- Email: [email protected]
Preferential expression of AChR ε-subunit in thymomas from patients with myasthenia gravis
Journal of Neuroimmunology 201–202 (2008) 28 – 32 www.elsevier.com/locate/jneuroim
Preferential expression of AChR ε-subunit in thymomas from patients with myasthenia gravis Calman A. MacLennan a,b , Angela Vincent a , Alexander Marx c , Nicholas Willcox a , Nils Eric Gilhus d , John Newsom-Davis a , David Beeson a,⁎ a b
c
Neurosciences Group, Weatherall Institute of Molecular Medicine, University of Oxford, Oxford OX3 9DS, UK Medical Research Council Centre for Immune Regulation, Division of Immunity and Infection, Medical School, University of Birmingham, Birmingham, B15 2TT, UK Institute of Pathology, University Hospital Mannheim, University of Heidelberg, D-68125 Mannheim, Germany d Department of Clinical Medicine, Section of Neurology, University of Bergen, Norway Received 17 April 2008; received in revised form 16 June 2008; accepted 17 June 2008
Abstract The role of antigen expression by thymomas in myasthenia gravis (MG) is not clear. Previous reports of acetylcholine receptor (AChR) mRNA expression by the highly sensitive reverse transcription-polymerase chain reactions (RT-PCR) produced varying results. To try to clarify this issue, we first used RT-PCR but then turned to the more accurate and quantitative RNase protection assays (RPA) to assess AChR subunit mRNA expression in thymomas from 25 patients (22 with MG). By RT-PCR, all five AChR subunits could be detected in many thymomas. However, by RPA, the mRNA for the adult-specific AChR ε-subunit was found in 13/25 (52%) thymomas, but not mRNA for the other subunits. AChR ε-subunit was more frequently detected in thymomas of A or AB histology (WHO classification) than those with B1–B3 histology. Overall, 6/6 with thymomas of A or AB histology were positive compared with only 8/19 with B histology (p = 0.02). Autoantibodies in the two patients with the highest levels of ε-subunit mRNA bound better to adult (α2βδε) AChR than to fetal (α2βδγ) AChR, whereas the other sera bound better to fetal AChR. The greater abundance of mRNA for AChR ε-subunit than for other subunits suggests that the AChR ε-subunit may play a distinctive role in autosensitization in MG-associated thymomas, particularly those of type A or AB. © 2008 Published by Elsevier B.V. Keywords: Myasthenia gravis; Thymoma; AChR ε-subunit; RNase protection
1. Introduction In myasthenia gravis (MG), autoantibodies against the native AChR conformation cause loss of AChR at the neuromuscular junction and fatigable muscle weakness (Newsom-Davis et al., 1993). The AChR is a pentameric transmembrane protein, with
Abbreviations: AChR, Muscle acetylcholine receptor; RT-PCR, Reverse transcription-polymerase chain reaction; RPA, RNase protection assay; MG, Myasthenia gravis; TE671-γ, TE671 cell line that expresses fetal AChR; TE671-ε, TE671 cell line transfected to expresses adult AChR. ⁎ Corresponding author. Tel.: +44 1865 222311; fax: +44 1865 222402. E-mail address: [email protected] (D. Beeson). 0165-5728/$ - see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.jneuroim.2008.06.016
two splice forms of the α-subunit, P3A− and P3A+ (Beeson et al., 1990) in humans; during development, the fetal γ-subunit is replaced by the ‘adult’-specific ε-subunit to form the mature AChR (α2,β,δ,ε). Around 10% of MG patients have thymomas (Marx et al., 1997) which are epithelial neoplasms, with varying lymphocytic components, classified as WHO types A, AB, B1–3 (Strobel et al., 2004). It has been suggested that defective negative selection of the thymocytes leads to the export of potentially autoreactive T cells (Marx and Müller-Hermelink, 1999), but the strong association of MG with antibodies to AChR and other muscle antigens suggests active sensitization of T cells to muscle antigens in the tumour (Nagvekar et al., 1998). Indeed, low levels of all five AChR subunit mRNAs have been detected
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
29
2.2. Reverse-transcription PCR (RT-PCR)
using reverse transcription followed by PCR (RT-PCR), or nested PCR (Wilisch et al., 1999). RT-PCR, however, is extremely sensitive and the results may be misleading. We studied 22 MG, and 3 non-MG thymomas with the more accurate and quantitative method of RNAse protection.
Total RNA was prepared by the phenol/guanidinium isothiocyanate method using RNAzol B (Cinna/Biotex Laboratories). RNA quality was assessed by visualizing under UV light 5 μg of each sample run on a formaldehyde/agarose gel, and by the intensity and integrity of bands for β-actin generated by RNase protection and RT-PCR assays. cDNA was synthesized from 10 μg of total RNA. PCR amplifications were performed according to the Perkin-Elmer/Cetus (Roche Molecular Systems Inc.) instructions, as previously described (MacLennan et al., 1997). All amplified products contained sequences from more than one exon, thus excluding by size any products that might derive from contaminating genomic DNA. Product specificity was confirmed by Southern blotting using an internal oligonucleotide as probe.
2. Methods 2.1. Patients Thymomas were obtained at surgery from 22 patients with MG (Oxford and London, UK) and 3 without MG (Bergen, Norway). The diagnosis of MG was based on clinical and electromyographic features and elevated serum AChR antibodies (Table 1A). None of the three patients without clinical MG showed electromyographic abnormalities. Normal thymus was obtained from children undergoing cardio-thoracic surgery. Thymoma tissue was carefully separated from the adjacent thymus and stored at − 80°C. The histology was reviewed by AM (Table 1A). All samples were taken with informed consent and local ethical committee approval.
