University of Southern California Dissertations and Theses

Investigation of the molecular mechanisms of ototoxicity

(USC Thesis Other)

Investigation of the molecular mechanisms of ototoxicity

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
INVESTIGATION OF THE MOLECULAR MECHANISMS
OF OTOTOXICITY
by
Litao Tao

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GENETIC, MOLECULAR AND CELLULAR BIOLOGY)

August 2014
Copyright 2014 Litao Tao

i
Dedication

This work is dedicated to
my parents, Chengjun Tao and Xiangying Yang,
my wife, Qinxian Pei,
and my daughter, April P. Tao,
for their unconditional love and support.

ii
Acknowledgements
I would like to thank my advisor, Dr. Neil Segil, for his encouragement and
support over the past seven years. Thank you for your guidance in
formulating scientific questions, for teaching me how to solve problems and
for demonstrating how to be an independent scientist.
I also want to thank current and previous members in the Segil lab for their
inspiring ideas and valuable suggestions. Thank you all for your thoughtful
discussions on my project and your patient explanations on experimental
skills. Particularly, thank you to Juan Llamas and Welly Makmura for taking
care of the animals, and thank you to Sum-Yan Ng and Dr. Robert Rainey
for collaborating on some experiments.
I truly appreciate the opportunity to conduct research in the intellectually
stimulating environment of House Research Institute and the University of
Southern California.

iii
Table of Contents
Dedication.........................................................................................................i
Acknowledgements......................................................................................... ii
List of Tables ..................................................................................................vi
List of Figures............................................................................................... vii
Abstract:..........................................................................................................ix
Chapter 1: Introduction....................................................................................1
1.1 Aminoglycoside antibiotic ototoxicity: .................................................1
1.2 Ototoxicity of Cisplatin .........................................................................3
1.3 Activation of JNK pathway in hair cells ...............................................5
1.4 Cell cycle machinery activation in postmitotic cells.............................7
1.5 Transcriptome and proteome study in cochlea....................................11
Chapter 2: Characterization of organotypic culture of cochleae from
neonatal mice.......................................................................................15
2.1 Organotypic culture of cochleae from neonatal mice..........................15
2.2 Gentamicin-induced hair cell death through apoptosis in cultures .....16
2.3 Discussion............................................................................................24
Chapter 3: Involvement of CDK2 in aminoglycoside ototoxicity ................26
3.1 Effect of pharmaceutical CDK inhibitors............................................26
3.2 Protective effect of CDK2 mutation on hair cell survival
following Gentamicin exposure ..........................................................29
3.3 Delay of Gentamicin-induced apoptosis..............................................32
3.4 Failure to detect cell cycle re-entry in Gentamicin-treated hair
cells......................................................................................................36
3.5 Regulation of c-Jun/AP1 transcription activity ...................................39
3.6 Roles of CDK2 in mature utricular hair cells......................................43
3.7 Drug sensitivity in mature animals......................................................51
3.8 Discussion............................................................................................55

iv
Chapter 4: CDK activity and Cisplatin ototoxicity .......................................59
4.1 Cisplatin-DNA adduct forming in hair cells .......................................60
4.2 Effect of pharmaceutical CDK inhibitors on cisplatin-induced
hair cell death ......................................................................................61
4.3 Discussion............................................................................................63
Chapter 5: Involvement of other factors in aminoglycoside ototoxicity.......65
5.1 No effect of knocking out Cdk4 on Gentamicin ototoxicity ...............65
5.2 Effects of knocking out CKIs

on hair cell sensitivity to
Gentamicin ..........................................................................................69
5.3 No effect of p53 KO on Gentamicin-induced hair cell death..............72
Chapter 6: Gene expression profile in hair cells after aminoglycoside
treatment..............................................................................................76
6.1 Experiment Design ..............................................................................76
6.2 Q-PCR validation.................................................................................77
6.3 Principle Component Analysis ............................................................79
6.4 Differential Gene Expression Analysis ...............................................81
6.5 Gentamicin-induced gene expression responses in hair cells .............82
Cell cycle machinery and cell cycle re-entry ........................................83
General Transcription apparatus............................................................88
DNA damage and repair........................................................................91
Protein translation and ribosome synthesis ...........................................94
Mitochondria....................................................................................... 101
Stress and cell death ........................................................................... 105
JNK pathway ...................................................................................... 108
NF- κB pathway................................................................................... 110
6.6 Pathway analysis of non-hair cell expression profiles ......................113
6.7 Similarity between aminoglycoside ototoxicity and
nephrotoxicity....................................................................................115
6.8 Discussion..........................................................................................120
Chapter 7: Effect of mCherry expression in hair cells ................................127
7.1 Less sensitivity to otoxins in mCherry-positive hair cells ................130
7.2 Hearing loss at young age in Atoh1-mCherry mice ..........................135
7.3 Discussion..........................................................................................136

v
Chapter 8: Materials and methods:..............................................................138
Organotypic culture and drug treatment..................................................138
Fluorescent immunostaining and western blotting..................................139
Cell purification and Quantitative-PCR ..................................................140
RNA sequencing and data analysis .........................................................141
ABR and Trans-tympanic injection.........................................................141
Quantification and statistics.....................................................................142
DNA extraction and Genotyping.............................................................143
References....................................................................................................144

vi
List of Tables
Table 1. Primer pairs for Q-PCR validation..................................................78
Table 2. Expression changes in genes involved in cell cycle........................86
Table 3. Expression changes in genes coding for general transcription
apparatus. ..........................................................................................90
Table 4. Expression changes of genes involved in DNA damage and
repair, cell cycle checkpoint. ............................................................93
Table 5. Expression changes in genes coding translation machinery
factors. ..............................................................................................98
Table 6. Expression changes of mitochondrial related genes. ................... 102
Table 7. Expression changes of stress response genes and cell death
genes. ..............................................................................................107
Table 8. Expression changes of genes involved in JNK pathway.............. 110
Table 9. Expression changes of genes involved in NF- κB pathway......... 112
Table 10. Altered gene expression in non-hair cells .................................. 114
Table 11. Gentamicin-induced similar gene expression changes in hair
cells and in kidney. .........................................................................118
Table 12. Primer pairs for Q-PCR.............................................................. 140
Table 13. Primer sequence for Genotyping................................................ 143

vii
List of Figures
Figure 1. Gentamicin accumulates in hair cells.............................................17
Figure 2. Gentamicin causes hair cell loss in a dose-dependent manner ......19
Figure 3. Disappearance of Atoh1-GFP reflects hair cell damage................21
Figure 4. Gentamicin induces hair cell death through apoptosis ..................23
Figure 5. Gentamicin damages hair cells but leave supporting cells
intact..................................................................................................23
Figure 6. CDK inhibitors protect hair cells against Gentamicin ...................28
Figure 7. Cdk2 KO organs are less sensitive to Gentamicin.........................31
Figure 8. CDK2 inhibitor delays Gentamicin-induced hair cell
apoptosis ...........................................................................................35
Figure 9. Failure to detect cell cycle re-entry by biochemistry methods ......38
Figure 10. Transcription activity of c-Jun is CDK dependent ......................42
Figure 11. Gentamicin induces hair cells loss in utricles ..............................45
Figure 12. CDK2 inhibitor CVT-313 protects mature utricular hair cells
against Gentamicin. ..........................................................................46
Figure 13. Cdk2 KO utricles are less sensitive to Gentamicin......................48
Figure 14. CDK2 has a similar regulatory role on c-Jun transcription
ativity ................................................................................................50

viii
Figure 15. Cdk2 KO animals are less sensitive to Gentamicin .....................54
Figure 16. Strong absorption of cisplatin by hair cells..................................60
Figure 17. Cisplatin-induced hair cell death is exacerbated by CDK
inhibitors ...........................................................................................62
Figure 18. Cdk4 knockout does not affect hair cell sensitive to
Gentamicin........................................................................................68
Figure 19. Effect of knocking out CKIs on Gentamicin sensitivity..............71
Figure 20. No effect of knocking out p53 on hair cell response to
Gentamicin........................................................................................74
Figure 21. RNA sequencing data is validated by Q-PCR .............................77
Figure 22. Principle Component Analysis shows the most significant
fators affecting gene expression. ......................................................80
Figure 23. Hair cells in Atoh1-mCherry organs are resistant to H
2
O
2

and cisplatin....................................................................................131
Figure 24. mCherry positive Hair cell are less sensitive to Gentamicin-
induced hair cell apoptosis ............................................................ 134
Figure 25. Atoh1-mCherry mice experience hearing loss and hair cell
damage at a very young age .......................................................... 136

ix
Abstract
Sensory hair cells are essential for transforming the mechanical vibrations of
sound into electric signals that our nervous system can interpret. However,
sensory hair cells are sensitive to a variety of stresses, including
aminoglycoside antibiotics, chemotherapy agents, and environmental noise.
The molecular mechanisms of ototoxicity have been under investigation for
several decades, yet little is known about the underlying signaling pathways.
In Chapter 2 and 3 of this study, we adapted an established technique of
organotypic culture of perinatal mouse cochlea to investigate the ototoxicity.
With this culture model, we found that CDK2 activity was involved in
aminoglycoside-induced hair cell death and that one consequence of its
activity was to mediate the transcriptional activity of c-Jun. The involvement
of CDK2 was further supported by evidence from utricular cultures and in
vivo experiments. Using knockout mice, we demonstrated that p19
ink4d
,
p21
kip1
and p53 were not involved in aminoglycoside ototoxicity. In Chapter
6, we profiled the gene expression changes in hair cells after aminoglycoside
treatment and found altered expression for genes involved in the general
transcription apparatus, translation machinery, mitochondria, DNA damage

x
and repair, stress and apoptosis, JNK pathway and NF- κB pathway. Finally,
we found that the sensitivity to ototoxins was changed in mCherry-H2B
positive hair cells. Our results create a framework for understanding the
molecular mechanisms of ototoxicity and the signaling pathways underlying
postmitotic hair cell apoptosis.

1
Chapter 1: Introduction
1.1 Aminoglycoside antibiotic ototoxicity:
Aminoglycosides are a group of molecules consisting of several six- or five-
carbon rings connected by glycosidic bonds with amino groups attached to
those rings, such as Gentamicin and streptomycin (Forge and Schacht, 2000).
Since the discovery of the first aminoglycoside antibiotic streptomycin in
1943, aminoglycoside antibiotics have been used worldwide due to their
broad anti-bacterial spectrum and excellent bactericidal effect.
Aminoglycoside antibiotics are found to interact specifically with bacterial
16S ribosomal RNA and consequentially inhibit protein synthesis in bacteria
(Blanchard et al., 1998). Most eukaryotic cells are not affected by
aminoglycoside antibiotics due to the different structures of eukaryotic
ribosomal RNA. However, sensory hair cells in the inner ear and proximal
tubular cells in the kidney are sensitive to aminoglycoside antibiotics,
resulting in hearing loss and kidney failure in patients that receive
aminoglycoside treatments.

2
Aminoglycoside antibiotics can be taken up specifically by sensory hair cells
and accumulate inside the cells. Texas Red-conjugated Gentamicin or
3
H-
labeled Gentamicin accumulates both in inner and outer hair cells in
organotypic culture and in live animals (Wang and Steyger, 2009;
Richardson et al., 1997). The entry and accumulation of Gentamicin require
functional mechanotransducer channels (Alharazneh et al., 2011). MyoVIIa
is an unconventional myosin that is expressed specifically in hair cells and is
essential for stereocilia organization and functional mechanotransducer
channels. Mutations in the MyoVIIa gene result in deaf animals secondary to
malfunctioning of stereocilia, and there is no accumulation of Gentamicin in
hair cells in MyoVIIa mutant mice (Richardson et al., 1997). Consistent with
this observation, blocking mechanotransducer channels inhibits the
accumulation of Gentamicin in hair cells (Alharazneh et al., 2011).
Aminoglycoside antibiotic accumulation in hair cells leads to hair cell
damage in cochlea and hearing loss. Hearing loss is usually permanent
because regeneration of hair cells is not possible in the mammalian organ of
Corti (Corwin and Cotanche, 1988). The molecular mechanisms of
aminoglycoside ototoxicity have been under investigation for decades.

3
Reactive Oxygen Species (ROS) produced by aminoglycosides is thought to
be a major cytotoxic factor. ROS level has been observed to increase after
aminoglycoside treatment both in vitro and in vivo (Priuska and Schacht,
1995; Clerici et al., 1996). Activation of the intrinsic mitochondrial
apoptosis pathway is thought to play a role in aminoglycoside-induced hair
cell death (Forge and Li, 2000; Cunningham et al., 2002). Overexpression of
anti-apoptotic Bcl2 protein protects hair cells from aminoglycoside
ototoxicity (Cunningham et al., 2004), suggesting mitochondria integrity is
important for hair cell survival. Inhibition of Caspase 9 and Caspase 3 also
confers protection to hair cells, while inhibition of Caspase 8 has little effect
(Matsui et al., 2002), indicating that the mitochondrial-Caspase signaling
pathway plays an essential role in aminoglycoside-induced hair cell
apoptosis.
1.2 Ototoxicity of Cisplatin
In addition to aminoglycoside antibiotics, many chemotherapeutic drugs
have been found to be ototoxic in clinical practice, especially platinum-
based antineoplatic drugs such as cisplatin and carboplatin (Rybak, 1999).
Cisplatin is an effective agent in treating testicular, ovarian, bladder, lung

4
and other cancers; however, the ototoxicity and nephrotoxicity of cisplatin
have been reported since its introduction to clinical practices in 1978.
Stable and inert under high chloride condition, as in the blood stream, the
chloride ligands of cisplatin are sequentially replaced by H
2
O in the cells
which have low chloride concentrations, to form highly reactive molecules,
the so-called positive-charged aqua complex. Those monohydrated or
dihydrated molecules are attracted by negatively charged molecules and then
react with the nucleophilic center, such as the N7 atom of guanine (Wang
and Lippard, 2005). After reacting with DNA, cisplatin molecules form
intrastrand adducts or interstrand crosslinks, which interrupt DNA
replication and transcription. Dividing cells, especially fast growing tumor
cells, are highly sensitive to cisplatin because the existence of excessive
DNA damage during DNA replication triggers apoptosis (Florea and
Büsselberg, 2011).
As non-dividing cells, hair cells are also sensitive to platinum-based drugs.
The oxidative stress induced by cisplatin is thought to be a major factor in
cisplatin ototoxicity. ROS production has been observed in cochlear explants

5
after cisplatin treatment (Clerici et al., 1996); and depletion of antioxidant
enzymes in cochlea tissue from animals treated with an ototoxic dose of
cisplatin has also been reported (Rybak et al., 2000). Administration of
antioxidants, such as superoxide dismutase (SOD) and glutathione (GSH),
ameliorates cisplatin ototoxicity (Rybak et al., 2000), as observed with
aminoglycoside antibiotics. In addition to ROS production, cisplatin-DNA
adducts can also be detected in hair cells, and build-up of unrepaired
cisplatin-DNA adducts contributes to the ototoxicity of cisplatin. (Van
Ruijven et al., 2005)
1.3 Activation of JNK pathway in hair cells
Activation of the mitochondrial apoptosis pathway is a downstream
executive event, which activates the Caspase cascade to execute
programmed cell death, and this pathway is regulated by upstream signals.
The c-Jun N-terminal Kinase (JNK) pathway is thought to be an important
upstream signaling pathway that regulates apoptosis in hair cells after
aminoglycoside antibiotic treatment (Pirvola et al., 2000).

6
In cells, JNK protein can be phosphorylated and activated by upstream
kinase MKK, and the then active JNK phosphorylates its substrates. The
transcription factor c-Jun, which heterodimerizes with c-Fos or ATF to form
Activator Protein 1 (AP1), is one target of active JNK and gets
phosphorylated by JNK at S63 and S73. S63/S73 phosphorylation on c-Jun
upregulates the transcriptional activity of AP1 and the expression of AP1
target genes. AP1 will bind to TRE (TAP Response Element, 5’-TGA(C/G)
TCA -3’) sites and initiate the transcription of genes containing this
consensus sequences. Several TRE sites exist in the promoter region of the
c-Jun gene, and the expression of c-Jun is auto-regulated by c-Jun/AP1 itself
(Angel et al., 1988). AP1 also regulates the expression of apoptotic factors,
such as Bim and FasL, implicating the JNK pathway in regulating apoptosis
(Hess et al., 2004).
In hair cells, activation of the JNK pathway has been found to help mediate
aminoglycoside-induced hair cell death. Phosphorylated JNK and
phosphorylated c-Jun can be detected in hair cells after Gentamicin
treatment both in organotypic cultures and in live animals, suggesting the
activation of JNK pathway in hair cells after Gentamicin treatment.

7
Additionally, when the JNK pathway is blocked by pharmaceutical
inhibitors, the organ of Corti is less sensitive to Gentamicin and more hair
cells survive (Pirvola et al., 2000; Ylikoski et al., 2002; Sugahar et al., 2006).
1.4 Cell cycle machinery activation in postmitotic cells
Some organs such as liver and skin can recover from injury because cells in
those organs can divide and repopulate the damaged organs. During the
process of cell division, Cyclin-Depdent Kinases (CDK) are activated
chronologically by the expression of CDK coactivators, cyclins. Cyclin D is
expressed in the G1 phase and activates CDK4 and CDK6, promoting the
passage of cells through the G1 phase; expression of Cyclin E starts at late
G1 phase and lasts through S phase; CDK2 is activated by cyclin E,
promoting the G1-S transition; Cyclin A is expressed in S phase and early
G2 phase; Cyclin A activates CDK2 in S phase and CDK1 in early G2 phase;
then cells start to express Cyclin B in G2 phase and Cyclin B activates
CDK1, helping cells pass through G2 and M phase to finish the cell cycle
(Morgan, 1997).

8
Unlike liver or skin, the organ of Corti of mammals cannot regenerate after
hair cell damage, because hair cells and supporting cells are terminally
differentiated non-dividing cells (Ruben, 1967). Hair cells and supporting
cells maintain their postmitotic state by actively expressing Cyclin-
dependent Kinase Inhibitors (CKI) such as p19
Ink4d
, p21
kip1
and p27
kip1
.
When CKI genes are mutated, hair cells lose the control of the postmitotic
state, re-enter the cell cycle, and eventually undergo apoptosis (Chen et al,
2003; Laine et al., 2007).
Like hair cells and supporting cells, neurons are also terminally
differentiated non-dividing cells. CDK activity (except CDK5 which is
active in neurons and is not involved in the cell cycle) is repressed in
neurons under physiological conditions (Kranenburg, 1995). However, CDK
activation has been observed both in neuronal cell culture after apoptotic
stimuli and in neurodegenerative disease model animals (Nguyen et al.,
2002; Rideout et al., 2003; Appert-Collin et al., 2006; Yang et al., 2003;
Biswas et al., 2005). When neuronal cells are challenged with various
stimuli such as DNA damaging reagents, ion concentrations shift and BDNF
withdraws, expression of cyclins is detected, CDK activity is upregulated

9
prior to neuronal cell apoptosis, and inhibition of CDK activity by
pharmaceutical inhibitors protects neuronal cells (Ghahremani et al., 2002;
Appert-Collin et al., 2006). In neuron degenerative disease, aberrant cyclin
expression and cell cycle re-entry can be detected in brain tissues.
Expression of Cyclin D and Cyclin B is detected in all stages of Alzheimer’s
disease (Yang et al., 2003), and disruption of the interaction between E2F
and Rb after Rb phosphorylation by CDKs activates the E2F-dependent
apoptosis pathway (Greene et al., 2007).
CDK activation is believed to precede neuronal cell apoptosis, but how CDK
activity regulates apoptosis in those postmitotic neurons is poorly
understood. One hypothesis is that upregulated CDK activity forces
postmitotic neurons to re-enter the cell cycle, activating cell cycle
checkpoints, which leads to the excessive DNA damage and consequently
triggers apoptosis (Kruman, 2004). In neurons and terminally differentiated
cells, which do not need to replicate their genomes and thus can dispense
with repairing the bulk of the genomes, Global-Genome Repair (GGR)
which scans and maintains the integrity of the whole genome, is strongly
attenuated, with transcribed genes still being efficiently repaired by

10
Transcription-Coupled Repair (TCR). When postmitotic neurons are forced
to re-enter the cell cycle and start DNA replication, accumulated DNA
damage in non-transcribed regions is hypothesized to lead to massive
replication errors. In addition, differentiated neurons predominantly express
a highly error-prone DNA polymerase β (Rao et al., 2001), so DNA
replication in differentiated neurons itself produces additional damages. This
damage from replication errors can activate cell cycle checkpoints and
consequently trigger apoptosis. An alternative hypothesis is that CDK
activity regulates pro-apoptosis or anti-apoptosis pathways directly. Pro-
apoptotic Bcl-2 family member BAD has been identified as a substrate of
CDK1, and phosphorylation at S128 in BAD by CDK1 disrupts the
interaction between BAD and 14-3-3, releasing BAD from sequestration in
cytoplasm (Konishi et al., 2002). Phosphorylation on FoxO by CDKs has
also been characterized, and active CDKs regulate the subcellular
localization of FoxO proteins, thus regulating the expression of FoxO target
genes (Huang et al., 2006). Another potential target of CDK activity in
postmitotic neurons is c-Jun/AP1. It has been shown that c-Jun can be

11
phosphorylated and regulated directly by CDK when neurons are treated
with DNA damaging reagents (Ghahremani et al., 2002).
CDK activation prior to cell death has also been observed in other
postmitotic cells such as thymocytes (Granés et al., 2004). However, it is not
clear whether CDK activity plays similar regulatory roles in hair cell
apoptosis. Studies described in this work are intended to elucidate whether
CDK activity regulates apoptosis in hair cells after drug treatment, and to
investigate how CDK activity regulates hair cell death.
1.5 Transcriptome and proteome study in cochlea
Investigating a single pathway or a single molecule with traditional
biochemistry methods is time consuming and labor intensive. High
throughput microarrays have been utilized as an approach for transcriptome
analysis for decades; however, next generation sequencing, a newly
developed technology, is more accurate in measurement, more informative
in analysis, and more powerful in exploiting unknown regions of the genome,
and its use is becoming increasingly common in research (Mardis, 2008). In
the field of hearing research, both microarray and RNA sequencing have

12
been used to profile the gene expression changes during development,
transdifferentiation and other biological events.
Our lab previously performed microarray analysis with purified progenitor
cells, hair cells and supporting cells from embryonic or neonatal mice to
investigate the gene expression changes during inner ear development. We
found genes that are up-regulated or down-regulated during cochlear
development, as well as genes highly expressed in hair cells, but silent in
supporting cells, or vice versa (unpublished data).
To investigate the gene expression changes in supporting cells after hair cell
damage, Mark Warchol and Michael Lovett performed microarray assays to
analyze the transcriptome of chick cochlea after hair cell damage by either
laser ablation or aminoglycoside antibiotics (Hawkins et al., 2007). They
found gene expression changes in the cochlea after hair cell damage as well
as changes during supporting cell transdifferentiation, implicating the
involvement of several transcription factors and pathways in the recovery
process.

13
Microarray assays have also been used to analyze how the cochlea responds
to ototoxins (Nagy et al., 2004). Microarray analysis performed on postnatal
day 5 (P5) rat following Gentamicin treatment revealed that 103 genes are
expressed at different levels after Gentamicin treatment, with 55 genes
down-regulated and 48 genes up-regulated.
Expression study by microarray or RNA sequencing provides information
regarding gene expression at the transcription level; however, a proteomic
approach is valuable because it provides information at the translation and
post-translation level. Antibody array experiments were performed by
Jamesdaniel et al. to investigate cisplatin-induced changes at the protein
level in rats injected with cisplatin (Jamesdaniel et al., 2008). By the
antibody array, 15 proteins were found to have increased >1.5 fold and 4
proteins decreased > 0.6 fold.
The aforementioned ototoxicity studies using high-throughput methods
demonstrate changes in gene expression in the cochlea in response to
ototoxin treatment; however, information provided by those studies is
minimal due to the limitation of technology and low purity of samples. Both

14
DNA microarray and antibody array require pre-knowledge of genes and
pathways that might be involved, making it difficult to exploit previously
unknown collections of genes or proteins. Additionally, whole cochleae
were used for sample preparation, so the samples consisted of different cell
populations such as hair cells, supporting cells, spiral ganglion,
mesenchymal cells and connective tissues. Hair cells account for less than
2% of the cell population in the cochlea, making it difficult to uncover
signals from hair cells within the heterogeneous samples.
To overcome these disadvantages in previous microarray studies, we
developed a protocol to purify fluorescence marked hair cells, and those
purified samples were used for RNA sequencing to investigate the
expression profile shifts in hair cells after aminoglycoside treatment.

15
Chapter 2: Characterization of organotypic culture of cochleae
from neonatal mice
The mouse is an excellent experimental model for investigating molecular
mechanisms and signaling pathways due to the ease of manipulation of the
mouse genome, and the fact that the entire mouse genome has been
sequenced and annotated.
However, the inner ear in adult mice is buried in the bony labyrinth, making
it difficult to dissect and culture the cochlea from animals older than
postnatal day 12 (P12) (Sobkowicz et al., 1993). For this reason, we have
adapted an ex vivo organotypic system to culture the cochlea from neonatal
mice to investigate the response of hair cells to a variety of ototoxicant
stimuli.
2.1 Organotypic culture of cochleae from neonatal mice
In this work, organotypic cultures are typically established on postnatal day
1 (P1). Mice are sacrificed and the cochleae are dissected in Ca
2+
- and Mg
2+
-
free PBS; then the spiral ganglion, lateral walls and other connective tissues
are carefully removed; finally, the intact cochleae are placed onto a 22 μm

16
pore size membrane (SPI supplies) floating on culture medium and
maintained in a low oxygen incubator (5% oxygen, 5% CO
2
). This
membrane method is preferred to the convential submerged cultures because
the organ of Corti frequently detaches from the bottom in submerged culture
during incubation, especially after ototoxic drug treatment. Under these
culture conditions, cultures can survive for at least five days without obvious
disruption of outer and inner hair cells (data not shown). (See methods in
Chaper 8 for complete description.)
2.2 Gentamicin-induced hair cell death through apoptosis in cultures
To test whether cochlea can take up aminoglycoside antibiotics from culture
medium, we cultured P1 cochlea in the presence of 0.5 mM Texas-Red
conjugated Gentamicin (GTTR) (Sandoval et al., 1998; Wang and Steyger,
2009) for 30 min. Hair cells in membrane cultures took up Gentamicin in a
basal-to-apical gradient, with stronger Texas-Red fluorescence in the basal
segment and weaker signal in the apical segment (Fig 1A). Organs were also
fixed with 4% paraformaldehyde and sectioned to examine the specificity of
drug absorption. As shown in Fig 1C, Gentamicin only accumulated in hair

17

Figure 1. Gentamicin accumulates in hair cells. Texas red (from GTTR or TR) and GFP fluorescence
(green) in surface preparations (A and B) or cross sections (C) of cochlea organs after 30 min incubation
with 0.5 mM GTTR. Strong GTTR fluorescence in the base and weak signal in the apex indicate the basal-
to-apical gradient (A and C). No accumulation of unconjucated Texas red (B). Background fluorescence in
non-hair cells suggests hair cell specific uptake (C). Cultures were also stained with Hoechst (blue). All
pictures were taken with the same microscope settings (same laser entensity, same exposure time, same
digital gain).

