 1998 Oxford University Press

Human Molecular Genetics, 1998, Vol. 7, No. 13 2051–2055

Mosaicism in sporadic neurofibromatosis 2 patients Lan Kluwe and Victor-F. Mautner1 Laboratory for Brain Tumor Biology, Department of Neurosurgery, University Hospital Eppendorf, Germany and 1Department of Neurology, General Hospital Ochsenzoll, Hamburg, Germany Received June 1, 1998; Revised and Accepted August 28, 1998

More than half of neurofibromatosis 2 (NF2) patients represent de novo mutations which could have occurred at either pre-zygotic or post-zygotic stages. A post-zygotic mutation can result in mosaicism. In four sporadic NF2 patients, we found NF2 mutations in only a portion of corresponding leukocytes. In two other sporadic patients, no mutations were found in leukocytes but constitutional NF2 mutations were suggested by identical mutations in different tumors from each patient. We screened leukocyte DNA from a total of 16 inherited and 91 sporadic NF2 patients, and found NF2 mutations in 13 (81%) of the former and in 46 (51%) of the latter cases. The 30% difference in the rate of detection of mutations (P = 0.051) might be partially explained by mosaicism in a portion of sporadic NF2 patients who carry the mutations in such a fashion that their leukocytes are unaffected. Among sporadic cases, we found mutations more frequently in patients with severe phenotypes (59%) than in patients with mild phenotypes (23%) (difference of 36%, P = 0.007). Mosaicism might be more common in the latter patient group since small populations of mutation-bearing cells can in some cases result in mild phenotypes and can also lead to difficulties in identifying mutations. No mutations were found in eight patients suspected of having NF2. Mosaicism with an extremely small population of affected cells may explain the incomplete phenotypes in some of these patients and the lack of mutations in their leukocytes. These findings suggest that mosaicism is relatively common in NF2 and may have important implications for diagnosis, prognosis and genetic counseling. INTRODUCTION Neurofibromatosis 2 (NF2) is an autosomal dominant disorder characterized by tumors of neural-crest origin cells. Bilateral vestibular schwannomas are the most frequent manifestation in this condition. Cerebral meningiomas, spinal tumors and tumors of the peripheral nerves as well as ophthalmologic abnormalities are also common in NF2 (1–4). NF2 has an incidence of 1:35 000–1:40 000 (5), and is caused by mutations in the NF2

gene on 22q12 (6,7). Various mutations in the NF2 gene have been identified in leukocyte DNA of 30–60% of NF2 patients (8–12) as well as in NF2-related tumors (13–16). Despite the almost ubiquitous distribution of the NF2 gene product, merlin or schwannomin, NF2-associated tumors are confined mainly to schwannomas and meningiomas. Growth control of Schwann cells and meningothelial cells is thus especially sensitive to inactivation of both copies of the NF2 gene. More than half of NF2 patients are sporadic and may therefore represent de novo or new mutations (2,5) which may have occurred either in the germline cells of their parents (pre-zygotic) or in any cell after fertilization and cell division (post-zygotic). A post-zygotic mutation can result in mosaicism (17). Mosaicism has been reported in a number of genetic diseases including neurofibromatosis 1 (18–21), congenital contractural arachnodactyly (22), androgen insensitivity syndrome (23) and NF2 (24,25). However, previous studies have focused mainly on single cases while the extent of mosaicism and its implications have not been investigated systematically. Recently we and others (12) have noticed that mutations are found more often in NF2 patients with family history or affected offspring than in sporadic patients with no affected offspring. Mosaicism in NF2 may thus be relatively common, and needs to be investigated further. In this study, we present direct evidence of mosaicism in four, and indicative evidence in two, sporadic NF2 cases. We further describe the different rates of detection of mutations in inherited versus sporadic cases, and in severe phenotypic versus mild phenotypic cases, and discuss a possible association with mosaicism. RESULTS To elucidate the mutation spectrum and investigate genotype– phenotype correlation in NF2, we screened leukocyte DNA from 107 unrelated NF2 patients. Using temperature gradient gel electrophoresis (TGGE) (26), exons 1–15 of the NF2 gene including exon–intron boundary sequences were scanned. Mutations were identified in 59 (55%) patients. Exon 8 amplified from patient 147.2, a patient with an NF2-affected mother, showed a typical mutant TGGE pattern: one normal homo-duplex, one mutant homo-duplex and two hetero-duplexes, all with equal intensity (Fig. 1a, lane 2). This pattern indicates that the ratio of normal to mutant alleles is 1:1, and thus that all cells in the patient carry the mutation on one allele. Sequencing revealed the transition 784C→T, a nonsense mutation resulting in R262X (Fig. 1b, upper panel). The affected