2.3. RNase protection assays (RPA) RNase protection assays (RPA) were performed essentially as described previously (MacClennan et al., 1997). cDNA
Table 1A Expression of AChR subunits in thymomas and serum antibody reactivity Patient code
MG 14 15 5 13 22 6 11 1 3 4 7 12 16 17 18 19 8 20 2 9 10 21
Age at thymectomy (years M/F)
63 F 46 F 64 M 56 M 47 F 64 F 53 F 50 F 17 M 57 M 59 F 19 M 66 F 48 M 50 M 45 F 23 F 47 M 43 M 35 M 45 F 32 M
No MG 25 58 M 23 37 F 24 66 F
Duration of MG at thymectomy (months)
Thymoma histology
Thymoma RPA
Thymoma RT-PCR
ε-subunit
ε-subunit
α-subunit
AChR antibody reactivity with
Adult AChR/ fetal AChR
γ-subunit
Fetal AChR
Adult AChR
Ratio
11 67 188 3 4 6 45 na 2 13 1 2 4 8 84 92 4 16 3 20 7 2
A⁎ A⁎ AB AB AB/B2 B1 B1 B1/B2 B2 B2 B2 B2 B2 B2 B2 B2† B2/B3⁎ B2/B3⁎ B3⁎ B3⁎ epithelial⁎ na
+++ + ++ + ++ ++ − + + − − − − − − − + + + − + −
+ + + + + + − + + + − − − + + + + + + − + +
+ − + + + + + + + + + + + + + + + + + − + +
+ − − − + + + + − + − − − − + + + − − − − −
5.5 9.6 1.4 na na 59.4 67.1 35.5 9.2 27.3 17.0 43.6 na na na na 36.0 na 81.0 52.0 544.2 na
7.9 3.6 6.7 na na 38.3 44.2 13.3 6.0 19.6 7.7 20.9 na na na na 25.2 na 43.4 30.1 302.5 na
1.44 0.38 4.79 na na 0.64 0.66 0.37 0.65 0.72 0.45 0.48 na na na na 0.7 na 0.54 0.58 0.56 na
na na na
AB B2 B2
+ − −
+ − +
− − +
+ − +
na 0.5 7.1
na 0.2 4.6
na 0.4 0.65
Samples ranked according to WHO histological classification. ⁎Patient on corticosteroid therapy at the time of resection. No further classification possible. †Pleural metastasis of similar appearance as primary tumor. “+” = AChR subunit mRNA detected. “−” = AChR subunit mRNA not detected. na, not available.
30
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
encoding fragments of the extracellular regions of the respective human AChR subunits (Beeson et al., 1993), and cytoplasmic β-actin, were subcloned into pGEM vectors. Anti-sense RNA probes were synthesized in vitro using 50 μCi [α-32P]CTP at N 3000 Ci/mmol (Amersham International) with either T7 or SP6 RNA polymerase. Plasmid DNA was removed by digestion with DNase I for 15 min at 37 °C. Protection assays were performed using the RPA II kit (Ambion Inc.) according to manufacturer's instructions, using 50 μg total RNA for each assay sample. Products were run on a standard 6% polyacrylamide urea sequencing gel and visualized by autoradiography. In order to optimize assay conditions, anti-sense probes were hybridized with calibrated dilutions of AChR subunit cRNA transcribed in vitro. RPAs were performed at least three times for each thymoma sample. 2.4. AChR antibody assays Sera taken around the time of thymectomy were assayed by radioimmunoprecipitation for reactivity with fetal and adult AChR as previously described (Beeson et al., 1996). Titers were expressed as nmol 125 I-α-bungarotoxin-binding sites precipitated per liter of serum.
3. Results Each of the five AChR subunit mRNAs could be detected by RT-PCR in normal thymus samples, as confirmed by Southern blotting and hybridization with an internal subunit-specific oligonucleotide (data not shown). These results confirm previous reports (Bruno et al., 2004) and also demonstrated the presence of AChR δ-subunit mRNA. We studied thymomas from 22 patients with MG and 3 without MG. Since AChR antibodies were present in all sixteen sera available for analysis, including two of those without MG (Table 1A), we pooled all the results for analysis. Expression of AChR α-, ε- and γ-subunits in thymomas by RT-PCR was positive in 21/25 (84%), 19/25 (76%) and 11/25 (44%) samples respectively (Fig. 1B). Both P3A+ and P3A− isoforms of the α-subunit were detected, although the P3A− signals were consistently stronger. AChR β- and δ-subunit mRNAs were also detected in many thymomas (data not shown). To obtain a more accurate measure of the mRNAs, we used RNase protection assays (RPA). None of the AChR subunits were detected by RPA in RNA derived from control thymus (n = 3), or from human liver, lung, brain or from yeast. By contrast, RNA derived from partially denervated muscle gave
Fig. 1. Detection of mRNA encoding β-actin and AChR α-, ε- and γ-subunits in thymoma samples and in control thymus, liver, lung, brain, yeast and skeletal muscle. A. RPA detection of mRNA. Autoradiographs of respective anti-sense probes protected by α-, ε- and γ-mRNA were exposed for 14 days at − 80°C, whereas for β-actin protection assay exposure was for 4 h at room temperature. Undigested anti-sense RNA probe is shown in the right-hand lane. B. RT-PCR amplifications of RNA samples. Products were size fractionated on 1.5% agarose gels and visualized under ultraviolet light after staining with ethidium bromide.
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
31
Table 1B Correlation with thymoma histology and between RPA and RT-PCR results
AChR ε-subunit detected N = 13 AChR ε-subunit not detected N = 12
Median age (range) M:F
WHO A and AB classification
WHO B2 classification
Positivity for AChR subunit by RT-PCR ε-subunit
α-subunit
γ-subunit
49 (17–63) 5:7 ns 49 (19–66) 6:3
6 p = 0.020 0
5 p = 0.047 9
13 (100) p = 0.005 6 (50)
11 (85) ns 10 (83)
6 (46) ns 6 (50)
Data were analysed by Fisher's exact test.
