Biallelic Loss of Proprioception-Related PIEZO2 Causes Muscular Atrophy with Perinatal Respiratory Distress, Arthrogryposis, and Scoliosis

Abstract

We report ten individuals of four independent consanguineous families from Turkey, India, Libya, and Pakistan with a variable clinical phenotype that comprises arthrogryposis, spontaneously resolving respiratory insufficiency at birth, muscular atrophy predominantly of the distal lower limbs, scoliosis, and mild distal sensory involvement. Using whole-exome sequencing, SNPchip-based linkage analysis, DNA microarray, and Sanger sequencing, we identified three independent homozygous frameshift mutations and a homozygous deletion of two exons in PIEZO2 that segregated in all affected individuals of the respective family. The mutations are localized in the N-terminal and central region of the gene, leading to nonsense-mediated transcript decay and consequently to lack of PIEZO2 protein. In contrast, heterozygous gain-of-function missense mutations, mainly localized at the C terminus, cause dominant distal arthrogryposis 3 (DA3), distal arthrogryposis 5 (DA5), or Marden-Walker syndrome (MWKS), which encompass contractures of hands and feet, scoliosis, ophthalmoplegia, and ptosis. PIEZO2 encodes a mechanosensitive ion channel that plays a major role in light-touch mechanosensation and has recently been identified as the principal mechanotransduction channel for proprioception. Mice ubiquitously depleted of PIEZO2 are postnatally lethal. However, individuals lacking PIEZO2 develop a not life-threatening, slowly progressive disorder, which is likely due to loss of PIEZO2 protein in afferent neurons leading to disturbed proprioception causing aberrant muscle development and function. Here we report a recessively inherited PIEZO2-related disease and demonstrate that depending on the type of mutation and the mode of inheritance, PIEZO2 causes clinically distinguishable phenotypes.

Copyright © 2016 American Society of Human Genetics. All rights reserved.

Figures

Figure 1

Identification of Homozygous Frameshift Variants in PIEZO2 in Ten Affected Individuals of Four Independent Families (A–D) Pedigrees and Sanger sequencing results of four independent families with homozygous frameshift mutations (A–C) or homozygous exon deletions (D) in PIEZO2. DNA was not available for individuals with gray outline. The unequivocal individual identifier is assembled of family, generation, and individual (e.g., A-III.1). Deletion of a genomic region comprising the PIEZO2 exons 6 and 7 in individual D-II.2 (D) was confirmed by PCR analysis and by identification of the breakpoint region. PIEZO2 exons 5, 6, 7, and 8 were amplified in multiplex with RYR1 exon 93 as PCR-positive control. Breakpoint regions were identified by Sanger sequencing of a long-range PCR product amplified from the genomic region between exon 5 and exon 8. (E) PIEZO2 protein structure and mutations identified: homozygous loss-of-function mutations (black) and dominant gain-of-function mutations for DA3 (blue), DA5 (green), and MWKS (red) are given. Protein isoforms refer to UniProt entry Q9H5I5. Bright regions indicate putative transmembrane domains (approximated scale); blue regions indicate missing or inserted regions in different isoforms. The bottom line shows exons in which homozygous frameshift mutations were identified.

Figure 2

Phenotypic Characteristics of Affected Individuals with Homozygous LoF Variants in PIEZO2 Photographs of individuals with homozygous frameshift variants/exon deletions in PIEZO2. (A) Frontal photographs. Facial muscle weakness was present in some of the examined affected individuals. Note weak facial expression/hypomimia in individual B-III.4 and C-II.2. (B) Foot deformities were present in all examined affected individuals including congenital bilateral talipes equinovarus, pes planus, sandal gap deformity, and other valgus and varus deformities. (C) Common hand abnormalities included arachnodactyly and camptodactyly. Note the characteristic hyperextension of the proximal interphalangeal joint with flexion of the metacarpophalangeal joint of the thumb (duck bill deformity) that was present in the majority of case subjects. (D) Spinal skeletal deformities were present in all affected individuals and have been in part surgically corrected. Affected individuals have a short stature (e.g., A-III.5).

Figure 3

Homozygous Frameshift Variants in PIEZO2 Cause Nonsense-Mediated Decay of PIEZO2 Transcripts Determination of PIEZO2 transcript levels by semiquantitative PCR. RNA was isolated from control and affected individual-derived fibroblasts that were either treated or non-treated with an NMD-inhibiting agent. 300 ng of RNA was reversely transcribed into cDNA. (A) RT-PCR of SRSF6 and PIEZO2. SRSF6 (PCR product size: 253 bp) is a physiological target of NMD used as control. A region spanning exons 51 and 52 of PIEZO2 (28 cycles; PCR product size: 306 bp) was amplified. Both cycloheximide (C) and emetine (E) treatment efficiently inhibited NMD. HPRT1 was amplified as loading control (product size: 168 bp). (B) Quantification of PIEZO2 transcripts. Bars show the mean and standard errors of triplicate NMD inhibition experiments. Asterisks indicate statistical significance (p < 0.05). (C) Oligonucleotides in PIEZO2 exons 12 and 14 were used to coordinately amplify transcripts containing (FL) or lacking (Δ13) exon 13 (38 cycles; FL-PCR product size: 430 bp; Δ13-PCR product size: 199 bp). (D) Δ13 PCR products were Sanger sequenced demonstrating skipping of PIEZO2 exon 13 in presence of the variant c.1550_1552delGCTinsCGAA. Abbreviations are as follows: C, cycloheximide; E, emetine; hom, homozygous; het, heterozygous; FL, full-length; Δ13, delta exon 13 (exon skipping).