*To whom correspondence should be addressed. Tel: +49 40 4717 2767; Fax: +49 40 4717 2765; Email: [email protected]

2052 Human Molecular Genetics, 1998, Vol. 7, No. 13 mother, patient 147.1 (sporadic), showed the same TGGE pattern (Fig. 1a, lane 3), but the relative intensities of the duplexes were different from those of her daughter: the two hetero-duplexes were weaker than the normal homo-duplex and the mutant homo-duplex was almost invisible. The same pattern was obtained again upon repeated amplification, indicating that the unusual TGGE pattern is not an artifact due to Taq polymerase error during the amplification. Sequencing of the fragment from patient 147.1 showed a very small amount of T at position 784, but the signal was too low to be evidence of a mutation (Fig. 1b, middle panel). These findings indicated that the sporadic patient 147.1 may have the mutation in only a portion of her leukocytes. To confirm the mutation, we extracted DNA from the heteroduplex-containing bands (which contain normal and mutant molecules in a 1:1 ratio), re-amplified the fragment and re-analyzed it using TGGE. As shown in Figure 1a (lane 4), a typical mutant pattern was then obtained. Sequencing confirmed the alteration of 784C→T in the re-amplified fragment (Fig. 1b, lower panel). Exon 4 of patient 150 (a sporadic patient) showed an abnormal band upon TGGE analysis (Fig. 2a, lane 2, arrow) which was absent in 20 other samples analyzed at the same time. The same pattern was obtained upon repeated amplification, indicating that the abnormal band is not due to Taq polymerase error. Direct sequencing revealed a possible alteration of 431insA (Fig. 2b, upper panel). However, the signals of the second sequence were too weak to be considered as evidence of the mutation. To confirm the mutation, we excised the hetero-duplex band from the TGGE gel, extracted DNA from the gel pieces and re-amplified the fragment. A mutant homo-duplex with the same intensity as the normal homo-duplex was then obtained upon TGGE analysis (Fig. 2a, lane 3). Subsequent sequencing confirmed the single base insertion of 431insA (Fig. 2b, lower panel). In two other sporadic patients, 39 and 196, abnormal bands upon TGGE suggested alterations in exon 2 (Fig. 3, lane 3, arrow) and exon 8, respectively. The same TGGE pattern was obtained

upon repeated amplification of the corresponding exon in each patient, indicating that the abnormal bands were not due to Taq polymerase error. Direct sequencing, however, failed to identify any significant change. In the former case, amplified exon 2 fragment was cloned using T-vector cloning. A fragment from three of 20 analyzed clones had the same abnormal band as the original fragment from the patient (Fig. 3, lane 4, arrow). Sequencing revealed an 8 bp insertion (Table 1). In the case of patient 196, the abnormal bands were at the same positions as those of patient 147.1, indicating the mutation of 784C→T also in this patient (Table 1). Sequencing of fragment amplified from the abnormal band confirmed this mutation. In two further sporadic cases, no NF2 mutations were found in blood DNA. However, identical mutations were found in three tumors from patient 13 (596C→T) and in two tumors from patient 253 (58A→T). Loss of heterozygosity analysis revealed that both patients had one tumor with NF2 allelic loss (Table 1). Patient 253 had one unilateral vestibular schwannoma at the age of 67. Patient 13 had bilateral vestibular schwannomas, but one is large while the other one is very small. As summarized in Table 2, mutations were found in 13 (81%) of 16 patients with family history or affected offspring, while only in 46 (51%) of 91 sporadic patients with no affected offspring. The difference in mutation detection rates was thus 30% (P = 0.051) between the two patient groups. Among sporadic NF2 patients, mutations were found more often in those with severe phenotypes (59%) than in those with mild phenotypes (23%). The difference of the mutation detection rates (36%, P = 0.007) is statistically significant. Seven additional patients did not fulfill the updated diagnostic criteria for NF2 (27) but all had two or more NF2-related tumors such as vestibular schwannomas, meningiomas and spinal tumors. One more patient had a vestibular schwannoma and a meningioma at the age of 25 and thus fulfilled the updated criteria for presumptive NF2. Screening of these eight patients did not reveal any mutation.

