Benign F d d Neonatal Convulsions: Evidence for Clinical and Genetic Heterogeneity Stephen G. Ryan, MD,* Max Wiznitzer, MDJ Charlotte Hollman, MD,$ hf. Cristina Torres, BS,“ Maria Szekeresova, &IS,? and Sandra Schneider, P h D i
The gene for autosomal dominant “benign” familial neonatal convulsions, a transient, primary epilepsy of infancy, has recently been assigned to chromosome 20q. To determine whether this disorder is genetically heterogeneous, we performed linkage analysis in two previously unreported pedigrees with benign familial neonatal convulsions in which clinical heterogeneity was evident. There were 14 affected persons in the first family, and none had seizures (febrile or afebrile) after the age of 2 months. The second family had 13 affected individuals and 2 obligate carriers; seizures frequently did not remit until 6 to 24 months, febrile convulsions occurred in at least 2 patients, apparent audiogenic seizures occurred in 4 patients, and 1 individual had refractory epilepsy until late adolescence. Linkage studies with the chromosome 20 markers D20S19 and D20S20 were performed in both families. The resulting data favored linkage of the disease and marker loci in Family 2 by a maximum odds ratio of 45:l at 6% recombination. In Family 1, however, the odds were greater than 20,OOO:l against linkage at 10% recombination or less. We conclude that the syndrome of benign familial neonatal convulsions is clinically and genetically heterogeneous. Further study will be necessary to clarify the relationship between phenotype and genotype in this disorder. Ryan SG, Wiznitzer M, Hollman C, Torres MC, Szekeresova M, Schneider S. Benign familial neonatal convulsions: evidence for clinical and genetic heterogeneity. Ann Neurol 1991;29:469-473
Genetic influences are thought to contribute to many forms of human epilepsy [l, 2}. Indeed, among the so-called “secondary” epilepsies are numerous mendelian or single-gene disorders in which seizures are symptomatic of a fixed or progressive encephalopathy El}. Most primary epilepsies, however, are probably multifactorial in etiology [ 3 } , and the mechanisms whereby genetic factors cause seizures in these disorders are not understood. The syndrome of benign familial neonatal convulsions (BFNC) is a rare example of a primary epilepsy {4} in which inheritance is unequivocally autosomal dominant 15-91. BFNC is characterized by the occurrence of unprovoked partial or generalized c h i c seizures in the neonatal period or early infancy. Results of routine diagnostic studies (including cranial computed tomography, interictal electroencephalography (EEG), and serum chemistry determinations) and subsequent intellectual development are normal. Seizures usually remit by the age of 6 months, but chronic childhood or adult epilepsy has been reported in about 10% of patients [7, 91.
Recently, the gene for BFNC was assigned to chromosome 20q by the detection of tight linkage between the BFNC locus and two polymorphic D N A markers, D20S19 and D20S20, in a single large pedigree [lo}. Subsequent studies in two additional families supported this finding, and raised the odds ratio favoring linkage between the disease and marker loci to greater than 8,000,000:l Ell}. We recently identified two large and previously unreported BFNC pedigrees whose clinical features suggested that the disorder was heterogeneous. We therefore performed linkage analysis in these families to determine whether distinct genetic loci were involved.
From the Departments of *Pediatrics and tperiodontics, The University of Texas Health Science Center, San Antonio, TX, and the $Departments of Pediatrics and Neurology, Case Western Reserve University, Cleveland, OH.
Received Sep 11, 1990, and in revked form Nov 1. Accepted for publication Nov 4,1990.
Materials and Methods The families were identified through the probands whose case histories are provided below. The parents (where available) of all individuals “at risk’ for BFNC were interviewed. Medical records were available in only a few instances, and in most cases the age at onset and remission were not precisely recalled. Nonetheless, for most individuals, we were able to obtain a clear history regarding: ( 1 ) the presence or absence
Address correspondence to Dr Ryan, Department of Pediatrics, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, TX 78284.
Copyright 0 1991 by the American Neurological Association 469
of seizures in the neonatal period or in infancy, (2) whether seizures persisted past the ages of 2 or 12 months, and ( 3 ) whether or not at least one febrile seizure had occurred. Genomic D N A was prepared by phenolichloroform extraction of lysed, Epstein-Barr virus-transformed peripheral blood lymphoblastoid cell lines El2, 131. The samples were digested with the restriction enzyme TaqI (New England Biolabs, Beverly, MA) and the resulting fragments were fractionated by 0.7% agarose gel electrophoresis and transferred by capillary action to nylon membranes (Nytran, Schleicher and Schuell, Keene, NH) according to protocols supplied by the manufacturers. Following prehybridization, the blots were hybridized with the radiolabelled [ 141 markers D20S 19 (pCMM6) [15] and D20S20 (pRMR6) 1161(purchased from the American Type Culture Collection, Rockville, MD) and washed in solutions of increasing stringency as suggested by the manufacturer. These two probes reveal D N A restriction fragment polymorphisms due to a variable number of tandem repeats of a short D N A sequence (D20S19) and the presence or absence of a TayI restriction site (D20S20). Because polymorphisms were poorly demonstrated with the latter probe, we tested fragments obtained from simultaneous digestion of the pRMR6 plasmid with Hind111 and EcoRI. A 0.7-kb fragment revealed the polymorphic system clearly and was used for typing. Autoradiograms of the membranes were then prepared [12), and marker genotypes assigned to each individual. D20S19 and D20S20 are closely linked, with an estimated recombination fraction of 0.016 [171. Since no recombination between the two markers was observed in our families, it was possible to assign a combined marker haplotype to each individual. Two-point linkage analysis was performed with the program LIPED [IS] on an IBM-XT compatible personal computer. A penetrance of 90% was assumed for the BFNC allele, based on a segregation ratio of 45% reported previously [lo].
