Abstract
Keywords
INTRODUCTION
Spinal muscular atrophy (SMA) is an autosomal recessive neuromuscular disorder characterized by the degeneration of motor neurons in the anterior horns of the spinal cord. SMA is caused by the homozygous inactivation of the SMN1 gene (Survival of Motor Neuron 1, MIM 600354) [1]. Estimated incidence is 1 in 6,000 to 1 in 10,000 live births, depending on the ethnicity [2]. SMA is associated with a wide spectrum of clinical severity and is classically classified into four groups (type I to type IV) based on age of onset and achieved motor milestones [3, 4]. Type I SMA is the most common and most severe subtype, defined by onset prior 6 months of age and lack of ability to sit independently. Very soon after the identification of the SMN1 gene, SMA cases with more severe phenotype than type I were described, extending the phenotypic spectrum of the disease [5]. In 1999, based on several publications of such severe SMA cases, a new form called type 0 was described [6]. Recently, Finkel et al. defined type 0 SMA by fetal onset and diagnosis at birth with the following clinical features: paucity of movement in limbs, face, trunk, no suck, muscle atrophy, areflexia, congenital contractures, and requirement for mechanical ventilation support at birth [7]. To date, only 37 patients corresponding to this very severe form of SMA, characterized by fetal onset and diagnosis at birth, have been reported [5, 8–29]. Little is known about their prenatal manifestations and the number of SMN2 copies of these patients has not been determined consistently.
The SMN2 gene (Survival of Motor Neuron 2, MIM 601627), a paralogous gene of SMN1, contains a C to T substitution at position +6 of exon 7 resulting in inefficient inclusion of exon 7 in the majority of SMN2 transcripts [30]. SMA patients, lacking full length transcripts originating from the SMN1 gene, are dependent on their SMN2 gene, which is not able to produce sufficient levels of SMN protein. SMN2 copy number is highly variable and has been shown to be the primary genetic modifier of SMA severity. Most SMA type I patients have two copies of SMN2, three SMN2 copies are common in SMA type II, while type III and IV generally have three or four [31]. The 0SMN1/1SMN2 genotype is rare (about 3.5% of SMA cases in the largest published series [31]) and is the most severe genotype that has been described in SMA patients.
To further delineate the phenotype of type 0 SMA and to characterize prenatal manifestations of this severe neonatal disease, we studied a retrospective cohort of SMA patients with the 0SMN1/1SMN2 genotype. Here, we report prenatal and postnatal data for 16 patients with this genotype and type 0 SMA and compare these data to those of the 37 type 0 SMA patients previously described.
MATERIALS AND METHODS
Patients
Patients with the 0SMN1/1SMN2 genotype were collected using the UMD-SMN1 France database for SMA patients. The UMD-SMN1 database oversight committee gave its approval for this database questioning. From 1999 until 2014, 1090 patients with an homozygous deletion of the SMN1 gene and assessment of SMN2 copy number were recorded through the database. Twenty of them had a single copy of the SMN2 gene. Referring clinicians of these 20 patients were contacted. Clinical information from obstetric and neonatology departments were available for 16 patients and systematically reviewed. Most patients underwent an examination both by a pediatrician and a clinical geneticist.
Determination of SMN1 and SMN2 copy number
At diagnosis, blood sample were obtained after written informed consent for each patient from parents. SMN1 and SMN2 copy numbers were determined by using assays based on the QMPSF (Quantitative Multiplex PCR of Short Fluorescent Fragments) method [32, 33] or by determination of the total SMN copy number, using competitive PCR, and of the SMN1 and SMN2 gene relative ratio, using a primer extension assay [34].
RESULTS
We report 16 unrelated patients (6 males and 10 females) with the 0SMN1/1SMN2 genotype(Table 1).
Postnatal features
The median gestational age at delivery was 39 weeks (range, 36–42 weeks).