Table 1. Mosaicism found in sporadic NF2 patients

Age at onset

Age at last examination

Symptomsa

Exon

Mutation

Codon

Consequence

Identified by analyzing

13

25

29

BVS, spinal and skin tumors

6

586C→T

196

Nonsense

Three tumors (one with NF2 allelic loss)

39

19

41

BVS, cerebral and spinal tumors

2

137del8bp

46

Frameshift

Cloned fragment

147.1

21

32

BVS, spinal and skin tumors

8

784C→T

262

Nonsense

Heteroduplex from leukocytes

150

23

34

BVS, cerebral and spinal tumors

4

431insA

144

Frameshift

Heteroduplex from leukocytes

196

16

44

BVS, spinal and skin tumors

8

784C→T

262

Nonsense

Heteroduplex from leukocytes

253

33

67

UVS, cerebral, spinal and skin tumors

1

58A→T

20

Nonsense

Two tumors (one with NF2 allelic loss)

Patient

aBVS,

bilateral vestibular schwannomas; UVS, unilateral vestilubar schwannomas.

2053 Human Genetics, 1998, 7, No. NucleicMolecular Acids Research, 1994, Vol. Vol. 22, No. 1 13 2053 Table 2. Mutation detection rate in NF2 patients Familial NF2 cases

Sporadic NF2 cases (severe, mild)

Cases of suspected NF2

Screened

16

91 (69, 22)

8

Mutation found

13

46 (41, 5)

0

Rate (%)

81

51 (59, 23)

0

Figure 1. A mosaic 784C→T in exon 8 of patient 147.1. (a) TGGE patterns. Four types of duplex are illustrated on the left with the mutation indicated by the asterisk. Lanes 1 and 5, negative control. Four bands with equal intensity in patient 147.2 (lane 2) indicated a 1:1 ratio of normal to mutant alleles. Two hetero-duplex bands of lesser intensity in patient 147.1 indicated excess of the normal allele (lane 3). The typical mutant TGGE pattern was obtained when the fragment was amplified from the hetero-duplex band of patient 147.1 (lane 4). (b) Sequencing of exon 8. A mixture of C/T at position 784 (arrow) is evident in patient 147.2 and in fragment re-amplified from the hetero-duplex of patient 147.1. In contrast, only a trace of T was visible in the fragment amplified from blood DNA of patient 147.1.

DISCUSSION In this study, we obtained direct evidence of somatic mosaicism in four sporadic NF2 cases, in which the patients had the NF2 mutations in only a portion of their leukocytes. In two other sporadic cases (patients 13 and 253), no mutations were found in

Figure 2. A mosaic 431insA in exon 4 of patient 150. (a) TGGE pattern of an amplified fragment. Lanes 1 and 4, negative control. The aberrant band in patient 150 (lane 2, arrow) indicated a mutation. By re-amplifying a fragment from this abnormal band, we obtained a mutant homo-duplex band and two hetero-duplexes (lane 3). The two hetero-duplexes migrated to the same position in the gel (lane 3), possibly because they have the same thermostability. (b) Sequencing of the fragment re-amplified from the hetero-duplex of patient 150 revealed, unambiguously, two overlapping sequences starting from position 431 (arrow), resulting from an ‘A’ insertion at this position. The second sequence was not, however, obvious in the original fragment amplified from genomic DNA of patient 150 (upper panel).

