Deafness in LIMP2-deficient mice due to early loss of the potassium channel KCNQ1/KCNE1 in marginal cells of the stria vascularis

Abstract

Our previous studies revealed a critical role of the lysosomal membrane protein LIMP2 in the regulation of membrane transport processes in the endocytic pathway. Here we show that LIMP2-deficient mice display a progressive high-frequency hearing loss and decreased otoacoustic emissions as early as 4 weeks of age. In temporal overlap to hearing impairment, fluorescence immunohistochemical studies revealed that the potassium channel KCNQ1 and its beta-subunit KCNE1 were almost completely lost in the luminal part of marginal cells in the stria vascularis, affecting first higher and later also lower frequency processing cochlear turns. Concomitant with this, the expression of megalin, a multiligand endocytic receptor, was reduced in luminal surfaces of marginal cells within the stria vascularis. KCNQ1/KCNE1 and megalin were also lost in the dark cells of the vestibular system. Although LIMP2 is normally expressed in all cells of the stria vascularis, in the organ of Corti and cochlear neurons, the lack of LIMP2 preferentially caused a loss of KCNQ1/KCNE1 and megalin, and structural changes were only seen months later, indicating that these proteins are highly sensitive to disturbances in the lysosomal pathway. The spatio-temporal correlation of the loss of KCNQ1/KCNE1 surface expression and loss of hearing thresholds supports the notion that the decline of functional KCNQ1/KCNE1 is likely to be the primary cause of the hearing loss. Our findings suggest an important role for LIMP2 in the control of the localization and the level of apically expressed membrane proteins such as KCNQ1, KCNE1 and megalin in the stria vascularis.

Figure 1. Hearing thresholds in LIMP2-deficient mice

ABR (A and B) and DPOAE (C) thresholds for wild-type (LIMP2^+/+, 2.5 months) and knockout mice (LIMP2^−/−, 2.5 months). Vertical bars give the standard deviations of the means. A, ABR thresholds on click stimulation, mean for wild-type and knockout mice (n = 12). The threshold loss of 15.7 dB is statistically significant (t test, P < 0.0001). B, ABR thresholds as a function of frequency for wild-type (LIMP2^+/+, circles, n = 5) and knockout mice (LIMP2^−/−, squares, n = 6) of 2.5 months of age. Knockout animals exhibit a statistically significant hearing loss of up to 46 dB at frequencies above 11.3 kHz. C, DPOAE thresholds (dB SPL) for wild-type (circles, n = 8) and LIMP2^−/− mice (squares, n = 11), legend as in B. Again, a statistically significant threshold loss of up to 29 dB for knockout animals at high frequencies became evident (>8 kHz).

Figure 2. Early hearing impairment in LIMP2-deficient mice

Click-ABR thresholds (A and B), frequency-specific ABR threshold (C) and DPOAE threshold (D) of LIMP2^+/+ control mice and LIMP2^−/− knockout mice of different ages. Mean (A, C and D) and individual data (B). A, knockout mice of 1 month of age (LIMP2^−/−, 1 mth) split into 2 groups of animals with hearing loss (‘bad’, dark grey) and animals without hearing loss (‘good’, light grey). Hearing of ‘good’ LIMP2^−/− was not different from wild-type controls (LIMP2^+/+, 2.5 mths, green), and ‘bad’ LIMP2^−/− was not statistically significantly different from LIMP2^−/− of 2.5 months of age (red). Data for controls and LIMP2^−/− aged 2.5 months were replotted from Fig. 1A. B, individual ABR thresholds on click stimulation illustrating the two groups of LIMP2^−/− animals with an early (‘young’, grey) and a late (‘aged’, red) hearing loss (diamonds, encircled data points). At 2.5 months of age, all LIMP2^−/− animals had developed a significant hearing loss. C, mean ABR thresholds in response to frequency-specific stimuli for LIMP2^+/+ and LIMP2^−/−. Data for 2.5-month-old controls (green) and knockout animals (red) were replotted from Fig. 1B (continuous lines). Thresholds of early deficient 1-month-old LIMP2^−/− (dark grey diamonds) follow exactly the threshold curve of 2.5-month-old LIMP2^−/− animals. Thresholds of ‘good’ hearing knockout animals (bright grey triangles) follow mostly the thresholds of the control animals but reveal the onset of a high frequency hearing loss at frequencies above 22 kHz. D, frequency-dependent threshold loss is also reflected in the DPOAE thresholds (details as in C), indicating the loss of outer hair cell function as a first reference for the progressive hearing loss.

Figure 3. Atrophy of the stria vascularis in LIMP2^−/− mice

Light microscopy of stria vascularis of wild-type (LIMP2^+/+ mice, panels A, C, E and G) and LIMP2^−/− mice (B, D, F and H) at different ages as indicated. In wild-type animals of all ages investigated the most conspicuous feature was the deep folds of the marginal cells (arrows) spanning most of the depth of the stria vascularis. In the LIMP2-deficient mice the histological organization of the stria vascularis became gradually destroyed with age and was fully lost at the age of 14 months (H). The first irregularities in the LIMP2-deficient stria vascularis were seen at the age of 2 months (D). B, I, M denote nuclei of basal, intermediate and marginal cells, respectively. F, nucleus of a type I-fibrocyte. The arrowheads point to accumulations of extracellular matrix around capillary profiles as well as to deposits without visible relation to capillaries (F). Scale bars represent 15 μm for all micrographs.

