Hair cell overexpression of Islet1 reduces age-related and noise-induced hearing loss

Abstract

Isl1 is a LIM-homeodomain transcription factor that is critical in the development and differentiation of multiple tissues. In the mouse inner ear, Isl1 is expressed in the prosensory region of otocyst, in young hair cells and supporting cells, and is no longer expressed in postnatal auditory hair cells. To evaluate how continuous Isl1 expression in postnatal hair cells affects hair cell development and cochlear function, we created a transgenic mouse model in which the Pou4f3 promoter drives Isl1 overexpression specifically in hair cells. Isl1 overexpressing hair cells develop normally, as seen by morphology and cochlear functions (auditory brainstem response and otoacoustic emissions). As the mice aged to 17 months, wild-type (WT) controls showed the progressive threshold elevation and outer hair cell loss characteristic of the age-related hearing loss (ARHL) in the background strain (C57BL/6J). In contrast, the Isl1 transgenic mice showed significantly less threshold elevation with survival of hair cells. Further, the Isl1 overexpression protected the ear from noise-induced hearing loss (NIHL): both ABR threshold shifts and hair cell death were significantly reduced when compared with WT littermates. Our model suggests a common mechanism underlying ARHL and NIHL, and provides evidence that hair cell-specific Isl1 expression can promote hair cell survival and therefore minimize the hearing impairment that normally occurs with aging and/or acoustic overexposure.

Figure 1.

Construction and characterization of Isl1-TG mice. A, A schematic diagram depicting the construction of the Pou4f3 promoter driven Isl1 expression, with an IRES signal that links an EGFP ORF as a separate marker. The total size of the construct is ∼14 kb. Positions of genotyping primers are also shown. B, C, In 20-month-old Is1l-TG mouse cochlea, GFP signal was detected in all OHCs (B) and inner hair cells (B, IHC) and was colocalized with MYO7A-labeled hair cells (C). D, E, In P1 Isl1-TG cochlea, Isl1 transgene expression was detected in all hair cells (D), whereas in the WT control cochlea, endogenous Isl1 was no longer detectable in hair cells (E). F, qRT-PCR of 3-month-old adult showed overexpression of Isl1 in the Isl1-TG compared with WT control inner ear tissues. Expression levels were normalized with control gene Pgk1. Ut, Utricle; Coch, cochlea. Isl1-TG, n = 2; WT, n = 3. Error bars indcate mean (±SEM). Scale bar, 10 μm.

Figure 2.

Normal differentiation of Isl1-TG inner ear. A–J, There is no significant difference in the key protein distribution between Isl1-TG and WT control inner ear and hair cells that include PTPRQ for hair bundles (A, B), acetylated tubulin (A, B, Tub), and neurofilament (G, H, NF) for nerve fibers; prestin for OHC, and PROX1 for supporting cells (C, D); PCNA for proliferating cells and GFP only in transgenic hair cells (E, F), p27KIP1 for supporting cells (G, H), SOX2 for supporting cells, and MYO7A for hair cells (I, J). K, L, FM1-43 uptake showed colocalization between GFP fluorescence and hair bundles in P2 Isl1-TG cochlea (K) and utricle (L). M, A diagram depicting signal intensity of genes in hair cells and supporting cells between Is1l-TG and WT cochleas. Only Isl1 and GFP showed significant increase in the Isl1-TG versus WT control hair cells. *p < 0.05; **p < 0.01, t test. Scale bar, 10 μm.

Figure 3.