leukocytes while identical mutations were found in multiple tumors in each patient. Both patients had one tumor with NF2 allelic loss, indicating that the analyzed tumors developed independently from different cells. Since allelic loss is usually the second genetic event for tumorigenesis, we assumed that the identified mutations were the constitutional ones. The finding of identical mutations in different tumors in each patient also supported our assumption. Patient 253 only had one unilateral vestibular schwannoma at the age of 67. This is unusual and may be due to mosaic NF2. These two cases demonstrated that for sporadic patients with no mutations found in their leukocytes,

2054 Human Molecular Genetics, 1998, Vol. 7, No. 13 cases, which is 30% (P = 0.051) in this study and 22% in the study of Evans et al. (12), may reflect to some extent the frequency of mosaicism in NF2. Because the cell lineage commitment in the human embryo is not well understood, the effect and outcome of a mosaicism is unpredictable. Further studies are needed to elucidate the implications of mosaicism for diagnosis, prognosis and genetic counseling (25) in NF2 as well as in other genetic diseases. MATERIALS AND METHODS Figure 3. TGGE pattern of a mosaic 8 bp deletion in exon 2 of patient 39. Lane 1, a positive control of a point mutation; lane 2, a negative control; lane 3, a fragment amplified from the patient, a weak abnormal band is visible (arrow); lane 4, a fragment amplified from the one clone appeared at the same position as the abnormal band in the original fragment amplified from the patient (arrow); all other lanes: a fragment amplified from clones.

analysis of multiple tumors may provide an additional possibility of finding the constitutional mutations. Approximately half of NF2 patients described in two previous studies (2,5) and 91 (85%) of the 107 NF2 patients examined in the present study had no family history. These patients thus represent new or de novo mutations. Some of these new mutations may have occurred after fertilization, resulting in mosaicism: mixed cell populations with and without mutation in the same individual (17). A mutation detected by our screening method may not be found in a sporadic NF2 case if the patient is mosaic for the mutation in such a fashion that the leukocytes are unaffected. Depending on how other tissues, especially those of neural-crest origin, are affected, the patient can develop severe, mild or incomplete disease phenotypes. Mosaicism may therefore account for the inability to find mutations in some sporadic patients and could partially explain the 30% (P = 0.051) difference in the rate of detection of mutations in inherited and sporadic NF2 cases. Among sporadic NF2 patients, mosaicism may be more common in those with mild phenotypes since, in some cases, mild phenotypes can be the result of reduced population sizes of mutation-bearing cells. A higher frequency of mosaicism may partially explain the lower rate of detecting mutations (23%) in patients with mild phenotypes. Furthermore, some incomplete NF2 phenotypes may also be result of mosaicism with very small populations of mutation-bearing cells. Some of the eight patients suspected of having NF2 may have such mosaicism. No mutation was found in any of these patients, possibly because their leukocytes are not affected. It has been observed that in some families the first person affected had mild or atypical NF2 phenotypes while other affected individuals in the following generation developed typical or severe NF2 phenotypes (5,24,28). This can be explained if we assume that the first person affected is mosaic. Mosaicism-based mild phenotypes should also be taken into account when considering genotype–phenotype correlation, since in mosaic cases, the mild phenotypes may not truly reflect the effect of corresponding mutations (24,26). Our data suggest that somatic mosaicism is not exceptional but rather relatively common in NF2. The exact frequency as well as the nature of mosaicism is difficult to estimate. The difference in the rate of detecting mutations between inherited and sporadic

