Oct 7, 1994 - The marker combinations and specific features of the closest microsatellites are described ..... in 10 per cent of our SMA families, i.e., apparently.
PRENATAL DIAGNOSIS, VOL.
15 407417 (1995)
PRENATAL PREDICTION IN FAMILIES WITH AUTOSOMAL RECESSIVE PROXIMAL SPINAL MUSCULAR ATROPHY (5ql1.2-q13.3): MOLECULAR GENETICS AND CLINICAL EXPERIENCE IN 109 CASES B. WJRTH, S. RUDNIK-SCHONEBORN, E. HAHNEN, D. ROHRIG AND K. ZERRES
Institute of Human Genetics, Wilhelmstr. 31, 53111 Bonn, Germany Received 7 October 1994 Revised I December 1994 Accepted 6 January 1995
SUMMARY Prenatal prediction in families at risk for autosomal recessive proximal spinal muscular atrophy (SMA) mainly of type I is often requested due to the high incidence and the fatal outcome of the disease. So far, only indirect genotype analysis can be performed in SMA families, since the gene has not yet been identified. We present our experience of 109 prenatal diagnoses obtained in 91 families by use of single- and multi-locus polymorphic microsatellites of the region 5q11.2-ql3.3. The marker combinations and specific features of the closest microsatellites are described in detail. From 137 requests for prenatal prediction of SMA between October 1991 and August 1994,28 families were excluded, mostly because the clinical diagnosis was uncertain or doubtful. Others had to be classified as ‘SMA-variants’ or showed autosomal dominant transmission of SMA. Of the 109 prenatal diagnoses performed, 29 fetuses were diagnosed to be at high risk (>99 per cent) of developing the disease, while in seven additional pregnancies no exact prediction could be made due to a recombination event in one parental haplotype. Altogether, recombinations between closely flanking markers were observed in 14 cases. In 35 cases, the parents decided to terminate the pregnancy. Of the remaining pregnancies, 32 could be followed beyond term. All infants were reported to develop normally without signs of SMA. Two children were born with transverse reduction defects of one hand, which was most likely related to early chorionic villus sampling at 9 and 10 weeks’ gestation. No further abnormalities could be detected. The limits of indirect genotype analysis and the problems of diagnostic accuracy and heterogeneity of proximal SMA are discussed. KEY WORDS: spinal muscular atrophy; SMA; prenatal diagnosis in SMA, risk of misdiagnosis in SMA.
INTRODUCTION Proximal spinal muscular atrophy (SMA) is one of the most frequent autosomal recessive diseases after cystic fibrosis (incidence greater than 1 in 10000) and the second most frequent neuromuscular disorder after Ducheme muscular dystrophy (DMD). The clinical picture of proximal SMA varies to a great extent regarding the age of Addressee for correspondence:Dr Brunhilde Wirth, Institute of Human Genetics, Bonn, Wilhelmstr. 31, 53111 Bonn, Gel-IIlaIly.
CCC 0197-3851/95/050407-11 0 1995 by John Wiley L Sons, Ltd.
onset and severity (Munsat and Davies, 1992; Zerres and Rudnik-Schoneborn, 1995). Prenatal prediction in families at risk for SMA is often requested due to the fatal outcome of the most frequent severe form (Werdnig-Hoffmann disease). After the gene of recessive infantile SMA was mapped to the long arm of chromosome 5 (Brzustowicz et al., 1990; Gilliam et al., 1990; Melki et al., 1990a,b), prenatal diagnosis by indirect genotype analysis could be offered to most families (Melki et al., 1992; Daniels et aZ., 1992). However, the gene remains to be identified, so a direct genotype analysis is not yet possible.