Neuromuscular system involvement
All patients had severe neuromuscular involvement at birth including profound hypotonia (n = 16/16), often associated with a froglike position (n = 7/9), and severe muscle weakness (n = 15/15) characterized by a lack of spontaneous movements of the limbs or only minimal movements of the extremities. Absence of deep tendon reflexes (n = 14/14), and presence of tongue fasciculation (n = 7/8) were common findings, that were suggestive of anterior horn cell involvement.
Respiratory distress
All patients had dramatic respiratory impairment. In most cases (n = 14/16), mechanical respiratory assistance was required within the first minutes or first hours of life. Patient 10 had progressive respiratory distress and required respiratory assistance at the age of 10 days. Patient 14 required intubation at 24 hours of life following an episode of severe bradycardia. All patients remained dependent on artificial ventilation.
Cranial nerve involvement
At birth, impairment of suck and swallow reflexes were constantly reported (n = 11/11). Weakness of the facial muscles was also frequently observed (n = 7/11). Extraocular muscle involvement was not reported (n = 0/9).
Autonomic nervous system dysfunction
Patients developed recurrent episodes of bradycardia (n = 7/7). Otherwise, vasomotor symptoms were reported in patient 10, consisting of excessivesweating.
Cerebral function
Eye contact was conserved in most cases (n = 9/14).
Fetal Akinesia Deformation Sequence (FADS)
The vast majority of patients had joint contractures at birth (n = 15/16), ranging from one affected joint to generalized contractures. Contractures involved either distal or proximal joints. Dysmorphic craniofacial features, present in more than half of the individuals (n = 8/13), mainly included micrognathia or high-arched palate. Other dysmorphic features such as abnormal ears, broad nasal bridge, up-slanting palpebral fissures or epicanthic folds were inconsistently observed. Peripheral edema was reported in n = 6/7 patients. Some patients had intrauterine growth retardation (n = 5/16). Other findings were reported such as single palmar crease (n = 3), narrow chest (n = 2), and excess nuchal skin (n = 2).
Congenital heart defects
Echocardiography was performed in 13 patients. Heart defects were found in 9 of them. The most commonly reported malformations were septal defects (Table 2).
Life expectancy
Infants did not survive beyond one month of age as in most cases, the decision to limit or withdraw intensive care was made after the diagnosis of SMA was confirmed. Median age at death was 15 days (range 6–33 days).
Additional anomalies
In a few patients, several additional anomalies were observed such as hypoplastic nails (patients 1, 5 and 11), inverted nipples (patient 2 and 11), xerosis (patient 5), keratosis pilaris (patient 8), and corpus callosum atrophy (patient 13). The link between these features and SMA is not obvious. In addition, patient 11 had microcephaly, thumb duplication, hirsutism, anterior anus, sacral dimple and overlapping calvarial sutures that evoked an additional syndrome, which has not been identified.
Prenatal manifestations
Diagnosis of SMA had been suspected prenatally for none of the patients. We retrospectively analyzed clinical/sonographic abnormalities that were noted prenatally in these patients (Table 1).
Decreased fetal movements (DFM)
DFM was the most commonly reported manifestation (n = 9/16). The median gestational age at which DFM occurred was 32 weeks (range, 24–34 weeks). For 6 patients, DFM was noticed exclusively by mothers (retrospectively in 2 cases). DFM was objectified by prenatal ultrasound scan in 3 cases. In patient 4, ultrasound at 24 weeks of gestation (WG), performed following the identification of talipes, showed reduced lower limb movements. In patient 8, third trimester ultrasound showed DFM, that was also perceived by the mother, but repeated ultrasound examinations did not confirm it. In patient 9, ultrasound scan was performed following a fall of the mother at 36 WG. At that time, DFM, which was perceived for several weeks by the mother, was objectified. For other patients, the fetus vitality was considered as normal on ultrasound scans.
Joint contractures
Joint contractures were reported in only one patient. In this patient (patient 4), left talipes equinovarus was observed at 23 WG and malposition of the right foot was noted at 24 WG.
Abnormal amniotic fluid volume
Polyhydramnios, a classic sign of FADS, was not observed on prenatal ultrasounds. However, for 2 patients, excess of amniotic fluid was noted at the time of rupture of membranes. In contrast, for patient 14 and 15, oligohydramnios was noted prenatally.
