Phenotype of CNTNAP1: a study of patients demonstrating a specific severe congenital hypomyelinating neuropathy with survival beyond infancy

Abstract

CHN is genetically heterogeneous and its genetic basis is difficult to determine on features alone. CNTNAP1 encodes CASPR, integral in the paranodal junction high molecular mass complex. Nineteen individuals with biallelic variants have been described in association with severe congenital hypomyelinating neuropathy, respiratory compromise, profound intellectual disability and death within the first year. We report 7 additional patients ascertained through exome sequencing. We identified 9 novel CNTNAP1 variants in 6 families: three missense variants, four nonsense variants, one frameshift variant and one splice site variant. Significant polyhydramnios occurred in 6/7 pregnancies. Severe respiratory compromise was seen in 6/7 (tracheostomy in 5). A complex neurological phenotype was seen in all patients who had marked brain hypomyelination/demyelination and profound developmental delay. Additional neurological findings included cranial nerve compromise: orobulbar dysfunction in 5/7, facial nerve weakness in 4/7 and vocal cord paresis in 5/7. Dystonia occurred in 2/7 patients and limb contractures in 5/7. All had severe gastroesophageal reflux, and a gastrostomy was required in 5/7. In contrast to most previous reports, only one patient died in the first year of life. Protein modelling was performed for all detected CNTNAP1 variants. We propose a genotype-phenotype correlation, whereby hypomorphic missense variants partially ameliorate the phenotype, prolonging survival. This study suggests that biallelic variants in CNTNAP1 cause a distinct recognisable syndrome, which is not caused by other genes associated with CHN. Neonates presenting with this phenotype will benefit from early genetic definition to inform clinical management and enable essential genetic counselling for their families.

Fig. 1

a Normal nerve, transverse section, H&E 40 × ; (b) Normal nerve, transverse section, Sol. Cyan 40 × ; (c) Normal nerve, longitudinal section, H&E 40 × ; (d) normal nerve, longitudinal section, Sol. Cyan 40 × . All normal nerve sections highlight a normal population of myelinated nerve fibres of appropriate thickness; (e, f) patient 1: Sural Nerve biopsy (TS) at 9 years of age showing marked loss of large diameter (thickly myelinated) nerve fibres, thinly myelinated fibres for axon size, and clusters of thinly myelinated fibres (axonal sprouting) on H&E (e) and Sol. Cyan (f) in keeping with hypomyelinating neuropathy; (g, h) patient 2: Sural Nerve biopsy (TS) at 10 months of age showing modest loss of large diameter (thickly myelinated) fibres, presence of thinly myelinated fibres, clusters of thinly myelinated fibres and a rare onion bulb on H&E (g) and Sol.Cyan (h) consistent with hypomyelination; (i, j) Patient 7: Sural Nerve biopsy (LS) following death at 3 months of age (autopsy specimen). Demonstrates widespread, almost complete loss of myelinated fibres on H&E (i) and Sol. Cyan (j) with a few residual thinly myelinated fibres. k Patient 2: Electron microscopy (EM) highlighting clusters of thinly myelinated fibres with only a few residual appropriately myelinated fibres. Inset highlights single onion bulb in keeping with active demyelination. l. Patient 7: EM highlighting extensive loss of myelinated nerve fibres with a residual rare thinly myelinated fibre. m Electron Microscopy of a normal nerve showing a normal complement of large myelinated fibres and thinly myelinated fibres with proportionate axons

Fig. 2

Facial features of patients 1,2,3,4,6,7 from left to right. Patient 1 is reported with the mildest phenotype, which is apparent from the photo comparisons and makes her appear more atypical. Note the consistent narrow down-slanting palpebral fissures, full rounded eyebrows, myopathic facies and mouth held wide open. Tracheostomy is shown in patients 2,3, 4 and 6

Fig. 3

Sanger sequencing of variants identified in CNTNAP1 in the affected brothers (patients 2 and 3) and their parents. Open symbols: unaffected; filled symbols: affected; square symbols: male and circular symbols: female The nucleotide and amino acid changes are indicated

Fig. 4

Representation of the CASPR protein showing functional domains and approximate location of all reported variants to date. Those marked with an * were identified in this study, others are marked with the paper reference. Variants are described according to NM_003632.2, NG_042091.1

Fig. 5

Comparative modelling of CASPR. a Predicted structure of CASPR residues 21–960. The sequence of CASPR (residues 1–1384) was modeled using the Phyre2 server; a multi-template, high-confidence model (~98% of residues modeled at > 90% confidence) was obtained for a contiguous region spanning residues 21–960, which includes most of the extracellular domain (1–1284) and includes all missense variants discussed here. The protein is shown in ribbon format coloured by secondary structure succession (N-terminal, blue to C-terminal, red); the sidechains of positions of missense variants are shown in stick format and labelled; domain annotation is taken from the InterPro database entry for CASPR (http://www.ebi.ac.uk/interpro/protein/P78357). b–d Modelling of the Laminin G-like 1 domain only (residues 174–355), based on template 3poyA, for wild-type CASPR and variants p.Cys323Arg and p.Leu212Pro respectively; protein is shown in ribbon format, coloured by secondary structure type (red, α-helix; yellow, β-strand; green, loop); the disulphide bond between cysteines 323 and 355 in B is shown by a yellow line; view is rotated compared to Fig. 4a for clarity