Family 1 The proband was a healthy, term neonate of MexicanAmerican ancestry, whose vaginal delivery followed an uncomplicated gestation. O n the ninth postnatal day, she had four clonic convulsions, each about 2 minutes long: Two were generalized, one was left-sided, and one was right-sided. Physical examination results, blood glucose levels, routine serum electrolyte levels (including calcium and magnesium), and an EEG were all normal. Neuroimaging was deferred because at least 3 affected relatives had had normal findings on cranial computed tomography. Phenobarbital was begun at a total daily dose of 5 mgikg, and was discontinued by the mother a month later. At the time of writing, the patient was 9 months old and developmentally normal, and had had no recurrence of seizures. Thirteen of the proband's relatives had neonatal seizures beginning between the second and fourteenth postnatal day (Fig 1, Table 1). No one in Generations I and 11 received anticonvulsant therapy, but all of the proband's affected cousins (Generation 111) received phenobarbital for up to 4 months. None had seizures while taking phenobarbital, and no seizures occurred in any family member after the age of 2 months, regardless of treatment status. Intellectual development appeared normal in all family members. 470 Annals of Neurology
V a l 29 N o 5
May 1991
I
Fig 1. Pedigree and genotypic data for Family 1 (arrow indicates proband). The D20S19 alleles are denoted by letters and the D20S20 alleles by numbers. Numerous instances of recombination between the disease and marher alleles are &dent.
Table 1. Clinical Features Suggestive of Heterogeneity in Two Pedigrees with Benign Familial Neonatal Convulsions
Persistence of seizures beyond 12 months Epilepsy in late childhoodiadolescence Audiogenic seizures Febrile seizures Asymptomatic obligate carriers
Family 1 (n = 14)
Family 2 (n = 15)
0
4
0
1
0 0 0
4 2 2
Family 2 The proband for Family 2 was born to parents of northern European ancestry via term, spontaneous vaginal delivery following an uncomplicated gestation. O n the second postnatal day, she had 11 partial clonic seizures. A loading dose of phenytoin was administered, and maintenance phenobarbital was begun at a dose of 5 mgikgiday. Physical examination findings, cranial computed tomograms, serum chemistry measurements, and EEGs were all normal. At least two additional seizures occurred during the next 3 weeks despite phenobarbital treatment (serum levels were not measured), but subsequently she was seizure-free. At 10 months old she was developmentally normal, and continued to take phenobarbital. Twelve relatives of the proband had a clear history of idiopathic neonatal convulsions (Fig 2;see Table 1).There were also two obligate carriers: Patients 11-9 and either 1-1 or 1-2. One infant (IV-3) had severe birth asphyxia due to placental abruption. Neither his neonatal seizures nor his subsequent microcephaly and developmental delay were thought to be manifestations of BFNC, but he was omitted from the linkage analysis because affectation status for BFNC was deemed uncertain. One relative (111-3) continued to have frequent convulsions until late adolescence despite phenobarbital therapy; his management was complicated by noncompliance and at least three episodes of convulsive status epilepticus. Although of borderline intelligence, he was fully employed at the time of writing. All other family members (except IV-3) appeared to be intellectually normal.
Table 2. Results oJ Linkage Analysis h d Scores
Recombination Fraction (Sex-Averaged)
Fig 2.Pedigree and genotypic data for Family 2 (arrow indicates proband). Marker alleles are denoted as in Figure I . Recombination between the disease and muker loci appears to have occurred in individuals 111-1 and III-16, but nonpenetrance could a h explain their marher genotypes. Patient IV-3 was omitted from the analysis because of uncertainty regarding his phenotype (see text).
Four affected patients (11-14, 111-3, 111-15, and IV-6) had a history strongly suggestive of audiogenic seizures. For example, in one 2’/2-month-old infant (IV-6) a prolonged (longer than 10 minutes) clonic seizure was apparently precipitated by the explosion of a firecracker. Seizures induced by such stimuli occurred in both sleeping and walung states. At least 4 affected relatives (11-14, 111-3, 111-6, and 111-8) continued to have afebrile seizures until the age of 12 months or beyond. At least 2 affected relatives (111-3 and IV-5) also had o n e or more febrile convulsions between the ages of 6 and 18 months. EEG results were available only for IV-5, IV-6, and IV-7; all three tracings were normal.
Results The marker genotypes are provided in Figs 1 and 2, and lod scores* from two-point linkage analysis are presented in Table 2. Both markers were informative in Family 1, but only D20S19 was informative in Pamily 2, as no affected parents were heterozygous for D20S20. The results were not influenced by the population frequencies of the marker alleles except for slight variation, in Family 2, as a function of the frequency of D20S19 allele B (data not shown). A value of 0.33 for this parameter, based on a survey of 27 unrelated individuals, was employed for the lod score computations in Table 2. For Family 1, the data strongly favored nonlinkage over linkage of the disease and marker loci; the odds against linkage at 10% recombination, for example, are greater than 21,OOO:l (lod score, -4.34). For Family 2, a maximal odds ratio of 45: 1 (lod score, 1.66) favoring linkage occurred at 6% recombina-
+
‘A Iod (“logarithm of odds”) score is a measure of the likelihood that a set of genotypic observations results froni linkage as opposed to chance. A lod score of 3.0, corresponding to a 1,000:1 odds ratio, is generally accepted as establishing linkage.