Congenital heart defects (CHD)
CHD was prenatally observed in only 2 cases. In patient 3, CHD was suspected on first trimester ultrasound, then ventricular septal defect and aortic dysplasia were shown on echocardiography at 24 WG. In patient 16, dilatation of the right atrium was observed on ultrasound scan at 15 WG.
Increased nuchal translucency (NT)
Increased NT at first trimester was frequently reported (n = 7/13). Patient 2 had hygroma and lymphedema at 12 WG that quickly regressed, patient 16 had NT measurement at 8.7 mm and fetal hydrops that regressed spontaneously. In patient 4, whose NT measured 3.4 mm, peripheral edema was noted at 22 WG. Other reported NT measurement were 3.1 (n = 2) and 3.8 in patients 1, 3 and 5 respectively. Karyotype was performed in 6 of these fetuses and was normal. Karyotype was performed after birth in the seventh patient (patient 1) and was normal.
DISCUSSION
Considering that the 0SMN1/1SMN2 genotype is the most severe genotype compatible with life, we analyzed 16 patients with this genotype in order to further delineate the phenotype of the most severe form of SMA. This constitutes the largest series of type 0 SMA. Because, in previous type 0 SMA reports, only few prenatal manifestations were reported, we aimed at consistently describing the prenatal period.
Among the 1090 SMA patients with SMN2 copy number assessment listed through the UMD-SMN1 France database, 20 had the 0SMN1/1SMN2 genotype, representing 1.8% of all SMA subtypes. Given that the incidence of SMA is 1 in 6,000 births in France [1], the incidence of the 0SMN1/1SMN2 genotype could be estimated at 1/300,000. Presentation of all analyzed patients with the 0SMN1/1SMN2 genotype was consistent with type 0 SMA, suggesting a strong correlation between the 0SMN1/1SMN2 genotype and type 0 SMA. Among the 37 type 0 SMA patients previously reported, SMN2 copy number had been assessed in only 10 patients. Nine of them had the 0SMN1/1SMN2 genotype [8, 27] and one had the 0SMN1/2SMN2 genotype [24].
Among the 20 patients with the 0SMN1/1SMN2 genotype recorded in the UMD-SMN1 France database, we were able to collect clinical data for 16 patients. In this series, the male-to-female ratio was 0.6. In contrast, for patients previously described and recruited on the basis of the clinical presentation, this ratio was 3. A sex-dependent modifier gene has been suggested by the observation of asymptomatic females with the 0SMN1/4SMN2 genotype [35, 36]. Similarly, we can hypothesize a difference in SMA severity between male and females, with more severe effects in males. Under this hypothesis, type 0 SMA males with the 0SMN1/2SMN2 genotype may be identified. In the same way, 0SMN1/1SMN2 genotype may be lethal in some males. Therefore, we suggest to systematically assess SMN2 copy number in patients with type 0 SMA.
Analysis of the 16 type 0 SMA patients reported in this series and of the 37 additional cases previously described shows that most infants with type 0 SMA, at birth, had profound hypotonia, severe muscle weakness, and respiratory distress requiring ventilator support within the first hours of life (Table 1). Only a few patients achieved some degree of spontaneous ventilation [13, 24]. One of these patients had the 0SMN1/2SMN2 genotype [24]. These few patients could be considered between type 0 and type I SMA. Type I SMA has been subclassified into type Ia, Ib and Ic [37]. Type Ia has been used by some authors, as a synonym of type 0, to define the most severe form of SMA. In the recent 209th ENMC International Workshop, a distinction was made between type 0 and type Ia, type 0 referring to clear fetal onset and type Ia to clear onset within the first two weeks of life [7]. Indeed, SMA is a continuum and a delineation of distinct subtypes may be difficult in clinical practice.
