Functional Analysis of Inner Ear Gap Junctions

Gap junction channels are formed by a family of connexin proteins. There are at least 3 connexins, i.e., Cx26, 30 and 43, co-expressed in the cochlear supporting cells. Each connexin mutation alone can induce hearing loss. These multiple connexin isoforms can form diverse connexin channel subtypes, including hybrid (homotypic, heterotypic, homomeric, and heteromeric) channels with asymmetric gating, between the cochlear supporting cells, even between the same type of cochlear supporting cells (Zhao, 2000; Zhao and Santos-Sacchi, 2000, see Appendices 1&11).Recently. channel reconstruction studies have demonstrated that some mutants are still able to form functional homotypic channels between the connexin transfectants (Choung et aI., 2002; Thonnissen et aI., 2002; Bruzzone et aI., 2003; Wang et aI., 2003), indicating that these mutants mainly impair hybrid channel subtypes in vivo. Either Cx30 knockout animal models or ablation of Cx26 in the cochlear supporting cells (the epithelial cell gap junction network) demonstrated supporting cells, eventually hair cells, degeneration after the onset of hearing. However, the cochlea has almost normal development in deficient mice (Cohen-Salmon et aI., 2002; Taubner et aI., 2003). The big concern of both reviewers with our prior application was lack of a clearly stated out overarching hypothesis in the grant proposal. We intend to elucidate the relations of connexin function to either cochlear supporting cells or connexin channel subtypes in these cells. Our working hypothesis is that multiple connexin isoforms and their channel subtypes may have cellular distribution in the cochlear supporting cells and selectively permeate to ions, cell signaling molecules, and nutrients. Concerted actions of these diverse gap junctions between multiple supporting cells can complete long-distant unidirectional intracellular transports, such as recycling K' from hair cell extracellular space back to the endolymph and bringing energy nutrients to cells in an opposite direction. Thus, different connexin isoforms and channel subtypes may provide cellular distribution pathways essential to the functioning of the cochlear sensory epithelium. Gap junctional permeability is an essential function for gap junctional coupling. Molecules up to I kDa, including all known second messengers and some endogenous metabolites, can pass through the connexin channels. Most cell signaling molecules and endogenous metabolites, such as IP), ATP, ADP, and Glutamate, are the charged molecules. Their gap junctional permeability can be assessed using biologically inert fluorescent probes. For example, IP3 possesses -2 changes and 417 Da, resembling a fluorescent dye Lucifer yellow (443 Da, -2 charges). Diminishment of gap junctional intercellular communication by IP) can decrease glucose mobilization in liver (Niessen et aI., 2000). SA 2 ( SAlb in the prior version) Fluorescence recovery after photobleaching (FRAP) to measure gap junctional diffusion. The first reviewer had a concern on how the FRAP is measured and the data are processed (Fig. 9). We described the detailed method for FRAP measurement and data processing in the Fig. 9 legend and the section of Method as well. For a FRAP measurement, grouped cells will be pre-loaded with a membrane impermeable fluorescent dye and then fluorescence in one cell will be quenched using a laser beam illumination. The fluorescence of the quenched cell will be recovered by dye diffusion from the neighboring cells through gap junctions. The intensity of recovering fluorescence in the quenched cell will be recorded and measured by time-lapse photography. The recovery fluorescent intensity will be plotted versus recording time to generate a recovery curve, which reflects the gap junctional permeability. The diffusion coefficient can be calculated from the time constant of the recovery curve. Theoretically, any value can be selected as a reference to normalize the FRAP data, since diffusion coefficient is determined by time-constant instead of the absolute values of fluorescent intensity. We will choose each pre-quenching fluorescent intensity before every FRAP measurement as the referent intensity to normalize the data for each FRAP measurement, so, the recovery curves can be directly comparable for each FRAP measurement. We re-plotted Fig. 9 and plotted FRAP measurements in each direction (or at either cell side) separately, normalizing the recovery fluorescent intensities in each FRAP to its own pre-bleaching point. The time constants or diffusion coefficients for dye diffusion through gap junctions in both direction, i.e., Cell A ~ Cell B and Cell B ~ Cell A, are quite different, .