Sensorineural hearing loss (SNHL) is a common sensory deficit, which affects 1 in 500 newborns, and can arise from etiologically diverse structural and functional inner ear abnormalities. The mammalian inner ear is an elegant labyrinth that contains a cochlea, the primary auditory organ, and a vestibular system that maintains body balance. Lateral cochlear duct cells, comprising Reissner’s membrane and the stria vascularis, are critical for production, maintenance and secretion of endolymph, a specialized fluid that supports hair cell function. A recent Developmental Cell paper by Alex Rohacek, a DSRB student from Douglas Epstein’s lab, highlights a complex splicing program that is necessary for proper development of the lateral cochlear duct cells and is, therefore, essential to form a functional hearing organ in mammals.
This research story began at the Children’s Hospital of Philadelphia (CHOP), during an evaluation of an 8-year- old female with congenital hearing loss. While her parents appeared to have normal hearing, her older brother was previously diagnosed with SNHL. The girl’s four other siblings were asymptomatic. This finding and the absence of a family history of hearing loss suggested an autosomal recessive mode of SNHL inheritance. To identify a potential causative mutation, the researchers used whole-exome DNA sequencing, and were able to zoom in on a gene called Epithelial Splicing Regulatory Protein 1 (ESRP1). It became clear that mutations in ESRP1 were segregating with hearing loss in the family
While these human studies performed at CHOP were essential to identify ESRP1 as a novel hearing loss gene, the follow-up experiments done by Alex and the rest of the UPenn team were fundamental to uncovering a functional role for ESRP1 during inner ear development. In particular, the analysis of homozygous knockout ESRP1 mouse embryos shed light onto the developmental progression of inner ear pathology and allowed the scientists to interrogate key alternative splicing events that could underlie SNHL.
As expected from the human findings, ESRP1 null embryos had profound inner ear defects, manifesting in cochlear duct truncation and immature or absent hair cells. To probe altered gene expression events in these mutant embryos, Alex and the team performed RNA-Seq analysis to identify the top misregulated transcripts. Interestingly, many genes previously implicated in hearing loss were differentially expressed between the control and ESRP1 null mice. Genes with the highest fold change included ion channel subunits Bsnd and Kcnq1, and their upstream regulator Nr3b2/Esrrβ, all of which happen to be expressed in a subset of cochlear duct cells called marginal cells.
These transcriptome data were instrumental for identifying both altered transcript levels and specific aberrant splicing events in ESRP1 null mice. Consistent with the role of ESRP1 as a master regulator of epithelial cell-type specific programing, many epithelial-specific isoforms were substituted for their mesenchymal counterparts in ESRP1 null embryos. Splicing of Fibroblast Growth Factor Receptor 2 (FGFR2), for example, involves a tight regulation of mutually exclusive exons IIIb (epithelial isoform) and IIIc (mesenchymal isoform). Alex and the team confirmed the switch from the normal FGFR2-IIIb to mesenchymal FGFR2-IIIc isoform in the cochlear epithelium of mice lacking ESRP1. According to Alex, making sense of alternative splicing data was by far one of the most challenging aspects of this project.
“I’d never really worked with post-transcriptional processing before and very little is known about splicing in the ear. Trying to tie the switches in isoform usage we were finding to the phenotype was really difficult,” he noted.
Analysis of specific nonsensory inner ear cell identities in ESRP1 null mutants revealed a curious cell fate switch phenotype characterized by a reduction of marginal cells and a concomitant increase in Reissner’s membrane cells, an imbalance that can, perhaps, explain the hearing loss phenotype in human patients. It was an in-depth look at the splicing switch in FGFR2 and Fgf signaling that led to surprising discoveries. It was hypothesized that the observed splicing switch in FGFR2 would render it unable to respond to Fgf10 ligand. Since the Fgf10 inner ear mutants had been published, Alex and the team expected to see the same phenotype, a loss of Reissner’s membrane, in their ESRP1 null mice but, instead, they observed the exact opposite - Reissner’s membrane was expanded.
“After a lot of literature searching and thought experiments, we reasoned that the other isoform of FGFR2 was still expressed and responded to a different ligand, Fgf9, effectively giving us a gain of function phenotype. We were able to rescue the expanded Reissner’s phenotype by removing an allele of Fgf9, validating our hypothesis. I’m honestly still amazed at how complex that whole interaction is and how well that experiment worked,” shared Alex.
Alex’s paper is a great example of original research that has been enabled by frequent collaborations between labs at UPenn and its next-door neighbor, CHOP. A combination of patient data with cutting-edge animal research has allowed for a better mechanistic understanding of how mutations in ESRP1 can lead to hearing loss.
Rohacek A, Bebee T, Tilton R, Radens C, McDermott-Roe C, Peart N, Kaur M, Zaykaner M, Cieply B, Musunuru K, Barash Y, Germiller J, Krantz I, Carstens R & Epstein D (2017) ESRP1 Mutations Cause Hearing Loss due to Defects in Alternative Splicing that Disrupt Cochlear Development.Dev Cell 43, 318–331.e5.