Hum Genet (1992) 90:356-359 human .. genetics

�9 Springer-Verlag 1992

Analysis of segregation and expression of an identified mutation at the neurofibromatosis type 1 locus

Markus Stark, Giinter Assum, Dieter Kaufmann, Hildegard Kehrer, Winfrid Krone

Abteilung Humangenetik, Universit~t Ulm, Albert-Einstein-Allee l 1, W-7900 Ulm, Federal Republic of Germany

Received: 28 July 1992 / Revised: l October 1992

Abstract. A previously identified complex mutat ion, af- fecting exon 28 of the neurof ibromatos is type 1 gene, was employed for the analysis of the expression pat tern in pr imary cultures of neurof ib roma cells and melano- cytes f rom a cafd-au-lait macule of the patient, respec- tively. Reverse transcription and subsequent polymerase chain react ion amplification of the segment carrying the muta t ion revealed that both alleles were expressed in both cell types analysed, thus excluding loss of hetero- zygosity in this part icular instance. Segregation of the al- leles of the intragenic Alu sequence length-polymorphism disclosed the paternal orgin of the muta ted allele. Detec- tion of this muta t ion was also used for presymptomat ic direct D N A diagnosis in the younger child of the patient.

natal diagnosis (Ward et al. 1990: Lazaro et al. 1992) and for the identification of the parental origin of the defective gene of sporadic patients (Jadayel et al. 1990; Stephens et al. 1992), identified mutat ions provide (besides direct D N A diagnosis) the additional oppor tuni ty to examine the expression of the two alleles in both unaffected and affected cell populat ions. In the present study we have used a 5-bp deletion with an adjacent base exchange, character ized previously (Stark et al. 1991), for ascer- taining the presymptomat ic diagnosis in an unaffected child of the patient , and to examine the expression of both alleles in cultured melanocytes f rom a caf~-au-lait macule ( C A L M ) and in fibroblast-like cells f rom a neu- rof ibroma.

Introduction

The cloning and sequencing of the complete c D N A and of parts of the non-coding regions of the gene whose de- fects cause neurof ibromatos is 1 (NF1; for review see Wallace and Collins, 1991) has opened many new pos- sibilities for basic and applied investigations of the nu- merous problems connected with this disease. The amino acid sequence deduced from the c D N A sequence revealed homology of a segment spanning about 360 amino acids to the functional domain of the p21ras-GTPase-act ivat- ing protein p l 2 0 G A P , and this kind of activity of the NF1 gene product , now designated as neurof ibromin, has been demons t ra ted (Ballester et al. 1990; Martin et al. 1990). One line of investigation of high priority is the search for muta t ions of the NF1 gene and the precise character izat ion of their structure and their phenotypical consequences . While a number of closely flanking mark- ers of the NF1 gene have already been utilized for pre-

This paper is dedicated with deep respect to Ulrich Wolf on his 60th birthday on 2 January 1993 Correspondence to: G. Assum

Materials and methods

The symptoms of the 35-year-old patient (II-2 in Fig. la) met the criteria of NFI. Cell cultures of a neurofibroma and from adjacent skin were prepared from fragmented tissue directly, or after, treat- ment with collagenase. A CALM and normally pigmented skin of the patient served to obtain cultures of melanocytes according to Halaban et al. (1986), which were propagated as described by Kaufmann et al. (1991). Preparation of metaphase slides and dif- ferential staining by GTG-banding for cytogenetic analysis of fi- broblast-like cells and peripheral blood cells was performed by standard procedures. Partial synchronization by release from a 24- h thymidine block was required to increase the yield of metaphases from the slowly growing melanocytcs.

DNA was isolated form peripheral blood cells according to the method described by Miller et al. (1988). Exon 28 [formerly exon 1 in Cawthon et al. (1990)] of the NFI gene was amplified by poly- merase chain reaction (PCR) under standard conditions (lnnis et al. 1990) and subsequently subjected to heteroduplex analysis as described by Keen et al. (199l). The fragments were visualized by silver staining (Merril et al. 1981). For PCR amplification the fol- lowing primers, complementary to intronic sequences immediately flanking the exon, were employed: Ex28L: TCTJTGTCTTTJWTG- TCATTTTCC; Ex28R: AGTCAAGAAAAGCAATGAATCGT. Thirty amplification cycles were run with denaturation at 93~ for 1 min, annealing at 55~ for I rain and extension at 72~ for I min.

