Berbamine Analogs Exhibit Differential Protective Effects From Aminoglycoside-Induced Hair Cell Death

ORIGINAL RESEARCH

published: 29 July 2020

doi: 10.3389/fncel.2020.00234

Berbamine Analogs Exhibit

Differential Protective Effects From

Aminoglycoside-Induced Hair

Cell Death

Alexandria M. Hudson¹, Gavin M. Lockard^2†, Ojas A. Namjosh i³, Joseph W. Wilson³,

Katie S. Kind t⁴, Bruce E. Bloug h³and Allison B. Cofﬁ n^1,2

*

¹Integrative Physiology and Neuroscience, Washington State University, Vancouver, WA, United States, ²College of Arts and

Sciences, Washington State University, Vancouver, WA, United States, ³RTI International, Research Triangle Park, NC,

United States, ⁴National Institute on Deafness and Other Communication Disorders, National Institutes of Health, Bethesda,

MD, United States

Hearing loss is the third most common chronic health condition in the United States

and largely results from damage to sensory hair cells. Major causes of hair cell damage

include aging, noise exposure, and medications such as aminoglycoside antibiotics. Due

to their potent antibacterial properties and low cost, aminoglycosides are often used for

the treatment of gram-negative bacterial infections, surpassing expensive antibiotics with

fewer harmful side effects. However, their use is coupled with permanent hearing loss

in over 20% of patients requiring these life-sustaining antibiotics. There are currently

no FDA-approved drugs that prevent hearing loss from aminoglycosides. A previous

study by our group identiﬁed the plant alkaloid berbamine as a strong protectant

of zebraﬁsh lateral line hair cells from aminoglycoside damage. This effect is likely

due to a block of the mechanotransduction channel, thereby reducing aminoglycoside

entry into hair cells. The present study builds on this previous work, investigating 16

synthetic berbamine analogs to determine the core structure underlying their protective

mechanisms. We demonstrate that nearly all of these berbamine analogs robustly protect

lateral line hair cells from ototoxic damage, with ED₅₀values nearing 20 nM for the most

potent analogs. Of the 16 analogs tested, nine strongly protected hair cells from both

neomycin and gentamicin damage, while one conferred strong protection only from

gentamicin. These data are consistent with prior research demonstrating that different

aminoglycosides activate somewhat distinct mechanisms of damage. Regardless of the

mechanism, protection required the entire berbamine scaffold. Phenolic alkylation or

acylation with lipophilic groups appeared to improve protection compared to berbamine,

implying that these structures may be responsible for mitigating damage. While the

majority of analogs confer protection by blocking aminoglycoside uptake, 18% of our

analogs also confer protection via an uptake-independent mechanism; these analogs

Edited by:

Michael E. Smith,

Western Kentucky University,

United States

Reviewed by:

Alan G. Cheng,

Stanford University, United States

Emma J. Kenyon,

University of Sussex, United Kingdom

*Correspondence:

Allison B. Cofﬁn

allison.cofﬁn@wsu.edu

^†Present address:

Gavin M. Lockard,

College of Pharmacy, Washington

State University, Spokane, WA,

United States

Specialty section:

This article was submitted to

Cellular Neuropathology, a section of

the journal

Frontiers in Cellular Neuroscience

Received: 09 April 2020

Accepted: 02 July 2020

Published: 29 July 2020

Citation:

Hudson AM, Lockard GM,

Namjoshi OA, Wilson JW, Kindt KS,

Blough BE and Cofﬁn AB

1

(2020) Berbamine Analogs Exhibit

Differential Protective Effects From

Aminoglycoside-Induced

Abbreviations: MET, mechanotransduction; H NMR, nuclear magnetic resonance spectra for proton; TLC, thin-layer

chromatography; EM, E2 embryo medium; dpf, days post-fertilization; DMSO, dimethylsulfoxide; DASPEI, 2-[4-

(dimethylamino)styryl]-N-ethylpyridinium iodide; MS-222, Tricaine methanesulfonate; GTTR, gentamicin conjugated with

Texas Red; OPC, optimal protective concentration; TRP channels, transient receptor potential channels; NMDA, N-methyl-

D-aspartate; AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; NF-κB, nuclear factor κ-light-chain-enhancer

of activated B cells; JNK, c-Jun N-terminal kinase pathways; ERK, extracellular receptor kinase.

Hair Cell Death.

Front. Cell. Neurosci. 14:234.

doi: 10.3389/fncel.2020.00234

Frontiers in Cellular Neuroscience | www.frontiersin.org

1

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

exhibited protection when delivered after aminoglycoside removal. Based on our studies,

berbamine analogs represent a promising tool to further understand the pathology of

aminoglycoside-induced hearing loss and can serve as lead compounds to develop

otoprotective drugs.

Keywords: hair cell, aminoglycoside, zebraﬁsh, lateral line, berbamine, ototoxicity, hearing loss,

mechanotransduction

INTRODUCTION

PROTO-1 and PROTO-2, both of which protected hair cells

from neomycin toxicity (Owens et al., 2008). Optimization of

Sensory hair cells can be damaged from aging, intense noise

exposure, and medications such as aminoglycoside antibiotics.

In mammals, damaged hair cells are not replaced, which can

result in permanent hearing loss. In humans, hearing loss

can be devastating since it can lead to social isolation and

decreased employment opportunities (Raviv et al., 2010; Jung

and Bhattacharyya, 2012; Bainbridge and Wallhagen, 2014; Mick

et al., 2014 ). Despite the adverse effects of aminoglycosides on

hearing, they are widely used in developing nations due to

their potent antibacterial properties and low cost, surpassing

expensive antibiotics with fewer harmful side effects. Currently,

there are no FDA-approved drugs that prevent aminoglycoside-

induced hearing loss. Our work seeks to develop a drug therapy

that robustly prevents aminoglycoside-induced hearing loss

by modifying the chemical scaffold of a naturally occurring

otoprotective compound.

There is strong evidence supporting mechanoelectrical

transduction (MET) as the primary path for aminoglycosides to

enter hair cells. The MET channels in hair cells are non-selective

cation channels with a pore diameter of 1.25 nm. These features

enable relatively large positively charged aminoglycosides to

permeate the channel (Farris et al., 2004). Additionally, there is

a membrane potential difference between the extracellular fluid

and the negatively polarized cytoplasm that increases cellular

uptake of positively charged aminoglycosides (Marcotti et al.,

2005; Myrdal and Steyger, 2005). Pharmacologically blocking

the MET channel dramatically decreases aminoglycoside uptake

(Marcotti et al., 2005) further supporting antibiotic entry via this

route. In addition to MET channels, there are also secondary

entry routes occurring via endocytosis or through other ion

channels (Portmann et al., 1974; Myrdal and Steyger, 2005;

Karasawa et al., 2008; Hailey et al., 2017 ). The current hypothesis

surrounding entry via endocytosis is that aminoglycosides are

initially sequestered by endosomes, then trafficked to lysosomes,

but different aminoglycosides (e.g., neomycin vs. gentamicin)

differ in their rates of uptake into subcellular compartments.

These data imply that sequestration of aminoglycosides in

lysosomes could potentially attenuate hair cell damage (Hailey

et al., 2017). Regardless of the entry route, aminoglycosides

accumulate in hair cells, leading to pathological consequences.

In light of our understanding of the mechanisms of

aminoglycoside toxicity, new targets for protection are arising.

Given that the MET channel is the primary entry route for

aminoglycosides, one option for protection is to block entry

of aminoglycosides through the channel. Prior work using a

zebrafish lateral line assay identified two such compounds,

PROTO-1 yielded ORC-13661, an otoprotective lead compound

that acts as a permeant MET channel blocker (Owens et al., 2008;

Chowdhury et al., 2018; Kitcher et al., 2019). In a separate study,

Kenyon et al. (2017) used zebrafish to identify an N-methyl-D-

aspartate (NMDA) receptor antagonist and a selective potassium

channel antagonist that also protected hair cells by attenuating

aminoglycoside entry. Here, we use a zebrafish lateral line assay

to assess the relative protection conferred from a modified

scaffold of an otoprotective plant alkaloid. Our modifications

are designed to diversify the alkaloid’s pharmacological activity

to modulate multiple aspects of hair cell death, leading to a

stronger therapy.

A previous study by our lab screened 502 natural compounds

using a zebrafish model for ototoxicity and identified four

otoprotective bisbenzylisoquinoline analogs: berbamine,

E6 berbamine, hernandezine, and isotetrandrine, with berbamine

being the most protective (Kruger et al., 2016). These analogs

share a macrocyclic bistetrahydroisoquinoline ring scaffold and

robustly protect hair cells from aminoglycoside damage, likely

by attenuating aminoglycoside entry. These data are consistent

with Ou et al. (2009, 2012), who demonstrated that quinoline

ring compounds such as tacrine and chloroquine reduce

aminoglycoside uptake by hair cells, leading to increased hair

cell survival. Berbamine also reduces aminoglycoside-induced

hair cell death in mice, likely by reducing aminoglycoside

loading into the cochlea (Kirkwood et al., 2017). However, high

concentrations of berbamine (30 µM) were toxic to murine

cochlear hair cells. Screening additional berbamine analogs offer

an excellent opportunity to identify moieties that are responsible

for berbamine’s protective activity while avoiding the toxicity

seen at high concentrations. This information will allow us to

develop a non-toxic compound that maintains the diverse and

protective pharmacological properties of berbamine to develop a

robust therapeutic compound.

Bisbenzylisoquinoline compounds have recently attracted

attention as therapeutics due to their innate chemical diversity

and wide-ranging pharmacological effects, making them

powerful scaffolds for drug optimization (Tian and Zheng,

2017; Teles et al., 2019). Berbamine, a plant alkaloid derived

from Berberidaceae (barberry plants), has its pharmacological

origins in traditional Chinese medicine as an anti-inflammatory

agent. More recently, berbamine has shown potential as a

therapeutic to reduce inflammation from a diverse array of

cancers, thus inhibiting tumor cell invasion (Ren et al., 2008;

Liang et al., 2016; Jia et al., 2017). Here, we investigate the

protective role of berbamine analogs in a zebrafish model of

Frontiers in Cellular Neuroscience | www.frontiersin.org

2

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

aminoglycoside-induced hair cell death. The study of structural

analogs offers the opportunity to identify structural moieties of

berbamine that may be responsible for its protective activity and

enable the development of a robust and targeted therapeutic, the

long-term goal of this work.

d, doublet; t, triplet; q, quartet; m, multiplet; br, broad

singlet. The relative integrals of peak areas agreed with those

expected for the assigned structures. Elemental analyses were

performed by AtlanticMicrolab, Incorporation (Norcross, GA,

USA). Element compositions are within 0.4% of the calculated

values. Thin-layer chromatography (TLC) was performed on

WHATMAN UV254 silica gel plates with a fluorescent

indicator and the spots were visualized under 254 and/or

365 nm illumination. Flash chromatography was performed on

a Teledyne ISCO CombiFlash chromatography system using

pre-packed silica gel columns purchased from Teledyne ISCO.