18
cells, as indicated by the overlapping of GTTR label (red) with Atoh1-GFP
(green), while fluorescence signals in other cells was undetectable. The
basal-to-apical hair cell specific accumulation of GTTR is consistent with
the Gentamicin-induced basal-to-apical gradient of hair cell damage
observed in vitro and in vivo (Wang and Steyger , 2009; Alharazneh et al.,
2011)
We tested several doses of Gentamicin to determine the dose suitable for
damage analysis. Gentamicin was added to the culture medium at different
concentrations and then replaced with fresh medium after a 3-hour
incubation. The live organs were imaged using a fluorescence dissection
microscope to assess the loss of Atoh1-GFP positive hair cells. When organs
were treated with 0.02 mM Gentamicin, the number of surviving outer hair
cells throughout the cochlea is comparable to untreated controls even at 72
hours, indicating that 0.02 mM is a sublethal dose. In the basal segment, 0.1
mM Gentamicin caused about 15% +/- 8% outer hair cell loss at 24 hours,
and about 90% +/- 6% outer hair cell loss at 72 hours (n=3). At 0.5 mM
concentration, Gentamicin killed more than 91% +/- 7% of outer hair cells at
24 hours in the base (n=3)(Fig 2). Only outer hair cells were counted,

19
Gentamicin Dose Response
0
20
40
60
80
100
120
0h 24h 48h 72h
Percentage of GFP+ OHC
Ctrl
0.02mM
0.1mM
0.5mM

Figure 2. Gentamicin causes hair cell loss in a dose-dependent manner. Quantification of outer hair cell
numbers in the base. Numbers are normalized with starting time point as 100%. 0.02 mM Gentamicin has
little effect; 0.1 mM Gentamicin kills hair cells at a slower rate (85% +/- 8% GFP positive OHC at 24h,
38% +/- 10% at 48h and 10% +/- 6% at 72h), while 0.5 mM Gentamicin causes the damage within 24 hours
(9% +/- 7% GFP positive OHC at 24h). Error bars represent standard deviations, n = 3.

because of the difficulty of distinguishing inner hair cells and inner
phallangeal cells after drug treatment due to misexpression of GFP in inner
phallangeal cells in the Atoh1-GFP animals. We chose 0.5 mM as the
concentration for subsequent experiments, because Gentamicin killed outer
hair cells in the basal segment effectively at this concentration.

20
We verify the hair cell loss indicated by the disappearance of GFP
fluorescence by performing fluorescence immunostaining using antibodies
against hair cell specific markers. In organs stained with antibody against
Parvalbumin, a calcium binding protein important for hair cell calcium
homeostasis (Yamagishi et al., 1993), Parvalbumin positive hair cells
overlapped with GFP positive cells (Fig 3A), suggesting that Parvalbumin
was gone when hair cells lost GFP. Treated organs were also stained with
antibody against MyoVI, an unconventional myosin important for hair cell
function, expressed specifically in hair cells (Friedman et al., 1999). When
hair cells started to lose GFP fluorescence at 9 hours, we observed
intensified staining of MyoVI in GFP negative hair cells. However, the
organs eventually lost both GFP and MyoVI staining at 24 hours (Fig 3B).
Redistribution of MyoVI in hair cells with disrupted cytoskeleton (Yang et
al., 2005) might account for the increased staining signal of MyoVI in hair
cells at 9 hours, making MyoVI a poor choice for damage assessment.
Phalloidin, which binds actin and marks hair cell bundles, is a good indicator
for healthy hair cells (Willott, 2001). Phalloidin staining verifed the
coincidence of GFP positive cells with hair bundles (Fig 3C). Taken together,

21
our evidence suggests that the disappearance of GFP fluorescence in hair
cells is a good indicator of hair cell damage and could be used to quantify
hair cell loss. This observation is pursued in the next section.

Figure 3. Disappearance of Atoh1-GFP reflects hair cell damage. A, Parvalbumin staining in Atoh1-
GFP cochlea 9 hours after Gentamicin treatment showing consistency between GFP and Parvalbumin. B,
MyoVI staining in Atoh1-GFP cochleae. At 9 hours, GFP negative hair cells have stronger staining for
MyoVI due to redistribution of MyoVI; 24 hours after Gentamicin treatment, organs lose both GFP and
MyoVI. C, Phalloidin staining 9 hours after Gentamicin treatment shows the absence of hair bundles when
hair cells lose GFP. Disappearance of GFP fluorescence indicates Gentamicin-induced hair cell damage.

22
Aminoglycoside antibiotic-induced hair cell death has long been believed to
involve the activation of the intrinsic mitochondrial apoptosis pathway
(Forge and Li, 2000; Cunningham et al., 2002). To determine whether hair
cells undergo the apoptosis pathway after Gentamicin treatment in our
organotypic cultures, we used immunohistochemistry to stain for activated
Caspase 3, a marker of apoptotic cells (Cohen, 1997), as well as Terminal
deoxynucleotidyl transferase dUTP nick end labeling (TUNEL staining)
(Gavrieli et al., 1992). After Gentamicin treatment, we observed positive
staining for active Caspase 3 in Atoh1-GFP positive hair cells, which
suggests that hair cells died through apoptosis (Fig 4A). In Gentamicin
treated organs, TUNEL staining signals overlapped with hair cell marker
MyoVI staining, indicating Gentamicin induced apoptosis mainly in hair
cells (Fig 4B). Taken together, results show that Gentamicin induces hair
cell death through the apoptosis pathway, consistent with what is observed
as previously described in literature (Forge and Li, 2000).

23

Figure 4. Gentamicin induces hair cell death through apoptosis. Active Caspase 3 staining (A) and
TUNEL staining (B) in cochlea 18 hours after Gentamicin treatment. Active Caspase 3 and TUNEL signal
are detected in hair cells (GFP or MyoVI positive) after Gentamicin treatment, confirming that Gentamicin
induces hair cell death through the apoptosis pathway.

Figure 5. Gentamicin damages hair cells but leaves supporting cells intact. Immunostaining for hair
cell marker MyoVI and supporting cell marker p27 in cochlea cross sections. Only one MyoVI positive hair
cell is observed in the Gentamicin-treated organ section, while p27 positive supporting cells still exist after
Gentamicin treatment. Missing hair cells and intact supporting cells in Gentamicin-treated organs indicate
the specific toxic effect of Gentamicin on hair cells.

24
We showed specific Gentamicin absorption by hair cells (Fig 1), indicating
that hair cells are the primary target of Gentamicin in the cochlea. We also
asked whether supporting cells were affected during Gentamicin treatment
by staining for supporting cell markers. MyoVI staining shows intact three
outer hair cells and one inner hair cells in sections from untreated organs and
missing hair cells in sections from Gentamicin treated organs (Fig 5, green
staining for hair cells). In contrast, staining of p27, a marker of supporting
cells, indicates that supporting cells are intact in Gentamicin treated organs,
although some disorganization is seen among supporting cells (Fig 5, red
staining for supporting cells), possibly as a consequence of hair cell loss.
The results suggest that Gentamicin primarily damages hair cells without
affecting supporting cells.
2.3 Discussion

Our results show that organotypic culture of neonatal mouse cochleae is a
reasonble model for ototoxicity research. In our floating membrane
organotypic culture system, ototoxic drug treatment induced hair cell
damage effectively. The hair cell damage and hair cell loss caused by
ototoxicants could be easily examined by comparing GFP-positive hair cell

25
numbers before and after treatment, and no fixation or immunostaing were
required, which made live organ observation possible. Neonatal cochlea
could be dissociated into single cell suspension by simple enzyme digestion
and trituration, and then hair cells could be purified by Fluorescence
Activated Cell Sorting (FACS) using Atoh1-GFP as a marker. Those merits
of the organotypic culture system make it a suitable model to investigate
ototoxicity through biochemistry and molecular biology approaches.

26
Chapter 3: Involvement of CDK2 in aminoglycoside ototoxicity
Previous work from our laboratory showed that mutations of certain CKIs
allowed hair cells to re-enter the cell cycle and undergo apoptosis (Chen et
al., 2003), indicating the importance of CDK activity regulation in avoiding
hair cell death. Involvement of the cell cycle machinery in the apoptosis of
postmitotic neurons (Folch et al., 2012) suggests that CDK activity could
also be a contributing factor in aminoglycoside-induced hair cell death. The
work presented here is focused on investigating whether CDK activity is
involved in aminoglycoside-induced hair cell death, as well as which CDK
is involved. We also investigated the molecular mechanisms of how CDK
activity regulated Gentamicin-induced hair cell apoptosis.
3.1 Effect of pharmaceutical CDK inhibitors
A number of pharmaceutical inhibitors are commercailly available to
specifically inhibit CDK activity, and one such inhibitor, Olomoucine, has
been identified as a purine analogue inhibitor for CDK1, CDK2 and CDK5.
(The CI
50
of Olomoucine is 7 μM to CDK1, 7 μM to CDK2, 3 μM to CDK5,
>1mM to CDK4 and >150 μM to CDK6) (Veselý et al., 2005). We tested
whether inhibiting CDK activity modulated aminoglycoside-induced hair

27
cell death by incubating cultures with Olomoucine at several concentrations
together with 0.5 mM Gentamicin. After 3 hours incubation, Gentamicin
was washed out and Olomoucine was kept in the medium for another 21
hours. In organs treated with Gentamicin in DMSO (vehicle control), more
than 80% +/- 1% (n=4) outer hair cells in the basal segment were damaged
24 hours after treatment; however, we observed more surviving outer hair
cells at the 24 hour time point in organs treated with Olomoucine and
Gentamicin (52% +/- 5% outer hair cell surviving when treated with 200 μM
Olomoucine, 71% +/- 12% with 400 μM and 92% +/- 6% with 600 μM, n=4)
(Fig 6A, B). We observed the dose-dependent protective effect of
Olomoucine against Gentamicin ototoxicity, suggesting that Gentamicin-
induced hair cell apoptosis requires CDK activity, specifically CDK1, CDK2
or CDK5 activity.
Since pan-CDK inhibitor Olomoucine is a general inhibitor of CDK1, CDK2
and CDK5 at a relatively low concentration, we were unable to differentiate
which CDK was responsible for Gentamicin-induced hair cell death.

28

Figure 6. CDK inhibitors protect hair cells against Gentamicin. Atoh1-GFP hair cells in cochlea after
Gentamicin and inhibitor treatment (A and C). Quantification of outer hair cells in the base (B and D).
There are more outer hair cells surviving the Gentamicin treatment when pan-CDK inhibitor Olomoucine is
present and this protective effect is dose-dependent (100 % of outer hair cell surviving, 20% +/- 1%, 52%
+/- 5%, 71% +/- 12% and 92% +/- 6% from top to bottom (A) or from left to right (B)). CDK2 inhibitor
CVT-313 has a similar protective effect (100%, 5% =/- 3%, 16% +/- 9%, 35% +/- 7% and 40% +/- 10%
from top to bottom (C) or from left to right (D)). Error bars represent standard deviation, n = 4.

29
We therefore tested a more specific CDK inhibitor CVT-313. CVT-313 is
more specific to CDK2, with a CI
50
of 0.5 μM for CDK2, 4.2 μM for CDK1,
215 μM for CDK4, and >1.25 mM for MAPK, PKA and PKC (Brooks et al.,
1997). We treated organ cultures with CVT-313 at several concentrations
together with 0.5 mM Gentamicin in culture medium for 3 hours after which
Gentamicin was washed out and CVT-313 was maintained. We observed
more than 95% +/- 3% (n= 4) of missing outer hair cells in the basal
segment of Gentamicin and DMSO treated organs. In organs treated with
Gentamicin and CVT-313, there were 16% +/- 9%, outer hair cells surviving
in the presence of 80 μM CVT-313, 35% +/- 7% outer hair cells in 160 μM
CVT-313 and 40% +/- 10%, outer hair cells in 320 μM CVT-313 (n=4) (Fig
6C, D). Data reveal that CDK2 inhibitor CVT-313 confers protection to hair
cells from Gentamicin-induced hair cell damage in a dose-dependent manner,
suggesting that CDK2 activity modulates Gentamicin-induced hair cell death.
3.2 Protective effect of CDK2 mutation on hair cell survival following
Gentamicin exposure
Pharmaceutical inhibitors can be toxic and they might modulate Gentamicin-
induced hair cell death through off-target pathways. We therefore tested the

30
involvement of CDK2 activity on Gentamicin-induced hair cell death using
Cdk2 knockout mice. Cdk2 knockout animals are viable with no reported
inner ear deficiency (Fig 7A Control and ABR data not shown).
We treated cochlea from Cdk2 knockout animals and wildtype littermates
with 0.5mM Gentamicin. We found comparable hair cell density in organs
from Cdk2 knockout animals and wildtype littermates before Gentamicin
treatment. Twenty four hours after Gentamicin treatment, we observed
greater number of surviving GFP positive outer hair cells in the basal
segment of Cdk2 knockout organ of Cotri (58% +/- 7% outer hair cells
remaining in organ of Corti from Cdk2 knockout mice versus 18% +/- 2% in
that from wildtype littermates, n=4) (Fig 7), suggesting hair cells are more
resistant to Gentamicin ototoxicity when CDK2 is knocked out. Our results
from experiments using CDK inhibitors and Cdk2 knockout animals indicate
that CDK2 activity is involved in Gentamicin-induced hair cell apoptosis.

31

Figure 7. Cdk2 KO organs of Corti are less sensitive to Gentamicin. Atoh1-GFP hair cells in Cdk2 WT
and Cdk2 KO organs 24 hours after Gentamicin treatment (A) and quantification of outer hair cells in the
base (B). The organ from the Cdk2 KO mouse retains more GFP positive outer hair cells (18% +/- 2%
outer hair cell surviving in WT versus 58% +/- 7% in KO), and the difference between Cdk2 WT and KO
organs is significant. Error bar represents standard deviation, and n = 4.

32
3.3 Delay of Gentamicin-induced apoptosis
It has been established that aminoglycoside antibiotics induce hair cell death
through the intrinsic mitochondrial apoptosis pathway (Cuningham et al.,
2004). To investigate whether attenuation of CDK2 activity by
pharmaceutical inhibitors, or by genetic mutations, affects the intrinsic
apoptosis pathway, we examined the loss of mitochondrial potential and
activation of Caspase in Gentamicin-induced hair cell apoptosis.
The most important step in the intrinsic apoptosis pathway is the collapse of
the mitochondrial membrane potential. The integrity of the mitochondrial
membrane is important to maintain the mitochondrial potential and to
sequester Cytochrome C from the cytoplasm, so the collapse of the
mitochondrial membrane leads to a loss of negative charged mitochondrial
potential and the release of Cytochrome C (Jiang and Wang, 2004). The
integrity of the mitochondrial membrane can be monitored by a specific dye,
Rhodamine 123, which only interacts with intact negatively charged
mitochondrial membrane (Johnson et al., 1980). Rhodamine 123 was put
into culture medium at 5.5 hours after Gentamicin treatment and then
cultures were returned to incubator for 30 minutes. The organs then were

33
washed with warmed PBS three times before observation under the
fluorescent microscope where the basal segment of the cochlea was
examined. In control organs, outer hair cells were brightly stained by
Rhodamine 123 due to the tremendous amount of mitochondria in hair cells
(Duvall et al., 1966), while Gentamicin-treated organs presented diminished
Rhodamine 123 fluorescence in outer hair cells, indicating the loss of
mitochondrial potential and membrane integrity in those cells. When CDK2
inhibitor CVT-313 was included in Gentamicin treatment, there were more
outer hair cells stained positively by Rhodamine 123 at the same time point,
suggesting fewer outer hair cells lost mitochondrial potential (Fig 8A). The
data indicates that inhibiting CDK2 activity delays Gentamicin-induced
collapse of the mitochondrial membrane in hair cells.
After the initiation of apoptosis by Cytochrome C release from mitochondria,
the apoptotic Caspase cascade is activated to execute the programmed cell
death process (Cuningham et al., 2002). We examined the onset of Caspase
cascade in hair cells to investigate the effect of CDK activity inhibition on
Caspase activation. FITC-VAD-FMK is an inhibitor to active Caspases, and
can be used as an indicator of Caspase cascade activation. When Caspase

34
proteins are activated by enzymatic cleavage, FITC-VAD-FMK can interact
with cleaved Caspase proteins and fluoresce green when excited with blue
light, making it a good indicator of Caspase cascade activation (Jayaraman,
2003). Organs from wildtype animals (not Atoh1-GFP transgenic mice)
were treated with Gentamicin for 3 hours and then incubated in Gentamicin
free medium in the presence of 5 μM FITC-VAD-FMK. The green
fluorescence in live cultures was examined at different time points. In
control organs, no green fluorescence was detected in hair cells, suggesting
no apoptosis in untreated organs. In organs treated with Gentamicin and
DMSO vehicle control, we first observed green fluorescent hair cells 8 hours
after Gentamicin treatment, which reached a plateau at 12 hours. In organs
treated with Gentamicin and CDK2 inhibitor CVT-313, no hair cells were
labeled by green fluorescence at 8 hour time point; 12 hours after treatment,
a few hair cells became green in Gentamicin and CVT-313 treated organs
compared to Gentamicin only treated organs; at the 24 hour time point, we
found comparable numbers of green fluorescent hair cells in Gentamicin
alone and Gentamicin plus CVT-313 groups (Fig 8B, C).

35

Figure 8. CDK2 inhibitor delays Gentamicin-induced hair cell apoptosis. (A) Mitochondrial potential
shown by Rhodamin 123. Some hair cells lose mitochondrial potential 6 hours after adding Gentamicin
(Gent 6h), but CVT-313 preserves the integrity of mitochondria (Gent/CVT 6h). Active Caspase shown by
FITC-VAD-FMK (B) and quantification of positively stained hair cells in the base (C). At 12 hours, many
hair cells have positive staining in Gentamicin treated organs (about 150 FITC positive hair cells per organ),
but many fewer hair cells are stained when CVT-313 is present (about 20 FITC positive hair cells per
organ) . Error bar shows data range, n = 3.

36
Collapse of the mitochondrial membrane potential and the activation of
Caspase cascade are considered “executive” steps for cells to die through
apoptosis (Cohen, 1997). Delays in those steps when CDK2 activity is
inhibited suggest that CDK2 activity is involved in upstream signaling
pathways that initiate these events.
3.4 Failure to detect cell cycle re-entry in Gentamicin-treated hair cells
In the central nervous system, mature neurons with targeted deletion of CDK
inhibitor genes involved in maintaining the postmitotic state lose control of
the postmitotic state, re-enter the cell cycle, replicate DNA and eventually
undergo apoptosis (Folch et al., 2012). Likewise, in the organ of Corti,
inappropriate activation of the cell cycle machinery by co-deletion of
p19
ink4d
and p21
kip1
leads to DNA damage and p53-mediated apoptosis in
hair cells (Laine et al., 2007). Based on these observations, we speculated
that CDK activity forces hair cells to re-enter the cell cycle and triggers
apoptosis.
To test this hypothesis, several methods were used to detect signs of cell
cycle re-entry in hair cells after Gentamicin treatment. First, we conducted

37
an EdU incorporation assay to detect DNA synthesis. No EdU incorporation
was detected in hair cells in control organs, which was expected given the
postmitotic state of hair cells. However, in organs treated with Gentamicin,
we observed no EdU incorporation in hair cells either, suggesting there was
no DNA synthesis in hair cells after Gentamicin treatment (Fig 9B). One
possible explanation for the absence of EdU incorporation could be that the
assay was not sensitive enough, or hair cells died before they entered S
phase. Second, we performed staining for Ki67, a marker for actively
dividing cells in all cell cycle phases. However, we also failed to detect Ki67
staining in hair cells after Gentamicin treatment (Fig 9A). By those
biochemical methods, we failed to detect the cell cycle re-entry in hair cells
after Gentamicin treatment. We were confident with our reagents and
experimental procedures since we observed positive staining for Ki67 and
EdU in other mitotic cells underneath the sensory epithelium.
However, our RNA sequencing data show that expression of genes involved
in cell cycle regulation, general transcription apparatus, and DNA damage
repair are induced in hair cells by Gentamicin treatment (discussed in
Chapter 6), strongly suggesting Gentamicin-induced cell cycle re-entry

38
attempt of hair cells. We think that Ki67 and EdU staining are hallmarks for
cells in late G1 phase or in S phase and that the hair cells die before starting
these events (more discussion in later sections of this chapter).

Figure 9. Failure to detect cell cycle re-entry by biochemistry methods. Immunostaining for Ki67 (A)
and EdU incorporation in hair cells after Gentamicin treatment (B). No positive staining for Ki67 or EdU is
found in hair cells after Gentamicin treatment at all stages. As a positive control for our reagents and
experimental procedures, positive staining for Ki67 and EdU is observed in cells underneath sensory
epithelium (blue for Hoechst and red for Ki67 or EdU).

39
3.5 Regulation of c-Jun/AP1 transcription activity
Our previous experiments have shown that CDK activity is involved in
upstream signaling pathways that regulate aminoglycoside-induced hair cell
apoptosis. Others have shown that the JNK pathway has been identified as
an important pathway to regulate the response of hair cells to
aminoglycosides (Pirvola et al., 2000). Based on this, we tested whether the
JNK pathway is a potential target that is regulated by CDK activity in
aminoglycoside-induced hair cell death.
We performed immunostaining to confirm that the JNK pathway is activated
in hair cells after Gentamicin treatment as previously reported. An antibody
against phosphorylated c-Jun (p-c-Jun) was used to probe the
phosphorylation of c-Jun on S73 by JNK. This post-translational
modification is indicative of the JNK pathway activation (Johnson and
Lapadat, 2002). We detected only background staining in hair cells of
control organs; and positive staining for p-c-Jun was detected in hair cell
nuclei in organs treated with Gentamicin for 6 hours (Fig 10A), suggesting
the activation of the JNK pathway in hair cells after Gentamicin treatment.
We then performed staining on Gentamicin-treated organs of Corti in the

40
presence of CDK inhibitor CVT-313 in order to test the effect of CDK
activity inhibition on c-Jun phosphorylation. At 6 hours post-treatment,
positive staining for p-c-Jun was still detected in hair cells when CDK
activity was inhibited, suggesting that CDK activity has no effect on c-Jun
phosphorylation at S73 by JNK (Fig 10A).
Given that c-Jun is a transcription factor and that c-Jun regulates the
transcription of itself, we asked whether inhibiting of CDK activity
modulates the transcription activity of c-Jun by measuring the messenger
level of Jun gene. Atoh1-GFP positive organs were dissociated 6 hours after
Gentamicin treatment, when Atoh1-GFP fluorescence was still detectable in
hair cells, GFP positive hair cells were purified by FACS sorting, and RNA
was extracted from the purified hair cell population, followed by cDNA
synthesis and quantitative PCR (Q-PCR). Q-PCR results showed that
expression of c-Jun was induced in hair cells by Gentamicin treatment,
indicating the up-regulated transcription activity of c-Jun, given the fact that
c-Jun is auto-regulated. When CDK activity was inhibited by CVT-313, c-
Jun expression could no longer be induced in hair cells by Gentamicin,
suggesting Gentamicin-induced expression of c-Jun is dependent on CDK

41
activity (Fig 10C). We also examined the expression of another c-Jun target
gene, a pro-apoptotic Bcl2 member Bim. Gentamicin treatment induced high
expression of Bim, and this induction was repressed when CDK activity was
inhibited by CVT-313 (Fig 10D). To exclude the possibility that CDK2
inhibitor CVT-313 inhibits transcription non-specifically, we examined the
expression of housekeeping gene Gapdh and no reduction of Gapdh
expression under the presence of CVT-313 was found (Fig 10E). In addition,
expression of another pro-apoptosis factor Bax was measured and we did not
find either significant induction of Bax expression by Gentamicin or
repression of Bax expression by CVT-313 (Fig 10F). The Q-PCR results
demonstrate that expression of c-Jun target gene c-Jun itself and Bim is
induced in hair cells by Gentamicin treatment and that these expression
increases are suppressed specifically by CDK2 inhibitor CVT-313,
suggesting the regulatory role of CDK2 activity in mediating c-Jun
transcription activity.
To confirm that expression of c-Jun but not phosphorylation of c-Jun is
modulated by CDK activity inhibition, whole organ lysates after Gentamicin
and inhibitor treatment were used for Western Blotting. The JNK pathway

42

Figure 10. Transcription activity of c-Jun is CDK dependent. A, Immunostaining for p-c-Jun in hair
cells 6h after Gentamicin and CVT-313 treattment. Positive staining of p-c-Jun is found in hair cells after
Gentamicin treatment no matter whether CVT-313 is present or not. B, Western blotting showing protein
level of c-Jun and p-c-Jun. Lysate from one organ (~10,000 cells) was loaded in each lane. Phosphorylation
of c-Jun is not affected by CVT-313 but is inhibited by JNK pathway inhibitor SP600125. However, CVT-
313 does reduce the protein level of c-Jun. C and D, Expression of c-Jun target genes, c-Jun and Bim. The
expression of c-Jun and Bim is induced by Gentamicin, but this induction is suppressed by CVT-313. E,
expression of housekeeping gene Gapdh; F, expression of pro-apoptosis factor Bax. Expression of these
genes is not induced by Gentamicin or suppressed by CVT-313, suggesting the specificity of CDK2’s
regulatory role for c-Jun transcription activity. FACS purified hair cells were used for cDNA library.