To investigate thc question of loss of heterozygosity (LOH) in CALM, melanocytes were grown in 96-well multiplatcs to a densi-

357

ty of about 10000 cells per well and subsequently washed with PBS. Then 19.3 lal Diethyl-pyrocarbonate (DEPC)-treated water, 0.2 ~1 Dithiothreitol (DTT, 100 mM) and 0.5 lal RNAsin (Promega, Madison, Wis.) were added and cells were freeze-thawn several times with liquid nitrogen until complete cell lysis was achieved. The lysate was centrifuged, the pellet was removed, and the super- natant was used to perform reverse transcription of RNA with the Super Script Preamplification System (BRL, Gaithersburg, Md) according to the supplier's protocol (Eisenbarth et al., manuscript in preparation). We used 15 gl of the resulting cDNA solution as template in a 50-~1 PCR (93~ 1 min/54~ 1 min/72~ i rain/30 cycles) with the following primers flanking exon 28: NF3750: ACGTTAAGTATTTTCTACCAAGC; NF4242: TACTTGGA- CAGCAGTAGAACC. The resulting products were analysed with the heteroduplex method and BglI restriction.

Results and discussion

The family history of the index patient NF71 (II-2 in Fig. la ) did not reveal any previous case of NF1. His par- ents were found to be free of signs of the disease upon a thorough dermatological examination. The 35-year-old patient NF71, whose mutation was identified as a 5-bp deletion with an adjacent base exchange (Stark et al. 1991) exhibited the typical symptoms of NFI: more than 100 cutaneous and subcutaneous neurofibromas, multi- ple CALMs, axillary freckles and Lisch nodules bilater- ally. In addition he had two pigmented nevocellular naevi on this back, he was not, however, affected by one of the more severe complications of the disease. The patient had transmitted the mutation to his daughter (III-1), who, at 5 years of age, showed multiple CALMs and Lisch nodules in both eyes.

Fig. 1. a Pedigree of the family of patient NF71 (11-2). b Heterodu- plex analysis of PCR-amplified exon 28 of the NF1 gene from the members of family NF71, showing two additional bands in the samples of the affected individuals, c Direct detection of the NF1- mutation on the basis of an altered BglI cleavage site

The patient's peripheral lymphocytes exhibited a nor- mal karyotype on examination of 50 GTG-banded meta- phases, with one grossly abnormal cell (44 ,XY, - 17 , - 19, +dmin; 4ctb). Thirty lymphcoyte metaphases from his affected daughter similarly analysed had a normal karyo- type. Cultured fiboblast-like cells from a cutaneous neu- rofibroma of the propositus, examined cytogenetically at in vitro passage no. 2, showed the unusually high rate of chromosomal breakage of 24% (9 ctb, 1 csb, 2 fragments; n = 50). In cultured melanocytes derived from the skin overlying a neurofibroma only 4% breaks were found. A cytogeneticaUy recognizable "second event" was not found upon analysis of an additional 20 neurofibroma- derived metaphases, screened at the 500-550 bands level after exposure of the culture to ethidium bromide.

The younger child (III-2) had two small pigment spots of unknown origin at age 18 months. Heteroduplex anal- ysis of the PCR-amplified exon 28 affected by the muta- tion showed, that this boy had not inherited the defec- tive allele (Fig. lb) . Thus, the identified mutation could immediately be utilized for presymptomatic diagnosis in this family. By the same method, it could also be shown that the healthy brothers of the patient were homozy- gous for the normal allele at the NF1 locus (Fig. lb , lanes 7 and 8), as were his parents (lanes 1 and 2). Since the mutation eliminates a BglI restriction site, the geno- types could also be ascertained after BglI digestion of the amplified exon (Fig. lc).

To identify the parental origin of the mutation of this sporadic patient, the segregation of the intragenic Alu- polymorphism described by Xu et al. (1991) was analysed. The pattern observed allows the conclusion that the mu- tation had occurred in a paternal germ cell (data not shown). Assuming in this informative situation a maxi- mum probability of crossover of 0.003 between the two intragenic markers (the NF1 mutation and the Alu-poly- morphism), the likelihood of paternal origin of the mu- tation is 99.7%. Since this polymorphic site is located near the breakpoint of the t(1;17) which, like our pa- tient's mutation, lies a short distance downstream of the GAP-related functional domain (GRD) , the distance between these two sites must be much smaller than the putative length of the NF1 gene (300 kb). Therefore, the likelihood of a recombination even occurring between these two sites can be assumed to be negligible. Taking together the results of the two studies addressed to the problem of the parental origin of the mutant allele in sporadic NF1 patients (Jadayel et al. 1990; Stephens et al. 1992) with our case and with the paternal Alu inser- tion de novo reported by Wallace et al. (1991), paternal origin occurred in 24 of the 26 cases examined.