Family 1
D20S19
0
-32
0.01 0.05 0.10 0.20 0.30 0.40
-6.91 -3.86 -2.47 -1.10 -0.43 -0.09
D20S20 -Ic
-6.08 -3.26 -2.05 -0.93 -0.38 -0.10
Two-Marker Haplotype -02
-11.30 -6.61 -4.34 -2.12 -0.97 -0.32
Family 2 D20S19
1.58 1.61 1.66 1.63 1.40 1.00 0.48
tion, with a one-lod confidence interval of 1 to 37%. When sex-specific recombination fractions were allowed to vary independently, a maximum lod score of 1.67 occurred at male and female rates of 10 and 396, respectively. Numerous instances of recombination between the disease and marker loci are evident in Family 1. In Family 2, only 2 individuals (aside from the obligate carriers) failed to display the BFNC phenotype despite inheritance of the linked haplotype. The carrier status of these individuals (111-1 and 111-16) is uncertain, since nonpenetrance cannot be distinguished from recombination between the disease and marker loci. No a€fected individuals lacked the linked haplotype.
Discussion Numerous reports of BFNC have appeared since the original description of Rett and TeubelT5-9, 191. The clinical picture is distinctive and relatively uniform, although it has been suggested that differences among pedigrees with regard to risk for epilepsy in later life might reflect genetic heterogeneity. Shevell and associates {9] reviewed the clinical features of 113 patients with BFNC in 15 families; 11 of 62 individuals in 6 families experienced nonfebrile convulsions beyond infancy, while the 5 1 individuals in the remaining 9 families experienced early and complete remission of seizures. They observed that the data were compatible with “two forms [of BFNC] with sharply different risks for subsequent epilepsy.” Similarly, Leppert and associates {lo] noted that the development of nonfebrile seizures after the age of 6 months may be restricted to certain pedigrees. Unfortunately, most published reports of BFNC do not provide detailed information on age at remission of seizures, presumably because of difficulty in obtaining reliable data. We encountered uncertainty with regard to age at remission as wel, but were nonetheless able to detect a trend toward later remission in Family 2 than in Family 1.
Ryan et al: Familial Neonatal Convulsions 471
Our linkage data strongly support the existence of two distinct genetic loci for BFNC. The maximum lod
score of 1.66 obtained for Family 2 suggests, but does not prove, that the BFNC locus in this kindred is linked to the marker loci. Given previous reports of linkage of BFNC to these markers [lo, 1I), however, it appears quite likely that the disease gene in Family 2 maps to chromosome 20q. In contrast, in Family 1, the data militate strongly against proximity of the disease and marker loci. Given the magnitude of the negative lod scores in Family 1, it is unlikely that erroneous phenotype assignment, nonpaternity, or an inaccurate penetrance estimate could account for these results. The data further suggest that the BFNC subtype linked to chromosome 20q may be associated with later remission and a higher risk for subsequent epilepsy (defined as afebrile seizures after age 2 years) than in the variant of this disorder present in our first family. Based on the original reports 15-71, late epilepsy occurred in at least 3 of 15, 2 of 14, and none of 6 individuals in the three families studied by Leppert and associates [lo]. Additional studies are needed to determine whether there is a consistent relationship between the genetic subtype of BFNC and clinical features such as ethnicity, age at remission, response to anticonvulsant therapy, audiogenic seizures, and the risk of late epilepsy. There are several precedents for the discovery of heterogeneity by linkage analysis of disorders presumed to be uniform. The earliest example is probably that of hereditary elliptocytosis, for which linkage to the Rh blood group locus has been established in some kindreds, but disproved in others r20). Recent examples in clinical neurology include hereditary motor and sensory neuropathy type I (in which loci on chromosome 1 and 17 have been implicated in different families) 121, 221 and tuberous sclerosis (in which linkage to the ABO blood group locus on chromosome 9 has been demonstrated in some families but disproved in others) [23}. The presence of heterogeneity in BFNC may complicate efforts to isolate the 20q gene, since a critical step in such a project will be a search for flanking markers {24]. This entails a search for individuals who demonstrate recombination between the disease and a closely linked marker locus. Given heterogeneity, only recombinants in pedigrees that unequivocally demonstrate linkage-hence, very large pedigrees-can be considered. In other words, the potential information present in small pedigrees is lost because the genetic type cannot be reliably determined. The presence of heterogeneity in BFNC, an epilepsy with highly distinctive clinical features and single-gene inheritance, suggests that disorders such as childhood absence and benign partial childhood (rolandicj epilepsy [4], for both of which dominant inheritance has 472
Annals of Neurology Vol 29
No 5
May 1991
been suggested [25, 261, could also exhibit this phenomenon. If this is the case, then linkage analysis in these disorders, which will almost certainly involve numerous small families, will be particularly challenging. Although such an approach may be successful, as in juvenile myoclonic epilepsy (recently assigned to chromosome Gp) 1271, heterogeneity could easily complicate efforts to assign genes for epilepsies that rarely, if ever, segregate in extensive pedigrees. Without firm clinical criteria for classifying families, evidence supporting linkage in some families could be negated by evidence against it in others. We have been able to demonstrate heterogeneity in BFNC because large families were available for study and because two markers (one moderately and the other highly polymorphic) mapping to the region of interest were also available. It is likely that phenotypic variation among BFNC families is directly related to genotypic differences, and may be more common than previously appreciated. Further study should confirm or refute this hypothesis.
