1. Cell Biology
  2. Neuroscience
Download icon

Pregnancy-associated plasma protein-aa supports hair cell survival by regulating mitochondrial function

  1. Mroj Alassaf
  2. Emily C Daykin
  3. Jaffna Mathiaparanam
  4. Marc A Wolman  Is a corresponding author
  1. University of Wisconsin, United States
Research Article
  • Cited 0
  • Views 615
  • Annotations
Cite this article as: eLife 2019;8:e47061 doi: 10.7554/eLife.47061

Abstract

To support cell survival, mitochondria must balance energy production with oxidative stress. Inner ear hair cells are particularly vulnerable to oxidative stress; thus require tight mitochondrial regulation. We identified a novel molecular regulator of the hair cells’ mitochondria and survival: Pregnancy-associated plasma protein-aa (Pappaa). Hair cells in zebrafish pappaa mutants exhibit mitochondrial defects, including elevated mitochondrial calcium, transmembrane potential, and reactive oxygen species (ROS) production and reduced antioxidant expression. In pappaa mutants, hair cell death is enhanced by stimulation of mitochondrial calcium or ROS production and suppressed by a mitochondrial ROS scavenger. As a secreted metalloprotease, Pappaa stimulates extracellular insulin-like growth factor 1 (IGF1) bioavailability. We found that the pappaa mutants’ enhanced hair cell loss can be suppressed by stimulation of IGF1 availability and that Pappaa-IGF1 signaling acts post-developmentally to support hair cell survival. These results reveal Pappaa as an extracellular regulator of hair cell survival and essential mitochondrial function.

https://doi.org/10.7554/eLife.47061.001

Introduction

Without a sufficient regenerative capacity, a nervous system’s form and function critically depends on the molecular and cellular mechanisms that support its cells’ longevity. Neural cell survival is inherently challenged by the nervous system’s high energy demand, which is required to support basic functions, including maintaining membrane potential, propagating electrical signals, and coordinating the release and uptake of neurotransmitters (Halliwell, 2006; Kann and Kovács, 2007; Howarth et al., 2012). Metabolic energy is primarily supplied by mitochondrial oxidative phosphorylation (Kann and Kovács, 2007). Although this process is essential to cell survival, a cytotoxic consequence is the generation of reactive oxygen species (ROS). Oxidative stress caused by ROS accumulation damages vital cell components including DNA, proteins, and lipids (Schieber and Chandel, 2014). Neural cells are particularly vulnerable to oxidative stress due not only to their energy demand and thereby ROS production, but also to their relatively insufficient antioxidant capacity (Halliwell, 1992). This heightened susceptibility to oxidative stress-mediated cell death is believed to underlie aging and neurodegenerative disorders, including Alzheimer’s disease (AD), Parkinson’s disease (PD), and Amyotrophic lateral sclerosis (ALS) (Perry et al., 2002; Barber et al., 2006; Mattson and Magnus, 2006; Blesa et al., 2015).

Hair cells of the inner ear are a population of neural cells that are particularly susceptible to oxidative stress-induced death (Gonzalez-Gonzalez, 2017) These specialized sensory cells relay sound and balance information to the central nervous system. Hair cell death or damage, which is irreversible in mammals, is the primary cause of hearing loss, and is exacerbated by aging, genetic predisposition, exposure to loud noise and therapeutic agents (Eggermont, 2017). Identifying the molecular and cellular mechanisms that promote the longevity of hair cells is a critical step towards designing therapeutic strategies that minimize the prevalence of hearing loss and its effect on quality of life. The insulin-like growth factor-1 (IGF1) signaling pathway is known to support mitochondrial function and cell survival (García-Fernández et al., 2008; Lyons et al., 2017). IGF1 deficiency has been shown to strongly correlate with age-related hearing loss in humans and animal models (Riquelme et al., 2010; Lassale et al., 2017). Recently, exogenous IGF1 supplementation was found to protect hair cells against death by exposure to the aminoglycoside neomycin (Hayashi et al., 2013; Yamahara et al., 2017). However, it remains unclear how endogenous IGF1 signaling is regulated to support hair cell survival and whether IGF1 signaling influences the hair cell’s essential mitochondrial functions.

IGF1 is synthesized both in the liver for systemic distribution and locally in tissues, including the nervous system (Bondy et al., 1992; Sjögren et al., 1999). IGF1’s biological functions are mediated by binding to cell surface IGF1 receptors (IGF1Rs), which act as receptor tyrosine kinases. When bound by IGF1, the IGF1R autophosphorylates and stimulates intracellular PI3kinase-Akt signaling (Feldman et al., 1997). Extracellularly, IGF1 is sequestered by IGF-binding proteins (IGFBPs), which restrict IGF1-IGF1R interactions (Hwa et al., 1999). To counter the inhibitory role of IGFBPs on IGF1 signaling, locally secreted proteases cleave IGFBPs to ‘free’ IGF1 and thereby stimulate local IGF1 signaling. One such protease, Pregnancy-associated plasma protein A (Pappa), targets a subset of IGFBPs to stimulate multiple IGF1-dependent processes, including synapse formation and function (Boldt and Conover, 2007; Wolman et al., 2015; Miller et al., 2018). Pappa has not been studied for its potential to act as an extracellular regulator of IGF1-dependent hair cell survival, mitochondrial function, or oxidative stress. Here, through analyses of lateral line hair cells in zebrafish pappaa mutants, we reveal a novel role for Pappaa in regulating mitochondrial function to support hair cell survival.

Results

IGF1R signaling affects hair cell survival and mitochondrial function in zebrafish

Hair cells of the zebrafish lateral line are found in superficial neuromasts and form a rosette-like structure that is surrounded by support cells (Raible and Kruse, 2000) (Figure 1A). These hair cells share functional, morphological, and molecular similarities with mammalian inner ear hair cells (Ghysen and Dambly-Chaudière, 2007). Acute exposure of larval zebrafish to the aminoglycoside neomycin triggers hair cell death and mitochondrial dysfunction (Harris et al., 2003; Esterberg et al., 2014; Esterberg et al., 2016). This experimental platform has been used to dissect the molecular and cellular mechanisms that support hair cell survival (Owens et al., 2008). A role for IGF1R signaling in the survival of zebrafish lateral line hair cells and their mitochondria has yet to be demonstrated. We hypothesized that if IGF1R signaling supports hair cell survival, then attenuating IGF1R signaling would further reduce hair cell survival following neomycin exposure. To test this, we used a transgenic line in which an inducible heat shock promoter drives ubiquitous expression of a dominant negative IGF1Ra [Tg (hsp70:dnIGF1Ra-GFP)] (Kamei et al., 2011). dnIGF1Ra-GFP expression was induced from 24 hr post fertilization (hpf) to 5 days post fertilization (dpf). At five dpf, larvae were exposed to neomycin for 1 hr and evaluated for hair cell survival 4 hr later. Larvae expressing dnIGF1Ra-GFP showed a greater reduction in hair cell survival compared to heat-shocked wild type and non-heat- shocked Tg (hsp70:dnIGF1Ra-GFP) larvae (Figure 1B).

Inhibition of IGF1R activity enhances neomycin-induced hair cell death.

(A) Schematic of lateral line neuromast. (B) Mean percentage of surviving hair cells following induction of dnIGF1Ra-GFP expression. To calculate hair cell survival percentage, hair cell number 4 hr post-neomycin treatment was normalized to mean hair cell number in non-heat-shocked, vehicle-treated larvae of the same genotype. **p<0.01, ***p<0.001, ****p<0.0001 two-way ANOVA, Holm-Sidak post test. N = 7–14 larvae per group (shown at base of bars), three neuromasts perlarva from two experiments. (C) brn3c:GFP labeled hair cells loaded with TMRE in NVP-AEW541 and vehicle treated larvae. (D–D’) Mean TMRE fluorescence (D) and mean TMRE fluorescence normalized to GFP fluorescence (D’) from Z-stack summation projections of brn3c:GFP labeled hair cells. N = 9 larvae per group. Total number of neuromasts included in the analysis = 26 (vehicle treated) and 27 (NVP-AEW541). **p<0.01, ****p<0.0001. Unpaired t test, Welch-corrected.

https://doi.org/10.7554/eLife.47061.002

Next, we evaluated mitochondrial activity in hair cells following attenuation of IGF1R signaling. A mitochondria’s transmembrane potential is closely linked to its functions (Zorova et al., 2018) and can be visualized by the fluorescent, potentiometric probe TMRE. This live cationic dye readily accumulates in mitochondria based on the negative charge of their membrane potential. (Crowley et al., 2016). To determine whether IGF1R attenuation affected mitochondrial membrane potential, we treated wild type larvae with NVP-AEW541, a selective inhibitor of IGF1R phosphorylation (Chablais and Jazwinska, 2010), and loaded the hair cells with TMRE. We found that pharmacological attenuation of IGF1R signaling in wild type hair cells resulted in increased TMRE fluorescence (Figure 1C–D’). Together, these results indicate that IGF1R signaling regulates mitochondrial activity and the survival of zebrafish lateral line hair cells.

Pappaa is expressed by lateral line neuromast support cells

We were next curious whether extracellular regulation of IGF1 signaling was important for hair cell survival. A strong candidate to stimulate extracellular IGF1 availability is Pappaa (Lawrence et al., 1999). In situ hybridization revealed pappaa expression in lateral line neuromasts that co-localizes with the position of support cells, which surround the hair cell rosette (Figure 2A–B). To determine whether pappaa was also expressed by hair cells, but at levels below detection by fluorescent in situ hybridization (Figure 2B), we performed RT-PCR on fluorescently sorted hair cells from five dpf Tg(brn3c:GFP) (Xiao et al., 2005) larvae (Figure 2C). Again, we found that hair cells did not express pappaa (Figure 2C). Our in situ analysis also showed pappaa expression in the ventral spinal cord, where motor neurons reside (Figure 2A). As a control for fluorescent cell sorting and detection of pappaa by RT-PCR, we performed RT-PCR for pappaa on fluorescently sorted motor neurons from five dpf Tg(mnx1:GFP) larvae (Rastegar et al., 2008), and confirmed pappaa expression by motor neurons (Figure 2C).

pappaa is expressed by neuromast support cells and motor neurons.

