1976" Newsom-Davis et al., 1989b; Sommer et al.. 1991).

In this study a new ELISPOT assay for enumeration of IL-2-secreting ceils was developed for detection of stimulated T cells and used in parallel with the IFN-7 assay. Results obtained using the IL-2 assay were com- parable with those of IFN-y-secreting cells. Of impor- tance is the correlation between the stimulated num- bers of IFN-7- and IL-2-secreting cells in response to the AChR. This correlation provides further evidence for the specificity of the receptor-induced stimulation of T cells from MG patients.

For enumeration of anti-AChR IgG antibody-secret- ing cells, an ELISPOT assay was employed. By this assay, anti-AChR IgG antibody-secreting cells were

16

' ~ 12

<

"2 6 .

E 4.

-~ o

-~ - 2 -5

i , , . . . . . . , . o

o

0 0 ~ ~ ' ~ 0 0 o ~/~-o

F " . . . . . . 0 5 10 15 2 0 25 30 35

Number of anti-AChR lgG secreting cells

Fig. 7. Relationship between the numbers of anti-AChR IgG anti- body-secreting cells and the serum levels in units (Lefvert et al.,

1978) of anti-AChR antibodies in patients with MG.

present in 91% of patients with MG and 8% of the controls. The value in the patient group closely corre- sponds to the prevalence of anti-acetylcholine receptor antibodies (r = 0.713, P = 0.0001, Fig. 7). The possibil- ity to calculate the numbers of antibody-producing cells will be of importance in further documenting the repertoire and the immune reactivity in the patients. Of interest is the correlation between the stimulated numbers of cytokine-secreting T cells and the numbers of anti-AChR IgG antibody-secreting B cells (Fig. 6). This correlation might reflect a close collaboration between T and B ceils in MG.

The ELISPOT assay for enumerating IL-2- and IFN-7-secreting ceils is a sensitive assay for the detec- tion of antigen-specific T cell activation. It is superior to the classical proliferation assay in detection of small numbers of stimulated cells and is, in contrast to the latter, also quantitative since it allows the direct visual- ization and counting of stimulated cells in terms of spots (Czerkinsky et al., 1988; Kabilan et al., 1990). ELISPOT assay for IFN-7-secreting cells has been used successfully in the determination and enumera- tion of antigen-specific T cells, such as idiotype- and anti-idiotype-specific T cells in MG (Yi et al., 1992), AChR-reactive T ceils in MG (Link et al., 1991) and myelin basic protein (Olsson et al., 1990) and myelin- oligodendrocyte glycoprotein-reactive T cells in multi- ple sclerosis (Sun et al., 1991). This T cell assay has a high sensitivity and is superior to the proliferation assay in detecting a very low frequency of activated cells. The method also measures an earlier stage of T cell stimulation than the proliferation assay. Thus, al- though it is not possible to directly compare results obtained by the ELISPOT assay and the classical pro- liferation assay, there are both practical and theoreti- cal advantages of the ELISPOT assay. Nevertheless, as ELISPOT assay can only be used to enumerate the numbers of cytokine- or antibody-secreting cells, it is currently not possible to quantitatively detect the

221

amount of cytokines or antibodies secreted by single cells.

In conclusion, this study has shown that PBMC from 90% of patients with MG contain T and B ceils reac- tive with the human AChR. The prevalence of such cells was about 1/5000 PBMC for T ceils and 1 /70 000 PBMC for B cells. These techniques will be of impor- tance in mapping the repertoire and the elucidation of the role of T and B cells in the disease.

Acknowledgements

This work was supported by grants from the Swedish Medical Research Council, the foundations of the Karolinska Institute, the Palle Ferb foundation and the King Gustaf V foundation.

References

Conti-Tronconi, B.M., Scotti, A., Sghirlanzoni, A. and Clementi, F. (1983) Specific involvement of peripheral T lymphocytes against acetylcholine receptors in myasthenia gravis. J. Neurol. Neuro- surg. Psychiatry 46, 832-836.

Czerkinsky, C.C, Nilsson, L..A., Nygren, H., Ouchterlony, I). and Tarkowski, A. (1983) A solid-phase enzyme-linked immunospot (ELISPOT) assay for enumeration of specific antibody-secreting cells. J. Immunol. Methods 65, 109-121.