Reduced age-related threshold shift. A–H, Age-related threshold shift is reduced in Isl1-TG mice, as measured either by ABRs (A, C, E, G) or DPOAEs (B, D, F, H). Thresholds were compared with 3-month-old WT and Isl1-TG mice, which had indistinguishable ABR and DPOAE at this age. Data are group means (±SEM). Group sizes were as follows: 3 months: n = 4 for control, n = 10 for Isl1-TG; 6 months: n = 4 for control, n = 10 for Isl1-TG; for 12 months: n = 4 for control, n = 9 for Isl1-TG; for 17 months: n = 4 for control, n = 11 for Isl1-TG. The up arrows indicate that at the highest SPLs tested some animals showed no response, which will cause an underestimation of threshold shift. For ANOVA analysis: 6 months: ABR: F = 11.31, **p = 0.0012; DPOAE: F = 0.560; p = 0.457; 12 months: ABR: F = 7.636, *p = 0.018; DPOAE: F = 1.773, p = 0.210; 17 months: ABR: F = 11.184; **p = 0.006; DPOAE: F = 11.554, **p = 0.005.

Figure 4.

Survival of hair cells and preservation of presynaptic ribbons in aged Isl1-TG cochleas. A, B, Confocal images from the 32 kHz region in 19-month-old cochleas from an Isl1-TG and a WT. Hair cells are immunostained with antibodies to myosin VIIA (cyan); ribbons are stained with antibodies to CtBP2 (red). Scale bar, 10 μm. C, H&E staining of a 17-month-old WT cochlea section to show OHC loss. D, H&E staining of a 17-month-old Isl1-TG cochlea section to show preservation of OHCs and the overall structure. E, F, H&E staining of 15-month-old WT and Isl1-TG mid-turn sections to show similar ganglion survival. G, Hair cell counts from 19-month-old animals of both genotypes show significant enhancement of survival in the basal half of the cochlea in the Isl1-TG. Data are group means (±SEM). Group sizes were as follows: n = 8 for TG; n = 7 for WT. H, At 19 months, the number of synaptic ribbons per hair cell was significantly reduced in the basal half of the cochlea in WT mice but preserved in the Isl1-TGs. I, J, At 3 months of age, the numbers of hair cells or ribbons were not significantly different in WT versus Isl1-TG mice. K, By genomic PCR, a fragment of 194 bp for the Cdh23^ahl allele was amplified. Upon PstI digestion, both Isl1-TG and WT showed a 170 bp fragment corresponding to the mutant Cdh23^ahl allele, whereas the 194 bp fragment remained in CBA/CaJ. n = 4 in each comparison. ANOVA analysis: G, F = 25.07, ***p < 0.001; H, F = 19.23, **p < 0.01.

Figure 5.

Noise-induced threshold shift is reduced in Isl1-TG mice compared with WT 2 weeks after noise exposure, as measured by either ABRs (A) or DPOAEs (B). Data are group means (±SEM). Group sizes were as follows: n = 19 for control, n = 30 for Isl1-TG. The up arrow indicates a lack of response in some animals at the highest SPLs tested, which could lead to an underestimated threshold shift. ANOVA analysis: ABR: F = 4.718, *p = 0.035; DPOAE: F = 1.205, p = 0.278.

Figure 6.

Hair cell survival in Isl1-TG after intense noise exposure. A, B, One month after 100 dB exposure, in both WT and Isl1-TG mice, most OHCs survived with a few missing OHCs. The number of ribbon counts was also similar (CtBP2 labeling). C, D, 2 months after 116 dB exposure, WT mice exhibited major OHC loss, whereas in Isl1-TG mice most OHCs survived. E, Twenty-one hours after 116 dB exposure, many WT OHCs were labeled with cleaved Casp9 (arrow), and many condensed nuclei (arrowheads) were seen in the basal half of the cochlea. F, In Isl1-TG cochlea from the similar region, no Casp9-positive hair cells were detected. Scale bar, 10 μm.

Figure 7.

Hair cell survival is enhanced after 116 dB noise in Isl1-TG mice. A, B, No significant difference in hair cell survival or ribbon counts was seen between Isl1-TG and WT mice 1 month after 100 dB exposure. C, After the 116 dB exposure, the OHC loss was significantly larger in some regions in WTs than in Isl1-TGs. ANOVA, F = 55.20, ***p < 0.001; n = 5 for TG; n = 6 for WT. D, There were no differences in the synaptic ribbon counts between Isl1-TG and WT.