(A) Whole mount in situ hybridization shows pappaa mRNA expression at four dpf by lateral line neuromasts (arrowheads). (B) Fluorescent in situ hybridization of pappaa (green) and brn3c:GFP labeled hair cells (white) shows pappaa mRNA expression by the support cells that surround hair cells. (C) RT-PCR of fluorescently sorted brn3c:GFP labeled hair cells and mnx1:GFP labeled motor neurons shows pappaa expression by motor neurons, but not hair cells. RT-PCR products represent pappaa cDNA fragment.

https://doi.org/10.7554/eLife.47061.005

Pappaa supports hair cell survival

We next sought to determine whether Pappaa’s regulation of IGF1 signaling supports hair cell survival. We examined hair cell survival after neomycin treatment of 5 dpf pappaa mutants (hereafter referred to as pappaap170). pappaap170 mutants harbor nonsense mutations upstream of Pappaa’s proteolytic domain and show reduced IGF1R activation in other neural regions of pappaa expression (Wolman et al., 2015; Miller et al., 2018). Following exposure to neomycin, hair cells of pappaap170 larvae showed reduced hair cell survival compared to wild type hair cells (Figure 3A–B). Notably, the support cells were unaffected by neomycin exposure in both genotypes (Figure 3—figure supplement 1). Next, we hypothesized that if Pappaa is acting through the IGF1 signaling pathway, then stimulating IGF1 signaling would improve hair cell survival in pappaap170 larvae. To test this hypothesis, we bathed wild type and pappaap170 larvae in recombinant human IGF1 protein. Pre-treatment with IGF1 for 24 hr prior to and during neomycin exposure improved hair cell survival in pappaap170 larvae at concentrations of IGF1 that had no effect on hair cell survival in wild type larvae (Figure 3C). Because Pappaa acts to increase IGF1 bioavailability by freeing IGF1 from IGFBPs (Boldt and Conover, 2007), we asked whether this role was important for hair cell survival following neomycin exposure. To test this, we bathed wild type and pappaap170 larvae in NBI-31772, an IGFBP inhibitor that stimulates IGF1 availability (Safian et al., 2016). Treatment with NBI-31772 for 24 hr prior to and during neomycin exposure improved hair cell survival in wild type and pappaap170 larvae (Figure 3D). Taken together, these results suggest that extracellular regulation of IGF1 bioavailability by Pappaa enhances hair cell survival.

Figure 3 with 1 supplement see all
Hair cell survival is reduced in zebrafish pappaap170 larvae.

(A) Representative images of brn3c:GFP labeled hair cells from vehicle or 10 µM neomycin treated larvae. Scale = 10 µm. (B) Mean percentage of surviving hair cells. To calculate hair cell survival percentage, hair cell number 4 hr post-neomycin treatment was normalized to mean hair cell number in vehicle treated larvae of the same genotype. **p<0.01, ***p<0.001, two-way ANOVA, Holm-Sidak post test. N = 11–15 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 45 (wild type; vehicle-treated), 39 (wild type; 1 µM neomycin), 39 (wild type; 10 µM neomycin), 39 (wild type; 30 µM neomycin), 45 (pappaap170; vehicle-treated), 39 (pappaap170; 1 µM neomycin), 39 (pappaap170; 10 µM neomycin), and 33 (pappaap170; 30 µM neomycin) from two experiments. (C) Mean percentage of surviving hair cells following co-treatment with IGF1 and 10 µM neomycin. To calculate hair cell survival percentage, hair cell counts after treatment were normalized to hair cell number in vehicle treated larvae of same genotype. *<0.05, **p<0.01 ****p<0.0001. Two-way ANOVA, Holm-Sidak post test. N = 7–11 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 33 (wild type; vehicle-treated), 27 (wild type; 10 µM neomycin), 24 (wild type; 10 µM neomycin +10 ng IGF1), 21 (wild type; 10 µM neomycin +30 ng IGF1), 24 (wild type; 10 µM neomycin +100 ng IGF1), 33 (pappaap170; vehicle-treated), 24 (pappaap170; 10 µM neomycin), 24 (pappaap170; 10 µM neomycin +10 ng IGF1), 21 (pappaap170; 10 µM neomycin +30 ng IGF1), and 21 (pappaap170; 10 µM neomycin +100 ng IGF1) (D) Mean percentage of surviving hair cells following co-treatment with NBI-31772 and 10 µM neomycin. To calculate hair cell survival percentage, hair cell counts after treatment were normalized to hair cell number in vehicle treated larvae of same genotype. **p<0.01 ****p<0.0001. Two-way ANOVA, Holm-Sidak post test. N = 7–9 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 15 (wild type; vehicle-treated), 27 (wild type; 10 µM neomycin), 21 (wild type; 10 µM neomycin +30 µM NBI-31772), 24 (wild type; 10 µM neomycin +100 µM NBI-31772), 15 (pappaap170; vehicle-treated), 24 (pappaap170; 10 µM neomycin), 21 (pappaap170; 10 µM neomycin +30 µM NBI-31772), and 24 (pappaap170; 10 µM neomycin +100 µM NBI-31772).

https://doi.org/10.7554/eLife.47061.006

Pappaa acts post-developmentally to promote hair cell survival

We next assessed when Pappaa acts to support hair cell survival. Zebrafish lateral line hair cells begin to appear at two dpf and are fully functional by four dpf (Raible and Kruse, 2000; Ghysen and Dambly-Chaudière, 2007). At five dpf, pappaap170 larvae are responsive to acoustic stimuli (Wolman et al., 2015), suggesting that their hair cells are functionally intact. In addition, pappaap170 hair cells appeared morphologically indistinguishable from hair cells in wild type larvae (Figure 4A). Therefore, we hypothesized that Pappaa acts post-developmentally to support hair cell survival. To test this idea, we asked whether post-developmental expression of Pappaa was sufficient to suppress the pappaa mutant’s enhanced hair cell loss when exposed to neomycin. We generated a transgenic line in which a temporally inducible heat shock promoter drives ubiquitous expression of Pappaa (Tg(hsp70:pappaa-GFP)). We found that induced expression of Pappaa, beginning at four dpf and through neomycin treatment at five dpf, resulted in the complete rescue of pappaap170 hair cell sensitivity to neomycin and raised pappaap170 hair cell survival to wild type levels (Figure 4B). Consistent with these results, we found that post-developmental attenuation of IGF1R signaling, through induction of dnIGF1Ra-GFP expression beginning at four dpf, was sufficient to reduce hair cell survival when exposed to neomycin at five dpf (Figure 4C).

Post-developmental regulation of IGF1 signaling by Pappaa is required for hair cell survival.

(A) Lateral view of brn3c:GFP labeled hair cells in 5dpf wild type and pappaap170 larvae. Scale = 10 µm. (B) Mean percentage of surviving hair cells following post-developmental induction of Pappaa expression in pappaap170 larvae. To calculate hair cell survival percentage, hair cell number 4 hr post-neomycin treatment was normalized to mean hair cell number in non-heat-shocked, vehicle-treated larvae of the same genotype. **p<0.01, ***p<0.001, ****p<0.0001, 2-way ANOVA, Holm-Sidak post test. N = 5–9 larvae per group (shown at base of bars), three neuromasts perlarva. (C) Mean percentage of surviving hair cells following post-developmental induction of dnIGF1Ra-GFP expression. To calculate hair cell survival percentage, hair cell number 4 hr post-neomycin treatment was normalized to mean hair cell number in non-heat-shocked, vehicle-treated larvae of the same genotype. ****p<0.0001, two-way ANOVA, Holm-Sidak post test. N = number of larvae per group (shown at base of bars), three neuromasts per larva.

https://doi.org/10.7554/eLife.47061.012

Pappaa loss causes increased mitochondrial ROS in hair cells

A role for Pappaa in hair cell survival is novel. To define how Pappaa activity influences hair cell survival, we evaluated cellular mechanisms known to underlie their neomycin-induced death. Neomycin enters hair cells via mechanotransduction (MET) channels found on the tips of stereocilia (Kroese et al., 1989). MET channel permeability has been correlated to hair cells’ neomycin sensitivity (Alharazneh et al., 2011; Stawicki et al., 2016). We hypothesized that pappaap170 hair cells may be more susceptible to neomycin-induced death due to an increase in MET channel-mediated entry. To assess MET channel entry we compared uptake of FM1-43, a fluorescent styryl dye that enters cells through MET channels (Meyers et al., 2003), by hair cells in wild type and pappaap170 larvae. FM1-43 fluorescence was equivalent between wild type and pappaap170 hair cells (Figure 5A–B’), suggesting that the increased death of pappaap170 hair cells was not due to increased MET channel permeability.

Figure 5 with 1 supplement see all
Pappaa regulates mitochondrial ROS generation.

(A, C, E) Still images of live brn3c:GFP hair cells loaded with the amphypathic styryl dye FM1-43 (A) or cytoplasmic or mitochondrial ROS indicators (C: CellROX, E: mitoSOX). Scale = 10 µm. (B, D, F) Mean dye fluorescence (D) and mean dye fluorescence normalized to GFP fluorescence (B’, D’, F’) from Z-stack summation projections of brn3c:GFP labeled hair cells. N = 5–6 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 18 (wild type; FM1-43), 18 (pappaap170; FM1-43), 21 (wild type; CellROX), 21 (pappaap170; CellROX), 18 (wild type; mitoSOX), and 22 (pappaap170; mitoSOX). *p<0.05, **<p < 0.01, ***p<0.001. Unpaired t test, Welch-corrected. (G) Mean percentage of surviving hair cells post Antimycin A treatment. To calculate hair cell survival percentage, hair cell counts after treatment were normalized to hair cell number in vehicle treated larvae of same genotype. **p<0.01 ****p<0.0001. Two-way ANOVA, Holm-Sidak post test. N = 9–10 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 30 (wild type; vehicle-treated), 30 (wild type; 100pM antimycin a), 27 (wild type; 300pM antimycin a), 30 (wild type; 500pM antimycin a), 27 (pappaap170; vehicle-treated), 27 (pappaap170; 100pM antimycin a), 27 (pappaap170; 300pM antimycin a), and 30 (pappaap170; 500pM antimycin a).

https://doi.org/10.7554/eLife.47061.015

We next questioned whether Pappaa affects essential organelle functions in hair cells, which are known to be disrupted by neomycin. Within the hair cell, neomycin triggers Ca2+ transfer from the endoplasmic reticulum to the mitochondria (Esterberg et al., 2014). This Ca2+ transfer results in stimulation of the mitochondrial respiratory chain, increased mitochondrial transmembrane potential, and increased ROS production (Görlach et al., 2015; Esterberg et al., 2016). The ensuing oxidative stress ultimately underlies the neomycin’s cytotoxic effect on hair cells. To explore whether excessive ROS production underlies pappaap170 hair cells’ increased sensitivity to neomycin, we evaluated cytoplasmic ROS levels with a live fluorescent indicator of ROS (CellROX) (Esterberg et al., 2016). pappaap170 hair cells displayed elevated ROS levels at baseline; prior to addition of neomycin (Figure 5C–D’). Given that the mitochondria are the primary generators of cellular ROS (Lenaz, 2001), we asked whether the elevated levels of cytoplasmic ROS observed in pappaap170 hair cells originated from the mitochondria. We evaluated mitochondrial ROS with the live fluorescent indicator mitoSOX (Esterberg et al., 2016), again without neomycin treatment, and observed increased signal in hair cells of pappaap170 compared to wild type (Figure 5E–F’). This increased mitochondrial ROS was not due to an overabundance of mitochondria within pappaap170 hair cells, as determined by measuring mitochondrial mass with mitotracker (Figure 5—figure supplement 1).