strong RPA bands for all AChR subunits except ε (as predicted as the AChR would be predominantly of fetal type). Remarkably, in thymomas, RPA detected only the ε-subunit mRNA, with 13 (52%) of the 25 samples positive (Fig. 1A). Though variable, band intensities were often comparable with those from muscle. We did not detect signals for the α-, β-, γ- or δ-subunit mRNA by RPA, even in the thymomas positive by RTPCR for AChR α- or γ-, β- or δ-subunit mRNAs. The results are detailed in Table 1A where they are listed according to the thymic pathology, and correlations with histology and RT-PCR results are summarized in Table 1B. All of the 13 thymomas that were positive for ε-subunit mRNA by RPA were also positive by RT-PCR, compared with only 6 of the 12 RPA-negatives (p = 0.005, Fisher's exact test). Interestingly, ε-subunit mRNA expression by RPA was more frequent in tumours classified as WHO type A or AB than in those classified as predominantly B1–3 (Table 1B, Fisher's exact test p = 0.020). Sera were not available on all patients. All of the 11 patients whose thymomas were negative by RPA for AChR εmRNA had lower levels of anti-adult than anti-fetal AChR antibodies, as is commonly seen in non-thymomatous MG. However, sera from two patients with the strongest ε-signals by RPA, #14 and #5, bound more strongly to adult AChR than to fetal AChR. 4. Discussion Despite many demonstrations of the AChR subunits in thymomas from MG patients, it has never been clear whether there is preferential expression of any one subunit, probably because the RT-PCR methods used were too sensitive. We used the RNase protection assay (RPA) which readily detects expression of all AChR subunit mRNAs in human muscle (MacClennan et al., 1997), and after denervation (Fig. 1A), but provides a more quantitative and accurate measure than RTPCR. Surprisingly, we only detected the AChR ε-subunit mRNA; this was found in 12/22 (55%) of thymomas from MG patients and in one of three without MG. Moreover, ε-subunit mRNA expression was more common in thymomas of the A or AB classification than in those with the B classifications. Our results using RT-PCR confirm earlier findings (Wilisch et al., 1999; Bruno et al., 2004) including the presence of both the P3A+ and P3A− isoforms of the α-subunit (Beeson et al., 1990). The lack of detection of the ε-subunit mRNAs in some thymomas, and lack of α-, β-, δ- and γ-subunit mRNAs by RPA, even in samples positive by RT-PCR, suggests that the levels of these subunits are relatively low in all cases.
The ε-subunit mRNAs that we detected by RPA are very unlikely to have derived from myoid cells, which are rare in most thymomas (Rosai and Levine, 1976). Our failure to detect ε by RPA in normal thymus, which contains myoid cells that express fetal AChR (Schluep et al., 1987), is consistent with this and implies, therefore, over-expression in thymomas. Conversely, the α-, β-, δ- and γ-AChR subunits that we detected in thymoma samples by RT-PCR might well have derived from myoid cells in adjacent normal thymic tissue (Leite et al., 2007). Thymomas show strong associations with autoimmune disorders (Buckley et al., 2001) and generate and export T cells (Muller-Hermelink and Marx, 2000). In theory, the tumors could either be directly selecting/sensitizing T cells against selfantigens that they express (Nagvekar et al., 1998) or failing to tolerize against antigens that they do not express (Marx and Muller-Hermelink, 2005). The over-expression of the ε-subunit in thymomas argues in favor of the former, as does the preference for adult AChR shown by serum antibodies from the two patients with the highest levels of ε-subunit mRNA. In the other five, the more prevalent antibodies against fetal AChR, which are common in MG, might be the end-result of determinant spreading (Vincent et al., 1998) initiated by earlier responses to the ε-subunit. Indeed, several features of the AChR ε distinguish it from the other subunits and suggest a special potential for breaking tolerance in MG. Firstly, transcription of the ε-subunit is under tight regulatory control, and is normally restricted to the subsynaptic nuclei of muscle fibers (Sanes and Lichtman, 2001), suggesting that aberrant over-expression in thymomas could be immunologically ‘dangerous'. Secondly, since its expression is so minimal in the normal thymus, it may tolerize incompletely with the result that patients would have some circulating ε-subunit-specific T cells that could colonize the thymoma. Thirdly, as the ε-subunit appears to be produced in excess over the other AChR subunits in thymomas, it is much more likely to be degraded rapidly than to be incorporated into whole AChR molecules, with the result that the resulting peptides would be available for presentation to developing T cells, either by the neoplastic epithelial cells (Gilhus et al., 1995) or the abundant adjacent dendritic cells. It has already been shown that the ε-subunit includes at least one important helper T cell epitope (Hill et al., 1999) and T cells from MG patients have been found to recognize epitopes unique to the ε-subunit (Ragheb et al., 2005). The autoimmune regulator (AIRE) plays a crucial role in driving the expression of self-antigens in medullary thymic epithelial cells. It has been shown to play a role in controlling thymic expression of the AChR α-subunit from the CHRNA locus, and possibly setting the threshold for self-tolerance
32
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
versus autoimmunity for the AChR (Giraud et al., 2007). However, AIRE expression is down-regulated in human thymoma. Whereas the presence of AIRE is responsible for activating a number of loci, and its loss correspondingly reduces their expression (Ströbel et al., 2007), for other gene loci AIRE may negatively regulate expression (Johnnidis et al., 2007). It may be that AIRE has a contrasting role in controlling thymic expression from the CHRNE locus. The relationship with tumour type is intriguing. In type A or AB thymomas the epithelial cells frequently show ‘medullary’ or ‘spindle’ morphology, and thymocytes may be focally abundant (type AB or ‘mixed’ thymoma) or absent (type A). In type B the pathology resembles disorganized thymic cortex with polygonal epithelial cells and abundant (B1 and B2) or some (B3) developing T cells; about 50% of type B thymomas are invasive. Around 30% of all patients with type B thymomas develop MG, but curiously only 5–25% of patients with A or AB thymomas (Ströbel et al., 2004). One could hypothesise that, in order to induce an immune response, higher levels of AChR ε-subunit expression are required in A/AB thymomas than in type B2 tumours, which provide a more immunogenic environment. Subsequent determinant spreading of the response to other AChR epitopes expressed by myoid cells in the adjacent thymus would generally occur leading to the typical serum antibodies found in most patients with MG. Acknowledgements This study was supported by the Wellcome Trust (grant reference number 035860/Z/92/Z), the Myasthenia Gravis Association, the Muscular Dystrophy Campaign, the Sir Jules Thorn Charitable Trust, the Medical Research Council, and a grant by the European Commission in the FP6 program to AM (LSHB-CT-2003-503410, Euro-Thymaide). References Beeson, D., Morris, A., Vincent, A., Newsom-Davis, J., 1990. The human nicotinic acetylcholine receptor α subunit exists as two isoforms: a novel exon. EMBO J. 9, 2101–2106. Beeson, D., Amar, M., Bermudez, I., Vincent, A., Newsom-Davis, J., 1996. Stable functional expression of the adult subtype of human muscle acetylcholine receptor following transfection of the human rhabdomyosarcoma cell line TE671 with cDNA encoding the ε subunit. Neurosci. Lett. 207, 57–60. Beeson, D., Brydson, M., Betty, M., Jeremiah, S., Povey, S., Vincent, A., NewsomDavis, J., 1993. Primary structure of the human muscle acetylcholine receptor. cDNA cloning of the γ and ε subunits. Eur. J. Biochem. 215, 229–238. Bruno, R., Sabater, L., Tolosa, E., Sospedra, M., Ferrer-Francesch, X., Coll, J., Foz, M., Melms, A., Pujol-Borrell, R., 2004. Different patterns of nicotinic acetylcholine receptor subunit transcription in human thymus. J. Neuroimmunol. 149, 147–159. Buckley, C., Newsom-Davis, J., Willcox, N., Vincent, A., 2001. Mature, longlived CD4+ and CD8+ T cells are generated by the thymoma in myasthenia gravis. Ann Neurol. 50, 64–72. Gilhus, N.E., Willcox, N., Harcourt, G., Nagvekar, N., Beeson, D., Vincent, A., Newsom-Davis, J., 1995. Antigen presentation by thymoma epithelial cells from myasthenia gravis patients to potentially pathogenic T cells. J. Neuroimmunol. 56, 65–76.
