Considerable progress has been made in identifying deafness genes, but still little is known about the genetic basis of
normal variation in hearing function. We recently carried out a Genome Wide Association Study (GWAS) of quantitative
hearing traits in southern European populations and found several SNPs with suggestive but none with significant
association. In the current study, we followed up these SNPs to investigate which of them might show a genuine association
with auditory function using alternative approaches. Firstly, we generated a shortlist of 19 genes from the published GWAS
results. Secondly, we carried out immunocytochemistry to examine expression of these 19 genes in the mouse inner ear.
Twelve of them showed distinctive cochlear expression patterns. Four showed expression restricted to sensory hair cells
(Csmd1, Arsg, Slc16a6 and Gabrg3), one only in marginal cells of the stria vascularis (Dclk1) while the others (Ptprd, Grm8,
GlyBP, Evi5, Rimbp2, Ank2, Cdh13) in multiple cochlear cell types. In the third step, we tested these 12 genes for replication
of association in an independent set of samples from the Caucasus and Central Asia. Nine out of them showed nominally
significant association (p,0.05). In particular, 4 were replicated at the same SNP and with the same effect direction while
the remaining 5 showed a significant association in a gene-based test. Finally, to look for genotype-phenotype relationship,
the audiometric profiles of the three genotypes of the most strongly associated gene variants were analyzed. Seven out of
the 9 replicated genes (CDH13, GRM8, ANK2, SLC16A6, ARSG, RIMBP2 and DCLK1) showed an audiometric pattern with
differences between different genotypes further supporting their role in hearing function. These data demonstrate the
usefulness of this multistep approach in providing new insights into the molecular basis of hearing and may suggest new
targets for treatment and prevention of hearing impairment.