We hypothesized that the elevated ROS in pappaap170 hair cells predisposed them closer to a cytotoxic threshold of oxidative stress. To test this idea, we asked whether pappaap170 showed reduced hair cell survival following pharmacological stimulation of mitochondrial ROS via exposure to Antimycin A, an inhibitor of the mitochondrial electron transport chain (Hoegger et al., 2008; Quinlan et al., 2011) We found that pappaap170 hair cells showed increased death by Antimycin A compared to wild type hair cells (Figure 5G). These results are consistent with the idea that pappaap170 hair cells are predisposed to oxidative stress-induced death due to elevated baseline levels of ROS.

Pappaa regulates mitochondrial Ca2+uptake and transmembrane potential

Mitochondrial ROS production is stimulated by Ca2+ entry into the mitochondria (Brookes et al., 2004; Görlach et al., 2015). Given the increased mitochondrial ROS in pappaap170 hair cells, we asked whether the mutants’ hair cell mitochondria exhibited increased Ca2+ levels. To address this, we used a transgenic line Tg(myo6b:mitoGCaMP3), in which a mitochondria-targeted genetically encoded Ca2+ indicator (GCaMP3) is expressed in hair cells (Esterberg et al., 2014). Live imaging of mitoGCaMP3 fluorescence revealed a 2-fold increase in fluorescent intensity in pappaap170 hair cells compared to wild type hair cells (Figure 6A–B). Mitochondrial Ca2+ uptake is driven by the negative electrochemical gradient of the mitochondrial transmembrane potential, a product of mitochondrial respiration. Ca2+ -induced stimulation of mitochondrial oxidative phosphorylation causes further hyperpolarization of mitochondrial transmembrane potential, leading to increased uptake of Ca2+ (Brookes et al., 2004; Adam-Vizi and Starkov, 2010; Ivannikov and Macleod, 2013; Esterberg et al., 2014; Görlach et al., 2015). Therefore, we hypothesized that pappaap170 mitochondria would have a more negative transmembrane potential compared to wild type. Using TMRE as an indicator of mitochondrial transmembrane potential (Perry et al., 2011), we found that pappaap170 mitochondria possess a more negative transmembrane potential compared to wild type (Figure 6C–D’). This increased TMRE signal is similar to our observations following pharmacological inhibition of IGF1R (Figure 1C–D’).

Mitochondrial Ca2+ levels and transmembrane potential are disrupted in pappaap170 hair cells.

(A) Still images from live myo6b:mitoGCaMP3 labeled hair cells. Scale = 10 µm. (B) Mean mitoGCaMP fluorescence; quantified from Z-stack summation projection. *p<0.05. Unpaired t test, Welch-corrected. N = 7–9 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 19 (wild type) and 26 (pappaap170). (C) Still images from live brn3c:GFP labeled hair cells loaded with TMRE. Scale = 10 µm. (D) Mean TMRE fluorescence (D) and mean TMRE fluorescence normalized to GFP fluorescence (D’) from Z-stack summation projections of brn3c:GFP labeled hair cells.. N = 6–8 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 17 (wild type) and 23 (pappaap170).. **p<0.01. Unpaired t test, Welch-corrected. (E) Mean percentage of surviving hair cells post Cyclosporin A treatment. To calculate hair cell survival percentage, hair cell counts post-treatment were normalized to hair cell numbers in vehicle treated larvae of same genotype. N = 7–10 larvae per group (shown at base of bars). Total number of neuromasts included in the analysis = 45 (wild type; vehicle-treated), 27 (wild type; 0.1 µM CsA), 21 (wild type; 1 µM CsA), 21 (wild type; 3 µM CsA), 42 (pappaap170; vehicle-treated), 30 (pappaap170; 0.1 µM CsA), 27 (pappaap170; 1 µM CsA), and 21 (pappaap170; 3 µM CsA) from two experiments. *p<0.05. Two-way ANOVA, Holm-Sidak post test.

https://doi.org/10.7554/eLife.47061.023

Given that mitochondria of pappaap170 hair cells exhibited elevated Ca2+ (Figure 6A–B) and a more negative transmembrane potential (Figure 6C–D’), we hypothesized that pharmacologically disrupting these mitochondrial features would have a more cytotoxic effect on pappaap170 hair cells. To test this idea, we exposed wild type and pappaap170 larvae to Cyclosporin A (CsA), an inhibitor of the mitochondrial permeability transition pore that causes buildup of mitochondrial Ca2+ and further hyperpolarizes mitochondria (Crompton et al., 1988; Esterberg et al., 2014). pappaap170 larvae showed reduced hair cell survival at concentrations of CsA, which had no effect on hair cell survival in wild type larvae (Figure 6E). Taken together, these results suggest that Pappaa loss disrupts mitochondrial functions that can predispose hair cells to death.

Pappaa regulates the expression of mitochondrial antioxidants

Oxidative stress can be caused by an imbalance in ROS production and antioxidant activity (Betteridge, 2000). IGF1 signaling positively correlates with antioxidant expression (Higashi et al., 2013; Wang et al., 2016). Therefore, we questioned whether the cytotoxicity of excessive ROS in pappaap170 was compounded by a reduced antioxidant system. To address this, we compared gene expression of antioxidants in wild type and pappaap170 hair cells by RT-qPCR. This analysis revealed reduced expression of mitochondrial antioxidants genes (gpx, sod1, and sod2) (Figure 7A) (Weisiger and Fridovich, 1973; Okado-Matsumoto and Fridovich, 2001; Higgins et al., 2002; Brigelius-Flohé and Maiorino, 2013), but not expression of catalase; an antioxidant that does not localize to mitochondria (Zhou and Kang, 2000). These results suggest that the pappaap170 hair cells’ elevated ROS can be attributed not only to increased mitochondrial calcium and transmembrane potential, but also to reduced mitochondrial antioxidants. Finally, we asked whether the increased mitochondria-generated ROS in pappaap170 hair cells was sufficient to explain their increased mortality rate when exposed to neomycin. We hypothesized that if this were the case, then reducing mitochondrial-ROS would suppress their increased mortality. To test this idea we exposed pappaap170 larvae to the mitochondria-targeted ROS scavenger mitoTEMPO (Esterberg et al., 2016) and observed up to complete protection of pappaap170 hair cells against neomycin-induced death (Figure 7B). These results suggest that abnormally elevated mitochondrial ROS underlies the enhanced hair cell death in neomycin treated pappaap170 zebrafish.

Pappaa regulates expression of mitochondrial antioxidants.

(A) Mean fold change in antioxidants transcript levels in hair cells at five dpf. cDNA was from FACsorted hair cells that were collected from 200 Tg(brn3c:GFP) dissected tails. N = 2–4 technical replicates/gene. *p<0.05, ***p<0.001. Multiple t tests, Holm-Sidak post test. Error bars = SEM. (B) Mean percentage of surviving pappaap170 hair cells following co-treatment with mitoTEMPO and 10 µM neomycin. To calculate hair cell survival percentage, hair cell counts 4 hr post-neomycin treatment were normalized to hair cell counts in vehicle treated pappaap170 larvae. **p<0.01, ***p<0.001, ****p<0.0001. Two-way ANOVA, Holm-Sidak post test. N = 4–10 (shown at base of bars). Total number of neuromasts included in the analysis = 30 (wild type; vehicle-treated), 30 (wild type; 10 µM neomycin), 21 (wild type; 10 µM neomycin +10 µM mitoTEMPO), 30 (wild type; 10 µM neomycin +50 µM mitoTEMPO), 27 (wild type; 10 µM neomycin +100 µM mitoTEMPO), 30 (pappaap170; vehicle-treated), 12 (pappaap170; 10 µM neomycin), 15 (pappaap170; 10 µM neomycin +10 µM mitoTEMPO), 21 (pappaap170; 10 µM neomycin +50 µM mitoTEMPO), and 30 (pappaap170; 100 µM neomycin +10 µM mitoTEMPO). Error bars = SEM.

https://doi.org/10.7554/eLife.47061.027

Discussion

Extracellular factors and the regulation of mitochondrial function and oxidative stress have been demonstrated to support the survival of various cell types (Hasan et al., 2003; Echave et al., 2009; Li et al., 2009; Wang et al., 2013; Genis et al., 2014; Pyakurel et al., 2015; Kim, 2017). For cells that do not have a capacity for regeneration, like hair cells, the regulation of these factors and intracellular processes are particularly important. Exogenous application of IGF1 was recently shown to protect hair cells against neomycin exposure (Hayashi et al., 2013). This finding identified a molecular pathway that could be potentially targeted to combat sensorineural hearing loss. Yet, significant questions remained regarding the mechanism by which IGF1 signaling serves this role. For example, it was unclear whether endogenous IGF1 signaling supports hair cell survival and how this pathway is extracellularly regulated to promote hair cell survival. Through zebrafish mutant analysis, we identified a novel extracellular regulator of IGF1 signaling that supports hair cell survival and mitochondrial function: the secreted metalloprotease Pappaa. Based on a series of in vivo experiments we propose a model by which Pappaa stimulates IGF1R signaling in hair cells to control mitochondrial function and oxidative stress, and thereby, promotes the longevity of these cells.

Pappaa regulated IGF1 signaling supports hair cell survival

Pappaa acts as an extracellular positive regulator of IGF1 signaling by cleaving inhibitory IGFBPs, thereby freeing IGF1 to bind and activate cell-surface IGF1Rs (Boldt and Conover, 2007). In the nervous system, Pappaa’s signaling role has been shown to support synaptic structure and function, but a cell protective function had not been explored. Our results demonstrate that pappaap170 hair cells showed increased mortality to neomycin (Figure 3A–B), and that this phenotype could be suppressed by pharmacological stimulation of IGF1 availability (Figure 3C and D). Given the novelty for Pappaa in supporting hair cell survival, it is interesting to consider whether Pappaa acts developmentally or post-developmentally in this context. In five dpf pappaap170, hair cells appeared to develop normally based on their cellular morphology (Figure 4A) and ability to mediate acoustic startle responses (Wolman et al., 2015). We found that post-developmental expression of Pappaa was sufficient to increase the pappaap170 hair cells’ survival when exposed to neomycin to near wild type levels (Figure 4B). Consistent with this post-developmental role for Pappaa, post-developmental attenuation of IGF1R signaling was also sufficient to increase neomycin-induced hair cell loss (Figure 4C), while stimulation of IGF1R signaling was sufficient to suppress pappaap170 hair cell loss (Figure 3C and D). Taken together, these findings suggest that Pappaa-IGF1R signaling acts post-developmentally to mediate resistance against toxins, like neomycin.

