Cumulative Mitochondrial Activity Correlates with Ototoxin Susceptibility in Zebrafish 2 Mechanosensory Hair Cells
Total Page:16
File Type:pdf, Size:1020Kb
1 TITLE: Cumulative mitochondrial activity correlates with ototoxin susceptibility in zebrafish 2 mechanosensory hair cells. 3 4 AUTHORS: 5 6 Sarah B. Pickett1,2, Eric D. Thomas1,2, Joy Y. Sebe1, Tor Linbo1, Robert Esterberg1,3, Dale W. 7 Hailey1,3, David W. Raible1,2,3* 8 1. Department of Biological Structure 9 2. Graduate Program in Neuroscience 10 3. Virginia Merrill Bloedel Hearing Research Center 11 University of Washington, 1959 NE Pacific Street, Box 357420, Seattle, WA 98195, USA 12 Correspondence: [email protected] 13 14 15 SUMMARY: 16 17 Mitochondria play a prominent role in mechanosensory hair cell damage and death. Although hair 18 cells are thought to be energetically demanding cells, how mitochondria respond to these demands 19 and how this might relate to cell death is largely unexplored. Using genetically encoded indicators, 20 we found mitochondrial calcium flux and oxidation are regulated by mechanotransduction 21 and demonstrate that hair cell activity has both acute and long-term consequences on 22 mitochondrial function. We tested whether variation in mitochondrial activity reflected differences in 23 vulnerability of hair cells to the toxic drug neomycin. We observed that susceptibility did not 24 correspond to the acute level of mitochondrial activity but rather to the cumulative history of that 25 activity. 26 27 INTRODUCTION: 28 29 Neurons are some of the most energy demanding cells in the body[1]. Because of this, they 30 are particularly vulnerable to disruptions in mitochondrial and metabolic function. Despite the 31 importance of proper mitochondrial function for all neuronal cell types, some subpopulations are 32 more vulnerable than others, such as those affected by Parkinson’s Disease, ALS, and other 33 neurodegenerative diseases[2,3]. This selective susceptibility of affected subpopulations has been 34 attributed in part to their high physiological activity and altered response to oxidative stress. In 35 addition to high activity levels, advanced age is also a major risk factor for neurodegenerative 36 diseases, suggesting a connection between aging and selective cell death. Increased selective 37 susceptibility over time could be attributed, at least in part, to the accumulation of oxidative 38 damage caused by reactive oxygen species (ROS), as postulated by the free radical theory of 39 aging[4–6]. Mitochondria have been identified as a major source of ROS and studies revealed that 40 ROS production increases with age[4,7]. If accumulation of mitochondrial oxidation and stress are 41 related to cellular activity, highly active or older cells that have experienced more accumulated 42 activity over time may be more vulnerable to acute damage. 1 43 Selective susceptibility is not limited to the central nervous system. Peripheral sensory cells 44 are also highly active as they constantly receive and filter sensory input. Among the peripheral 45 receptors, selective cell death has been well documented for hair cells, the mechanosensory cells 46 of the mammalian inner ear that mediate hearing and balance. Auditory hair cell loss is a common 47 feature of hearing impairment and can occur as a result of exposure to loud noise or clinical 48 therapeutic compounds, and with aging[8]. Hair cell loss occurs in a characteristic pattern relative 49 to both the frequency tuning of the cells and the cell type (inner vs. outer hair cells). After exposure 50 to aminoglycoside antibiotics or with advancing age, for example, hair cells tuned to higher sound 51 frequencies are lost before those tuned to lower frequencies. Similarly, outer hair cells are more 52 susceptible to damage than inner hair cells[9–15]. The cause of selective susceptibility for hair 53 cells remains unclear, although some evidence suggests that differences in hair cell metabolism, 54 free radical damage, or calcium (Ca2+) handling may be contributors, thus implicating mitochondrial 55 involvement[16–18]. A role for mitochondria is further supported by evidence of mitochondrial 56 dysfunction during hair cell damage. Across species, dying hair cells exhibit swollen mitochondrial 57 morphology and generate ROS in response to ototoxic agents, including aminoglycoside antibiotics 58 and copper[19–29]. 59 In this study, we investigate the relationship between mitochondrial activity and selective 60 toxicity of aminoglycosides using the zebrafish lateral line system. Sensory input from the lateral 61 line is mediated by externally located clusters of mechanosensitive hair cells, called neuromasts. 