Figure 4. Immunohistochemical assessment of the stria vascularis

A, LIMP2 (green) was expressed in marginal, intermediate and basal cells of the stria vascularis in 14-day-old wild-type mice (arrows point to some immunopositive cells). B, LIMP2 expression (red) in 3-month-old wild-type mice (left) and its complete absence in LIMP2-deficient mice (right). C and D, expression of KCNJ10/K_ir4.1 (red) in the intermediate cells of 3-month-old wild-type (+/+) and LIMP2^−/− knockout mice; note the severe shrinkage of the stria vascularis in the LIMP2-deficient mice despite persistence of expression of KCNJ10/K_ir4.1 (arrows point to examples of stained cells). E and F, in contrast, K_v1.1 (red), expressed in basal and probably intermediate cells (arrows point to some stained cells, E), is strongly reduced in same aged LIMP2^−/− knockout mice (F). The inset in F indicates some positive K_v1.1 staining (red) in the stria vascularis of the apical cochlear turn in 3-month-old LIMP2^−/−. Dotted line marks the approximate width of the stria vascularis. Nuclei are stained with DAPI (blue). Scale bar in A, 10 μm; in B, 50 μm; in C–F, 20 μm.

Figure 5. KCNQ1/KCNE1 expression in the stria vascularis

A, KCNQ1 (red) was expressed in the luminal part of the marginal cells of the stria vascularis, shown for the midbasal cochlear turns in 3-month-old wild-type mice. B, with the exception of a small area, the surface expression was absent in LIMP2-deficient mice. The inset in B indicates punctate KCNQ1 labelling (red) in the stria vascularis of the apical cochlear turn in 3-month-old LIMP2-deficient mice. Arrows point to examples of stained cells. C and D, KCNE1 (red) in the apical part of marginal cells of the stria vascularis of 3-month-old wild-type (+/+) and LIMP2^−/− knockout mice. In LIMP2^−/− knockout mice KCNE1 persisted in small membrane fragments. Arrows point to some immunopositive marginal cells. Nuclei were stained with DAPI (blue). All scale bars represent 20 μm.

Figure 6. No alteration of K_v1.1 but decrease of KCNQ1 expression in the stria vascularis coincident with hearing loss

Immunofluorescence of 1-month-old LIMP2-deficient mice. A and B, no alteration of the expression of K_v1.1 (red) in the stria vascularis was observed either in ‘good’-hearing LIMP2-deficient mice or in ‘bad’-hearing LIMP2-deficient mice. C and D, KCNQ1 (red) expression was observed in the marginal cells of ‘good’-hearing mice but not in ‘bad’-hearing mice. Note small patches of persisting surface expression (D, arrow). The inset in D indicates a larger immunopositve patch of KCNQ1 staining (red, the arrows point to examples of stained cell areas) in the stria vascularis of the apical cochlear turn in a 1-month-old “bad” hearing LIMP2^−/− mouse. Dotted line marks the approximate width of the stria vascularis. Nuclei are stained with DAPI (blue). All bars represent 20 μm.

Figure 7. Decrease of megalin expression in the stria vascularis coincident with hearing loss

A and B, megalin expression (green) in 3-month-old wild-type mice (+/+) and its complete absence in LIMP2-deficient mice (−/−). C and D, megalin expression was observed in the marginal cells of 1-month-old ‘good’-hearing LIMP2-deficient mice but not in same age ‘bad’-hearing LIMP2-deficient mice. The inset in D indicates a larger immunopositive patch of megalin staining (green) in the stria vascularis of the apical cochlear turn in 1-month-old “bad”-hearing LIMP2-deficient mice. Arrows point to examples of stained cell areas. The dotted line marks the approximate widths of the stria vascularis. Nuclei are stained with DAPI (blue). All bars represent 20 μm.

Figure 8. Decrease of KCNQ1 and megalin expression in dark cells of the vestibular system

A and B, strong expression of KCNQ1 (open arrow points to a labelled cell) was noted in the luminal surface of dark cells (closed arrow) neighbouring the epithelia of the crista ampullaris of ‘good’-hearing mice (A), while the expression was lost in dark cells of ‘bad’-hearing mice (B). C and D, megalin expression was also typically noted in dark cells (C). Megalin was observed in the luminal and abluminal surface of the cells (open arrows point to some stained cells). Megalin was completely lost from dark cells in ‘bad’-hearing LIMP2^−/− mice, coincident with the loss of KCNQ1. Dotted line delineates the region of the dark cells. Nuclei are stained with DAPI in blue. All bars represent 20 μm.