Identification of a genetic variant underlying familial cases of recurrent benign paroxysmal positional vertigo

Abstract

Benign paroxysmal positional vertigo (BPPV) is the most common cause of vertigo in humans, yet the molecular etiology is currently unknown. Evidence suggests that genetic factors may play an important role in some cases of idiopathic BPPV, particularly in familial cases, but the responsible genetic variants have not been identified. In this study, we performed whole exome sequencing [including untranslated regions (UTRs)] of 12 families and Sanger sequencing of additional 30 families with recurrent BPPV in Caucasians from the United States (US) Midwest region, to identify the genetic variants responsible for heightened susceptibility to BPPV. Fifty non-BPPV families were included as controls. In silico and experimental analyses of candidate variants show that an insertion variant rs113784532 (frameshift causing truncation) in the neural cadherin gene PCDHGA10 (protocadherin-gamma A10) is an exceedingly strong candidate (p = 1.80x10-4 vs. sample controls; p = 5.85x10-19 vs. ExAC data; p = 4.9x10-3 vs. NHLBI exome data). The mutant protein forms large aggregates in BPPV samples even at young ages, and affected subjects carrying this variant have an earlier onset of the condition than those without [average 44.0±14.0 (n = 16) versus 54.4±16.1 (n = 36) years old, p = 0.054]. In both human and mouse inner ear tissues, PCDHGA10 is expressed in ganglia, hair cells and vestibular transitional epithelia. Fluorescent RNA in situ hybridization using mouse inner ear tissues shows that expression increases with age. In summary, our data show that a variant in the PCDHGA10 gene may be involved in causing or aggravating some familial cases of recurrent idiopathic BPPV.

Fig 1. Pedigrees of the 12 BPPV families selected for whole exome sequencing (WES) including UTR.

Filled arrowheads indicate probands. Circle and square symbols represent female and male individuals, respectively. Symbols with slashes indicate deceased individuals. The BPPV status of deceased individuals is unknown unless the circle or square is shaded or filled. Affected and unaffected family members are denoted in filled and unfilled symbols, respectively. Asterisks indicate the family members selected for WES, and # indicates those carrying the PCDHGA10 variant rs113784532. FB: Familial BPPV.

Fig 2. Tiered selection of variants from whole exome sequencing of 12 BPPV families.

Fig 3. Frameshift variant causes premature stop, truncating the PCDHGA10 short isoform.

A, Verification of the NM_032090.1:c.2476_2477dup in PCDHGA10 by Sanger sequencing after BpmI restriction digestion. B, Protein structures of PCDHGA10 long, short and mutant isoforms. EC, extracellular; TM, transmembrane; Const: Constant domain. C, Fluorescent immunostaining shows that wildtype PCDHGA10 has no large aggregates in the cytosol of non-BPPV samples at young ages (the vestibular ganglia is shown). D, PCDHGA10 with the rs113784532 mutation forms large intracellular aggregates in BPPV samples even at young ages. Arrows indicate aggregates. M30 and M32, male at 30 and 32 years of age, respectively. Scale bar, 25μm.

Fig 4. Fluorescent immunostaining of Pcdhga10 in the murine vestibule at 2 months of age.

Arrows indicate hair cells, arrowheads transitional epithelial cells and hollow arrow vestibular ganglia. Scale bar, 25μm.

Fig 5. Fluorescent in situ hybridization of the long isoform of Pcdhga10 mRNA in mouse vestibular ganglia.

At P0 (A) and 2M (C), the fluorescent signals of the antisense probe of the long isoform are absent (or similar to that in the negative controls at the same ages) (B & D), whereas at 18M (E), the fluorescent signal of the antisense probe is much stronger than the negative control (F). P0, postnatal day 0; 2M and 18M, 2 and 18 months old, respectively. VeG, vestibular ganglia. Scale bar, 25μm. *** and ### denote p<0.001 when the 18M group is compared with P0 and 2M, respectively (n = 4 mice/group).

Fig 6. Fluorescent in situ hybridization of the short isoform of Pcdhga10 mRNA in mouse vestibular ganglia.

Fluorescent signals of the antisense probe of the short isoform increase with age, whereas signals of the negative controls are at background levels. P0, postnatal day 0; 2M and 18M, 2 and 18 months old, respectively. VeG, vestibular ganglia. Scale bar, 25μm. *** denotes p<0.001 when compared with P0, # denotes p<0.05 when the 2M group is compared with 18M (n = 4 mice/group).