Giraud, M., Taubert, R., Vandiedonck, C., Ke, X., Lévi-Strauss, M., Pagani, F., Baralle, F.E., Eymard, B., Tranchant, C., Gajdos, P., Vincent, A., Willcox, N., Beeson, D., Kyewski, B., Garchon, H.J., 2007. An IRF8-binding promoter variant and AIRE control CHRNA1 promiscuous expression in thymus. Nat. 448, 934–937. Hill, M., Beeson, D., Moss, P., Jacobson, L., Bond, A., Corlett, L., Newsom-Davis, J., Vincent, A., Willcox, N., 1999. Early-onset myasthenia gravis: a recurring T-cell epitope in the adult-specific acetylcholine receptor ε subunit presented by the susceptibility allele HLA-DR52a. Ann Neurol. 45, 224–231. Johnnidis, J.B., Venanzi, E.S., Taxman, D.J., Ting, J.P., Benoist, C.O., Mathis, D.J., 2005. Chromosomal clustering of genes controlled by the aire transcription factor. Proc. Natl. Acad. Sci. U S A. 102, 7233–7238. Leite, M.I., Jones, M., Ströbel, P., Marx, A., Gold, R., Niks, E., Verschuuren, J.J., Berrih-Aknin, S., Scaravilli, F., Canelhas, A., Morgan, B.P., Vincent, A., Willcox, N., 2007. Myasthenia gravis thymus: complement vulnerability of epithelial and myoid cells, complement attack on them, and correlations with autoantibody status. Am. J. Pathol. 171, 893–905. MacLennan, C., Beeson, D., Buijs, A.M., Vincent, A., Newsom-Davis, J., 1997. Acetylcholine receptor expression in human extraocular muscles and their susceptibility to myasthenia gravis. Ann Neurol. 41, 423–431. Marx, A., Wilisch, A., Schultz, A., Gattenlöhner, S., Nenninger, R., MüllerHermelink, H.K., 1997. Pathogenesis of myasthenia gravis. Virchows Arch. 430, 355–364. Marx, A., Müller-Hermelink, H.K., 1999. From basic immunobiology to the upcoming WHO-classification of tumors of the thymus. The Second Conference on Biological and Clinical Aspects of Thymic Epithelial Tumors and related recent developments. Path Res. Pract. 195, 515–533. Muller-Hermelink, H.K., Marx, A., 2000. Thymoma. Curr. Opin. Oncol. 12, 426. Nagvekar, N., Moody, A.M., Moss, P., Roxanis, I., Curnow, J., Beeson, D., Pantic, N., Newsom-Davis, J., Vincent, A., Willcox, N., 1998. A pathogenic role for the thymoma in myasthenia gravis; autosensitization of IL-4producing T cell clones recognizing extracellular AChR epitopes presented by minority class II isotypes. J. Clin. Invest. 101, 2268–2277. Newsom-Davis, J., Vincent, A., Willcox, H.N., 1993. Autoimmune disorders of the neuromuscular junction, In: Lachmann, P.J., Peters, D.K., Rosen, F.S., Walport, N.J. (Eds.), 5th ed. Clinical aspects of immunology, vol. 3. Blackwell Scientific Publications, Oxford, pp. 2091–2111. Ragheb, S., Mohamed, M., Lisak, R.P., 2005. Myasthenia gravis patients, but not healthy subjects, recognize epitopes that are unique to the epsilonsubunit of the acetylcholine receptor. J. Neuroimmunol. 159, 137–145. Rosai, J., Levine, G.D., 1976. Tumors of the thymus. Atlas of Tumor Pathology, 2nd series. US Armed Forces Institute of Pathology. Fascicle 13. Sanes, J.R., Lichtman, J.W., 2001. Induction, assembly, maturation and maintenance of a postsynaptic apparatus. Nat. Rev. Neurosci. 2, 791–805. Schluep, M., Willcox, N., Vincent, A., Dhoot, G.K., Newsom-Davis, J., 1987. Acetylcholine receptors in human thymic myoid cells in situ: an immunohistological study. Ann Neurol. 22, 212–222. Ströbel, P., Bauer, A., Puppe, B., Kraushaar, T., Krein, A., Toyka, K., Gold, R., Semik, M., Kiefer, R., Nix, W., Schalke, B., Muller-Hermelink, H.K., Marx, A., 2004. Tumor recurrence and survival in patients treated for thymomas and thymic squamous cell carcinomas: a retrospective analysis. J. Clin. Oncol. 22, 1501–1509. Ströbel, P., Murumägi, A., Klein, R., Luster, M., Lahti, M., Krohn, K., Schalke, B., Nix, W., Gold, R., Rieckmann, P., Toyka, K., Burek, C., Rosenwald, A., Müller-Hermelink, H.K., Pujoll-Borrell, R., Meager, A., Willcox, N., Peterson, P., Marx, A., 2007. Deficiency of the autoimmune regulator AIRE in thymomas is insufficient to elicit autoimmune polyendocrinopathy syndrome type 1 (APS-1). J. Pathol. 211, 563–571. Vincent, A., Willcox, N., Hill, M., Curnow, J., MacLennan, C., Beeson, D., 1998. Determinant spreading and immune responses to acetylcholine receptors in myasthenia gravis. Immunol. Rev. 164, 157–168. Wilisch, A., Gutsche, S., Hoffacker, V., Schultz, A., Tzartos, S., Nix, W., Schalke, B., Schneider, C., Müller-Hermelink, H.K., Marx, A., 1999. Association of acetylcholine receptor α-subunit gene expression in mixed thymoma with myasthenia gravis. Neurol. 52, 1460–1466.
