RTN2 deficiency results in an autosomal recessive distal motor neuropathy with lower limb spasticity

Abstract

Heterozygous RTN2 variants have been previously identified in a limited cohort of families affected by autosomal dominant spastic paraplegia (SPG12-OMIM:604805) with a variable age of onset. Nevertheless, the definitive validity of SPG12 remains to be confidently confirmed due to the scarcity of supporting evidence. In this study, we identified and validated seven novel or ultra-rare homozygous loss-of-function RTN2 variants in 14 individuals from seven consanguineous families with distal hereditary motor neuropathy (dHMN) using exome, genome and Sanger sequencing coupled with deep-phenotyping. All affected individuals (seven males and seven females, aged 9-50 years) exhibited weakness in the distal upper and lower limbs, lower limb spasticity and hyperreflexia, with onset in the first decade of life. Nerve conduction studies revealed axonal motor neuropathy with neurogenic changes in the electromyography. Despite a slowly progressive disease course, all patients remained ambulatory over a mean disease duration of 19.71 ± 13.70 years. Characterization of Caenorhabditis elegans RTN2 homologous loss-of-function variants demonstrated morphological and behavioural differences compared with the parental strain. Treatment of the mutant with an endoplasmic/sarcoplasmic reticulum Ca2+ reuptake inhibitor (2,5-di-tert-butylhydroquinone) rescued key phenotypic differences, suggesting a potential therapeutic benefit for RTN2-disorder. Despite RTN2 being an endoplasmic reticulum (ER)-resident membrane shaping protein, our analysis of patient fibroblast cells did not find significant alterations in ER structure or the response to ER stress. Our findings delineate a distinct form of autosomal recessive dHMN with pyramidal features associated with RTN2 deficiency. This phenotype shares similarities with SIGMAR1-related dHMN and Silver-like syndromes, providing valuable insights into the clinical spectrum and potential therapeutic strategies for RTN2-related dHMN.

Keywords: dHMN; hereditary spastic paraplegia; neurodegeneration; polyneuropathy.

Figure 1

The illustration depicts family pedigrees and the RTN2 gene, highlighting the biallelic and monoallelic variants identified in this study and previous studies respectively. (A) Pedigrees of 14 individuals from seven consanguineous families with segregating homozygous RTN2 variants are shown. (B) The localization of the presented variants across the gene is displayed. With homozygous loss-of-function variant (black) identified in this study shown above the gene schematic, and previously reported heterozygous variants (grey) shown below the gene.

Figure 2

Images displaying distal limbs and muscle MRI from families with RTN2-related distal hereditary motor neuropathy. Muscle weakness and wasting of the distal lower > upper limbs in affected individuals from Families 1 (A–F and M–R), 2 (G) and 3 (H–L). Axial fat T1-weighted images of the lower limbs in a 17-year-old Iranian boy [A(i and ii)] and his 27-year-old sister from Family 1 [A(iii and iv)]. The thigh muscles of the boy are normal [A(i)]. There is fatty infiltration of his leg muscles with relative sparing of the medial head of gastrocnemius and extensor digitorum longus [A(ii)]. Changes are more severe in his older sister with generalized loss of muscle bulk and fatty infiltration of the hamstrings [A(iii)]. There is more extensive change in her legs, with relative sparing of the medial head of the gastrocnemius, tibialis anterior, extensor digitorum longus and peroneus longus [A(iv)].

Figure 3

Loss-of-function ret-1 (RTN2 homologue) Caenorhabditis elegans phenotyping and drug screen data. (A) Hierarchical clustering of behavioural fingerprints of untreated wild-type (N2) and C. elegans RTN2 orthologue, ret-1, loss-of-function (LoF) strains, alongside treatment with the panel of 15 bioactive compounds. The heatmap shows the entire set of 8289 behavioural features extracted by Tierpsy for untreated worms and worms exposed to 100 μM of each compound for 4 h. The top dendrogram shows the relationship of the individual extracted features within the entire feature set, with the tracking period bar denoting when during image acquisition the feature was extracted: pre-stimulation (pink), during stimulation with blue light (blue) and post-stimulation (green). The left dendrogram shows the phenotypic similarity of the worms and the colour map (top left) represents the normalized z-score of the features. Despite the strains generally clustering separately from each other, treatment of ret-1 LoF mutant with 2,5-di-tert-butylhydroquinone (DTHBQ) caused this strain to cluster alongside N2, suggesting phenotypic rescue of the mutant. (B) Position of N2 and ret-1 LoF worms in phenospace with respect to the top principal components in the behavioural feature set upon drug treatment. Untreated worms are denoted by stars and treated N2 or treated ret-1(syb4955). Error bars represent the standard error of the mean. Treatment of the ret-1 LoF strain with DTHBQ moved the mutant towards the untreated wild-type in phenomic space. (C) Key morphological, postural and locomotive features were significantly different between the untreated ret-1 LoF mutant (black boxes) and untreated N2 (grey boxes). From left to right: ret-1 mutants are longer; have decreased angular velocity of the head; increased angular velocity of the midbody; decreased midbody curvature; and increased midbody speed. White boxes show ret-1(syb4955) exposed to 100 μM DTHBQ for 4 h, rescuing of 3/5 of the key behavioural features. (D) Overall speed of ret-1 LoF mutant and N2 worms after 4 h exposure to DTHBQ. Treatment of N2 with DTHBQ resulted in an attenuation of the photophobic escape response (increased speed) following stimulation of the worms with three 10 s bursts of high-intensity blue light 100 s apart (denoted by blue shaded regions on the line plot), whereas the ret-1 LoF mutant showed resistance to the effects of DTHBQ. The mean peak speed (box plot, right) was calculated as an average of the maximum detected speed across the three independent pulses of blue light. Individual points marked on all box plots are average values from multiple worms (n = 3) in a single well. The different point colours indicate data from independent experimental days. P-vales were calculated using block permutation t-tests (n = 10 000 permutations). Permutations were shuffled randomly within, but not between, the independent days of image acquisition in order to control for day-to-day variation in the experiments and corrected for multiple comparisons using the Benjamini-Hochberg procedure to control the false discovery rate at 5%. P > 0.05 was considered not statistically significant (ns).