=87.95 sand 60.06 s, respectively, and had an almost 30% difference. In prior data processing and plot, all fluorescent intensities in two FRAP measurements were normalized to the pre-bleaching intensity ofthe first measurement . st like the reviewer' dicated the rever intensit in the secon FRAP measurement was only 50% of the first FARP measurement. This indicated that our experiment and FRAP measurement are reliable. The time-constant was the same in both old and ncw data normalizations. Actually. in this case. wc also did the third FRAP measurement, or repeated on measurement of the first measured cell (Cell A). We added this repeated measurement in new Fig. 9C. The repeated data perfectly matched the first measurement. This further implicated the difference in gap junctional diffusion in either direction arising from the difference in direction of gap junctional diffusion rather than the possible damage in gap junctional coupling caused by photobleaching. We also added detailed method for FRAP data processing, experiment concerns, and problem solving in the Method section. lt is true that the fluorescent dye ofSNARF is a pH indicator. However, we did not find any detectable change in fluorescent intensity after quenching a single uncoupled cell in preliminary experiments, indicating that cytoplasmic pH might not change significantly. We will carefully choose experiment parameters to avoid changes in the cell pH (see a section of FRAP measurement in Method). In addition, we will use another pH-insensitive dye, Calcein AM, to do measurement as a control. Other concerns and possible alternative strategies are also described in the method section. SA 3 (old SA2) for connexin hemichannel study: The first reviewer required a stronger case to support our hypothesis. We added new data (three new figures; Figs. 12-14) in the preliminary data section. The new data clearly show that connexin hemichannels in the cochlear supporting cells have the same charge selectivity in their permeability as that which appeared in the intact connexin channels. For example, new Fig. 12 shows that a Hensen cell pair had synametric distribution ofhemichannels with different permeabilities to anionic and cationic dyes in two cells. The permeability of gap junctional channels between the two cells showed similar charge selectivity, impermeable to anionic dye Alexa Fluor 350. Otherwise, the Alexa Fluor dye would have filled the right cell through the gap junctions between the cells. This case also shows that a clear rectification of the dye diffusion through the gap junctions. The new data in Fig. 14C shows that some hemichannels in supporting cells had dye uptake incubating in a regular extracellular solution, which contained 2 mM CaH (Fig. 14C). The cells had an influx of a cationic dye of Propidium iodide but did not permeate to another coincubated anionic dye, Lucifer yellow. This is a good control, indicating that dye filling did not arise from cell membrane damage induced leakage. The new data also implied that the hemichannels in the cochlear supporting cells may be able to play physiological functions under normal physiological conditions. Our new data further suggest that hemichannel study can elucidate connexin channel structure-function and may also lead to find interesting unknown connexin functions in the cochlea. The second reviewer had some concerns on patch clamp recording for sealing resistance measurement in Fig. 15. We added the pre-and post-measurement seal resistances and electrode resistances (access resistances) in the legend of Fig. 15. The seal resistances were above 1 GQ and recording quality was quite good. Actually, in our patch clamp recording, the electrode resistance (or access resistance, Rs), membrane resistance (or seal resistance, Rm), input capacitance (Cin), and other patch clamping parameters were routinely recorded in 2-4 Hz during experiment (Zhao, 2001; Zhao and Santos-Sacchi, 1998,2001, see Appendices I-Ill). We will follow the same procedure in this proposed proposal to monitor R" Rm,and Cin continuously (See Methods). Before and after each measurement, the Rs, Rm,and Cn will be measured to assess the recording quality and the pipette sealing condition. IfRm drops down to 500 MQ, which would show deterioration of pipette sealing, the recording will be stopped and the data will be excluded trom the analysis. We also clearly stated information about the anti-connexin antibody in the Fig. 15 legend and in the method