Preferential expression of AChR ε-subunit in thymomas from patients with myasthenia gravis Calman A. MacLennan a,b , Angela Vincent a , Alexander Marx c , Nicholas Willcox a , Nils Eric Gilhus d , John Newsom-Davis a , David Beeson a,⁎ a b
c
Neurosciences Group, Weatherall Institute of Molecular Medicine, University of Oxford, Oxford OX3 9DS, UK Medical Research Council Centre for Immune Regulation, Division of Immunity and Infection, Medical School, University of Birmingham, Birmingham, B15 2TT, UK Institute of Pathology, University Hospital Mannheim, University of Heidelberg, D-68125 Mannheim, Germany d Department of Clinical Medicine, Section of Neurology, University of Bergen, Norway Received 17 April 2008; received in revised form 16 June 2008; accepted 17 June 2008
Abstract The role of antigen expression by thymomas in myasthenia gravis (MG) is not clear. Previous reports of acetylcholine receptor (AChR) mRNA expression by the highly sensitive reverse transcription-polymerase chain reactions (RT-PCR) produced varying results. To try to clarify this issue, we first used RT-PCR but then turned to the more accurate and quantitative RNase protection assays (RPA) to assess AChR subunit mRNA expression in thymomas from 25 patients (22 with MG). By RT-PCR, all five AChR subunits could be detected in many thymomas. However, by RPA, the mRNA for the adult-specific AChR ε-subunit was found in 13/25 (52%) thymomas, but not mRNA for the other subunits. AChR ε-subunit was more frequently detected in thymomas of A or AB histology (WHO classification) than those with B1–B3 histology. Overall, 6/6 with thymomas of A or AB histology were positive compared with only 8/19 with B histology (p = 0.02). Autoantibodies in the two patients with the highest levels of ε-subunit mRNA bound better to adult (α2βδε) AChR than to fetal (α2βδγ) AChR, whereas the other sera bound better to fetal AChR. The greater abundance of mRNA for AChR ε-subunit than for other subunits suggests that the AChR ε-subunit may play a distinctive role in autosensitization in MG-associated thymomas, particularly those of type A or AB. © 2008 Published by Elsevier B.V. Keywords: Myasthenia gravis; Thymoma; AChR ε-subunit; RNase protection
1. Introduction In myasthenia gravis (MG), autoantibodies against the native AChR conformation cause loss of AChR at the neuromuscular junction and fatigable muscle weakness (Newsom-Davis et al., 1993). The AChR is a pentameric transmembrane protein, with
Abbreviations: AChR, Muscle acetylcholine receptor; RT-PCR, Reverse transcription-polymerase chain reaction; RPA, RNase protection assay; MG, Myasthenia gravis; TE671-γ, TE671 cell line that expresses fetal AChR; TE671-ε, TE671 cell line transfected to expresses adult AChR. ⁎ Corresponding author. Tel.: +44 1865 222311; fax: +44 1865 222402. E-mail address: [email protected] (D. Beeson). 0165-5728/$ - see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.jneuroim.2008.06.016
two splice forms of the α-subunit, P3A− and P3A+ (Beeson et al., 1990) in humans; during development, the fetal γ-subunit is replaced by the ‘adult’-specific ε-subunit to form the mature AChR (α2,β,δ,ε). Around 10% of MG patients have thymomas (Marx et al., 1997) which are epithelial neoplasms, with varying lymphocytic components, classified as WHO types A, AB, B1–3 (Strobel et al., 2004). It has been suggested that defective negative selection of the thymocytes leads to the export of potentially autoreactive T cells (Marx and Müller-Hermelink, 1999), but the strong association of MG with antibodies to AChR and other muscle antigens suggests active sensitization of T cells to muscle antigens in the tumour (Nagvekar et al., 1998). Indeed, low levels of all five AChR subunit mRNAs have been detected
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
29
2.2. Reverse-transcription PCR (RT-PCR)
using reverse transcription followed by PCR (RT-PCR), or nested PCR (Wilisch et al., 1999). RT-PCR, however, is extremely sensitive and the results may be misleading. We studied 22 MG, and 3 non-MG thymomas with the more accurate and quantitative method of RNAse protection.
Total RNA was prepared by the phenol/guanidinium isothiocyanate method using RNAzol B (Cinna/Biotex Laboratories). RNA quality was assessed by visualizing under UV light 5 μg of each sample run on a formaldehyde/agarose gel, and by the intensity and integrity of bands for β-actin generated by RNase protection and RT-PCR assays. cDNA was synthesized from 10 μg of total RNA. PCR amplifications were performed according to the Perkin-Elmer/Cetus (Roche Molecular Systems Inc.) instructions, as previously described (MacLennan et al., 1997). All amplified products contained sequences from more than one exon, thus excluding by size any products that might derive from contaminating genomic DNA. Product specificity was confirmed by Southern blotting using an internal oligonucleotide as probe.
2. Methods 2.1. Patients Thymomas were obtained at surgery from 22 patients with MG (Oxford and London, UK) and 3 without MG (Bergen, Norway). The diagnosis of MG was based on clinical and electromyographic features and elevated serum AChR antibodies (Table 1A). None of the three patients without clinical MG showed electromyographic abnormalities. Normal thymus was obtained from children undergoing cardio-thoracic surgery. Thymoma tissue was carefully separated from the adjacent thymus and stored at − 80°C. The histology was reviewed by AM (Table 1A). All samples were taken with informed consent and local ethical committee approval.