NF2 patients and affected relatives were contacted through our neurofibromatosis centre in collaboration with the German neurofibromatosis lay organization, von Recklinghausen Gesellschaft at Hamburg. The diagnosis of NF2 was based on the updated NIH criteria (27). The protocol was approved by the institutional review board and all participants provided informed consent. Leukocyte DNA from 107 unrelated NF2 patients was screened for mutations in the NF2 gene. Patients who had two or more NF2-related tumors but did not fulfill the updated NIH diagnostic criteria (27) were defined as having incomplete NF2 phenotypes or possibly having NF2. DNA was extracted from whole blood and tumors using a QIAamp Blood PCR Kit and a QIAamp tissue PCR Kit respectively, according to the manufacturer’s instructions (Qiagen, Germany). Exons 1–15 of the NF2 gene including exon–intron boundary sequences were screened using TGGE as described previously (26). Exons showing altered banding patterns on TGGE were re-amplified and, after purification, both strands were sequenced using a PRISM Dye Primer or later a BigDye Sequencing Kit and an automated sequencer ABI373 (ABI, Foster City). In cases where sequencing did not reveal any significant change, the hetero-duplexes were excised from TGGE gels. DNA was extracted from the excised gel pieces, re-amplified, re-analyzed on TGGE, and sequenced. In one patient (patient 39), T vector cloning was employed using a pGEM-T vector (Promega). The exon 2 fragment was amplified directly from several colonies and analyzed by TGGE. Fragments from three colonies had the same altered bands as the original fragment from the patient and were sequenced. Significance of different mutation detection rates in different patient groups was tested using the z-test as suggested by the computer program SigmaStat. Analysis of allelic loss in tumors was performed using six microsatellite markers flanking or within the NF2 gene: CRYB2, D22S193, NF2CA1, NF2CA3, D22S268 and D22S430 (29). One of each primer pair was labeled with an ABI fluorescent dye. PCRs were performed using paired blood–tumor DNA. All six amplified markers (0.5 µl of PCR product each) were pooled and analyzed using an ABI Genetic Analyzer 310. ACKNOWLEDGEMENTS We gratefully acknowledge the assistance of B. Heinrich, the support of M. Westphal and clinical information provided by S. Thomas (Nordstadt-Krankenhaus, Hannover, Germany). We also thank D. Scoles (CSMC, Los Angeles, CA) for reading the manuscript. This study was supported by Wilhelm-SanderStiftung 93.052.1, 93.052.2 and in part by Hamburg Stiftung zur Förderung der Krebskämpfung 116, 127, 130.