43
inhibitor SP600125 was used as a control for phosphorylation inhibition. We
saw no increase in c-Jun protein levels or phosphorylation levels after
Gentamicin treatment, which is expected since Gentamicin only activates the
JNK pathway in hair cells, and hair cells only account for 2% of the
population in the cochlea. As a control, JNK pathway inhibitor SP600125
inhibited both c-Jun expression and c-Jun phosphorylation, as expected.
However, CDK2 inhibitor CVT-313 reduced c-Jun expression but did not
change c-Jun phosphorylation levels (Fig 10D), which is consistent with our
observations by immunostaining and Q-PCR.
Our results implicate the involvement of CDK activity in regulating c-Jun
transcriptional activity, but how CDK activity regulates c-Jun transcriptional
activity without any disruption of c-Jun phosphorylation remains unknown
(more discussion in later section of this chapter).
3.6 Roles of CDK2 in mature utricular hair cells
Organotypic culture of cochlea from neonatal mice is a useful model to test
the response of hair cells to ototoxins. However, it is estimated that the
mouse starts to hear after P14 (Ehret, 1976). Neonatal cochlea is not mature

44
until this stage and hair cells are not considered fully functional. Given the
unfeasibility of culturing cochlea from mice older than P12 (Sobkowicz et
al., 1993), we tested utricles from older mice to investigate the response of
mature hair cells in aminoglycsoide-induced hair cell death. Aminoglycoside
antibiotics can also induce sensory hair cell death in utricles (Cunningham,
2006). Postnatal day 21 mice can swim straight and walk with regular gaits,
suggesting that utricles, which are responsible for head movement detection,
have matured and possess functional hair cells at this stage (Sawada et al.,
1994; Hunt et al., 1987).
We designed experiments to test whether inhibition of CDK activity confers
similar protective effects on utricular hair cells against Gentamicin
ototoxicity as observed in cochlea hair cells. It has been observed that a
higher Gentamicin dose is required to achieve significant hair cell damage in
utricles (Taleb et al., 2008), necessitating a different strategy in our
approach to treating utricle cultures. Utricles from three weeks old animals
were treated with 2.0 mM Gentamicin for 24 hours, and stained with an
antibody against MyoVIIa, a marker for sensory hair cells (Friedman et al.,
1999). Gentamicin-treated utricles lost more than half of the MyoVIIa hair

45
cells both in striola and extra-striola regions 24 hours after Gentamicin
treatment, compared to untreated controls (Fig 11 and Fig 12). Hair cell
density was decreased to 6.0 +/- 1.8 cells/1000 μm
2
in Gentamicin treated
utricles from 10.9 +/- 1.5 cells/1000 μm
2
in control organs in the striola
region; to 8.3 +/-2.2 cells/1000 μm
2
from 16.9 +/- 1.0 cells/1000 μm
2
in the
extra-striola region (from 3 organs). When CDK inhibitor CVT-313 was
included in culture medium, we observed more hair cells both in striola and
extra-striola regions, with a density of 8.8 +/- 1.6 cells/1000 μm
2
in the
striola region and a density of 10.8 +/- 1.9 cells/1000 μm
2
in the extra-striola
region (from 5 organs), suggesting that hair cells are more resistant to
Gentamicin ototoxicity when CDK activity is inhibited (Fig 12A, B and C).

Figure 11. Gentamicin induces hair cell loss in utricles. Low magnification pictures of MyoVIIa staining
of untreated and Gentamicin-treated utricles showing hair cell loss in striola (dash line circle) and extra-
striola regions. Compared to untreated organs, treated utricles lose hair cells in both striola and extra-striola
regions.

46

Figure 12. CDK2 inhibitor CVT-313 protects mature utricular hair cells against Gentamicin. A,
MyoVIIa staining. B and C, Hair cell quantification in striola and extra-striola regions. Gentamicin kills
utricular hair cells in both striola and extra-striola regions, but CVT-313 protects utricular hair cells against
Gentamicin in both regions. At least three organs were included in each treatment group. To count hair
cells, sampling squares larger than 3000 μm2 were drawn in these organs and more than 7 sampling squares
were counted for each region in each group. Error bars show standard deviation among sampling squares.

47
We sought to confirm our CDK inhibitor results by asking whether utricles
from Cdk2 knockout mice are less sensitive to Gentamicin treatment.
Untreated utricles from both Cdk2 wildtype littermates and Cdk2 knockout
animals had comparable hair cell densities, 9.3 +/- 1.8 cells/1000 μm
2
in
wildtype organs versus 9.9 +/- 1.4 cells/1000 μm
2
in Cdk2 knockout organs
in striola regions; 15.7 +/- 1.8 cells/1000 μm
2
in wildtype versus 16.1 +/- 0.8
cells/1000 μm
2
in Cdk2 knockout in extra-striola regions (from 2 organs)
(Fig 13). 24 hours after Gentamicin treatment, we observed more hair cells
in utricles from Cdk2 knockout mice than that from wildtype littermates (4.4
+/- 1.1 cells/1000 μm
2
in wildtype versus 9.0 +/- 1.2 cells/1000 μm
2
in
knockout in striola region; 5.1 +/- 0.8 cells/1000 μm
2
in wildtype versus 9.5
+/- 1.8 cells/1000 μm
2
in knockout in extra-striola region, from 3 organs)
(Fig 13A, B and C), suggesting that hair cells are less sensitive to
Gentamicin ototoxicity when Cdk2 is knocked out. Taken together, the
results show that, similar to neonatal cochlea hair cells, hair cells in mature
utricles are sensitive to Gentamicin treatment and CDK2 activity is involved
in Gentamicin-induced utricular hair cell death.

48

Figure 13. Cdk2 KO utricles are less sensitive to Gentamicin. A. MyoVIIa staining. B and C, hair cell
quantification in striola and extra-striola regions. There are more hair cells after Gentamicin treatment in
Cdk2 KO organs than that in WT, suggesting lower sensitivity of Cdk2 KO utricular hair cells. For each
genotype, 2 organs were untreated (control) and 3 organs were treated. For each region, at least 5 sampling
squares were counted. Error bars stand for standard deviation among sampling squares.

49
We next asked whether CDK activity regulates c-Jun transcriptional activity
in Gentamicin-induced hair cell death in mature utricles, as observed in
neonatal cochlea. Immunostaining with the antibody against phosphorylated
c-Jun showed positive staining in hair cells after Gentamicin treatment (Fig
14A, arrowheads), confirming JNK pathway activation in utricular hair cells
after Gentamicin treatment, consistent with our previous results from
neonatal cochlea, as well as other reports (Ylikoski et al., 2002). We
performed Q-PCR using cDNA libraries from FACS purified uticular hair
cells and found induced expression of c-Jun in utricular hair cells after
Gentamicin treatment, and repression of c-Jun expression when organs were
treated with CDK inhibitor CVT-313 (Fig 14B), implicating CDK activity in
c-Jun transcriptional activity regulation in hair cells in mature utricles,
consistent with our previous finding in the cochlea.

50

Figure 14. CDK2 has a similar regulatory role on c-Jun transcription activity. A, Immunostaining for
p-c-Jun. Positive stained utricular hair cells are only observed in Gentamicin-treated organs (arrowheads),
confirming JNK pathway activation by Gentamicin. B, Expression of c-Jun by qPCR. Gentamicin induced
c-Jun expression is suppressed by CVT-313, suggesting the regulatory role of CDK2 in c-Jun transcription
activity. FACS purified utricular hair cells were used for cDNA library synthesis.

51
3.7 Drug sensitivity in mature animals
We have shown that CDK activity is involved in aminoglycoside-induced
hair cell death in both neonatal cochlea and mature utricles. We now sought
to confirm our in vitro results by investigating how hair cells respond to
Gentamicin in vivo, using a trans-tympanic injection method. We tested
whether injecting Gentamicin directly into the middle ear of Cdk2 knockout
mice has a different effect in wildtype littermates. Systemic administration
of Gentamicin was not used because this method has previously been shown
to result in the death of mice before hearing loss could be analyzed (Wu et
al., 2001).
Since it is estimated that the mouse starts to hear sound around P14 (Ehret,
1976), we perform Auditory Brainstem Response (ABR) test to examine the
hearing ability of 3-week-old Cdk2 knockout mice and wildtype littermates
(CD1 background), before and after Gentamicin injection. Mice of each
genotype had similar baseline ABR thresholds prior to Gentamicin
administration. 25 μl (50mg/ml) Gentamicin solution was injected into the
right middle ear through the tympanic membrane using a 30G needle, and
the left ear served as a control. Mice were given 2 weeks to recover and the

52
healing of tympanic membrane was checked using a microscope. We
observed elevated ABR thresholds at all frequencies tested in the injected
ear of Cdk2 WT littermates and the shifts were more than 22 +/- 4 dB at 12
kHz and 26 +/- 4 dB at 16 kHz (n=7), suggesting severely compromised
hearing ablility (Fig 15A). The threshold shifts in the injected ears of Cdk2
KO mice were much smaller than that in wildtype littermates (3 +/- 2 dB
elevation at 12 kHz and 4 +/- 4 dB at 16 kHz, n =7) (Fig 15A), indicating
moderate hearing loss in Cdk2 KO mice. Our results indicate that Cdk2 KO
mice are less susceptible to Gentamicin-induced hearing loss in vivo.
ABR data indicate that Cdk2 wildtype ears challenged with Gentamicin
manifest a significant hearing threshold shift, while knockout ears are less
sensitive. We therefore assessed hair cell integrity by performing hair cell
counts in MyoVIIa –stained cochlea surface preparations after the final ABR
testing. The immunostaining results from control ears in both Cdk2 wildtype
and knockout mice showed no hair cell loss in apical, middle and basal
segments, and that hair cell numbers in both genotypes were comparable
(Fig 15B WT Control and KO Control). We observed a complete loss of
outer hair cells in the basal turn of injected WT ears, and a pattern of

53
damage that followed a basal-to-apical gradient (0% outer hair cells
surviving in the base, 68% +/- 18% in the middle, and 87% +/- 15% in the
apex, n = 4) (Fig 15B). This result is consistent with the basal-to-apical
sensitivity gradient reported in literature (Forge and Schacht, 2000) and
consistent with the severe hearing loss suggested by ABR at high
frequencies. In contrast, the injected ears from Cdk2 KO animals showed
less severe hair cell damage in the basal segment, while the middle and
apical segments displayed normal numbers and arrangement of sensory hair
cells (94% +/- 1% outer hair cell remaining in the base, 92% +/- 4% in the
middle, and 93% +/- 5 % in the apex, n = 3) (Fig 14B), consistent with the
moderate hearing loss suggested by ABR. These results suggest that, as is
the case in perinatal in vitro assary, mature cochlea hair cells are also less
vulnerable to Gentamicin ototoxicity when Cdk2 is knocked out.
Both ABR and immunostaining results show the reduced sensitivity of Cdk2
knockout mice to Gentamicin ototoxicity, which is consistent with the in
vitro organotypic culture results. Our in vivo Gentamicin injection
experiment provides evidence of the involvement of CDK2 activity in
aminoglycoside-induced hair cell damage in mature hair cells.

54

Figure 15. Cdk2 KO animals are less sensitive to Gentamicin. A, Threshold shift by ABR test. Cdk2 KO
animals have smaller shifts than WT animals. Error bars stand for standard deviation, n = 7 for each
genotype. B, MyoVIIa staining. Gentamicin induces severe damage in WT animals (OHC wiped out in the
base and missing in the middle), but Cdk2 KO animals only lose some OHC in the base. Both ABR and
immunostaining show lower sensitivity to Gentamicin in Cdk2 KO animals.

55
3.8 Discussion
Cell cycle machinery activation and CDK activity increase have been
observed in response to neuronal cell apoptotic signals, and linkage between
CDK activity and apoptosis regulation in neurons have been reported
(Konishi et al., 2002; Huang et al., 2006; Ghahremani et al., 2002). Here, we
report that CDK2 activity is also involved in aminoglycoside antibiotic-
induced hair cell death. Disruption of CDK2 activity, by either
pharmaceutical inhibitors (Fig 6) or genetic mutants (Fig 7), protects the hair
cell from Gentamicin ototoxicity in neonatal cochlea, and similar protective
effects exist in mature utricular hair cells (Fig 12 and Fig 13). Additionally,
in vivo testing of 3-week-old Cdk2 KO animals showed that they are less
sensitive to Gentamicin than their WT littermates (Fig 15).
Using immunostaining and Q-PCR, we demonstrate that c-Jun transcription
activity is under the control of CDK2 activity. However, how CDK2
regulates c-Jun transcriptional activity remains poorly defined. Direct
phosphorylation of c-Jun on S73 (the same site recognized by the antibody
we used) by CDK4 has been demonstrated in T lymphocytes and dendritic
cells (Vanden Bush and Bishop, 2011). In other studies, it has been found

56
that CDK activity regulates both c-Jun phosphorylation and c-Jun
transcriptional activity (Ghahremani et al., 2002; Besirli and Johnson, 2003).
In this study, we found that CDK2 activity is required for transcription of the
c-Jun target gene, but c-Jun phosphorylation on S73 is not affected when
CDK2 activity is inhibited. It is possible that CDK2 regulates c-Jun
phosphorylation on other sites; or alternatively, CDK2 may modulate the
interaction between c-Jun and other transcription factors, such as c-Fos or
Atf.
Other pathways under the control of CDK activity in hair cells may exist.
FoxO transcriptional activity and Bad phosphorylation have been identified
as regulatory targets of CDK activity in neurons (Konishi et al., 2002;
Huang et al., 2006), and these molecules could also be regulated by CDK
activity in the process of aminoglycoside-induced hair cell death.
Although we failed to detect the cell cycle re-entry in hair cells by EdU
incorporation and Ki67 staining, these biochemical methods for cell cycle
hallmarks may not be suitable to detect the Gentamicin-induced cell cycle
re-entry attempt in hair cells. As you will see in Chapter 6, our data from

57
RNA sequencing strongly suggest that mRNA expression profiles in hair
cells change for cell cycle re-entry upon Gentamicin treatment. Expression
shifts at transcription level could be the early event as hair cells adapt to exit
their quiescent state. However, due to the simultaneous activation of other
pathways initiating apoptosis, such as the JNK pathway demonstrated here,
hair cells die before they reach the G1/S checkpoint or other cell cycle
hallmarks. To detect checkpoint activation, we stained Gentamicin-treated
organs of Corti with antibody against phosphorylated Chk2 (Checkpoint2) to
detect checkpoint activation in hair cells, however, we failed to detect any
positive staining in hair cells after Gentamicin treatment (data not shown),
suggesting that the checkpoint is not activated in hair cells by Gentamicin
treatment. Additionally, we did not observe any significant difference
between cochlear organs from wildtype or p53 knockout animals in response
to Gentamicin treatment (discussed in detail in Chapter 5), suggesting that
Gentamicin-induced hair cell death is p53 independent. Consistent with this
result, we did not find significant expression changes of p53 target genes by
RNA sequencing. The small effect in p53 knockout mice and the absence of
induced expression of p53 target genes both suggest the dispensable role of

58
p53 in Gentamicin-induced hair cell death, implicating that the G1/S
checkpoint is not involved in Gentamicin ototoxicity, given the important
role of the p53 pathway in G1/S checkpoint activation (Stewart and
Pietenpol, 2001). These data together further support the hypothesis that hair
cells die before G1/S checkpoint activation even though hair cells make the
attempt to re-enter the cell cycle after Gentamicin treatment.

59
Chapter 4: CDK activity and Cisplatin ototoxicity
Cisplatin (cis-diamminedichloroplatinum) is a therapeutic drug developed to
treat cancers. The anti-cancer effect of cisplatin stems from its ability to
cause massive DNA damage in the form of interstrand or intrastrand
crosslinks, which are lethal for fast dividing cells due to the stalling of DNA
replication by DNA damage (Wang and Lippard, 2005).
Sensory hair cells are terminally differentiated, quiescent cells and are
maintained in a postmitotic state by actively expressing CKIs (Chen et al,
2003; Laine et al., 2007). In addition to aminoglycoside antibiotics, hair
cells are sensitive to cisplatin. Since there is no DNA replication in
postmitotic hair cells, oxidative stress induced by cisplatin has been thought
to be a primary cause for its ototoxicity. Ototoxicity of aminoglycoside
antibiotics is also ascribed to oxidative stress, so drugs in those two classes
might share some similarities in signal transduction. In previous chapters,
we discussed the involvement of CDK2 activity in aminoglycoside-induced
hair cell death in vitro and in vivo. We therefore asked whether CDK2
activity plays a similar role in the process of cisplatin-induced hair cell death.

60
4.1 Cisplatin-DNA adduct forming in hair cells
To examine the uptake of cisplatin by cells in the cultured cochlea, we
treated organs with 0.1mM cisplatin for 3 hours and then fixed for
histological analysis. Sections of neonatal organ of Corti cultured for 24
hours with/without cisplatin, were stained with an antibody against cisplatin-
DNA adducts. We observed a stronger staining signal in hair cell nuclei and
a weaker signal in surrounding cells (Fig 16), suggesting strong absorption
of cisplatin by hair cells.

Figure 16. Strong absorption of cisplatin by hair cells. Immunostaining for cisplatin-DNA adducts in
organ sections after cisplatin treatment. Intense staining in hair cell nuclei (arrow heads) indicates strong
absorption of cisplatin by hair cells.

61
4.2 Effect of pharmaceutical CDK inhibitors on cisplatin-induced hair
cell death
Organotypic cultures of cochleae from neonatal mice were treated with
freshly made 0.1mM cisplatin for 24 hours, at which point Atoh1-GFP
positive hair cells were counted to analyze the hair cell damage caused by
cisplatin. We observed hair cell damage in the basal segment (70% +/- 2%
outer hair cells surviving, n=4), but not in middle and apical segments (Fig
17A). To test whether pan-CDK inhibitor Olomoucine confers protective
effect against cisplatin ototoxicity, Olomoucine was included in the culture
medium at a concentration of 600 μM, a dose that confers protection to hair
cells from Gentamicin ototoxicity (Fig 6). 24 hours after treatment, almost
all outer hair cells in the basal segment were lost (2 % +/- 1% outer hair cells
surviving, n=4) (Fig 17A), suggesting hair cells were more sensitive to
cisplatin in the presence of pan-CDK inhibitor Olomoucine.
We also tested CDK2 inhibitor CVT-313. We observed a larger hair cell loss
response to cisplatin in the presence of CVT-313 in the basal segment, and
the damage was extended to middle and apical segments in the presence of

62

Figure 17. Cisplatin-induced hair cell death is exacerbated by CDK inhibitors. Atoh1-GFP hair cells in
cochlea after cisplatin and CDK inhibitor treatment. Cisplatin alone kills some hair cells in the base (70%
+/- 2%, n=4), but the presence of CDK inhibitor Olomoucine (Olo, A) or CVT-313 (CVT, B) exacerbates
the cytotoxicity of cisplatin ( 2% +/- 1% in the abase for Gentamicin and Olomoucin, n=4; < 2% in all
segments for Gentamicin and CVT-313, n=3 ).

63
CVT-313 (< 2% outer hair cells in all segments, n= 3) (Fig 17B), suggesting
that hair cells are more vulnerable to cisplatin when CDK2 activity is
inhibited. Unlike in aminoglycoside-induced hair cell death, both pan-CDK
inhibitor Olomoucine and CDK2 specific inhibitor CVT-313 sensitize hair
cells to cisplatin, and exacerbate hair cell damage caused by cisplatin. Our
results suggest that CDK activity does not modulate cisplatin-induced hair
cell death in the same way as aminoglycoside-induced hair cell death, and
that different pathways are activated in hair cells by cisplatin and
aminoglycoside antibiotics, even though both of them induce oxidative
stress.
4.3 Discussion
Oxidative stress has long been thought to be a major factor in
aminoglycoside antibiotics and cisplatin ototoxicity; however, CDK
inhibitors have different effects. While CDK inhibition protects hair cells
from aminoglycoside antibiotics, CDK inhibitors sensitize hair cells to
cisplatin, suggesting a difference in the underlying cytotoxic mechanisms.

64
Given the link between DNA damage and cell cycle re-entry from neuronal
apoptosis studies, cell cycle re-entry is believed to be required for neurons to
restore their DNA repairing capacity (Kruman, 2004). In the cochlea, we
observed the presence of cisplatin-DNA adducts in hair cells, indicating the
existence of DNA damage. We speculate that in the presence of cisplatin,
hair cells reactivate the cell cycle machinery to repair DNA damage, and that
inhibition of CDK activity disrupts the repair attempt, leading to exacerbated
cisplatin-induced hair cell death.

65
Chapter 5: Involvement of other factors in aminoglycoside
ototoxicity
The temporal activation and deactivation of CDKs during the cell cycle is
the driving force for cells to go through phases of the cell cycle; however,
factors other than CDKs and their activators also contribute to cell cycle
progression and cell cycle arrest. Involvement of CDK2 activity in
aminoglycoside-induced hair cell death has been demonstrated; we therefore
asked whether other cell cycle regulating factors, CDK4, p19
ink4d
, p21
kip1
,
and checkpoint protein p53, affected aminoglycoside ototoxicity using
knockout animals.
5.1 No effect of knocking out Cdk4 on Gentamicin ototoxicity
CDK4 and CDK6 are activated by D-type cyclins in G1 phase. At a
relatively low concentration, pan-CDK inhibitor Olomoucine inhibits CDK1,
CDK2 and CDK5, but not CDK4 or CDK6 (Veselý et al., 2005). CDK2
inhibitor CVT-313 also has a much higher CI
50
for CDK4 and CDK6
(Brooks et al., 1997). Therefore, we could not reach a definitive conclusion

66
as to whether CDK4 or CDK6 also contributed to aminoglycoside-induced
hair cell death from previous CDK inhibitor experiments.
To do this, we utilized Cdk4 knockout mice to determine whether CDK4
activity modulates aminoglycoside-induced hair cell death. Cdk4 knockout
mice are viable and no development deficiency in the inner ear has been
reported in Cdk4 knockout mice (Tsutsui et al., 1999). We treated cochleae
from Cdk4 knockout mice and wildtype littermates in vitro with 0.5mM
Gentamicin for 3 hours followed by 21-hour incubation in fresh medium
using the same culture system previously described. We observed nicely
organized outer and inner hair cells with comparable densities in untreated
cochlea from both genotypes, suggesting there is no deficiency in the organ
of Corti caused by knocking out of Cdk4. However, more than 90 percent of
outer hair cells in the basal segment were lost in Gentamicin-treated cochlea
regardless of the genotype (10% +/- 3 % outer hair cell surviving in WT
versus 4% +/- 2 % in KO, n=4)(Fig 18A), suggesting knocking out of Cdk4
does not confer any protective effect to hair cells against 0.5 mM
Gentamicin. We further treated organs of Corti with 0.1mM Gentamicin for
3 hours and then incubated for another 69 hours in fresh medium. Cochlea

67
from both genotypes lost outer hair cells in the basal segment at a
comparable rate (87% +/- 6% outer hair cell surviving in WT versus 86% +/-
1% in KO at 24h, 31% +/- 2% in WT versus 37% +/- 2% in KO at 48h and
10% +/- 4% in WT versus 11% +/- 6% in KO at 72h, n=4)(Fig 18B),
suggesting no protective effect against lower dose of Gentamicin by
knocking out Cdk4. Our results suggest that, unlike Cdk2 knockout, Cdk4
knockout does not modulate hair cell sensitivity to Gentamicin ototoxicity.
Insignificant difference between Cdk4 knockout and wildtype littermates in
response to Gentamicin treatment may be caused by the redundancy between
CDK4 and CDK6. CDK4 and CDK6 are both activated by D cyclins in G1
phase and they share great similarity in sequence, structure and substrates
(Malumbres et al., 2009). Given the fact that mice with mutation in either
Cdk4 or Cdk6 genes are relatively normal in development (Barriere et al.,
2007; Malumbres et al., 2004), it is believed that CDK4 and CDK6 play
redundant roles in cells. We therefore need more experiments to show
whether CDK4 and/or CDK6 are involved in aminoglycoside-induced hair
cell death.

68

Figure 18. Cdk4 knockout does not affect hair cell sensitivity to Gentamicin. Quantification of outer
hair cells in the base in Cdk4 WT and KO organs after 0.5mM (A) or 0.1mM (B) i treatment. There is no
significant difference between Cdk4 WT and KO cochleae after Gentamicin treatment, suggesting that
Cdk4 knockout does not affect Gentamicin ototoxicity. Error bars stand for standard deviation, n = 4 for
each genotype in each experiment.

69
5.2 Effects of knocking out CKIs

on hair cell sensitivity to Gentamicin
CDKs are activated by activating cyclins and their activities are inhibited by
Cyclin-dependent Kinase Inhibitors (CKI). Previously in our lab, we found
p19
Ink4d
and p21
kip1
to be important for hair cells to maintain the postmitotic
state (Chen et al., 2003; Laine et al., 2007). When p19
Ink4d
and p21
kip1
are
disrupted, hair cells re-enter the cell cycle, evidenced by DNA synthesis;
however, hair cells die from apoptosis before they finish mitosis (Laine et al.,
2007). To investigate whether CKIs p19
Ink4d
and p21
kip1
were involved in
aminoglycoside-induced hair cell death, we examined the hair cell damage
after Gentamicin treatment in cochlea from p19
Ink4d
knockout and p21
kip1

knockout mice.
The protective effect of pharmaceutical CDK inhibitors has been
demonstrated previously, we therefore speculated that CKI deletion would
exacerbate aminoglycoside-induced hair cell death. We again treated the
organ of Corti with 0.5mM Gentamicin for 3 hours, using previously
described culture conditions, and found that almost 90% of outer hair cells in
the basal segment were damaged at 24 hours in all genotypes (9% +/- 5%
outer hair cells surviving in WT and 9% +/- 7% in KO, n=4) (Fig 19A). We

70
made additional hair cell counts at 9 hours to track possible sensitization by
p19
Ink4d
knockout or p21
kip1
knockout. At this time point, we observed
comparable outer hair cells remaining in the basal segments in organs from
both p21
kip1
knockout mice and wildtype littermates (78% +/- 1% in WT and
79% +/- 4% in KO, n = 4) (Fig 19A). The absence of a significant difference
between p21
kip1
knockout and wildtype organs indicates that loss of p21
kip1

did not modulate hair cell death induced by 0.5mM Gentamicin. Similarly,
we did not observe any significant difference between p19
Ink4d
knockout and
wildtype organs after 0.5mM Gentamicin treatment ( 41 % outer hair cell
surviving in WT and 39% in KO at 9h; 5% in WT and 1% in KO at 24h, n=2)
(Fig 19B), suggesting that p19
Ink4d
is dispensable under this condition.
We further tested a lower dose of Gentamicin. When cochleae were treated
with 0.1 mM Gentamicin for 3 hours, it took 72 hours instead of 24 hours to
observe hair cell damage as severe as in organs treated with a higher dose
(Fig 2). We did not test the protective effect of pharmaceutical CDK
inhibitors with 0.1 mM Gentamicin because the pharmaceutical CDK
inhibitors themselves caused hair cell damage at 72 hours. We treated

71

Figure 19. Effect of knocking out CKIs on Gentamicin sensitivity. Quantification of of outer hair cells in
the base. At 0.5mM, Gentamicin kills comparable amounts of outer hair cells in p21 WT and KO organs (n
=4) (A), in p19 WT and KO organs (n =2) (B). However, organs from p19 KO animals are more sensitive
to 0.1 mM Gentamicin (61% +/- 6% surviving OHC in WT versus 29% +/- 8% in KO at 48h, n=4) (C).
Error bars represent standard deviation.