To support hair cell survival, pappaa’s expression pattern suggests that Pappaa is likely to act in a paracrine manner. Although hair cells require Pappaa for survival, they do not express pappaa. Rather, pappaa is expressed by the adjacent support cells (Figure 2A–C). Support cells have been shown to secrete factors that promote hair cell survival (May et al., 2013; Yamahara et al., 2017) and our results suggest that Pappaa is one such factor. To understand Pappaa’s cell autonomy it will be necessary to define in which cells IGF1 signaling activation is required. It is possible that Pappaa does not act directly on hair cells, rather it may influence support cells to promote hair cell survival. Moreover, it will be interesting to define the molecular cues that trigger Pappaa activity, their cellular source, and to determine whether Pappaa acts directly in response to such cues or serves a more preventative role for hair cells.

Pappaa affects mitochondrial function and oxidative stress

To understand how a Pappaa-IGF1 signaling deficiency increased neomycin-induced hair cell loss, we examined the hair cells’ mitochondria, which are known to be disrupted by neomycin (Esterberg et al., 2016). The mitochondria in pappaap170 hair cells showed multiple signs of dysfunction, including elevated ROS (Figure 5E–F’), transmembrane potential (Figure 6C–D’), and Ca2+ load (Figure 6A–B). Consistent with these observations, reduced IGF1 signaling has been associated with increased ROS production and oxidative stress (García-Fernández et al., 2008; Lyons et al., 2017). Three lines of evidence suggest that mitochondrial dysfunction, and particularly the elevated ROS levels, underlie the increased hair cell loss in pappaap170. First, pappaap170 hair cells showed enhanced sensitivity to pharmacological stimulators of mitochondrial ROS production (Figures 5G and 6E). Second, pappaap170 hair cells showed reduced expression of mitochondrial antioxidant genes (Figure 7A). Third, attenuation of mitochondrial ROS was sufficient to suppress neomycin-induced hair cell loss in pappaap170 (Figure 7B).

Based on results presented here, we can only speculate on the primary subcellular locus and defect that triggers mitochondrial dysfunction in pappaap170 hair cells. The challenge lies in the tight interplay between mitochondrial transmembrane potential, Ca2+ load, and ROS production and clearance (Brookes et al., 2004; Adam-Vizi and Starkov, 2010; Ivannikov and Macleod, 2013; Esterberg et al., 2014; Görlach et al., 2015). The oxidative phosphorylation that generates ROS relies on maintaining a negative mitochondrial transmembrane potential. Negative transmembrane potential is achieved by pumping protons out of the mitochondrial matrix as electrons move across the electron transport chain. Protons then move down the electrochemical gradient through ATP synthase to produce ATP. Given that ROS is a byproduct of oxidative phosphorylation, a more negative transmembrane potential yields more ROS (Kann and Kovács, 2007; Zorov et al., 2014). Mitochondrial Ca2+ is a key regulator of transmembrane potential and the resultant ROS generation, as it stimulates the activity of key enzymes involved in oxidative phosphorylation (Brookes et al., 2004). And, Ca2+ uptake by the mitochondria is driven by the electrochemical gradient of a negative transmembrane potential. Thus, Ca2+ and transmembrane potential are locked in a feedback loop (Brookes et al., 2004; Adam-Vizi and Starkov, 2010; Ivannikov and Macleod, 2013; Esterberg et al., 2014; Görlach et al., 2015). Because mitochondria in pappaap170 hair cells have a more negative transmembrane potential (Figure 6C–D’) and experience Ca2+ overload (Figure 6A–B), this likely sensitizes the mitochondria to any further increase in Ca2+ levels. In support of this idea, pappaap170 hair cells were hypersensitive to Cyclosporin A (Figure 6E), which increases mitochondrial Ca2+ levels by blocking the mitochondrial permeability transition pore (Smaili and Russell, 1999).

Given that oxidative stress is caused by the imbalance between ROS production and clearance (Betteridge, 2000), pappaap170 mitochondrial dysfunction may have been triggered by their weak antioxidant system (Figure 7A). As the main site for ROS generation, mitochondria are primary targets of ROS-induced damage. Specifically, ROS can damage the mitochondrial oxidative phosphorylation machinery leading to excessive electron ‘leaks’ and the ensuing ROS formation (Marchi et al., 2012). To prevent such oxidative damage, mitochondria are equipped with antioxidants that can rapidly neutralize ROS by converting them into water molecules. Failure of antioxidants to clear ROS, either due to reduced enzymatic activity or reduced expression levels, has been shown to damage several components of the mitochondria causing oxidative stress that culminates in cell death (Williams et al., 1998; Armstrong and Jones, 2002; Aquilano et al., 2006; Velarde et al., 2012). Therefore, it is no surprise that mitochondria-targeted antioxidants have shown great promise in clinical studies to treat neurodegeneration (Sheu et al., 2006). Indeed, mitochondria-targeted antioxidant treatment was successful in preventing ROS-induced hair cell death (Figure 7B). Given that ROS can modulate the activity of Ca2+ channels and consequently raise mitochondrial Ca2+ load and transmembrane potential (Chaube and Werstuck, 2016), it is possible that the reduced mitochondrial antioxidants levels in pappaap170 hair cells triggered their mitochondrial dysfunction resulting in a vicious cycle of further ROS production. Alternatively, pappaap170 mitochondria may suffer from overproduction of ROS that is compounded by insufficient clearance. Further experimental dissections of mitochondria in pappaap17mutant hair cells are needed to define the primary locus by which a deficiency in Pappaa-IGF1 signaling alters subcellular processes in hair cells.

Conclusions and outlook

Here, we define a novel role for Pappaa in hair cell survival. Although ample evidence exists about the protective role of IGF1, little is known about how IGF1’s extracellular availability is regulated to promote cell survival. Our results demonstrate that extracellular regulation of IGF1 by Pappaa is critical for maintaining mitochondrial function, and in turn, survival of hair cells. Future work should explore whether Pappaa plays a similar role in other neural cell types, including motor neurons and the retinal cells (Miller et al., 2018), where IGF1 signaling is known to affect cell survival (Lewis et al., 1993; Sakowski et al., 2009) Furthermore, future experimentation will be needed to resolve the cellular autonomy of Pappaa-IGF1R signaling to cell survival and to define the primary subcellular locus by which this signaling axis influences mitochondrial activity and oxidative stress.

Materials and methods

Key resources table
Reagent type
(species) or
resource
DesignationSource or referenceIdentifiersAdditional
information
Gene (Danio rerio)pappaap170Wolman et al., 2015RRID:ZFIN_ZDB-GENO-190322-4single nucleotide nonsense mutationC > T at position 964 in Exon 3
Strain, strain background (Danio rerio)Tg(myo6b:mitoGCaMP3)Esterberg et al., 2014RRID:ZFIN_ZDB-GENO-141008-1
Strain, strain background (Danio rerio)Tg(brn3c:GFP)Xiao et al., 2005RRID:ZFIN_ZDB-ALT-050728-2
Strain, strain background (Danio rerio)Tg(mnx1:GFP)Rastegar et al., 2008RRID:ZFIN_ZDB-GENO-140605-2
Strain, strain background (Danio rerio)Tg(hsp70:dnIGF1Ra-GFP)Kamei et al., 2011RRID:ZFIN_ZDB-GENO-110614-4
Strain, strain background (Danio rerio)Tg(hsp70:pappaa-GFP)this paperMaterials and methods subsection maintenance of zebrafish
Strain, strain background (Danio rerio)TLFZebrafish International Resource Center (ZIRC)RRID:ZFIN_ZDB-GENO-990623-2
Antibodyanti-myosinVI (rabbit polyclonal)Proteus biosciencesRRID:AB_100136261:200
Antibodyanti-SOX2 (rabbit polyclonal)AbcamRRID:AB_23411931:200
Antibodyanti-GFP (rabbit polyclonal)ThermoFisher ScientificRRID:AB_2215691:500
AntibodyAlexa 488, secondary (rabbit polyclonal)ThermoFisher ScientificRRID:AB_25762171:500
Peptide, recombinant proteinIGF-1Cell sciencesCatalog number: DU100
OtherFM1-43ThermoFisher
Scientific
Catalog number:
T3136 PubChem CID:6508724
3 uM for 30 s
OtherCellROX deep redThermoFisher ScientificCatalog number: C1042210 uM for 60 min
OthermitoSOX redThermoFisher ScientificCatalog number:
M36008
1 uM for 30 min
Othermitotracker greenFMThermoFisher ScientificCatalog number:
M7514 PubChem
CID:70691021
100 nM for 5 min
OtherTMREThermoFisher ScientificCatalog number: T669 PubChem CID:276268225 nM for 20 min
Other0.25% trypsin-EDTASigma-AldrichCatalog number: T3924
OtherTRIzolInvitrogenCatalog number: 15596026
OtherSso fast Eva Green SupermixBioRadCatalog number: 1725200
Chemical compound, drugNeomycinSigma-AldrichCatalog number: N1142 PubChem CID:24897464
Chemical compound, drugNBI-31772Fisher ScientificCatalog number: 519210
Chemical compound, drugantimycin-aSigma-AldrichCatalog number: A8674 PubChem
CID: 24891355
Chemical compound, drugCsAAbcamCatalog number: ab120114 PubChem
ID: 5284373
Chemical compound, drugmitoTEMPOSigma-AldrichCatalog number: SML0737 PubChem
CID: 134828258
Chemical compound, drugNVP-AEW541SelleckCatalog number: S1034 PubChem
CID: 11476171
Commercial assay or kitSuperScript II Reverse TranscriptaseInvitrogenCatalog number: 18064014
Software, algorithmFluoview (FV10-ASW 4.2)OlympusRRID:SCR_014215
Software, algorithmImageJPMID: 22743772RRID:SCR_003070
Software, algorithmGraphPad PrismGraphPadRRID:SCR_002798

Maintenance of zebrafish

Request a detailed protocol

To generate pappaa+/+ and pappaap170 larvae for experimentation, adult pappaap170/+ zebrafish (on a mixed Tubingen long-fin (TLF), WIK background) were crossed into the following transgenic zebrafish backgrounds: Tg(brn3c:GFP)s356t, Tg(hsp70:dnIGF1Ra-GFP)zf243, Tg(hsp70:pappaa-GFP),Tg(mnx1:GFP)mI5, and Tg(myo6b:mitoGCaMP3)w78 and then incrossed. To establish the Tg(hsp70:pappaa-GFP) line, used gateway cloning vectors (Invitrogen) to generate a transgenesis construct under the control of a ubiquitous heat shock promoter (hsp70). The pappaa sequence was generated from cDNA of 120 hpf TLF embryos. Expression was assessed by fluorescence of a co-translated and post-translationally cleaved green GFP after induction by exposure in a heated water bath for 1 hr. The stable Tg(hsp70:pappaa-GFP) line was maintained on a pappaap170/+ background.. Embryonic and larval zebrafish were raised in E3 media (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO4, pH adjusted to 6.8–6.9 with NaHCO3) at 29°C on a 14 hr/10 hr light/dark cycle through 5 days post fertilization (dpf) (Kimmel et al., 1995; Gyda et al., 2012). All experiments were done on larvae between 4–6 dpf. Genotyping of pappaap170 larvae was performed as previously described (Wolman et al., 2015).