The mechanisms generating the benign lesions char- acteristic of NF1, such as neurofibromas, CALM and Lisch nodules, are still unknown. While loss of the nor- mal allele at the NF1 locus, detected as loss of hetero- zygosity (LOH) for flanking and/or intragenic markers, seems to be involved in the causation of neurofibrosar- comas in NF1 patients (Skuse et al. 1989; De Clue et al. 1992), the role of L O H as a possible cause of the benign lesions remains to be investigated. Because of their cellu- lar heterogeneity this is a difficult task in the case of neu-

358

Fig.2. a Control melanocytes; b fibroblasts from neurofibroma; c melanocytes from CALM. Expression of both alleles of the NF1 locus in cultured fibroblasts from a neurofibroma and in CALM melanocytes from patient NF71 as demonstrated by heteroduplex analysis of a PCR-amplified cDNA fragment spanning exon 28

rofibromas. Pure cell cultures can, however, be obtained of melanocytes, which are the progenitor cells of CALM. In fact, a number of morphological and biochemical pa- rameters have been shown to be altered in both melano- cytes derived from the normally pigmented skin of NF1 patients and in those from CALMs (Martuza et al. 1985; Kaufmann et al. 1989, 1991). In this study CALM mela- nocytes and fibroblast-like cells from a neurofibroma of the patient NF71 were examined for the occurrence of L O H at the NF1 locus. Starting from reverse transcribed RNA preparations, exon 28 was amplified by PCR. Fig- ure 2 shows the result of heteroduplex analyses perform- ed with the D N A fragments so obtained. Both alleles of the NF1 locus were clearly expressed in both types of cells in a quantitatively comparable level. This could also be demonstrated by BglI-digestion of the amplification product which yields the pattern expected for hetero- zygosity (data not shown).

The two-step model of tumorigenesis which postu- lates (in its conservative version) the occurrence of two consecutive mutations at the alleles of a recessive tumor suppressor gene (Knudson 1971; Ohno 1971; Comings 1973), provides a framework for hypotheses on tumori- genesis also in neurofibromatosis. In a naive, straightfor- ward application of this model to NF1, the homozygous loss or functional inactivation of the neurofibromin gene would be necessary and sufficient to generate the pro- genitor cell of a neurofibrosarcoma or a rhabdomyosar- coma. Explanation of how the benign lesions, like neu- rofibromas, Lisch nodules and CALM arise would then require assumed second events different from the com- plete functional elimination of the wild type allele of the NF1 locus. Somatic mutations at different gene loci may, in this context, be just as important as epigenetic regulat- ory mechanisms stimulated (or inhibited) by external in- fluences, like trauma or the random fluctuations of the cellular microenvironment (Riccardi 1979). It is also conceivable that the generation of the benign lesions re- quires homozygous deficiency at the NF1 locus and addi- tional steps, such as mutations of the p53 gene, to bring about the malignant degeneration (Menon et al. 1990). Studies on the monoclonality of the tumors, and those of LOH, are instrumental in deciding on some of these

alternatives. While both of these issues, monoclonality and LOH, are difficult to ascertain in neurofibromas and in primary cultures derived from them, because of the heterogeneity of their cellular composition, primary cul- tures of melanocytes from a CALM should represent a uniform cell population with regard to the expression of an NFl-related defect. Since only exon 28 was exam- ined, our finding of the expression of both alleles of the NF1 gene in CALM melanocytes does not exclude the occurrence of a second mutation somewhere else in the wild type copy, resulting in compound somatic hetero- zygosity. The complete loss, or the failure to be expres- sed, of the wild-type allele does not, however, seem to be a prerequisite for the generation of a CALM in NF1.

The second event that causes the disseminated CALM to emerge must not necessarily occur at the NF1 locus it- self. in fact statistical evidence was recently provided (Easton et al., submitted) that modifying genes may con- tribute significantly to the variable expressivity of NF1, and that the modifiers influencing CALM may be differ- ent from those affecting neurofibroma formation.

Acknowledgements. Once again we want to thank the patient and his family for their kind cooperation. Excellent technical assistance by Ulrike Stehle, Annegret Schulz, and Anna Siegel is greatfully acknowledged. This work was supported by the Deutsche For- schungsgemeinschaft (Kr 362/17).

References

Ballester R, Marchuk D, Boguski M, Saulino A, Letcher R, Wig- ler M, Collins F (1990) The NF1 locus encodes a protein func- tionally related to mammalian GAP and yeast IRA proteins Cell 63:851-859

Cawthon RM, Weiss R, Xu G, Viskochil D, Culver M, Stevens J. Robertson M, Dunn D, Gesteland R, O'Connell P, White R (1990) A major segment of the neurofibromatosis type 1 gene: cDNA sequence, genomic structure, and point mutations. Cell 62 : 193-20l

Comings D (1973) A general theory of carcinogenesis. Proc Natl Acad Sci (USA) 70:3324-3328

DeClue JE, Papageorge AG, Fletcher JA, Diehl SR, Rather N, Vass WC, Lowy DR (1992) Abnormal regulation of mamma- lian p21ras contributes to malignant tumor growth in yon Reck- linghausen (type 1) neurofibromatosis. Cell 69:265-273