2055 Human Genetics, 1998, 7, No. NucleicMolecular Acids Research, 1994, Vol. Vol. 22, No. 1 13 2055 REFERENCES 1. Evans, D.G.R., Huson, S.M., Donnai, D., Neary, W., Blair, V., Teare, D., Newton, V. and Harris, R. (1992) A clinical study of type 2 neurofibromatosis. Q. J. Med., 84, 603–618. 2. Parry, D.M., Eldridge, R., Kaiser-Kupfer, M.I., Bouzas, E.A., Pikus, A. and Patronas, N. (1994) Neurofibromatosis 2 (NF2): clinical characteristics of 63 affected individuals and clinical evidence for hetergeneity. Am. J. Med. Genet., 52, 450–461. 3. Mautner, V.F., Tatagiba, M., Lindenau, M., Fünfsterer, C., Pulst, S.M., Kluwe, L. and Zanella, F.E. (1995) Spinal tumors in patients with neurofibromatosis type 2: MR imaging study of frequency, multiplicity, and variety. Am. J. Roentol., 165, 951–955. 4. Mautner, V.F., Lindenau, M., Baser, M.E., Hazim, W., Tatagiba, M., Haase, W., Samii, M., Wais, R. and Pulst, S.-M. (1996) The neuroimaging and clinical spectrum of neurofibromatosis 2. Neurosurgery, 38, 880–886. 5. Evans, D.G.R., Huson, S.M., Donnai, D., Neary, W., Blair, V., Theare, D., Newton, V., Strachan, T., Ramesden, R. and Harris, R. (1992) A genetic study of type 2 neurofibromatosis in the north west of England and the UK: I prevalence, mutation rate, fitness and confirmation of maternal transmission effect on severity. J. Med. Genet., 29, 841–846. 6. Rouleau, G.A., Merel, P., Lutchman, M., Sanson, M., Zucman, J., Marineau, C., Hoang-Xuan, K., Demczuk, S., Desmaze, C., Plougastel, B., Pulst, S.M., Lenoir, G., Bijlsma, E., Fahsold, R., Dumanski, J., de Jong, P., Parry, D., Eltridge, R., Aurias, A., Delattre, O. and Thomas, G. (1993) Alteration in a new gene encoding a putative membrane-organizing protein causes neurofibromatosis type 2. Nature, 363, 515–521. 7. Trofatter, J.A., MacCollin, M.M., Rutter, J.L., Murrell, J.R., Duyao, M.P., Parry, D.M., Eldridge, R., Kley, N., Menon, A.G., Pulaski, K., Haase, V.H., Ambrose, C.M., Munroe, D., Bove, C., Haines, J.L., Martuza, R.L., MacDonald, M.E., Seizinger, B.R., Short, M.P., Buckler, A.J. and Gusella, J.F. (1993) A novel moesin-, ezrin-, radixin-like gene is a candidate for the neurofibromatosis 2 tumor suppressor. Cell, 72, 791–800. 8. Merel, P., Hoang-Xuan, K., Sanson, M., Bijlsma, E., Rouleau, G., LaurentPuig, P., Pulst, S., Baser, M., Lenoir, G., Sterkers, J.M., Philippon, J., Resche, F., Mautner, V.F., Fischer, G., Hulsebos, T., Aurias, A., Delattre, O. and Thomas, G. (1995) Screening for germ-line mutations in the NF2 gene. Genes Chromosomes Cancer, 12, 117–127. 9. Kluwe, L., Bayer, S., Baser, M.E., Hazim, W., Haase, W., Fünfsterer, C. and Mautner, V.F. (1996) Identification of NF2 germ-line mutations and comparison with neurofibromatosis 2 phenotypes. Hum. Genet., 98, 534–538. 10. Parry, D.M., MacCollin, M., Kaiser-Kupfer, M.I., Pulaski, K., Nicholson, H.S., Bolesta, M., Eldridge, R. and Gussela, J.F. (1996) Germ-line mutations in the neurofibromatosis 2 gene: correlations with disease severity and retinal abnormalities. Am. J. Hum. Genet., 59, 529–539. 11. Ruttledge, M.H., Andermann, A.A., Phelan, C.M., Claudio, J.O., Han, F.-y., Chretien, N., Rangaratnam, S., MacCollin, M., Short, P., Parry, D., Michels, V., Riccardi, V., Weksberg, R., Kitamura, K., Brandburn, J.M., Hall, B.D., Propping, P. and Rouleau, G.A. (1996) Type of mutation in the neurofibromatosis type 2 gene (NF2) frequently determines severity of disease. Am. J. Hum. Genet., 59, 331–342. 12. Evans, D.G.R., Trueman, L., Wallace, A., Collins, S. and Strachan, T. (1997) Genotype/phenotype correlations in type 2 neurofibromatosis (NF2): evidence for more severe disease associated with truncating mutations. J. Med. Genet., 35, 450–455. 13. Jacoby, L.B., MacCollin, M.M., Louis, D.N., Mohney, T., Rubio, M.-P., Pulaski, K., Trofatter, J.A., Kley, N., Seizinger, B.R., Ramesh, V. and Gusella, J.F. (1994) Exon scanning for mutation of the NF2 gene in schwannomas. Hum. Mol. Genet., 3, 413–419.

