P2X4 expression in the cochlea
After testing the dilution range 1:50 to 1:2000, 1:1000 dilution was chosen to have the best signal to background ratio (data not shown) for the anti-P2X4 antibody (Figure 1). High levels of expression of P2X4 in the rat OoC were observed (Figure 1D, F), but less in the spiral ligament (Figure 1C, arrow) and spiral ganglion (Figure 1E, arrow) In the OoC, there was a relatively higher expression in IHC & and to a lesser extent in OHCs (Figure 1D, F-H). Immunolabeling of P2X4 throughout was abolished by preabsorbing P2X4 antibody with excess peptide molecules in the controls (Figure 1I). The expression of P2X4 was evident in the hair cells from E20.5 but was more prominent in the IHCs than OHCs. When compared the OoC at different rat ages,(E20.5, P4, P8, P21); Figure C-F), immature IHC at E20.5 expressed P2X4 above the background, but the signal was relatively weak. At P4 and P8, expression of P2X4 was clearly evident in IHCs and OHCs. By P21, a week after the hearing onset, P2X4 exhibited a similar expression pattern to the adult cochlea with strong expression of P2X4 in the IHCs. At P8, P2X4 expression was prominent in IHC & OHC (Figure 2A-B). Some cells lining the cochlear scala tympani and vestibuli also expressed detectable levels of P2X4 (Figure 2A-B, arrows). It is not possible to identify these cells but they appear morphologically very similar to Iba1-expressing macrophages observed in the postnatal mouse cochlea (Kishimoto, Okano, Nishimura, Motohashi, & Omori, 2019). To confirm the identity of cells expressing P2X4, two cell-type specific markers were used: myosin VIIa, which is consistently expressed in IHC and OHCs (Jung et al., 2019; Xiong et al., 2019) and Sox2, which is a transcription factor expressed in nuclei of all types of supporting cells (Smeti et al., 2011). P2X4-labelled cells co-expressed myosin VIIa, confirming these to be the IHCs and OHCs (Figure 2G), while P2X4 was not observed in cells expressing Sox2 (Figure 2H).
Polarity of P2X4 subcellular distribution within IHC and OHC in adult rat cochlea
We next investigated the subcellular localizations of IHCs and OHCs in the adult rat cochlea to correlate the distribution of P2X4 with the distinct functional domains of IHC and OHC. Analysis of z-stack images of OoC enabled compartmentalisation of the hair cells into four different sub-domains from the apical surface to the basal pole of the cell; sub-cuticular (Figure 3A), cytoplasmic (Figure 3B), nuclear (Figure 3C) and subnuclear zones (Figure 3D). At the sub-cuticular plate level of IHC, P2X4 immunolabeling appeared as bright irregular clusters in the cell cytoplasm immediately underneath the cuticular plate (Figure 3A). In the supranuclear cytoplasm, between the cuticular plate and nucleus, there were similar clusters of P2X4 immunolabelling, but these appeared larger and brighter (Figure 3B). At the nucleus and sub-nucleus levels, the cytoplasmic immunolabeling for P2X4 appeared brightest and the most abundant (Figure 3C, D). This pattern is also evident in the 3D re-constructions (Figure 3E). Orthogonal views of the images were also generated with ImageJ where the stack of images was viewed in a XY, YZ, XZ planes, where the X, Y, Z planes correspond to the left and right (Le-R), medial and lateral (M-L) and the apical and basal (A-B), respectively (Figure 3F). Orthogonal visualization confirmed more intense P2X4 immunolabeling along the medial side of the IHCs (Figure 3F’’, arrow). This corresponds to the large, patchy signal appearance in the 3D re-construction (Figure 3E). More intense signal was also observed at the apical part of the image (Figure 3F’’’). Signal distribution for P2X4 along the apical-basal and medial-lateral axes of the cells was quantified using ImageJ (Supplementary Figure 4) to confirm these visual observations that the P2X4 expression was more concentrated at the basal end of the IHCs (Figure 3G) and at the medial side of the IHCs (Figure 3H).
Similar analyses in the OHCs (Figure 4) showed P2X4 immunolabeling at all four levels predominately in the cell cytoplasm, however the characteristic pattern of P2X4 localisation was quite different from that observed in IHCs. At the sub-cuticular plate level, the P2X4 expression appeared to be more concentrated than observed in IHC (Figure 4A). Interestingly, the cluster of P2X4 labelling often appeared immediately underneath the ‘cuticular-free zone’, a small region on the lateral aspect of the cell that does not stain with phalloidin (Figure 4 arrow). At the cytoplasmic level, regions of P2X4 appeared more scattered, but some medium-sized clusters were observed (Figure 4B). At the nucleus level, the P2X4 immunoreactivity was less obvious (Figure 4C), but more intense in the basal sub-nucleus level of OHC (Figure 4D). When reconstructed in 3D, a prominent cluster of P2X4 immunolabeling was observed at the apical part of the cytoplasm, and it was not as homogenously distributed through the whole cell compared to P2X4 immunolabeling in IHC (Figure 4E). Examined using the orthogonal view, the most intense signal for P2X4 (Figure 4F’’, asterisks), appearing as a prominent cluster, was observed at the lateral side of each cell underneath the CP free zone (Figure 4E). Z-stack images obtained for the OHCs were quantified using ImageJ (See supplementary figure 5 for details) to confirm that the P2X4 expression was more concentrated at both the apical and the basal end of the OHCs (Figure 4G). The gradient in OHC was very subtle in the medial to lateral direction (Figure 4L-M), compared to the clear trend observed for IHC (Figure 3H). Immunolabeling for of P2X4 appeared as more discrete dots in the OHCs compared to the IHCs, allowing additional “particle analysis” (Figure 5A & B; see Supplementary Figure 3 for more detail). The total area occupied by P2X4 immunolabeling was the greatest in the sub-cuticular zone, compared to the three other zones (Figure 5C). In contrast, the total amount of staining was the highest in the cytoplasmic zone compared to the other three zones (Figure 5D).