Czerkinsky, C., Andersson, G., Ekre, H.P., Nilsson, L.,~., Klareskog, L. and Ouchterlony, 0 . (1988) Reverse ELISPOT assay for clonal analysis of cytokine production. I. Enumeration of gamma-inter- feron-secreting cells. J. Immunol. Methods 110, 29-36.

Drachman, D.B. (1981) The biology of myasthenia gravis. Annu. Rev. Neurosci. 4, 195-225.

Engel, A.G. (1980) Morphological and immunopathologic findings in myasthenia gravis and in congenital myasthenic syndromes. J. Neurol. Neurosurg. Psychiatry 43, 577-589.

Hinman, C.L. and Hudson, R:A. (1983) An enzyme-linked im- munosorbent assay for antibody against acetylcholine receptor. J. Neurosci. Methods 9, 141-145.

Hohlfeld, R., Toyka, K.V., Heininger, K., Gross-Wilde, H. and Kalies, I. (1984) Autoimmune human T lymphocytes specific for acetylcholine receptor. Nature 310, 244-246.

Hohlfeld, R., Kalies, I., Kohleisen, B., Heininger, K., Conti-Tronconi, B.M. and Toyka, K.V. (1986) Myasthenia gravis: Stimulation of antireceptor autoantibodies by autoreactive T cell lines. Neurol- ogy 36, 618-621.

Hohlfeld, R., Toyka, K.V., Tzartos, S.J., Carson, W. and Conti- Tronconi, B.M. (1987) HumanT-helper lymphocytes in myasthe- nia gravis recognize the nicotinic receptor a subunit. Proc. Natl. Acad. Sci. USA 84, 5379-5383.

Hohlfeld, R., Toyka, K.V., Miner, L.L., Walgrave, S.L and Conti- Tronconi, B.M. (1988) Amphipathic segment of the nicotinic receptor alpha subunit contains epitopes recognized by T lym- phocytes in myasthenia gravis. J. Clin. Invest. 81,657-660.

Jailkhani, B.L., Asthana, D., Jaffery, N.F., Khajuria, A., Subbalaxmi, B. and Ahuja, G.K. (1986) ELISA for detection of IgG and IgM antibodies to human nicotinic acetylcholine receptor in myasthe- nia gravis. Indian J. Med. Res. 83, 187-195.

Kabilan, L., Andersson, G., Lolli, F., Ekre, H.-P., Olsson, T. and Troye-Blomberg, M. (1990) Detection of intracellular expression and secretion of interferon-7 at the single-cell level after activa-

tion of human T cells with tetanus toxoid m vitro. Eur. J. [mmunol. 20, 1085-1[189.

LeB.,ert, A.K., Bergstr6m, K., Matell, G., Osterman, P.O. and Pirska- nen, R. 11978) Determination of acetylcholine receptor antibody in myasthenia gravis: clinical usefulness and pathogenetic impli- cations. J. Neurol. Neurosurg. PsychiatD, 4l, 394-403.

LeDert, A.K. and Holm, G. 11987) Idiotypic network in myasthenia gravis demonstra ted by human monoclonal B-cell lines. Scand. J. Immunol. 26, 573-578.

Lindstrom, J., Campbell, M. and Nave, B. 119781 Specificities of antibodies to acetylcholine receptors. Muscle Nerve 1, i40-145.

Lindstrom. J. (19861 Probing nicotinic acetylcholine receptors w, ith monoclonal antibodies. Trends Neurosci. 9, 401 407.

Link, H., Olsson, O., Sun, J.-B., Wang, W.-Z., Andersson, G., Ekre, H.-P., Brenner, T., Abramsky, O. and Olsson, T. (1991) Acetyl- choline receptor-reactive T and B cells in myasthenia gravis and controls. J. Clin. Invest. 87, 2191-2196.

Lu, C.-Z, Link, H.. Mo, X.-A., Xiao. B.-G., Zhang, Y.-L. and Zhen, Q. (1991) Anti-presynaptic membrane receptor antibodies in myasthenia gravis, J. Neurol. Sci. 102, 39 45.

McQuillen. D.P., Koethe, S.M. and McQuillen, M.P. 11983) Cellular response to human acetylcholine receptor in patients with myas- thenia gravis. J. Neuroimnlunol. 5, 59-65.

