Reply: ATAD1 encephalopathy and stiff baby syndrome: a recognizable clinical presentation

Sir,

We welcome the letter by Wolf et al. (2018). They report a third family with ATAD1-related lethal encephalopathy, further confirming ATAD1 as a disease gene for the well recognizable stiff baby syndrome. ATAD1 encodes thorase, an AAA+ ATPase that helps control postsynaptic AMPA receptor internalization by disassembling complexes between glutamate receptor interacting protein (GRIP1) and the AMPA receptor subunit GluA2 (Zhang et al., 2011). The common clinical features shared by all affected members in the three reported families comprise hypertonia, absence of spontaneous movements, almost no motor development, and death within the first months of life (Ahrens-Nicklas et al., 2017; Piard et al., 2018; Wolf et al., 2018). Myoclonic jerks and seizures have been noticed in the affected children reported by Ahrens-Nicklas et al. (2017) and the female described by Wolf et al. (2018), but in none of the three affected infants we reported (Piard et al., 2018). We agree with Wolf et al. (2018) that ATAD1-related encephalopathy should be a first-line diagnosis in extremely stiff neonates with or without seizures. However, we are aware of another infantile epileptic encephalopathy, known as rigidity and multifocal seizure syndrome, lethal neonatal (RMFSL; MIM 614498), showing an overlapping clinical picture. This phenotype is characterized by intractable seizures, hypertonia, and early death and caused by biallelic mutations in BRAT1 (Puffenberger et al., 2012; Saunders et al., 2012; Saitsu et al., 2014). We therefore recommend ATAD1 and BRAT1 to be preferentially analysed by Sanger-sequencing or in next-generation sequencing panels in the diagnostic work-up of stiff neonates.

The ATAD1 variant identified by Wolf et al. (2018), a homozygous c.162G>C transversion predicting the amino acid substitution p.(Gln54His), has not been tested in the patient’s parents to demonstrate that both are heterozygous carriers. Moreover, Wolf et al. (2018) did not comment on minor allele frequency or pathogenicity predictions of the ATAD1 variant. By doing so, we noticed that the c.162G>C change is absent from population databases (ExAC and gnomAD Browsers) and predicted to be damaging by three in silico tools combining previous pathogenicity scores [CADD (Kircher et al., 2014), REVEL (Ioannidis et al., 2016), and M-CAP (Jagadeesh et al., 2016)]. As the G-to-C change affects the last nucleotide of exon 2, splicing of the ATAD1 pre-mRNA has been postulated to be altered (Wolf et al., 2018). We investigated the effect of this sequence change on splicing using four splice site prediction programs: all of them detected the wild-type donor site (http://www.umd.be/HSF3/HSF.shtml; http://www.cbs.dtu.dk/services/NetGene2/; http://www.fruitfly.org/seq_tools/splice.html; http://bioinfo.itb.cnr.it/oriel/splice-view.html). Three recognized the donor site in the mutant sequence with reduced splicing efficiency, while one predicts the complete loss of the donor site. The effect of the ATAD1 mutation c.162G>C on pre-mRNA splicing and protein function is difficult to predict in the absence of mRNA and functional studies. Nonetheless, it is tempting to speculate that both aberrant ATAD1 mRNAs, possibly subjected to nonsense-mediated mRNA decay, and mutant ATAD1^Gln54His protein had been produced in the affected infant and likely caused (nearly) complete ATAD1 loss-of-function. Similarly, the homozygous ATAD1 mutation reported by Ahrens-Nicklas et al. (2017) also was a loss-of-function mutation as the c.826G>T change predicted the introduction of a premature termination codon [p.(Glu276*)]. A drastically decreased ATAD1 mRNA amount and the absence of thorase protein in lymphoblastoid cells derived from one patient further underscored ATAD1 null alleles in the affected individuals of this family (Ahrens-Nicklas et al., 2017). In contrast, we would like to emphasize that the ATAD1 mutation c.1070_1071delAT/p.(His357Argfs*15) reported in our family turned out to have a gain-of-function effect, a rarely observed mechanism in autosomal recessively inherited disorders (Piard et al., 2018). Although at first glance the frameshift variant gives the impression of an ATAD1 loss-of-function mutation, the functional work presented in our study indicated that the encoded thorase mutant^{His357Argfs*15} is expressed in patient cells and gained a novel function of its C-terminal end. We demonstrated decreased levels of surface AMPA receptors in unstimulated mutant^{His357Argfs*15}-expressing neuronal cells compared to wild-type-expressing neurons. We put forward the hypothesis that the thorase mutant^{His357Argfs*15} may inhibit the recycling back and/or reinsertion of AMPA receptors to the surface following endocytosis resulting in a decrease in the steady-state levels of these receptors at the cell surface (Piard et al., 2018). In contrast, biallelic ATAD1 loss-of-function mutations likely increase the population of excitatory postsynaptic AMPA receptors in neurons as shown in Atad1^−/− mice (Zhang et al., 2011). In the latter scenario, perampanel, an AMPA receptor antagonist, was shown to decrease tonicity and prevent seizures in two individuals with the homozygous ATAD1 nonsense mutation p.(Glu276*) (Ahrens-Nicklas et al., 2017). Taken together, these data demonstrate clear benefits of the individualized medicine approach in the two severely affected children reported by Ahrens-Nicklas et al. (2017), as also pointed out by Wolf and colleagues (2018). However, we respectfully but strongly disagree with the statement by Wolf et al. (2018) to ‘… ensue a therapeutic trial with perampanel, even before genetic confirmation’ in an extremely stiff infant. As both loss- as well as gain-of-function ATAD1 mutations give rise to the shared neurological features of stiffness and absence of spontaneous movements, we recommend (i) genetic testing in the affected neonates, including ATAD1 and BRAT1, to unravel the genetic basis of the disease; and (ii) determining the consequences of the ATAD1 mutation, at least on mRNA and protein level in cells obtained from patients. Absence or reduced amount of mutated ATAD1 transcripts and/or thorase protein in patient-derived cells provide evidence for ATAD1 deficiency, thus opening up the therapeutic potential of perampanel in these patients. On the other hand, therapy with perampanel is not an option in individuals with an ATAD1 gain-of-function variant, such as the p.(His357Argfs*15) frameshift mutation reported in our study (Piard et al., 2018). Wolf et al. (2018) pointed out the importance of detailed clinical work-up for rare neurological disorders and we agree. Beyond that, the work by Ahrens-Nicklas et al. (2017) and our work (Piard et al., 2018) highlight the need for determining the molecular basis of rare and devastating neurological disorders and their underlying pathophysiological mechanisms to develop a targeted therapy.

Funding

This work was supported by a grant from the Deutsche Forschungsgemeinschaft (KU 1240/10-1 to K.K.) and National Institutes of Health – National Institute on Drug Abuse, P50DA000266 to V.L.D and T.M.D.