408
B. WIRTH ET AL.
Genetic linkage studies localized the SMA gene to the region 5q11.2-ql3.3 by a great number of flanking polymorphic microsatellites (Brzustowicz et al., 1990, 1992; Melki et al., 1990a,b, 1993, 1994; Gilliam et al., 1990; Lien et al., 1991; Daniels et al., 1992; Momson et al., 1992, 1993a; Soares et al., 1993; Francis et al., 1993; Wirth et al., 1994, 1995; Burghes et al., 1994; Clermont et al., 1994; DiDonato et al., 1995). The closest flanking markers to the SMA locus are D5S823/A31 (Wirth et al., 1995a) and D5S557/2AE9.1 (Francis et al., 1993; Wirth et al., 1994), which encompass approximately 850 kb. More recently, a large number of multicopy polymorphic microsatellites which recognize 1-4 alleles or sublocil chromosomes were identified (Kleyn et al., 1993; Brahe et al., 1994; Burghes et al., 1994; Melki et al., 1994; DiDonato et al., 1995; Wirth et al., 1995a). The various subloci are clustered within the SMA candidate region. In this paper we report our experience of indirect genotype analysis in infantile SMA using the closest polymorphic microsatellites, the results of 109 prenatal diagnoses in 91 SMA families, the attitudes of the families towards prenatal prediction, and the diagnostic difficulties encountered in SMA.
MATERIALS AND METHODS F m i l y assessment In more than 300 SMA families, a genotype analysis was undertaken for a later prenatal diagnosis. Between October 1991 and August 1994, there were 137 requests for prenatal diagnosis concerning SMA. Ninety-one women underwent a prenatal prediction in 109 pregnancies, whereas in 28 additional pregnancies prenatal diagnosis was withdrawn. Apart from diagnostic doubts, other reasons for exclusion were non-paternity, evidence of autosomal dominant inheritance, and ‘non-5q linkage’ (see Discussion). In each index patient, the clinical diagnosis of proximal SMA was made according to the diagnostic criteria defined by the International SMA Consortium (Munsat and Davies, 1992). The families were classified into SMA types 1-111 corresponding to the achieved motor milestones of the types proposed by the International SMA Consortium. Of the 91 index patients, 70 belonged to type I, 20 to type 11, and one to type 111.
DNA isolation
DNA was usually isolated from venous EDTA blood (Miller et al., 1988), in few cases from frozen muscle tissue (Sambrook et al., 1989), in two cases from stained muscle slides (Eggennann et al., 1994), and once from dried blood spots (Rubin et al., 1989). In two of four further cases, we succeeded in isolating DNA from paraffinembedded tissue (Eggermann et al., 1994). In most cases the prenatal diagnosis was carried out on DNA isolated from chorionic villus biopsies (CVS) performed at 9-14 weeks’ gestation, twice from amniotic fluid cultures, and once from a placenta biopsy at 19 weeks’ gestation. Probes and conditions of polymerase chain reaction (PCR)
The probes used for indirect genotype analysis were all polymorphic microsatellites: (a) single locus microsatellites (listed from centromere to telomere): Lambda 559iDSS76 (Sherrington et al., 1991a), WlSCATTD5S507 (Wirth et al., 1994), EF(TG/AG)/DSSl25 (Sherrington et al., 1991b), AMF114ye7/DSS464 (Melki et al., 1993), EF13/14/DSS680 (Morrison et al., 1993a); LAS96D5S681 (Morrison et al., 1993b), JK348/ D5S435 (Soares et al., 1993; Wirth et al., 1994), AFM265wf5D5S629 (Clermont et al., 1994), A31/D5S823 (Wirth et al., 1995a), 2AE9.1/ D5S557 (Francis et al., 1993), 38.3 (Thompson et al., 1993), AFM281yh9DSS637 (Clermont et al., 1994), MIT-I105/DSS351 (Hudson et al., 1992), RB110/11WMAPl B (Brzustowicz et al., 1992), RB104/106/3’MAPIB (Lien et ~ l . ,1991), JK53CND5S112 (Daniels et al., 1992), YNCT/ D5S127 (Shemngton et al., 1991b), 6741GTI D5S39 (Mankoo et al., 1990); (b) multicopy polymorphic microsatellites: Ag 1-CA/D5S 1556 (DiDonato et al., 1994), C212/D5F149 (Melki et al., 1994) (see also Table I). PCR ampl$cation of single-locus microsatellites