Presented in part at the International Conference on Genetics and Epilepsy, Minneapolis, MN, July 20-22, 1990, and to the Child Neurology Society, Atlanta, GA, October 18-20, 1990. We are indebted to the members of the two families described herein for their generous cooperation, to Robin Leach, PhD, for many helpful suggestions. and to Sue Wiggins for preparation of the typescript.
References 1. Engel J. Seizures and epilepsy. Philadelphia: FA Davis, 1989: 112-134 2. Anderson VE, Hauser WA, Penry JK, Sing CF, eds. Genetic basis of the epilepsies. New York: Raven Press, 1982 3. Andermann E. Multifactional inheritance of generalized and focal epilepsy. In: Anderson VE, Hauser WA, Penry JK, Sing CF, eds. Genetic basis of the epilepsies. New York: Raven Press, 1982:103-112 4. Commission on Classification and Terminology of the International League against Epilepsy: Proposal for classification of epilepsies and epileptic syndromes. Epilepsia 1985;26:268-278 5. Quattlebaum TG. Benign familial convulsions in the neonatal period and early infancy. J Pediatr 1979;95:257-259 6. Bjerre I, Corelius E. Benign familial neonatal convulsions. Acta Pediatr Scand 1968;57:557-56 1 7. ZonanaJ, Sifvey K, Strimling H. Familial neonatal and infantile seizures: an autosomal-dominant disorder. Am J Med Genet 1984;18:455-459 8. Dobrescu 0, Labrisseau A. Benign familial neonatal convulsions. Can J Neurol Sci 1982;9:345-347 9. Shevell MI, Sinclair DB, Metrakos K. Benign familial neonatal seizures: clinical and electroencephalographic characteristics. Pediatr Neurol 1986;2:272-275 10. Leppert M, Anderson VE, Quattlebaum T, et al. Benign familial neonatal convulsions linked to genetic markers on chromosome 20. Nature 1989;337:647-648 11. Leppert M, Bjerre I, Quattlebaum TG, et al. Benign familial neonatal convulsions linked to markers on chromosome 20. Epilepsia 1989;30:687
12 Sawbrook J, Fritsch EF, Maniatis T Molecular cloning a laboratory manual 2nd ed Cold Spring Harbor, NY Cold Spring Harbor Press, 1989 13 Anderson MA, Gusella JF Use of cpclosporin A in establishing Epstein-Barr virus transformed human lymphoblastoid cell lines In Vitro 1984,20 856-858 14 Feinberg AP, Vogelstein P A technique for radiolabelling DNA restriction endonuclease fragments to high spechc activity Anal Biochem 1984,137 266-267 15 Nakamura Y , Martin C , Leppert M, et al Isolation and mapping of a polymorphic DNA sequence (pCMM6) on chromosome 20 (D20Sl7) Nucleic Acids Res 1988,16 5222 16 Myers R, Nakamura Y ,Leppen M, et al. Isolation and mapping of a polymorphic DNA sequence (pRMR6) on chromosome 20 (DZOSZO) Nucleic Acids Res 1988,16 9883 17 Nakamura Y, Leppert M, OConnell P A genetic linkage map of markers for human chromosome 20 Genomics 1989,j 945-947 18 Ott J Analysis of human genetic hnkage Baltunore The Johns Hopkins University Press, 1985 86 19 Rett A, Teubel R Neugeborenen-krampfe im rahmen einer epileptisch belasteten familie Wiener Klin Wochenschr 1964, 76 609-613
20. Morton NE. The detection and estimation of linkage between thc genes for elliptocytosis and the Rh blood type. Am J Hum Genet 1756;8:80-96 2 I . Bird TD, OR J, Giblett ER. Evidence for linkage of CharcotMarie-Tooth neuropathy to the Duffy locus on chromosome 1. Am J Hum Genet 1982;34:388-394 22. Vance JM, Nicholson GA, Yamaoka LH, et al. Linkage of Charcot-Marie-Tooth neuropathy type l a to chromosome 17. E x p Ncurol 1989;104:186-189 23. Haines JL, Amos J, Atwood J, et al. Linkage heterogeneity in tuberous sclerosis. Cytogenet Cell Genet 1987;5 1:1010 24. White R, Lalouel JM. Genetic markers in medicine. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic basis of inherited disease. New York: McGraw-Hill, 1989:277-288 25. Metrakos K, Metrakos JD. Genetics of convulsive disorders 11. Genetics and electroencephalogrdpIc studies in centrencephalic epilepsy. Neurology 1761;11:474-483 26. Bray PF, Wiser WC. Evidence for a genetic etiology of temporal-central abnormalities in focal epilepsy. N Engl J Med 1964;271:926-933 27. Greenberg D, Delgado-Escueta AV, Widelitz H, et al. Juvenile myoclonic epilepsy may be linked to the BF and HLA loci on human chromosome 6. Am J Med Genet 1788;31:185-192
Ryan et al: Familial Neonatal Convulsions
473
The gene for autosomal dominant “benign” familial neonatal convulsions, a transient, primary epilepsy of infancy, has recently been assigned to chromosome 20q. To determine whether this disorder is genetically heterogeneous, we performed linkage analysis in two previously unreported pedigrees with benign familial neonatal convulsions in which clinical heterogeneity was evident. There were 14 affected persons in the first family, and none had seizures (febrile or afebrile) after the age of 2 months. The second family had 13 affected individuals and 2 obligate carriers; seizures frequently did not remit until 6 to 24 months, febrile convulsions occurred in at least 2 patients, apparent audiogenic seizures occurred in 4 patients, and 1 individual had refractory epilepsy until late adolescence. Linkage studies with the chromosome 20 markers D20S19 and D20S20 were performed in both families. The resulting data favored linkage of the disease and marker loci in Family 2 by a maximum odds ratio of 45:l at 6% recombination. In Family 1, however, the odds were greater than 20,OOO:l against linkage at 10% recombination or less. We conclude that the syndrome of benign familial neonatal convulsions is clinically and genetically heterogeneous. Further study will be necessary to clarify the relationship between phenotype and genotype in this disorder. Ryan SG, Wiznitzer M, Hollman C, Torres MC, Szekeresova M, Schneider S. Benign familial neonatal convulsions: evidence for clinical and genetic heterogeneity. Ann Neurol 1991;29:469-473