2.3. RNase protection assays (RPA) RNase protection assays (RPA) were performed essentially as described previously (MacClennan et al., 1997). cDNA
Table 1A Expression of AChR subunits in thymomas and serum antibody reactivity Patient code
MG 14 15 5 13 22 6 11 1 3 4 7 12 16 17 18 19 8 20 2 9 10 21
Age at thymectomy (years M/F)
63 F 46 F 64 M 56 M 47 F 64 F 53 F 50 F 17 M 57 M 59 F 19 M 66 F 48 M 50 M 45 F 23 F 47 M 43 M 35 M 45 F 32 M
No MG 25 58 M 23 37 F 24 66 F
Duration of MG at thymectomy (months)
Thymoma histology
Thymoma RPA
Thymoma RT-PCR
ε-subunit
ε-subunit
α-subunit
AChR antibody reactivity with
Adult AChR/ fetal AChR
γ-subunit
Fetal AChR
Adult AChR
Ratio
11 67 188 3 4 6 45 na 2 13 1 2 4 8 84 92 4 16 3 20 7 2
A⁎ A⁎ AB AB AB/B2 B1 B1 B1/B2 B2 B2 B2 B2 B2 B2 B2 B2† B2/B3⁎ B2/B3⁎ B3⁎ B3⁎ epithelial⁎ na
+++ + ++ + ++ ++ − + + − − − − − − − + + + − + −
+ + + + + + − + + + − − − + + + + + + − + +
+ − + + + + + + + + + + + + + + + + + − + +
+ − − − + + + + − + − − − − + + + − − − − −
5.5 9.6 1.4 na na 59.4 67.1 35.5 9.2 27.3 17.0 43.6 na na na na 36.0 na 81.0 52.0 544.2 na
7.9 3.6 6.7 na na 38.3 44.2 13.3 6.0 19.6 7.7 20.9 na na na na 25.2 na 43.4 30.1 302.5 na
1.44 0.38 4.79 na na 0.64 0.66 0.37 0.65 0.72 0.45 0.48 na na na na 0.7 na 0.54 0.58 0.56 na
na na na
AB B2 B2
+ − −
+ − +
− − +
+ − +
na 0.5 7.1
na 0.2 4.6
na 0.4 0.65
Samples ranked according to WHO histological classification. ⁎Patient on corticosteroid therapy at the time of resection. No further classification possible. †Pleural metastasis of similar appearance as primary tumor. “+” = AChR subunit mRNA detected. “−” = AChR subunit mRNA not detected. na, not available.
30
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
encoding fragments of the extracellular regions of the respective human AChR subunits (Beeson et al., 1993), and cytoplasmic β-actin, were subcloned into pGEM vectors. Anti-sense RNA probes were synthesized in vitro using 50 μCi [α-32P]CTP at N 3000 Ci/mmol (Amersham International) with either T7 or SP6 RNA polymerase. Plasmid DNA was removed by digestion with DNase I for 15 min at 37 °C. Protection assays were performed using the RPA II kit (Ambion Inc.) according to manufacturer's instructions, using 50 μg total RNA for each assay sample. Products were run on a standard 6% polyacrylamide urea sequencing gel and visualized by autoradiography. In order to optimize assay conditions, anti-sense probes were hybridized with calibrated dilutions of AChR subunit cRNA transcribed in vitro. RPAs were performed at least three times for each thymoma sample. 2.4. AChR antibody assays Sera taken around the time of thymectomy were assayed by radioimmunoprecipitation for reactivity with fetal and adult AChR as previously described (Beeson et al., 1996). Titers were expressed as nmol 125 I-α-bungarotoxin-binding sites precipitated per liter of serum.
3. Results Each of the five AChR subunit mRNAs could be detected by RT-PCR in normal thymus samples, as confirmed by Southern blotting and hybridization with an internal subunit-specific oligonucleotide (data not shown). These results confirm previous reports (Bruno et al., 2004) and also demonstrated the presence of AChR δ-subunit mRNA. We studied thymomas from 22 patients with MG and 3 without MG. Since AChR antibodies were present in all sixteen sera available for analysis, including two of those without MG (Table 1A), we pooled all the results for analysis. Expression of AChR α-, ε- and γ-subunits in thymomas by RT-PCR was positive in 21/25 (84%), 19/25 (76%) and 11/25 (44%) samples respectively (Fig. 1B). Both P3A+ and P3A− isoforms of the α-subunit were detected, although the P3A− signals were consistently stronger. AChR β- and δ-subunit mRNAs were also detected in many thymomas (data not shown). To obtain a more accurate measure of the mRNAs, we used RNase protection assays (RPA). None of the AChR subunits were detected by RPA in RNA derived from control thymus (n = 3), or from human liver, lung, brain or from yeast. By contrast, RNA derived from partially denervated muscle gave
Fig. 1. Detection of mRNA encoding β-actin and AChR α-, ε- and γ-subunits in thymoma samples and in control thymus, liver, lung, brain, yeast and skeletal muscle. A. RPA detection of mRNA. Autoradiographs of respective anti-sense probes protected by α-, ε- and γ-mRNA were exposed for 14 days at − 80°C, whereas for β-actin protection assay exposure was for 4 h at room temperature. Undigested anti-sense RNA probe is shown in the right-hand lane. B. RT-PCR amplifications of RNA samples. Products were size fractionated on 1.5% agarose gels and visualized under ultraviolet light after staining with ethidium bromide.
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
31
Table 1B Correlation with thymoma histology and between RPA and RT-PCR results
AChR ε-subunit detected N = 13 AChR ε-subunit not detected N = 12
Median age (range) M:F
WHO A and AB classification
WHO B2 classification
Positivity for AChR subunit by RT-PCR ε-subunit
α-subunit
γ-subunit
49 (17–63) 5:7 ns 49 (19–66) 6:3
6 p = 0.020 0
5 p = 0.047 9
13 (100) p = 0.005 6 (50)
11 (85) ns 10 (83)
6 (46) ns 6 (50)
Data were analysed by Fisher's exact test.
strong RPA bands for all AChR subunits except ε (as predicted as the AChR would be predominantly of fetal type). Remarkably, in thymomas, RPA detected only the ε-subunit mRNA, with 13 (52%) of the 25 samples positive (Fig. 1A). Though variable, band intensities were often comparable with those from muscle. We did not detect signals for the α-, β-, γ- or δ-subunit mRNA by RPA, even in the thymomas positive by RTPCR for AChR α- or γ-, β- or δ-subunit mRNAs. The results are detailed in Table 1A where they are listed according to the thymic pathology, and correlations with histology and RT-PCR results are summarized in Table 1B. All of the 13 thymomas that were positive for ε-subunit mRNA by RPA were also positive by RT-PCR, compared with only 6 of the 12 RPA-negatives (p = 0.005, Fisher's exact test). Interestingly, ε-subunit mRNA expression by RPA was more frequent in tumours classified as WHO type A or AB than in those classified as predominantly B1–3 (Table 1B, Fisher's exact test p = 0.020). Sera were not available on all patients. All of the 11 patients whose thymomas were negative by RPA for AChR εmRNA had lower levels of anti-adult than anti-fetal AChR antibodies, as is commonly seen in non-thymomatous MG. However, sera from two patients with the strongest ε-signals by RPA, #14 and #5, bound more strongly to adult AChR than to fetal AChR. 4. Discussion Despite many demonstrations of the AChR subunits in thymomas from MG patients, it has never been clear whether there is preferential expression of any one subunit, probably because the RT-PCR methods used were too sensitive. We used the RNase protection assay (RPA) which readily detects expression of all AChR subunit mRNAs in human muscle (MacClennan et al., 1997), and after denervation (Fig. 1A), but provides a more quantitative and accurate measure than RTPCR. Surprisingly, we only detected the AChR ε-subunit mRNA; this was found in 12/22 (55%) of thymomas from MG patients and in one of three without MG. Moreover, ε-subunit mRNA expression was more common in thymomas of the A or AB classification than in those with the B classifications. Our results using RT-PCR confirm earlier findings (Wilisch et al., 1999; Bruno et al., 2004) including the presence of both the P3A+ and P3A− isoforms of the α-subunit (Beeson et al., 1990). The lack of detection of the ε-subunit mRNAs in some thymomas, and lack of α-, β-, δ- and γ-subunit mRNAs by RPA, even in samples positive by RT-PCR, suggests that the levels of these subunits are relatively low in all cases.