72
p19
Ink4d
knockout cochlea with 0.1 mM Gentamicin and found that hair cells
were damaged faster in p19
Ink4d
knockout than p19
Ink4d
wildtype (87% +/- 5%
outer hair cells remaining in WT versus 79% +/- 8% in KO at 24h, 61% +/-
6% in WT versus 29% +/- 8% in KO at 48h, 17% +/- 7%in WT versus 5%
+/- 4% in KO at 72h, n =4) (Fig 19C), suggesting that the p19
Ink4d
knockout
sensitizes hair cells to a lower dose of aminoglycoside antibiotics.
5.3 No effect of p53 KO on Gentamicin-induced hair cell death
As a transcription factor, p53 plays an important role in cell cycle arrest and
apoptosis initiation after checkpoint activation (Stewart and Pietenpol, 2001).
In postmitotic neurons, the p53-dependent cell death pathway after cell cycle
re-entry has been identified (Folch et al., 2012), suggesting the importance
of p53 in neuronal cell death. In hair cells, the double knockout of p19
Ink4d

and p21
kip1
causes aberrant cell cycle re-entry, which leads to p53 activation
and apoptosis (Laine et al., 2007), implicating the involvement of p53
pathway in hair cell death after loss of the postmitotic state.
Given the regulatory role of p53 in cell cycle arrest and apoptosis, we
examined Gentamicin-induced hair cell death in p53 knockout organs to

73
determine whether aminoglycoside-induced hair cell death was p53
dependent. We treated the organ of Corti with 0.5mM Gentamicin for 3
hours and found no significant difference between p53 knockout and
wildtype organs at 24 hours. Only 9% GFP positive outer hair cells remained
in the basal segment of p53 wildtype organs, and 5% of outer hair cells in
p53 knockout organs (n =2)(Fig 20A). We also assessed hair cell damage in
p53 knockout and wildtype organs treated with 0.1 mM Gentamicin for 3
hours, and found that both p53 knockout and wildtype organs lost their hair
cells at a comparable rate (77% outer hair cell surviving in WT versus 80%
% in KO at 24h, 17% in WT versus 13% in KO at 48h, 7% in WT versus 4%
in KO at 72h, n =2) (Fig 20B), suggesting no difference between genotypes.
Absence of a significant difference between p53 knockout and wildtype
organs in response to Gentamicin treatment suggests aminoglycoside-
induced hair cell death is p53 independent. Consistent with this result,
expression of genes involved in the p53 pathway in hair cells is not
significantly affected by Gentamicin treatment in our RNA sequencing data
(discussed in Chapter 6). Particularly, expression of p53 target gene Bax in

74

Figure 20. No effect of knocking out p53 on hair cell response to Gentamicin. Quantification of outer
hair cells in the base. 0.5mM Gentamicin kills outer hair cells at a similar rate in p53 WT and KO organs
(A). There is no difference in responding to 0.1mM Gentamicin between p53 WT and KO organs (B).
These results suggest that Gentamicin-induced hair cell death is p53 independent. Two organs were
counted for each genotype in each experiment.

75
hair cells is not induced by Gentamicin treatment in both RNA sequencing
and Q-PCR analysis (Fig 10), supporting the hypothesis that p53 is
dispensable in Gentamicin-induced hair cell death.

76
Chapter 6: Gene expression profile in hair cells after
aminoglycoside treatment
To investigate how hair cells alter gene expression in response to
aminoglycoside antibiotics and to analyze which pathways are involved in
aminoglycoside ototoxicity, we utilized RNA sequencing to profile the gene
expression changes in FACS purified hair cells after Gentamicin treatment.
6.1 Experiment Design
Cochlea from neonatal transgenic Atoh1-GFP mice were dissected and
cultured as described in Materials and Methods. After overnight recovery,
organs were treated with Gentamicin at 0.5 mM for 3 hours and then
dissociated for hair cell purification by FACS. GFP disappearance and
Caspase activation were first observed at 8 hours after Gentamicin treatment
(Fig 8). Therefore, we purified GFP positive hair cells and GFP negative
non-hair cell populations 3 hours after treatment. At least 50,000 cells were
collected for each sample population, and three replicates were prepared.
RNA extraction, cDNA synthesis, RNA sequencing, read trimming, and
alignments were done as described in Chapter 8 Materials and Methods.

77
6.2 Q-PCR validation
We picked 30 genes with p value < 0.01 and fold change > 1.2 for Q-PCR
analysis to validate our RNA sequencing data. Primer pairs for selected
genes are listed in Table 1. Expression patterns of all the selected genes in
Q-PCR are consistent with those in RNAseq, verifying that RNAseq results
are reliable (Fig 21).

Figure 21. RNA sequencing data is validated by Q-PCR. Expression fold change of selected genes.
Genes with greater than 1 fold change in RNA sequencing (black bars) also have greater than 1 fold change
in Q-PCR (grey bars), indicating the consistency between RNA sequencing and Q-PCR.

78
Table 1. Primer pairs for Q-PCR validation. Primer sequences obtained from PrimerBank (Wang and
Seed, 2003) or designed by ourselves.
Gene
Symbol Forward Reverse
Alkbh3 GAGCCAGTCTGCTACTCAGC AACACAAATTGTCGGTCACATTG
Apex2 GGATGGATGGCTTGCTCAGTA ACTTCAGGGAGTAAGAAGGAGG
Atf2 CCGTTGCTATTCCTGCATCAA TTGCTTCTGACTGGACTGGTT
Bnip1 AGGCTATGCAGACTCTAGTCAG CAGTTCTCGGCGGTTGTACT
Cables1 CGCCTCAACTCGTTCACTCAG GGAGAGGGGCATTCTCTTCAA
Ccar1 AGATGAGTATGACCCAATGGAGG CCTTGCAGTACCGGCTGAC
Ccne2 ATGTCAAGACGCAGCCGTTTA GCTGATTCCTCCAGACAGTACA
Ccnt1 AACAAGCGGTGGTATTTTACTCG CCTGCTGGCGGTAAGAGAG
Chuk GTCAGGACCGTGTTCTCAAGG GCTTCTTTGATGTTACTGAGGGC
Clp1 ATGAGCGAGGAATCCAATGATG CTCCAACTGAACCGATTGAGAG
Ddit3 CTGGAAGCCTGGTATGAGGAT CAGGGTCAAGAGTAGTGAAGGT
Ercc8 GAGGAAGATGAAGCTATGGAA CTTCAGGGGTTTCTCTTTGTC
Fbxo6 TCCCTATGGAAGCGCAAGAGT CTCCGTTGGAGTCTATCCGC
Gadd45g GGGAAAGCACTGCACGAACT AGCACGCAAAAGGTCACATTG
Klf11 CATGGACATTTGTGAGTCGATCC CCTTTGGTAGATCAGGTGCAG
Mapk8 AGCAGAAGCAAACGTGACAAC GCTGCACACACTATTCCTTGAG
Mrpl1 GTAAGGTGCCTCCGTAGAGTC AACAAGGGTAAAGAGATGCCTG
Mrps22 ACGTTCAACAAGACACTTTGGT AGGCTGGTGGCATCATCAAC
Nfkb1 ATGGCAGACGATGATCCCTAC TGTTGACAGTGGTATTTCTGGTG
Nfkbib GCGGATGCCGATGAATGGT TGACGTAGCCAAAGACTAAGGG
Nop2 CGAAAGGCCCGAAAACAGAAG GGAGACTTATTTGGCTTAGGGAC
Nop58 TTTGAAACGTCCGTTGGCTAC GAATGCTGCTAATGCTTCTGC
Pde12 ATGTGCTCAATGTGGACGC GGGAAGCCAGCCATGATGTAG
Polr1b CAGGCCATACCTCCCTTTGAA TCCCTCGGTAGGTGCTCTTTC
Polr2h CTGCACTGTGAGAGTGAATCTT GCTATGACCAACCGGAACTTG
Polr3b CTGGTGAAACAGCACATAGACT CATGGGGTCAGCATCACTTGT
Rad52 CTTTGTTGGTGGGAAGTCTGT CGGCTGCTAATGTACTCTGGAC
Taf4b TTGCAGCTATTGGACCAAGGA GTGGCTGTTAGGCTGGAAGT
Tfb2m GGCCCATCTTGCATTCTAGGG CAGGCAACGGCTCTATATTGAAG
Tsr1 ATGGCTAAAGTCGCTGATACCA ACAGCTAGTGTATAGGTGGGAAG

79
6.3 Principle Component Analysis
We conducted Principle Component Analysis to capture the variability
among samples as much as possible and to identify the most significant
factors contributing to the variability. As shown in Fig 22, the most
significant contributing factor is the cell type. Hair cell samples and non-hair
cell samples are well separated along the X axis and 49.91% of the total
variability in our dataset is captured on the X axis (Fig 22, hair cells versus
non-hair cells), suggesting the huge difference between gene expression
profiles in hair cells and non-hair cell population. This result is expected
since it has been known that hair cells are highly differentiated cells and they
express a set of hair cell specific genes (unpublished microarray data from
our lab). The second significant contributing factor is the Gentamicin
treatment. Gentamicin-treated hair cell samples and untreated hair cells
samples are far from each other along the Y axis and 13.53% of the
variability is captured on the Y axis (Fig 22, hair cells red versus blue),
suggesting the significant gene expression changes induced by Gentamicin
in hair cells. In contrast, non-hair cell samples are not divided by
Gentamicin treatment. These results are consistant with our observation (Fig

80
1) and the description in the literature (Wang et al., 2009; Alharazneh et al.,
2011) that Gentamicin is absorbed specifically by hair cells in cochlea.

Figure 22. Principle Component Analysis shows the most significant fators affecting gene expression.
Each sample is represented by one dot. Samples from untreated organs are shown in red, and samples from
treated organs are shown in blue. 71.88% of the variability in gene expression among 12 samples is
captured in three dimensions (X axis for 49.91%, Y axis for 13.53% and Z axis for 8.44%). Hair cell
samples and non-hair cell samples are well separated along X axis, indicating huge gene expression
differences caused by cell type. Untreated and treated hair cell samples are separated along Y axis,
suggesting a significant gene expression shift in hair cells induced by Gentamicin treatment. The variability
captured on Z axis comes mainly from the heterogeneity within non-hair cell samples.

81
Another observation from Principle Component Analysis is low
heterogeneity in hair cell samples and high heterogeneity in non-hair cell
samples. Three untreated hair cell samples are close to each other (Fig 22,
hair cells red), indicating low heterogeneity among these samples. Similarly,
Gentamicin-treated hair cell samples cluster together tightly (Fig 22, hair
cells blue), suggesting high similarity among these replicates. The high
similarity among Gentamicin-treated samples also indicates the uniformed
gene expression response to Gentamicin treatment in hair cells. In contrast,
untreated and treated non-hair cell samples are spread along Y and Z axis
(Fig 22, non-hair cells), suggesting low similarity and high heterogeneity in
these samples. These results are expected since hair cells were purified by
GFP marker, which was more stringent than how the non-hair cell
population was collected.
6.4 Differential Gene Expression Analysis
The Gentamicin-induced differential gene expression was analyzed with the
embedded reads quantification software in Partek Flow. After filtering the
results with criteria of a p value less than 0.01, and a fold change greater
than 1.2, 1917 genes are up-regulated and 1792 genes are down-regulated in

82
hair cells after Gentamicin treatment. In the non-hair cell population, with
the same criteria, 436 genes are found up-regulated and 262 genes down-
regulated in Gentamicin treated samples. From those numbers, we can tell
that Gentamicin treatment affects gene expression in hair cells more
significantly than that in non-hair cell populations, consistent with our
previous Principle Component Analysis. It is also possible that changes in
gene expression in the non-hair cell population are secondary to changes in
the hair cells. With these lists of up- or down-regulated genes, we conducted
Functional Annotation Enrichment Analysis and pathway analysis to
investigate the gene expression responses to Gentamicin treatment in hair
cells and in non-hair cell population.
6.5 Gentamicin-induced gene expression responses in hair cells
Ingenuity IPA software, DAVID v6.7 database, BioBase ExPlain 3.1 and
GeneCard database v3.11 were used to analyze gene expression changes,
and we found the following biological processes and pathways were affected
in hair cells by Gentamicin treatment.

83
Cell cycle machinery and cell cycle re-entry
Cell cycle machinery is repressed in postmitotic cells, such as hair cells and
neurons. However, cell cycle machinery activation has been observed in
some postmitotic cells after exposure to stress (Folch et al., 2012). It has
been discovered that CDK activity regulates apoptosis pathways by
phosphorylating apoptosis regulators, such as c-Jun, FoxO and BAD
(Ghahremani et al., 2002; Huang et al., 2006; Konishi et al., 2002). In
previous chapters, we also discussed the involvement of CDK2 activity in
Gentamicin-induced hair cell death. Pharmaceutical CDK2 inhibitors protect
hair cells against Gentamicin treatment and CDK2 knockout confers
resistance to Gentamicin ototoxicity. Based on these observations, we
speculate that cell cycle machinery activation has a similar regulatory effect
in aminoglycoside-induced hair cell death. Therefore, we examined the
expression of genes involved in cell cycle machinery (Table 2) to investigate
whether Gentamicin treatment caused significant expression changes in
these genes.
Cell cycle progression is promoted by CDK activity, which is regulated by
the expression of appropriated CDK activator cyclins. In Gentamicin-treated

84
hair cells, expression of Cyclin B2 (Ccnb2, 2.2 fold) and Cyclin E2 (Ccne2,
1.5 fold) is increased. In dividing cells, Cyclin B2 is expressed in G2 phase
and activates CDK1, promoting G2 progression. Expression of Cyclin E2
starts in late G1 and lasts into S phase; CDK2/Cyclin E complex promotes
the G1/S transition. Regulation of these cyclins indicates that CDK activity
is up-regulated in hair cells after Gentamicin treatment. Additionally,
expression of some CDK Activating Kinases (CAK) is induced, such as
Cdk7, Cyclin H (CcnH), Cdk20, Cyclin K (CcnK) and Mnat. Even though
these kinases are also involved in transcription regulation, increased
expression of these genes facilitates the up-regulation of CDK activity.
Consistent with these results, expression of CKI genes, p21
kip1
(Cdkn1a, -1.5
fold) and p57
kip2
(Cdkn1c, -2.2 fold), is repressed in Gentamicin-treated hair
cells. Decreased expression of CKI is another indication of CDK activity up-
regulation. These results together suggest that CDK activity is up-regulated
in hair cells after Gentamicin treatment.
In addition to the expression changes of CDK related genes, expression of
genes involved in cell cycle progression regulation is also affected by
Gentamicin treatment. E2f6, Gadd45gip1, Thap1 and Suv39h2, these genes

85
are involved in E2F transcription regulation and G1/S transition; and their
expression is induced in hair cells by Gentamicin. Expression of genes
involved in S phase, G2 phase and M phase in hair cells is also affected by
Gentamicin treatment. It is not surprising to find the expression changes of
genes involved in phases other than G1. In cortical neurons, acute removal
of Rb results in cell cycle re-entry, and up-regulated genes are involved in
all cell cycle stages (Andrusiak et al., 2012).
The expression increases of cyclins and other cell cyle machinery factors in
hair cells after Gentamicin treatment, together with our CDK inhibitor and
Cdk2 knockout experiments (Chapter 3), strongly suggest that hair cells
attempt to re-enter the cell cycle by up-regulating CDK activity. However,
we failed to detect EdU incorporation in hair cells after Gentamicin
treatment (Fig 9), suggesting that there is no DNA synthesis in hair cells. In
addition, we could not detect checkpoint activation in Gentamicin-treated
hair cells (data not shown) and there is no difference between the
Gentamicin sensitivity in the p53 WT organ of Corti versus the p53 KO
organ, indicating that the G1/S checkpoint is not activated by Gentamicin.
Taken together, RNA sequencing data and the involvement of CDK2

86
activity in Gentamicin-induced hair cell death demonstrate that hair cells
attempt to re-enter the cell cycle when challenged by aminoglycoside
antibiotics, but the absence of checkpoint activation and the absence of DNA
replication indicate that hair cells die before they reach the G1/S checkpoint.
In IPA Up-stream analysis, we found that the state of Myc was predicted as
activated with an activation Z score of 4.18 and an overlapping p-value of
1.81E-06. Myc protein is a transcription factor and regulates expression of
about 15% of all the genes, including positively regulating the expression of
cell cycle progression genes and negatively regulating the expression of cell
cycle arrest genes (Dang et al., 2006). The predicted activation of Myc
indicates that the expression pattern of Myc target genes in hair cells is
similar to that in dividing cells in which Myc is active, supporting our
hypothesis that hair cells attempt to re-enter the cell cycle after Gentamicin
treatment.
Table 2. Expression changes in genes involved in the cell cycle.
Gene symbol Function p value Fold change
E2f5 Cell cycle entry 0.000867 -1.34193
Arl2 Cell cycle regulation 0.004731 -1.32077
Ccpg1 Cell cycle regulation 0.002023 -1.21159
Ccng2 Cell cycle regulation 0.007048 -1.81896
Gas1 Cell cycle regulation 0.002549 -1.52164

87
Table 2. Continued
Gene symbol Function p value Fold change
Blcap Cell cycle regulation 0.003651 -1.39359
Tfdp2 Cell cycle regulation 0.006879 -1.30984
Foxo4 Cell cycle regulation 0.003413 -1.47883
Cdkn1a CKI 0.007986 -1.50225
Cdkn1c CKI 0.000596 -2.27339
Ccnh CAK 0.004257 1.32942
Ccnk CAK 0.004885 1.21868
Cdk7 CAK 6.44E-05 1.39676
Mnat1 CAK 0.00464 1.32402
Cdk20 CAK 6.81E-05 1.50149
Ccnd2 G1 phase 0.009613 -2.17425
Suv39h2 G1/S phase 0.000618 1.82814
Thap1 G1/S phase 6.88E-07 2.94531
Gadd45gip1 G1/S phase 0.000395 1.35737
E2f6 G1/S phase 8.16E-06 1.30529
Cdc73 G1/S phase 0.000052 1.46645
Cables1 G1/S phase 0.000395 1.99902
Ccne2 G1/S phase 0.031776 1.51167
Lats2 G1/S phase 0.006026 -1.49745
Cables2 G1/S phase 0.001179 -1.28215
Gmnn S pahse 0.002183 -1.56245
Tlk2 S phase 2.47E-05 1.25584
Ccnf S phase 0.008913 -1.83873
Ccnb2 S/G2 phase 0.003239 2.16567
Mdc1 S, G2, M phase 0.000263 -1.47327
Rprm G2 pahse 0.001875 -1.2871
Zc3hc1 G2/M phase 0.000856 1.54521
Cdc5l G2/M phase 0.000806 1.37072
Hmg20b G2/M phase 0.000433 -1.23246
Lzts1 G2/M phase 0.008991 -1.2956
Clasp1 M phase 0.000015 -1.55826
Mad2l2 M phase 0.003026 -1.27901
Cep250 M phase 0.003587 -1.22762
Crocc M phase 0.000203 -2.22394
Cit M phase 0.004669 -2.19637
Ckap5 M phase 0.001982 -1.34103
Ncapd3 M phase 0.001489 -1.28807
Chtf8 M phase 0.00015 -1.62119
Cdc14a M phase 0.005453 1.37782
Fbxo5 M phase 0.004915 1.69927
Haus2 M phase 0.000592 1.36318
Nedd1 M phase 0.003755 1.21945
Anapc10 M phase 0.001417 1.36399

88
General Transcription apparatus
Through Functional Annotation Enrichment Analysis, we found that the
expression of general transcription factors used to assemble the RNA Pol I,
Pol II and Pol III complexes was induced in hair cells by Gentamicin
treatment (Table 3). RNA Pol I is responsible for rRNA transcription and
accounts for more than 50% of total RNA synthesis in cells. Expression of
RNA Pol I subunits Polr1a, Polr1b and Polr1e are up-regulated in
Gentamicin treated hair cells. Other basal transcription factors for rRNA
transcription are also induced, such as Taf1a, Taf1b, Taf1c and Taf1d.
RNA Pol II is involved in mRNA synthesis and is tightly regulated by
various mechanisms. Expression of Pol II subunit, Polr2h, is increased after
Gentamicin treatment, and basal transcription factors facilitating
transcription by Pol II, Taf2, Taf3, Taf4b, Taf5, Taf8, Taf9b, Leo1 and
Cdc73, are also expressed at a higher level in Gentamicin-treatment hair
cells. In addition, elongation factors, Ell2, Elp3, Elp4, Tcerg1, and
termination factor Ttf2 are up-regulated.

89
The transcription of 5s rRNA and tRNA is dependent on RNA Pol III.
Expression of subunits of RNA Pol III, Polr3b, Polr3d and Polr3f, is up-
regulated. Snapc1 and Brf2, general transcription factors for Pol III, are also
induced.
The up-regulation of RNA polymerase holoenzyme subunits and general
transcription factors indicates the tendency to increase the transcription for
all types of RNAs in Gentamicin-treated hair cells. However, it has been
discovered that under oxidative stress, transcription is inhibited globally
(Berthiaume et al., 2006). It is unexpected to find that the expression of
those general transcription factors is up-regulated in hair cells after
Gentamicin treatment, since oxidative stress has been thought to be a major
factor of Gentamicin ototoxicity. In fibroblasts, it has been observed that
RNA polymerase activity is suppressed when cells exit the cell cycle and
increased when quiescent cells are stimulated with serum or growth factors
(Johnson et al., 1976; Mauck, 1977; Scott et al., 2001). Therefore, one
possible explanation for our observation, that expression of basal
transcription factors is up-regulated in hair cells after Gentamicin treatment,

90
Table 3. Expression changes in genes coding for the general transcription apparatus.
Gene
symbol Function p value
Fold
change
Polr1a RNA Pol I subunit 0.000959 1.25334
Polr1b RNA Pol I subunit 9.13E-05 1.7852
Polr1e RNA Pol I subunit 4.9E-07 1.52511
Polr2h RNA Pol II subunit 0.000544 1.50516
Polr3b RNA Pol III subunit 0.000142 1.48559
Polr3d RNA Pol III subunit 0.008918 1.24701
Polr3f RNA Pol III subunit 0.009981 1.21421
Polr3e RNA Pol III subunit 0.000276 -1.31977
Gtf2e1 General transcription factor for RNA polymerase II 0.000496 1.33358
Gtf2f1 General transcription factor for RNA polymerase II 0.007396 1.22849
Gtf2h1 General transcription factor for RNA polymerase II 1.57E-05 1.89369
Gtf2h2 General transcription factor for RNA polymerase II 0.000246 1.46909
Taf2 General transcription factor for RNA polymerase II 0.003703 1.32614
Taf3 General transcription factor for RNA polymerase II 0.006964 1.51685
Taf4b General transcription factor for RNA polymerase II 5.07E-05 2.4327
Taf5 General transcription factor for RNA polymerase II 0.00109 1.41099
Taf8 General transcription factor for RNA polymerase II 2.24E-05 1.58281
Taf9b General transcription factor for RNA polymerase II 0.000252 1.85098
Brf2 General transcription factor for RNA polymerase III 0.000241 1.51361
Rrn3 Transcription initiation factor for RNA polymerase I 0.000228 1.23064
Taf1a Transcription initiation factor for RNA polymerase I 0.000205 1.65207
Taf1b Transcription initiation factor for RNA polymerase I 5.66E-05 1.48456
Taf1c Transcription initiation factor for RNA polymerase I 0.003047 1.23859
Taf1d Transcription initiation factor for RNA polymerase I 0.0002 2.03952
Taf6 Transcription initiation factor for RNA polymerase I 0.00016 -1.20964
Snapc1 Transcription initiation factor for RNA polymerase II and III 0.001814 1.53146
Ccnh Transcription elongation factor for RNA polymerase II 0.004257 1.32942
Cdk7 Transcription elongation factor for RNA polymerase II 6.44E-05 1.39676
Mnat1 Transcription elongation factor for RNA polymerase II 0.00464 1.32402
Leo1 Transcription elongation factor for RNA polymerase II 0.001455 1.27087
Cdc73 Transcription elongation factor for RNA polymerase II 0.000052 1.46645
Ell2 Transcription elongation factor for RNA polymerase II 0.005531 1.6064
Elp3 Transcription elongation factor for RNA polymerase II 0.00011 1.20866
Elp4 Transcription elongation factor for RNA polymerase II 0.000437 1.54864
Tcerg1 Transcription elongation factor for RNA polymerase II 4.45E-05 1.65498
Ttf1 Transcription termination factor for RNA polymerase I 0.004302 1.25057
Ttf2 Transcription termination factor for RNA polymerase II 0.002652 1.3741
Dr1 Negative regulator of transcription by RNA polymerase II 0.000592 1.75297

91
is cell cycle re-entry. As discussed earlier, hair cells attempt to re-enter the
cell cycle after Gentamicin treatment. Up-regulating the expression of
general transcription factors might be another attempt of hair cells to prepare
for elevated transcription needs after cell cycle re-entry.
DNA damage and repair
Both Functional Annotation Enrichment Analysis and pathway analysis have
shown that expression of DNA damage and repair factors is up-regulated in
hair cells after Gentamicin treatment, and those genes are summarized in
Table 4.
For interstrand crosslink repair, Ercc4, Dclre1a and Dclre1b are up-
regulated. For mismatch repair, Msh3 and Pms1 are induced. A lot of genes
involved in double strand break are expressed at a higher level, such as
Brcc3, Brip1, Gen1, Fancl, Lig4, Rad52, Nbn and Xrcc4. Another group of
genes is involved in excision repair: Polb and Prmt6 for polymerase activity
to fill the gap; Ape2 for base excision repair sub-pathway; Xpa, Xab2 and
Rad23b for global nucleotide excision repair; Ercc6 and Ercc8 for
transcription-coupled nucleotide excision repair. In addition to higher

92
expression of DNA repair factors, cell cycle checkpoint related proteins are
expressed at higher level after Gentamicin treatment.
Since ROS production induced by aminoglycoside antibiotics has been
considered a major factor of aminoglycoside cytotoxicity and DNA
molecules are vulnerable targets of ROS, it seems that the high level of
DNA repair factor expression is caused by the increased level of DNA
damage. However, DNA damage usually induces post-translational
modifications on those factors rather than transcription up-regulation
(Oberle and Blattner, 2010). Furthermore, we failed to detect the up-
regulation of DNA repair genes at the transcription level when we treated
the organs with the DNA damaging reagent cisplatin (unpublished data from
Ng, Sum-Yan). We hypothesize that up-regulation of DNA repair genes is
the result of attempted cell cycle re-entry by hair cells after Gentamicin
treatment. Consistent with this hypothesis, increased expression of DNA
repair genes has also been observed in neurons when they exit the quiescent
state (Andrusiak et al., 2012).