Genetic influences are thought to contribute to many forms of human epilepsy [l, 2}. Indeed, among the so-called “secondary” epilepsies are numerous mendelian or single-gene disorders in which seizures are symptomatic of a fixed or progressive encephalopathy El}. Most primary epilepsies, however, are probably multifactorial in etiology [ 3 } , and the mechanisms whereby genetic factors cause seizures in these disorders are not understood. The syndrome of benign familial neonatal convulsions (BFNC) is a rare example of a primary epilepsy {4} in which inheritance is unequivocally autosomal dominant 15-91. BFNC is characterized by the occurrence of unprovoked partial or generalized c h i c seizures in the neonatal period or early infancy. Results of routine diagnostic studies (including cranial computed tomography, interictal electroencephalography (EEG), and serum chemistry determinations) and subsequent intellectual development are normal. Seizures usually remit by the age of 6 months, but chronic childhood or adult epilepsy has been reported in about 10% of patients [7, 91.
Recently, the gene for BFNC was assigned to chromosome 20q by the detection of tight linkage between the BFNC locus and two polymorphic D N A markers, D20S19 and D20S20, in a single large pedigree [lo}. Subsequent studies in two additional families supported this finding, and raised the odds ratio favoring linkage between the disease and marker loci to greater than 8,000,000:l Ell}. We recently identified two large and previously unreported BFNC pedigrees whose clinical features suggested that the disorder was heterogeneous. We therefore performed linkage analysis in these families to determine whether distinct genetic loci were involved.
From the Departments of *Pediatrics and tperiodontics, The University of Texas Health Science Center, San Antonio, TX, and the $Departments of Pediatrics and Neurology, Case Western Reserve University, Cleveland, OH.
Received Sep 11, 1990, and in revked form Nov 1. Accepted for publication Nov 4,1990.
Materials and Methods The families were identified through the probands whose case histories are provided below. The parents (where available) of all individuals “at risk’ for BFNC were interviewed. Medical records were available in only a few instances, and in most cases the age at onset and remission were not precisely recalled. Nonetheless, for most individuals, we were able to obtain a clear history regarding: ( 1 ) the presence or absence
Address correspondence to Dr Ryan, Department of Pediatrics, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, TX 78284.
Copyright 0 1991 by the American Neurological Association 469
of seizures in the neonatal period or in infancy, (2) whether seizures persisted past the ages of 2 or 12 months, and ( 3 ) whether or not at least one febrile seizure had occurred. Genomic D N A was prepared by phenolichloroform extraction of lysed, Epstein-Barr virus-transformed peripheral blood lymphoblastoid cell lines El2, 131. The samples were digested with the restriction enzyme TaqI (New England Biolabs, Beverly, MA) and the resulting fragments were fractionated by 0.7% agarose gel electrophoresis and transferred by capillary action to nylon membranes (Nytran, Schleicher and Schuell, Keene, NH) according to protocols supplied by the manufacturers. Following prehybridization, the blots were hybridized with the radiolabelled [ 141 markers D20S 19 (pCMM6) [15] and D20S20 (pRMR6) 1161(purchased from the American Type Culture Collection, Rockville, MD) and washed in solutions of increasing stringency as suggested by the manufacturer. These two probes reveal D N A restriction fragment polymorphisms due to a variable number of tandem repeats of a short D N A sequence (D20S19) and the presence or absence of a TayI restriction site (D20S20). Because polymorphisms were poorly demonstrated with the latter probe, we tested fragments obtained from simultaneous digestion of the pRMR6 plasmid with Hind111 and EcoRI. A 0.7-kb fragment revealed the polymorphic system clearly and was used for typing. Autoradiograms of the membranes were then prepared [12), and marker genotypes assigned to each individual. D20S19 and D20S20 are closely linked, with an estimated recombination fraction of 0.016 [171. Since no recombination between the two markers was observed in our families, it was possible to assign a combined marker haplotype to each individual. Two-point linkage analysis was performed with the program LIPED [IS] on an IBM-XT compatible personal computer. A penetrance of 90% was assumed for the BFNC allele, based on a segregation ratio of 45% reported previously [lo].