The ε-subunit mRNAs that we detected by RPA are very unlikely to have derived from myoid cells, which are rare in most thymomas (Rosai and Levine, 1976). Our failure to detect ε by RPA in normal thymus, which contains myoid cells that express fetal AChR (Schluep et al., 1987), is consistent with this and implies, therefore, over-expression in thymomas. Conversely, the α-, β-, δ- and γ-AChR subunits that we detected in thymoma samples by RT-PCR might well have derived from myoid cells in adjacent normal thymic tissue (Leite et al., 2007). Thymomas show strong associations with autoimmune disorders (Buckley et al., 2001) and generate and export T cells (Muller-Hermelink and Marx, 2000). In theory, the tumors could either be directly selecting/sensitizing T cells against selfantigens that they express (Nagvekar et al., 1998) or failing to tolerize against antigens that they do not express (Marx and Muller-Hermelink, 2005). The over-expression of the ε-subunit in thymomas argues in favor of the former, as does the preference for adult AChR shown by serum antibodies from the two patients with the highest levels of ε-subunit mRNA. In the other five, the more prevalent antibodies against fetal AChR, which are common in MG, might be the end-result of determinant spreading (Vincent et al., 1998) initiated by earlier responses to the ε-subunit. Indeed, several features of the AChR ε distinguish it from the other subunits and suggest a special potential for breaking tolerance in MG. Firstly, transcription of the ε-subunit is under tight regulatory control, and is normally restricted to the subsynaptic nuclei of muscle fibers (Sanes and Lichtman, 2001), suggesting that aberrant over-expression in thymomas could be immunologically ‘dangerous'. Secondly, since its expression is so minimal in the normal thymus, it may tolerize incompletely with the result that patients would have some circulating ε-subunit-specific T cells that could colonize the thymoma. Thirdly, as the ε-subunit appears to be produced in excess over the other AChR subunits in thymomas, it is much more likely to be degraded rapidly than to be incorporated into whole AChR molecules, with the result that the resulting peptides would be available for presentation to developing T cells, either by the neoplastic epithelial cells (Gilhus et al., 1995) or the abundant adjacent dendritic cells. It has already been shown that the ε-subunit includes at least one important helper T cell epitope (Hill et al., 1999) and T cells from MG patients have been found to recognize epitopes unique to the ε-subunit (Ragheb et al., 2005). The autoimmune regulator (AIRE) plays a crucial role in driving the expression of self-antigens in medullary thymic epithelial cells. It has been shown to play a role in controlling thymic expression of the AChR α-subunit from the CHRNA locus, and possibly setting the threshold for self-tolerance
32
C.A. MacLennan et al. / Journal of Neuroimmunology 201–202 (2008) 28–32
versus autoimmunity for the AChR (Giraud et al., 2007). However, AIRE expression is down-regulated in human thymoma. Whereas the presence of AIRE is responsible for activating a number of loci, and its loss correspondingly reduces their expression (Ströbel et al., 2007), for other gene loci AIRE may negatively regulate expression (Johnnidis et al., 2007). It may be that AIRE has a contrasting role in controlling thymic expression from the CHRNE locus. The relationship with tumour type is intriguing. In type A or AB thymomas the epithelial cells frequently show ‘medullary’ or ‘spindle’ morphology, and thymocytes may be focally abundant (type AB or ‘mixed’ thymoma) or absent (type A). In type B the pathology resembles disorganized thymic cortex with polygonal epithelial cells and abundant (B1 and B2) or some (B3) developing T cells; about 50% of type B thymomas are invasive. Around 30% of all patients with type B thymomas develop MG, but curiously only 5–25% of patients with A or AB thymomas (Ströbel et al., 2004). One could hypothesise that, in order to induce an immune response, higher levels of AChR ε-subunit expression are required in A/AB thymomas than in type B2 tumours, which provide a more immunogenic environment. Subsequent determinant spreading of the response to other AChR epitopes expressed by myoid cells in the adjacent thymus would generally occur leading to the typical serum antibodies found in most patients with MG. Acknowledgements This study was supported by the Wellcome Trust (grant reference number 035860/Z/92/Z), the Myasthenia Gravis Association, the Muscular Dystrophy Campaign, the Sir Jules Thorn Charitable Trust, the Medical Research Council, and a grant by the European Commission in the FP6 program to AM (LSHB-CT-2003-503410, Euro-Thymaide). References Beeson, D., Morris, A., Vincent, A., Newsom-Davis, J., 1990. The human nicotinic acetylcholine receptor α subunit exists as two isoforms: a novel exon. EMBO J. 9, 2101–2106. Beeson, D., Amar, M., Bermudez, I., Vincent, A., Newsom-Davis, J., 1996. Stable functional expression of the adult subtype of human muscle acetylcholine receptor following transfection of the human rhabdomyosarcoma cell line TE671 with cDNA encoding the ε subunit. Neurosci. Lett. 207, 57–60. Beeson, D., Brydson, M., Betty, M., Jeremiah, S., Povey, S., Vincent, A., NewsomDavis, J., 1993. Primary structure of the human muscle acetylcholine receptor. cDNA cloning of the γ and ε subunits. Eur. J. Biochem. 215, 229–238. Bruno, R., Sabater, L., Tolosa, E., Sospedra, M., Ferrer-Francesch, X., Coll, J., Foz, M., Melms, A., Pujol-Borrell, R., 2004. Different patterns of nicotinic acetylcholine receptor subunit transcription in human thymus. J. Neuroimmunol. 149, 147–159. Buckley, C., Newsom-Davis, J., Willcox, N., Vincent, A., 2001. Mature, longlived CD4+ and CD8+ T cells are generated by the thymoma in myasthenia gravis. Ann Neurol. 50, 64–72. Gilhus, N.E., Willcox, N., Harcourt, G., Nagvekar, N., Beeson, D., Vincent, A., Newsom-Davis, J., 1995. Antigen presentation by thymoma epithelial cells from myasthenia gravis patients to potentially pathogenic T cells. J. Neuroimmunol. 56, 65–76.