Easton DF, Ponder MA, Huson SM, Ponder BAJ (1992) An anal- ysis of variation in expression of neurofibromatosis type 1: evi- dence for modifying genes. Am J Hum Genet (submitted)

Halaban R, Gosh S, Duray P, Kirkwood JM, Lerner AB (1986) Human melanocvtes cultured from nevi and melanomas. J In- vest Dermatol 87:95-101

Innis MA, Gelfand DH, Sninsky J J, White TJ (eds) (1990) PCR- protocols: a guide to methods and applications. Academic Press. New York London

Jadayel D, Fain P, Upadhyaya M, Ponder MA, Huson SM, Carey J, Fryer A, Mathew CGP, Barker DF, Ponder BAJ (1990) Pa- ternal origin of new mutations in von Recklinghausen neurofi- bromatosis. Nature 343 : 558-559

Kaufmann D, Krone W, Hochsattel R, Martin R (1989) A cell cul- ture study on melanocytes from patients with neurofibroma- tosis 1. Arch Dermatol Res 281:510-513

Kaufmann D, Wiandt S, Veser J, Krone W (1991) Increased melanogenesis in cultured epidermal melanocytes from pa- tients with neurofibromatosis 1 (NF1). Hum Genet 87:144- 150

359

Keen J, Lester D, Inglehearn C, Curtis A, Bhattacharya S (1991) Rapid detection of single base mismatches as heteroduplexes on hydrolink gels. Trends Genet 7 : 5

Knudson AG jr (1971) Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci USA 68 : 820-823

Lazaro C, Ravella A, Casals T, Volpini V, Estivill X (1992) Pre- natal diagnosis of sporadic neurofibromatosis 1. Lancet 339: 119-120

Martin GA, Viskochil D, Bollag G, McCabe PC, Crosier WJ, Haubruck H, Conroy L, Clark R, O'Conell P, Cawthon RM, Innis MA, McCormick F (1990) The GAP-related domain of the neurofibromatosis type 1 gene product interacts with ras p21. Cell 63 : 843-849

Martuza RL, Philippe I, Fitzpatrick TB, Zwaan J, Keki Y, Leder- man J (1985) Melanin macroglobules as a cellular marker of neurofibromatosis: a quantitative study. J Invest Dermatol 85 : 347-350

Menon AG, Anderson KM, Riccardi VM, Chung RY, Whaley JM, Yandell DW, Farmer GE, Freimann RN, Lee JK, Li FP, Barker DF, Ledbetter DH, Kleider A, Martuza RL, Gusella JF, Seizinger BR (1990) Chromosome 17p deletions and p53 gene mutations associated with the formation of malignant neurofibrosarcomas in yon Recklinghausen neurofibromatosis. Proc Natl Acad Sci USA 87 : 5435-5439

Merril CR, Goldman D, Sedman SA (1981) Ultrasensitive stain for proteins in poylacrylamid gel shows regional variation in ce- rebrospinal fluid proteins. Science 211:1437-1438

Miller GA, Dykes DD, Polesky JF (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16 : 1215

Ohno S (1971) Genetic implication of karyological instability of malignant somatic cells. Physiol Rev 51:496-526

Riccardi VM (1979) Cell-cell interaction as on epigenetic determi- nant in the expression of mutant neural crest cells. Birth De- fects 15 : 89-98

Skuse GR, Kosciolek BA, Rowley PT (1989) Molecular genetic analysis of tumors in von Recklinghausen neurofibromatosis: loss of heterozygosity for chromosome 17. Genes Chromos Cancer 1 : 36-41

Stark M, Assure G, Krone W (1991) A small deletion and an adja- cent base exchange in a potential stem-loop region of the neu- rofibromatosis 1 gene. Hum Genet 87 : 685-687

Stephens K, Kayes L, Riccardi VM, Rising M, Sybert PV, Pagon RA (1992) Preferential mutation of the neurofibromatosis type 1 gene in paternally derived chromosomes. Hum Genet 88 : 279-282

Wallace MR, Collins FS (1991) Molecular genetics of von Reck- linghausen neurofibromatosis. Hum Genet 20 : 267-307

Wallace MR, Anderson LB, Saulino AM, Gregory PE, Glover TW, Collins FS (1991) A de novo Alu insertion results in neurofibromatosis type 1. Nature 353 : 864-866

Ward K, O'Connel P, Carey JC, Leppert M, Jolley S, Plaetke R, Ogden B, White R (1990) Diagnosis of neurofibromatosis 1 by using tightly linked, flanking DNA markers. Am J Hum Genet 46 : 943-949

Xu G, Nelson L, O'Connell P, White R (1991) An Alu poly- morphism intragenic to the neurofibromatosis type 1 gene (NF1), Nucleic Acids Res 19 : 13