93
Table 4. Expression changes of genes involved in DNA damage and repair, and cell cycle checkpoint.
Gene symbol Function p value Fold change
Atrip check point 0.000593 1.30435
Eef1e1 check point 0.000287 1.3179
Fbxo31 check point 2.71E-05 1.37891
Fbxo6 check point 0.006493 1.6366
Gadd45a check point 0.007003 1.30902
Kat5 check point 2.23E-06 1.70692
Rad17 check point 0.00152 1.26994
Rint1 check point 0.000725 1.40107
Fancg check point 0.001105 -1.59617
Atm check point 0.001358 -1.69878
Cep164 check point 0.003237 -1.25828
Mdc1 check point 0.000263 -1.47327
Brip1 DNA helicase 0.004553 1.33288
Fbxo18 DNA helicase 0.000753 -1.4394
Lig4 DNA ligase 5.82E-05 1.41918
Polb DNA polymerase 0.006362 1.25865
Rev1 DNA polymerase 0.00064 -1.59406
Apex2 base excision repair 0.000399 1.54913
Prmt6 base excision repair 1.49E-05 2.56831
Neil2 base excision repair 0.001502 -1.52559
Ercc8 nucleotide excision repair 2.55E-05 1.55954
Rad23b nucleotide excision repair 5.89E-05 1.32263
Xab2 nucleotide excision repair 0.000169 1.23787
Xpa nucleotide excision repair 0.001081 1.34703
Ddb1 nucleotide excision repair 6.77E-06 -1.21701
Baz1b double-strand break repair 0.0021 1.31671
Brcc3 double-strand break repair 0.007395 1.22303
Eya4 double-strand break repair 0.000308 1.29523
Fancl double-strand break repair 0.004927 1.5099
Gen1 double-strand break repair 0.002639 1.79456
Rad52 double-strand break repair 7.15E-07 1.9868
Rnf168 double-strand break repair 0.000349 1.37152
Sfpq double-strand break repair 0.000749 1.25277
Xrcc4 double-strand break repair 0.005052 1.32484
Xrcc3 double-strand break repair 5.32E-05 -1.76403
Aplf double-strand break repair 0.000295 -1.41476
Eya3 double-strand break repair 0.001292 -1.22167
Nono double-strand break repair 0.007133 -1.20573
Nbn double-strand break repair, check point 0.000924 1.42728
C230052I12Rik inter-strand cross link repair 0.009441 1.41927
Dclre1a inter-strand cross link repair 0.000196 1.8369

94
Table 4. Continued.
Gene symbol Function p value Fold change
Ercc4 inter-strand cross link repair 0.001488 1.20008
Msh3 mismatch repair 0.005681 1.29044
Pms1 mismatch repair 0.003893 1.38358
Ercc6 nucleotide excision repair 0.002766 1.2879

Protein translation and ribosome synthesis
The excellent bactericidal effect of aminoglycoside antibiotics stems from
the specific inhibition of the prokaryotic ribosome by aminoglycoside
antibiotics (Blanchard et al., 1998). Even though the eukaryotic ribosome is
different from prokaryotic ribosome in sequence and in structure, inhibition
of cytoplasmic ribosomes and mitochondrial ribosomes in eukaryotic cells
by aminoglycoside antibiotics has been observed (Francis et al., 2013;
Shulman et al., 2014). Since protein synthesis is the single most energy-
consuming cellular process (Firczuk et al., 2013), the translation machinery
is controlled through post-translation modification and transcription
regulation by several mechanisms (Jastrzebski et al., 2007; Lempiainen and
Shore, 2009). Based on these observations, we hypothesized that ribosome
inhibition by aminoglycoside antibiotics affected expression of genes
involved in translation. To test this hypothesis, we exmined the expression

95
of genes involved in both cytoplasmic and mitochondrial protein synthesis in
untreated and Gentamicin-treated hair cells (listed in Table 5).
We first discovered significant expression changes in genes involved in
cytoplasmic protein synthesis. Expression of factors involved in rRNA
processing is up-regulated in Gentamicin-treated hair cells, such as nucleolar
proteins, exosome components and small subunit processome components.
Expression of genes involved in tRNA processing is also induced in hair
cells by Gentamicin treatment, including tRNA methyltransferases,
ribonucleases, adenosine deaminase, pseudouridine synthase, queuine
tRNA-ribosyltransferases and other tRNA modification factors. In addition,
expression of translation initiation factors, such as Eif1, Eif2b, Eif3a, Eif4e
and Eif5, is also higher in Gentamicin-treated hair cells. In contrast,
expression of ribosomal proteins, such as Rpl8, Rpl10, Rps4x and Rps5, is
collectively down-regulated.
We also found similar significant expression changes in genes responsible
for mitochondrial protein synthesis. Expression of genes involved in
mitochondria tRNA modification and charging, such as Mto1 and Yars , are

96
induced in hair cells by Gentamicin. Genes responsible for mitochondria
rRNA maturation, such as Mrm1 and Gtpbp5, are expressed at a higher level
after Gentamicin treatment. Additionally, mitochondria translation
regulation factors are up-regulated, including Mtif3, Mtrf1 and Gadd45gip1.
In contrast to the down-regulation of cytoplasmic ribosomal protein
expression, expression of mitochondria counterparts is increased, including
Mrpl1, Mrps2 and 13 other ribosomal proteins.
In postmitotic cells, global protein synthesis inhibition has been reported
(Saelens et al, 2001; Wiita et al, 2013). However, increased transcription of
genes involved in protein synthesis has also been observed during apoptosis
in some cells (Koh et al., 2007; Wiita et al, 2013). In cortical neurons treated
with U18666A (an intracellular cholesterol transportaion inhibitor),
expression of translation factors, mitochondrial ribosomal proteins and
cytoplasmic ribosomal proteins is increased (Koh et al., 2007). In another
study of myeloma apoptosis, protein synthesis factors are expressed at
higher levels after proteasome inhibitor bortezomib treatment, including
translation factors and ribosomal proteins. The difference in our data set is
that expression of cytoplamic ribosomal proteins is repressed in hair cells

97
after Gentamicin treatment. After Gentamicin treatment, hair cells up-
regulate the expression of proteins required for rRNA maturation, tRNA
charging and translation initiation, but down-regulate the expression of
cytoplasmic ribosomal proteins, suggesting that other than the apoptosis
response, expression changes in these genes might reflect Gentamicin-
induced cytoplasmic translation inhibition. Supporting this hypothesis is a
great similarity between the translation machinery expression in hair cells
and that in kidney after Gentamicin treatment (Ozaki et al., 2010) (Further
discussion in a later section of this Chapter).
However, it is unclear why expression of mitochondrial ribosomal proteins
is up-regulated in Gentamicin-treated hair cells. This difference may stem
from Gentamicin-induced differential damage in these two sets of translation
machinery or from different regulatory mechanisms controlling ribosomal
protein expression.

98
Table 5. Expression changes in genes coding for translation machinery factors.
Gene symbol Function p value Fold change
Rpl10 ribosomal protein 0.009703 -1.4795
Rpl13 ribosomal protein 0.009974 -1.40745
Rpl13a ribosomal protein 0.007175 -1.46215
Rpl21 ribosomal protein 0.008394 -1.4207
Rpl22 ribosomal protein 0.00723 -1.28674
Rpl23 ribosomal protein 0.00545 -1.47058
Rpl27a ribosomal protein 0.007478 -1.50796
Rpl8 ribosomal protein 0.006787 -1.35991
Rplp0 ribosomal protein 0.009469 -1.48504
Rps4x ribosomal protein 0.005457 -1.54131
Rps5 ribosomal protein 0.003573 -1.5848
Uba52 ribosomal protein 0.00916 -1.36464
Fau ribosomal protein 0.00856 -1.53234
Ddx51 ribosome maturation 1.29E-05 1.72628
Dimt1 ribosome maturation 0.000412 1.48432
Dis3 ribosome maturation 5.06E-06 2.03283
Ebna1bp2 ribosome maturation 0.000299 1.32017
Ftsj2 ribosome maturation 0.003082 1.46254
Gar1 ribosome maturation 0.00045 1.41333
Gtpbp10 ribosome maturation 0.000125 1.23443
Krr1 ribosome maturation 0.0002 1.30312
Mphosph10 ribosome maturation 1.23E-07 1.64849
Nob1 ribosome maturation 4.79E-06 1.5027
Nop14 ribosome maturation 9.54E-07 1.52282
Nop2 ribosome maturation 1.95E-05 1.60851
Rrp9 ribosome maturation 0.000261 1.44643
Tsr2 ribosome maturation 2.41E-05 1.31477
Utp11l ribosome maturation 0.004222 1.43135
Utp15 ribosome maturation 0.000414 1.25165
Utp18 ribosome maturation 4.37E-06 1.76142
Wdr55 ribosome maturation 0.00661 1.39685
Dcaf13 ribosome maturation 1.31E-05 1.71708
Rpf1 ribosome maturation 1.45E-05 1.95375
Exosc10 ribosome maturation 9.99E-05 1.38
Exosc2 ribosome maturation 0.003093 1.27762
Exosc7 ribosome maturation 0.00065 1.24322
Exosc9 ribosome maturation 0.000604 1.4878
Naf1 ribosome maturation 0.009184 1.23168
Imp3 ribosome maturation 0.002665 -1.29848
Dhx29 translation factor 0.009092 1.2483
Eif1ad translation factor 0.000494 1.23629

99
Table 5. Continued.
Gene symbol Function p value Fold change
Eif2ak1 translation factor 0.006183 1.22713
Eif2b1 translation factor 0.001436 1.20179
Eif2b2 translation factor 3.21E-06 1.57234
Eif2b3 translation factor 2.38E-05 1.27123
Eif2b4 translation factor 0.001361 1.27076
Eif3a translation factor 0.00054 1.25462
Eif4enif1 translation factor 0.000471 1.27425
Eif4e translation factor 0.000293 1.31175
Eif1 translation factor 2.77E-05 1.4947
Eif2s1 translation factor 0.003545 1.33145
Eif5 translation factor 0.000235 1.44185
Eif3f translation factor 0.000149 -1.48772
Eif3h translation factor 0.00597 -1.3386
Eif4b translation factor 0.000897 -1.32013
Ctu1 tRNA processing 0.004 1.25266
Cdk5rap1 tRNA processing 0.000108 1.8195
Clp1 tRNA processing 9.15E-05 1.5308
Ctu2 tRNA processing 0.003901 1.31458
Trmt5 tRNA processing 0.004097 1.26207
Trub2 tRNA processing 0.003355 1.30586
Wdr4 tRNA processing 0.000045 1.27258
Adat2 tRNA processing 0.008091 1.30853
Elac2 tRNA processing 0.00457 1.2189
Pop1 tRNA processing 4.36E-05 3.15943
Pus3 tRNA processing 0.000854 1.31119
Pus10 tRNA processing 0.000497 1.65332
Pus7l tRNA processing 0.000658 1.59777
Pusl1 tRNA processing 0.00854 1.28197
Qtrt1 tRNA processing 9.78E-06 1.53324
Qtrtd1 tRNA processing 5.58E-05 1.46489
Rars tRNA processing 2.32E-05 1.34906
Rpp14 tRNA processing 0.000165 1.3047
Rpp30 tRNA processing 0.002785 1.35845
Rpp38 tRNA processing 0.000732 1.97906
Trmt61a tRNA processing 0.00122 1.33658
Trdmt1 tRNA processing 2.13E-06 1.95721
Trit1 tRNA processing 0.00655 1.21405
Trmt11 tRNA processing 2.59E-06 1.84756
Trmt6 tRNA processing 0.000991 1.48207
Trnt1 tRNA processing 0.004607 1.46614
Tyw3 tRNA processing 0.004686 1.31744

100
Table 5. Continued.
Gene symbol Function p value Fold change
Urm1 tRNA processing 0.001374 1.21238
Elac1 tRNA processing 0.000644 -1.59421
Qars tRNA processing 8.97E-06 -1.39612
Tsen54 tRNA processing 7.04E-05 -1.91495
Lactb mitochondrial ribosomal protein 8.74E-05 1.526
Mrpl1 mitochondrial ribosomal protein 3.42E-05 1.79447
Mrpl16 mitochondrial ribosomal protein 2.66E-05 1.47633
Mrpl17 mitochondrial ribosomal protein 0.000329 1.2684
Mrpl37 mitochondrial ribosomal protein 0.000329 1.32927
Mrpl39 mitochondrial ribosomal protein 0.000651 1.21263
Mrpl44 mitochondrial ribosomal protein 0.000346 1.30152
Mrpl9 mitochondrial ribosomal protein 0.003039 1.25232
Mrps2 mitochondrial ribosomal protein 0.006297 1.22214
Mrps22 mitochondrial ribosomal protein 8.65E-06 1.65178
Mrps26 mitochondrial ribosomal protein 0.000152 1.34783
Mrps27 mitochondrial ribosomal protein 0.004935 1.32814
Mrps28 mitochondrial ribosomal protein 0.003261 1.26174
Mrps30 mitochondrial ribosomal protein 0.000899 1.45369
Mrps31 mitochondrial ribosomal protein 0.000092 1.4415
Mrp63 mitochondrial ribosomal protein 0.001813 -1.24396
Mrpl12 mitochondrial ribosomal protein 0.006071 -1.23746
Mrpl14 mitochondrial ribosomal protein 0.000332 -1.5309
Mrpl23 mitochondrial ribosomal protein 0.003565 -1.38508
Mtif3 mitochondrial translation 0.005568 1.23179
Gadd45gip1 mitochondrial translation 0.000395 1.35737
Mtrf1 mitochondrial translation 0.001739 1.53799
Mtrf1l mitochondrial translation 0.000217 1.81782
Ptcd3 mitochondrial translation 0.000105 1.62439
Aadat tRNA processing in mitochondrion 2.68E-05 1.70856
Ddx28 tRNA processing in mitochondrion 7.21E-06 1.94949
Mto1 tRNA processing in mitochondrion 7.46E-05 1.65278
Nars2 tRNA processing in mitochondrion 6.01E-05 1.59786
Osgepl1 tRNA processing in mitochondrion 0.000112 1.8943
Rnmtl1 tRNA processing in mitochondrion 3.03E-05 1.76029
Rpusd4 tRNA processing in mitochondrion 0.001931 1.33929
Yars2 tRNA processing in mitochondrion 1.58E-06 2.02635
Mtfmt tRNA processing in mitochondrion 0.003493 -1.27496
Pars2 tRNA processing in mitochondrion 0.002771 -1.41355
Gtpbp5
ribosome maturation in
mitochondrion 8.84E-05 1.4594
Mrm1
ribosome maturation in
mitochondrion 0.003249 1.43193

101
Mitochondria
It has long been believed that ROS induced by aminoglycoside antibiotics is
a major factor in aminoglycoside cytotoxicity. There are two sources of ROS,
catalytic reactions caused by the aminoglycoside-ion chelating complex, and
damaged mitochondria. Mitochondrial metabolism in hair cells is affected
immediately by Gentamicin treatment (Tiede et al., 2009), suggesting that
mitochondria are a primary target of aminoglycoside antibiotics. Some
patients carrying mitochondrial rRNA mutations are supersensitive to
aminoglycoside antibiotics (Prezant et al., 1993), further supporting the idea
that mitochondria are attacked by aminoglycosides. Therefore, we examined
the expression changes of mitochondria related genes (Table 6).
In addition to the up-regulated expression of genes involved in
mitochondrial protein synthsis, expression of another >120 mitochondrial
related genes is affected in hair cells by Gentamicin treatment. Factors
involved in mitochondrial transcription and mRNA synthesis, such as Tfam,
Tfb2m and Polg2, are up-regulated. Expression changes are also found for
genes involved in metabolisms in mitochondria, including carbohydrate
metabolism, lipid metabolism and the respiratory chain. Expression of

102
mitochondrial transporters, such as Abcd2 and Slc25a42, is de-regulated.
Another group of de-regulated genes is involved in stress and apoptosis,
such as Dnajc11, Gfer and Chchd4 for chaperone activity, Msrb2, Oxr1 and
Coq6 for antioxidantive activity, Bcl2l2, Mcl1, Parl and Triap1 for apoptosis.
The expression changes in mitochondrial related genes suggest that
mitochondria function is affected by Gentamicin treatment. In another study
investigating global cellular response during apoptosis, it has been shown
that expression of mitochondrial related genes is de-regulated during
apoptosis, including genes involved in mitochondrial translation,
transcription, metabolism, transport, stress, and apoptosis (Ozaki et al.,
2010). Therefore, based on our RNA sequencing data, we could not reach a
conclusion whether mitochondria are direcetly damaged by Gentamicin or
mitochondrial dysfunction is secondary to apoptosis.
Table 6. Expression changes of mitochondrial related genes.
Gene symbol Function p value Fold change
Lyrm1 Apoptosis 0.002245 1.33208
Mcl1 Apoptosis 0.001166 1.41694
Parl Apoptosis 0.000205 1.3389
Triap1 Apoptosis 0.002636 1.21442
Bcl2l2 Apoptosis 0.000312 -1.32403
Nit1 Apoptosis 0.008444 -1.23194
Pacs2 Apoptosis 0.000342 -1.36848
Top1mt DNA replication in mitochondrion 0.000325 -1.66992

103
Table 6. Continued.
Gene symbol Function p value Fold change
Polg DNA replication in mitochondrion 0.000205 -1.30074
Abhd10 mitochondrial metabolism 3.96E-06 1.454
Acot8 mitochondrial metabolism 9.41E-05 1.29188
Agk mitochondrial metabolism 0.00059 1.21103
Agps mitochondrial metabolism 0.001096 1.3124
Cecr5 mitochondrial metabolism 0.000478 1.24134
Coq10b mitochondrial metabolism 0.001703 1.63207
Coq6 mitochondrial metabolism 0.000673 1.50808
Cox10 mitochondrial metabolism 0.003019 1.51111
Cox11 mitochondrial metabolism 0.000159 1.92261
Dbt mitochondrial metabolism 0.000195 1.53627
Foxred1 mitochondrial metabolism 0.000432 1.35249
Fpgs mitochondrial metabolism 0.003783 1.44845
Golph3 mitochondrial metabolism 0.001755 1.20237
Gpd1 mitochondrial metabolism 0.000292 2.49198
Hccs mitochondrial metabolism 0.002293 1.27629
Lias mitochondrial metabolism 0.000521 1.41775
Lipe mitochondrial metabolism 0.000513 1.51588
Lipt1 mitochondrial metabolism 0.000639 1.42576
Mccc1 mitochondrial metabolism 0.00124 1.29656
Mecr mitochondrial metabolism 0.000237 1.61723
Ndufaf4 mitochondrial metabolism 0.00015 1.5771
Oxnad1 mitochondrial metabolism 0.003137 1.41817
Pde12 mitochondrial metabolism 1.77E-06 1.88993
Pdhx mitochondrial metabolism 0.000419 1.26152
Pdss1 mitochondrial metabolism 3.84E-06 2.11078
Pdss2 mitochondrial metabolism 1.48E-05 1.69608
Sco1 mitochondrial metabolism 7.86E-07 1.71241
Sco2 mitochondrial metabolism 0.002131 1.90669
Sdhaf1 mitochondrial metabolism 0.003661 1.35775
Secisbp2 mitochondrial metabolism 9.48E-05 1.43298
Slc25a38 mitochondrial metabolism 7.79E-05 1.70549
Tmem70 mitochondrial metabolism 0.000144 1.49919
D2hgdh mitochondrial metabolism 0.001635 1.21479
Ucp1 mitochondrial metabolism 0.000247 1.99598
Slc25a14 mitochondrial metabolism 0.006547 1.22836
Abat mitochondrial metabolism 0.002967 -1.38497
Atp5d mitochondrial metabolism 0.000283 -1.20397
Cds2 mitochondrial metabolism 0.000165 -1.65079
Ndufb9 mitochondrial metabolism 2.45E-06 -2.05738

104
Table 6. Continued.
Gene symbol Function p value Fold change
Ndufv3 mitochondrial metabolism 0.004651 -1.38591
Acat1 mitochondrial metabolism 0.000711 -1.29627
Acaa2 mitochondrial metabolism 0.009898 -1.25423
Acsl3 mitochondrial metabolism 0.000638 -1.27564
Acss3 mitochondrial metabolism 0.005365 -1.71278
Adc mitochondrial metabolism 8.58E-06 -1.58137
Cbr4 mitochondrial metabolism 0.000181 -1.33011
Cyb5r3 mitochondrial metabolism 0.00335 -1.30441
Cox4i1 mitochondrial metabolism 0.00064 -1.35569
Cox4i2 mitochondrial metabolism 0.002532 -2.5024
Cox7a2l mitochondrial metabolism 0.00033 -1.54036
Ech1 mitochondrial metabolism 0.005525 -1.31235
Idh3g mitochondrial metabolism 0.00177 -1.26541
Me2 mitochondrial metabolism 0.007216 -1.30279
Mlycd mitochondrial metabolism 2.05E-05 -1.38796
Mmab mitochondrial metabolism 0.007634 -1.39394
Nudt8 mitochondrial metabolism 0.006533 -1.33444
Oat mitochondrial metabolism 0.008936 -1.63595
Atp5g2 mitochondrial metabolism 0.000201 -1.32047
Atp5h mitochondrial metabolism 0.007986 -1.28541
Uqcrh mitochondrial metabolism 0.000704 -1.40782
Pcx mitochondrial metabolism 0.00279 -1.65712
Pdk2 mitochondrial metabolism 0.003137 -1.54417
Pdk3 mitochondrial metabolism 0.001667 -1.55856
Sucla2 mitochondrial metabolism 0.000478 -1.20618
Car5b mitochondrial metabolism 0.000126 -1.34881
Capza2 mitochondrial morphologenesis 0.002239 1.23912
Rhot1 mitochondrial morphologenesis 0.00279 1.2683
Spata5 mitochondrial morphologenesis 1.45E-07 2.70772
Tmem11 mitochondrial morphologenesis 0.000131 1.46863
Usp30 mitochondrial morphologenesis 0.000212 1.295
Wasf1 mitochondrial morphologenesis 0.005891 1.22938
Polg2 mitochondrial mRNA prcessing 0.003826 1.38377
Tfam mitochondrial mRNA prcessing 0.001149 1.37182
Tfb2m mitochondrial mRNA prcessing 3.11E-06 1.60785
Gm9897 mitochondrial mRNA prcessing 0.003544 2.12039
Mtpap mitochondrial mRNA prcessing 2.24E-05 1.52836
Hnrpll mitochondrial mRNA prcessing 0.000167 1.31324
Abce1 mitochondrial mRNA prcessing 5.93E-05 1.29047
Abcd3 mitochondrial transportator 0.00016 1.20011
Ccdc90a mitochondrial transportator 2.82E-07 2.09789

105
Table 6. Continued.
Gene symbol Function p value Fold change
Slc25a26 mitochondrial transportator 0.000472 1.62755
Slc25a29 mitochondrial transportator 0.009938 1.27133
Slc35b3 mitochondrial transportator 8.67E-06 1.78679
Slc3a1 mitochondrial transportator 0.007556 1.35495
Star mitochondrial transportator 0.000183 1.99837
Yrdc mitochondrial transportator 2.84E-05 1.82816
Slc25a25 mitochondrial transportator 0.000492 2.36215
Abcb10 mitochondrial transportator 0.000385 -1.49971
Sfxn5 mitochondrial transportator 2.96E-05 -2.21538
Slc25a11 mitochondrial transportator 0.00083 -1.28068
Slc25a23 mitochondrial transportator 6.04E-05 -1.56565
Slc25a39 mitochondrial transportator 0.003457 -1.34156
Slc25a42 mitochondrial transportator 2.42E-05 -2.82819
Ucp2 mitochondrial transportator 1.58E-05 -1.76587
Vdac1 mitochondrial transportator 0.002928 -1.3764
Clpp protein metabolism in mitochondrion 0.005099 1.29899
Clpx protein metabolism in mitochondrion 7.24E-05 1.42542
Yme1l1 protein metabolism in mitochondrion 0.00707 1.26357
Ppif protein metabolism in mitochondrion 0.001135 -1.36891
Alkbh7 Stress in mitochondrion 0.003729 1.37287
Dnajc11 Stress in mitochondrion 3.44E-05 1.23341
Gfer Stress in mitochondrion 0.001933 1.32408
Msrb2 Stress in mitochondrion 0.00391 1.60307
Oxr1 Stress in mitochondrion 6.06E-06 1.72376
Oma1 Stress in mitochondrion 0.00205 1.45429
Chchd4 Stress in mitochondrion 2.66E-05 1.21718
Snn Stress in mitochondrion 0.008276 -1.35268

Stress and cell death
Since aminoglycoside antibiotics induce hair cell death through apoptosis
(Forge and Li, 2000; Cunningham et al., 2002), we examined expression
changes of genes involved in stress response and apoptosis. Gentamicin

106
treatment affects the expression of stress response genes and apoptosis
related genes in hair cells (Table 7).
Expression of Ddit3, the key transcription regulator of the ER stress
response, is up-regulated (2.75 fold). However, expression of other ER stress
markers, such as Hspa5 (Grp78, -1.3 fold) and Hsp90b1 (Grp94, -1.5 fold),
is not induced by Gentamicin, suggesting the absence of ER stress at this
time point. Atg12, Atg14 and Atg5, responsible for autophagosome
maturation, are expressed at higher level, indicating possible activation of
autophagy in Gentamicin-treated hair cells. As stated in previous sections,
expression of genes involved in cell cycle checkpoint, DNA damage and
repair, is up-regulated in response to aminoglycoside-induced stress.
In Gentamicin-treated hair cells, expression of both anti- and pro-apoptosis
genes is affected. Up-regulated anti-apoptosis genes include 4632434I11Rik,
Bnip1 and Ebag9, and some other anti-apoptosis genes are down-regulated,
including Bag1, Bcl2l12, Bcl2l2 and Tmbim6. Genes involved in Caspase
cascade are also affected at transcription level, such as Casp2 (-1.8 fold),
Casp9 (-1.2 fold) and Casp3 (1.3 fold). Expression of genes involved in the