Each genomic DNA sample (30 ng) was amplified in a 15pI reaction mixture consisting of 4 pmol of primers; 1 2 0 each ~ ~of dATP, dTTP, and dGTP; 1 2 p dCTP; ~ 0.124 of 32P-dCTP at 10 mCi/ml ( 3000 Ci/mmol; Hartmann); IX Cetus PCR Buffer; and 0.4U of Taq polymerase (Gibco BRL or Amersham). Samples were
-
409
PROXIMAL SPINAL. MUSCULAR ATROPHY
Table I-Characterization of polymorphic microsatellites closely flanking the SMA gene ~-~~ ~
Marker ken Lambda 559 W 15CATT EF(TG/AG) AMFl14ye7 EF13/14 LAS96 JK348 AFM265wf5 A3 1 Agl-CA c212 2AE9.1 38.3 AFM281yh9 MIT-I105 RB 110/111 RB 104/106 JK53CA YNCT 6741GT 5qtel
~
Annealing temperature ("C)
Size of PCR product (bp)
Genetic distance to SMA (cM)
Heterozygosity frequency
D5S76 D5S507 D5S125 D5S464 D5S680 D5S681 D5S435 D5S629 D5S823 D5S1556 D5F149 D5S557
55 62 55 55
94-1 10 95-103 143-1 47 199 141-173 142-1 56 149-167 -116 128-150 90-122 166-203 148-1 72
5 5 1
0.77 0.27
1
D5S637 D5S351 5'MAP 1B 3'MAPlB D5S112 D5S127 D5S39
55
0.45 0.57 068 0.69 0-86* 0.54 0.98 0.99 0.54 ND 0.77* 0.72 0.76 0.72 0.75 0-88 0.83
Locus
-
55 55 55 55 60 62 60 55 55 60-61 60-62 55
55 62 55
-
-ND
250 196-234 100 * 200 100-120 96-1 14 212-220
-
2 2 1 2 1 0 0 2 ND 2t 0 1 1 3 4 5
0-50
*Heterozygosity frequency given in the original reference; all the others have been calculated from 200 independent persons, mainly of German origin. ?Genetic distances given in the original reference; all the others have been calculated from a linkage study in over 100 SMA families (Wirth et al., 1994 and this paper). ND=Not determined.
overlaid with one drop of mineral oil. Amplification conditions were as follows: one cycle of denaturation at 93°C for 7min; 27 cycles including 1 min denaturation at 94"C, 2 min annealing at the appropriate temperature (between 55 and 62"C), and 25min extension at 72°C; and one cycle of final extension of 1Omin at 72°C in a Perkin Elmer Cetus Cycler 480. Two p1 of the reaction mixture was mixed with 4pl of formamide loading buffer, boiled for 2 min, and run on an 8 per cent denaturing polyacrylamide gel at 45-50 mA, 1900V, 50°C for 2-4 h in a sequencing apparatus Model S2 (Gibco BRL). Gels were dried on 3MM Watman paper and exposed to Kodak X-OMAT AR film at room temperature for several hours or overnight. To minimize the number of PCRs (important when small amounts of material from the fetus or affected child were obtained), we com-
bined at least two pairs of primers in one assay (Table 11). PCR amplification of multicopy polymorphic microsatellites Five pmol of forward primer for each PCR was end-labelled in 1 x T,-kinase buffer, 0.1 U of T,-kinase (MBI Fermentas or USB/Amersham), and 0.18 pl of gamma '2P-dATP (5000 CUmmol; Hartmann) for 45min at 37°C. The reaction mixture was heat-inactivated at 95°C for 1Omin. After that, each genomic DNA sample (30ng) was amplified in a 15p1 reaction mixture consisting of 5 pmol of end-labelled primer and 5 pmol of reverse primer, 120pM each dNTP, 1 X Cetus PCR buffer, and 0.4U of Taq polymerase. The PCR amplification and running conditions were as described above.