section as well. Almost all commercial available anti-Cx26/30 antibodies are anti-segments of the cytoplasmic loop of connexins. We will put such antibodies into the patch pipette or do intracellular perfusion to load the antibody into the cells. To show how to isolate the hemichannel response, we added a detailed method in the Method section. The cochlear supporting cells have K and Ca and other voltage-dependent ion channels on the non-junctional membrane. The prevalent ionic currents on the non-junctional membrane of the supporting cells are the voltagedependent outward rectifying K currents (Nenov et aI., 1998; Zhao, 2000). We will use ionic blocking solution (IBS, see Method), which contains 20 mM TEA, 20 mM CsCl, and 1.25 mM CoCh, to block voltage-dependent K and Ca channels. This can block ionic currents on the non-junctional membrane of supporting cells very well (See new Figs. 16B and 17A after uncoupling, or Zhao, 2000 and Zhao and Santos-Sacchi, 1998,2000, .. ' does not inactivate, its responses are easy to distinguish from the fast-developing, quickly inactive K and Ca currents. However, we wil1 stil1 use a connexin antibody or uncoupling agent, such as oetano!. to verity the connexin hemichannel responses. For detailed method and experiment concerns see the section of Methods. We added new SA4 on directly recording the ionic passage through gap junctions to replace old SA3. The detailed method and information about how to record and identity the transjunctional current appears in the new SA4 figures (Figs. 16& 17) and text, and in the corresponding section of Methods as wel!. Old SA3 for testing how K- enter into the supporting cel1s has been moved to Future Direction and replaced with new SA4. This makes the proposal focused and limited as reviewers suggested. However. we still like to discuss briefly some concerns of the first reviewer on old SA3. The reviewer proposed that since the transduction current carried by hair cells is in nA range, equally large conductance need to be present in Deiters' cells. Indeed, bending the stalk of a Deiters' cell could evoke inward current in a nA range. For example, , the evoked inward current of a single Deiters' cell was about 0.65 nA in our old Fig. 13 (which has been removed in this application). Considering Deiters' cells are coupled together by gap junctions and work as a group, the sinking capability of grouped Deiters' cells will be very large and enough to sink the hair cell transduction current. It will be the same for the VR-OAC channel, if it is responsible for the evoked inward current. How the K ions enter into the cochlear supporting cells is an interesting and important question. If possible, we would like to address this issue in this proposal. Background and Significance: The reviewers had no concerns on this section. However, we updated the new lectures and studies since the last submission. The new studies have revealed that knockout of Cx30 or targeted ablation of Cx26 in the cochlear sensory epithelial cells can induce elimination of endolymphatic potential (EP) and potassium concentration, and supporting cells, eventual1y hair cel1s, begin to degenerate after the onset of hearing. However, the cochlea of the deficient mice has almost normal development (Cohen-Salmon et aI., 2002; Teubner et aI., 2003). These data suggest that the cochlear epithelial gap junction network, which we propose to investigate in this project, playa crucial role in hearing functions. This also implies that gap junctional intracellular communication may be important for providing energy. The cell degeneration in the connexin deficient mice occurred after the onset of hearing. This may be because the cells need more energy supplies for performing functions. It is also possible that because gap junctional coupling is required for removing endogenous metabolites, ototoxins develop. In any case, the directional1y, selectively junctional transport among the cochlear supporting cells is required for hearing, which we will test in this project. The new studies also show that some deafness-associated Cx26 mutants still can form functional homotypic channels between the connexin transfectants in vitro (Choung et aI., 2002; Thonnissen et aI., 2002; Bruzzone et aI., 2003; Wang et aI., 2003). This indicates that these mutants mainly impair their constituting hybrid channel subtypes, and the hybrid channel subtypes play important roles in hearing function. This also strongly supports our working hypothesis that the hybrid channels may have cel1ular distribution and specific functions in the cochlear supporting cells.