107
death receptor pathway and the Trp53 pathway, such as Rassf1, Tnfrsf10b,
tnfrsf18, Trp53bp2 and Ccar1, changes upon Gentamicin treatment.
Altered expression of stress and cell death genes is expected in Gentamicin-
treated hair cells. In IPA Up-stream analysis, we found that the state of
Trp53 was inhibited with an activation Z score of -3.17 and overlap p-value
of 4.70E-09, suggesting that the p53 pathway is not activated in Gentamicin-
treated hair cells. This result is consistent with our previous observation that
p53 knockout organs have similar sensitivity to Gentamicin ototoxicity as
wildtype organs.
Table 7. Expression changes of stress response genes and cell death genes.
Gene symbol Function p value Fold change
4632434I11Rik anti-apoptosis 8.41E-05 2.94643
Bnip1 anti-apoptosis 0.000174 1.50276
Ebag9 anti-apoptosis 0.008552 1.26425
Bag1 anti-apoptosis 4.31E-06 -1.4104
Bcl2l12 anti-apoptosis 0.00761 -1.36995
Bcl2l2 anti-apoptosis 0.000312 -1.32403
Tmbim6 anti-apoptosis 3.29E-05 -1.51497
Mcl1 apoptosis regulation 0.001166 1.41694
Pdcd2 apoptosis regulation 2.95E-05 1.58397
Stk17b apoptosis regulation 0.000325 2.07653
Stk3 apoptosis regulation 0.003602 1.25325
Klf11 apoptosis regulation 3.83E-06 3.18778
Zdhhc16 apoptosis regulation 0.000947 1.27241
Ppm1f apoptosis regulation 0.002186 -1.94099
Rnf130 apoptosis regulation 0.001545 -1.35487
Atg14 autophage 1.08E-05 1.744883
Atg12 autophage 0.007599 1.31901
Atg5 autophage 0.000371 1.41132

108
Table 7. Continued.
Gene symbol Function p value Fold change
Dap autophage 0.007849 -1.35459
Lrdd Caspase cascade 0.004729 1.44689
Casp3 Caspase cascade 0.002432 1.30768
Pdcl3 Caspase cascade 0.000535 1.32184
Rffl Caspase cascade 0.004349 1.33833
Rnf34 Caspase cascade 2.98E-06 1.50082
Thoc1 Caspase cascade 3.93E-05 1.4374
Triap1 Caspase cascade 0.002636 1.21442
Casp2 Caspase cascade 0.001698 -1.81613
Casp9 Caspase cascade 0.004844 -1.28884
Tns4 Caspase cascade 0.002387 -1.25262
Rassf1 cell death through death receptor 0.000204 1.99947
Fem1b cell death through death receptor 0.001244 1.33195
Tnfrsf10b cell death through death receptor 0.000927 -1.43136
Tnfrsf18 cell death through death receptor 9.63E-06 -2.22731
Aimp2 cell death through p53 pathway 0.001103 1.31391
Ccar1 cell death through p53 pathway 9.12E-05 1.62621
Ppp1r15a cell death through p53 pathway 0.009454 1.73969
Rybp cell death through p53 pathway 0.002196 1.37985
Topors cell death through p53 pathway 7.89E-07 2.06667
Trp53bp2 cell death through p53 pathway 0.001889 1.35363
Hipk2 cell death through p53 pathway 0.002977 -1.44062
Hif1a hypoxia 0.003081 1.29486
Ddit3 stress 2.62E-05 2.75188
Hspa5 stress 0.004463 -1.3884
Pxdn stress 0.001212 -2.26354
Prdx2 stress 0.005064 -1.21185
Uaca stress 0.000132 -1.43138

JNK pathway
The protein kinase/c-Jun N-terminal kinase (SAPK/JNK) pathway is
involved in cell growth, differentiation, survival and apoptosis, and this
pathway is activated by stress stimuli (Bogoyevitch and Kobe, 2006). In hair

109
cells, it has been discovered that the JNK pathway is activated after
aminoglycoside treatment and that inhibition of the JNK pathway protects
hair cells from aminoglycoside ototoxicity (Pirvola et al., 2000; Ylikoski,
2002; Sugahar et al., 2006). Therefore, we were interested in the expression
of genes involved in the JNK pathway (Table 8).
JNK proteins are phosphorylated by upstream kinases after stress, and the
then active p-JNK phosphorylates downstream substrates, such as c-Jun,
ATF2, ELK1, p53 and NFAT4 (Bogoyevitch and Kobe, 2006). For upstream
factors, expression of positive regulators, Gadd45a, Gadd45g, Map3k5,
Map3k6 and negative regulator, Pdcd4 is induced in hair cells by
Gentamicin. The expression of Jnk1 (Mapk8, 1.4 fold) itself is also up-
regulated. In addition, the expression of JNK target genes is affected by
Gentamicin; Atf2 (1.4 fold) and c-Jun (Jun, 2.7 fold) are expressed at higher
levels, while Elk1 (-1.4 fold) is expressed at a lower level. Even though the
activity of the JNK pathway component is highly regulated through post-
translational phosphorylation (Bogoyevitch and Kobe, 2006), the increased
expression of c-Jun indicates the activation of the JNK pathway because the
expression of the c-Jun gene is under the control of c-Jun itself, which is

110
Table 8. Expression changes of genes involved in the JNK pathway.
Gene symbol Function p value Fold change
Atf2 substrate of JNK 0.000845 1.42346
Dusp8 upstream regulator of JNK 0.000361 4.92245
Elk1 substrate of JNK 0.000957 -1.45779
Gadd45a upstream regulator of JNK 0.007003 1.30902
Gna12 upstream regulator of JNK 0.000171 -1.28694
Gng2 upstream regulator of JNK 0.004548 -2.27777
Gng7 upstream regulator of JNK 1.87E-06 -1.66464
Jun substrate of JNK 0.00245 2.73299
Kras upstream regulator of JNK 0.001369 1.30701
Map3k3 MAPKKK 0.000638 -1.79641
Map3k4 MAPKKK 0.001049 -1.5858
Map3k5 MAPKKK 0.000481 1.3673
Map3k9 MAPKKK 0.001964 -1.48858
Map3k10 MAPKKK 0.000321 -1.36354
Map4k2 MAPKKKK 0.00576 -1.30819
Map4k5 MAPKKKK 2.18E-05 1.36399
Mapk8 JNK 4.11E-05 1.38856
Mapk8ip1 scaffold protein for JNK 0.002792 -1.31827
Mink1 MAPKKKK 0.000199 -1.55777
Nfatc1 substrate of JNK 0.009843 1.20859
Pik3c2a upstream regulator of JNK 0.000132 1.64071
Pik3c2b upstream regulator of JNK 0.001625 -1.61926
Pik3ca upstream regulator of JNK 4.12E-05 -1.2901
Pik3r2 upstream regulator of JNK 0.004985 -1.22969
Pik3r3 upstream regulator of JNK 0.005551 1.34233
Rras upstream regulator of JNK 0.001567 -1.90346
Sos1 upstream regulator of JNK 0.000608 1.29741
Tab1 upstream regulator of JNK 0.00524 -1.23636
Traf2 upstream regulator of JNK 0.000645 1.37403
Braf upstream regulator of JNK 0.005642 1.24368

activated by JNK phosphorylation. Interestingly, expression of another AP1
component, c-Fos (Fos, -2.2 fold), is down-regulated in Gentamicin treated
hair cells. It has been suggested that Atf2 and c-Fos compete for
dimerization with c-Jun. In potassium deprivation-induced neuronal

111
apoptosis, c-Jun dimerizes with Atf2 and controls downstream target gene
expression, while c-Fos transcription is reduced (Yuan et al., 2009).
Increased expression of c-Jun and Atf2 and decreased expression of c-Fos in
Gentamicin treated hair cells suggest that there is a similarity between
Gentamicin-induced hair cell death and neuronal cell death induced by
potassium deprivation.
NF- κB pathway
The NF- κB pathway is another pathway important for cell survival and cell
death (Kucharczak et al., 2003). In the canonical NF- κB pathway, NF- κB is
sequestered in the cytoplasm by I κB. Upon activation by signals from
membrane receptors, I κB Kinase (IKK) phosphorylates I κB and thus
interrupts the interaction of NF- κB and I κB, and NF- κB thereafter
translocates into nucleus and initiates the expression of downstream target
gene (Karin and Ben-Neriah, 2000). In hair cells, it has been found that
inhibition of NF- κB caused hair cell death (Nagy et al., 2005) and activation
of NF- κB protected hair cells from aminoglycoside antibiotics (Jiang et al.,
2005). Therefore, we examined the expression changes of genes involved in
NF- κB pathway.

112
Table 9. Expression changes of genes involved in the NF- κB pathway.
Gene symbol Function p value Fold change
Btrc I κB degradation 0.003418 1.32168
Chuk I κB degradation 0.000336 1.33418
Ikbkg I κB degradation 0.000947 -1.14412
Nfkbib I κB subunit 0.000177 1.47658
Nfkb1 NF- κB subunit 0.000599 1.4881
Csnk2a1 phsphorylate NF- κB 0.005818 -1.30308
Il33 regulator of NF- κB transcription 0.000195 -3.25076
Prkaca regulator of NF- κB transcription 8.98E-06 -1.41853
Ube2v1 upstream regulator of NF- κB 2.43E-05 -1.31343
Akt1 upstream regulator of NF- κB 0.002857 -1.33926
Akt3 upstream regulator of NF- κB 0.00218 -1.26203
Atm upstream regulator of NF- κB 0.001358 -1.69878
Azi2 upstream regulator of NF- κB 7.02E-08 -1.41411
Bcl10 upstream regulator of NF- κB 0.006697 1.25087
Bmpr2 upstream regulator of NF- κB 0.000179 -1.30109
Fgfr1 upstream regulator of NF- κB 0.005303 -1.4831
Igf1r upstream regulator of NF- κB 0.000916 -1.3136
Igf2r upstream regulator of NF- κB 0.000139 -1.87728
Insr upstream regulator of NF- κB 0.001631 -1.53154
Irak4 upstream regulator of NF- κB 0.000675 -1.37758
Kras upstream regulator of NF- κB 0.001369 1.30701
Malt1 upstream regulator of NF- κB 2.12E-05 1.40772
Map3k3 upstream regulator of NF- κB 0.000638 -1.79641
Map3k14 upstream regulator of NF- κB 0.000325 2.0297
Pdgfrb upstream regulator of NF- κB 0.003063 -2.62884
Pik3c2a upstream regulator of NF- κB 0.000132 1.64071
Pik3c2b upstream regulator of NF- κB 0.001625 -1.61926
Pik3ca upstream regulator of NF- κB 4.12E-05 -1.2901
Pik3r2 upstream regulator of NF- κB 0.004985 -1.22969
Pik3r3 upstream regulator of NF- κB 0.005551 1.34233
Plcg2 upstream regulator of NF- κB 0.004878 1.3201
Prkcb upstream regulator of NF- κB 0.005603 -2.58088
Rras upstream regulator of NF- κB 0.001567 -1.90346
Tank upstream regulator of NF- κB 0.00208 1.34189
Tbk1 upstream regulator of NF- κB 0.002482 1.33465
Tgfbr2 upstream regulator of NF- κB 0.001416 -1.96806
Tlr9 upstream regulator of NF- κB 0.00124 1.73688
Traf2 upstream regulator of NF- κB 0.000645 1.37403
Traf3 upstream regulator of NF- κB 0.004047 -1.21531

113
Factors in the NF- κB pathway with altered expression in hair cells after
Gentamicin treatment are listed in Table 9. The upstream signaling factors,
Traf2, Tank, Bcl10 and Malt1, involved in TNF receptor signal transduction,
are induced in hair cells by Gentamicin. Expression of I κB kinase, Ikk1
(Chuk) and I κB ubiquitination enzyme, Btrc, is up-regulated. In addition,
NF- κB1 (Nfkb1, 1.5 fold) and I κB β (Nfkbib, 1.5 fold) are also expressed at
higher level. Since expression of both NF- κB and I κB β is induced when the
NF- κB pathway is active (Ten et al., 1992; Renner and Schmitz, 2009), the
elevated messenger level of NF- κB and I κB β indicates the activation of the
NF- κB pathway in hair cell after Gentamicin treatment.
6.6 Pathway analysis of non-hair cell expression profiles
There are fewer genes affected by Gentamicin treatment in non-hair cells.
Thourgh IPA pathway analysis, we found several pathways were over-
represented, mainly biosynthesis pathways (example genes in Table 10). The
first over-represented pathway is the cholesterol biosynthesis pathway, and
genes involved in cholesterol synthesis are generally up-regulated. In
addition to the cholesterol synthesis pathway, many of other biosynthesis
pathways are affected with up-regulated gene expression, such as

114
Geranylgeranyl diphosphate biosynthesis, mevalonate pathway, ketogenesis,
purine nucleotide de novo synthesis, and pyrimidine ribonucleotide de novo
synthesis.
Table 10. Altered gene expression in non-hair cells. This is not a complete list of differentially expressed
genes.
Gene Symbol Functions p Value Fold change
Cyp51 cholesterol synthesis 0.001207 1.5303
Dhcr24 cholesterol synthesis 0.00016 1.81536
Dhcr7 cholesterol synthesis 0.008023 1.43552
Fdft1 cholesterol synthesis 0.003584 1.44386
Fdps cholesterol synthesis 0.001996 1.56075
Hmgcr cholesterol synthesis 0.006199 1.31223
Hmgcs1 cholesterol synthesis 0.000234 1.44677
Hsd17b7 cholesterol synthesis 0.00312 1.7418
Idi1 cholesterol synthesis 0.000314 1.76379
Lss cholesterol synthesis 0.006246 1.51283
Mvk cholesterol synthesis 0.002229 1.85485
Nsdhl cholesterol synthesis 0.004794 1.58728
Sc4mol cholesterol synthesis 0.003643 1.48995
Sqle cholesterol synthesis 0.000277 1.67235
Acat2 geranylgeranyldiphosphate synthesis 0.001888 1.51789
Fdps geranylgeranyldiphosphate synthesis 0.001996 1.56075
Sc4mol zymosterol synthesis 0.003643 1.48995
Adsl purine de novo synthesis 0.008632 1.36849
Atic purine de novo synthesis 4.34E-05 1.34064
Gart purine de novo synthesis 0.001085 1.37505
Cad pyrimidine de novo synthesis 0.000264 1.71232
Ctps pyrimidine de novo synthesis 0.002654 1.25618
Entpd6 pyrimidine de novo synthesis 0.002378 1.20357
Nme1 pyrimidine de novo synthesis 0.008075 1.34783

The elevated biosynthesis activity might be attributed to elevated activities
in supporting cells, even though we pooled supporting cells with other GFP

115
negative cells as non-hair cell population. In a Texas Red-Conjugated
Gentamicin uptake experiment, hair cells are the only cell type showing
fluorescence in cultured cochleae (Fig 1), therefore gene expression in
supporting cells is most likely to be affected by Gentamicin due to their cell-
cell contact with hair cells. During inflammation, cholesterol synthesis is
increased (Omoigui, 2007), and the up-regulated expression of cholesterol
synthesis genes could be the initial inflammation response in supporting
cells after hair cell damage.
6.7 Similarity between aminoglycoside ototoxicity and nephrotoxicity
Nephrotoxicity is another side effect of aminoglycoside antibiotics and
tremendous work has been done to investigate the nephrotoxicity of
aminoglycoside antibiotics (Mingeot-Leclercq et al., 1999; Lopez-Novoa et
al., 2011). In IPA Up-stream analysis, we found that there was a great
similarity in gene expression alterations between Gentamicin-treated hair
cells and Gentamicin-treated kidney cells with an overlapping p-value of
8.02E-08. In the kidney, 211 genes involved in intracellular vesicle
movement, membrane transport, fatty acid metabolism, protein translation,
proliferation, cell death and stress/DNA repair mechanisms, have been

116
identified by microarray as de-regulated genes with significant changes after
Gentamicin treatment (Ozaki et al, 2010), and 50 of these genes overlap
with our hair cell dataset, showing consistent expression alterations (Table
11).
As we discussed before, expression of genes involved in protein translation
is affected in hair cells by Gentamicin. We compared the expression changes
of these genes in hair cells with that in kidney, Tars, involved in tRNA
charging, Nop58, Ncl, involved in ribosome biogenesis, Mrpl23,
mitochondrial ribosome protein, and Eif2b2, Eif2b4, Eef1a1, translation
factors. These genes are affected at the transcription level in Gentamicin-
treated hair cells in the same way as observed in Gentamicin-treated kidney,
suggesting that Gentamicin causes similar damage to the translation
machinery in both hair cells and kidney cells.
We observed up-regulation of genes involved in the cell cycle and DNA
damage and repair in hair cells after Gentamicin. However, the molecular
analysis of kidney response was carried out on whole kidney consisting of at
least nine cell types, making direct comparison of gene expression changes

117
difficult. Purified hair cells are a postmitotic population with high
homogeneity, while the kidney consists of both dividing cells and quiescent
cells.
One significant difference between gene expression changes in hair cells and
in kidney is the expression of ER stress genes, Hspa5 and Hsp90b1. As we
stated before, both Hspa5 and Hsp90b1 are expressed at lower levels in hair
cells after Gentamicin treatment. But in kidney, Gentamicin treatment
induces up-regulation of these genes. In addition, we did not observe
increased expression for other Hsp genes, such as Hsp90aa1, Hspa1a,
Hspa1b, Hspb1 and Hyou1. One possible explanation for the difference
between Hsp gene expression in hair cells and kidney cells is that
Gentamicin causes different kinds of stress in hair cells and in kidney cells.
Or the inconsistent expression of Hsp genes is caused by different timepoints
at which samples were collected (3 hours of Gentamicin treatment for hair
cells purification versus 7 days of Gentamicin injection for kidney sample
preparation).

118
Despite the difference, great similarities exist between Gentamicin-induced
gene expression changes in hair cells and in kidney, suggesting that there are
some similarities between underlying mechanisms responsible for
aminoglycoside ototoxicity and nephrotoxicity. This is consistent with the
observation that expression of many gene isoforms is shared between these
tissues (Gabashvili et al., 2007; Quick et al., 1973).
Table 11. Gentamicin-induced similar gene expression changes in hair cells and in kidney.
Gene symbol Function p Value
Fold
Change
Findings in
nephrotoxicity Consistent
Zfp622 transcription regulator 4.77E-06 1.483148 Upregulates Yes
Tsc22d1 transcription regulator 0.000157 1.315102 Upregulates Yes
Tars tRNA synthesis 0.001263 1.292011 Upregulates Yes
St6galnac3 glycolipid synthesis 0.005244 -1.46649 Downregulates Yes
Smn1 RNA splicing 0.000161 1.45206 Upregulates Yes
Sfpq RNA splicing 0.000749 1.252767 Upregulates Yes
S100a1 calcium binding 0.006299 -1.37041 Downregulates Yes
Ptbp1 RNA splicing 3.83E-05 1.636352 Upregulates Yes
Ppp1r15a phosphatase 0.009454 1.739694 Upregulates Yes
Ppp1r10 phosphatase 5.84E-06 4.524746 Upregulates Yes
Phax RNA export 0.006074 1.438425 Upregulates Yes
Pelo mitosis 0.000148 1.319478 Upregulates Yes
Pcx metabolism 0.00279 -1.65712 Downregulates Yes
Pafah1b1
inactivates platelet-
activating factor 0.000164 1.346669 Upregulates Yes
Nop58 ribosome biogenesis 7.68E-08 1.941254 Upregulates Yes
Ncl ribosome biogenesis 0.00327 1.206809 Upregulates Yes
Naa15 N-acetyltransferase 0.000102 1.461136 Upregulates Yes
Mybbp1a transcription regulator 0.000956 1.214954 Upregulates Yes
Mrpl23 mitochondrial ribosome 0.003565 -1.38508 Downregulates Yes
Mdm2 p53 pathway regulation 0.005424 1.28435 Upregulates Yes
Lig4 DNA repair 5.82E-05 1.41918 Upregulates Yes
Kti12 unkown 0.008232 1.649551 Upregulates Yes
2310044G17Rik unkown 0.005972 1.37605 Upregulates Yes
Jun transcription factor 0.00245 2.732988 Upregulates Yes
Hacl1 metabolism 0.002033 -1.43919 Downregulates Yes

119
Table 11. Continued.
Gene symbol Function p Value
Fold
Change
Findings in
nephrotoxicity Consistent
Gstm1 antioxidant 0.00249 -1.32453 Downregulates Yes
Gnl3 p53 pathway regulation 7E-06 2.127794 Upregulates Yes
Gfer growth factor 0.001933 1.324081 Upregulates Yes
Gadd45a MAPK pathway regulator 0.007003 1.30902 Upregulates Yes
Eif2b4 translation 0.001361 1.270756 Upregulates Yes
Eif2b2 translation 3.21E-06 1.572336 Upregulates Yes
Eef1a1 translation 2.35E-05 -2.0177 Downregulates Yes
Eed transcription repressor 8.35E-06 1.601639 Upregulates Yes
Eaf1 transcription factor 0.000222 1.369313 Upregulates Yes
Dnaja2 chaperone 0.001429 1.282581 Upregulates Yes
Decr2 metabolism 0.003445 -1.21269 Downregulates Yes
Ddx50 RNA helicase 0.00147 1.313465 Upregulates Yes
Ddx39 RNA splicing 4.98E-06 1.991224 Upregulates Yes
Ddit3 transcription factor 2.62E-05 2.751878 Upregulates Yes
Cul2 ubiquitin ligase 7.77E-06 1.510931 Upregulates Yes
Cnksr3 sodium transport 0.00051 1.240931 Upregulates Yes
Chka phospholipid synthesis 6.28E-07 2.029311 Upregulates Yes
Cdk13 transcription elongation 0.000244 1.365288 Upregulates Yes
Ccnl1 RNA splicing 4.46E-06 2.482795 Upregulates Yes
Cand1 ubiquitin ligase 0.003349 1.215329 Upregulates Yes
Azin1 metabolism 0.001597 1.388636 Upregulates Yes
Asl urea cycle 0.000314 -1.27227 Downregulates Yes
Arf6 Ras superfamily 0.000976 1.388019 Upregulates Yes
Acsl3 metabolism 0.000638 -1.27564 Downregulates Yes
Abat metabolism 0.002967 -1.38497 Downregulates Yes
Zfp354a unkown 0.001648 1.544937 Downregulates No
Zfp36l1 growth factor response 0.001571 -1.65211 Upregulates No
Tubb5 cytoskeleton 0.001836 -2.0123 Upregulates No
Spp1 cell-matrix interaction 0.004118 -4.30391 Upregulates No
Pdia4 protein folding 0.002689 -1.72559 Upregulates No
Nedd4 ubiquitin ligase 2.45E-06 -1.27335 Upregulates No
Hyou1 stress 0.002956 -1.22211 Upregulates No
Hspa5 stress 0.004463 -1.3884 Upregulates No
Hsp90b1 chaperone 5.87E-05 -1.56328 Upregulates No
Egr1 transcription regulator 0.004467 -1.61013 Upregulates No
Ctsd protease 0.00174 -1.33383 Upregulates No
Calr calcium binding 0.000413 -1.4266 Upregulates No

120
6.8 Discussion
We profiled the gene expression changes in hair cells in response to
aminoglycoside antibiotics by sequencing messenger RNA from purified
hair cells after Gentamicin treatment. This high throughput method provided
us with a more comprehensive view of the transcriptional response to
aminoglycoside ototoxicity.
We found that hair cells attempted to re-enter the cell cycle by activating the
cell cycle machinery after Gentamicin treatment. Expression of Cyclin B2,
Cyclin E2, and CAKs are increased in Gentamicin-treated hair cells,
indicating an attempt to activate CDK, the driving force for cell cycle
progression. Additionally, genes involved in E2F transcriptional activity
regulation, S and G2 phase progression control, and mitosis regulation are
also affected by Gentamicin. In neurons with Rb gene knockout, up-
regulated expression of genes involved in cell cycle phases other than G1
phase has been observed (Andrusiak et al., 2012), which is consistent with
our data here. The IPA Up-stream analysis reveals that there is a similarity
in Myc target gene expression patterns between Gentamicin-treated hair
cells and other dividing cells. These results together indicate that hair cells

121
are attempting to re-enter the cell cycle following the treatment with
Gentamicin.
In addition to the altered expression of cell cycle related genes, up-regulated
expression of the general transcription apparatus supports the idea of cell
cycle re-entry. Transcription of all three RNA polymerase subunits, as well
as related basal transcription factors, is increased in hair cells after
Gentamicin treatment, suggesting an attempt to enlarge the pool of factors
representing the transcription apparatus. In studies of cell cycle exit and re-
entry with fibroblasts, transcription activity is repressed in G0 phase and
stimulated when cells re-enter the cell cycle (Johnson et al., 1976; Mauck,
1977; Scott et al., 2001). The up-regulated expression of genes to expand the
transcription apparatus reservoir could be a related attempt to re-enter the
cell cycle in hair cells after Gentamicin treatment.
Another observation supporting the cell cycle re-entry hypothesis is the
induction of DNA damage and repair genes and cell cycle checkpoint genes.
Generally, DNA damage stimulates DNA repair factors at the post-
translational level, but not the transcription level. In support of this

122
hypothesis, we failed to detect the induction of DNA repair gene expression
in hair cells by the DNA damaging reagent cisplatin (unpublished data from
Ng, Sum-Yan). Based on this analysis, we suggest that up-regulation of
DNA repair gene expression is not caused by actual DNA damage, but rather
a result of cell cycle re-entry attempt in hair cells after Gentamicin treatment.
In research comparing DNA repair in proliferating and quiescent cells,
expression of many repair factors has been found reduced when cells exit the
cell cycle (Nouspikel and Hanawalt, 2002). On the other hand, expression of
DNA repair genes is up-regulated in postmitotic neurons when they are
forced to re-enter the cell cycle (Andrusiak et al., 2012). Therefore, we think
the up-regulated expression of DNA damage repair genes could be attributed
to an attempt at cell cycle re-entry of hair cells after Gentamicin treatment.
All taken together, altered expression of cell cycle related genes, the general
transcription apparatus, DNA damage and repair genes, suggest that the cell
cycle machinery is activated in hair cells by Gentamicin. However, we failed
to detect DNA synthesis (Fig 9) and cell cycle checkpoint activation (data
not shown) before hair cell death, suggesting that the cell cycle machinery
activation does not lead to G1/S transition or DNA syntheis in hair cells, but

123
rather contributes to apoptosis through direct or indirect involvement in
regulating cell death pathways. Consistent with this hypothetical model, we
found inhibition of CDK activity protected hair cell from aminoglycoside
ototoxicity (Chapter 3).
Disruption of protein synthesis by aminoglycoside antibiotics has been
proposed in a ribotoxic stress model, and this hypothesis is supported by the
observation of cytoplasmic protein synthesis inhibition by Gentamicin in
hair cells (Francis et al., 2013). Protein synthesis is tightly regulated so that
cells have adequate supplies to fulfill the needs of the cell, and not much
energy is wasted to synthesize much more than needed. Protein synthesis is
regulated by the quantity of active ribosomes and the rate limiting translation
initiation step (Lempiainen and Shore, 2009; Sonenberg and Hinnebusch,
2009). Another property of the protein synthesis process is that translation of
mRNA into polypeptide requires all three major types of RNAs (mRNA,
tRNA and rRNA), so factors controlling RNA processing and modification
also contribute to protein synthesis regulation, thus the synthesis of these
components needs to be coordinated at the transcriptional level and the
translation level (Frugier et al., 2005; Granneman and Baserga, 2005). Here

124
we demonstrated that expression of cytoplasmic and mitochondrial
translation machinery genes in hair cells are affected by Gentamicin. The
expression alterations in ribosomal protein genes, translation initiation
factors, tRNA and rRNA processing factors indicate the de-regulation of
protein synthesis in hair cells after Gentamicin treatment. Even though
transcription up-regulation of translation machinery has also been observed
in other apoptotic cells (Koh et al., 2007; Wiita et al, 2013), hair cells have a
unique transcriptional response, with down-regulation of cytoplasmic
ribosomal proteins and up-regulation of other cytoplamic machinery factors,
suggesting that this transcription response is specific to Gentamicin
treatment. The similarity between expression patterns of translation
machinery in Gentamicin-treated hair cells and kidney supports the
hypothesis that Gentamicin induces a specific transcriptional response in
genes involved in peptide translation. This hypothesis is consistent with the
ribotoxic stress model, even though the underlying regulation mechanism is
not clear.
In contrast, expression changes of genes involved in mitochondrial
translation machinery, transcription regulation, metabolism, transport, stress

125
and cell death, are similar to the transcription response in apoptotic cells
(Ozaki et al., 2010). It is unclear whether the expression changes in
mitochondrial related genes are due to the direct effect of Gentamicin or a
secondary response to the apoptosis pathway.
Involvement of the JNK pathway and the NF- κB pathway in hair cell death
has been implicated (Pirvola et al., 2000; Ylikoski et al., 2002; Sugahar et
al., 2006; Nagy et al., 2005; Jiang et al., 2005). When the JNK pathway is
inhibited, hair cells are less sensitive to aminoglycoside antibiotics. In
contrast, inhibition of NF- κB pathway leads to hair cell death, and activation
of the NF- κB pathway protects hair cells from aminoglycoside ototoxicity.
From our expression data, we found both pathways are stimulated,
suggesting that both pro- and anti-death signaling pathways are activated in
hair cells by Gentamicin.
The analysis of gene expression in hair cells after Gentamicin treatment
reveals that several mechanisms/pathways participate in aminoglycoside
ototoxicity, including but not limited to cell cycle re-entry, ribotoxic stress,
JNK pathway and NF- κB pathway. Outcomes from all of those pathways

126
can add up and lead to the hair cell apoptosis, making inhibition of one
single pathway not an optimized approach to protect hair cells. Also,
aminoglycoside antibiotics have properties specific for each single drug,
such as different affinity for ribosome (Scheunemann et al., 2010), so they
might affect some pathways preferentially. This might explain why different
aminoglycoside antibiotic drugs have differential ototoxicity (Nakashima et
al., 2000) and why inhibiting a certain pathway only works for some drugs
but not others (Coffin et al., 2013).