410
B. WIRTH ET
AL
Table 11-Combination of primers which can be used as duplex polymerase chain reactions Markers
RUIl Annealing temperature ('C) time (h)
A31 and MIT-I105 MIT-I105 and RB110/111 JK348 and AFM281yh9 EF(TG/AG) and Lambda 599 EF13114 and AFMl14ye7 EF(TG/AG) and JK53CA RB110/111 and YNCT Agl-CA and C212
60 60
55 55 55
55 62 60
RESULTS AND DISCUSSION
Molecular genetics of infantile SMA Since 1991 we have performed linkage studies in more than 400 SMA families using all available polymorphic microsatellites (Wirth et al., 1993, 1994, 1995a and b). Initially, we performed the prenatal diagnosis only with single-locus microsatellites, while more recently we have also used multilocus polymorphic microsatellites: these were always used in combination with the single-locus microsatellites. From all the multicopy markers described up to now, only Agl-CA (DiDonato et al., 1994) and C212 (Melki et al., 1994) were used for prenatal diagnosis because they show a smaller number of alleles, especially in SMA type I, and are therefore easier to evaluate. In general, we used the primers as described in the original references (see the Materials and Methods section and Table I). For marker JK348/ D5S435 (Soares et al., 1993), we recommend the use of the described combination of primers which give more specific and stronger signals Wirth et al., 1994). As we had problems amplifying the AFM265wf5D5S629 primers (Clermont et al., 1994), we developed new primers (S'CCTGGGCAACAAGAGCAA3' and STCTTTCAGTCAGCCAGATATCC3') which give a smaller PCR product (1 16 bp instead of 250 bp) and better signals. In order to obtain quickly information about the genotypic status of the family members on the basis of a few PCRs only, we usually performed duplex PCRs (Table 11). We found that among the single-locus microsatellites, the markers with the highest heterozygosity frequency in SMA families (mainly
3 3 4 2 4 2 2 3
Location relative the SMA gene
to
Proximal and distal Both distal Proximal and distal Both proximal Both p r o d Proximal and distal Both distal Unknown (no recombinations to SMA)
German families) are RB110/111 and MIT-I105 on the distal side and JK348 and AFM265wf5 on the proximal side. The closest flanking markers A3 1 and 2AE9.1 are less informative (0.54), so in many cases a genotype analysis using the next informative markers was required. The marker AFMl14ye7 revealed a 'null' allele in 10 per cent of our SMA families, i.e., apparently the child did not receive any alleles from one of the parents who always showed only one allele (very likely hemizygous). In two cases, the children showed new alleles suggesting the occurrence of a new mutation, since the haplotypes with all the other markers were compatible with those of the parents. A similar observation which also suggested new mutations was made in two cases for 2AE9.1. We found evidence of linkage disequilibrium between the SMA locus and alleles 1 and 6 of JK348 (P-10 (A31) and 1 (JK348)-6 (A31) have mainly or exclusively been found on SMA chromosomes in our material. However, these data require further analysis; therefore they are not yet applicable for prenatal prediction, screening tests, or carrier detection in families in which the affected child is not available for investigation. With regard to the new multicopy markers, the interpretation of the results might be diEcult sometimes (question of simple or double dosage of alleles, evaluation of allele combinations transmitted from each parent). Furthermore, loss of alleles and deletions were detected by both markers C212
PROXIMAL SPINAL MUSCULAR ATROPHY
41 1
and Agl-CA (Melki et al., 1994; DiDonato et al., 1994; Wirth et al., 1995b). The deletions were found to be statistically associated with SMA type I and can occur de now, which has important implications for the reliability of prenatal prediction. Consequently, these multicopy microsatellites are only of limited value for prenatal diagnosis and should only be used in combination with singlelocus microsatellites which are clearly flanking the SMA gene. The position of the SMA gene relative to the multicopy markers is not yet known. However, a strong association between the SMA locus and certain alleles of Agl-CA has been reported in the French-Canadian and American population (DiDonato et al., 1994) and was also found in our own material with both markers Agl-CA and C212 (Wirth et al., 199513). Therefore, in cases where a recombination of a closely flanking marker can be observed without recombination of Agl-CA, the risk of the fetus having inherited the mutated gene is probably small but cannot be determined more precisely. The genetic distances for each of the markers versus the SMA gene are given in Table I. Since the closest flanking markers showed a maximum recombination frequency to the SMA gene of 5 per cent, the risk of a false prediction ranged between 5 5 per cent for markers informative on only one side and 0.025 per cent for informative markers on both sides. However, in approximately 90 per cent of our families we were able to find informative markers in a genetic distance of 1-2cM. When multicopy markers were used, all families became informative; the recombination frequency should be almost zero, since no recombinants have been found so far.