Melms, A., Schalke, B.C.G., Kirchner, Th., Muller-Hermelink, H.K.. Albert. E. and Wekerle. H. (1988) Thymus in myasthenia gravis. Isolation of T-lymphocyte lines specific for the nicotinic acetyl- choline receptor from thymuses of myasthenic patients. J. Clin. Invest. 81 ,902-908.

M6ller, S.A. and Borrebaeck. C A R . 11985) A filter immuno-plaque assay for the detection of antibody-secreting cells in vitro. J. Immunol. Methods 79, t95-204.

Newsom-Davis, J., Harcourt, G., Sommer. N., Beeson, D., Willcox, N., and Rothbard. J.B. (1989a) T cell reactivity in myasthenia gravis. J. Autoimmun. 2 (Suppl.), 101-108.

Newsom-Davis, J., Harcourt. G. and Beeson, D. (1989b) Myasthenia gravis: nicotinic acetylcholine receptors as targets for autoim- mune attack. Biochem. Soc. Trans. 17, 635-637.

Noda, M., Furutani, Y., Takahashi, H., Toyosato, M., Tanabe, T.. Shimizu. S., Kilq,,otani, S., Kayano, T., Hirose, T., Inayama, S. and Numa, S. (10831 Cloning and sequence analysis of calf cDNA

and hunlan genomic DNA encodiug ~-subunil precui>ot ot mu,,- cle acetylcholine receptor. Nature 305, 818 823.

Olsson. T., Zhi, W.-W., H{ijeberg, B.. Kostulas, V.. Jlang. Y.-P.. Andersson, G., Ekre, H.-P. and Link, H. (19901 Autoreactive T I.,,mphocytes in multiple sclerosis determined by antigen-induced secretion of interferon-3,. J. Clin. Invest. 86, 981 985.

Oosterhuis, H.H.G.H. (1964) Studies in myasthenia gravis. Part 1. A clinical stud.~ of 180 patients. J. Neurol. Sci. l, 512 546.

Pmtti, M.P., Manfredi, A.A., Wu, X.D., Moiola, L.. Howard, J.F. and Conti-Tronconi. B.M. 119911 Myasthenia gravis. T epitopes on the ¢5 subunit of ht .man muscle acet,vlcholine receptor. J. lmmunol. 146 . . . . 3-__61.

Richman. D.P., Patrick, J. and Arnason, B.G.W. 119761 Cellular immunity in myasthenia gravis. Response to purified acet_~,l- choline receptor and autologous thymocytes. N. Engl. J. Med. 294, 694-698.

Sommer, N., Harcourt, G.C., Willcox, N., Beeson, D. and Newsom- Davis. J. (1991) Acetylcholine receptor-reactive T lymphocyte.', from health,,' subjects and mya~thenia gravis patients. Neurology 41, 1270 1276.

Sun, J.-B., Link, It., Olsson, T., Xiao, B.-G., Andersson, G., Ekre, H.-P., Linington, C. and Diener, P. 119911T and B cell responses to myelin-oligodendrocyte gbcoprotein in multiple sclerosis. J. lmmunol . 146, 1490-1495.

Sundewall, A.-C. and Lefvert. A.K. (19911 Autoantibody repertoire in p r imal , biliary cirrhosis studied by monoclona/ cell lines. In: Anti-Acetylcholine Receptor Antibodies in Patients With Pri- mary' Billary Cirrhosis and Myasthenia Gravis. Thesis. Karolinska Institutet, Stockholm.

Tzartos, S.J., Kokla, A., Walgrave, S.L. and Conti-Tmnconi , B.M. 119881 Localization of the main immunogenic region of human muscle acetylcholine receptor to residues 67-76 of the a subunit. Proc. Natl. Acad. Sci. USA 85, 2899-2903.

Yi, Q., ,~hlberg. R. and Lefl,ert, A.K. (19921 T cells with specificity for idiotypic determinants on human monoclonal autoantibodies in myasthenia gravis. Res. Immunol. 143, 149-156.

,~hlberg, R.E., Pirskanen, R. and Lefl'ert, A.K. (19911 Defective T lymphocyte function in non-thymectomized patients with myas- thenia gravis. Clin. lmmunol . Immunopathol . 60, 93-105.