62 These cells allow fish to detect changes in water flow and to navigate their environment. They also 63 share many similarities with hair cells of the inner ear (see Nicolson, 2017 for review)[30]. This 64 includes conservation with human deafness genes as well as susceptibility to compounds that are 65 ototoxic [31–35]. Moreover, the surface location of the lateral line system provides a unique 66 advantage in that we can monitor cellular changes that occur in vivo during physical or chemical 67 manipulation with subcellular resolution. Like auditory hair cells, lateral line hair cells of older 68 zebrafish are more susceptible to aminoglycoside-induced cell death[36]. Unlike cochlear hair 69 cells, lateral line hair cells do not exhibit intrinsic frequency selectivity and, although lateral line hair 70 cells can be classified based on their polarity, they are not otherwise functionally sub-classified by 71 type as are mammalian auditory and vestibular hair cells[37,38]. For this reason, it has been 72 particularly puzzling why some lateral line hair cells are more susceptible than others. Here we 73 have imaged cumulative mitochondrial responses to hair cell stimulation or application of toxic 74 aminoglycosides. We demonstrate that hair cell mechanotransduction (MET) activity has both 75 acute and long-term effects on mitochondrial activity. Moreover, we demonstrate that cumulative 76 changes in mitochondrial activity correspond with hair cell susceptibility to damage. 77 78 RESULTS: 2 79 80 Mitochondria respond to acute hair cell stimulation 81 To examine mitochondrial responses to hair cell MET activity, we conducted in vivo time- 82 lapse imaging studies of lateral line hair cells expressing genetically encoded Ca2+ indicators. To 83 visualize hair cell responses to stereocilia deflection, we used the Tg[myo6b:RGECO]vo10Tg line in 84 which the red Ca2+ indicator RGECO is cytoplasmically expressed in hair cells (referred to as 85 cytoRGECO)[39]. We simultaneously monitored mitochondrial Ca2+ in the same cells using a 86 mitochondrially targeted GCaMP3 also expressed in hair cells (Tg[myo6b:mitoGCaMP3]w119) 87 (referred to as mitoGCaMP)[25,29]. Hair cells were imaged at a single plane and mechanically 88 stimulated at 10 Hz using a pressure wave applied via a waterjet pipette. Both cytoRGECO and 89 mitoGCaMP fluorescence intensity increased, indicating that hair cell stimulation causes an influx 90 of Ca2+ into both the cytoplasm and mitochondria of the activated cells. Sample frames from four 91 different videos are shown in Figure 1 for cytoRGECO (panel A) and mitoGCaMP (panel B). These 92 data demonstrate that mitochondria respond to hair cell stimulation. The two signals differ, 93 however, in their kinetics. While cytoplasmic Ca2+ levels increase and decrease rapidly with the 94 onset and offset of waterjet stimulation, mitochondrial Ca2+ levels exhibit a delayed rise and decay 95 more slowly (Fig. 1C). This is shown in traces of Ca2+ responses (Fig. 1C, 1D-F) and in summary 96 plots of rise times (Fig. 1D; cyto vs mito: 8.9 0.7s vs 13 0.9s; mean SE; n = 21 cells). The 97 integrated area (Fig. 1E; cyto vs mito: 5.6 0.7 vs 11 1.3; mean SE; n = 19 cells) and peak 98 change in the response (Fig. 1F; cyto vs mito: 1.3 0.04 vs 1.5 0.06; mean SE; n = 19 vs 21 99 cells) were significantly greater for mitochondria than for cytoplasm. 100 101 Mitochondrial activity in the absence of mechanotransduction 102 To assay acute mitochondrial activity another way, we used the cationic dye JC-1, a marker 103 of mitochondrial membrane potential. JC-1 fluorescence shifts from green to red upon aggregation 104 in energized mitochondria[40,41]. As a result, JC-1 provides a ratiometric readout of relative 105 mitochondrial membrane potential as an indicator of mitochondrial activity. To assess whether MET 106 activity altered mitochondrial activity, we examined JC-1 in wildtype or heterozygous larvae 107 (WT/Het) and in cadherin23 mutant larvae, also referred to as sputnik mutants[42]. As in 108 mammalian hair cells, Cadherin23 is a critical component of the tip links necessary for function of 109 the MET apparatus; thus sputnik mutant hair cells are MET inactive[43]. Example images of JC-1- 110 labelled hair cells are shown in figures 2A and B. JC-1 was analyzed as a ratio of red to green 111 fluorescence intensity. Compared to age-matched WT/Het siblings, sputnik mutants exhibited 112 significantly lower JC-1 fluorescence ratios, indicating that the mitochondria are depolarized (Fig. 113 2C) (WT/Het: 0.25 ± 0.24 n = 8 fish, Mutant: 0.05 ± 0.07 n = 8 fish, Mann-Whitney U test, p < 0.01, 3 114 mean ratio ± SD). These results suggest that mitochondrial activity is reduced in the absence of 115 MET. 116 117 Measuring mitochondrial aging and oxidation with mitoTimer 118 The waterjet and calcium imaging studies reveal that mitochondria respond to hair cell MET 119 activity. Ca2+ flux can have multiple effects on mitochondrial function, including regulation of 120 electron transport during oxidative phosphorylation (OXPHOS) and generation of ROS[44]. We 121 next wanted to examine whether acute MET activity causes persistent effects on the state of 122 mitochondria. To look at cumulative mitochondrial activity over time, we used a transgenic 123 zebrafish expressing the reporter mitoTimer in all hair cells (Tg[myo6b:mitoTimer]w208; here referred 124 to as mitoTimer).