have to be expected generally within the first few months after birth and intrafamilial variability is small, especially in families with severe infantile SMA (Rudnik-Schoneborn et al., 1994a), we considered the 25 infants in SMA I families to be at extremely low risk of developing SMA at a later age. There were seven sibs of SMA I1 patients whose clinical information was obtained at 6-13 months of age. Since they have shown a normal development up to now, it is unlikely that they will reveal symptoms of SMA in the next few months (Rudnik-Schoneborn et al., 1994a). Two children were born with transverse limb reduction defects of one hand; this is most likely related to the early CVS procedure at 9 and 10 weeks’ gestation. CVS has been discussed as an aetiological factor for limb reduction defects (Firth et al., 1994): the risk is believed to be 3-6 times higher after CVS at 9-10 weeks’ gestation than in the normal population (Olney et al., 1994). The severity of transverse deficiencies was found to decrease with increasing gestation at the time of the CVS procedure (Firth et al., 1994); therefore early CVS should be avoided. Despite closely flanking single-locus microsatellites, we observed 14 recombinations out of 109 cases (12.8 per cent) occurring in one parental haplotype between flanking markers. In seven cases, it was not possible to decide whether the genotypic status of the fetus was SMNSMA or SMNnormal and so six of these pregnancies were terminated; however, in the remaining seven pregnancies the status was SMNnormal or normal/ normal. Altogether, 33 per cent of pregnancies were at high risk (36 out of 109), which is higher than the expected 25 per cent as partly related to the recombination events in some families.
Results of prenatal diagnosis
Attitudes towardr prenatal prediction
In 109 pregnancies a prenatal prediction was performed; the results are summarized in Table 111. There were 36 pregnancies where a high risk of an affected fetus was predicted; these were terminated in all but one case. Seventy-three pregnancies, in which the fetus was diagnosed to be unaffected, were followed beyond term if possible. Follow-up information was obtained in 32 cases by means of a questionnaire regarding the outcome of the pregnancy. The infants born were between 3 and 12 months of age when the clinical status was documented and none of them presented signs of SMA. Since the first signs of muscle weakness in SMA I
SMA is one of the diseases with a strong demand for prenatal diagnosis. In a survey on the attitudes towards prenatal prediction of SMA in 51 patients older than 15 years and 78 parents, 68.2 per cent of the patients and 90.3 per cent of the parents stated that termination of pregnancy in SMA is acceptable. When asked personally, only about 50 per cent of both patients and parents would demand prenatal diagnosis if it were available (Zerres et al., 1993). A higher percentage of parents whose children were already deceased expressed a positive attitude towards prenatal diagnosis (60 per cent) as compared with those
Sitting unaided not achieved Sitting possible, no standing or walking
Unaided walking or standing possible
I I1
I11 109
1
1 91
84 24
70 20 22
0
17 5 42
0
36 6
1 29
9
20 8 0
5 4
7
0
6 1
32
0
25 7
41
0
33 8
*No signs of SMA.
36
1
26 9
No. of Unaffected Affectedhigh risk No. of prenatal Healthy No answer families diagnoses W/W MIW WIRec. MIM MIRec. infants* possible Terminations
Outcome
W =Wild type; M=mutant haplotype segregating with SMA; Rec. =haplotype presenting a recombination between informative flanking markers.