The zebrafish has become an increasingly popular vertebrate

model in biomedical research, in part due to their small

transparent larvae that are produced in large clutches, providing

an excellent system for high-throughput drug screening (Ou

et al., 2009, 2010; Coffin et al., 2010; Esterberg et al., 2013; Kenyon

et al., 2017 ). Zebrafish and other fishes possess a sensory system

known as the lateral line which is composed of hair cells clustered

into groups (neuromasts). These hair cells are structurally and

functionally similar to mammalian hair cells and show similar

responses to ototoxic damage. The lateral line functions to detect

vibrations in the water and to initiate an appropriate behavioral

response such as predator avoidance or schooling. One advantage

of using this model is that the hair cells are located externally,

allowing for easy pharmacological manipulation in vivo. Using

this advantage, chemical genetic screens have been performed

to identify compounds that protect against aminoglycoside-

induced hair cell loss. To date, the zebrafish lateral line has served

as a platform for the discovery of several drug candidates with

otoprotective potentials, such as PROTO- 1 (the precursor to

ORC-13661), berbamine, tacrine, afimoxifene, and linopirdine

(e.g., Ton and Parng, 2005; Owens et al., 2008; Ou et al., 2009;

Vlasits et al., 2012; Kruger et al., 2016; Kenyon et al., 2017 ).

Furthermore, some of these drugs confer similar protections in

the mammalian inner ear (Ou et al., 2009; Kenyon et al., 2017;

Kirkwood et al., 2017; Majumder et al., 2017 ; Chowdhury et al.,

2018; Kitcher et al., 2019 ; O’Reilly et al., 2019 ).

Animal Husbandry

All experiments were approved by the Institutional Animal Care

and Use Committee at Washington State University (protocol #

6024) or by the Animal Care and Use Committee at the National

Institutes of Health (NIH; protocol #1362–13). Zebrafish (Danio

rerio) were maintained in aquatic facilities at Washington State

University Vancouver or the NIH (calcium imaging experiments

only). Zebrafish embryos were obtained by breeding pairs of

adults and raising the embryos at 28^◦C. Unless stated otherwise,

all experiments were conducted using E2 Embryo Medium

(EM; 1 mM MgSO₄, 120 µM KH₂PO₄, 74 µM Na₂HPO₄,

1 mM CaCl₂, 500 µM KCl, 15 mM NaCl₂in distilled water

at a pH of 7.2; Westerfield, 2000). Larval zebrafish between

5–6 days post-fertilization (dpf) in the wildtype *AB strain

background were used for experiments unless specifically noted.

Tg(myo6b:EGFP), Tg(Brn3c:GFP) or Tg(α-tubulin: tdTomato)

transgenic lines were used to visualize hair cells for a subset

of experiments. We selected fish between 5–6 dpf because

their hair cells show adult-like responses to ototoxic damage

(Murakami et al., 2003; Santos et al., 2006). For calcium imaging

experiments to detect mechanosensitive responses in lateral line

hair bundles, the previously described transgenic line was used:

Tg(-6myo6b:GCaMP6s-CAAX)^idc1Tg(Zhang et al., 2018). This

line was maintained in the wildtype Tübingen strain background.

Here, we report the synthesis and study of 16 berbamine

analogs as potential otoprotective therapeutics. We demonstrate

that alkylated and acylated berbamine analogs attenuate

aminoglycoside-induced hair cell death in the zebrafish lateral

line. Tetrahydroisoquinoline monomers were inactive and failed

to attenuate cell death. The majority of active analogs confer

protection by attenuating aminoglycoside entry into hair cells.

Interestingly, a subset of analogs also confer protection via an Berbamine Analog Dose-Response

uptake-independent mechanism. The compounds that show a

multimodal mechanism of protection may be stronger leads

for developing a therapy because they can potentially modulate

multiple targets to inhibit hair cell death. This work demonstrates

the potential clinical effectiveness of berbamine analogs as a

therapeutic to prevent aminoglycoside ototoxicity.

To create a dose-response curve, zebrafish (in 6-well plates)

were pretreated with varying log and half-log concentrations

(0.0032 µM to 9.98 µM) of a berbamine analog for one hour

before aminoglycoside treatment (neomycin or gentamicin, both

from Sigma-Aldrich, St. Louis, MO, USA, N1142 and G1397,

respectively). Analogs were reconstituted in dimethylsulfoxide

(DMSO) and diluted in EM before zebrafish administration. The

dilutions were controlled so the final % of DMSO was the same

across all analogs (0.5%). Control fish were treated with DMSO

(vehicle) in the absence of analog. Larval zebrafish were exposed

to one of two aminoglycoside treatment paradigms, defined here

as either ‘‘acute’’ or ‘‘chronic’’ exposure. For acute treatment,

zebrafish were exposed to 200 µM neomycin for 30 min in the

presence of an analog, followed by four rinses in EM and a

60-min recovery in EM with no drugs present. Chronic exposure

consisted of 100 µM gentamicin for 6 h followed by four rinses

in EM and a 15-min recovery. We selected these aminoglycosides

and exposure paradigms because they both induce hair cell

MATERIALS AND METHODS

Compound Synthesis

All solvents and chemicals for compound synthesis were

purchased from Acros Organics, Sigma-Aldrich, Combi-Blocks,

or AstaTech, Inc. and were used as received. All evaporations

were carried out in vacuo with a rotary evaporator. Nuclear

magnetic resonance spectra for proton (¹H NMR) were

recorded on a Bruker 300 MHz NMR spectrometer. The

chemical shift values are expressed in ppm (parts per million)

relative to tetramethylsilane as an internal standard: s, singlet;

Frontiers in Cellular Neuroscience | www.frontiersin.org

3

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

damage with distinct time courses and mechanisms (Owens et al.,

2009; Coffin et al., 2013a,b).

followed by co-treatment with analog and 50 µM GTTR (made

according to Steyger et al., 2003) for 18 min (Steyger et al.,

2003; Alharazneh et al., 2011; Kruger et al., 2016). The fish were

then rinsed 4× in EM before euthanasia with 0.002% buffered

MS-222 and fixation in 4% paraformaldehyde, then imaged on

a Leica SP8 confocal microscope using a 20× dry objective

with 5× digital zoom. The gain and laser settings were constant

across all groups. The image stacks were collapsed into maximum

projection images before image analysis via ImageJ. The images

were analyzed by subtracting the background and measuring the

fluorescence intensity (arbitrary units; a.u.) of each neuromast

and taking an overall average per group (Kruger et al., 2016).

The IO2 neuromast was selected for consistency throughout

the experiments.

Hair Cell Assessment

The majority of our hair cell assessment was done using 2-

(4-(dimethylamino)styryl)-N- ethylpyridinium iodide (DASPEI;

ThermoFisher, Eugene, OR, USA, D426), a mitochondrial dye

that robustly labels the hair cells in the lateral line of zebrafish

and other fishes (Harris et al., 2003; Coffin et al., 2009; Brown

et al., 2010). After aminoglycoside and/or analog treatments,

zebrafish were incubated for 15 min at 28^◦C with 0.005%

DASPEI. The fish were subsequently rinsed 2× with EM and

anesthetized with 0.001% buffered tricaine methanesulfonate

(MS-222, Syndel, Ferndale, WA, USA) before visualization.

We assessed the relative fluorescence intensity for 10 head

neuromasts on every zebrafish (IO1, IO2, IO3, IO4, M2, O2, MI1,

MI2, SO1, SO2; Raible and Kruse, 2000) using 50x magnification

on a Leica M165F fluorescent dissecting microscope. Each

neuromast was given a score to quantify the fluorescence: 2

(bright labeling), 1 (modest labeling), and 0 (no labeling);

the individual neuromast scores were added to provide an

overall score between 0–20 per fish (Harris et al., 2003; Coffin

et al., 2009; Owens et al., 2009; Kruger et al., 2016 ). Scores

were averaged per treatment group. To normalize the data, all

treatment group averages were divided by the average value of

the controls for that experiment and multiplied by 100 to arrive

at the percentage of surviving hair cells. DASPEI labeling is a

viable method of assessing hair cells with comparable results

to directly counting immunolabeled hair cells (Harris et al.,

2003; Coffin et al., 2013a; Kruger et al., 2016). For a subset

of experiments, we used Tg(Brn3c:GFP) larvae, where hair cells

express membrane-specific GFP. These experiments provided

observation of cellular morphology following aminoglycoside

and analog treatment. Animals were treated as described above

for analog and aminoglycoside exposure, then euthanized with

0.002% buffered MS-222, fixed in 4% paraformaldehyde, and

imaged on a Leica SP8 confocal microscope.

Washout Experiment

To further understand the mechanism(s) of protection by

berbamine analogs, we exposed zebrafish to one of three

treatment paradigms that varied in the timing of aminoglycoside

administration relative to berbamine analog administration. The

first paradigm relied on a pre-co treatment where fish were

exposed to the analog both before and during gentamicin

exposure. In the second paradigm, there was a co-treatment of

the analog and aminoglycoside only during gentamicin exposure.

In the third paradigm, post-treatment, the analog was applied

after gentamicin removal. In all three treatments, we used a

30-min exposure to 200 µM gentamicin followed by a 5.5-

h recovery in EM or (for the post-treatment time course)

berbamine analog (Kruger et al., 2016). This aminoglycoside

time course was selected since gentamicin-induced hair cell

damage continues after ototoxin removal, allowing us to assess

the optimal protective paradigm in a slower phase of hair cell

death (Owens et al., 2009). Zebrafish were then rinsed 4× in

EM and anesthetized with MS-222 before the DASPEI hair

cell assessment.

Hair Cell Viability Assessment

To confirm that our analogs were not compromising the viability

of the hair cells or the function of the MET channel after

prolonged analog exposure and subsequent removal, we assessed

the hair cells using a combination of two vital dyes, YO-PRO-

1 and FM 1–43FX (both from ThermoFisher Scientific, Eugene,

OR, USA, Y3603 and F35355, respectively). YO-PRO-1 is a

nucleic acid stain that specifically labels lateral line hair cells in

live larvae, allowing us to visualize nuclear morphology (Santos

et al., 2006). It is also a nontoxic dye that can be retained

for 24 h making it suitable for long-term hair cell assessments

(Thomas et al., 2015; Kruger et al., 2016; Neveux et al., 2016).

FM1–43FX is a membrane probe that readily enters hair cells via

the MET channel (Gale et al., 2001; Corey et al., 2004; Alharazneh

et al., 2011) making it a good proxy for MET channel function

following analog removal. Zebrafish were exposed to 3 µM YO-

PRO-1 for 1 h followed by three rinses in EM. After labeling, the

zebrafish were then incubated with analog for 24 h, rinsed 4×

with EM, and left to recover for an hour to allow for dissociation

of the analog from the MET channel (e.g., Kirkwood et al.,

2017). Zebrafish were then incubated with 1.5 µM FM1–43FX

Protection After 24 h

To determine whether we were observing true protection of

hair cells from aminoglycoside damage rather than a delay in

hair cell death, we exposed zebrafish to the optimal protective

concentration (OPC) of each of our analogs for 1 h followed

by a co-treatment with analog and 200 µM gentamicin for

30 min, with a 24-h recovery. This time course was selected to

examine the protective capacity of our analogs in a prolonged

phase of aminoglycoside exposure since gentamicin-induced hair

cell damage continues after drug removal (Owens et al., 2009).

Zebrafish were then rinsed 4× in EM and anesthetized with

MS-222 before DASPEI labeling and hair cell assessment.

