Partial loss-of-function variant in neuregulin 1 identified in family with heritable peripheral neuropathy

Abstract

Neuregulin 1 signals are essential for the development and function of Schwann cells, which form the myelin sheath on peripheral axons. Disruption of myelin in the peripheral nervous system can lead to peripheral neuropathy, which is characterized by reduced axonal conduction velocity and sensorimotor deficits. Charcot-Marie-Tooth disease is a group of heritable peripheral neuropathies that may be caused by variants in nearly 100 genes. Despite the evidence that Neuregulin 1 is essential for many aspects of Schwann cell development, previous studies have not reported variants in the neuregulin 1 gene (NRG1) in patients with peripheral neuropathy. We have identified a rare missense variant in NRG1 that is homozygous in a patient with sensory and motor deficits consistent with mixed axonal and de-myelinating peripheral neuropathy. Our in vivo functional studies in zebrafish indicate that the patient variant partially reduces NRG1 function. This study tentatively suggests that variants at the NRG1 locus may cause peripheral neuropathy and that NRG1 should be investigated in families with peripheral neuropathy of unknown cause.

Keywords: NRG1; functional genomics; neuregulin 1; peripheral neuropathy; zebrafish.

Figure 1:. Recessively inherited R551Q variant in the cytoplasmic tail region of NRG1 segregates with Charcot-Marie-Tooth neuropathy in an Australian-Iraqi family.

(A) Segregation of the homozygous NRG1 missense variant [NM_013956.5: c.1652G>A p.(Arg551Gln)] with disease is demonstrated. Limited clinical information is available for individual II:7, her current affection status is unknown, although she reported symptoms in keeping with a CMT neuropathy prior to genetic testing. Double lines indicate consanguinity. Proband indicated with arrow. Note II:3 is a first cousin of the proband. (B) Human NRG1 gene structure. Six “types” of NRG1s are generated through use of alternative transcriptional start sites (arrows), giving each type a unique N-terminal exon (exons and introns to scale, unless indicated with slashes; white boxes, 5’- or 3’-UTRs). We have identified an R551Q missense variant (red line) in the last exon (exon “a”), which encodes the cytoplasmic tail. (C) NRG1 mRNA splicing diagram, illustrating the diversity of transcripts generated through alternative starts sites and alternative splicing. The R551Q variant affects transcripts utilizing the “a” exon, for example NRG1 type III beta1a, a type III isoform that includes the EGF beta exon (red), linker 1, and terminates with the “a” exon. Grey dashed lines indicate alternative splicing paths utilized in RefSeq transcripts in NCBI gene database. Asterisks indicate alternative stop codons. (D) Linear protein diagrams, illustrating the location of the R551Q substitution in isoforms types I-III (transmembrane domains, orange). (E) NRG1 protein isoform topology diagram showing the role of proteolytic processing in generating mature signaling proteins. Type III has an N-terminal TM domain and can remain localized to the membrane after proteolysis; axonal Nrg1 type III is required for myelination of peripheral nerves in vertebrates(Morrissey et al., 1995; Perlin et al., 2011; Taveggia et al., 2005). Subsequent proteolysis can release the C-terminal cytoplasmic tail containing sequences encoded by the “a” exon, in which we have identified the R551Q variant.

Figure 2:. R551Q patient variant has reduced function in peripheral nerve rescue assay.

(A) Peripheral nerve rescue experimental design. A construct expressing human NRG1 type III is microinjected into embryos at the one-cell stage. This construct also expresses GFP in heart tissue (cmlc2:EGFP). After 4.5 days of development, larvae are visually screened for high expression of the rescue construct using the GFP heart signal, and myelin basic protein (mbp) mRNA is detected through in situ hybridization. Larvae are imaged and the length of mbp expression is analyzed, then larvae are genotyped individually. (B1) Control; expression of NRG1 type III in the macrophage lineage (mpeg1) does not affect mbp expression in wild type (wt) embryos. Normal CNS (black arrows) and PNS (white arrowheads) mbp expression extending the full length of the larva (L=100%). (B2) In nrg1 st153mutant embryos (mut), NRG1 expression in the macrophage lineage does not rescue PNS mbp expression (L=0%). (B3) NRG1 expression in neurons fully rescues PNS mbp expression in many injected mutants and (B4) partially rescues myelination in others. (B5) NRG1 type III R551Q expression in neurons only partially or (B6) minimally rescues PNS mbp expression. (C) Table showing the number of larvae in each rescue category out of total mutant larvae injected with the indicated construct. Percentages show the proportion of larvae in each rescue category relative to the total number of larvae in that condition. (D) Summary chart generated from the data in C by characterizing the 100% and 80% categories as “full rescue”, the 0% and 20% categories as “minimal”, and 40% and 60% as “partial”. Whereas both NRG1 type III constructs can rescue myelination, the patient variant is significantly less active than wildtype (Chi-square test, p=0.0277).