De novo SPAST mutations may cause a complex SPG4 phenotype

Sir, Spastic paraplegia type 4 (SPG4) is caused by mutations in the SPAST gene, and is the most common form of autosomal dominantly inherited pure hereditary spastic paraplegia (HSP) worldwide (Hazan et al., 1999; Salinas et al., 2008; Finsterer et al., 2012; Dong et al., 2018; Koh et al., 2018; Parodi et al., 2018). Age of onset and severity of SPG4 are known to be highly variable. In a recent issue of Brain, Parodi et al. (2018) described an impressive series of 842 patients with SPG4. They found that the mutation type and the patient’s sex modified the age of onset and disease severity. We have carefully read their paper, with specific interest in the variable phenotype of SPG4 and the mode of inheritance. The authors reported that most of the patients had a positive family history; only 5.7% of all cases occurred sporadically, similar to all large (population) studies on SPG4. Parodi et al. (2018) did not explicitly mention whether the sporadic cases in their cohort were due to de novo SPAST mutations. In general, the identification of sporadic patients with SPG4 is attributed to common mechanisms like incomplete penetrance, somatic mosaicism, non-paternity, and inadequate clinical assessment of the parents. Altogether, true de novo occurrence of SPAST mutations, i.e. where both parents of the patient are proven not to have the mutation in lymphocytes, appears rare and has—to our knowledge—never been studied separately. Confronted with the unanticipated complex phenotypes of the first young children that we recently diagnosed with de novo SPAST mutations (see below) and the lack of knowledge on the topic, we thoroughly reviewed the literature using Pubmed, searching for papers using the terms ‘SPG4’, ‘SPAST’, ‘spastin’, and ‘hereditary spastic paraplegia’. Reference lists were screened for additional papers. All papers were then searched for the description of cases with sporadic HSP in the abstract and the results and discussion sections, assuming that cases with de novo mutations would be identified by this strategy. Additionally, every paper was also screened by an electronic search for the word ‘novo’ in the pdf file of the paper. We found that, since the discovery of the SPAST gene (Hazan et al., 1999), a large number of papers has been published on the clinical and molecular features of SPG4, including many case reports (not all included in the reference list) but also numerous large studies on selected cohorts of hundreds of patients with familial and isolated forms of HSP from all different parts of the world (Fonknechten et al., 2000; Crippa et al., 2006; Depienne et al., 2006; Magariello et al., 2006, 2010; Salinas et al., 2008; Shoukier et al., 2009; De Bot et al., 2010; Finsterer et al., 2012; Elert-Dobkowska et al., 2015; Park et al., 2015; Balicza et al., 2016; Chrestian et al., 2016; Mészárosová et al., 2016; Burguez et al., 2017; Dong et al., 2018; Koh et al., 2018; Parodi et al., 2018). From the literature, we were able to identify only 14 patients who were reported to have true de novo SPAST mutations (Table 1) (Crippa et al., 2006; Blair et al., 2007; Depienne et al., 2007; Battini et al., 2011; Aulitzky et al., 2014; Elert-Dobkowska et al., 2015; Mészárosová et al., 2016; Polymeris et al., 2016; Burguez et al., 2017; Gillespie et al., 2018). Interestingly, all of these patients suffered from pure HSP. Six of 14 patients were identified in studies that had pure HSP as an inclusion criterion, which may partly explain the absence of divergent phenotypes in the literature. Disease onset was (far) below 10 years in nine patients (Patients 2, 5, 8–12 and 14) and after 25 years in four patients (Patients 1, 3, 7 and 13). Of note, Patient 6 was an asymptomatic mother with a somatic mosaicism for a SPAST mutation and three children affected with SPG4. Patients with de novo SPAST mutations reported in the literature Patients with de novo SPAST mutations reported in the literature After this literature study, we searched the exome database of the Department of Human Genetics of the Radboud University Medical Centre for other patients in whom de novo mutations in the SPAST gene had been identified. In all cases mutation analysis of both parents had been performed and found to be negative. Referring clinicians were asked to contribute to this study by sharing clinical, neuroimaging and molecular data of their patients, after having obtained informed consent of the patient and/or the parents, in order to get a first impression of the clinical significance of the de novo occurrence of SPAST mutations. Doing so, we identified 13 patients (Table 2) and found that six patients (Patients 1, 3, 8, 10, 11 and 13) had pure HSP (note, Patient 13 may be too young for final phenotyping). Contrasting this classic phenotype in six patients, however, the other seven patients showed an exceptional phenotype. Patients 2, 4–7, 9 and 12 suffered from a complex, generalized and progressive motor disorder with onset in the first or second year of life. Additionally, they almost invariably also suffered from severe, progressive loss of cognitive functions, loss of spoken language, and severe dysphagia requiring tube feeding. Three of these patients also had seizures. MRI studies in these 13 Dutch patients had been performed at different ages, on different machines, and following different protocols; irrespective of these disadvantages we could conclude that cerebral imaging generally did not show any abnormalities, neither in the patients with pure HSP nor in the more severely affected children. Clinical and molecular characteristics of 13 Dutch patients with SPG4 caused by de novo SPAST mutations Age at 31 December 2018, in years unless otherwise stated. The parents of most young patients only gradually became aware of the fact that the gross motor development of their children progressed too slowly, and their legs were stiff; therefore, it was impossible to determine the exact