In most patients, cranial nerve involvement was present at birth, as attested by inability to suck/swallow, facial weakness and, much less frequently, weakness of extra-ocular muscles [12, 17]. Several patients developed episodes of bradycardia [10, 26], reflecting autonomic nervous system dysfunction. Interestingly, the SMNΔ7 model of SMA also has bradycardia likely attributable to aberrant autonomic signaling and develops marked dilated cardiomyopathy resulting at least partially from early bradycardia [38]. Digital necrosis has also been described in type 0 or type 1 SMA patients whose survival has been extended because of supportive care [25, 26]. This digital necrosis is probably linked to autonomic dysfunction. In type 0 SMA patients, neurologic involvement at birth was therefore comparable to that of type I SMA in end-stage, particularly as regards brainstem involvement. Eye contact was conserved in most cases (n = 9/14 patients in this series), reflecting normal cerebral functions, as observed in other types of SMA. However, five patients had poor arousal, suggesting as a first hypothesis a possible perinatal insult caused by respiratory distress at birth. In addition, type 0 SMA patients showed characteristics of FADS, with predominantly joint contractures, but also micrognathia, high-arched palate, peripheral edema, and less frequently intrauterine growth retardation, narrow chest or fractures. Therefore, in a newborn with respiratory distress, profound hypotonia, severe muscle weakness, and joint contractures, tongue fasciculation should be investigated and may lead to diagnose type 0 SMA.
CHD appears as a common finding in type 0 SMA patients (n = 9/13 in this series, n = 17/18 in previous reports [5, 29]) and includes most often atrial or ventricular septal defects, but also aortic coarctation, patent ductus arteriosus, mitral hypoplasia, hypoplastic left heart, common atrium, dilated right ventricle, anomalous pulmonary venous return, dilated right atrium, aortic dysplasia, pulmonary and aortic stenosis, and asymmetric ventricular hypertrophy. Cardiac defects have also previously been reported in severe SMA mouse models [39]. Septum defects are hard to detect during fetal development, explaining why most of CHD were only discovered at birth. Considering that incidence of CHD, even including mild forms, is about 7.5% live births [40], the proportion of type 0 SMA patients with CHD in this series (n = 9/13; 69%) shows an association between type 0 SMA and CHD (p < 0.00001, proportion test). Therefore, this study further supports the hypothesis raised by Rudnik-Schöneborn et al. based on the observation of 3 patients with CHD out of 4 type 0 SMA patients, that SMN, which is ubiquitously expressed, is involved in cardiogenesis [23]. In these rare cases of SMA with only one SMN2 copy, the very low amount of SMN might be insufficient to assure normal cardiogenesis.
Association between SMA and increased NT has been questioned before. Parra et al. prospectively studied NT measurements of 19 fetuses affected by SMA. The only fetus who had an increased NT was the only one with 1 SMN2 copy [41]. Two other publications reported SMA in fetuses after a prenatal diagnosis because of type 0 SMA history in sibling. These two fetuses had increased NT [11, 22]. Other authors retrospectively studied NT measurement in 29 fetuses with SMA, including mainly type I, and found no case of increased NT [42]. Thus the association between increased NT and SMA remained an open question. Here, we report 7 new cases of type 0 SMA with increased NT (n = 7/13; 54%) (Table 1). Increased NT, which is defined by a NT measurement above the 95th percentile, thus appears to be associated with type 0 SMA (p < 0.00001, proportion test). Interestingly, among the 11 type 0 SMA patients with increased NT (n = 7/13 in this series and n = 4/5 in previous reports [11, 20]), 9 were reported to have CHD. In addition, 2 of the 3 fetuses with increased NT that were diagnosed prenatally because of an affected type 0 SMA sib, also had CHD [22, 41]. Increased NT in type 0 SMA patients could be explained by CHD as previously suggested by Parra et al. [41]. Finally, since type 0 is very rare (approximately 1/300,000 considering the correlation with the 0SMN1/1SMN2 genotype) compared to the frequency of increased NT and normal karyotype (approximately 4%, if we consider that, among fetuses with increased NT, nearly 20% have chromosomal aberrations [43]), the positive predictive value of increased NT is extremely low (approximately 4.5×10- 5). Therefore, systematic screening for SMA in any increased NT with normal karyotype does not seem to be appropriate.