GTTR Loading Assay

To quantify the uptake of aminoglycosides in the presence

of berbamine analog, we exposed our zebrafish to gentamicin

conjugated with Texas Red (GTTR), which predominantly enters

hair cells via the MET channel (Steyger et al., 2003; Alharazneh

et al., 2011). Zebrafish were first administered a one-hour

pretreatment of the optimally protective concentration of analog

Frontiers in Cellular Neuroscience | www.frontiersin.org

4

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

for 30 s, rinsed 4× in EM, and anesthetized with MS-222

before live imaging on a Leica SP8 confocal microscope. All

imaging was done sequentially on live anesthetized fish to resolve

discrete signals between the fluorophores since FM1–43FX has

a broad emission spectrum that overlaps with YO-PRO-1. In a

separate experiment, Tg(myo6b:EGFP) zebrafish also were used

to confirm that berbamine analogs are not ototoxic. We exposed

our Tg(myo6b:EGFP) zebrafish to the OPC of our analogs for 24 h

followed by 4× rinses in EM. The zebrafish were then euthanized

with MS-222 followed by fixation with 4% paraformaldehyde

before confocal imaging of neuromast IO2 for all zebrafish. The

gain was kept constant across all groups.

a Nikon CFI Fluor 60× 1.0 NA water immersion objective. To

coordinate stimulation with image acquisition, the fluid-jet was

driven by a voltage-step command from the imaging software

(Prairie view) during image acquisition. To simultaneously image

the calcium signals in all hair bundles a piezoelectric motor

(PICMA P-882.11–888.11 series, PI instruments, Auburn, MA,

USA) attached to the objective was used to allow rapid imaging

in five planes along the Z-axis at 0.5 µm intervals. Z-stacks were

acquired at a 50 Hz frame rate yielding a 10 Hz volume rate.

Calcium imaging experiments were performed on a minimum

of four animals and seven neuromasts on two independent days.

For image processing and quantification, the five plane Z-stack at

each timepoint was projected into a single plane. The GCaMP6s

signals in the projected images were then measured in FIJI. In

FIJI, a circular region of interest with a ∼1.5 µm diameter was

placed on the center of each bundle and a peak finding program

was used to determine the ∆F/F₀for each bundle. We quantified

both the average bundle ∆F/F₀per neuromast and the temporal

representation of the average ∆F/F₀signals.

Fluorescent Analog Localization

We examined a fluorescently tagged berbamine analog to

determine if analogs can enter hair cells. We used Tg(α-tubulin:

tdTomato) zebrafish (to visualize hair cells) and pre-treated them

with a fluorescent 7-(diethylamino)coumarin-tagged berbamine

analog (BA-17) for one hour in EM before confocal imaging. The

absorption and emission maxima of the coumarin tag are 409 nm

and 473 nm, respectively. To test that the fluorescent tag did not

interfere with the protective effects of the analog, we pretreated

zebrafish for one hour with the fluorescently tagged analog before

acute neomycin exposure. Zebrafish were then rinsed 4× with

EM and anesthetized with MS-222 before DASPEI labeling and

hair cell assessment. To examine if the punctate labeling from

BA-17 colocalized with lysosomes, we used AB zebrafish labeled

with LysoTracker (ThermoFisher, Eugene, OR, USA, L7528).

Zebrafish were pretreated with BA-17 for 1 h followed by a 3-min

exposure to 50 nM LysoTracker before confocal imaging of the

IO2 neuromast.

Data Analysis

Data were analyzed via one-way ANOVA using GraphPad

Prism version 8. Post hoc comparisons were performed using

Bonferroni or Tukey corrections for multiple comparisons

(see figure legends for details). For calcium imaging, statistical

differences in hair bundle calcium responses were determined

using a paired, two-tailed Student’s t-test. Data are presented as

mean ± SEM. ED₅₀analysis was conducted by normalizing all

y-values and analyzing dose-response stimulus for agonist

concentration vs. normalized response using GraphPad

Prism version 8.

Functional Calcium Imaging in Zebraﬁsh Hair

Bundles

RESULTS

The protocol to image GCaMP6s-based calcium signals in Berbamine Analog Synthesis

zebrafish hair bundles has been described previously in detail

(Lukasz and Kindt, 2018 ). Briefly, before imaging, individual

larvae were anesthetized with 0.03% MS-222 (Sigma) and then

pinned onto a Sylgard-filled recording chamber. To suppress the

movement, alpha-bungarotoxin (125 µM, Tocris, Minneapolis,

MN, USA) was injected into the heart. Larvae were then

immersed in extracellular imaging solution (in mM: 140 NaCl,

2 KCl, 2 CaCl₂, 1 MgCl_2,and 10 HEPES, pH 7.3) without MS-222.

A fluid-jet (a glass pipette connected to an HSPC-1 device, ALA

Scientific, Farmingdale, NY, USA) was used to apply a 500-ms

anterior or posterior stimulus to deflect the apical bundles of

trunk neuromasts (L1-L4). After baseline GCaMP6s responses

were recorded, the extracellular imaging solution was exchanged

with the same solution containing a berbamine analog. The

solution in the fluid-jet pipette was also exchanged. Ten minutes

after the solutions were exchanged, the same hair bundles were

restimulated and the GCaMP6s responses were reexamined.

Calcium imaging was performed blind concerning the protective

capacity of each berbamine analog.

Berbamine is a bisbenzylisoquinoline comprised of two

structurally different benzyltetrahydroisoquinolines in which the

benzyl groups are linked on one end and the isoquinoline groups

are linked on the other, forming a macrocycle (Figure 1A). We

selected an analog synthesis strategy based on the chemistry

of the parent berbamine compound, to explore alkyl and

phenyl substitutions and the contribution of the dual ring

structure to otoprotective activity. Berbamine has one readily

available synthetic handle, the phenol, which was used to

synthesize our analogs (Figure 1A). This synthetic strategy

maintained both the O-acylated (E6 berbamine) and O-alkylated

(isotetrandrine) version of berbamine, based on our prior data

demonstrating that both of these structures confer otoprotection

(Kruger et al., 2016). Eight acylated berbamine analogs were

synthesized as shown in Figure 1B (BA-1 to BA-8). Analogs

were synthesized using the Steglich esterification method by

reacting berbamine with a carboxylic acid in the presence of

N,N⁰-dicyclohexylcarbodiimide and 4-dimethylaminopyridine.

Four alkylated berbamine analogs were synthesized (BA-9 to

BA-12). These were synthesized using the standard Williamson

etherification conditions by reacting berbamine with sodium

hydride, and then alkylating the intermediate phenoxide anion

To capture calcium-dependent GCaMP6s signals in apical

hair bundles, we used a Bruker Swept-field confocal system

equipped with a Rolera EM-C2 CCD camera (QImaging) and

Frontiers in Cellular Neuroscience | www.frontiersin.org

5

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

aminoglycosides (Figures 2, 3) and were more robust

protectants than berbamine. Twelve total analogs significantly

protected against acute neomycin damage, while 13 conferred

protection from chronic gentamicin. The berbamine monomer

benzyltetrahydroisoquinolines conferred little to no protection;

BA-14 was significantly protective from gentamicin, and

BA-15 was significantly protective from neomycin, but in

both cases, protection was modest compared to the other

analogs. We observed some toxicity for some analogs for

the highest concentration tested (9.98 µM) but all protective

concentrations were orders of magnitude lower than the

ototoxic concentration.

Given our long-term goal of clinical translation, we were

especially interested in the analogs with both high efficacy and

high potency. As shown in Figure 3, for all 11 acylated or

alkylated analogs that attenuated neomycin damage, the OPC

(optimal protective concentration) was between 0.032 µM and 1

µM with an ED₅₀range of 0.02 µM–0.24 µM (Figures 3A,C).

BA-7 (purple line) was the most potent analog with an OPC

of 0.032 µM and an ED₅₀of 0.02 µM. All 12 acylated or

alkylated analogs reduced gentamicin toxicity with OPCs varying

between 0.32 µM and 9.98 µM and an ED₅₀range of 0.03

µM–1.83 µM (Figures 3B,D). Compound BA-8 (pink line) was

an exceptionally robust protectant, with an OPC of 0.32 µM

against gentamicin, yet no protection from neomycin. Overall,

the phenyl ring appears to be important, but BA-9, which is

an alkylated version of berbamine (and an extension of the

alkyl group of isotetrandrine) is also robustly protective, so

the phenyl group may not be needed. The statistical analyses

for all neomycin and gentamicin dose-response curves can be

found in Table 1. Collectively, these data suggest that acylated

and alkylated berbamine analogs are important for hair cell

protection due to the bisbenzylisoquinoline ring system and the

hydrophobic functional groups.

FIGURE 1 | Synthesis of berbamine analogs. (A) Design strategy, including

the synthetic handle. (B) Speciﬁc analogs synthesized for otoprotection

studies. BA-13 through BA-16 are monomers, while BA-17 is a ﬂuorescent

analog.

Hair Cells Are Viable and Functional After

Prolonged Analog Exposure

with an alkyl bromide in the presence of a catalytic amount of

sodium iodide. These two approaches enable the study of various

analogs modified on the phenol position.

Next, we wanted to assess whether our analogs were

compromising the long-term viability of the hair cells or

the functionality of the MET channel after prolonged analog

exposure and removal. After a 24-h incubation in each analog

at the OPC, we assessed hair cell viability using the vital dye

YO-PRO-1 and found no qualitative reduction in fluorescence

intensity (Figure 4A). FM 1–43FX was used to assess MET

channel function since this dye is known to enter hair cells via

the MET channel (Gale et al., 2001; Alharazneh et al., 2011). Of

the analogs tested that showed protection from neomycin and/or

gentamicin, none of them appeared to reduce FM 1–43FX entry,

suggesting that MET channel integrity is preserved after analog

removal (Figure 4A). To further complement this experiment,

berbamine analogs were individually administered for 24 h to

myo6b:EGFP fish to observe hair cell morphology or to AB

larval zebrafish for hair cell survival assessment with DASPEI

labeling (Figures 4B,C). We did not observe changes to hair

cell morphology or reductions in DASPEI scores. Overall, our

viability and functional assays indicate that at their OPC, our

analogs do not cause hair cell toxicity.

Four benzyltetrahydroisoquinolines (BA-13 to B-

16) were also studied and considered monomers of the

bisbenzylisoquinoline ring system. BA-15 and BA-16 were

commercially available. BA-13 and BA-14 were synthesized

using a slightly modified literature procedure based on Perrey

et al. (2013). 4-Methoxyphenylacetic acid was converted

into amides with either 4-methoxyphenethylamine or 3,

4-dimethoxyphenethylamine

using

propanephosphonic

acid anhydride as a coupling reagent and Hünig’s base.

The resulting amides were then cyclized into intermediate

dihydroisoquinolines using phosphorous oxychloride in toluene

at 90^◦C followed by reduction of the resulting imine with sodium

borohydride to form the final compounds.

Berbamine Analogs Confer

Dose-Dependent Protection

We examined these 16 different berbamine analogs and

found that all 12 acylated or alkylated analogs (BA-1 to

BA-12) protected hair cells from at least one of the

Frontiers in Cellular Neuroscience | www.frontiersin.org

6

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

FIGURE 2 | Berbamine analogs confer protection from neomycin. Analog BA-4 is shown here as an example. (A) Zebraﬁsh labeled with the mitochondrial dye

DASPEI to visualize neuromasts in the lateral line (white dots). Scale bar = 1 mm. (B) Zebraﬁsh labeled with DASPEI after 200 µM neomycin resulting in reduced

ﬂuorescence. (C) Zebraﬁsh labeled with DASPEI after pre-co treatment with analog BA-4 and 200 µM neomycin. (D,E) Representative IO3 neuromasts from

Tg(Brn3c:GFP) zebraﬁsh. (D) Zebraﬁsh exposed to 200 µM neomycin exposure, resulting in hair cell loss. Scale bar = 10 µm. (E) Zebraﬁsh exposed to a pre-co

treatment of analog BA-4 and 200 µM neomycin.