14. Jacoby, L.B., MacCollin, M., Barone, R., Ramesh, V. and Gusella, J.F. (1996) Frequency and distribution of NF2 mutations in schwannomas. Genes Chromosomes Cancer, 17, 45–55. 15. Ruttledge, M.H., Sarrazin, J., Rangaratnam, S., Phelan, C.M., Twist, E., Merel, P., Delattre, O., Thomas, G., Nordenskjöld, M., Collins, V.P., Dumanski, J.P. and Roulear, G.A. (1994) Evidence for the complete inactivation of the NF2 gene in the majority of sporadic meningiomas. Nature Genet., 6, 180–184. 16. Wellenreuther, R., Kraus, J.A., Lenartz, D., Menon, A.G., Schramm, J., Louis, D.N., Ramesh, V., Gusella, J.F., Wiestler, O.D. and von Deimling, A. (1996) Analysis of the neurofibromatosis 2 gene reveals molecular variants of meningioma. Am. J. Pathol., 146, 827–832. 17. Bernards, A. and Gusella, J.F. (1994) The importance of genetic mosaicism in human disease. N. Engl. J. Med., 331, 1447–1449. 18. Colman, S.D., Rasmussen, S.A., Ho, V.T., Abernathy, C.R. and Wallace, M.R. (1996) Somatic mosaicism in a patient with neurofibromatosis type 1. Am. J. Hum. Genet., 58, 484–490. 19. Ainsworth, P.J., Chakraborty, P.K. and Weksberg, R. (1997) Example of somatic mosaicism in a series of de novo neurofibromatosis type 1 cases due to a maternally derived deletion. Hum. Mutat., 9, 452–457. 20. Wu, B.L., Boles, R.G., Yaari, H., Wermowicz, S., Schneider, G.H. and Korf, B.R. (1997) Somatic mosaicism for deletion of the entire NF1 gene identified by FISH. Hum. Genet., 99, 209–213. 21. Lázaro, C., Ravella, A., Gaona, A., Bolpini, V. and Estivill, X. (1994) Neurofibromatosis type 1 due to germ-line mosaicism in a clinically normal father. N. Engl. J. Med., 331, 1403–1407. 22. Putnam, E.A., Park, E.S., Aalfs, C.M., Hennekam, R.C. and Milewicz, D.M. (1997) Parental somatic and germ-line mosaicism for a FBN2 mutation and analysis of FBN2 transcript levels in dermal fibroblasts. Am. J. Hum. Genet., 60, 818–827. 23. Holterhus, P.-M., Brüggenwirth, H.T., Hiort, O., Kleinkauf-Houcken, A., Krose, K., Sinnecker, G.H.G. and Brinkmann, A.O. (1997) Mosaicism due to a somatic mutation of the androgen receptor gene determines phenotype in androgen insensitivity syndrome. J. Clin. Endocrinol. Metab., 82, 3584–3589. 24. Bourn, D., Carter, S.A., Evans, D.G.R., Goodship, J., Coakham, H. and Strachan, T. (1994) A mutation in the neurofibromatosis type 2 tumorsuppressor gene giving rise to widely different clinical phenotypes in two unrelated individuals. Am. J. Hum. Genet., 55, 69–73. 25. Bijsma, E.K., Wallace, A.F. and Evans, D.G. (1997) Misleading linkage results in an NF2 presymptomatic test owing to mosaicism. J. Med. Genet., 34, 934–936. 26. Kluwe, L., MacCollin, M., Tatagiba, M., Thomas, S., Hazim, W., Haase, W. and Mautner, V.F. (1998) Phenotypic variability associated with 14 splice-site mutations in the NF2 gene. Am. J. Med. Genet., 77, 228–233. 27. Gutman, D.H., Aylsworth, A., Carey, J.C., Korf, B., Marks, J., Pyertz, R.E., Rubenstein, A. and Viskochil, D. (1997) The diagnostic evaluation and multidisciplinary management of neurofibromatosis 1 and neurofibromatosis 2. J. Am. Med. Assoc., 278, 51–57. 28. Kluwe, L., Pulst, S.M., Köppen, J. and Mautner, V.F. (1995) A 163-bp deletion at the C-terminus of the schwannomin gene associated with variable phenotypes of neurofibromatosis type 2. Hum. Genet., 95, 443–446. 29. Evans, D.G.R., Mason, S., Huson, S.M., Ponder, M., Harding, A.E. and Strachan, T. (1997) Spinal and cutaneous schwannomatosis is a variant form of type 2 neurofibromatosis: a clinical and molecular study. J. Neurol. Neurosurg. Psychiatry, 62, 361–366.