Pharmacology

Request a detailed protocol

The following treatments were performed on Tg(brn3c:GFP) larvae through the addition of compounds to the larvae’s E3 media at five dpf unless otherwise noted. Neomycin sulfate solution (Sigma-Aldrich N1142) was added at 1–30 µM for 1 hr. Cyclosporin A (Abcam ab120114; dissolved in DMSO) was added at 0.3–3 µM for 1 hr. Antimycin A (Sigma-Aldrich A8674; dissolved in DMSO) was added at 100–500 pM for 24 hr, beginning at four dpf. MitoTEMPO (Sigma-Aldrich SML0737; dissolved in DMSO) was added at 10–100 µM 30 min prior to a 1 hr exposure to 10 µM neomycin. To stimulate IGF1 signaling: larvae were pre-treated with either NBI-31772 at 30–100 µM (Fisher Scientific 519210; dissolved in DMSO), or recombinant IGF1 at 10–100 ng/mL (Cell Sciences DU100; dissolved in 10 µM HCl and diluted in 0.1 mg BSA in E3) for 24 hr prior (beginning at four dpf) and then exposed to 10 µM neomycin for 1 hr on five dpf. To inhibit IGF1 signaling, larvae were treated with NVP-AEW541 at 30 µM (Selleck S1034; dissolved in DMSO) for 24 hr (beginning at four dpf), before live imaging. Following each treatment period, larvae were washed 3 times with E3 and left to recover in E3 for 4 hr at 28°C before fixation with 4% paraformaldehyde (diluted to 4% w/v in PBS from 16% w/v in 0.1M phosphate buffer, pH 7.4). For mitoTEMPO, NBI-31772, and IGF1 treatments, the compounds were re-added to the E3 media for the 4 hr recovery period post neomycin washout. Vehicle-treated controls were exposed to either 0.9% sodium chloride in E3 (neomycin control), 0.1 mg BSA in E3 (IGF1 control) or 0.1% DMSO in E3 for the other compounds.

Hair cell survival

Request a detailed protocol

Hair cell survival experiments were performed in Tg(brn3c:GFP)), TLF, Tg(hsp70:dnIGF1R-GFP), or Tg(hsp70:pappaa-GFP) larvae where hair cells are marked by GFP (brn3c) or anti-myosin VI antibody (TLF, hsp70:dnIGF1Ra-GFP, and hsp70:pappaa-GFP). For each larva, hair cells were counted from the same three stereotypically positioned neuromasts (IO3, M2, and OP1) (Raible and Kruse, 2000) and averaged. The percent of surviving hair cells was calculated as: [(mean number of hair cells after treatment)/ (mean number of hair cells in vehicle treated group)] X 100. For the heat shock experiments, treatment groups were normalized to the non-heat-shocked, vehicle-treated group for each genotype. Normalizations were genotype specific to account for a slight increase in hair cell number (~2 per neuromast) in pappaap170 larvae at five dpf.

Induction of transgenic dominant negative IGF1Ra and Pappaa expression

Request a detailed protocol

To induce expression of a dominant negative form of IGF1Ra, we used Tg(hsp70:dnIGF1Ra-GFP) larvae, which express dnIGF1Ra-GFP under the control of zebrafish hsp70 promoter (Kamei et al., 2011). dnIGF1Ra-GFP expression was induced by a 1 hr heat shock at 37°C, which was performed once per 12 hr from either 24 hpf to five dpf or from 4 dpf to five dpf. To induce expression of Pappaa, Tg(hsp70:pappaa-GFP) larvae in the pappaap170 background were heat shocked once for 30 min at 37°C at four dpf. To control for possible effects of heat shocking, non-transgenic wild type and pappaap170 larvae were exposed to the same treatment.

Single cell dissociation and fluorescence activated cell sorting

Request a detailed protocol

For each genotype, 30 five dpf Tg(mnx1:GFP) and 200 five dpf Tg(brn3c:GFP) larvae were rinsed for 15 min in Ringer’s solution (116 mM NaCl, 2.9 mM KCl, 1.8 mM CaCl2, 5 mM HEPES, pH 7.2) (Guille, 1999). pappaap170 larvae were identified by lack of swim bladder inflation (Wolman et al., 2015). To collect motor neurons we used whole Tg(mnx1:GFP) larvae and to collect hair cells we used tails dissected from Tg(brn3c:GFP) larvae. Samples were pooled into 1.5 mL tubes containing Ringer’s solution on ice, which was then replaced with 1.3 mL of 0.25% trypsin-EDTA (Sigma-Aldrich) for digestion. Tg(mnx1:GFP) samples were incubated for 90 min and Tg(brn3c:GFP) were incubated for 20 min. Samples were titrated gently by P1000 pipette tip every 15 min for motor neurons and every 5 min for hair cells. To stop cell digestion, 200 µL of 30% FBS and 6 mM CaCl2 in PBS solution (Steiner et al., 2014) was added, cells were centrifuged at 400 g for 5 min at 4°C, the supernatant was removed, the cell pellet was rinsed with Ca2+-free Ringer’s solution and centrifuged again. The cell pellet was then resuspended in 1X Ca2+-free Ringer’s solution (116 mM NaCl, 2.9 mM KCl, 5 mM HEPES, pH 7.2) and kept on ice until sorting. Immediately before sorting, cells were filtered through a 40 µm cell strainer and stained with DAPI. A two-gates sorting strategy was employed. DAPI was used to isolate live cells, followed by a forward scatter (FSC) and GFP gate to isolate GFP+ cells. Sorted cells were collected into RNAse-free tubes containing 500 µL of TRIzol reagent (Invitrogen) for RNA extraction.

RNA extraction and RT-PCR

Request a detailed protocol

Total RNA was extracted from whole larvae and FACS sorted motor neurons and hair cells using TRIzol. cDNA was synthesized using SuperScript II Reverse Transcriptase (Invitrogen 18064014). Real-time Quantitative PCR (RT-qPCR) was performed using Sso fast Eva Green Supermix (Biorad 1725200) in a StepOnePlus Real-Time PCR System (Applied Biosystems) based on manufacture recommendation. Reactions were run in 3–4 technical replicates containing cDNA from 50 ng of total RNA/reaction. The primer sequences for the antioxidant genes were previously described (Jin et al., 2010) and are as follows: For sod1, forward: GTCGTCTGGCTTGTGGAGTG and reverse: TGTCAGCGGGCTAGTGCTT; for sod2, forward: CCGGACTATGTTAAGGCCATCT and reverse: ACACTCGGTTGCTCTCTTTTCTCT; for gpx, forward: AGATGTCATTCCTGCACACG and reverse: AAGGAGAAGCTTCCTCAGCC; for catalase, forward: AGGGCAACTGGGATCTTACA and reverse: TTTATGGGACCAGACCTTGG. b-actin was used as an endogenous control with the following primer sequences: forward TACAGCTTCACCACCACAGC and reverse: AAGGAAGGCTGGAAGAGAGC (Wang et al., 2005). Cycling conditions were as follows: 1 min at 95°C, then 40 cycles of 15 s at 95°C, followed by 1 min at 60°C (Jin et al., 2010): Relative quantification of gene expression was done using the 2−ΔΔCt method (Livak and Schmittgen, 2001). PCR amplification for pappaa fragment was performed by using forward primer: AGACAGGGATGTGGAGTACG, and reverse primer: GTTGCAGACGACAGTACAGC. PCR conditions were as follows: 3 min at 94°C, followed by 40 cycles of 94°C for 30 s, 57°C for 1 min, and 70°C for 1 min (Wolman et al., 2015). The PCR product was run on a 3% agarose gel.

Live imaging

Request a detailed protocol

All experiments were done on 5–6 dpf pappaap170 and pappaa+/+ larvae at room temperature. Images were acquired with an Olympus Fluoview confocal laser scanning microscope (FV1000) using Fluoview software (FV10-ASW 4.2). To detect oxidative stress, Tg(brn3c:GFP) larvae were incubated in 10 µM CellROX Deep Red (Thermofischer Scientific C10422; dissolved in DMSO) and 1 µM mitoSOX Red (Thermofischer Scientific M36008; dissolved in DMSO) in E3 for 60 min and 30 min, respectively. To detect mitochondrial transmembrane potential, Tg(brn3c:GFP) larvae were incubated in 25 nM TMRE (Thermofischer Scientific T669; dissolved in DMSO) for 20 min. To investigate the effects of inhibiting IGF1 signaling on mitochondrial transmembrane potential, larvae were incubated in 25 nM TMRE for 20 min following treatment with NVP-AEW541. To detect MET channel function, Tg(brn3c:GFP) larvae were incubated in 3 µM FM1-43 (Thermofischer Scientific T3136; dissolved in DMSO) for 30 s. To measure mitochondrial mass, larvae were incubated in 100 nM mitotracker green FM (Thermofischer Scientific M7514; dissolved in DMSO) for 5 min. Following the incubation period, larvae were washed three times in E3, anesthetized in 0.002% tricaine (Sigma-Aldrich) in E3, and mounted as previously described (Stawicki et al., 2014). Fluorescent intensity of the reporter was measured using ImageJ (Schneider et al., 2012) by drawing a region of interest around brn3c:GFP-labaled hair cells of the neuromast from Z-stack summation projections that included the full depth of the hair cells. Background fluorescent intensity was measured by drawing a ROI away from the neuromast in the same Z-stack summation projection. The corrected total cell fluorescence (CTCF) was used to subtract background fluorescence from each reporter. The CTCF formula was as follows: Integrated Density - (Area of selected cells X Mean fluorescence of background) (McCloy et al., 2014). The mean CTCF of each live dye was reported both independently and as the ratio to the mean CTCF of GFP fluorescence.