Giraud, M., Taubert, R., Vandiedonck, C., Ke, X., Lévi-Strauss, M., Pagani, F., Baralle, F.E., Eymard, B., Tranchant, C., Gajdos, P., Vincent, A., Willcox, N., Beeson, D., Kyewski, B., Garchon, H.J., 2007. An IRF8-binding promoter variant and AIRE control CHRNA1 promiscuous expression in thymus. Nat. 448, 934–937. Hill, M., Beeson, D., Moss, P., Jacobson, L., Bond, A., Corlett, L., Newsom-Davis, J., Vincent, A., Willcox, N., 1999. Early-onset myasthenia gravis: a recurring T-cell epitope in the adult-specific acetylcholine receptor ε subunit presented by the susceptibility allele HLA-DR52a. Ann Neurol. 45, 224–231. Johnnidis, J.B., Venanzi, E.S., Taxman, D.J., Ting, J.P., Benoist, C.O., Mathis, D.J., 2005. Chromosomal clustering of genes controlled by the aire transcription factor. Proc. Natl. Acad. Sci. U S A. 102, 7233–7238. Leite, M.I., Jones, M., Ströbel, P., Marx, A., Gold, R., Niks, E., Verschuuren, J.J., Berrih-Aknin, S., Scaravilli, F., Canelhas, A., Morgan, B.P., Vincent, A., Willcox, N., 2007. Myasthenia gravis thymus: complement vulnerability of epithelial and myoid cells, complement attack on them, and correlations with autoantibody status. Am. J. Pathol. 171, 893–905. MacLennan, C., Beeson, D., Buijs, A.M., Vincent, A., Newsom-Davis, J., 1997. Acetylcholine receptor expression in human extraocular muscles and their susceptibility to myasthenia gravis. Ann Neurol. 41, 423–431. Marx, A., Wilisch, A., Schultz, A., Gattenlöhner, S., Nenninger, R., MüllerHermelink, H.K., 1997. Pathogenesis of myasthenia gravis. Virchows Arch. 430, 355–364. Marx, A., Müller-Hermelink, H.K., 1999. From basic immunobiology to the upcoming WHO-classification of tumors of the thymus. The Second Conference on Biological and Clinical Aspects of Thymic Epithelial Tumors and related recent developments. Path Res. Pract. 195, 515–533. Muller-Hermelink, H.K., Marx, A., 2000. Thymoma. Curr. Opin. Oncol. 12, 426. Nagvekar, N., Moody, A.M., Moss, P., Roxanis, I., Curnow, J., Beeson, D., Pantic, N., Newsom-Davis, J., Vincent, A., Willcox, N., 1998. A pathogenic role for the thymoma in myasthenia gravis; autosensitization of IL-4producing T cell clones recognizing extracellular AChR epitopes presented by minority class II isotypes. J. Clin. Invest. 101, 2268–2277. Newsom-Davis, J., Vincent, A., Willcox, H.N., 1993. Autoimmune disorders of the neuromuscular junction, In: Lachmann, P.J., Peters, D.K., Rosen, F.S., Walport, N.J. (Eds.), 5th ed. Clinical aspects of immunology, vol. 3. Blackwell Scientific Publications, Oxford, pp. 2091–2111. Ragheb, S., Mohamed, M., Lisak, R.P., 2005. Myasthenia gravis patients, but not healthy subjects, recognize epitopes that are unique to the epsilonsubunit of the acetylcholine receptor. J. Neuroimmunol. 159, 137–145. Rosai, J., Levine, G.D., 1976. Tumors of the thymus. Atlas of Tumor Pathology, 2nd series. US Armed Forces Institute of Pathology. Fascicle 13. Sanes, J.R., Lichtman, J.W., 2001. Induction, assembly, maturation and maintenance of a postsynaptic apparatus. Nat. Rev. Neurosci. 2, 791–805. Schluep, M., Willcox, N., Vincent, A., Dhoot, G.K., Newsom-Davis, J., 1987. Acetylcholine receptors in human thymic myoid cells in situ: an immunohistological study. Ann Neurol. 22, 212–222. Ströbel, P., Bauer, A., Puppe, B., Kraushaar, T., Krein, A., Toyka, K., Gold, R., Semik, M., Kiefer, R., Nix, W., Schalke, B., Muller-Hermelink, H.K., Marx, A., 2004. Tumor recurrence and survival in patients treated for thymomas and thymic squamous cell carcinomas: a retrospective analysis. J. Clin. Oncol. 22, 1501–1509. Ströbel, P., Murumägi, A., Klein, R., Luster, M., Lahti, M., Krohn, K., Schalke, B., Nix, W., Gold, R., Rieckmann, P., Toyka, K., Burek, C., Rosenwald, A., Müller-Hermelink, H.K., Pujoll-Borrell, R., Meager, A., Willcox, N., Peterson, P., Marx, A., 2007. Deficiency of the autoimmune regulator AIRE in thymomas is insufficient to elicit autoimmune polyendocrinopathy syndrome type 1 (APS-1). J. Pathol. 211, 563–571. Vincent, A., Willcox, N., Hill, M., Curnow, J., MacLennan, C., Beeson, D., 1998. Determinant spreading and immune responses to acetylcholine receptors in myasthenia gravis. Immunol. Rev. 164, 157–168. Wilisch, A., Gutsche, S., Hoffacker, V., Schultz, A., Tzartos, S., Nix, W., Schalke, B., Schneider, C., Müller-Hermelink, H.K., Marx, A., 1999. Association of acetylcholine receptor α-subunit gene expression in mixed thymoma with myasthenia gravis. Neurol. 52, 1460–1466.