Family 1 The proband was a healthy, term neonate of MexicanAmerican ancestry, whose vaginal delivery followed an uncomplicated gestation. O n the ninth postnatal day, she had four clonic convulsions, each about 2 minutes long: Two were generalized, one was left-sided, and one was right-sided. Physical examination results, blood glucose levels, routine serum electrolyte levels (including calcium and magnesium), and an EEG were all normal. Neuroimaging was deferred because at least 3 affected relatives had had normal findings on cranial computed tomography. Phenobarbital was begun at a total daily dose of 5 mgikg, and was discontinued by the mother a month later. At the time of writing, the patient was 9 months old and developmentally normal, and had had no recurrence of seizures. Thirteen of the proband's relatives had neonatal seizures beginning between the second and fourteenth postnatal day (Fig 1, Table 1). No one in Generations I and 11 received anticonvulsant therapy, but all of the proband's affected cousins (Generation 111) received phenobarbital for up to 4 months. None had seizures while taking phenobarbital, and no seizures occurred in any family member after the age of 2 months, regardless of treatment status. Intellectual development appeared normal in all family members. 470 Annals of Neurology
V a l 29 N o 5
May 1991
I
Fig 1. Pedigree and genotypic data for Family 1 (arrow indicates proband). The D20S19 alleles are denoted by letters and the D20S20 alleles by numbers. Numerous instances of recombination between the disease and marher alleles are &dent.
Table 1. Clinical Features Suggestive of Heterogeneity in Two Pedigrees with Benign Familial Neonatal Convulsions
Persistence of seizures beyond 12 months Epilepsy in late childhoodiadolescence Audiogenic seizures Febrile seizures Asymptomatic obligate carriers
Family 1 (n = 14)
Family 2 (n = 15)
0
4
0
1
0 0 0
4 2 2
Family 2 The proband for Family 2 was born to parents of northern European ancestry via term, spontaneous vaginal delivery following an uncomplicated gestation. O n the second postnatal day, she had 11 partial clonic seizures. A loading dose of phenytoin was administered, and maintenance phenobarbital was begun at a dose of 5 mgikgiday. Physical examination findings, cranial computed tomograms, serum chemistry measurements, and EEGs were all normal. At least two additional seizures occurred during the next 3 weeks despite phenobarbital treatment (serum levels were not measured), but subsequently she was seizure-free. At 10 months old she was developmentally normal, and continued to take phenobarbital. Twelve relatives of the proband had a clear history of idiopathic neonatal convulsions (Fig 2;see Table 1).There were also two obligate carriers: Patients 11-9 and either 1-1 or 1-2. One infant (IV-3) had severe birth asphyxia due to placental abruption. Neither his neonatal seizures nor his subsequent microcephaly and developmental delay were thought to be manifestations of BFNC, but he was omitted from the linkage analysis because affectation status for BFNC was deemed uncertain. One relative (111-3) continued to have frequent convulsions until late adolescence despite phenobarbital therapy; his management was complicated by noncompliance and at least three episodes of convulsive status epilepticus. Although of borderline intelligence, he was fully employed at the time of writing. All other family members (except IV-3) appeared to be intellectually normal.
Table 2. Results oJ Linkage Analysis h d Scores
Recombination Fraction (Sex-Averaged)
Fig 2.Pedigree and genotypic data for Family 2 (arrow indicates proband). Marker alleles are denoted as in Figure I . Recombination between the disease and muker loci appears to have occurred in individuals 111-1 and III-16, but nonpenetrance could a h explain their marher genotypes. Patient IV-3 was omitted from the analysis because of uncertainty regarding his phenotype (see text).
Four affected patients (11-14, 111-3, 111-15, and IV-6) had a history strongly suggestive of audiogenic seizures. For example, in one 2’/2-month-old infant (IV-6) a prolonged (longer than 10 minutes) clonic seizure was apparently precipitated by the explosion of a firecracker. Seizures induced by such stimuli occurred in both sleeping and walung states. At least 4 affected relatives (11-14, 111-3, 111-6, and 111-8) continued to have afebrile seizures until the age of 12 months or beyond. At least 2 affected relatives (111-3 and IV-5) also had o n e or more febrile convulsions between the ages of 6 and 18 months. EEG results were available only for IV-5, IV-6, and IV-7; all three tracings were normal.
Results The marker genotypes are provided in Figs 1 and 2, and lod scores* from two-point linkage analysis are presented in Table 2. Both markers were informative in Family 1, but only D20S19 was informative in Pamily 2, as no affected parents were heterozygous for D20S20. The results were not influenced by the population frequencies of the marker alleles except for slight variation, in Family 2, as a function of the frequency of D20S19 allele B (data not shown). A value of 0.33 for this parameter, based on a survey of 27 unrelated individuals, was employed for the lod score computations in Table 2. For Family 1, the data strongly favored nonlinkage over linkage of the disease and marker loci; the odds against linkage at 10% recombination, for example, are greater than 21,OOO:l (lod score, -4.34). For Family 2, a maximal odds ratio of 45: 1 (lod score, 1.66) favoring linkage occurred at 6% recombina-
+
‘A Iod (“logarithm of odds”) score is a measure of the likelihood that a set of genotypic observations results froni linkage as opposed to chance. A lod score of 3.0, corresponding to a 1,000:1 odds ratio, is generally accepted as establishing linkage.