Immunohistochemistry and in situ hybridization

Request a detailed protocol

For whole-mount immunostaining, larvae at five dpf were fixed in 4% paraformaldehyde for 1 hr at room temperature then rinsed three times with PBS. Larvae were blocked for 1 hr at room temperature in incubation buffer (0.2% bovine serum albumin, 2% normal goat serum, 0.8% Triton-X, 1% DMSO, in PBS, pH 7.4). Larvae were incubated in primary antibodies in IB overnight at 4°C. Primary antibodies were as follows: hair cells (anti-myosinVI, 1:200, rabbit polyclonal, Proteus biosciences, RRID:AB_10013626) or using Tg(brn3c:GFP) larvae (anti-GFP, 1:500, rabbit polyclonal; ThermoFisher Scientific, RRID:AB_221569), and support cells (anti-SOX2 ab97959, 1:200, rabbit polyclonal; Abcam, RRID:AB_2341193) (He et al., 2014). Following incubation of primary antibodies, larvae were incubated in AlexaFluor-488- conjugated secondary antibody in IB for 4 hr at room temperature. (goat anti-rabbit polyclonal, 1:500; ThermoFisher Scientific, RRID:AB_2576217). After staining, larvae were mounted in 70% glycerol in PBS. Images were acquired with an Olympus Fluoview confocal laser scanning microscope (FV1000) using Fluoview software (FV10-ASW 4.2).

For whole-mount in situ hybridization: digoxygenin-UTP-labeled antisense riboprobes for pappaa (Wolman et al., 2015) were used as previously described (Halloran et al., 1999; Chalasani et al., 2007). Images of colorimetric in situ reactions were acquired using a Leica Fluorescence stereo microscope with a Leica DFC310 FX digital color camera. Images of fluorescent in situ reactions were acquired using an Olympus Fluoview confocal laser scanning microscope (FV1000).

Statistics

All data were analyzed using GraphPad Prism Software 7.0b (GraphPad Software Incorporated, La Jolla, Ca, USA, RRID:SCR_002798). Prior to use of parametric statistics, the assumption of normality was tested using Shapiro-Wilk’s test. Parametric analyses were performed using a two-tailed unpaired t-test with Welch’s correction, multiple t tests with a Holm-Sidak correction, or 2-way ANOVA with a Holm-Sidak correction. Data are presented as means ± standard error of the mean (SEM; N = sample size). Significance was set at p<0.05. N for each experiment is detailed in the figure legends. All data presented are from individual experiments except for data in Figures 1B, 3B and 6E. Data collected from multiple experiments were normalized to their respective controls prior to pooling.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
    Glutathione peroxidases
    1. R Brigelius-Flohé
    2. M Maiorino
    (2013)
    Biochimica et Biophysica Acta (BBA) - General Subjects 1830:3289–3303.
    https://doi.org/10.1016/j.bbagen.2012.11.020
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
    Inhibition by cyclosporin A of a Ca2+-dependent pore in heart mitochondria activated by inorganic phosphate and oxidative stress
    1. M Crompton
    2. H Ellinger
    3. A Costi
    (1988)
    The Biochemical Journal 255:357–360.
  16. 16
  17. 17
  18. 18
    Hearing Loss: Causes, Prevention, and Treatment
    1. JJ Eggermont
    (2017)
    London, UK: Academic Press.
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
    Mitochondria and neuronal activity
    1. O Kann
    2. R Kovács
    (2007)
    American Journal of Physiology-Cell Physiology 292:C641–C657.
    https://doi.org/10.1152/ajpcell.00222.2006
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
    Mechanisms of Mitochondrial Free Radical Production and their Relationship to the Aging Process
    1. CL Quinlan
    2. JR Treberg
    3. MD Brand
    (2011)
    In: E. J Masoro, S. N Austad, editors. Handbook of the Biology of Aging (7). San Diego: Academic Press. pp. 47–61.
    https://doi.org/10.1016/b978-0-12-378638-8.00003-8
  67. 67
  68. 68
  69. 69
  70. 70
  71. 71
  72. 72
  73. 73
  74. 74
  75. 75
  76. 76
  77. 77
  78. 78
  79. 79
  80. 80
  81. 81
  82. 82
  83. 83
  84. 84
    Mitochondrial superoxide simutase. site of synthesis and intramitochondrial localization
    1. RA Weisiger
    2. I Fridovich
    (1973)
    The Journal of Biological Chemistry 248:4793–4796.
  85. 85
  86. 86
  87. 87
  88. 88
  89. 89
  90. 90
  91. 91

Decision letter

  1. David W Raible
    Reviewing Editor; University of Washington, United States
  2. Didier Y Stainier
    Senior Editor; Max Planck Institute for Heart and Lung Research, Germany

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "Pregnancy-associated plasma protein-aa promotes neuron survival by regulating mitochondrial function" for consideration by eLife. Your article has been reviewed by a Senior Editor, a Reviewing Editor, and three reviewers. The following individuals involved in review of your submission have agreed to reveal their identity: David Raible (Reviewer #1).

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered for publication in eLife at this time. While there was general enthusiasm for the studies, there were a number of concerns raised by the reviewers that we feel would not be able to be addressed in the standard period given for revision. However, given the overall enthusiasm we would consider a revised version of the manuscript if you can meet the referees' concerns. Please note that this would be considered a new submission.

Reviewer #1:

Alassaf et al., examine the effects of mutation in the IGF signaling regulator Pappaa on hair cell survival after exposure to neomycin. They find that pappaa mutants are more sensitive to damage. They present evidence that hair cells in mutants have higher levels of ROS production, increases in mitochondrial polarization and increases in mitochondrial calcium, all previously implicated in neomycin-induced hair cell death. They also present evidence that manipulating IGF signaling alters the effects of the pappaa mutation on hair cell survival. Finally, they show that pappaa mutants lose motor neurons over time. Several issues need to be addressed before the manuscript could be considered for publication.

The authors propose that IGF signaling is important for hair cell survival but do not test this directly. They have used the dn-igfr transgenic line in previous publications – does this method of inhibiting IGF signaling also alter hair cell survival? Does the dn-igfr or the IGF antagonist NVP-AEW541 shift the neomycin dose-response curve? Does SC79 and IGF-1 protect wildtype hair cells against neomycin?

Other points:

The authors should refer to hair cells in the abstract rather than describing them as neurons.

The authors suggest that the failure to add new hair cells is because of degeneration (Figure 1). However, growth of the lateral line depends on the ability of larvae to feed. Are mutants able to feed and grow normally? If not, this phenotype is likely indirect.

The analysis to determine whether Pappaa is expressed in hair cells (Figure 2) is inconclusive. Expression centrally in neuromasts (2A) usually corresponds to hair cells. Co-localization with GFP is inconclusive (2B). There is no direct comparison between hair cells and support cells by RT-PCR. Comparison to whole larva is not relevant.

Experiments shown in 3G are inconclusive. Previous work has shown that mitoTEMPO protects hair cells from neomycin exposure in wildtype animals. Are pappaa mutants differentially sensitive to this treatment?

Additional experiments are needed for the analysis of IGFR signaling (Figure 5A). The specificity of the phospho-IGF1R staining should be tested. This staining should be sensitive to NVP-AEW541. It is also important to determine whether staining is actually in hair cells.

Controls are missing from pharmacological experiments (Figure 5). The effect of NVP-AEW541 on hair cells in the absence of neomycin should be shown (5D), and conditions repeated on pappaa mutants. The effect of SC79 (5E) and IGF-1 (5F) on wildtype animals should be shown as discussed above. In particular the authors should test whether these reagents are differentially effective in wildtype and pappaa mutant embryos.

Reviewer #2:

The topic of this study is of great interest to both auditory neuroscientists and to the broader neuroscience community. With that said, hair cells, while they share properties with neurons, are highly specialized and have their own distinct properties. As the data on mitochondrial function and the contribution of IGF1R signaling to survival were collected from hair cells, the emphasis on "neurons" in place of "hair cells" in the title and the manuscript is misleading. Additionally, I have major concerns about how the authors performed and analyzed the live imaging experiments.

Point by point concerns are addressed below:

- Naturally occurring cell death (Figure 1): Couldn't this observation also indicate a failure to proliferate?

- FM1-43 uptake (Figure 3 A-B): The dye was diluted with DMSO, which confounds the results as DMSO could be facilitating entry into hair cells.

- Image analysis (Figure 3 B,D,F; Figure 4 C,D; Figure 5C): The method used for analyzing differences in the fluorescence of indicators relative to brn3c:GFP fluorescence (F indicator/ FGFP) is unconventional. I'm concerned that using GFP fluorescence to normalize for variable neuromast size would give misleading results; variances in hair-cell GFP fluorescence appears completely independent of variances in indicator fluorescence (see Figure 3E for an example). Is this a technique that has been previously used? If not, the authors need to clearly explain their rationale. Additionally, the authors need to provide more detail about how their images were processed and analyzed in ImageJ (e.g. were z-stack projections used? how was background fluorescence subtracted? which specific plugins were used for the analysis?)

- Subsection “Pappaa regulates mitochondrial Ca2+ uptake and transmembrane potential”: replace "a doubling" with " ~2-fold greater".

- Subsection “Pappaa-IGF1 receptor signaling is required for neuron survival”: "based on their numbers" – the observation that there are more hair cells in pappaap170 mutants relative to their siblings could indicate that the neuromasts are not developing normally.

- Subsection “Pappaa-IGF1 receptor signaling is required for neuron survival”: "regulation of mitochondrial function in neurons" – the previous paragraph in the discussion suggests, since Pappaa is expressed in spinal motor neurons but not hair cells, that Pappaa may promote motor neuron survival via mechanisms that differ from hair cells. How do the authors extrapolate that the evidence they show for mitochondrial dysfunction contributing to hair-cell loss in the mutant also applies to loss of neurons?

- Pharmacology: 1% DMSO seems unusually high (usually 5 to 10x less is used as a carrier for free-swimming larvae). Why did the authors choose that concentration?

Reviewer #3:

The manuscript "Pregnancy-associated plasma protein-aa promotes neuron survival by regulating mitochondrial function" is a thorough assessment of both hair cell and neuron presence and measures of mitochondrial health in vivo with loss of the extracellular protein Pappaa using the zebrafish system. Using in vivo imaging of mitochondrial/cytoplasmic ROS, mitochondrial calcium, and mitochondrial membrane potential, the authors demonstrate that loss of Pappaa leads to alterations in these measures of mitochondrial health. These abnormal mitochondrial measures correlate with loss of neurons and hair cells in older larvae as well as increased susceptibility of hair cells to calcium and neomycin-mediated death. Strikingly, ROS reduction decreases hair cell loss in pappaa mutants, confirming the relationship between mitochondrial ROS and hair cell loss. Overall, this is a thorough description of a novel role for the extracellular factor Pappaa in hair cell and neuron presence. The manuscript is well written and flows nicely. The data presented are solid and offer potential new insights into the role of IGF in maintenance of neurons and hair cells in vivo. However, a few additional experiments would solidify this role, specifically making the connection between cell survival and pappaa mutation and addressing the known role of IGF in mitochondrial biology. Suggestions for modifications to strengthen this work to make it suitable for publication are below.