In this series as in previous reports, DFM appears as the most common prenatal manifestation (n = 35/46). DFM was mostly reported exclusively by mothers, with a median gestational age of 32 WG (range 24–38 WG). DFM may occur at the end of the second trimester or during the third trimester.
Importantly, no arthrogryposis was reported prenatally in type 0 SMA patients, indicating that joint contractures may occur in late stages of pregnancy. The lack of early joint contractures is consistent with the evidence that early fetal movements are not impaired in SMA [27]. Thus, the most severe form of SMA compatible with life does not lead to early FADS that could be recognizable in the first or second trimester of pregnancy. Nevertheless, lethal forms of FADS related to SMA may exist. To our knowledge, only one case of SMA has been diagnosed, after the termination of pregnancy, in a patient manifesting prenatal arthrogryposis (gender and SMN2 copy number not available) [44]. The 0SMN1/0SMN2 genotype has never been identified and is suspected to lead to early embryonic lethality, consistent with multiple organ dysfunction. Moreover, our observation of a gender imbalance in the most severe forms of SMA compatible with life suggests the possibility of more severe prenatal presentations in males with the 0SMN1/1SMN2 genotype. SMA screening in lethal FADS with suggestive neuropathological findings could potentially lead to the identification of such exceptional genotypes.
Polyhydramnios was reported for 11 type 0 SMA patients in previous reports [10, 29]. However, results in this series show that polyhydramnios is not a frequent manifestation (Table 1), suggesting that swallowing ability may be preserved until late stage of pregnancy. Thus, cranial nerves involvement, which is present from birth, seems to occur at the end of pregnancy, as already observed for DFM and joint contractures.
Finally, we wondered if evocative prenatal phenotype could guide the diagnosis of SMA. We therefore analyzed the combination of DFM, joint contracture, polyhydramnios, CHD, and increased NT in this series. In two patients, none of these manifestations were observed. Only one patient displayed more than two manifestations (Table 1). We thought that the association of DFM during the third trimester of pregnancy with increased NT with normal karyotype and/or detection of CHD and/or polyhydramnios might represent a warning sign of SMA. But this association was observed in only 4 out of the 16 patients. Moreover, DFM is a subjective and late manifestation. There is thus no typical phenotype of type 0 SMA during pregnancy. Because manifestations are not specific and not consistent, it appears to be hard to suspect the diagnosis of type 0 SMA prenatally.
In conclusion, type 0 SMA highlights the wide spectrum of severity of neurologic impairment in SMA and the predominant role of SMN2 as a modifier gene. Moreover, involvement of the heart further extends the phenotypic spectrum of SMA. Interestingly, restoring SMN expression solely in motor neurons does not extend survival in both transgenic mouse models and zebrafish [45, 46]. Considering SMA not only as a motor neuron disease but as a multi-organ disease [47, 48] is of high importance given the ongoing clinical trials that aim at increasing SMN expression. Thus, therapeutic drug should allow SMN restoration not only in the nervous system but also in the peripheral tissues [38, 49].
CONFLICT OF INTEREST
The authors have no conflict of interest to report.
Footnotes
ACKNOWLEDGMENTS
UMD-SMN1 database was supported by AFM (Association Française contre les Myopathies)
We thank JP. Bonnefont and A. Totoescu-Serefian for their help in data collection, S. Fehrenbach for her technical help, C. Le Clezio for her help in statistical analyzes, M Tosi for his help in editing the manuscript, F. Amram, H. Bourdial, F. Castela, V. Drouin-Garraud, AL. Duigou, C. Fichtner, S. Galène-Gromez, B. Guillois, M. Mathieu-Dramard, AG. Le Moing, F. Lesage, S. Lyonnet, S. Magnin, J. Nakhleh, S. Odent, C. Oheix, M. Oualha, M. Rio, L. Trestard, and C. Vanhulle for their cooperation.