Total
Definition
SMA trpe
Results of genotype analysis
Table 111-Results of genotype analysis and outcome of 109 prenatal diagnoses in families with infantile SMA
F
4 9
5l
m s
h,
e
PROXIMAL SPINAL MUSCULkR ATROPHY
parents who had children with more chronic courses who were still alive (40 per cent). Among the participating patients, no significant differences were found between SMA types I1 and 111. There was only one family who decided not to undergo prenatal diagnosis; this was because a recombination in the maternal haplotype between the flanking markers occurred in the affected child which made an exact prediction in a subsequent pregnancy impossible. In another case, the family decided to continue the pregnancy, although the fetus had inherited the maternal chromosome with the mutation, while in the paternal chromosome a recombination between JK348 and AE9.1 had occurred. Agl-CA showed the alleles of the non-mutated paternal chromosome. Follow-up information is not yet available. The applicability of prenatal diagnosis in SMA leads in general to termination of pregnancies in which the fetus is at high risk of being affected. In the 35 families where the first pregnancy was terminated, a large number decided to have another child as soon as possible: 14 women became pregnant again in the following 5-24 months, the average being 8 months. Fortunately, in all but one case the second prenatal genotype analysis resulted in the diagnosis of an unaffected child. The interval was less than 5 months after termination of pregnancy in eight of the remaining 21 women, so that these families cannot be considered for a second CVS at present. Therefore, only 13 of 35 women have not become pregnant again considering an interval of more than 5 months after termination of pregnancy, whereas the proportion of women who underwent a second prenatal diagnosis after giving birth to a healthy child is much smaller (approximately 7 per cent). Our data indicate that those couples who experienced at least one termination of pregnancy decide in favour of a further pregnancy within due course. Problems of prenatal prediction and risk of misdiagnosis With regard to the reliability of prenatal diagnosis in infantile SMA, several aspects have to be taken into account.
( I ) The accuracy of clinical diagnosis-Only 91 families who fulfilled the diagnostic criteria of the International SMA Consortium were considered for prenatal prediction, whereas 28 had to be
413
excluded because of various clinical or genetic reasons. A major problem is the distinction of conditions from classical infantile SMA which may also have associated anterior horn cell involvement and are probably not linked to chromosome 5q (‘SMA plus’ forms, Table IV). Diagnostic doubts or the presence of additional or atypical features were the most frequent reasons for exclusion (n=16). Similar experiences were reported by Cobben et al. (1993a). ( 2 ) Heterogeneity in proximal SMA-In none of our multiplex families did we ever fmd a situation indicating non-linkage to 5q. In other studies, 5-10 per cent of unlinked SMA families have been reported (Penchaszadeh et al., 1991); Daniels et al. (1992) even suggested that the probability of an isolated case with mild SMA being autosomal recessive was only 40 per cent, the remainder of the risk being attributable to non-inherited causes. Brzustowicz et al. (1993) found genetic heterogeneity in three out of 38 multiplex chronic SMA families. Since clinical data are missing, the results are not convincing; the authors themselves mentioned diagnostic doubts in their families. So far, only a few families (less than 1 per cent?) have been described in which the SMA locus was identical in two affected sibs and one unaffected sib, indicating the existence of a second autosomal recessive gene locus (Cobben et al., 1993b, 1994; MacKenzie et al., 1994). In our linkage studies with chromosome 5q markers, 5 per cent of families showed identical haplotypes in the affected and at least one unaffected sib (Rudnik-Schoneborn et al., 1994b). This has been interpreted mainly as a result of new autosomal dominant mutations and was supported by further observations. In two families, prenatal prediction could not be offered since they gave evidence of affected subjects in two generations. Autosomal dominant proximal SMA is not located on chromosome 5 q (Kausch et al., 1991); therefore the risk of a false prediction is difiicult to assess. Phenocopies or recessive new mutations of SMA also have to be considered as explanations for non-5q linkage. While phenocopies have not yet been described in typical SMA, recessive de novo mutations have been the focus of recent attention (Melki et al., 1994; Wirth et al., 199%). It remains to be clarified whether undetected new recessive mutations are relevant risk factors for prenatal misdiagnosis.
Weakness and hypotonia usually from birth Cerebellar signs (nystagmus, vision impairment, ataxia, mental retardation) Rapid course (death usually at