Essential revisions:

1) The authors argue that loss of Pappaa function causes decreased numbers of hair cell and neurons; however, while number of hair cell and motor neurons is decreased by 9 dpf, no measures of cell death in the absence of chemical ablation are shown. Additionally, no measures of proliferation are shown. To clarify the basic function of Pappaa in neuron presence, a measure of proliferation for hair cells and motor neurons as well as a measure of cell death in these populations should be shown. Numerous methods exist to assay these cellular processes in vivo in zebrafish larvae.

2) IGF signaling has been implicated in the regulation of mitochondrial dynamics, mitophagy and mitochondrial biogenesis. Disrupting these processes could lead to mitochondrial stress, which would then lead to the calcium level, ROS, and matrix potential defects observed in pappaa mutants. Essentially sensitizing the system. To determine if defects in these aspects of mitochondrial biology are the primary drivers of the hair cell and neuron loss, the authors should look at mitochondrial size, mitochondrial load, and mitochondrial localization using either a transient transgenic approach or mitotracker.

3) The majority of the work presented was done in hair cells and the results are then assumed to represent what is also happening in motor neurons. To determine if this is indeed the case, at least a subset of the measures of mitochondrial calcium, potential, and ROS should be done in motor neurons of pappaa mutants. Additionally, the ability of exogenous IGF or AKT to rescue motor neurons should be shown to confirm or refute their similarity to hair cells.

4) In subsection “Pappaa-IGF1 receptor signaling is required for neuron survival” the authors argue that hair cells and motor neurons develop normally to day 5, indicating a role for Pappaa in maintenance rather than development; however, it is not shown if Pappaa is maternally expressed and, if it is, how long the protein lasts. This would be essential for this argument as it is entirely possible that maternal Pappaa allows for normal development rather than Pappaa not being necessary for development. Alternatively, a maternal zygotic mutant could also answer this question.

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "Pregnancy-associated plasma protein-aa supports hair cell survival by regulating mitochondrial function" for consideration by eLife. Your article has been reviewed by Didier Stainier as the Senior Editor, a Reviewing Editor, and two reviewers. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

The revised manuscript "Pregnancy-associated plasma protein-aa supports hair cell survival by regulating mitochondrial function" carefully describes the function of IGF signaling in hair cell resistance to cell death. The authors have convincingly shown that loss of Pappaa results in increased susceptibility to neomycin-mediated hair cell death and that this is likely through increased susceptibility to mitochondrial stress pathways. Additionally, the authors have strengthened the work by showing that Pappaa functions in the IGF availability pathway through using the IGFBP inhibitor NBI-31772. Together, their work argues for a role for IGF signaling in the maintenance of hair cell survival through supporting mitochondria/inhibiting mitochondrial ROS stress.

Essential revisions:

1) The reviewers again raised concerns about the numbers of replicates, with no clear way to know how many times experiments were performed. In addition, experiments appear to be often underpowered with small n's. These are of particular concern for imaging experiments presented in Figure 5 and Figure 6. The authors need to clearly state the numbers of samples used for calculations, the numbers of experiments performed, and provide justification.

2) In a previous version of this manuscript, exogenous IGF was shown to suppress neomycin-mediated hair cell death. It is unclear why this was taken out as it clearly compliments Figure 3C which shows that enhancing IGF signaling can suppress hair cell death. It should at least be included in the supplement unless there is a specific reason to exclude it.

3) Throughout the manuscript, the authors use "% surviving cells" as a measure of hair cell survival. This implies that they counted the hair cells in the same neuromasts before and after treatment. Their methods state that they normalized number of cells after treatment to average number of cells in the vehicle treated control. While this is a suitable way to control for the data, this should be spelled out in the Results section rather than hidden in the Materials and methods section.

1) Subsection “Pappaa regulates survival of hair cell sensory neurons”, the authors state: "This live probe (re:TMRE) emits more fluorescence as the membrane potential becomes more negative." This is not true. This vital cationic dye just accumulates to a higher degree in mitochondria that have a more negative matrix potential. The dye does not change the its emission profile to my knowledge. The language should be changed to reflect this.

2) In the revised manuscript, the authors provide "absolute fluorescence levels" and refer to these articles (Ogun and Zallocchi, 2014; Majumder et al., 2017). Both studies quantified relative FM1-43 fluorescence by performing background subtraction of fluorescence, selecting the region of interest (e.g. neuromast), and measuring the average fluorescent intensity i.e. total fluorescent intensity/area. If this method is what the authors also used, they should update their Material and Methods section and add citations accordingly.

https://doi.org/10.7554/eLife.47061.032

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

Reviewer #1:

[…] The authors propose that IGF signaling is important for hair cell survival but do not test this directly. They have used the dn-igfr transgenic line in previous publications – does this method of inhibiting IGF signaling also alter hair cell survival? Does the dn-igfr or the IGF antagonist NVP-AEW541 shift the neomycin dose-response curve? Does SC79 and IGF-1 protect wildtype hair cells against neomycin?

Per this suggestion, we used the hsp70:dnIGF1Ra-GFP transgenic line to reduce IGF1R signaling and ask whether IGF1 signaling mediates hair cell survival. As shown in Figure 1B and Figure 4C, we found that induction of dnIGF1Ra increases hair cell death in larvae exposed to neomycin and therefore shifts the neomycin dose-response curve. We have excluded the IGF-1 and SC79 supplementation experiments from the manuscript because we wanted to focus mainly on the influence of endogenous IGF1 signaling and the pathway’s extracellular regulation (Pappaa, Figure 3C), rather than exogenous supplementation by IGF1 and SC79.

The authors should refer to hair cells in the Abstract rather than describing them as neurons.

The manuscript has been edited to reflect these changes and now solely focuses on the effect of Pappaa-IGF1 signaling on hair cells.

The authors suggest that the failure to add new hair cells is because of degeneration (Figure 1). However, growth of the lateral line depends on the ability of larvae to feed. Are mutants able to feed and grow normally? If not, this phenotype is likely indirect.

We thank the reviewer for calling our attention to this important point. pappaa mutants do show reduced feeding. Therefore, we have excluded this result and its interpretation from the manuscript.

The analysis to determine whether pappaa is expressed in hair cells (Figure 2) is inconclusive. Expression centrally in neuromasts (2A) usually corresponds to hair cells. Co-localization with GFP is inconclusive (2B). There is no direct comparison between hair cells and support cells by RT-PCR. Comparison to whole larva is not relevant.

We have modified Figure 2 to provide more conclusive expression data. The colorimetric pappaa in situ expression shows clear expression by the neuromasts. We believe the expression centrally in each neuromast is by the inner support cells, not the hair cells. By analyzing Z-stack confocal images slice by slice in Brn3c-GFP larvae stained for pappaa expression by FISH, we do not observe pappaa mRNA in hair cells. We performed the RT-PCR of pappaa on FAC sorted hair cells to rule out the possibility that hair cells express pappaa, but at a level below FISH detection. RT-PCR on whole larvae and FAC sorted motor neurons provided positive controls. Unfortunately, we do not have a transgenic line that only labels support cells to use in a complementary experiment.

Experiments shown in 3G are inconclusive. Previous work has shown that mitoTEMPO protects hair cells from neomycin exposure in wildtype animals. Are pappaa mutants differentially sensitive to this treatment?

Results of the mitoTEMPO experiment are now in Figure 7B. Consistent with previous work (Esterberg et al., 2016), we find mitoTEMPO protects hair cells from neomycin exposure in wildtype and pappaa mutants. It is difficult to determine relative sensitivity of the two genotypes to mitoTEMPO because mitoTEMPO produces an all or nothing survival effect in both genotypes at doses ranging from 10-100 μM. Rather than assessing relative sensitivity, the rationale behind our experiment was to determine whether the excessive ROS production by pappaa mutant hair cells was the root cause of their increased death when exposed to neomycin. If mitoTEMPO did not or only partially rescued hair cell loss in pappaa mutants, then this outcome would have suggested the mutant hair cells have additional defects that underlie their increased neomycin sensitivity. Given that mitoTEMPO provided up to complete protection against neomycin, we concluded that ROS was the primary reason.

Additional experiments are needed for the analysis of IGFR signaling (Figure 5A). The specificity of the phospho-IGF1R staining should be tested. This staining should be sensitive to NVP-AEW541. It is also important to determine whether staining is actually in hair cells.

The anti-pIGF1R antibody used was validated by Chablais and Jazwinska (2010) in zebrafish using NVP-AEW541. However, we decided to exclude the result (reduced p-IGF1R labeling in pappaa mutant hair cells) because it does not significantly add to the manuscript.

Controls are missing from pharmacological experiments (Figure 5). The effect of NVP-AEW541 on hair cells in the absence of neomycin should be shown (5D), and conditions repeated on pappaa mutants. The effect of SC79 (5E) and IGF-1 (5F) on wildtype animals should be shown as discussed above. In particular the authors should test whether these reagents are differentially effective in wildtype and pappaa mutant embryos.

We did not observe hair cell death when larvae were treated with NVP-AEW541 alone. That said, these controls and experiments are now moot, since we have decided against including the results in the manuscript. As detailed above, we instead used the hsp70:dnIGF1Ra-GFP transgenic line to reduce IGF1R signaling and ask whether IGF1 signaling mediates hair cell survival (Figure 1B and Figure 4C). Notably, expression of dnIGF1Ra alone (no neomycin) did not affect hair cell survival. Because we decided to focus more on the extracellular regulation of endogenous IGF1 signaling rather than exogenous supplementation of IGF1 and SC79, we included results from pharmacological stimulation of IGF1 bioavailability by NBI-31772 (Figure 3C).

Reviewer #2:

[…] Point by point concerns are addressed below:

- Naturally occurring cell death (Figure 1): Couldn't this observation also indicate a failure to proliferate?

The hair cells in pappaa mutants develop on time and are capable of regeneration. Therefore, we do not believe that a failure to proliferate is causing the lack of age-related increase in hair cells. However, as reviewer 1 pointed out, growth of lateral line hair cells depends on the larvae’s ability to grow and feed normally. Given that pappaa mutant larvae show reduced ability to feed, we have decided to exclude this result and interpretation from the current manuscript.

- FM1-43 uptake (Figure 3 A-B): The dye was diluted with DMSO, which confounds the results as DMSO could be facilitating entry into hair cells.

FM1-43 was dissolved in DMSO, as per manufacturer recommendation. We used FM1-43 at a low final concentration (3uM) made from a 5 mM stock in DMSO. The final concentration of DMSO would have been at 0.06%. This method has been used previously to demonstrate that FM1-43 (dissolved in DMSO) only enters hair cells via MET channels when exposure time is less than 1 minute (Gale et al., 2001).

- Image analysis (Figure 3 B,D,F; Figure 4 C,D; Figure 5C): The method used for analyzing differences in the fluorescence of indicators relative to brn3c:GFP fluorescence (F indicator/ FGFP) is unconventional. I'm concerned that using GFP fluorescence to normalize for variable neuromast size would give misleading results; variances in hair-cell GFP fluorescence appears completely independent of variances in indicator fluorescence (see Figure 3E for an example). Is this a technique that has been previously used? If not, the authors need to clearly explain their rationale.

To provide further analysis, we have now included the absolute fluorescence levels. This follows the field standard (Ogun and Zallocchi, 2014; Majumder et al., 2017). Additionally, we have included the normalized fluorescence (F indicator/F GFP), which we feel is important to control for potential variability in labeling across samples.

Additionally, the authors need to provide more detail about how their images were processed and analyzed in ImageJ (e.g. were z-stack projections used? how was background fluorescence subtracted? which specific plugins were used for the analysis?)

We have provided more detail on the image processing in subsection “Live imaging”. We have added the text “Fluorescent intensity of the reporter was measured using ImageJ (Schneider et al., 2012) by drawing a region of interest around brn3c:GFP-labaled hair cell cluster of the neuromast from Z-stack summation projections that included the full depth of the hair cells. Background fluorescent intensity was measured by drawing a ROI away from the neuromast in the same Z-stack summation projection. The corrected total cell fluorescence (CTCF) was used to subtract background fluorescence from each reporter. The CTCF formula was as follows: Integrated Density – (Area of selected cells X Mean fluorescence of background) (McCloy et al., 2014). Relative fluorescent intensity was reported as the ratio to GFP fluorescence.” No specific plugins were used.

- Subsection “Pappaa regulates mitochondrial Ca2+ uptake and transmembrane potential”: replace "a doubling" with " ~2-fold greater".

This change has been made.

- Subsection “Pappaa-IGF1 receptor signaling is required for neuron survival”: "based on their numbers" – the observation that there are more hair cells in pappaap170 mutants relative to their siblings could indicate that the neuromasts are not developing normally.

We do not believe pappaa affects neuromast development because (1) the hair cell and support cell morphologies are normal (Figure 4A and Figure 3—figure supplement 1B), (2) the mutants respond to acoustic stimuli and do not show hallmark behavioral signs of dysfunctional hair cells (Wolman et al., 2015) and (3) we show that Pappaa is dispensable for hair cell development in the context of hair cell sensitivity to neomycin (Figure 3A-B). The increased number of hair cells in neuromasts of pappaa mutants is very modest: ~2 per neuromast. If we did not account for this difference and used absolute hair cell counts to reflect mortality, then the hair cell loss would appear artificially inflated in the pappaa mutants compared to wild type controls. Therefore, we normalized hair cell loss to vehicle treated control wild type or pappaa mutants to most accurately represent and compare the relative hair cell loss between the two genotypes.

- Subsection “Pappaa-IGF1 receptor signaling is required for neuron survival”: "regulation of mitochondrial function in neurons" – the previous paragraph in the discussion suggests, since Pappaa is expressed in spinal motor neurons but not hair cells, that Pappaa may promote motor neuron survival via mechanisms that differ from hair cells. How do the authors extrapolate that the evidence they show for mitochondrial dysfunction contributing to hair-cell loss in the mutant also applies to loss of neurons?

The current manuscript no longer includes results or discussion of Pappaa’s influence on motor neuron survival. Therefore, this section has been omitted.

- Pharmacology: 1% DMSO seems unusually high (usually 5 to 10x less is used as a carrier for free-swimming larvae). Why did the authors choose that concentration?

“1% DMSO” was a typo. Changes were made to reflect the real concentration (0.1%).

Reviewer #3:

[…] Essential revisions:

1) The authors argue that loss of Pappaa function causes decreased numbers of hair cell and neurons; however, while number of hair cell and motor neurons is decreased by 9 dpf, no measures of cell death in the absence of chemical ablation are shown. Additionally, no measures of proliferation are shown. To clarify the basic function of Pappaa in neuron presence, a measure of proliferation for hair cells and motor neurons as well as a measure of cell death in these populations should be shown. Numerous methods exist to assay these cellular processes in vivo in zebrafish larvae.

Because hair cells in pappaa mutants develop on time and are capable of regeneration, we do not believe that failure to proliferate is causing the lack of age-related increase in hair cells. However, as reviewer 1 has pointed out, growth of lateral line hair cells depends on the larvae’s ability to grow and feed normally. Given that pappaa mutant larvae show reduced ability to feed, we have decided to exclude this experiment from our manuscript. We also decided to focus solely on hair cells. Results describing motor neuron degeneration have been excluded from the revised manuscript.

2) IGF signaling has been implicated in the regulation of mitochondrial dynamics, mitophagy and mitochondrial biogenesis. Disrupting these processes could lead to mitochondrial stress which would then lead to the calcium level, ROS, and matrix potential defects observed in pappaa mutants. Essentially sensitizing the system. To determine if defects in these aspects of mitochondrial biology are the primary drivers of the hair cell and neuron loss, the authors should look at mitochondrial size, mitochondrial load, and mitochondrial localization using either a transient transgenic approach or mitotracker.

We analyzed mitochondrial load using mitotracker and found no significant difference (Figure 5—figure supplement 1). A future direction of this project is to evaluate mitochondrial size, morphology, and localization in pappaa mutants by electron microscopy. We believe results from these analyses are beyond the scope of the current study.

3) The majority of the work presented was done in hair cells and the results are then assumed to represent what is also happening in motor neurons. To determine if this is indeed the case, at least a subset of the measures of mitochondrial calcium, potential, and ROS should be done in motor neurons of pappaa mutants. Additionally, the ability of exogenous IGF or AKT to rescue motor neurons should be shown to confirm or refute their similarity to hair cells.

In the manuscript’s revised form, we only show results for hair cells and we have limited our conclusions to hair cells.

4) In subsection “Pappaa-IGF1 receptor signaling is required for neuron survival” the authors argue that hair cells and motor neurons develop normally to day 5, indicating a role for pappaa in maintenance rather than development; however, it is not shown if pappaa is maternally expressed and, if it is, how long the protein lasts. This would be essential for this argument as it is entirely possible that maternal pappaa allows for normal development rather than pappaa not being necessary for development. Alternatively, a maternal zygotic mutant could also answer this question.

In our revised manuscript, we have included an experiment in which we induce Pappaa expression at 4 dpf (Figure 4B). We found that post-developmental expression of Pappaa is sufficient to rescue pappaa mutant hair cell sensitivity to neomycin.

[Editors' note: the author responses to the re-review follow.]

[…] Essential revisions:

1) The reviewers again raised concerns about the numbers of replicates, with no clear way to know how many times experiments were performed. In addition, experiments appear to be often underpowered with small n's. These are of particular concern for imaging experiments presented in Figure 5 and Figure 6. The authors need to clearly state the numbers of samples used for calculations, the numbers of experiments performed, and provide justification.

We have added more details concerning the specific number of neuromasts and animals for each treatment group in the figures and in the figure legends. Please note that the N in this paper refers to the number of animals per group (shown at base of bars) and not neuromasts. Multiple neuromasts per animal were analyzed and averaged. To address the number of experimental trials, we’ve added the following in the methods “All data presented are from individual experiments except for data in Figure 1B, Figure 3B, and Figure 6E. Data collected from multiple experiments were normalized to their respective controls prior to pooling”. We also included the number of experiments in the relevant figure legends.

Because each live imaging experiment was done on the same day to control for any technical variabilities, we were limited in the number of animals tested. Please note that multiple neuromasts were analyzed per animal resulting in an average of 20 neuromasts per genotype (or treatment group). However, we do agree that some experiments had a particularly small number of animals tested (4 or fewer). Therefore, we have repeated experiments presented in Figure 5—figure supplement 1 and Figure 6B and 6D-D’ to include a bigger sample size.

2) In a previous version of this manuscript, exogenous IGF was shown to suppress neomycin-mediated hair cell death. It is unclear why this was taken out as it clearly compliments Figure 3C which shows that enhancing IGF signaling can suppress hair cell death. It should at least be included in the supplement unless there is a specific reason to exclude it.

Thank you for the suggestion. We’ve incorporated the exogenous IGF1 experiment to the revised manuscript (now figure 3C).

3) Throughout the manuscript, the authors use "% surviving cells" as a measure of hair cell survival. This implies that they counted the hair cells in the same neuromasts before and after treatment. Their methods state that they normalized number of cells after treatment to average number of cells in the vehicle treated control. While this is a suitable way to control for the data, this should be spelled out in the Results section rather than hidden in the Materials and methods section.

We have changed the Y-axis titles of relevant graphs to better clarify this.

https://doi.org/10.7554/eLife.47061.033

Article and author information

Author details

  1. Mroj Alassaf

    1. Department of Integrative Biology, University of Wisconsin, Madison, United States
    2. Neuroscience Training Program, University of Wisconsin, Madison, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Writing—original draft
    Competing interests
    No competing interests declared
  2. Emily C Daykin

    Department of Integrative Biology, University of Wisconsin, Madison, United States
    Contribution
    Data curation
    Competing interests
    No competing interests declared
  3. Jaffna Mathiaparanam

    Department of Integrative Biology, University of Wisconsin, Madison, United States
    Contribution
    Resources, Methodology
    Competing interests
    No competing interests declared
  4. Marc A Wolman

    Department of Integrative Biology, University of Wisconsin, Madison, United States
    Contribution
    Conceptualization, Resources, Funding acquisition, Project administration, Writing—review and editing
    For correspondence
    mawolman@wisc.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8929-779X

Funding

Greater Milwaukee Foundation

  • Marc A Wolman

Saudi Ministry of Education

  • Mroj Alassaf

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

The authors would like to thank Dr. David Raible (University of Washington-Seattle) for the myo6b:mitoGCaMP3 fish line and Dr. Corinna Burger (University of Wisconsin Department of Neurology) for use of the RT-qPCR cycler.

Ethics

Animal experimentation: This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocols (L00457-A1) of the University of Wisconsin.

Senior Editor

  1. Didier Y Stainier, Max Planck Institute for Heart and Lung Research, Germany

Reviewing Editor

  1. David W Raible, University of Washington, United States

Publication history

  1. Received: March 22, 2019
  2. Accepted: June 14, 2019
  3. Accepted Manuscript published: June 17, 2019 (version 1)
  4. Version of Record published: June 26, 2019 (version 2)

Copyright

© 2019, Alassaf et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 615
    Page views
  • 93
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Evolutionary Biology
    Omaya Dudin et al.
    Research Article Updated
    1. Cell Biology
    2. Chromosomes and Gene Expression
    Leah F Rosin et al.
    Research Article Updated