Localization of P2X4 to subcellular organelles within IHCs and OHCs.
The P2X4 immunostaining was in clusters and appeared to be vesiculated. To determine if these were associated with other membranous intracellular organelles, we looked at co-localisation of P2X4 with endosomes, lysosomes, Golgi bodies and mitochondria using immunohistochemistry (Table 1). Early endosomes are derived from the plasma membrane (Gindhart & Weber, 2009) and distinguished from late endosomes and other vesicles by the expression of early endosome antigen 1 (EEA-1) (Patki et al., 1997), including in IHCs and OHCs (Schug et al., 2006). Endosomes and the Golgi apparatus are part of the intracellular protein transportation and recycling pathway. EEA-1 labelling in IHCs had a diffuse appearance, with vesicular labelling more concentrated in the apical part of the cell (Figure 6A). In OHCs, EEA-1 labelled vesicles appeared throughout (Figure 6B). There was some co-labelling between EEA-1 and P2X4 (Figure 6A, B arrow) in both IHC and OHC, with qualitatively more co-occurrence observed in OHC. To quantify the co-localization of EEA-1 with P2X4, the JACoP plugin (Bolte & Cordelières, 2006) in ImageJ was used (see Supplementary Figure 3 for details). Z-stack images covering either entire OHCs or IHCs were selected for analysis. JACoP quantifies the co-occurrence of P2X4 and EEA-1 as two “Mander’s coefficients” calculated as M1 and M2 coefficients with a value range between 0-1.0. M1 represents the proportion of EEA-1 co-localized with P2X4 signal over the total P2X4 signal. M2 represents the proportion of the EEA-1 co-localized with P2X4 over the total signal of EEA-1. The average M1 values for each organelle marker in IHC and OHC are summarised in Figure 6J & K.
Taking the same approach, we analysed the co-occurrence of P2X4 with LAMP-1, GM130, Tom20 and Wheat Germ Agglutinin (WGA). LAMP-1 is a protein found on lysosomes and lysosome-endosome fusion vesicles and is commonly used as a marker for lysosomes (Huotari & Helenius, 2011). Lysosomes are distributed throughout the cell in the IHCs and OHCs, but large lysosomes are often found at the apical, lateral side of the cell (Spicer, Thomopoulos, & Schulte, 1999). OHCs have a greater number of lysosomes compared to IHCs (Spicer, Thomopoulos, & Schulte, 1998; Wiwatpanit et al., 2018). LAMP-1 labelling in IHCs had a more diffuse appearance with lower signal levels, and minimally co-occurred with P2X4 (Figure 6C) where the OHCs had a vesicular appearance (Figure 6D). There was a clear overlap of the P2X4 immunolabelling and LAMP-1 in OHC (Figure 6D, Table 2).
GM130 is a marker for Golgi matrix protein of 130kDa, which typically targets the cis-component of Golgi (Nakamura, Lowe, Levine, Rabouille, & Warren, 1997). The Golgi apparatus is located mainly around the apical part of the cytoplasm in HCs (Schug et al., 2006; Spicer et al., 1998, 1999). In the rat cochlea, cytoplasmic expression of GM130 was observed in the IHCs and OHCs with vesicular, string-like structures (Figure 6E, F), consistent with previous reports (Schug et al., 2006). Notably, the co-occurrence of the GM130 and P2X4 in both the IHCs and OHCs was minimal (Figure 6E, F), 11.7% ± 2.4%, and in the OHCs, 27% ± 2%. TOM20 is a protein expressed on the mitochondrial outer membrane (Balaker, Ishiyama, Lopez, Ishiyama, & Ishiyama, 2013) and was used here as the marker for mitochondria. There was some overlap of TOM20 and P2X4 signal in the IHCs (Figure 6G, arrow). However, there was little co-localization between P2X4 and TOM20 in both OHCs and IHC (Figure 6H). TOM20 was co-occurred with P2X4 in the IHCs 13.7% ± 2.5% and in OHCs 11.3% ± 1.2%. Finally, WGA is naturally occurring molecule known to bind to glycoproteins found in the cell membrane, and fluorescent conjugates are commonly used as a marker for cell membrane (Emde, Heinen, Gödecke, & Bottermann, 2014). The WGA labelled the OHC membrane but did not stain IHC, similar to a previous study (Gil-Loyzaga & Brownell, 1988). Therefore, the association with the IHC membrane was inconclusive and therefore not included in this study. We observed the minimal overlap between WGA and P2X4 in the OHCs(Figure 6I). WGA was co-localized with P2X4 in OHCs 4.3% ± 0.2% (Figure 6K).
In summary, in IHCs, EEA-1 and GM130 have the highest percentage of co-localization with P2X4, at 26% and 27%, respectively, compared to other organelle markers, suggesting cytoplasmic P2X4 were likely associated with endosomes and Golgi apparatus. The co-localization pattern in OHCs was slightly different from that with IHCs; EEA-1 and LAMP-1 have a higher percentage of co-localization with P2X4 at 42.3% and 32.%, respectively. This suggests that P2X4 associate with endosomes and lysosomes in OHCs (Figure 6 K&L, Table 2).