Family 1
D20S19
0
-32
0.01 0.05 0.10 0.20 0.30 0.40
-6.91 -3.86 -2.47 -1.10 -0.43 -0.09
D20S20 -Ic
-6.08 -3.26 -2.05 -0.93 -0.38 -0.10
Two-Marker Haplotype -02
-11.30 -6.61 -4.34 -2.12 -0.97 -0.32
Family 2 D20S19
1.58 1.61 1.66 1.63 1.40 1.00 0.48
tion, with a one-lod confidence interval of 1 to 37%. When sex-specific recombination fractions were allowed to vary independently, a maximum lod score of 1.67 occurred at male and female rates of 10 and 396, respectively. Numerous instances of recombination between the disease and marker loci are evident in Family 1. In Family 2, only 2 individuals (aside from the obligate carriers) failed to display the BFNC phenotype despite inheritance of the linked haplotype. The carrier status of these individuals (111-1 and 111-16) is uncertain, since nonpenetrance cannot be distinguished from recombination between the disease and marker loci. No a€fected individuals lacked the linked haplotype.
Discussion Numerous reports of BFNC have appeared since the original description of Rett and TeubelT5-9, 191. The clinical picture is distinctive and relatively uniform, although it has been suggested that differences among pedigrees with regard to risk for epilepsy in later life might reflect genetic heterogeneity. Shevell and associates {9] reviewed the clinical features of 113 patients with BFNC in 15 families; 11 of 62 individuals in 6 families experienced nonfebrile convulsions beyond infancy, while the 5 1 individuals in the remaining 9 families experienced early and complete remission of seizures. They observed that the data were compatible with “two forms [of BFNC] with sharply different risks for subsequent epilepsy.” Similarly, Leppert and associates {lo] noted that the development of nonfebrile seizures after the age of 6 months may be restricted to certain pedigrees. Unfortunately, most published reports of BFNC do not provide detailed information on age at remission of seizures, presumably because of difficulty in obtaining reliable data. We encountered uncertainty with regard to age at remission as wel, but were nonetheless able to detect a trend toward later remission in Family 2 than in Family 1.
Ryan et al: Familial Neonatal Convulsions 471
Our linkage data strongly support the existence of two distinct genetic loci for BFNC. The maximum lod
score of 1.66 obtained for Family 2 suggests, but does not prove, that the BFNC locus in this kindred is linked to the marker loci. Given previous reports of linkage of BFNC to these markers [lo, 1I), however, it appears quite likely that the disease gene in Family 2 maps to chromosome 20q. In contrast, in Family 1, the data militate strongly against proximity of the disease and marker loci. Given the magnitude of the negative lod scores in Family 1, it is unlikely that erroneous phenotype assignment, nonpaternity, or an inaccurate penetrance estimate could account for these results. The data further suggest that the BFNC subtype linked to chromosome 20q may be associated with later remission and a higher risk for subsequent epilepsy (defined as afebrile seizures after age 2 years) than in the variant of this disorder present in our first family. Based on the original reports 15-71, late epilepsy occurred in at least 3 of 15, 2 of 14, and none of 6 individuals in the three families studied by Leppert and associates [lo]. Additional studies are needed to determine whether there is a consistent relationship between the genetic subtype of BFNC and clinical features such as ethnicity, age at remission, response to anticonvulsant therapy, audiogenic seizures, and the risk of late epilepsy. There are several precedents for the discovery of heterogeneity by linkage analysis of disorders presumed to be uniform. The earliest example is probably that of hereditary elliptocytosis, for which linkage to the Rh blood group locus has been established in some kindreds, but disproved in others r20). Recent examples in clinical neurology include hereditary motor and sensory neuropathy type I (in which loci on chromosome 1 and 17 have been implicated in different families) 121, 221 and tuberous sclerosis (in which linkage to the ABO blood group locus on chromosome 9 has been demonstrated in some families but disproved in others) [23}. The presence of heterogeneity in BFNC may complicate efforts to isolate the 20q gene, since a critical step in such a project will be a search for flanking markers {24]. This entails a search for individuals who demonstrate recombination between the disease and a closely linked marker locus. Given heterogeneity, only recombinants in pedigrees that unequivocally demonstrate linkage-hence, very large pedigrees-can be considered. In other words, the potential information present in small pedigrees is lost because the genetic type cannot be reliably determined. The presence of heterogeneity in BFNC, an epilepsy with highly distinctive clinical features and single-gene inheritance, suggests that disorders such as childhood absence and benign partial childhood (rolandicj epilepsy [4], for both of which dominant inheritance has 472
Annals of Neurology Vol 29
No 5
May 1991
been suggested [25, 261, could also exhibit this phenomenon. If this is the case, then linkage analysis in these disorders, which will almost certainly involve numerous small families, will be particularly challenging. Although such an approach may be successful, as in juvenile myoclonic epilepsy (recently assigned to chromosome Gp) 1271, heterogeneity could easily complicate efforts to assign genes for epilepsies that rarely, if ever, segregate in extensive pedigrees. Without firm clinical criteria for classifying families, evidence supporting linkage in some families could be negated by evidence against it in others. We have been able to demonstrate heterogeneity in BFNC because large families were available for study and because two markers (one moderately and the other highly polymorphic) mapping to the region of interest were also available. It is likely that phenotypic variation among BFNC families is directly related to genotypic differences, and may be more common than previously appreciated. Further study should confirm or refute this hypothesis.