FIGURE 3 | Berbamine analogs confer protection from aminoglycosides. (A,B) Dose-response curves for berbamine analog polymers and berbamine (parent

compound) against 200 µM neomycin and 100 µM gentamicin. (C,D) Dose-response curves for the remainder of the berbamine analog polymers, and the

berbamine analog monomers (BA-13 to BA-16) against 200 µM neomycin and 100 µM gentamicin. Zebraﬁsh were pretreated for 1 h with each berbamine analog at

variable doses followed by co-treatment with analog and aminoglycoside. Data are normalized to controls. Twelve analogs protected hair cells from neomycin

damage, while 11 protected from gentamicin. Ten of the berbamine analogs were protective against both neomycin and gentamicin in a dose-dependent manner.

Hair cells were assessed via DASPEI. Data are presented normalized to vehicle controls. Data were analyzed by one-way ANOVA, N = 4–12, except for the

gentamicin dose-response curves for BA-9, BA-10, BA-11, and BA-13, where N = 2 for the gentamicin-only groups. We are still conﬁdent in the data for these

groups, based on the additional gentamicin data shown here and the data in Supplementary Figure S1. Bars are ± SEM. Refer to T a ble 1 for statistics.

Given that our protective analog concentrations are not

inherently ototoxic, we next wanted to test whether berbamine

analogs were truly otoprotective or if they only delayed hair cell

death. We therefore pre-treated larvae with the OPC of analog

for one hour before a 30-min administration of gentamicin,

then assessed hair cells 24 h later. This time course was selected

Frontiers in Cellular Neuroscience | www.frontiersin.org

7

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

TABLE 1 | Statistics for berbamine analog dose-response curves for 200 µM

neomycin and 100 µM gentamicin for the data shown in Figure 3.

TABLE 1 | Continued

Compound Neomycin:

Neomycin:

p-value by

concentration

(µM)

Gentamicin:

ANOVA

Gentamicin:

p-value by

concentration

(µM)

Compound Neomycin:

Neomycin:

p-value by

concentration

(µM)

Gentamicin:

ANOVA

Gentamicin:

p-value by

concentration

(µM)

Name

ANOVA

Name

ANOVA

Berbamine F_(6,60)= 4.02

P = 0.0019

0.32*

F_(6,66)= 2.14

P = 0.06

ns

BA-1

BA-2

F_(6,69)= 94.79

P < 0.0001

0.1; **

0.32; ^∗∗∗∗

1; ^∗∗∗∗

F_(6,70)= 48.91

P < 0.0001

0.032; *

0.32; ^∗∗∗∗

1; ^∗∗∗∗

0.032; ^∗∗∗∗

0.1; ^∗∗∗∗

0.32; ^∗∗∗

1; ^∗∗∗∗

3.16; ^∗∗∗∗

9.98; ^∗∗∗∗

0.0032; *

0.032; ^∗∗∗∗

0.1; ^∗∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

3.16; ^∗∗∗∗

9.98; ^∗∗∗∗

0.01*

Compound names in bold indicate protection from both neomycin and gentamicin,

while compound names in black (not bold) protect from neomycin or gentamicin, and

compound names in gray indicate non-protective compounds. Data were analyzed by

one-way ANOVA (columns 2 and 4) followed by Bonferroni-corrected t-tests for multiple

comparisons (columns 3 and 5). In columns 3 and 5, numbers in black indicate signiﬁcant

increases in hair cell survival (compared to aminoglycoside-only controls), while numbers

in gray indicate signiﬁcant reductions in hair cell survival relative to aminoglycoside-only

groups. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, **p < 0.01, *p < 0.05. ns, not signiﬁcant.

F_(6,66)= 72.23 0.0032; **

P < 0.0001

F_(8,89)= 50.76

P < 0.0001

0.01; ^∗∗∗∗

0.032; ^∗∗∗∗

0.1; ^∗∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

BA-3

BA-4

F_(6,70)= 70.76 0.032; ^∗∗∗

P < 0.0001

F_(8,63)= 61.08

P < 0.0001

0.1; ^∗∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

because after an acute gentamicin exposure hair cell damage

continues even after drug removal (Owens et al., 2009). All of the

analogs tested (BA-2, BA-6, BA-7, and BA-8) showed over 90%

hair cell survival 24 h post-gentamicin exposure compared to

untreated controls (Figure 5). When we compared the protection

of our analogs in chronic (6 h) vs. acute (30 min + 24 h)

gentamicin treatment, we saw that BA-6 and BA-7 exhibited

10–20% more protection in the acute gentamicin exposure assay

(compare Figures 3B, 5). These data demonstrate that berbamine

analogs robustly protect hair cells from aminoglycoside damage

after both acute and chronic aminoglycoside treatment.

F_(6,70)= 51.23 0.0032; **

P < 0.0001

F_(6,60)= 19.36

P < 0.0001

0.01; **

0.032^∗∗∗∗

0.1; ^∗∗∗∗

0.32; ^∗∗∗

1; ^∗∗∗∗

0.032; *

0.1; ^∗∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

BA-5

BA-6

F_(6,68)= 15.29

P < 0.0001

0.1; **

F_(6,42)= 4.34

P = 0.017

0.0032; *

1; ^∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

F_(6,70)= 94.12 0.1; ^∗∗∗∗

P < 0.0001

F_(6,76)= 41.35

P < 0.0001

0.0032; ^∗∗∗∗

0.032; ^∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

A Subset of Berbamine Analogs Reduce

GTTR Loading

F_(6,53)= 113.8 0.032; ^∗∗∗∗

F_(8,77)= 10.83

P < 0.0001

1; ^∗∗∗

Next, we investigated if our analogs protect hair cells by

blocking aminoglycoside entry. Previous research shows that

aminoglycosides predominantly enter hair cells via the MET

channel. Furthermore, quinoline-ringed structures have been

shown to protect hair cells by reducing MET channel-mediated

aminoglycoside loading (Ou et al., 2012; Kirkwood et al., 2017).

Since berbamine is a quinoline ring structure we predicted its

analogs could also protect by reducing aminoglycoside uptake

via the MET channel. We found that all 12 of the acylated

or alkylated berbamine analogs (BA-1 to BA-12) attenuated

GTTR uptake by 50% or more relative to negative controls

(Figures 6A,B). The berbamine monomers (BA-13 to BA-16) did

not attenuate uptake compared to negative controls. Calcium,

which reduces the open probability of the MET channel, was

used as a positive control. While calcium reduced GTTR by

70%, 10 of our analogs were more effective uptake blockers,

as was the parent compound berbamine (Figure 6B, Kruger

et al., 2016). BA-7 was the most effective uptake blocker

with over 95% attenuation of GTTR fluorescence (Figure 6B).

We then conducted a linear regression to determine how

strongly hair cell protection was correlated with GTTR uptake

(Figure 6C). Given a correlation coefficient of 0.23, there is no

relationship, suggesting that protection may not be solely reliant

on aminoglycoside uptake block. Overall our GTTR uptake

analysis indicates that a subset of analogs block GTTR loading

into the hair cells but the magnitude of uptake block is not

directly correlated with hair cell protection.

BA-7

P < 0.0001

0.1; ^∗∗∗∗

0.32; ^∗∗∗∗

1; ^∗∗∗∗

3.16; ^∗∗∗∗

BA-8

F_(6,72)= 19.58 0.0032; ^∗∗∗∗

F_(8,59)= 19.85

P < 0.0001

0.32; ^∗∗∗∗

1; ^∗∗∗∗

3.16; ^∗∗∗∗

P < 0.0001

0.01; ^∗∗∗∗

0.032; ^∗∗∗∗

BA-9

F_(4,40)= 332.3 0.1; ^∗∗∗∗

F_(4,38)= 32.05

P < 0.0001

0.1; *

9.98; *

P < 0.0001

1; ^∗∗∗∗

9.98; ^∗∗∗∗

BA-10

BA-11

BA-12

F_(4,39)= 111.5 0.1; ^∗∗∗∗

F_(4,33)= 9.09

P < 0.0001

ns

P < 0.0001

1; ^∗∗∗∗

9.98; ^∗∗∗∗

0.1; **

F_(4,26)= 87.80

P < 0.0001

F_(4,23)= 7.09

P = 0.0007

1; ^∗∗∗∗

9.98; **

1; ^∗∗∗

F_(4,37)= 5.90

P = 0.0009

F_(4,52)= 67.35

P < 0.0001

0.1; ^∗∗∗∗

1;^∗∗∗∗

9.98^∗∗∗∗

BA-13

BA-14

F_(4,32)= 0.60

Ns

F_(4,22)= 5.18

P = 0.004

0.1; **

F_(4,44)= 0.468

F_(4,48)= 10.54

P < 0.0001

0.1; ^∗∗∗

1; ^∗∗∗∗

9.98; ^∗∗∗

BA-15

BA-16

F_(4,48)= 8.03

P < 0.0001

F_(4,40)= 5.99

P = 0.0007

0.1; **

1; **

F_(5,34)= 1.80

P = 0.139

F_(5,34)= 1.29

P = 0.289

ns

(Continued)

Frontiers in Cellular Neuroscience | www.frontiersin.org

8

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

FIGURE 4 | Hair cells are viable after analog exposure. (A) Zebraﬁsh were pretreated with the nuclear marker YO-PRO-1 for 1 h followed by treatment with the

optimally protective concentration of each analog for 24 h. Zebraﬁsh were then rinsed in embryo medium (EM) and treated with FM 1–43FX to assess

mechanoelectrical transduction (MET) channel function after prolonged analog exposure and removal. Analog treatment did not alter FM 1–43FX or

YO-PRO-1-labeling. Scale bar = 10 µm. (B) Examples of Tg(myo6b:EGFP) neuromasts treated with the optimally protective concentration of each analog for 24 h

before hair cell assessment. In each case, neuromast morphology appeared normal. Scale bar = 10 µm. (C) Zebraﬁsh were exposed to the optimal protective

concentration (OPC) of berbamine analog for 24 h before hair cell assessment with DASPEI. None of the analogs impact hair cell survival. Data are normalized to

controls and analyzed by one-way ANOVA, F_(4,48)= 0.6564, p = 0.6252. N = 11–12, bars are ± SEM.

Berbamine Analogs Differentially Affect

Mechanoelectrical Transduction Channel

Activity

By contrast, the application of the berbamine derivative

BA-9 reduced the GCaMP6s signals by ∼95%, indicating

significant inhibition of mechanotransduction (Figures 7B,C).

Our calcium imaging experiments complement the GTTR

uptake data and suggest that when berbamine analogs block

GTTR uptake, their otoprotective effects are due to the MET

channel block, thereby reducing aminoglycoside entry into

hair cells.

Our results indicate that several of our otoprotective berbamine

analogs attenuate aminoglycoside toxicity by blocking GTTR

uptake through MET channels. To confirm that protective

analogs impair mechanotransduction, we mechanically

stimulated lateral line hair cells and measured GCaMP6s-evoked

calcium signals in hair bundles (Lukasz and Kindt, 2018).

This previously established method allows for sensitive and

quantitative measurements of mechanotransduction in lateral

line hair bundles (Figure 7A). Therefore, these GCaMP6s signals

can provide a quantitative readout of mechanosensitive-calcium

responses in the presence and absence of berbamine analogs.

For our analysis, we tested two representative analogs, one

protective analog (BA-9, a potent GTTR blocker) and one analog

that was not protective (BA-16, which does not block GTTR

uptake). We first measured evoked GCaMP6s signals in hair

bundles during fluid-jet stimulation before analog treatment

(control, blue line, Figures 7B–E). After the application of

the berbamine monomer BA-16, there was no difference

in the magnitude of the GCaMP6s signals in hair bundles

as compared to the pretreatment condition, indicating that

BA-16 does not impair mechanotransduction (Figures 7D,E).

Berbamine Analogs May Also Protect via

Uptake-Independent Mechanisms

We have shown that many of our berbamine analogs can

confer protection when co-administered with gentamicin.

Given that 12 of our berbamine analogs are likely to be

otoprotective due to reduced GTTR uptake, we hypothesized

that no protection would be conferred if we administered

the analogs after gentamicin removal. To test this hypothesis,

we used an acute 30-min gentamicin treatment time course.

We compared three gentamicin treatment paradigms: (1) 1-

h pre-treatment with the analog, followed by co-treatment

with analog and gentamicin (pre-co); (2) co-treatment with the

analog-only during gentamicin treatment (no pre-treatment; co)

and (3) analog post-treatment for 5.5 h after gentamicin washout

(post). All experiments used the OPC of each analog (Figure 8).

Frontiers in Cellular Neuroscience | www.frontiersin.org

9

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

in a dose-dependent manner against 200 µM neomycin and

toxic at a concentration of 9.98 µM, consistent with other

protective analogs (Figure 9A). To determine whether our

analogs enter hair cells, we pre-treated Tg(α-tubulin: tdTomato)

zebrafish with the OPC of BA-17 for one hour before confocal

imaging. BA-17 was visible in hair cells as well as some punctae

in surrounding neuromasts, suggesting the analog can enter

multiple cell types (Figure 9B). To begin to understand cellular

localization of berbamine analogs, we asked if our analogs enter

lysosomes. For these experiments, we pre-treated AB zebrafish

for 1 h with BA-17 followed by incubation in LysoTracker and

found that some BA-17 punctae colocalize with the lysosomal

marker (Figure 9C). These data demonstrate that berbamine

analogs can enter neuromasts cells and localize in subcellular

compartments, including lysosomes.

DISCUSSION

Our former work demonstrated that berbamine, a natural

plant alkaloid, robustly protected zebrafish hair cells from

aminoglycoside-induced damage (Kruger et al., 2016). In

a separate study, Kirkwood et al. (2017) further confirmed

berbamine’s otoprotective properties in the zebrafish lateral

line and in mouse cochlear explants. The present study set

out to develop berbamine analogs that are more potent and

efficacious than the parent compound and to investigate

the extent of their protection against aminoglycoside-

induced hair cell damage. The medicinal chemistry campaign

strategy focused on functionalizing the phenol group to

synthesize analogs as an initial probe of the structure-activity

relationship. Leveraging available chemistry in this manner

is a robust strategy employed in the early development of

natural products. Like many natural products, the berbamine

scaffold is somewhat complex making de novo synthesis

far more time consuming and expensive than substitution

at the available phenol. E6 berbamine and isotetrandrine,

which also confer protection from aminoglycosides (Kruger

et al., 2016), are both substituted at the phenol suggesting

further exploration at that position would likely produce

active compounds. Moreover, since MET inhibition was

the likely target of activity, increasing analog size and

hydrophobicity by substitution at the phenol should make

analogs better-MET blockers.

Overall, we show that many of our berbamine analogs

effectively protect hair cells from aminoglycoside toxicity,

demonstrating more robust protection than berbamine. Details

of all 16 non-fluorescent analogs are shown in Figure 10,

where we classify analogs based on the relative degree of

protection. Of the 16 analogs, nine demonstrated at least 40%

protection against both neomycin and gentamicin-induced hair

cell death (‘‘robust’’ or ‘‘strong’’ protection, see Figure 10).

Two analogs protected against neomycin (>30% protection)

but conferred little protection from gentamicin (BA-5, BA-

12) and one only conferred protection from gentamicin but

not neomycin (BA-8). Twelve analogs reduced GTTR and

thus likely confer protection by blocking aminoglycoside

entry through the MET channel, which is the primary site

FIGURE 5 | Hair cell protection persists after 24 h. Zebraﬁsh were

pretreated with the OPC of analog for 1 h followed by 200 µM gentamicin

treatment for 30 min. Fish were allowed to recover for 24 h before the

assessment. All analogs tested signiﬁcantly prevented hair cell death after

24 h suggesting true protection rather than a delay in cell death onset. Hair

cells were assessed via DASPEI. Data are presented normalized to vehicle

controls. Data were analyzed by one-way ANOVA, ****p < 0.0001,

***p < 0.001, F_(4,32)= 12.98, N = 5–10, bars are ± SEM.

Consistent with our hypothesis, eight of our analogs (BA-1,

BA-2, BA-4, BA-8, BA-9, BA-10, BA-11, and BA-12) only

protected hair cells when the analog was administered before

and during (pre-co) or concurrent with (co) aminoglycoside

administration (Figures 8A–C, Supplementary Figure

S1, Supplementary T able S1 ). One analog, BA-5, which

was previously not protective when co-administered with

gentamicin in the chronic assay, was protective when

administered before and during gentamicin exposure (pre-

co). This suggests that protection occurs via aminoglycoside

uptake block in an acute gentamicin administration but

that protection diminishes in a longer treatment paradigm.

Surprisingly, BA-3, BA-6, and BA-7 showed protection in all

three treatment paradigms. Analogs that show protection

even when administered after aminoglycoside removal

suggest that hair cell protection may also occur by an

uptake-independent mechanism such as interaction with

intracellular pathways.

Berbamine Analogs Enter Hair Cells

Given that a subset of analogs protects hair cells by both

uptake-dependent and uptake-independent mechanisms,

we next wanted to know if the analogs enter hair cells.

For these experiments, we used the approach outlined in

Figure 1 to synthesize BA-17, a berbamine analog with a

fluorescent 7-(diethylamino)coumarin tag. To assess for

protection by this fluorescent analog, zebrafish were exposed to

varying concentrations of analog before and concurrently with

aminoglycoside exposure. We found that BA-17 was protective

Frontiers in Cellular Neuroscience | www.frontiersin.org

10

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

FIGURE 6 | Berbamine analogs reduce aminoglycoside uptake by hair cells. (A) Twelve of the analogs signiﬁcantly reduced uptake of gentamicin conjugated with

Texas Red (GTTR), as did berbamine. Zebraﬁsh were pretreated with the optimally protective analog concentration (or with 25 µM berbamine, Kruger et al., 2016) for

1 h followed by a co-treatment with 50 µM GTTR for 18 min. Dimethylsulfoxide (DMSO) and high calcium are negative and positive controls, respectively. (B)

Fluorescence intensity quantiﬁcation using the OPC of analog. The intensity was measured in arbitrary units and the data were normalized by subtracting the

background ﬂuorescence. Data were analyzed via one-way ANOVA, F_(17,101)= 38.71, ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, asterisks denote

signiﬁcantly different from vehicle controls, N = 4–8 for most compounds, except for N = 2 for compounds BA-1 and BA-5 and N = 15 for berbamine, bars

are ± SEM. (C) Linear regression was conducted for analogs that protect hair cells from gentamicin to identify a relationship between analog potency and uptake

(y = −0.0203x + 0.9162). The Pearson correlation coefﬁcient of 0.2268 denotes no signiﬁcant relationship. Berbamine data are replotted from Kruger et al. (2016).

of aminoglycoside entry into hair cells (Marcotti et al., 2005;

Alharazneh et al., 2011). BA-9, a potent uptake blocker

and protective compound against both neomycin and

gentamicin, reduced MET-evoked calcium activity to baseline

levels, further supporting our hypothesis that protection

is largely due to MET channel block. The reduction of

MET-evoked calcium activity was not observed with BA-

16, a non-protective compound. Our results are consistent with

Kirkwood et al. (2017), who demonstrated that berbamine

is a permeant MET channel blocker in mouse cochlear

hair cells.

Attenuating aminoglycoside uptake into hair cells has gained

traction as a feasible otoprotective strategy (Ou et al., 2012;

O’Sullivan et al., 2017; Kitcher et al., 2019 ). d-Tubocurarine

and phenoxybenzamine have previously been shown to protect

hair cells by reducing aminoglycoside entry (Kirkwood et al.,

2017; Majumder et al., 2017). d-Tubocurarine is a structural

analog of berbamine with a dimethyl quaternary ammonium

Frontiers in Cellular Neuroscience | www.frontiersin.org

11

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

FIGURE 7 | Berbamine analogs differentially reduce mechanotransduction channel activity. Mechanosensitive-calcium signals from hair bundles of posterior lateral

line neuromasts. (A) Side view of a neuromast expressing hair cell-speciﬁc membrane-localized GCaMP6s. Left: an imaging plane through the apical hair bundles

can be used to measure mechanosensitive-calcium signals during ﬂuid-jet stimulation. Right: a top-down view of the imaging plane showing the hair bundles

expressing GCaMP6s. (B) Traces comparing the average GCaMP6s signals before and after application of 1 µM BA-16, N = 10 neuromasts. (C) Dot plots

quantifying the peak GCaMP6s response per neuromast. BA-16 does not alter evoked calcium activity in hair cells. (D) Traces comparing the average GCaMP6s

signals before and after application of 1 µM BA-9, N = 7 neuromasts. (E) Dot plots quantifying the peak GCaMP6s response per neuromast. BA-9 reduces

MET-evoked calcium activity compared to baseline levels. Each dot in (C,E) averages the response of all hair bundles in each neuromast. A paired t-test was used in

(C,E), ^∗∗∗p < 0.001. Error bars in (C,E) and dashed lines on either side of the solid line in (B,D) represent ± SEM.

ion in one of the tetrahydroisoquinoline rings and with

one of the other methoxy groups replaced with a second

phenol (Supplementary Figure S2). Ionic compounds such

as d-tubocurarine might be expected to block the MET

channel since similar effects were observed with analogs of

other scaffolds (Supplementary Figure S2). ORC-13661 is

being developed as an otoprotectant and also protects both

zebrafish and rodent hair cells from aminoglycosides, likely

by blocking aminoglycoside uptake (Chowdhury et al., 2018).

Quaternary analogs of ORC-13661 were much more active

than the non-quaternary versions. ORC-13661 as well as our

berbamine analogs contain lipophilic groups that appear to

improve protective activity. Protection is likely due to the

hydrophobic nature of the compounds and their ability to

diffuse through plasma membranes. This attribute makes these

compounds better candidates for future drug development since

they could potentially cross the blood-brain barrier and/or the

blood labyrinth barrier (Abbott et al., 2006; Koo et al., 2015;

Nyberg et al., 2019).

The majority of our protective analogs were not only

lipophilic but also contained an aryl group. Absence of the

aryl moiety reduced our analogs’ overall protection from

aminoglycosides, especially from gentamicin, as we found for

compound BA-5. Loss of an aryl substituent and its impact on

protective activity from aminoglycosides in zebrafish was also

observed in Chowdhury et al. (2018). In that study, the original

screening hit contained a p -chlorophenyl group and replacement

led to a loss of activity. In our study, analog BA-8 also contained a

chlorophenyl substituent but it did not seem to be important for

neomycin protection, in contrast to the Chowdhury et al. (2018)

study. However, the chlorophenyl substituent was important

for robust gentamicin protection. Likely, differences seen in

neomycin protection for the p-chlorophenyl compounds (ORC-

13661 and BA-8) may have more to do with how the substituent

interacts in combination with its compound scaffold to permeate

the MET channel rather than channel blocking activity being

solely reliant on the moiety.

Nevertheless, aryl groups are present in other protective

compounds such as tacrine, berbamine, carvedilol, ORC-13661,

and phenoxybenzamine, suggesting the importance of the aryl

moiety for otoprotection (Ou et al., 2012; Kruger et al., 2016;

Chowdhury et al., 2018; O’Reilly et al., 2019 ). It is likely that

the phenyl group, among other aryl groups, may be necessary

for protection due to their unusual stability and ability to

delocalize charges within their structure by resonance. This

would allow for temporary charges to reside within the structure

which may form favorable electrostatic interactions with its

target and could explain why quaternary versions of protective

drugs such as ORC-13661 are more active. This is further

supported by evidence from Chowdhury et al. (2018) and

O’Reilly et al. (2019). Both studies found that aminoglycoside

otoprotection required a basic center with a positive charge to

engage in electrostatic interactions, such as basic nitrogen or

a quaternary ammonium salt. An electronegative atom could

also yield a temporary dipole and thus a transient charge, which

may explain why analogs such as BA-2 and BA-8 were robust

Frontiers in Cellular Neuroscience | www.frontiersin.org

12

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

FIGURE 8 | Differential patterns of analog protection from gentamicin. Zebraﬁsh were either pre-treated with the OPC of analog for 1 h, then co-treated with analog

and gentamicin for 30 min (“pre-co” condition), co-treated with analog and gentamicin without a pretreatment period (“co” condition) or post-treated with analog for

5.5 h following gentamicin removal (“post” condition). All treatments used 200 µM gentamicin. Examples are shown for analogs that represent each pattern of

protection. (A) Analog BA-1 was only protective when administered in the “pre-co” or “co” conditions suggesting that protection occurs via aminoglycoside uptake

block. (B,C) Analogs BA-6 and BA-7 signiﬁcantly protected hair cells in all three conditions, suggesting an intracellular mechanism of protection in addition to

aminoglycoside uptake block. (D) Analog BA-16 is not protective of any treatment paradigm. Hair cells were assessed via DASPEI. Data are presented normalized to

untreated controls, which are not shown but represent 100% hair cell survival. Data were analyzed via one-way ANOVA, ****p < 0.0001, **p < 0.01 (as compared to

gentamicin-only controls), N = 7–12, bars are + SEM. Refer to Supplementary T a ble S1 for statistics.

MET blockers. However, aromaticity is not the only feature

that contributes to otoprotection. The majority of our aromatic

compounds were not protective, such as BA-13 through BA-16.

These compounds were half the weight of the other analogs

tested. Additionally, analogs BA-13 through BA-16 have a carbon

center, similar to some non-protective carvedilol derivatives that

Frontiers in Cellular Neuroscience | www.frontiersin.org

13

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

non-endocytic pathways, resulting in drug delivery from the

extracellular space to lysosomes, and that intracellular trafficking

differs between aminoglycosides (Hailey et al., 2017). Some

of our analogs likely target cellular pathways that are unique

to neomycin vs. gentamicin, thus accounting for the observed

differences in protection.

We observed that the fluorescently modified analog BA-17

does enter hair cells but whether the entry is necessary for

protection remains to be determined. An interesting observation

from this experiment was the punctate/granular labeling of

BA-17 in hair cells, which may co-localize with lysosomes

(Figure 9C), supporting findings from Hailey et al. (2017)

that both aminoglycosides that enter hair cells ultimately end

up in lysosomes in zebrafish. Future experiments will further

examine the analogs’ mechanism of protection by observing

how long the fluorescently tagged analog stays in hair cells

and how that time scale compares to the protection that it

exhibits. We also found that there is no correlation between the

degree of uptake block and the magnitude of hair cell protection.

Interestingly, some of the analogs that weren’t protective (BA-

13, BA-16) seemed to enhance GTTR entry. This information

could be potentially useful as it suggests that certain moieties

may exacerbate toxicity and should be avoided in future drug

optimization. Additionally, our analog washout experiments

suggest that our analogs disassociate from the channel and do

not alter MET channel function after the analog is removed,

suggesting that hair cell viability is preserved.

While the MET channel is the primary site of aminoglycoside

entry into hair cells, several other ion channels are also candidates

for aminoglycoside entry. Transient receptor potential (TRP)

channels such as TRPV1, TRPV4, and TRPA1 are particularly

interesting because they have pore diameters large enough to

allow passage of bulky aminoglycoside molecules (Myrdal and

Steyger, 2005; Karasawa et al., 2008; Stepanyan et al., 2011 ).

Kenyon et al. (2017) discovered a select group of aromatic

compounds that protect against aminoglycoside-induced hearing

loss and are also known to have biological activity for potassium

channels, NMDA and AMPA receptors, and other channels.

Of the 13 protective analogs identified in their study, only six

were MET channel blockers, yet additional compounds reduced

fluorescent drug entry, implying that the other compounds may

exert their protective effects through other receptors or channels

(Kenyon et al., 2017). Bisbenzylisoquinoline compounds, in

addition to inhibiting prostaglandins and leukotrienes (Teh et al.,

1990) are also proficient vasodilators and modulators of voltage-

gated calcium channels (Medeiros et al., 2011; Alaoui et al.,

2015), leading to the hypothesis that our analogs may reduce

aminoglycoside entry through alternate channels in tandem with

the MET channel, which would explain the varying degrees

of GTTR fluorescence attenuation shown by our analogs. This

would also explain why some of our multimodal protective

analogs such as BA-3 and BA-6 are not the strongest of uptake

blockers, unlike analogs such as BA-9 and BA-12, which only

protect by attenuating aminoglycoside entry. Future studies

will further characterize the binding targets of our berbamine

analogs, which can also provide insight into the mechanisms of

aminoglycoside-induced hair cell death.

FIGURE 9 | Berbamine analogs enter hair cells. (A) The dose-response

curve for the ﬂuorescent analog (BA-17) demonstrates signiﬁcant protection

from 200 µM neomycin. Hair cells were assessed via DASPEI. Data are

normalized to controls. One-way ANOVA, F_(4,49)= 50.22, p < 0.0001.

N = 10–12, ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, bars are ± SEM. (B) Tg(α-tubulin:

tdTomato) zebraﬁsh were pretreated with BA-17 for 1 h before imaging on a

confocal microscope. Fluorescent analog (green) is present in hair cells (red).

Scale bar = 10 µm. (C) AB zebraﬁsh were pretreated with BA-17 for 1 h

followed by a 3-min exposure to 50 nM LysoTracker before confocal imaging.

Fluorescent analog (green) does overlap with lysosomes (red) in some areas

(white arrowheads). Scale bars in (B,C) = 10 µm.

substituted the center nitrogen with carbon or nonbasic nitrogen

(O’Reilly et al., 2019).

BA-8 robustly protects hair cells from gentamicin damage

yet confers no protection from neomycin. Also, the majority

of our protective analogs show differing degrees of protection

from neomycin vs. gentamicin, supporting the literature that

neomycin and gentamicin activate partially distinct pathways

of hair cell damage (Owens et al., 2009; Coffin et al., 2013a,b ;

Wiedenhoft et al., 2017 ). BA-5 was only protective when

administered before and during gentamicin exposure (pre-co),

suggesting protection occurs via transient aminoglycoside uptake

block during acute gentamicin administration. In the chronic

gentamicin assay, BA-5 was not protective against gentamicin.

This could mean that BA-5 may disassociate from the MET

channel in a longer treatment paradigm thus rendering it

non-protective. These data are consistent with prior studies

showing that acute neomycin and acute gentamicin activate

similar damage mechanisms, while chronic gentamicin also

operates via additional pathway(s) (Coffin et al., 2009, 2013a,b;

Owens et al., 2009). Additionally, there is more recent evidence

that aminoglycosides enter hair cells via both endocytic and

Frontiers in Cellular Neuroscience | www.frontiersin.org

14

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

FIGURE 10 | Summary of berbamine analog protection and compound characteristics. Compound name, structure, molecular weight, protection class, likely

mechanism of protection, ED₅₀, and OPCs against neomycin and gentamicin for all 16 berbamine analogs. The general protection classes were deﬁned based on

thresholds of protection levels by percent. Robust = 60% increase in hair cell protection compared to aminoglycoside group; Strong = over 40% increase in hair cell

protection compared to aminoglycoside group; Modest = over 30% increase in hair cell protection compared to aminoglycoside group; No protection = less than

30% of hair cell protection compared to aminoglycoside group. The likely mechanism was determined by the GTTR experiments in Figure 6 and the time course

experiments shown in Figure 8 and Supplementary Figure S1.

Based on our findings that the majority of our analogs

protect by blocking aminoglycoside uptake, we expected to only

observe protection when the berbamine analog was administered

either before or at the same time as the antibiotic. We were

surprised that three analogs (BA-3, BA-6, and BA-7) exhibited

protection when delivered after aminoglycoside exposure but

before hair cell death, suggesting both uptake-dependent and

uptake-independent protective mechanisms. Interestingly, all

three analogs were alkylated with bulky lipophilic groups

such as alkyl-substituted phenyl groups and cyclohexyl analog.

Analogs with halide substitution or small straight-chain alkyl

substitution did not have the same profile. Berbamine has

been used as an anti-inflammatory in traditional Chinese

medicine and has gained traction as an anti-cancer agent (Du

et al., 2018). Previous studies have shown that berbamine

and its analogs can inhibit tumor growth by modulating

nuclear factor kappa-light-chain-enhancer of activated B cells

(NF-κB) signaling (Liang et al., 2016; Jia et al., 2017). A

recent study showed that berbamine has anti-inflammatory

activity that robustly inhibits the release of chemokines and

cytokines including TNF-α, IL-1β, IL-6 (Jia et al., 2017).

Berbamine also suppressed the phosphorylation of p65, IκB,

c-Jun N-terminal kinase (JNK) pathways, and extracellular

receptor kinase (ERK)1/2 pathways (Jia et al., 2017), all of which

are implicated in aminoglycoside-induced hair cell loss (Huth

et al., 2011 ). Possibly the berbamine analogs that confer uptake-

independent protection may reduce pro-inflammatory cytokines

or modulate pathways such as NF-κB, JNK, or ERK1/2. Future

work will examine the berbamine modulation of intracellular

pathways in aminoglycoside-damaged hair cells.

In our prior work, 25 µM berbamine was the optimal

concentration that prevented aminoglycoside toxicity and

Frontiers in Cellular Neuroscience | www.frontiersin.org

15

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

reduced GTTR loading (Kruger et al., 2016 and Figure 6A).

One advantage of our analogs compared to berbamine is that

the analogs are protective at a much lower concentration than

the parent compound (over 100-fold more potent, in some

cases), making the analogs suitable for clinical translation.

One caveat of some of our analogs is that they are toxic

to hair cells at high concentrations (9.98 µM) in zebrafish

(Figures 3C,D). This toxicity was also observed when berbamine

was tested in mice at concentrations greater than or equal to

30 µM (Kirkwood et al., 2017). However, analog toxicity was

observed at much higher concentrations than the optimally

protective concentration. For example, the OPC for BA-7 is

300-fold lower than the ototoxic concentration. Toxicity of

berbamine and some of its analogs at higher concentrations

may be due to the para-methylene phenol moiety that is

common to many bisbenzylisoquinolines, which can lead to

reactive intermediates (Tian and Zheng, 2017). Furthermore,

many clinical drugs such as Amphetamine and Dapsone are toxic

at high concentrations but highly effective at low concentrations

(Schulz and Schmoldt, 2003). An advantage of using analogs

with natural scaffolds is they can be readily modified to enhance

desired characteristics such as physiological stability, solubility,

and efficiency (Ji et al., 2009). Moving forward we plan to

determine the pharmacokinetics of our lead analogs and further

investigate the protective mechanism(s) of these leads in rodents

as a prelude to clinical trials.

for neomycin) demonstrate protection from neomycin and

gentamicin and strong GTTR uptake block (Figure 10). Our

research provides more insight into hair cell pathology and

aminoglycoside-induced hearing loss, yielding new approaches

to preserve hearing.

DATA AVAILABILITY STATEMENT

The raw data supporting the conclusions of this article will be

made available by the authors, without undue reservation.

ETHICS STATEMENT

The animal study was reviewed and approved by Institutional

Animal Care and Use Committee, Washington State University.

AUTHOR CONTRIBUTIONS

AH and AC conceived the concept, designed the bulk of the

experiments and wrote the bulk of the manuscript. ON, JW and

BB developed the analog synthesis strategy and made the analogs.

KK designed, conducted, and analyzed the calcium imaging

experiments. AH, GL and AC conducted all other experiments

and analyzed the associated data. All authors contributed to

writing and editing.

In summary, our study demonstrates that berbamine analogs

with phenolic moieties robustly protect against aminoglycoside-

induced hair cell death, primarily by attenuating aminoglycoside

entry into hair cells, with some analogs also conferring uptake-

independent protection. Several of the berbamine analogs

described here were more potent than ORC-13661, a compound

in clinical trials as an otoprotectant. The multiple modes of

protection of some of our analogs also bolster their clinical

use since there are likely multiple relevant targets, meaning

that a single compound may confer protection via multiple

mechanisms to increase protective efficacy.

Our former work confirmed that berbamine does not

interfere with the antibacterial efficacy of aminoglycosides

(Kruger et al., 2016) suggesting that berbamine analogs are

not likely to interfere with clinical aminoglycoside use but

that hypothesis remains to be tested. The zebrafish lateral

line is an excellent pre-clinical tool for drug discovery and

lead compound optimization for further translational work;

future research will address the extent of analog protection

in mammals. From our zebrafish analog screen, we identified

several analogs, particularly BA-3, BA-4, and BA-7, as leads

for future studies; these potent compounds (ED₅₀< 0.25 µM

FUNDING

This research was funded by the National Institutes of Health

Award R15DC013900 and by funds from Washington State

University, both to AC. This work was also supported by

National Institute on Deafness and Other Communication

Disorders Intramural Research Program Grant 1ZIADC000085-

01 to KK.

ACKNOWLEDGMENTS

We thank R. Boney for assistance with early experiments, E.

Cooper, M. Sokolova, R. Tsay, and M. Macpherson for fish

husbandry, K. Lawton for assistance with the calcium imaging

experiments, and P.S. Steyger for insightful discussions. We also

thank two anonymous reviewers for helpful comments.

SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online

at: https://www.frontiersin.org/articles/10.3389/fncel.2020.002

34/full#supplementary-material.

REFERENCES

Alharazneh, A., Luk, L., Huth, M., Monfared, A., Steyger, P. S., Cheng, A. G.,

et al. (2011). Functional hair cell mechanotransducer channels are required

for aminoglycoside ototoxicity. PLoS One 6:e22347. doi: 10.1371/journal.pone.

0022347

Bainbridge, K. E., and Wallhagen, M. I. (2014). Hearing loss in an

aging american population: extent, impact, and management. Annu.

Rev. Public Health 35, 139–152. doi: 10.1146/annurev-publhealth-032013-

182510

Abbott, N. J., Rönnbäck, L., and Hansson, E. (2006). Astrocyte-endothelial

interactions at the blood-brain barrier. Nat. Rev. Neurosci. 7, 41–53.

doi: 10.1038/nrn1824

Alaoui, C. E., Choukairi, Z., Lamri, L., Berrada, F., and Fechtali, T.

(2015). Inhibition of voltage-gated calcium channels by natural alkaloids:

pharmacological and therapeutic effects. Int. J. Sci. Res. 4, 1289–1294.

Frontiers in Cellular Neuroscience | www.frontiersin.org

16

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

Brown, A. D., Mussen, T. D., Sisneros, J. A., and Coffin, A. B. (2010). Reevaluating

the use of aminoglycoside antibiotics in behavioral studies of the lateral line.

Hear. Res. 272, 1–4. doi: 10.1016/j.heares.2010.10.014

Chowdhury, S., Owens, K. N., Herr, R. J., Jiang, Q., Chen, X., Johnson, G.,

et al. (2018). Phenotypic optimization of urea-thiophene carboxamides

to yield potent, well tolerated and orally active protective agents

against aminoglycoside-induced hearing loss. J. Med. Chem. 61, 84–97.

doi: 10.1021/acs.jmedchem.7b00932

Coffin, A. B., Ou, H., Owens, K. N., Santos, F., Simon, J. A., Rubel, E. W., et al.

(2010). Chemical screening for hair cell loss and protection in the zebrafish

lateral line. Zebrafish 7, 3–11. doi: 10.1089/zeb.2009.0639

Coffin, A. B., Reinhart, K. E., Owens, K. N., Raible, D. W., and Rubel, E. W. (2009).

Extracellular divalent cations modulate aminoglycoside-induced hair cell death

in the zebrafish lateral line. Hear. Res. 253, 42–51. doi: 10.1016/j.heares.2009.

03.004

Kirkwood, N. K., O’Reilly, M., Derudas, M., Kenyon, E. J., Huckvale, R.,

van Netten, S. M., et al. (2017). D-Tubocurarine and berbamine: alkaloids

that are permeant blockers of the hair cell’s mechano-electrical transducer

channel and protect from aminoglycoside toxicity. Front. Cell. Neurosci. 11:262.

doi: 10.3389/fncel.2017.00262

Kitcher, S. R. R., Kirkwood, N. K., Camci, E. D., Wu, P., Gibson, R. M.,

Redila, V. A., et al. (2019). ORC-13661 protects sensory hair cells

from aminoglycoside and cisplatin ototoxicity. JCI Insight 4:e126764.

doi: 10.1172/jci.insight.126764

Koo, J.-W., Quintanilla-Dieck, L., Jiang, M., Liu, J., Urdang, Z. D.,

Allensworth, J. J., et al. (2015). Endotoxemia-mediated inflammation

potentiates aminoglycoside-induced ototoxicity. Sci. Transl. Med. 7,

298ra118–298ra118. doi: 10.1126/scitranslmed.aac5546

Kruger, M., Boney, R., Ordoobadi, A. J., Sommers, T. F., Trapani, J. G., and

Coffin, A. B. (2016). Natural bizbenzoquinoline derivatives protect zebrafish

lateral line sensory hair cells from aminoglycoside toxicity. Front. Cell.

Neurosci. 10:83. doi: 10.3389/fncel.2016.00083

Liang, Y., He, X., Li, X., Zhang, X., Zhang, X., Zhang, L., et al. (2016). 4-

chlorbenzoyl berbamine, a novel derivative of the natural product berbamine,

potently inhibits the growth of human myeloma cells by modulating

the NF-kB and JNK signalling pathways. Cancer Invest. 34, 496–505.

doi: 10.1080/07357907.2016.1235709

Lukasz, D., and Kindt, K. S. (2018). In vivo calcium imaging of lateral-line hair cells

in larval zebrafish. J. Vis. Exp. 141:10.3791/58794. doi: 10.3791/58794

Majumder, P., Moore, P. A., Richardson, G. P., and Gale, J. E. (2017).

Protecting mammalian hair cells from aminoglycoside-toxicity: assessing

phenoxybenzamine’s potential. Front. Cell. Neurosci. 11:94. doi: 10.3389/fncel.

2017.00094

Coffin, A. B., Rubel, E. W., and Raible, D. W. (2013a). Bax, Bcl2, and

P53 differentially regulate neomycin- and gentamicin-induced hair cell

death in the zebrafish lateral line. J. Assoc. Res. Otolaryngol. 14, 645–659.

doi: 10.1007/s10162-013-0404-1

Coffin, A. B., Williamson, K. L., Mamiya, A., Raible, D. W., and Rubel, E. W.

(2013b). Profiling drug-induced cell death pathways in the zebrafish lateral line.

Apoptosis 18, 393–408. doi: 10.1007/s10495-013-0816-8

Corey, D. P., García-Añoveros, J., Holt, J. R., Kwan, K. Y., Lin, S.-

Y., Vollrath, M. A., et al. (2004). TRPA1 is

a candidate for the

mechanosensitive transduction channel of vertebrate hair cells. Nature 432,

723–730. doi: 10.1038/nature03066

Du, G.-H., Song, X.-Y., Kong, L.-L., and Chen, N.-H. (2018). Berbamine,’’ in

Natural Small Molecule Drugs from Plants, ed. G. Du (Singapore: Springer),

485–489.

Esterberg, R., Coffin, A. B., Ou, H., Simon, J. A., Raible, D. W., and Rubel, E. W.

(2013). Fish in a dish: drug discovery for hearing habilitation. Drug Discov.

Today Dis. Models 10:10.1016/j.ddmod.2012.02.001. doi: 10.1016/j.ddmod.

2012.02.001

Farris, H. E., LeBlanc, C. L., Goswami, J., and Ricci, A. J. (2004). Probing the pore

of the auditory hair cell mechanotransducer channel in turtle. J. Physiol. 558,

769–792. doi: 10.1113/jphysiol.2004.061267

Gale, J. E., Marcotti, W., Kennedy, H. J., Kros, C. J., and Richardson, G. P. (2001).

FM1–43 dye behaves as a permeant blocker of the hair-cell mechanotransducer

channel. J. Neurosci. 21, 7013–7025. doi: 10.1523/JNEUROSCI.21-18-07

013.2001

Marcotti, W., van Netten, S. M., and Kros, C. J. J. (2005). The aminoglycoside

antibiotic dihydrostreptomycin rapidly enters mouse outer hair cells through

the mechano-electrical transducer channels. J. Physiol. 567, 505–521.

doi: 10.1113/jphysiol.2005.085951

Medeiros, M. A. A., Pinho, J. F., De-Lira, D. P., Barbosa-Filho, J. M.,

Araújo, D. A. M., Cortes, S. F., et al. (2011). Curine, a bisbenzylisoquinoline

2+

channels and decreases intracellular Ca

alkaloid, blocks L-type Ca

transients in A7r5 cells. Eur. J. Pharmacol. 669, 100–107. doi: 10.1016/j.ejphar.

2011.07.044

Mick, P., Kawachi, I., and Lin, F. R. (2014). The association between hearing loss

and social isolation in older adults. Otolaryngol. Head Neck Surg. 150, 378–384.

doi: 10.1177/0194599813518021

Hailey, D. W., Esterberg, R., Linbo, T. H., Rubel, E. W., and Raible, D. W.

(2017). Fluorescent aminoglycosides reveal intracellular trafficking routes in

mechanosensory hair cells. J. Clin. Invest. 127, 472–486. doi: 10.1172/jci85052

Harris, J. A., Cheng, A. G., Cunningham, L. L., MacDonald, G., Raible, D. W., and

Rubel, E. W. (2003). Neomycin-induced hair cell death and rapid regeneration

in the lateral line of zebrafish (danio rerio). J. Assoc. Res. Otolaryngol. 4,

219–234. doi: 10.1007/s10162-002-3022-x

Huth, M. E., Ricci, A. J., and Cheng, A. G. (2011). Mechanisms of aminoglycoside

ototoxicity and targets of hair cell protection. Int. J. Otolaryngol. 2011:937861.

doi: 10.1155/2011/937861

Ji, H.-F., Li, X.-J., and Zhang, H.-Y. (2009). Natural products and drug discovery.

Can thousands of years of ancient medical knowledge lead us to new and

powerful drug combinations in the fight against cancer and dementia? EMBO

Rep. 10, 194–200. doi: 10.1038/embor.2009.12

Murakami, S. L., Cunningham, L. L., Werner, L. A., Bauer, E., Pujol, R.,

Raible, D. W., et al. (2003). Developmental differences in susceptibility to

neomycin-induced hair cell death in the lateral line neuromasts of zebrafish

(danio rerio). Hear. Res. 186, 47–56. doi: 10.1016/s0378-5955(03)00259-4

Myrdal, S. E., and Steyger, P. S. (2005). TRPV1 regulators mediate gentamicin

penetration of cultured kidney cells. Hear. Res. 204, 170–182. doi: 10.1016/j.

heares.2005.02.005

Neveux, S., Smith, N. K., Roche, A., Blough, B. E., Pathmasiri, W., and Coffin, A. B.

(2016). Natural compounds as occult ototoxins? Ginkgo biloba flavonoids

moderately damage lateral line hair cells. J. Assoc. Res. Otolaryngol. 18, 275–289.

doi: 10.1007/s10162-016-0604-6

Nyberg, S., Abbott, N. J., Shi, X., Steyger, P. S., and Dabdoub, A. (2019). Delivery of

therapeutics to the inner ear: the challenge of the blood-labyrinth barrier. Sci.

Transl. Med. 11:eaao0935. doi: 10.1126/scitranslmed.aao0935

Jia, X.-J., Li, X., Wang, F., Liu, H.-Q., Zhang, D.-J., and Chen, Y. (2017). Berbamine

exerts anti-inflammatory effects via inhibition of NF-KB and MAPK

signaling pathways. Cell. Physiol. Biochem. 41, 2307–2318. doi: 10.1159/0004

75650

Jung, D., and Bhattacharyya, N. (2012). Association of hearing loss with decreased

employment and income among adults in the united states. Ann. Otol. Rhinol.

Laryngol. 121, 771–775. doi: 10.1177/000348941212101201

Karasawa, T., Wang, Q., Fu, Y., Cohen, D. M., and Steyger, P. S. (2008).

TRPV4 enhances the cellular uptake of aminoglycoside antibiotics. J. Cell Sci.

121, 2871–2879. doi: 10.1242/jcs.023705

Kenyon, E. J., Kirkwood, N. K., Kitcher, S. R. R., O’Reilly, M., Derudas, M.,

Cantillon, D. M., et al. (2017). Identification of ion-channel modulators that

protect against aminoglycoside-induced hair cell death. JCI Insight 2:e96773.

doi: 10.1172/jci.insight.96773

O’Reilly, M., Kirkwood, N. K., Kenyon, E. J., Huckvale, R., Cantillon, D. M.,

Waddell, S. J., et al. (2019). Design, synthesis and biological evaluation of

a new series of carvedilol derivatives that protect sensory hair cells from

aminoglycoside-induced damage by blocking the mechanoelectrical transducer

channel. J. Med. Chem. 62, 5312–5329. doi: 10.1021/acs.jmedchem.8b

01325

O’Sullivan, M. E., Perez, A., Lin, R., Sajjadi, A., Ricci, A. J., and Cheng, A. G.

(2017). Towards the prevention of aminoglycoside-related hearing loss. Front.

Cell. Neurosci. 11:325. doi: 10.3389/fncel.2017.00325

Ou, H. C., Cunningham, L. L., Francis, S. P., Brandon, C. S., Simon, J. A.,

Raible, D. W., et al. (2009). Identification of FDA-approved drugs and

bioactives that protect hair cells in the zebrafish (danio rerio) lateral line

and mouse (mus musculus) utricle. J. Assoc. Res. Otolaryngol. 10, 191–203.

doi: 10.1007/s10162-009-0158-y

Frontiers in Cellular Neuroscience | www.frontiersin.org

17

July 2020 | Volume 14 | Article 234

Hudson et al.

Otoprotection of Berbamine Analogs

Ou, H. C., Keating, S., Wu, P., Simon, J. A., Raible, D. W., and Rubel, E. W.

(2012). Quinoline ring derivatives protect against aminoglycoside-induced hair

cell death in the zebrafish lateral line. J. Assoc. Res. Otolaryngol. 13, 759–770.

doi: 10.1007/s10162-012-0353-0

Ou, H. C., Santos, F., Raible, D. W., Simon, J. A., and Rubel, E. W. (2010).

Drug screening for hearing loss: using the zebrafish lateral line to screen for

drugs that prevent and cause hearing loss. Drug Discov. Today 15, 265–271.

doi: 10.1016/j.drudis.2010.01.001

Owens, K. N., Coffin, A. B., Hong, L. S., O’Connell Bennett, K., Rubel, E. W., and

Raible, D. W. (2009). Response of mechanosensory hair cells of the zebrafish

lateral line to aminoglycosides reveals distinct cell death pathways. Hear. Res.

253, 32–41. doi: 10.1016/j.heares.2009.03.001

Owens, K. N., Santos, F., Roberts, B., Linbo, T., Coffin, A. B., Knisely, A. J.,

et al. (2008). Identification of genetic and chemical modulators of zebrafish

mechanosensory hair cell death. PLoS Genet. 4:e1000020. doi: 10.1371/journal.

pgen.1000020

Steyger, P. S., Peters, S. L., Rehling, J., Hordichok, A., and Dai, C. F. (2003). Uptake

of gentamicin by bullfrog saccular hair cells in vitro. J. Assoc. Res. Otolaryngol.

4, 565–578. doi: 10.1007/s10162-003-4002-5

Teh, B. S., Seow, W. K., Li, S. Y., and Thong, Y. H. (1990). Inhibition of

prostaglandin and leukotriene generation by the plant alkaloids tetrandrine

and berbamine. Int. J. Immunopharmacol. 12, 321–326. doi: 10.1016/0192-

0561(90)90088-5

Teles, M. M. R. S., Vieira Pinheiro, A. A., Da Silva Dias, C., Tavares, J. F., Barbosa

Filho, J. M., and Leitão Da Cunha, E. V. (2019). The alkaloids: chemistry

and biology. Alkaloids Chem. Biol. 82, 147–304. doi: 10.1016/bs.alkal.2018.

11.002

Thomas, A. J., Wu, P., Raible, D. W., Rubel, E. W., Simon, J. A., and Ou, H. C.

(2015). Identification of small molecule inhibitors of cisplatin-induced hair cell

death: results of a 10,000 compound screen in the zebrafish lateral line. Otol.

Neurotol. 36, 519–525. doi: 10.1097/mao.0000000000000487

Tian, Y., and Zheng, J. (2017). Chapter six

- metabolic activation and

Perrey, D. A., German, N. A., Gilmour, B. P., Li, J.-X., Harris, D. L., Thomas, B. F.,

et al. (2013). Substituted tetrahydroisoquinolines as selective antagonists

for the orexin 1 receptor. J. Med. Chem. 56, 6901–6916. doi: 10.1021/jm4

00720h

Portmann, M., Darrouzet, J., and Coste, C. H. (1974). Distribution within

the cochlea of dihydrostreptomycin injected into the circulation: an

autoradiographic and electron microscopic study. Arch. Otolaryngol. 100,

473–475. doi: 10.1001/archotol.1974.00780040487014

toxicities of bis-benzylisoquinoline alkaloids. Adv. Mol. Toxicol. 11, 241–272.

doi: 10.1016/B978-0-12-812522-9.00006-3

Ton, C., and Parng, C. (2005). The use of zebrafish for assessing ototoxic and

otoprotective agents. Hear. Res. 208, 79–88. doi: 10.1016/j.heares.2005.05.005

Vlasits, A. L., Simon, J. A., Raible, D. W., Rubel, E. W., and Owens, K. N. (2012).

Screen of FDA-approved drug library reveals compounds that protect hair cells

from aminoglycosides and cisplatin. Hear. Res. 294, 153–165. doi: 10.1016/j.

heares.2012.08.002

Raible, D. W., and Kruse, G. J. (2000). Organization of the lateral line system in

embryonic zebrafish. J. Comp. Neurol. 421, 189–198. doi: 10.1002/(sici)1096-

9861(20000529)421:2 <189::aid-cne5 >3.0.co;2-k

Raviv, D., Dror, A. A., and Avraham, K. B. (2010). Hearing loss: a common

disorder caused by many rare alleles. Ann. N Y Acad. Sci. 1214, 168–179.

doi: 10.1111/j.1749-6632.2010.05868.x

Westerfield, M. (2000). The Zebrafish Book. A Guide for the Laboratory Use of

Zebrafish (Danio Rerio). 4th Edn. Univsity of Oregon Press: Eugenecpubn.

Wiedenhoft, H., Hayashi, L., and Coffin, A. B. (2017). PI3K and inhibitor

of apoptosis proteins modulate gentamicin-induced hair cell death in the

zebrafish lateral line. Front. Cell. Neurosci. 11:326. doi: 10.3389/fncel.2017.

00326

Ren, Y., Lu, L., Guo, T. B., Qiu, J., Yang, Y., Liu, A., et al. (2008).

Novel immunomodulatory properties of berbamine through selective

down-regulation of STAT4 and action of ifn-gamma in experimental

autoimmune encephalomyelitis. J. Immunol. 181, 1491–1498. doi: 10.

4049/jimmunol.181.2.1491

Santos, F., MacDonald, G., Rubel, E. W., and Raible, D. W. (2006). Lateral line hair

cell maturation is a determinant of aminoglycoside susceptibility in zebrafish

(Danio Rerio). Hear. Res. 213, 25–33. doi: 10.1016/j.heares.2005.12.009

Schulz, M., and Schmoldt, A. (2003). Therapeutic and toxic blood concentrations

of more than 800 drugs and other xenobiotics. Pharmazie 58, 447–474.

doi: 10.1016/bs.alkal.2018.11.002

Zhang, Q., Li, S., Wong, H.-T. C., He, X. J., Beirl, A., Petralia, R. S., et al. (2018).

Synaptically silent sensory hair cells in zebrafish are recruited after damage.

Nat. Commun. 9:1388. doi: 10.1038/s41467-018-03806-8

Conﬂict of Interest: The authors declare that the research was conducted in the

absence of any commercial or financial relationships that could be construed as a

potential conflict of interest.

This is an open-access article distributed under the terms of the Creative Commons

Attribution License (CC BY). The use, distribution or reproduction in other forums

is permitted, provided the original author(s) and the copyright owner(s) are credited

and that the original publication in this journal is cited, in accordance with accepted

academic practice. No use, distribution or reproduction is permitted which does not

comply with these terms.

Stepanyan, R. S., Indzhykulian, A. A., Vélez-Ortega, C. A., Boger, E. T.,

Steyger, P. S., Friedman, T. B., et al. (2011). TRPA1-Mediated Accumulation

of Aminoglycosides in Mouse Cochlear Outer Hair Cells. J. Assoc. Res.

Otolaryngol. 12, 729–740. doi: 10.1007/s10162-011-0288-x

Frontiers in Cellular Neuroscience | www.frontiersin.org

18

July 2020 | Volume 14 | Article 234

Article Doi

Berbamine Analogs Exhibit Differential Protective Effects From Aminoglycoside-Induced Hair Cell Death

DOI: 10.3389/fncel.2020.00234

Source and publish data:

Authors:

Article abstract of DOI:10.3389/fncel.2020.00234

Full text of DOI:10.3389/fncel.2020.00234

Products guided by the article

R&D Labs maybe for 41898-71-9

Relevant to this article

Hot Product