The important regulatory role of UTRs in the pathogenicity of human disease has already been outlined [15, 16]. Reporting variants in these non-coding regions is increasingly recommended [17], based on a better accessibility by GS. UTRs are non-coding segments flanking the protein-coding sequence of a gene. Although transcribed into mRNA, they are not translated into protein. Instead, UTRs are pivotal for post-transcriptional gene regulation, and alterations within these regions have been shown to exert profound effects on protein expression [16]. Pathogenic variants within 5’UTR have previously been described for other hereditary neuropathies [18]. The here newly described variant (c.−243G > A, p.?) in the 5’UTR of PRX is predicted to possibly result in altered splicing of the non-coding exon 1. It affects a highly conserved (phyloP score 3.516) nucleotide position and is predicted to impact the splice donor site of the non-coding exon 1 of PRX. Its potential consequences at protein level cannot be determined without functional analyses. However, splicing alterations within 5′UTR have been shown to compromise mRNA stability and impair translational efficiency. In combination with the maternally inherited stop variant (c.2145 T > A, p.Cys715Ter) in exon 7 of the PRX gene and specific clinical features, we hypothesize this change to be associated with autosomal recessive CMT4F.

Loss of function has been described as the underlying mechanism in PRX-associated CMT [4, 6], which is also the predominantly described effect of pathogenic variants in 5’UTR [15], further corroborating its possible role here.

The clinical phenotype, characterized by predominant sensory ataxia, with electrophysiological evidence of a severe sensorimotor neuropathy, is consistent with PRX-associated CMT4F [4, 6, 19, 20]. This distinct clinical profile is further reflected in a moderate-to-severe impairment level of the CMT Pediatric Scale, which incorporates measures of sensory ataxia. In contrast, muscle strength was only mildly affected, as demonstrated by near-normal results in the MRC sum score and 6-min walk test. Notably, the sensorimotor trajectory in our patient closely parallels a 5-year-old patient harboring a homozygous deletion in exon 7 of PRX [7].

To date, only a single nerve ultrasound examination in a child with PRX-associated CMT4F has been published, reporting nerve sizes within normal range [7]. However, comparing normative HRNUS in childhood [11], these measurements reported by Cartwright et al. [21] fall within the upper reference range. Indeed, the nerve dimensions of this 5-year-old patient were similar to those of our patient at 4 years 2 months (Table 2 supplement), who exhibited the most pronounced enlargement from 6 years onward (Fig. 2). This age-dependent pattern parallels pediatric patients with CMT4D (publication in submission), an equally early-onset sensorimotor neuropathy with comparable histopathological features (Fig. 2). Further evidence of nerve enlargement in CMT4F can be found in an autopsy case of an adult patient [19] and histological analyses [4, 22].

Longitudinal follow-up of the nerve size pattern, as depicted by high-resolution nerve ultrasound for the cross sectional area (CSA) of the median nerve at the forearm (A), elbow (B) and distal third of the upper arm (C), as well as the vagus nerve (D) and diameter of the cervical roots C5 (E) and C6 (F), in comparison to other CMT4 subtypes, differentiated for age. Consecutive measurements in a single patient are connected by a line. A significant, age-related enlargement pattern is visualized for the CMT4F patient, similar to in-house data of pediatric patients with CMT4D. Age-related normative values are used for comparison5−8. Longitudinal follow-up of muscle echogenicity, as depicted for the anterior tibial (G) and peroneus tertius muscle (H), derived as a mean grey scale value of three measurements with standardized imaging settings. Consecutive measurements in a single patient are connected by a line. For comparison, in-house data from pediatric patients with other CMT4 subtypes are demonstrated

In summary, this report underscores the potential pathogenicity of untranslated regions in PRX-associated CMT4F and broadens the knowledge on clinical and imaging characterization of this rare disease. As major limitations of this study, the genetic findings need to be validated by functional RNA analysis and the imaging results by ultrasound studies in further CMT4F patients (Fig. 3).

Panel A shows the pedigree of the family with the index patient carrying compound heterozygous variants in the PRX gene. Sanger sequencing of the non-coding exon 1 containing the identified variant in the 5′UTR of PRX. The variant position is highlighted in grey. The figure shows exemplary sequencing chromatograms for the index, father, and mother. The exonic region is highlighted in blue, and the intronic region is highlighted in yellow. Both the index and the father carry the variant c.−243G > A in a heterozygous state, whereas the mother is homozygous for the wild-type allele at this position. Panel B shows a schematic representation of the PRX gene and protein structure. White boxes represent non-coding exons, and black boxes indicate coding exons. Latin numerals indicate the exons, whereas Roman numerals indicate the introns. The positions of the two variants identified in the index are indicated. Below, a schematic overview of all previously reported variants in the coding region of the PRX gene and their respective positions within L(ong)-Periaxin are indicated. Most variants are located in the exon 7 of the gene which is the largest exon encoding 90% of L-periaxin which contains 4 major domains: PDZ which is important for the interaction with proteins, peptides, lipidosomes, NLS (nuclear localization signal) which mediates the export of L-periaxin from the nucleus to cytoplasm as well as the repeat and acidic domain which enable L-periaxin to bind to the cytoskeleton of Schwann cells and imply a stabilizing role in the myelin formation31. Loss of the acidic domain has been characterized as the main aetiological factor causing CMT4F. A detailed list of the previously reported pathogenic variants, indicating the nucleotide and protein position, is enclosed in Table 3 of the appendix