127
Chapter 7: Effect of mCherry expression in hair cells
Green Fluorescent Protein (GFP) was first discovered in Aequorea jellyfish
by Shimomura in 1961, and many other fluorescent proteins were
subsequently identified from other bioluminescent organisms. In the mid-
1990s, the scientific community started to utilize those proteins for
biological fluorescent imaging applications. With modifications of the
chromophores, the utilization of fluorescent proteins was expanded in
research, including subcellular localization, protein-protein interaction, cell
labeling and linage tracking (Campbell, 2008). One limitation of fluorescent
proteins concerns their cytotoxicity when expressed at high level.
Particularly, the cytotoxicity of DsRed is found to be associated with the
suppression of Bcl-xL translation (Zhou et al., 2011). With high fluorescent
intensity and relatively low cytotoxicity, enhanced Green Fluorescent
Protein (EGFP) and mCherry are common fluorescent proteins used in
biological studies. When fluorescent proteins themselves are expressed in
HeLa cells, the cytotoxicity of mCherry is comparable to EGFP (Shemiakina
et al., 2012). However, when fluorescent proteins are fused with other

128
proteins or peptides, cytotoxicity should be tested carefully (Link et al.,
2006).
The transgenic Atoh1-GFP mouse line, was created by Jane E. Johnson’s
group and first used to analyze inner ear expression by Chen et al (Lumpkin
et al., 2003). In this transgenic line, GFP gene is under the control of Atoh1
enhancers and GFP protein is expressed specifically in cells that express the
Atoh1 gene (Lumpkin et al., 2003). Additionally, a nuclear localization
signal peptide coding sequence is inserted into the GFP gene to produce
nuclear-localized GFP. The Atoh1-GFP reporter gene faithfully mimics
expression of endogenous Atoh1 expression in hair cells. In cochlea, GFP
fluorescence appears around E13.5 and disappears around P6 in hair cells
(Lumpkin et al., 2003). In utricles, GFP is still detectable in some hair cells
in adult animals (data not shown). For ototoxicity studies using neonatal
animals, Atoh1-GFP can be used as a hair cell marker for a sensitive real-
time in vitro assay of hair cell survival in response to ototoxic drugs (this
work). Additionally, Atoh1-GFP allows the purification of hair cells by
FACS for a direct analysis of gene expression by Q-PCR or RNA
sequencing.

129
Some transgenic mouse lines have also been established to label supporting
cells in the inner ear, such as the p27-GFP (White et al., 2006), and Lfng-
GFP (Korrapati et al., 2013) mouse lines. To double label both hair cells and
supporting cells in different colors for purification, Atoh1-mCherry
transgenic mice were created in our lab. In this mouse line, the mCherry
gene and H2b gene are fused and placed under the control of Atoh1
enhancers, and the red mCherry-H2b fusion protein is expressed specifically
in sensory hair cells. In contrast to the Atoh1-GFP line, mCherry
fluorescence persists in adult animals in both cochlea and utricles (Fig 25).
In our study of ototoxicity with neonatal mice, hair cell damage to the organ
of Corti from Atoh1-GFP mice is comparable to that in organs from wild
type animals (Chapter 1). However, organs from Atoh1-mCherry mice are
resistant to ototoxicants. Additionally, Atoh1-mCherry animals begin to lose
hearing at a very young age, indicating the expression of mCherry-H2b
fusion protein in hair cells is toxic. How the expression of Atoh1-mCherry in
hair cells affects the susceptibility to ototoxins is discussed later.

130
7.1 Less sensitivity to otoxins in mCherry-positive hair cells
To test how mCherry-positive hair cells responded to ototoxin treatment, we
treated organotypic cultures of cochleae from different transgenic mice with
several ototoxic reagents, such as H
2
O
2
, cisplatin and Gentamicin.
We first tested H
2
O
2
at 1 mM to examine the difference between transgenic
lines. Atoh1-GFP organs started hair cell loss, indicated by GFP
fluorescence disappearance, at 1 hour and the damage was severe at 3 hours
(data not shown). However, Atoh1-mCherry organs still had nicely
organized hair cells marked by mCherry even 3 hours after treatment. At this
time point, almost all supporting cells were damaged by H
2
O
2
, while hair
cells were intact (n =2) (Fig 23A). The results suggest that hair cells are not
sensitive to H
2
O
2
when mCherry-H2b is expressed.
We then treated organs with cisplatin at 0.2mM for 3 hours and counted hair
cells 24 hours later. In wildtype organs, more than 70% (n=1) of the outer
hair cells were lost, while Atoh1-mCherry organs retained half of
Parvalbumin positive/mCherry-positive hair cells (47%, n=1), suggesting the
resistance of mCherry-positive hair cells to cisplatin (Fig 23B).

131

Figure 23. Hair cells in Atoh1-mCherry organs are resistant to H
2
O
2
and cisplatin. A, Atoh1-mCherry
hair cells and Lfng-GFP supporting cells 3 hours after 1mM H
2
O
2
treatment. H
2
O
2
kills supporting cells,
but hair cells are relatively intact (n=2), suggesting mCherry hair cells are resistant to H
2
O
2
. B,
Parvalbumin staining for hair cells after cisplatin treatment. WT organs lose more hair cells than Atoh1-
mCherry organs (28% outer hair cell remaining in WT versus 47% in mCherry-positive), suggesting
mCherry-positive hair cells are less sensitive to cisplatin. C, Staining for cisplatin-DNA adducts showing
similar cisplatin-DNA adduct levels in WT and mCherry-positive hair cells, indicating no difference in
cisplatin absorption.

132
We confirmed the uptake of cisplatin by hair cells using immunostaining
with the antibody against cisplatin-DNA adducts. We observed staining for
cisplatin-DNA adducts in both wildtype hair cells and mCherry-positive hair
cells (Fig 23C), indicating that there is no difference in drug absorption
between wildtype and mCherry hair cells.
In addition to H
2
O
2
and cisplatin treatment, organs were also treated with 0.5
mM Gentamicin for 3 hours. As shown by Parvalbumin staining, wildtype
organs lost more than 85% +/- 8% (n=2) of outer hair cells in the base, while
Atoh1-mCherry organs retained more than half of outer hair cells (54% +/-
7%, n=2) (Fig 24A), suggesting Atoh1-mCherry positive hair cells are less
sensitive to Gentamicin. Texas-Red conjugated gentamicn (GTTR) was used
to verify the entry of Gentamicin in hair cells. GTTR fluorescence was
observed specifically in hair cells after 30 min incubation in both wildtype
and mCherry organs (Fig 24B), suggesting there is no difference in drug
absorption.
To further investigate how different hair cells responded to Gentamicin
treatment, organs were incubated with active Caspase indicator FITC-VAD-

133
FMK. Nine hours after Gentamicin treatment, green fluorescence was
detected in hair cell from wildtype organs, indicating the activation of the
Caspase cascade in those hair cells by Gentamicin. In contrast, there was no
FITC signal observed in hair cells from Atoh1-mCherry organs (Fig 24C),
suggesting no Caspase activation at this time point. Together these results
show that mCherry positive hair cells are less sensitive to Gentamicin.
7.2 Hearing loss at young age in Atoh1-mCherry mice
In organotypic cultures, expression of mCherry-H2b has protective effects
against ototoxins, we therefore tested whether Atoh1-mCherry mice had
better hearing at older age.
We performed Auditory Brainstem Response (ABR) to assess the hearing
ability. As early as three weeks, threshold shifts were evident in Atoh1-
mCherry mice across the spectrum tested. By 5 weeks, Atoh1-mCherry mice
(CD1 background) were almost deaf at all frequencies tested while wildtype
littermates and Lfng-GFP mice had normal hearing (n=3 for each genotype)
(Fig 25A). Two transgenic founder lines were tested to confirm the effect of

134

Figure 24. mCherry-positive hair cells are less sensitive to Gentamicin-induced hair cell apoptosis.
A, Parvalbumin staining showing relatively mild damage in Atoh1-mCherry organ (15% of outer hair cells
remaining in WT versus 54% in mCherry positive organs, n=2). B, Fluorescence of GTTR showing similar
uptake of Gentamicin in WT and Atoh1-mCherry hair cells. GTTR and mCherry fluorescence can be
differentiated as shown. C, FTIC-VAD-FMK fluorescence showing Caspase activation in WT organs 9
hours after Gentamicin treatment but not in Atoh1-mCherry organs.

135
transgene expression. Both lines showed worse hearing in Atoh1-mCherry
mice than that in WT littermates (data not shown), suggesting the expression
of the transgene causes deafness in animals.
ABR is an experiment to examine the function of the whole auditory system.
We further investigated whether hair cells are damaged in Atoh1-mCherry
mice by staining with antibody against hair cell marker MyoVI. As shown
by the MyoVI staining, wildtype littermates had intact outer and inner hair
cells at basal, middle and apical segments at the age of 3 weeks. In 3 week-
old Atoh1-mCherry mice, however, there were only several outer hair cells
remaining in the basal segment and less than half of the outer hair cells were
stained in the middle segment (Fig 25B). These results are consistent with
the ABR results, suggesting the hearing loss observed in Atoh1-mCherry
mice is caused by hair cell loss.
7.3 Discussion
Both ABR experiments and hair cell immunostaining results with mature
animals showed that Atoh1-mCherry mice had worse hearing due to hair cell

136

Figure 25. Atoh1-mCherry mice experience hearing loss and hair cell damage at a very young age.
A, ABR results showing a dramatic threshold shift in Atoh1-mCherry mice compared to WT or Lfng-GFP
mice (n=3 for each genotype), suggesting Atoh1-mCherry mice lose hearing ability. B, Staining for MyoVI
showing severe hair cell loss in Atoh1-mCherry animals in the base and middle.

137
loss than wildtype littermates, inconsistent with the protective effect of
mCherry-H2b expression in neonatal cochlea against ototoxin. However, we
need more physiology experiments to verify whether mCherry positive hair
cells are as active as wildtype hair cells in neonatal organs. One possibility is
that mCherry-H2b expression disturbs the normal activity of hair cells,
leading to lesser sensitivity to ototoxins in neonatal cochleae and hair cell
loss in mature animals.

138
Chapter 8: Materials and methods:
Organotypic culture and drug treatment
Inner ears from P1 mice were dissected out under sterile condition and then
transferred onto a plate filled with sterile Ca
2+
- and Mg
2+
-free PBS.
Cochleae were dissected using a stereo-dissecting microscope, and then
spiral ganglion and lateral wall were removed carefully. Cochleae were
mounted on membranes (13mm diameter, 1.0 μm pore size, SPI supplies)
floating on DMEM/F12 medium (Invitrogen) supplemented with 1% N2
(Gibco) and 100U/ml penicillin (Sigma). Organ cultures were maintained
under low oxygen condition in an incubator (37°C, 5% CO
2
, and 5% O
2
).
After overnight recovery, organs were treated with 0.5 mM Gentamicin
(Sigma) for 3 hours, followed by a fresh medium change. For CDK inhibitor
experiments, Olomoucine (Alexis Biochemicals) and CVT-313
(CalBiochem) were added into the medium together with Gentamicin and
kept in medium after Gentamicin was removed. For the active Caspase
detection assay and mitochondria potential detection, FITC-VAD-FMK
(Promega) or Rhodamine 123 (Sigma) was added at a 1:1000 dilution after
Gentamicin was washed out.

139
For organotypic culture of utricles from 3 week-old mice, the inner ear was
pulled out from the skull under sterile condition, and then the utricles were
dissected as described in literature (Cunningham, 2006). Otoconia were
removed by a gentle stream of PBS from an insulin syringe, then the organs
were mounted on a membrane and cultured under the same conditions as
described above. Utricles were treated with 2.0mM Gentamicin without
being washed out.
Fluorescent immunostaining and western blotting
Fixed organs were permeabilized with 0.2% (for neonatal cochea) or 1.0%
(for cochlea and utricles from older animals) Triton X-100 in PBS at room
temperature for 2 hours, then blocked with 10% donkey serum in PBS for
another 2 hours. After this and organs were incubated at 4°C overnight with
rabbit anti-MyoVI, rabbit anti-MyoVIIa (Proteus Biosciences), and mouse
anti-Parvalbumin (Sigma) at dilutions of 1:500. After washing with PBS six
times, Alex 594 or Alex 488 conjugated secondary antibodies (Molecular
Probes) were added at a dilution of 1:500 and the organs incubated at room
temperature for 3 hours. For hair cell staining in adult cochlea, inner ears
with an opening at the apex were fixed with 4% paraformaldehyde, rotating

140
at room temperature overnight, followed by decalcification with 300 mM
EDTA/4% paraformaldehyde for 24 hours. Whole inner ears were stained as
described above, then the cochleae were removed and mounted on slides.
Cell purification and Quantitative-PCR
To make single cell suspensions from neonatal cochlea, organs were
digested with 0.05% trypsin (Invitrogen) and 1 mg/ml Collagenase
(Worthington) at 37°C for 8 min and then triturated with a p200 pipette
around 300 times. The suspension was filtered by cell strainer to get rid of
clumps before FACS sorting with BD FACSAria II. Purity of the sorted
cells was confirmed by immunostaining and Q-PCR. RNA was extracted
from purified cell populations using a Zymo Quick-RNA Microprep kit for
RNA sequencing or for Q-PCR. The cDNA library was made by qScript
reverse transcriptase supermix (Quanta Biosciences). Primers are listed in
Table 12.
Table 12. Primer pairs for Q-PCR.
Gene Forward Reverse
Rpl19 GGTCTGGTTGGATCCCAATG CCCGGGAATGGACAGTCA
Gapdh AGGTCGGTGTGAACGGATTTG TGTAGACCATGTAGTTGAGGTCA
c-Jun CCTTCTACGACGATGCCCTC GGTTCAAGGTCATGCTCTGTTT
Bim CCCGGAGATACGGATTGCAC GCCTCGCGGTAATCATTTGC
Bax TGAAGACAGGGGCCTTTTTG AATTCGCCGGAGACACTCG

141
RNA sequencing and data analysis
An Illumina True-Seq mRNA-seq kit was used to make libraries from hair
cell RNA. Six samples were pooled into one lane and sequenced by Illumina
Hi-Seq 2000 for single end for 50 cycles (50bp reads). The sequencing data
were trimmed and aligned against mouse assembly mm10 using Partek Flow.
Differential gene expression was done by imbedded GSA in Partex Flow.
Ingenuity Pathway Analysis (IPA, Ingenuity), ExPlain 3.1 (BioBase
Biological Database), DAVID v6.7 and Expression2Kinase (Ma’ayan Lab,
Mountain Sinai School of Medicine) were used to do gene annotation,
function clustering, and pathway analysis.
ABR and Trans-tympanic injection
Animals were anaesthetized with an intraperitoneal injection of ketamine
hydrochloride (50 mg/kg) and xylazine hydrochloride (10 mg/kg) and
maintained at 37°C during the procedure. ABR recording was performed as
described previously (Chen et al., 2003). Under anesthetic conditions, 25 μl
50 mg/ml Gentamicin solution was injected through the tympanic membrane
using a 30G needle into right ear and the mice were kept lying on the left

142
side until awake. All animal care and procedures were approved by the
IACUC at the House Research Institue.
Quantification and statistics
Pictures of the whole cochlea explants were assembled in Photoshop and the
length was measured in ImageJ (average length for P1 cochleae was 5600
microns according to our measurements). The first 400 microns were
excluded from hair cell quantification as the hair cells in the hook region are
resistant to ototoxic treatment. The rest was divided into four segments with
equal length (1300 microns for the basal segment if whole organ pictures
were not available). Outer hair cells in each segment were counted and the
numbers normalized to the length of the segment.
For hair cell counts in utricles, sampling squares were drawn randomly in
striola or extra-striola regions. MyoVIIa positive hair cells were counted in
each sampling square and the numbers normalized to the areas of sampling
squares.
For statistic analysis, at least 3 independent samples were prepared and a
student’s t test was used to check the significance.

143
DNA extraction and Genotyping
Tails were clipped from pups at 5mm from the end, and then digested with
Zandy buffer and Proteinase K (1 mg/ml, Fisher Scientific) at 63°C for 4
hours. DNA was extracted from digested tails by using the phenol-
chloroform method and DNA was resolved into 50 μL dH
2
O. Primers for
genotyping are summarized in Table 13.
Table 13. Primer sequences for Genotyping. Amplicons of different lengths can be differentiated by
electrophoresis with proper concentration of agrose gel.
Allele Forward Reverse Length
Cdk2 WT GTGACCCTGTGGTACCGAGCACCT GGTTTTGCTCTTGGATGTGGGCATGG 130
Cdk2 KO GTGACCCTGTGGTACCGAGCACCT CCCGTGATATTGCTGAAGAGCTTGG 400
Cdk4 WT CGGAAGGCAGAGATTCGCTTAT CCAGCCTGAAGCTAAGAGTAGCTGT 195
Cdk4 KO CGGAAGGCAGAGATTCGCTTAT ATATTGCTGAAGAGCTTGGCGG 315
p19 WT CAAGATGCCTCCGGTACTAG TCCCTTCTTCAATGGACAGG 375
p19 KO CACGAGATTTCGATTCCACC AAGCTGACCACGGAGCTATG 500
p21 WT AAGCCTTGATTCTGATGTGGGC TGACGAAGTCAAAGTTCCACC 872
p21 KO AAGCCTTGATTCTGATGTGGGC GCTATCAGGACATAGCGTTGG 700
p53 WT ATAGGTCGGCGGTTCAT CCCGAGTATCTGGAAGACAG 600
p53 KO CTTGGGTGGAGAGGCTATTC AGGTGAGATGACAGGAGATC 280

144
References
Alharazneh, A., Luk, L., Huth, M., Monfared, A., Steyger, P.S., Cheng,
A.G. and Ricci, A.J. (2011). Functional Hair Cell Mechanotransducer
Channels Are Required for Aminoglycoside Ototoxicity. PLoS One, 6(7):
e22347.
Andrusiak, M.G., Vandenbosch, R., Park, D.S. and Slack, R.S. (2012).
The Retinoblastoma Protein Is Essential for Survival of Postmitotic Neurons.
The Journal of Neuroscience, 32(42):14809 –14814.
Angel, P., Hattori, K., Smeal, T. and Karin, M. (1988). The jun proto-
oncogene is positively autoregulated by its product, Jun/AP-1. Cell, 55(5):
875–885.
Appert-Collin, A., Hugel, B., Levy, R., Niederhoffer, N., Coupin, G.,
Lombard, Y., André, P., Poindron, P. and Gies, J.P. (2006). Cyclin
dependent kinase inhibitors prevent apoptosis of postmitotic mouse
motoneurons. Life Sciences, 79: 484-490.
Barriere, C., Santamarıa, D., Cerqueira, A., Galan, J., Martın, A.,
Ortega, S., Malumbres, M., Dubus, P. and Barbacid, M. (2007). Mice
thrive without Cdk4 and Cdk2. Molecular Oncology, 1:72-83.
Berthiaume, M., Boufaied, N., Moisan, A. and Gaudreau, L. (2006).
High levels of oxidative stress globally inhibit gene transcription and histone
acetylation. DNA and Cell Biology, 25(2): 124-134.
Besirli, C.G. and Johnson, E.M. (2003). JNK-independent Activation of c-
Jun during Neuronal Apoptosis Induced by Multiple DNA-damaging Agents.
The Journal of Biological Chemistry, 278(25):22357-22366.
Biswas, S.C., Liu, D.X. and Greene, L.A. (2005). Bim is a direct target of
a neuronal E2F-dependent apoptotic pathway. J. Neurosci., 25(37):8349-
8358.

145
Blanchard, S.C., Fourmy, D., Eason, R.G. and Puglisi, J.D. (1998).
rRNA Chemical Groups Required for Aminoglycoside Binding.
Biochemistry, 37:7716–7724.
Bogoyevitch, M.A. and Kobe, B. (2006). Uses for JNK: the Many and
Varied Substrates of the c-Jun N-Terminal Kinases. Microbiol. Mol. Biol.
Rev. 70(4):1061.
Brooks, E.E., Gray, N.S., Joly, A., Kerwar, S.S., Lum, R., Mackman,
R.L., Norman, T.C., Rosete, J., Rowe, M., Schow, S.R., Schultz,
P.G., Wang, X., Wick, M.M. and Shiffman, D. (1997). CVT-313, a
specific and potent inhibitor of CDK2 that prevents neointimal proliferation.
The Journal of Biololgical Chemistry. 272(46):29207-29211.
Campbell, R.E. (2008). Fluorescent proteins. Scholarpedia, 2008.
Chen, P., Zindy, F., Abdala, C., Liu, F., Li, X., Roussel, M.F. and Segil,
N. (2003). Progressive hearing loss in mice lacking the Cyclin-dependent
Kinase Inhibitor Ink4d. Nature Cell Biology, 5: 422-426.
Clerici, W.J., Hensley, K., DiMartino, D.L. and Butterfield, D.A. (1996).
Direct detection of ototoxicant-induced reactive oxygen species generation
in cochlear explants. Hearing Research, 98:116–124.
Coffin, A.B., Rubel, E.W. and Raible, D.W. (2013). Bax, Bcl2, and p53
differentially regulate neomycin- and Gentamicin-induced hair cell death in
the zebrafish lateral line. J. Assoc. Res. Otolaryngol., 14(5):645-659.
Cohen, G.M. (1997). Caspases: the executioners of apoptosis. Biochemical
Journal, 326:1-16.
Corwin, J.T. and Cotanche, D.A. (1988). Regeneration of sensory hair
cells after acoustic trauma. Science, 240(4860):1772-1774.
Cunningham, L.L. (2006). The adult mouse utricle as an in vitro
preparation for studies of ototoxic-drug-induced sensory hair cell death.
Brian Research, 1091:277-281.

146
Cunningham, L.L., Cheng, A.G. and Rubel, E.W. (2002). Caspase
activation in hair cells of the mouse utricle Exposed to neomycin. The
Journal of Neuroscience, 22(19): 8532-8540.
Cuningham, L.L., Matsui, J.I., Warchol, M.E. and Rubel, E.W. (2004).
Overexpression of Bcl-2 prevents neomycin-induced hair cell death and
Caspase-9 activation in adult mouse utricle in vitro. Journal of Neurobiology,
60(1): 89-100.
Dang, C.V., O'Donnell, K.A., Zeller, K.I., Nguyen, T., Osthus, R.C.
and Li, F. (2006). The c-Myc target gene network. Semin. Cancer
Biol., 16(4):253-264.
Duvall, A.J., Flock, A. and Wersall, J. (1966). The Ultrastructure of the
sensory hairs and associated organelles of the cochlear inner hair cell, with
reference to directional sensitivity. The Journal of Cell Biology, 29:497-505.
Ehret, G. (1976). Development of absolute auditory thresholds in the house
mouse (Mus musculus). Journal of the American Audiology Society, 1:179-
184.
Firczuk, H., Kannambath, S., Pahle, J., Claydon, A., Beynon, R.,
Duncan, J., Westerhoff, H., Mendes, P. and McCarthy, J. (2013). An in
vivo control map for the eukaryotic mRNA translation machinery.
Molecular Systems Biology, 9:635.
Florea, A. and Büsselberg, D. (2011). Cisplatin as an Anti-Tumor Drug:
Cellular Mechanisms of Activity, Drug Resistance and Induced Side Effects.
Cancers, 3: 1351-1371.
Folch, J., Junyent, F., Verdaguer, E., Auladell, C., Pizarro, J.G., Beas-
Zarate, C., Pallàs, M. and Camins, A. (2012). Role of cell cycle re-entry in
neurons: a common apoptotic mechanism of neuronal cell death. Neurotox
Res., 22(3):195-207.
Forge, A. and Li, L. (2000) Apoptotic death of hair cells in mammalian
vestibular sensory epithelia. Hearing Research, 139: 97-115.