Presented in part at the International Conference on Genetics and Epilepsy, Minneapolis, MN, July 20-22, 1990, and to the Child Neurology Society, Atlanta, GA, October 18-20, 1990. We are indebted to the members of the two families described herein for their generous cooperation, to Robin Leach, PhD, for many helpful suggestions. and to Sue Wiggins for preparation of the typescript.
References 1. Engel J. Seizures and epilepsy. Philadelphia: FA Davis, 1989: 112-134 2. Anderson VE, Hauser WA, Penry JK, Sing CF, eds. Genetic basis of the epilepsies. New York: Raven Press, 1982 3. Andermann E. Multifactional inheritance of generalized and focal epilepsy. In: Anderson VE, Hauser WA, Penry JK, Sing CF, eds. Genetic basis of the epilepsies. New York: Raven Press, 1982:103-112 4. Commission on Classification and Terminology of the International League against Epilepsy: Proposal for classification of epilepsies and epileptic syndromes. Epilepsia 1985;26:268-278 5. Quattlebaum TG. Benign familial convulsions in the neonatal period and early infancy. J Pediatr 1979;95:257-259 6. Bjerre I, Corelius E. Benign familial neonatal convulsions. Acta Pediatr Scand 1968;57:557-56 1 7. ZonanaJ, Sifvey K, Strimling H. Familial neonatal and infantile seizures: an autosomal-dominant disorder. Am J Med Genet 1984;18:455-459 8. Dobrescu 0, Labrisseau A. Benign familial neonatal convulsions. Can J Neurol Sci 1982;9:345-347 9. Shevell MI, Sinclair DB, Metrakos K. Benign familial neonatal seizures: clinical and electroencephalographic characteristics. Pediatr Neurol 1986;2:272-275 10. Leppert M, Anderson VE, Quattlebaum T, et al. Benign familial neonatal convulsions linked to genetic markers on chromosome 20. Nature 1989;337:647-648 11. Leppert M, Bjerre I, Quattlebaum TG, et al. Benign familial neonatal convulsions linked to markers on chromosome 20. Epilepsia 1989;30:687
12 Sawbrook J, Fritsch EF, Maniatis T Molecular cloning a laboratory manual 2nd ed Cold Spring Harbor, NY Cold Spring Harbor Press, 1989 13 Anderson MA, Gusella JF Use of cpclosporin A in establishing Epstein-Barr virus transformed human lymphoblastoid cell lines In Vitro 1984,20 856-858 14 Feinberg AP, Vogelstein P A technique for radiolabelling DNA restriction endonuclease fragments to high spechc activity Anal Biochem 1984,137 266-267 15 Nakamura Y , Martin C , Leppert M, et al Isolation and mapping of a polymorphic DNA sequence (pCMM6) on chromosome 20 (D20Sl7) Nucleic Acids Res 1988,16 5222 16 Myers R, Nakamura Y ,Leppen M, et al. Isolation and mapping of a polymorphic DNA sequence (pRMR6) on chromosome 20 (DZOSZO) Nucleic Acids Res 1988,16 9883 17 Nakamura Y, Leppert M, OConnell P A genetic linkage map of markers for human chromosome 20 Genomics 1989,j 945-947 18 Ott J Analysis of human genetic hnkage Baltunore The Johns Hopkins University Press, 1985 86 19 Rett A, Teubel R Neugeborenen-krampfe im rahmen einer epileptisch belasteten familie Wiener Klin Wochenschr 1964, 76 609-613
20. Morton NE. The detection and estimation of linkage between thc genes for elliptocytosis and the Rh blood type. Am J Hum Genet 1756;8:80-96 2 I . Bird TD, OR J, Giblett ER. Evidence for linkage of CharcotMarie-Tooth neuropathy to the Duffy locus on chromosome 1. Am J Hum Genet 1982;34:388-394 22. Vance JM, Nicholson GA, Yamaoka LH, et al. Linkage of Charcot-Marie-Tooth neuropathy type l a to chromosome 17. E x p Ncurol 1989;104:186-189 23. Haines JL, Amos J, Atwood J, et al. Linkage heterogeneity in tuberous sclerosis. Cytogenet Cell Genet 1987;5 1:1010 24. White R, Lalouel JM. Genetic markers in medicine. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic basis of inherited disease. New York: McGraw-Hill, 1989:277-288 25. Metrakos K, Metrakos JD. Genetics of convulsive disorders 11. Genetics and electroencephalogrdpIc studies in centrencephalic epilepsy. Neurology 1761;11:474-483 26. Bray PF, Wiser WC. Evidence for a genetic etiology of temporal-central abnormalities in focal epilepsy. N Engl J Med 1964;271:926-933 27. Greenberg D, Delgado-Escueta AV, Widelitz H, et al. Juvenile myoclonic epilepsy may be linked to the BF and HLA loci on human chromosome 6. Am J Med Genet 1788;31:185-192
Ryan et al: Familial Neonatal Convulsions
473