age at which symptoms occurred. For that reason, we decided to denote ‘in the first 2 y of life’ here, in all these cases; only when the history was evidently different, and an age at onset could be given more precisely, we indicated this. Whole exome sequencing led to the molecular diagnosis in all patients, except Patients 1, 3, 8 and 9, in whom targeted SPAST mutation analysis was performed. Known SPAST mutations: c.1496G>A (Park et al., 2015); c.1775T>A (Parodi et al., 2018); c.1360G>A (Fogli et al., 2009). The proposed modifier c.131C>T (p.Ser44Leu) (Svenson et al., 2004) was only found in the SPAST gene in Patients 1 and 12. Novel mutations: c.1326A>T; c.1226C>T; c.1169T>G; c.1072G>T; and c.1483G>C. The same codon is affected by a known mutation, 1324G>A, Glu442Lys (Ribai et al., 2008; Parodi et al., 2018). Alternatively, c.1226C>T may also lead to p.Ala409_Tyr415del. Complex HSP. AAO = age at onset; AAMD = age at molecular diagnosis; CHD = congenital hip dysplasia; mo = months; y = years. Clinical and molecular characteristics of 13 Dutch patients with SPG4 caused by de novo SPAST mutations Age at 31 December 2018, in years unless otherwise stated. The parents of most young patients only gradually became aware of the fact that the gross motor development of their children progressed too slowly, and their legs were stiff; therefore, it was impossible to determine the exact age at which symptoms occurred. For that reason, we decided to denote ‘in the first 2 y of life’ here, in all these cases; only when the history was evidently different, and an age at onset could be given more precisely, we indicated this. Whole exome sequencing led to the molecular diagnosis in all patients, except Patients 1, 3, 8 and 9, in whom targeted SPAST mutation analysis was performed. Known SPAST mutations: c.1496G>A (Park et al., 2015); c.1775T>A (Parodi et al., 2018); c.1360G>A (Fogli et al., 2009). The proposed modifier c.131C>T (p.Ser44Leu) (Svenson et al., 2004) was only found in the SPAST gene in Patients 1 and 12. Novel mutations: c.1326A>T; c.1226C>T; c.1169T>G; c.1072G>T; and c.1483G>C. The same codon is affected by a known mutation, 1324G>A, Glu442Lys (Ribai et al., 2008; Parodi et al., 2018). Alternatively, c.1226C>T may also lead to p.Ala409_Tyr415del. Complex HSP. AAO = age at onset; AAMD = age at molecular diagnosis; CHD = congenital hip dysplasia; mo = months; y = years. Most of the SPAST mutations that occurred de novo in the patients identified in the literature study (Table 1) and our cohort (Table 2) are also present in families with multiple generations with pure HSP. The common c.1496G>A mutation was even found in nine of the 27 (33%) patients identified with a de novo SPAST mutation; remarkably, five of the seven Dutch patients with the complex SPG4 phenotype harboured this mutation. In addition to a de novo mutation, the proposed modifier c.131C>T (p.Ser44Leu) (Svenson et al., 2004) was found in the SPAST gene in 2 of the 13 Dutch patients, one with a pure HSP phenotype (Patient 1) and one with a complex phenotype (Patient 12). The low number of cases and the relative high occurrence of c.131C>T in controls (between 1–2% in Europe; Database ExAC, Lek et al., 2016) does not allow for any conclusions. We have no explanation for the fact that some SPAST mutations can lead to pure HSP as well as to the highly complex phenotype presented here. The SPAST mutations that may be associated with complex SPG4 when occurring de novo never seem to cause such a severe phenotype in individuals from large, multi-generation SPG4 families. This observation pleads against the presence of a simple, independent explanation in addition to the de novo occurrence of the SPAST mutation, such as a genetic or environmental modifier, for the existence of the complex SPG4 phenotype. Based on all available data, only a minority of patients with de novo SPAST mutations appear to suffer from the complex SPG4 phenotype. The high number of patients (22/27, 81%) with a de novo SPAST mutation and an extremely early onset of disease, in our series (13/13) as well as the literature (9/14), however, may point to a relationship between age of onset and the de novo occurrence of the mutation. Of note, this observation may also simply reflect the greater need to study parental carriership (and thus the chance to find a de novo mutation) in paediatric practices compared to adult neurogenetic services. With regard to the apparent (extreme) rarity of the complex SPG4 phenotype, it seems reasonable to assume that the true incidence may be higher since many paediatric cases with such a severe phenotype may never have had SPAST mutation analyses and thus may have been missed in the era before next generation sequencing techniques became available. This possibility might be illustrated by the fact that we identified most Dutch patients in a relatively short period. Recognition and acceptance of the complex SPG4 phenotype at the far end of the clinical spectrum of SPG4 are the first steps towards an appropriate diagnosis. Future studies will hopefully contribute to the further delineation of the complex phenotype, and the understanding of its underlying molecular mechanisms. In conclusion, we recognized a complex SPG4 phenotype with an infantile onset, complex, generalized, and progressive motor disorder and severe cognitive decline and loss of spoken language. This phenotype has never been reported in SPG4 families, and it appears to us that only de novo SPAST mutations can cause such a complex phenotype. Our clinical, observational study does not allow us to explain the remarkable findings. Nevertheless, we gladly add our data to the recent work of Parodi et al. (2018), and are willing to open a scientific discussion on what we think is a new clinical SPG4 phenotype. The data that support the findings of this study are available from the corresponding author, upon reasonable request. No funding was received towards this work. The authors report no competing interests.