147
Forge, A. and Schacht, J. (2000). Aminoglycoside antibiotics. Audiology &
Neuro-Otology, 5: 3-22.
Francis, S.P., Katz, J., Fanning, K.D., Harris, K.A., Nicholas, B.D., Lacy,
M., Pagana, J., Agris, P.F. and Shin, J.B. (2013). A Novel Role of
Cytosolic Protein Synthesis Inhibition in Aminoglycoside Ototoxicity. The
Journal of Neuroscience, 33(7):3079-3093.
Friedman, T.B., Sellers, J.R. and Avraham K.B. (1999). Unconventional
myosins and the genetics of hearing loss. American Journal of Medical
Genetics, 89(3): 147-157.
Frugier, M., Ryckelynck, M. and Giegé, R. (2005). tRNA-balanced
expression of a eukaryal aminoacyl-tRNA synthetase by an mRNA-
mediated pathway. EMBO Reports, 6(9):860-865.
Gabashvili, I.S., Sokolowski, B.H., Morton, C.C. and Giersch, A.B.
(2007). Ion Channel Gene Expression in the Inner Ear. Journal of the
Association for Research in Otolaryngology, 8: 305-328.
Gavrieli, Y., Sherman, Y. and Ben-Sasson, S.A. (1992). Identification of
programmed cell death in situ via specific labeling of nuclear DNA
fragmentation. Journal of Cell Biology, 119:493–501.
Ghahremani, M.H., Keramaris, E., Shree, T., Xia, Z., Davis, R.J.,
Flavell, R., Slack, R.S. and Park, D.S. (2002). Interaction of the c-jun/JNK
pathway and Cyclin-dependent Kinases in death of embryonic cortical
neurons evoked by DNA damage. The Journal of Biological Chemistry,
277(38): 35586-35596.
Granés, F., Roig, M.B., Brady, HJ.M. and Gil-Gómez, J. (2004). Cdk2
activation acts upstream of the mitochondrion during glucocorticoid induced
thymocyte apoptosis. European Journal of Immunology, 34(10): 2781-2790.
Granneman, S. and Baserga, S.J. (2005). Crosstalk in gene expression:
coupling and co-regulation of rDNA transcription, pre-ribosome assembly
and pre-rRNA processing. Current Opinion in Cell Biology, 17:281–286.

148
Greene, L.A., Liu, D.X., Troy, C.M. and Biswas, S.C. (2007). Cell
cycle molecules define a pathway required for neuron death in development
and disease. Biochim. Biophys. Acta., 1772(4):392-401.
Hawkins, R.D., Bashiardes, S., Powder, K.E., Sajan, S.A., Bhonagiri, V.,
Alvarado, D.M., Speck, J., Warchol, M.E. and Lovett, M. (2007). Large
Scale Gene Expression Profiles of Regenerating Inner Ear Sensory Epithelia.
PloS ONE, 2(6):e525.
Hess, J., Angel, P. and Schorpp-Kistner, M. (2004). AP-1 subunits:
quarrel and harmony among siblings. Journal of Cell Science, 117(25):
5965-5973.
Huang, H., Regan, K.M., Lou, Z., Chen, J. and Tindall, D.J. (2006).
CDK2-dependent phosphorylation of FoxO1 as an apoptotic response to
DNA damage. Science, 314:294-297.
Hunt, M.A., Miller, S.W., Nielson, H.C. and Horn, K.M. (1987).
Intratympanic injection of sodium arsanilate (atoxyl) solution results in
postural changes consistent with changes described for labyrinthectomized
rats. Behav. Neurosci., 101(3):427-428.
Jamesdaniel, S., Ding, D., Kermany, M.H., Davidson, B.A., Knight 3rd,
P.R., Salvi, R. and Coling, D.E. (2008). Proteomic analysis of the balance
between survival and cell death responses in cisplatin-mediated ototoxicity.
Journal of Proteome Research, 7: 3516-3524.
Jastrzebski, K., Hannan, K.M., Tchoubrieva, E.B., Hannan, R.D.
and Pearson, R.B. (2007). Coordinate regulation of ribosome biogenesis
and function by the ribosomal protein S6 kinase, a key mediator of mTOR
function. Growth Factors, 25(4):209-26.
Jayaraman S. (2003). Intracellular determination of activated caspases
(IDAC) by flow cytometry using a pancaspase inhibitor labeled with FITC.
Cytometry Part A, 56A:104-122.

149
Jiang, H., Sha, S.H. and Schacht, J. (2005). NF- κB Pathway Protects
Cochlear Hair Cells From Aminoglycoside-Induced Ototoxicity. Journal of
Neuroscience Research, 79:644–651.
Jiang, X. and Wang, X. Cytochrome C-mediated apoptosis. Annual Review
of Biochemistry, 73:87-106.
Johnson, G.L. and Lapadat, R. (2002). Mitogen-activated protein kinase
pathways mediated by ERK, JNK, and p38 protein kinases.
Science, 298(5600):1911-1912.
Johnson, L.F., Levis, R., Abelson, H.T., Green, H. and Penman, S.
(1976). Changes in RNA in relation to growth of the fibroblast. IV.
Alterations in theproduction and processing of mRNA and rRNA in resting
and growing cells. The Journal of Cell Biology, 71:933-938.
Johnson, L.V., Walsh, M.L. and Chen L.B. (1980). Localization of
mitochondrial in living cells with rhodamine 123. Proceedings of the
National Academy of Sciences, 77(2):990-994.
Karin, M. and Ben-Neriah, Y. (2000). Phosphorylation Meets
Ubiquitination: The Control of NF- κB Activity. Annual Review of
Immunology, 18:621–663.
Koh, C.H., Peng, Z.F., Ou, K., Melendez, A., Manikandan, J., Qi, R.Z.
and Cheung, N.S. (2007). Neuronal apoptosis mediated by inhibition of
intracellular cholesterol transport: microarray and proteomicsanalyses in
cultured murine cortical neurons. J. Cell Physiol., 211(1):63-87.
Konishi, Y., Lehtinen, M., Donovan, N. and Bonni, A. (2002). Cdc2
Phosphorylation of BAD Links the Cell Cycle to the Cell Death Machinery.
Molecular Cell, 9:1005-1016.
Korrapati, S., Roux, I., Glowatzki, E. and Doetzlhofer, A. (2013).
Notch signaling limits supporting cell plasticity in the hair cell-
damaged early postnatal murine cochlea. PLoS One, 8(8):e73276.

150
Kranenburg, O., Scharnhorst, V., Van der Eb, A.J. and Zantema, A.
(1995). Inhibition of Cyclin-dependent Kinase activity triggers neuronal
differentiation of mouse neuroblastoma cells. Journal of Cell Biology,
131(1): 227-234.
Kruman, I.I. (2004). Why do neurons enter the cell cycle? Cell cycle, 3(6):
769-733.
Kucharczak, J., Simmons, M.J., Fan, Y. and Gélinas, C. (2003). To be,
or not to be: NF-kappaB is the answer--role of Rel/NF-kappaB in the
regulation of apoptosis. Oncogene, 22(56):8961-8982.
Laine, H., Doetzlhofer, A., Mantela, J., Ylikoski, J., Laiho, M., Roussel,
M.F., Segil, N. and Pirvola, U. (2007). p19Ink4d and p21Cip1 collaborate
to maintain the postmitotic state of auditory hair cells, their codeletion
leading to DNA damage and p53-mediated apoptosis. The Journal of
Neuroscience, 27(6):1434-1444.
Lempiainen, H. and Shore, D. (2009). Growth control and ribosome
biogenesis. Current Opinion in Cell Biology, 21:855–863.
Link, C.D., Fonte, V., Hiester, B., Yerg, J., Ferguson, J., Csontos,
S., Silverman, M.A. and Stein, G.H. (2006). Conversion of green
fluorescent protein into a toxic, aggregation-prone protein by C-terminal
addition of a short peptide. J. Biol. Chem., 281(3):1808-1816.
Liwak, U., Faye, M.D. and Holcik M. (2012). Translation control in
apoptosis. Exp. Oncol., 34(3): 218–230.
Lopez-Hernandez, F.J. (2011). New insights into the mechanism of
aminoglycoside nephrotoxicity: an integrative point of view. Kidney
International, 79: 33–45.
Lopez-Novoa, J.M., Quiros, Y., Vicente, L., Morales, A.I. and
Lumpkin, E.A., Collisson, T., Parab, P., Omer-Abdalla, A., Haeberle, H.,
Chen, P., Doetzlhofer, A., White, P., Groves, A., Segil, N. and Johnson,
J.E. (2003). Math1-driven GFP expression in the developing nervous system
of transgenic mice. Gene Expression Pattern, 3: 389-395.

151
Malumbres, M., Harlow, E., Hunt, T., Hunter, T., Lahti, J.M., Manning,
G., Morgan, D.O., Tsai, L.H. and Wolgemuth, D.J. (2009). Cyclin-
dependent kinases: a family portrait. Nature Cell Biology, 11(11):1275-1276.
Malumbres, M., Sotillo, R., Santamarı ́a, D., Galán, J., Cerezo, A.,
Ortega, S., Dubus, P. and Barbacid, M. (2004). Mammalian Cells Cycle
without the D-Type Cyclin-Dependent Kinases Cdk4 and Cdk6. Cell,
118(4):493-504.
Mardis, E.R. (2008). Next-Generation DNA Sequencing Methods. Annual
Review of Genomics and Human Genetics, 9: 387-402
Matsui, J.I., Ogilvie, J.M. and Warchol, M.E. (2002). Inhibition of
Caspases prevents ototoxic and ongoing hair cell death. The Journal of
Neuroscience, 22(4): 1218-1227.
Mauck, J.C. (1977). Solubilized DNA-dependent RNA polymerase
activities in resting and growing fibroblast. Biochemistry, 16(4):793-797.
Mingeot-Leclercq, M. and Tulkens, P.M. (1999). Aminoglycosides:
Nephrotoxicity. Antimicrobial Agents and Chemotherapy, 43(5): 1003-1012.
Morgan, D.O. (1997). Cyclin-Dependent Kinase: Engines, Clocks, and
Microprocessors. Annual Review of Cell and Developmental
Biology, 13:261-291
Nagy, I., Bodmer, M., Brors, D. and Bodmer, D. (2004). Early gene
expression in the organ of Corti exposed to Gentamicin. Hearing Research,
195:1-8.
Nagy, I., Monge, A., Albinger-Hegyi, A., Schmid, S. and Bodmer, D.
(2005). NF- κB is Required for Survival of Immature Auditory Hair Cell In
Vitro. Journal of the Association for Research in Otolaryngology, 6: 260-
268.
Nakashima, T., Teranishi, M., Hibi, T., Kobayashi, M. and Umemura,
M. (2000). Vestibular and cochlear toxicity of aminoglycosides--a review.
Acta. Otolaryngol., 120(8): 904-11

152
Nguyen1, M.D., Mushynski, W.E. and Julien, J.P. (2002). Cycling at the
interface between neurodevelopment and neurodegeneration. Cell Death &
Differentiation, 9: 1294-1306.
Nouspikel, T. and Hanawalt, P.C. (2002). DNA repair in terminally
differentiated cells. DNA Repair, 1(1): 59-75.
Oberle, C., and Blattner, C. (2010). Regulation of DNA Damage Response
to DSBs by Post-Translational Modifications. Current Genomics, 11(3):
184-198.
Omoigui, S. (2007). The Interleukin-6 inflammation pathway from
cholesterol to aging – Role of statins, bisphosphonates and plant polyphenols
in aging and age-related diseases. Immunity & Ageing, 4:1.
Ozaki, N., Matheis, K.A., Gamber, M., Feidl, T., Nolte, T., Kalkuhl, A.
and Deschl, U. (2010). Identification of genes involved in Gentamicin-
induced nephrotoxicity in rats – A toxicogenomic investigation.
Experimental and Toxicologic Pathology, 62: 555–566.
Parsons, T.D., Obaid, A.L. and Salzberg, B.M. (1992). Aminoglycoside
antibiotics block voltage-dependent calcium channels in intact vertebrate
nerve terminals. The Journal of General Physiology, 99(4): 491-504.
Pirvola, U., Ling, X.Q., Virkkala, J., Saarma, M., Murakata, C.,
Camoratto, A.M., Walton, K.M. and Ylikoski, J. (2000). Rescue of
hearing, auditory hair cells, and neurons by CEP-1347/KT7515, an inhibitor
of c-Jun N-Terminal Kinase activation. The Journal of Neuroscience, 20(1):
43-50.
Prezant, T.R., Agapian, J.V., Bohlman, M.C., Bu, X., Oztas, S., Qiu,
W.Q., Arnos, K.A., Cortopassi, G.A., Jaber, L. and Rotter, J.I. (1993).
Mitochondrial ribosomal RNA mutation associated with both
antibioticinduced and non-syndromic deafness. Nature Genetics, 4:289–294.
Priuska, E.M. and Schacht, J. (1995). Formation of free radicals by
Gentamicin and iron and evidence for an iron/Gentamicin complex.
Biochemical Pharmacology, 50:1749–1752.

153
Quick, C.A, Fish, A. and Brown, C. (1973). The relationship between
cochlea and kidney. The Laryngoscope, 83(9): 1469-1482.

Rao, K.S., Annapurna, V.V. and Raji, N.S. (2001). DNA Polymerase- β
may be the main player for defective DNA repair in aging rat neurons.
Annals of the New York Academy of Science, 928: 113-120.
Renner, F. and Schmitz L.M. (2009). Autoregulatory feedback loops
terminating the NF- κB response. Trends in Biological Sciences, 34(3): 128-
135.
Richardson, G.P., Forge, A., Kros, C.J. Fleming, J., Brown, S.M. and
Steel, K.P. (1997). Myosin VIIA is required for aminoglycoside
accumulation in cochlear hair cells. The Journal of Neuroscience, 17(24):
9506-9519.
Rideout, H.J., Wang, Q., Park, D.S. and Stefanis, L. (2003). Cyclin
dependent Kinase activity is required for apoptotic death but not inclusion
formation in cortical neurons after proteasomal inhibition. The Journal of
Neuroscience, 23(4): 1237-1245.
Ruben, R.J. (1967). Development of the inner ear of the mouse: a
radioautographic study of terminal mitoses. Acta Oto-Laryngologica,
supplementum 220: 1-44.
Rybak, L.P. (1999). Ototoxicity and antineoplastic drugs. Current Opinion
in Otolaryngology & Head and Neck Surgery, 7(5): 239-243.
Rybak, L.P., Husain, K., Morris, C., Whitworth, C. and Somani, S.
(2000). Effects of protective agents against cisplatin ototoxicity. The
American Journal of Otology, 21: 513-520.
Saelens, X., Kalai, M. and Vandenabeele, P. (2001). Translation inhibition
in apoptosis: caspase-dependent PKR activation and eIF2-alpha
phosphorylation. J. Biol. Chem., 276(45):41620-41628.

154
Sandoval, R., Leiser, J. and Molitoris, B.A. (1998).
Aminoglycoside antibiotics traffic to the Golgi complex in LLC-PK1 cells. J.
Am. Soc. Nephrol., 9(2):167-74.
Sawada, I., Kitahara, M. and Yazawa, Y. (1994). Swimming test for
evaluating vestibular function in guinea pigs. Acta Otolaryngol
Suppl., 510:20-23.
Scheunemann, A.E., Graham, W.D., Vendeix, F.A.P. and Agris, P.F.
(2010). Binding of aminoglycoside antibiotics to helix 69 of 23S rRNA.
Nucleic Acids Research, 38(9):3094-3105.
Scott, P.H., Cairns, C.A., Sutcliffe, J.E., Alzuherri, H.M., McLees, A.,
Winter, A.G. and White, R.J. (2001). Regulation of RNA polymerase III
transcription during cell cycle entry. The Journal of Biological Chemistry,
276(2): 1005-1014.
Shemiakina, I.I., Ermakova, G.V., Cranfill, P.J., Baird, M.A., Evans,
R.A., Souslova, E.A., Staroverov, D.B., Gorokhovatsky, A.Y., Putintseva,
E.V., Gorodnicheva, T.V., Chepurnykh, T.V., Strukova, L., Lukyanov,
S., Zaraisky, A.G., Davidson, M.W., Chudakov, D.M. and Shcherbo, D.
(2012). A monomeric red fluorescent protein with low cytotoxicity. Nat.
Commun., 3:1204.
Shulman, E., Belakhov, V., Wei, G., Kendall, A., Meyron-Holtz, E.G.,
Ben-Shachar, D., Schacht, J. and Baasov, T. (2014). Designer
Aminoglycosides that selectively inhibite cytoplasmic rather than
mitochondrial ribosomes show decreased ototoxicity: a strategy for the
treatment of genetic diseases. The Journal of Biological Chemistry,
288(4):2318-2330.
Sobkowicz, H.M., Loftus, J.M. and Slapnick, S.M. (1993).
Tissue culture of the organ of Corti. Acta. Otolaryngol. Suppl., 502:3-36
Sonenberg, N. and Hinnebusch, A.G. (2009). Regulation of translation
initiation in eukaryotes: mechanisms and biological targets. Cell, 136(4):
731-745.

155
Stewart, Z.A., and Pietenpol, J.A. (2001). P53 signaling and cell cycle
checkpoints. Chemical Research in Toxicology, 14(3): 243-263.
Sugahar, K., Rubel, E.W. and Cunningham, L.L. (2006). JNK signaling
in neomycin-induced vestibular hair cell death. Hearing Research, 221: 128-
135.
Swan, S.K. (1997). Aminoglycoside nephrotoxicity. Seminars in
Nephrology, 17:27–33.
Taleb, M., Brandon, C.S., Lee, F.S., Lomax, M.I., Dillmann, W.H.
and Cunningham. L.L. (2008). Hsp70 inhibits aminoglycoside-induced
hair cell death and is necessary for the protective effect of heat shock. J.
Assoc. Res. Otolaryngol., 9(3):277-289.
Taylor, R.W. and Turnbull, D.M. (2005). Mitochondrial DNA Mutations
in Human Disease. Nature Reviews Genetics. 6(5): 389–402.
Ten, R.M., Paya, C.V., Israël, N., Le Bail, O., Mattei, M.G., Virelizier,
J.L., Kourilsky, P. and Israël, A. (1992). The characterization of the
promoter of the gene encoding the p50 subunit of NF-kappa B indicates that
it participates in its own regulation. EMBO J., 11(1):195-203.
Tiede, L., Steyger, P.S., Nichols, M.G. and Hallworth, R. (2009).
Metabolic Imaging of the Organ of Corti – A Window on Cochlea
Bioenergetics. Brain Research, 1277: 37–41.
Tsutsui, T., Hesabi, B., Moons, D.S., Pandolfi, P.P., Hansel, K.S., Koff,
A. and Kiyokawa, H. (1999). Targeted disruption of CDK4
delays cell cycle entry with enhanced p27(Kip1) activity. Mol. Cell
Biol., 19(10):7011-7019.
Vanden Bush, T.J. and Bishop, G.A. (2011). CDK-Mediated Regulation of
Cell Functions via c-Jun Phosphorylation and AP-1 Activation. PLoS ONE,
6(4): e19468.
van Ruijven, M.W., deGroot, J.C., Hendriksen and F., Smoorenburg,
G.F. (2005). Immunohistochemical detection of platinated DNA in the
cochlea of cisplatin-treated guinea pigs. Hearing Research, 203: 112-121.

156
Veselý, j., Havli ček, L., Strnad, M., Blow, J.J., Donella-Deana, A., Pinna,
L., Letham, D.S., Kato, J., Detivaud, L., Leclerc, S. and Meijer, L.
(2005). Inhibition of Cyclin-Dependent Kinases by Purine Analogues.
European Journal of Biochemistry, 224(2):771-786.
Wang, D., and Lippard, S.J. (2005). Cellular processing of platinum
anticancer drugs. Nature Review, 4: 307-320.
Wang, Q. and Steyger, P.S. (2009). Trafficking of Systemic Fluorescent
Gentamicin into the Cochlea and Hair Cells. Journal of the Association for
Research Otolaryngology, 10:205-219.
Wang, X.W. and Seed, B. (2003). A PCR primer bank for quantitative gene
expression analysis. Nucleic Acids Research, 31(24): e154.
White, P.M., Doetzlhofer, A., Lee, Y.S., Groves, A.K. and Segil, N.
(2006). Mammalian cochlear supporting cells can divide and trans-
differentiate into hair cells. Nature, 441(7096):984-987.
Wiita, A.P., Ziv, E., Wiita, P.J., Urisman, A., Julien, O., Burlingame,
A.L., Weissman, J.S. and Wells, J.A. (2013). Global cellular response to
chemotherapy-induced apoptosis. Elife, 2:e01236.
Willott, J.F. (2001). Cochlea hair cell density and inner-ear staining
techniques. Handbook of Mouse Auditory Research: From Behavior to
Molecular Biology. CRC Press LLC.
Wu, W.J., Sha, S.H., McLaren, J.D., Kawamoto, K., Raphael, Y. and
Schacht, J. (2001). Aminoglycoside ototoxicity in adult CBA, C57BL and
BALB mice and the Sprague –Dawley rat. Hearing Research, 158: 165-178.
Yamagishi, M., Ishizuka, Y., Fujiwara, M., Nakamura, H., Igarashi,
S., Nakano, Y. and Kuwano, R. (1993). Distribution of Calcium Binding
Proteins in Sensory Organs of the Ear, Nose and Throat. Acta Oto-
laryngologica, 113(506): 85-89.

157
Yang, L.E., Maunsbach, A.B., Leong, P.K. and McDonough, A.A. (2005).
Redistribution of myosin VI from top to base of proximal tubule microvilli
during acute hypertension. Journal of American Society of Nephrology,
16(10):2890-2896.
Yang, Y., Mufson, E.J. and Herrup, K. (2003). Neuronal cell death is
preceded by cell cycle events at all stages of Alzheimer's disease. J.
Neurosci., 23(7):2557-2563.
Ylikoski, J., Liang, X.Q., Virkkala, J. and Pirvola, U. (2002). Blockade
of c-Jun N-terminal kinase pathway attenuates Gentamicin-induced cochlear
and vestibular hair cell death. Hearing Research, 166:33-43.
Yuan, Z., Gong, S., Luo, J., Zheng, Z., Song, B., Ma, S., Guo, J., Hu, C.,
Thiel, G., Vinson, C., Hu, C., Wang, Y. and Li, M. (2009). Opposing
Roles for ATF2 and c-Fos in c-Jun-Mediated Neuronal Apoptosis.
Molecular and Cellular Biology, 29(9): 2431-2442.
Zhou, J., Lin, J., Zhou, C., Deng, X. and Xia, B. (2011). Cytotoxicity of
red fluorescent protein DsRed is associated with the suppression of Bcl-xL
translation. FEBS Letter, 585: 821-827.

Abstract (if available)

Abstract Sensory hair cells are essential for transforming the mechanical vibrations of sound into electric signals that our nervous system can interpret. However, sensory hair cells are sensitive to a variety of stresses, including aminoglycoside antibiotics, chemotherapy agents, and environmental noise. The molecular mechanisms of ototoxicity have been under investigation for several decades, yet little is known about the underlying signaling pathways. In Chapter 2 and 3 of this study, we adapted an established technique of organotypic culture of perinatal mouse cochlea to investigate the ototoxicity. With this culture model, we found that CDK2 activity was involved in aminoglycoside‐induced hair cell death and that one consequence of its activity was to mediate the transcriptional activity of c-Jun. The involvement of CDK2 was further supported by evidence from utricular cultures and in vivo experiments. Using knockout mice, we demonstrated that p19ⁱⁿᵏ⁴ᵈ, p21ᵏⁱᵖ¹ and p53 were not involved in aminoglycoside ototoxicity. In Chapter 6, we profiled the gene expression changes in hair cells after aminoglycoside treatment and found altered expression for genes involved in the general transcription apparatus, translation machinery, mitochondria, DNA damage and repair, stress and apoptosis, JNK pathway and NF-κB pathway. Finally, we found that the sensitivity to ototoxins was changed in mCherry-H2B positive hair cells. Our results create a framework for understanding the molecular mechanisms of ototoxicity and the signaling pathways underlying postmitotic hair cell apoptosis.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The mechanisms of somatic cell reprogramming

PDF

The role of HES/HEY transcriptional repressors in specification and maintenance of cell fate in the mouse organ of Corti

PDF

Induction of hypersignaling as a therapeutic approach for treatment of BCR-ABL1 positive Acute Lymphoblastic Leukemia (ALL) cells

PDF

Development and evolution of skeletal joints: lessons learned from studying zebrafish

PDF

The molecular mechanism underlying the autoimmune-associated PTPN22 R620W variation and the quest for therapeutics

PDF

SUMOylation regulates RNA polymerase III -- dependent transcripton via MAF1

PDF

Maf1 is a novel target of the tumor suppressor PTEN and a negative regulator of lipid metabolism

PDF

Variation in Galr1 expression determines the susceptibility to excitotoxin-induced neuronal death in mice

PDF

Investigating the role of SASH1 gene located on chromosome 6 in ovarian cancer

PDF

Neuroendocrine regulation of the transcription factor SKN-1/Nrf2 in oxidative stress response

PDF

Identify Werner protein molecular partners in S phase alt cell

PDF

RUNX2 & sex steroids: molecular mechanisms in regulating bone turnover

PDF

Tissue interactions & molecular pathways in specification of the ectomesenchyme from cranial neural crest

PDF

Characterization of three novel variants of the MAVS adaptor

PDF

Molecular strategies towards inhibition of KSHV-associated malignancies

PDF

The role of endoplasmic reticulum protein GRP78 in normal hematopoeises and PTEN-null leukemogenesis

PDF

The calcium-sensing receptor in the specification of normal and malignant hematopoietic cell localization in the bone marrow microenvironment

PDF

Kaposi's sarcoma associated herpes-virus induces cellular proxi expression to modulate host gene expression that benefits viral infection and oncogenesis

PDF

Differential role of two coactivators, CCAR1 and CARM1, for dysregulated beta-catenin activity in colorectal cancer cell growth and gene expression

PDF

The role of GRP78 in the regulation of apoptosis and prostate cancer progression

Asset Metadata

Creator Tao, Litao (author)

Core Title Investigation of the molecular mechanisms of ototoxicity

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Genetic, Molecular and Cellular Biology

Publication Date 07/01/2014

Defense Date 05/20/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag aminoglycoside antibiotics,c-Jun,cyclin-dependent kinase 2,OAI-PMH Harvest,ototoxicity,RNA sequencing

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Tower, John G. (committee chair), Comai, Lucio (committee member), Schauwecker, P. Elyse (committee member), Segil, Neil (committee member)

Creator Email litao4tao@gmail.com,litaotao@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-426847

Unique identifier UC11286582

Identifier etd-TaoLitao-2597.pdf (filename),usctheses-c3-426847 (legacy record id)

Legacy Identifier etd-TaoLitao-2597.pdf

Dmrecord 426847

Document Type Dissertation

Format application/pdf (imt)

Rights Tao, Litao

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA