University of Southern California Dissertations and Theses

Using novel small molecule modulators as a tool to elucidate the role of the Myocyte Enhancer Factor 2 (MEF2) family of transcription factors in leukemia

(USC Thesis Other)

Using novel small molecule modulators as a tool to elucidate the role of the Myocyte Enhancer Factor 2 (MEF2) family of transcription factors in leukemia

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

USING NOVEL SMALL MOLECULE
MODULATORS AS A TOOL TO ELUCIDATE
THE ROLE OF THE MYOCYTE ENHANCER
FACTOR 2 (MEF2) FAMILY OF
TRANSCRIPTION FACTORS IN LEUKEMIA

by

Michael A. Philips

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
(MOLECULAR BIOLOGY)

December 2013

Copyright 2013 Michael A. Philips
ii
EPIGRAPH
“I'm not saying I'm gonna change the world, but I guarantee that
I will spark the brain that will change the world.”
-Tupac Shakur
iii
DEDICATION
To my father, best friend, and hero Reed Philips (4/7/1949-3/30/2011)

iv
ACKNOWLEDGMENTS
First off, I need to thank my advisor Dr. Lin Chen. He took a chance on me
when I entered the graduate program here at USC. He only knew me as a baseball
player, not a scientist, and he took a large risk in admitting me to his lab. After I
joined the lab, Lin always had his door open, even when it was literally closed. He
was always available to talk, even when he was busy with other tasks. Lin had to
put forth a great effort into my maturation as a scientist, and I cannot express my
appreciation for his mentoring and leadership along this long journey.
I am also grateful for the input and discussions with my committee members
from both my oral qualifying exam and my dissertation. I appreciate the time and
effort put in by Drs. John Tower, Baruch Frenkel, Steven Finkel, and Xiaojiang Chen
to help guide me to my final goal over the past six years.
Thank you to everyone in the Lin Chen lab, both past and present. You have
all taught me so much about science, and myself. Without all of you, I never would
have made it out of my first year. Every question I had was answered without
hesitance, and every request I made was honored. I cannot imagine having a
graduate experience in any other lab, because the interactions with all of you over
the past six years have made me the scientist I am today. Dr. Yongqing Wu
mentored me during my rotation, and if not for him, I do not know if I would have
been interested in structural biology. He taught me everything from biology, to
chemistry, and even some Chinese, and for that I am grateful. Drs. Aidong Han,
Cosma Delisanti, and Raja Dey were only around the lab for a short time with me,
v
but their views on research helped shape my hands and mind early in graduate
school.
Former graduate students and friends Drs. Nimanthi Jayathilaka and Reza
Kalhor cannot be thanked enough for their endless support, both emotionally and
scientifically, throughout my time at USC. These are two of the brightest scientists I
have had the pleasure to meet, and I am so honored to call them my colleagues and
friends.
Former technician Katie Daugherty, current technician Melissa Hansen, and
current graduate student Kaori Noridomi have added to a great lab environment
over the years. These three ladies have been friends to me, and I cannot understate
how important that is when you come to work in the same room day after day, year
after year, especially when things just are not going as planned. Thank you for being
there for me.
My sincerest of thanks to everyone in the Molecular and Computational
Biology Department, as well as many other departments in the sciences at USC. If I
ever needed to use equipment or training in a technique that was foreign to my lab,
everyone was always happy and willing to help. The ladies in the front office,
starting with Christina, Linda, Eleni, Laura, Haley, and Cathy, were always so helpful
while I was trying to deal with the heaps of obligatory paperwork over the years. I
believe that our department acts as a family because we really are in this together.
Finally, my friends and family have been my rock. These past six years have
seen many ups and downs, and I know the highs would not have been as high
without you and the lows would have been downright unbearable. My friends,
vi
those from high school, undergrad, grad school, and from the area in which I live
now, have been nothing short of amazing. They always offered me a means to cope
with failure, disappointment, and frustration relating to graduate school and life in
general. Without my friends down here in the Los Angeles area, I honestly do not
think I would have made it to the point where I am today. They are the best friends
anyone could ask for, and I am so happy that they are mine. My family has been my
biggest cheerleaders during this entire experience. My mom, Susie, sister, Melissa,
and nephew, Trevor, always pushed me to succeed and never let me give up. There
were many times along my journey where I questioned my choice to attend
graduate school, but that thought never occurred to my family, and their confidence
never wavered.
To my dad, Reed, you gave me all the tools necessary to be successful in life.
You were not able to watch me finish grad school, and will not be around for several
other of my life goals, but everything that I have done or will do would never have
been possible without you. You were my first coach, smartest teacher, and best
friend and role model I have ever had. You are everything I want to be as a man, and
I miss you every day. I love you so much, and I dedicate this to you.

vii
TABLE OF CONTENTS
Epigraph ii
Dedication iii
Acknowledgments iv
List of Tables viii
List of Figures viii
Abstract x
1 Introduction
1.1 Myocyte Enhancer Factor 2 (MEF2) family of proteins 1
1.2 Epigenetic Regulation 3
1.3 Histone deacteylase (HDAC) family of proteins 5
1.4 Histone deacetylase inhibitors 9
1.5 Class IIa HDAC:MEF2 interaction 13

2 MEF2 Structural studies and their insights into cofactor interactions
2.1 Introduction 17
2.2 Early structural studies of MEF2 bound to DNA and co-factors 19
2.3 Results 27
2.4 Discussion 37
2.5 Materials and Methods 44

3 Anti-leukemic activity of MEF2 modulators (MEF2m) resulting from
the inhibition of the MEF2:class IIa HDAC interaction
3.1 Introduction 46
3.2 Results 63
3.3 Discussion 89
3.4 Materials and Methods 94

4 Genome-wide association studies on the effect of MEF2m in B cells
4.1 Introduction 99
4.2 Results 107
4.3 Discussion 119
4.4 Materials and Methods 123

5 Conclusion 124

Bibliography 129
viii
LIST OF TABLES

Table 2.1: Statistics of crystallographic analysis 31

LIST OF FIGURES

1.1 Sequence conservation of MEF2 2
1.2 MEF2 as a central regulator of differentiation and signal 4
responsiveness
1.3 Panel of human HDAC classes 6
1.4 HDAC inhibitors promote the acetylation of histones and non- 8
histone proteins by inhibiting the activity of HDAC enzymes
1.5 Structure of the Cabin1/MEF2B/DNA complex 10
1.6 The overall structure of the Cabin1-binding site of MEF2B 12
1.7 Structural and biochemical analyses of the HDAC9/MEF2/DNA 14
complex
1.8 Structural comparisons of different MEF2/co-repressor complexes 16

2.1 Properties of MEF2 proteins 18
2.2 The structure of MEF2A (2-78) bound to DNA 20
2.3 The MEF2A and SRF core/DNA complexes 22
2.4 Mechanisms of ligand binding to the hydrophobic groove of MEF2 24
2.5 Structure-based sequence alignments 26
2.6 Overall structure and packing of MEF2 dimer bound to DNA 28
2.7 Comparison with the previous crystal structures 33
2.8 DNA binding by the intact MADS-box/MEF2 domain 36
2.9 Cofactor binding site 38
2.10 Structure of the p300 TAZ2 domain bound to MEF2 on DNA 41
2.11 Protein –protein interactions at Interface I 43

3.1 MEF2D/DAZAP1 is a more potent transcriptional activator 48
than MEF2D
3.2 The self-renewal-associated signature is found in human MLL- 50
rearranged AML and activated as a hierarchy of gene expression
3.3 Ectopic expression of Mef2c into Irf8
− /−
BM progenitors induces 52
acute myelomonocytic leukemia in recipient mice
3.4 Mapping of 565 unique retroviral insertion sites onto the mouse 54
genome assembly from Celera Genomics
3.5 Structural characterization of the binding of BML-210 to MEF2A 57
3.6 Development of BML-210 analogs 59
3.7 Luciferase assay using GFP-HDAC4 (3-209) 61
3.8 Anti-cancer effects of the benzamide derivative NKL-30 in 63
leukemic and non-leukemic B cells
ix
3.9 Anti-cancer effects of benzamide derivatives on leukemia 67
3.10 Combination treatment with MEF2m and Imatinib Mesylate 69
3.11 HDAC activity of common HDAC inhibitors and benzamide 71
derivatives
3.12 Treatment of a control B-cell line with HDAC inhibitors 74
3.13 NKL-30 dose-dependently induces protein expression in 76
MEF2 target genes in a leukemia cell line
3.14 NKL-30 treatment results in accumulation of phosphorylated 79
HDAC4 in the cytoplasm of leukemia cells
3.15 Nalm6 leukemia mouse model and in vivo study protocol 82
3.16 NKL-30 treatment alone does not affect mouse body weight 82
3.17 In vivo efficacy of NKL-30 in Nalm6 leukemia mouse model 83
3.18 MEF2 is found outside of the nucleus in HL-60 leukemia cells 85
3.19 NCI60 mRNA expression data lends support to an HDAC3- 87
independent mechanism of action for benzamide derivatives
3.20 Proposed model of the mechanism of MEF2m NKL-30 91

4.1 An overview of the Tuxedo protocol 101
4.2 Metabolic analysis of up-regulated genes in GM12878 cells 108
treated with NKL-30 2 μM for 24 hours
4.3 Metabolic analysis of up-regulated genes in Nalm6 leukemia 111
cells treated with NKL-30 2 μM for 24 hours
4.4 Ikaros interaction with down-regulated genes 113
4.5 Metabolic analysis of genes with lower expression in Nalm6 115
leukemia cells compared with healthy GM12878 B cells
4.6 Network 1 of genes with higher expression in Nalm6 leukemia 117
cells compared with healthy GM12878 B cells
4.7 Network 2 of genes with higher expression in Nalm6 leukemia 118
cells compared with healthy GM12878 B cells
4.8 Cross-talk of networks in a comparison of GM12878 and 120
Nalm6 gene expression

x

Abstract

The Myocyte Enhancer Factor 2 (MEF2) family of transcription factors are
DNA-bound proteins that regulate gene expression via their interaction with co-
regulatory proteins. This interaction between MEF2 and its co-factors is necessary
for the development and proper function of many mammalian cell types, and is a
major factor in the onset and progression of many cancers, specifically leukemia.
Understanding the MEF2:co-factor interaction in leukemia, which is often times an
epigenetic event, is of utmost importance to those searching for therapies that have
high specificity and low chance of side effects.
In the first part of this paper, the structural characteristics of the MEF2
family that allow its interaction with co-regulators is examined. The study revealed
that the structure of MEF2 proteins are not majorly altered by the binding of a co-
regulator, such as class IIa histone deacetylases (HDACs), Cabin1, or p300, at the
protein-protein interface, and are inherently able to be bound while interacting with
DNA. This overruled previous notions that the hydrophobic groove, or “ b in d in g p ock et ” , of M E F 2 w as n ot p r ese nt unless i t was in te r act in g w it h a co -regulator.
xi
The second, and most involved, study of this paper examines the interaction
of MEF2 and class IIa HDACs in leukemia. This study was aided by the availability of
a small molecule benzamide compound NKL-30, which we call a MEF2 modulator
(MEF2m), that was used as a tool to observe the consequences of inhibiting the
protein-protein interaction both in vitro and in vivo. Our MEF2m were able to
preferentially decrease the viability of leukemia cells over non-cancerous control B
cells. We were also able to show that our compound was able to ablate the
MEF2:class IIa HDAC interaction and result in the cytoplasmic shuttling of these co-
repressors. Additionally, we were able to show that the expression of the protein
NR4A1/Nur77, a MEF2 target gene associated with apoptosis, increased after
treatment with NKL-30, likely because of the derepression of MEF2 afforded by the
exodus of class IIa HDACs from the nucleus. Finally, we were able to show in a
mammalian leukemia model utilizing Nalm6 cells, that mice treated with NKL-30
enjoy an increased lifespan compared to those receiving vehicle control injections.
The third and final part of this work involves the gene expression analysis of
B cells that have been treated with our MEF2m. Both healthy and leukemic B cells
were treated with our small molecule compound or a solvent control and then data
was collected and analyzed using mRNA-seq. We have shown that our MEF2m often
preferentially activates gene expression, possibly due to the derepression seen on
MEF2 and additional off-target effects. While these differentially regulated genes
seen after drug treatment do not fit a single profile or pathway, we do know that the
treatment of leukemia cells with our MEF2m usually results in death. Therefore,
this data is crucial in our quest for understanding this phenomenon. This work is
xii
still ongoing and can be massaged for more data and relationships with the help of
future replication and analyses.
In summation, the work presented in this dissertation addresses the
possibility of a small molecule compound to specifically target leukemia cells for
death over healthy cells due to its ability to bind to MEF2 proteins, disallowing the
binding of co-repressors, and resulting in an epigenetic alteration event. The
mechanism and consequence of this event, while not yet completely understood, is
detailed within the following chapters and the data suggest that we are on the right
track toward understanding a way to develop targeted therapies for previously
untreatable or high mortality-rate diseases.

1

Chapter 1
Introduction

1.1 Myocyte Enhancer Factor 2 (MEF2) family of proteins

The MEF2 family of mammalian transcription factors is comprised of four
proteins (MEF2A, B, C, and D) that collectively regulate specific gene expression in
diverse developmental programs and adaptive responses (Potthoff and Olson,
2007). In vertebrates, MEF2 is involved in the control of differentiation,
proliferation, and survival/apoptosis of a wide range of cell types including muscle,
lymphocytes and neurons (Potthoff and Olson, 2007; Youn and Liu, 2000; Pan et al.,
2004; Mao et al., 1999). The four vertebrate family members share a highly
conserved N-terminal region (residues 1-93) containing the MADS-box, which is the
primary DNA-binding domain, immediately followed by the MEF2-specific domain
(Shore and Sharrocks, 1995). Transcriptional activation domains are contained in
the divergent C-terminal regions (Figure 1.1). MEF2 was first discovered as a
2

Figure. 1.1 Sequence conservation of MEF2. The percentage amino acid identity
within the MADS, MEF2 and transcriptional activation domains of different MEF2
proteins from various organisms relative to human (h) MEF2A. N-termini are to the
left (Potthoff and Olson, 2007).

3
regulator of gene expression in vertebrate muscle cells, however, a single Mef2 gene
can also be found in other organisms such as Saccharomyces cerevisiae, Drosophilia,
and Caenorhabitis elegans (Gossett et al., 1989; Dodou and Treisman, 1997; Bour et
al., 1995; Sandmann et al., 2006). In adult tissues, these transcription factors also
serve as crucial regulators of stress responses and adaptive programs in response to
environmental signals (Kim et al., 2008; Shalizi et al. 2006; Yang et al. 2009).
MEF2 binds to DNA as either a homo- or hetero-dimer and interacts with
transcription co-factors through the highly conserved N-terminal MADS-box/MEF2
domain

(Wu et al., 2010). MEF2 can then be modulated through post-translational
modifications like acetylation, phosphorylation, and sumoylation or through the
binding of its co-activators such as CBP/p300 and myocardin, or co-repressors such
as Cabin1 and HDACs (Kim et al., 2008; Shalizi et al., 2006; Yang et al., 1998; Youn et
al., 1999; Youn et al., 2000; Wei et al., 2008 Bertos et al., 2001; Gregoire et al., 2006;
McKinsey et al., 2001; De Luca et al., 2003)(Figure 1.2). The type of target gene
activated or repressed by MEF2 is highly reliant upon the type of post-translational
modification of MEF2 and its interaction with various cofactors.

1.2 Epigenetic Regulation

Chromatin is a dynamic macromolecular complex of DNA and proteins that is
the basis for life itself. Epigenetic regulation of this complex by enzymes that act
upon proteins (histone methyltransferases [HMT], demethylases, histone

4

Figure. 1.2. MEF2 as a central regulator of differentiation and signal
responsiveness. MAP kinase signaling activates MEF2. Calcium-dependent signals
also activate MEF2 by stimulating calcium-dependent kinases that phosphorylate
class II HDACs, thereby promoting their dissociation from MEF2 and derepressing
MEF2 target genes. MEF2 recruits numerous co-factors to drive the differentiation
of the various cell types shown. Although MAPK and HDAC signaling pathways have
been implicated in the modulation of numerous MEF2-dependent developmental
programs, these signaling pathways have not yet been shown to operate in all the
cell types under MEF2 control (Potthoff and Olson, 2007).

5
acetyltransferases [HAT], and histone deacetylases [HDAC]), or those that act on
DNA (cytosine methyltransferases), enable the cell to make heritable changes in
gene expression without changing the DNA sequence itself. These changes allow the
cell to proceed through integral processes such as transcription, replication, and
repair (Kouzarides 2007).
The field of epigenetics got its first true start in the 1960s when a group of
scientists postulated that acetylation and methylation could play a role in regulating
RNA synthesis (Allfrey et al. 1964). Research focused on HATs and HDACs have
shown that the acetyl moiety added to lysine residues on histones by HAT neutralize
the positive charge of histones, weakening the interaction with the negatively
charged DNA. The result of this is a more relaxed and transcriptionally permissible
conformation. HDACs antagonize this state by removing the acetyl moieties, leading
to a stronger interaction of histone proteins with DNA and resulting in a more
transcriptionally repressed, condensed chromatin conformation (Ververis and
Karagiannis, 2012). In this work, however, I focus on the epigenetic regulation of
HDACs on non-histone proteins, specifically MEF2.

1.3 Histone Deacetylase (HDAC) family of proteins

The HDAC family consists of eighteen proteins that are responsible for
catalyzing the deacetylation of lysine residues on histone and non-histone proteins,
resulting in the regulation of cell proliferation, differentiation, and apoptosis. This
family is divided into four different classes based on their homology to yeast
6

Figure. 1.3. Panel of human HDAC classes. Structure, length and cellular
loc ali z at ion of HDAC en z y mes (D ell’ A v ersana 2 0 1 2 ).
7
orthologs Rpd3, Hda1, and Sir2: class I (HDAC 1, 2, 3, and 8), class II (HDAC 4, 5, 6, 7,
9, and 10), class III (Sirt1-Sirt7, the sirtuins), and class IV (HDAC 11) (Figure 1.3).
Classes I, II, and IV all share a common catalytic domain which uses Zn
+
as a
cofactor, whereas class III uses NAD
+
, and is therefore often excluded when
discussing classical HDAC proteins (de Ruijter et al., 2003).
Class II HDACs are unique in that they are able to translocate between the
nucleus and cytoplasm, the regulation of which is controlled by nuclear import and
export signals as well as phosphorylation-specific binding sites for 14-3-3 proteins.
Class II can further be divided into class IIa (HDAC 4, 5, 7, 9), and class IIb (HDAC 6,
10). Class IIa HDACs are highly expressed in muscle, heart, brain, and lymphocyte
cells (Fischel et al., 1999; Grozinger et al., 1999; Wang et al., 1999). These proteins
are regulated on many levels via transcriptional, translational, and post-
translational control such as selective proteolysis, sumoylation, ubiquitination, and
phosphorylation at multiple sites (Yang and Seto, 2008). Class IIa HDACs contain a
large, non-catalytic N-terminal region that has been shown to regulate interactions
with proteins such as MEF2 (Han et al., 2005). The N-terminal region of these class
IIa HDACs has a short amphipathic helix that binds to the highly hydrophobic MADS
box/MEF2 pocket in the MEF2 protein when bound to DNA. This binding, and
successive epigenetic modification of the MEF2 transcription factor can then
negatively affect the regulation of MEF2 target genes.

8

Figure. 1.4. HDAC inhibitors promote the acetylation of histones and non-
histone proteins by inhibiting the activity of HDAC enzymes. HDAC inhibitor-
mediated modification of histones and non-histone proteins (examples shown) can
result in increased or decreased gene expression, influencing other DNA-based
processes, including DNA replication and repair. As a result of these processes,
HDAC inhibitors are able to elicit a multitude of biological effects on cells, such as
apoptosis, cell-cycle arrest, and angiogenesis (Ververis et al., 2013).

9
1.4 Histone deacetylase inhibitors
Histone deacetylase inhibitors (HDACi) are a very popular, yet poorly
understood, epigenetic modulator. These chemicals are targeted to HDAC catalytic
domains and increase the acetylation level of many proteins throughout the cell.
These molecules have been shown to induce growth arrest, differentiation, and
apoptosis in transformed cells (Marks et al., 2000; Romanski et al., 2004; Sasaki et
al., 2008) (Figure 1.4). The anti-tumor and therapeutic effects in several types of
cancers, neurodegenerative diseases, and inflammation have all been documented,
however, there simply is not a single mechanism of action that can explain these
results (Bolden et al., 2006; Minucci et al., 2006; Chuang et al., 2009; Paris et al.,
2008).
The term HDAC inhibitor encompasses a wide range of compounds. These
compounds all result in HDAC activity inhibition and many common cellular
activities such as anti-proliferation and cell death, which could be due to the
common activity of HDAC inhibition (Humeniuk et al., 2009). However, distinct
classes of HDACi with highly divergent chemical structures show distinct cellular
activity that cannot be explained by the general inhibition of HDAC, or the results of
altered acetylation levels of histone or non-histone proteins after drug treatment
alone. To restate this sentiment, two distinct classes of HDACi may have a similar
effect in HDAC inhibition assays, but show distinct cellular activities (Chou et al.,
2008).

10

Figure 1.5. Structure of the Cabin1/MEF2B/DNA complex. Overall structure of
the Cabin1 (red), MEF2B (monomer A in green, monomer B in blue) and DNA
(magenta) complex (Han et al., 2003).

11
One explanation for this phenomenon is that certain HDACi target a specific
subclass of HDAC, either through differential thermodynamics or kinetics in
inhibiting different HDACi, but the detailed mechanisms of functional specificity of
different classes of HDACi are not well understood; there is no concrete evidence of
differential binding, and kinetics cannot explain the observed activity because the
drugs are always present in assays. Other possibilities include the notion that
certain compounds, while possessing the general and rather non-specific
background activity of inhibiting HDACs, also have much more specific targets that
explain, either alone or in combination with HDAC inhibition, their cellular
activities. We believe that the latter is most likely the case, since we have not yet
obtained a compound that is completely devoid of HDAC inhibition function.
Many of the leading HDACi, for example Trichostatin A (TSA) and
suberoylanilide hydroxamic acid (SAHA), target the zinc-containing catalytic
domain of class I, II, and IV HDACs, leaving much to be desired in the search for
small molecule drugs with HDAC class selectivity. This class selectivity is important
since the broad inhibition of HDACs, which are found in many cell types where they
deacetylate histone and non-histone proteins, can result in a myriad of gene
expression and protein interaction events that can lead the cell toward any number
of regulatory processes. This work includes data on novel compounds that
preferentially affect the epigenetic regulation of certain protein targets by class IIa
HDACs.

12

Figure 1.6. The overall structure of the Cabin1-binding site of MEF2B. A) The
Cabin1-binding pocket composed of β-strands S1, S2, S3 and α-helix H2 of each
monomer and the diagonal orientation of the Cabin1 helix in the pocket. B) A deep
groove of the Cabin1-binding site lined with hydrophobic residues (yellow patch)
(Han et al., 2003).
13
1.5 Class IIa HDAC:MEF2 interaction
MEF2 and its co-repressors take part in physical protein-protein interactions,
since many cannot bind DNA in a sequence-specific manner on their own. MEF2 is
the most characterized binding partner of the class IIa HDACs (Yang and Set, 2008).
Class IIa HDACs take part in a similar binding interaction with MEF2 as Cabin1, via
either co- r ep r essor ’s M E F 2 -binding motif and the MADS-box/MEF2 domain on the
MEF2 protein (Han et al., 2003) (Figures 1.5 and 1.6). This interaction involves an
α-helix on Cabin1 or HDAC that inserts into a hydrophobic pocket on the DNA-
bound MEF2 dimer (Han et al., 2005) (Figures 1.7, and 1.8). Once bound, HDACs can
deacetylate specific lysine residues on MEF2 (Gregoire et al., 2006).
The binding of class IIa HDACs to MEF2, regulated in a calcium-dependent
manner, represses the transcription of the coinciding MEF2 target gene. Once
activated by calcium, class IIa HDACs become phosphorylated and disassociate from
MEF2 where they are then bound by 14-3-3 proteins and actively shuttled to the
cytoplasm. This allows co-activators such as CBP/p300 to bind the DNA-bound
MEF2 dimer, resulting in transcriptional activation of the coinciding MEF2 target
gene (Youn et al., 2000). This cofactor:MEF2 binding, and therefore the
transcriptional repression and activation, is reversible and is an integral part of the
normal function of the cell.
The HDAC:MEF2 interaction is the topic of the greater part of this study. The
binding of class IIa HDACs to MEF2 resulting in repressed gene expression in
antagonism with the binding of cofactors that promote transcriptional activation
14

Figure 1.7. Structural and biochemical analyses of the HDAC9/MEF2/DNA
complex. A) Electron density of part of the HDAC9 peptide bound to MEF2. The
simulated- an ne ali ng o mit map is cont our ed at the 2. 0 σ l evel. B) O v er all stru ctur e of
the MEF2-binding motif of HDAC9 (gray) bound to the MEF2 dimer (green and light
green) on DNA (blue backbone trace). The secondary structures of one monomer
are labeled. The DNA sequence used in crystallization is listed below the Figure. C)
Detailed interactions at the HDAC9/MEF2 interface. Interacting residues (stick
model) from HDAC9 and MEF2 are colored according to their proteins (in Ribbons)
and labeled by black and blue fonts, respectively. Only one MEF2 monomer is
labeled. D) The long aliphatic side-chains of polar resides surrounding Leu147, such
as Lys144, Lys146 and Gln148, make extensive van der Waals contacts to Leu67 and
Thr70 of MEF2 (Han et al., 2005).
15
provides an important biological system that can be further studied in the cases of
disease. Since this is an epigenetic alteration event, many agree that it can be
targeted by small molecule inhibitors, however most research lies in HDACi that
target the enzymatic activity of the HDAC proteins. The manipulation of this
interaction in leukemia cells, with the help of small benzamide molecules as tools,
helped to elucidate the importance of this protein-protein interaction, as well as
provided insight into the mechanism of action of the benzamides, within the context
of cancerous lymphocytes.

16

Figure 1.8. Structural comparisons of different MEF2/co-repressor complexes.
Backbone superposition of the HDAC9/MEF2/DNA complex (MEF2 in green and
HDAC9 in gray) and the Cabin1/MEF2/DNA complex (MEF2 in cyan and Cabin1 in
magenta): the H2 –S3 loops of each monomer are labeled as loop1 and loop2,
respectively. The helical shift between HDAC9 and Cabin1 is evident in the figure
(Han et al., 2005).

17

Chapter 2
MEF2 structural studies and their
insights into cofactor interactions

2.1 Introduction
In 1989, a group of scientists at the University of Texas, M.D. Anderson
Cancer Center discovered what they termed “a new myocyte-specific enhancer-
binding factor that recognizes a conserved element associated with multiple muscle-
specific genes ” . This group grew skeletal myoblasts in a medium that lacked growth
factors. They found that these cells started to activate transcription of muscle-
specific genes, including the muscle creatine kinase (mck) gene. The transcriptional
activation was determined to be triggered by the binding of MEF2 to an A+T-rich
enhancer sequence 1,204 to 1,095 bp upstream of the mck gene and the expression
of MEF2 was associated with the activation of the differentiation program (Gossett
et al., 1989) (Figure 2.1 B). However, it was not until 2000 that the first x-ray

18

Figure 2.1. Properties of MEF2 proteins. (a) Sequence alignment of the DNA-
binding/dimerization domains of MEF2 proteins (four human members MEF2A-D;
Drosophila D-MEF2), human serum response factor (SRF) and yeast MCM1.
I nvar ia nt (∗ ) and hig hly conserv ed (+ ) amino a cid r esi d ues f or al l k now n M A DS -box
proteins and secondary structure elements for MEF2A are indicated (Kabsch and
Sander, 1983). (b) DNA site consensus sequences for MEF2A (MEF site) determined
by in vitro selection (adapted from Santelli and Richmond, 2000).

19
structure of MEF2 was solved, providing a wealth of insight as to how the protein
was folded in three-dimensional space and its interactions with DNA (Santelli and
Richmond, 2000). With an extremely high-resolution crystal structure solved at 1.5
Å, the first MEF2A-DNA complex has provided the framework for all subsequent
structural studies on the MEF2 family, including several from our own lab at USC
an d Dr . L in C hen’s p r eviou s lab at the U ni v ersit y of C olo r ad o, Bou l d er.

2.2 Early structural studies of MEF2 bound to DNA and
co-factors
MEF2 transcription factors belong to the MADS-box superfamily, named for
their similarity to MCM1, Agamous, Deficiens, and SRF proteins (Potthoff and Olson,
2007). This 58-amino acid-long MADS-box, followed immediately by the 28-amino
acid-long MEF2-specific domain in MEF2 family members is a conserved N-terminal
region showing more than 85% identity that is responsible for DNA binding,
dimerization, and protein-protein interactions (Gossett et al., 1989; Santelli and
Richmond, 2000; Han et al., 2003, Han et al., 2005; Potthoff and Olson, 2007; Wu et
al., 2010; He et al., 2011)(Figure 2.1 A). The MEF2 domain is not found in any other
MADS-box superfamily members other than the MEF2 family, hence its name. This
MEF2 domain allows either homo- or hetero-dimerization between MEF2 family
members, but not with any other MADS-box proteins (Molkentin et al., 1996;
Santelli and Richmond, 2000). MEF2 proteins also possess at least one activation

20

Figure 2.2. The structure of MEF2A (2-78) bound to DNA. The 2-fold axes of the
protein dimer and DNA (silver) are aligned in the crystals. The three-layer
organization of a protein monomer extends from the DNA as an N-extension
(brown, N- ext) an d an α - helix ( r ed , α I ) o f the a n ti p arallel coiled coil in t he “ b ot tom
lay er” , f oll ow ed b y t w o strand s (blue, β I - β I I ) o f the fou r - strand β -sheet making up
the “ mid d le layer ” . Th e t op an d b ott om lay ers co mprise t he M A DS - b ox . Th e α -helix
(y ellow , α I I ) o f the M E F d om ai n for ms the “ top lay er” w it h its counte r p art in the 2 -
fold related subunit (green) (Santelli and Richmond, 2000).

21
domain in their C-terminal regions, however these regions are divergent between
family members.
In 2000, Santelli and Richmond published the first crystal structure of a
MEF2 family member bound to DNA. This data revealed the variations on DNA
binding that occur between MEF2 family members and other MADS-box superfamily
members. MEF2A (2-78), which includes the entire MADS-box and a majority of the
MEF2 domain (also called the MEF or MEF2S domain), was shown to be folded in
the same three- lay er o r g an iz at ion ob serv ed f or S RF a nd M C M 1 (Fi g ur e 2.2) . T he α I helix and the N-terminal random coil (N-extension) are the major protein factors
r esp onsib le fo r b in d in g DN A . T his “ DNA conta cti ng l ayer ” in ad d ition to the middle
lay er, com p r ise d of two β - strand s (β I an d β I I ), ma k e up the M A DS -box. T he two α I helices in the protein dimer bind predominantly to the narrow minor groove on
DNA at the MEF2 site. T he α I I helix, ei g ht amin o acid s shy of a compl et e M E F2
d om ai n, mak es up the t op lay er and p ack s again st t he β -hairpin in a different
orientation as that of SRF, resulting from extra hydrophobic and hydrogen-bonding
interactions of one MEF2A monomer to the other, not seen in SRF proteins (Figure
2.3). Of particular note is the observation that the Lys154 found in SRF is replaced
by Glu14 in MEF2A. The exchange of oppositely charged side-chains was shown by
mutagenesis to contribute strongly to the lack of significant DNA bending for MEF2A
(Nurrish and Treisman, 1995; West et al., 1997). The differences in DNA binding
result in the overall angle bend of the DNA bound to SRF to be 72° compared to 17°
for the MEF2A-bound DNA.

22

Figure 2.3. The MEF2A and SRF core/DNA complexes. T he α I helices (r ed ) w ere
used f or the a lig nmen t of the complexes. Th e p ack in g of the β -sheet (green) on the
coiled coil is similar, although not identical, in the two MADS domains, whereas the
interaction of the N-extension (yellow) with DNA (blue-cyan) and the orientation of
the SRF C- te r mina l an d M E F d om ai ns (mag en ta ) o n t he β -sheet are substantially
different for the two complexes. The DNA is bent over an angle four times greater
for SRF than MEF2A (72° versus 17°)(Santelli and Richmond, 2000).

23
Previously, it was reported that the MEF2 domain is necessary and sufficient
to impart transcriptional silencing after interaction with co-repressors such as
histone deacetylase 4 and the alternatively spliced MITR from histone deacetylase 9
(M iska et al. , 1 9 9 9 ; S p ar r ow et al. , 1 9 9 9 ). The α I I helix w as show n b y S an te lli an d Richmond to contribute significantly to the dimerization interface of MEF2
primarily through hydrophobic c ont act s w it h t he α I I heli x and β -hairpin of the other
M E F 2 subunit . T he p ai r of α I I helices forms a pocket that allows the binding of
proteins such as MITR and homologous regions of HDAC2-4, amongst others, and
allows transactivation of MEF2. Without this dimerization, interactions with co-
factors and subsequent transcriptional regulation are impossible as shown through
mu ta ti ona l an aly ses i n w hich the VLL seq uence in the ce nt er o f α I I to A S R r emov es
the activation potential of MEF2C (Santelli and Richmond, 2000).
MEF2A recognizes a specific DNA binding site by selecting a narrow minor
groove, formed by the eight central base pairs of the MEF2 site, and by making
specific contacts with two base pairs per half-site. The minor groove is in close
contact with Gly2 and Arg3, and in each half site the side chain of Arg3 lies in the
minor groove with the opposite orientation as the homologous arginine for SRF
(Figure 2.3). Amino acid residues to either side of this conserved Arg3 are variable
in different MADS-box factors, resulting in differing structures of SRF, MCM1, and
MEF2A with respect to the conformation of the polypeptide chain in this segment of
the N-extension. The entire side chain of Arg3 of MEF2C and SRF, for example,
directly contacts the minor groove of DNA and affects the sequence of the central
A-T base-pairs above necessary for the α I helix , w her ea s A r g 3 o f M C M 1 does not.
24

Figure 2.4. Mechanisms of ligand binding to the hydrophobic groove of MEF2.
MEF2 is represented by the surface model. The consensus MEF2-binding sequence
is listed below the graph. Positions on the amphipathic helix poised to interact with
distinct hydrophobic pockets of MEF2 are indicated by ball and stick and are labeled
as A –E. The diagonally arranged hydrophobic pockets, A, B and C (labeled in red
font), interact with hydrophobic residues (colored in red) of the MEF2-binding
motifs. The other two pockets, D and E (labeled in blue font), interact with residues
of the MEF2-binding motif (colored in blue). The two residues sandwiching the
central leucine, indicated by letter Z, often engage in extensive van der Waals
interactions with helix H2 of MEF2. Note that the MEF2 groove has a dyad
symmetry. The helix can bind MEF2 in two possible orientations by switching
positions between A and C and D and E. The orientation of a co-regulator bound to
MEF2 may be specified by other interactions in a fully assembled promoter complex
(Han et al., 2005).
25
Following the publications of the first crystal structure (Santelli and
Richmond, 2000) and NMR solution structure (Huang et al., 2000) of MEF2A:DNA,
crystal structures of MEF2 family members bound with co-factors quickly appeared
in high impact journals. Complexes of Cabin1:MEF2B (Han et al., 2003) and
HDAC9:MEF2B (Han et al., 2005) solved by Dr. Aidong Han, a previous member of
my lab, were accepted by the scientific community based upon extremely accurate
and convincing X-ray diffraction data, analysis, and structural modeling. The overall
structure of these Co-factor:MEF2:DNA ternary complexes resembles a multi-
layered pyramid built on the DNA. Han et al. also found that the MEF2-binding
sur f ace of C ab in 1 has a “ p seud o - d y ad sy mm et r y ” that matches t he two -fold
symmetry of the MEF2 dimer, which agrees with the 1:2 binding ratio of Cabin1 to
MEF2 observed in solution. Also, by comparing this Cabin1:MEF2B:DNA structure
with the MEF2A:DNA structure solved previously, it was determined that the
binding of Cabin1 had little effect on the structure and function of the MADS box of
MEF2B.
A major improvement on the previously solved MEF2A:DNA solution and
crystal structures was the fully folded MEF2S domain of MEF2B in the
Cabin1:MEF2B:DNA crystal structure. The MEF2S structure implies that it stabilizes
dimerization by interacting with the MADS box of the reciprocal subunit in the
dimer, while also enhancing DNA-binding through direct stabilization of the DNA-
binding helix H1, as can be seen in figure 1.5. Importantly, the MEF2S domain
provides the major protein surface for other transcription factors interactions and a
stable docking site for Cabin1 and other transcriptional co-regulators.
26

Figure 2.5. Structure-based sequence alignments. Structure-based sequence
alignments of the MADS box of MEF2A –D, SRF and MCM1 (top panel), the MEF2S
domain of MEF2A –D (middle panel) and the MEF2-binding motif of Cabin1, HDAC4,
HDAC5, HDAC9 and p300 (bottom panel). Residues in MEF2 involved in DNA
binding (magenta), MADS-box –MEF2S interactions (blue) and Cabin1 binding (red)
are colored appropriately. MEF2-binding residues in Cabin1, HDAC4, HDAC5 and
HDAC9 are colored g r ee n. α - helices a nd β -strands are shown as bars and arrows,
respectively (Han et al., 2003).

27
In these two studies by Han et al., the general mechanism of recruitment of
co-repressors by MEF2 was revealed as follows: a short amphipathic helix from
class IIa HDACs or Cabin1 bind to a hydrophobic groove on the MEF2 dimer, also
r efer r ed to a s a “ b in d i ng pock et ” (Fi g ur e 2.4) . S y ste mati c m utat iona l an aly ses wer e
performed to pinpoint important residues on MEF2B (1-93), Cabin1, and HDAC4
(highly conserved MEF2-binding motif with respect to other class IIa HDACs and
Cabin1) proteins at their binding interface (Figure 2.5). Mutations of key residues
on Cabin1 at the Cabin1:MEF2 binding interface (Leu72 and Ile76) as well as
homologous residues on HDAC4 (Leu175, Val179, and Leu180) were able to
decrease or ablate the protein binding to MEF2 altogether (Figures 1.5 and 1.7).
Additionally, mutations on MEF2B on residues Tyr69 and Thr70 that make
extensive contacts with Cabin1 and HDAC9 as shown in their crystal structures,
disrupts the binding of HDAC4, lending evidence toward the theory that these
proteins bind MEF2 in a similar, yet distinct, fashion.

2.3 Results
The results that follow were published in a paper by Wu et al. in a 2010 issue
of the Journal of Molecular Biology. I want to make a note that my personal work on
this paper was completed during my 10-week rotation during my first year at USC.
Under the guidance of Dr. Yongqing Wu, I was responsible for growing the crystal of
MEF2A (2-91) bound to DNA and the diffraction data collection, however, my
28
a b

c d

29
Figure 2.6. Overall structure and packing of MEF2 dimer bound to DNA.
(a) Asymmetric unit containing two independent MEF2:DNA complexes stacking
head-to-head is shown in cartoon diagram. Two monomers in one complex are
colored in green and red, respectively, while in the other complex they are
colored in cyan and violet, respectively. DNA is colored in blue throughout the
illustration. (b) MEF2:DNA complex forming an intertwined dimer is shown in
cartoon diagram along the DNA axis. One monomer is colored in red and the
other in green. All three α helices for both the monomers are labeled in
corresponding colors. (c) Top view of the same complex shown in (b). Here all
three β sheets along with the three α helices of both the monomers are labeled in
corresponding colors. DNA is omitted in tis view for clarity. (d) Surface
representation with the same orientation as shown in (c). Here MADS-box and
MEF2 domain are colored in orange and red in one monomer. Corresponding
colors in the other monomer are blue and green. This representation shows that
the MADS-box and MEF2 domain of the two monomers for an intimately folded
domain.

30
rotation ended before any data analysis occurred. Therefore, I will only address the
data of the structure of the complex and its implications and not the further
modeling proposed in the second half of this journal article.
Since the original crystal structure of MEF2A bound to DNA by Santelli and
Richmond did not have a fully formed MEF2 domain, and the NMR study published
the same year (Huang et al., 2000) did not offer structural information on the same
amino acid residues (79-86) because they were disordered, many people
questioned the true structure and orientation of the MADS-box/MEF2 domain.
While Han et al. was able to show the formation of a concave hydrophobic groove on
MEF2B that interacts with an amphipathic helix of its co-repressors in two separate
ternary complexes, the question remained. Many still believed that the hydrophobic
binding pocket of MEF2 only formed with the interaction of a binding partner.
However, our lab believed that the binding pocket was always present as seen in the
cofactor-bound structures, so we set out to solve the first high-resolution structure
of the intact MADS-box/MEF2 domain of MEF2 bound to DNA without a binding
p artne r , called the “ a p o” stru ctur e. M E F 2 A can b e used w it hou t hesi tance here
because the binding pocket is in the N-terminal region that is highly conserved in all
MEF2 proteins.
Human MEF2A (1-95) was cloned into a bacterial vector, expressed, and then
purified via ion exchange and size exclusion columns on an FPLC machine. Double
stranded DNA containing a consensus MEF2-binding site (CTATTTATAA) was also
purified before proceeding to the process of growing crystals. Once each of these
31

Data Set
R esolution ( A ) 50- 2 . 8 7 (l ast b in 2 . 9 7 A - 2 . 8 7 A )
1
R
s y m
0 . 0 5 2 ( 0 . 5 8 5 )
2
C om p let en ess (%)
9 9 . 6 (9 9 . 5 )
I/σ (I )
4 3 . 7 (4 . 5 )
R ed und an cy 7 . 0 (6 . 9 )

Refinement
R esolution (A ) 4 4 . 0 6- 2 . 9 (l ast b in 3 . 1 2 A - 2 . 9 0 A )
3 , 4
R- f act or (%)
0 . 2 2 0 7 (0 . 3 0 1 7 )
3 , 4
R
f r e e
(%)
0 . 2 7 8 6 (0 . 3 7 3 5 )
r . m.s de v ia ti ons
Bond len g ths (A ) 0 . 0 0 8
Bond an g les (º) 1 . 2 2 7
A v er age B - F act or (A 2
)
8 5 . 4 6

Table 2.1. Statistics of crystallographic analysis.

1. R sym= Σ|I− I  |/ ΣI, where I is the observed intensity and I  is the statistically weighted
average intensity of multiple observations of symmetry-related reflections.
2. The number in parentheses is for the outer shell (last bin).
3. R work= Σ||Fo| - |Fc||/ Σ|Fo| where Fo and Fc are the observed and calculated structure
factor amplitudes, respectively.
4. R free is calculated for 5% of the data that were withheld from refinement.
32
components was purified, they were mixed in a 2:1 molar ratio of MEF2A:DNA since
MEF2 binds DNA as a dimer, and grown by the hanging drop method at 18° C. The
cubic crystals grew to 200 μm in seven days before being harvested in a
cryoprotectant buffer. The crystals were then processed for data collection on
camp us in Dr . Xia oji an g C hen’s lab at ou r ho me X -ray source.
The MEF2A (1-95) protein was solved via X-ray crystallography as MEF2A
(2-91). The first Met residue gets cleaved after processing in the bacterial cell, and
residues 92-95 showed no electron density after data analysis from the X-ray
source. The MEF2A (2-91):DNA crystal had an orthorhombic Bravais Lattice and
adopted the space group P2 12 12 (a=77.80 Å, b=77.96 Å, c=106.79 Å) and diffracted
to 2.87 Å (Figure 2.6). The structure was solved with molecular replacement using
the original MEF2A (2-78):DNA structure by Santelli and Richmond as a partial
search model. The complete data set, including an acceptable R-factor of 22.06%
(R free=27.86%) for the diffraction resolution can be seen in Table 2.1.
The asymmetric unit of the crystals shows two separate MEF2A (2-91):DNA
complexes that stack head-to-head (Figure 2.6a). Helix H2 of each MEF2A:DNA
complex seems to mimic the amphipathic helix of Cabin1 and HDAC9 to bind the
reciprocal complex as seen in previous structures by Han et al. However, the crystal
packing effects are much smaller than those seen in the previous structures, giving
more confidence to the fact that this “ b i nd in g p ock et -bin d in g p ock et ” in te r act ion is a
fairly normal phenomenon. Also, by using multi-angle light scattering, it was found

33

a b

c

34
Figure 2.7. Comparison with the previous crystal structures. (a) C
α

superposition of the present crystal structure of MEF2A (2-91, red) on that of
MEF2A (2-78, cyan) showing the identical core structure of the MADS-box (residues
2-58) and helix H2 and the additional structure of strand S3 and H3 observed here.
(b) Side view of the C-terminal region of the MEF2 domain (residues 79-91, stick
model) and its corresponding electron density (sigma-a weighted 2fofc density at 1
e/Å
3
). The rest of the MEF2 structure is displayed in cartoon diagram. The DNA is
not shown. (c) Superposition of the C
α
backbone of MEF2 in the present crystal
structure (red) with that in the HDAC9:MEF2B:DNA complex i(green) and
Cabin1:MEF2B:DNA complex (blue). This view is in the same orientation as in
Figure 2.6c.

35
that MEF2 (2-95) exists invariably as a dimer, suggesting that the packing
interaction mediated by helix H2 is unstable in solution.
One of several major findings gleaned from solving the apo structure of
MEF2A:DNA was that the MADS-box and MEF2 domain form as a single folded
structure. Previously, it was thought that the MADS-box and MEF2 domain folded as
two separate domains. Secondly, the four MEF2A molecules in the asymmetric unit
show nearly identical structures, despite being in different crystal environments.
This suggests that our structure of MEF2A (2-91) reveals the native folding of the
MEF2 dimer when bound to DNA, and was not a product induced by crystal packing.
Our structure can be compared with that of the fully folded structure of the MADS-
box/MEF2 domain of MEF2B in the Cabin1:MEF2B:DNA complex and the
HDAC9:MEF2B:DNA complex, both solved by Han et al. (Figure 2.7). From this
comparison, it can be seen that the MADS-box/MEF2 domain can form in the
absence of co-repressors and the structure is highly preserved upon co-repressor
binding. A third major finding from our research it that our structure is the first to
show well-defined electron density of DNA in complex with MEF2, which allowed
detailed analysis centering on the way in which the MADS-box/MEF2 domain of
MEF2A interacts with the minor groove of straight B-form DNA (Figure 2.8).
Additionally, our structure indicates that the cofactor binding site, formed by the
central beta sheet acting as the floor and helix H2 from each monomer acting as
each rim of the hydrophobic groove, is preformed in the apo state of the
MEF2A:DNA complex (Figure 2.6). Finally, after comparison with the structures of
Cabin1:MEF2B:DNA and HDAC9:MEF2B:DNA solved by Han et al., helix H2 of MEF2,
36

a b

c

Figure 2.8. DNA binding by the intact MADS-box/MEF2 domain. (a) DNA in stick
model shown in sigma-a weighted 2F o-F c map at a contour level of 2.5 e/Å
3
. (b)
Interactions of the N-terminal tail and helix H1 of the MADS-box with the minor
groove of the DNA. (c) Interactions between the C-terminal end of the MEF2 domain
of one monomer (red) with DNA.

37
the major structural element that interacts with co-factors, shifts significantly
along the axis of the helix with the beta sheet staying relatively unmoved (Figure
2.9). T his is b asi cally sta ti ng that t he “ f loo r ” of the M E F 2 b in d in g p ock et r emai ns
unchanged in either the cofactor- b ound o r ap o f or m, wh ereas t he “ r im” of the
p ock et , f or med b y an H 2 α -helix from each MEF2 monomer, shifts toward each
other to provide a more optimal conformation for cofactor binding. To aid even
further toward optimization of cofactor binding, amino acid residues in the
hydrophobic groove on MEF2 can shift the orientation of their side chains
(compare Figure 2.9a with 2.9b and 2.9c with respect to His76, and 2.9d with
respect to Asp63 in each C
α
trace).

2.4 Discussion
While these previous studies have defined a general mechanism by which
MEF2 interacts with its co-repressors, other mechanisms are possible. For
example, in Figure 2.10, the crystal structure of three MEF2A:DNA complexes
bound by a single TAZ2 domain located in the CH3 region near the C-terminus of
the histone acetyltransferase p300 is shown. Each of the three MEF2A:DNA
complexes binds the TAZ2 domain at three distinct and non-overlapping sites
and produces a trefoil structure in the crystal, with the p300 TAZ2 domain at the
center (He et al., 2011). Interface I is formed between helix α4 of the TAZ2
domain and one of the MEF2A dimers bound to DNA. The helix α 4 of T A Z2 acts
similarly to the amphipathic helices of Cabin1 and HDAC9 and packs diagonally
between helices H2 of the MEF2 dimer (Figure 2.11). Interfaces II and III,
38

a b

c d

39
Figure 2.9. Cofactor binding site. (a) Cofactor binding site of the present structure
in apo MEF2A (2-91) displaying the key residues involved in binding with the
cofactor. Two monomers in the intertwined dimer are colored in green and orange.
The orientation of the cartoon representation is same as in Figure 2.6c. (b) and (c)
are similar representations for the Cabin1:MEF2B:DNA complex and the
HDAC9:MEF2B:DNA complex, respectively. Different side chain orientations of
His76 are marked by arrows in corresponding colors of the monomers in the three
representations (a), (b) and (c). (d) Superposition of the C
α
of the three MEF2
structures (apo in red, HDAC9:MEF2B complex in green and Cabin1:MEF2B complex
in blue) showing different side chain orientations for Asp63. A pair of arrows also
indicates the shift of helix H2 in the apo structure with respect to the HDAC9 and
Cabin1 complexes.

40
however, bind to the DNA-bound MEF2A dimer in substantially different modes
than Interface I, and they both do so in a fashion composed of two separated foci
of binding interactions. A t Inte r f ace I I , helix α 1 an d the fir st z in c lo op of TA Z2 l ie above the H2 helices of the MEF2 dimer at a cross angle of ~60°, instead of lying
inside the hydrophobic groove constructed within the H2 helices. Interface III is
f or med b et w ee n heli ces α1 an d α 3 of the p 3 0 0 T A Z do main an d the t hir d M E F 2 A : DNA com p lex. T he α helices of the TA Z 2 d om ai n p ack again s t helix H2 of
each MEF2A monomer, in an almost parallel orientation. As can be seen from this
crystal structure, not only can cofactors bind MEF2 in different modes, but a
single cofactor like p300 can bind MEF2 in at least three different modes at the
same time unveiling the possibility of higher-order enhanceosomal complexes
allowing combinatorial control of transcriptional regulation.
Differential binding modes seen between the cofactors mentioned above
allow for the possibility to target specific cofactor interactions with MEF2 family
members, without affecting other MEF2 interactions. The development and use
of synthetic small molecule compounds that can bind to MEF2 and inhibit the
binding of its cofactors is an invaluable tool that can be used to understand the
mechanism of specific MEF2:cofactor interactions in health and disease. The
ability of these compounds to allow us to monitor MEF2:cofactor interactions
both in vitro and in vivo provide a great resource in the attempt to modulate the
activity of MEF2 and may possibly result in the treatment of diseases associated
with aberrant MEF2 expression or regulation.

41

Figure 2.10. Structure of the p300 TAZ2 domain bound to MEF2 on DNA.
Overall structure in ribbon style. The MEF2 dimers are colored in yellow, magenta
and blue, respectively; p300 is in green; the DNA backbone is shown in gold and its
sequence is listed below. The secondary structural elements of one MEF2 dimer are
labeled (He et al., 2011).

42
In the simplest of terms, MEF2 acts as an adapter that allows cofactors,
which cannot bind DNA with sequence specificity on their own, to associate with
the genome and alter gene expression patterns. The structural research
performed on MEF2 has provided great insight on the general mechanism of
cofactor interaction with DNA-bound MEF2. This insight has allowed us to start
to understand how a single protein can have so many diverse roles, across so
many diverse cell types. One can imagine that each of various cofactors can
slightly alter the expression of a MEF2 target gene after associating with the DNA
via MEF2. Furthermore, each cofactor can be distributed not only temporally and
spatially under cellular rest, but can also be redistributed during cellular
signaling. These two previous thoughts allow for numerous possibilities of
cellular actions, all going through the middleman, MEF2.
Together, the structural analyses provided by Santelli and Richmond and
the four p ap ers pub lished ou t of D r . L in C hen’s lab at b oth Uni v ersity o f C olo r a d o, Boulder and here at USC describe the features of the MEF2 ligand-binding pocket
that are not only important for understanding the mechanism of cofactor binding,
but also provide a great deal of information that is useful for structure-guided
virtual screening of small molecules that can bind to MEF2 and modulate its
function. In Chapter 3, I will delve into greater depth on compounds that have
been developed in our lab that can bind MEF2 competitively with HDAC4 in vitro
and can modulate the function of MEF2 in vivo.

43

Figure 2.11. Protein–protein interactions at Interface I. Structural comparison
of Interface I (p300 in green and MEF2 in gold) with the Cabin1:MEF2:DNA complex
(red) and the HDAC9:MEF2:DNA complex (cyan). The structures are superimposed
b y the Cα b ack b one of the β strand s of the M E F 2 cor e (H e e t a l. , 2 0 1 1 ).

44
2.5 Materials and Methods

Protein expression and purification
Human MEF2A (residues 1-95) was cloned in pET30b. The MEF2A protein was
expressed in Escherichia coli Rosetta BL21(DE3) pLysS and purified by Sp Sepherase
column (GE) and Superdex 200 column (GE) as described previously. The final
concentration of MEF2A is 32 mg/ml in storage buffer 10mM pH7.6, HEPES, 250mM
NaCl, 1mM EDTA and 1mM DTT.

DNA preparation
DNA (5 ’ A A C T A T T T A T A A GA 3 ’) and it s com p limen ta r y one ( 5 ’ T T C TT A TA A A T A GT- 3 ’) w ere p ur ch ase d f r om I nt eg r at ed DN A T echnolog ie s (C or alv il le, I A ) at 1 µ mo le
scale in the crude but desalted form. The crude DNA was dissolved in a buffer
(100mM NaCl, 10mM NaOH, pH 12.0) and purified by a Mono Q cation exchange
column on FPLC (Amersham biosciences, Piscataway, NJ). The peak fractions were
pooled and neutralized to pH 7.0 by HEPES prior to over night dialysis against
water. The desalted DNA sample was lyophilized to powder, resuspended in water,
and quantified at 260 nm. Complementary DNA strands were annealed at 95°C in
the annealing buffer (100mM NaCl, 5mM HEPES pH 7.6) at concentration of
1.36mM.

45
Crystallization, data collection, and structure determination
The MEF2A:DNA complex was prepared by mixing protein and DNA at 2:1 molar
ratio in storage buffer at a final concentration of 10 mg/ml. Crystals were grown by
the hanging drop method at 18°C using a reservoir buffer of 50 mM acetic acid
buffer pH 4.7, 142mM NaCl, 5mM MgCl 2, 10mM CaCl 2, 3.3% Glycerol and 22.5% PEG
3K. The cubic crystals grew to 200 μm in one week. Crystals were stabilized in the
harvest/cryoprotectant buffer: 20 mM Mg(OAc) 2, 50 mM acetic acid buffer pH 4.7,
30% PEG 3K, and 25% (w/v) glycerol and flash frozen with liquid nitrogen for cryo-
crystallography. The data were collected at a home X-ray source (Micromax) and
Raxis IV++ detector. The crystal diffracted to only 2.8 Å. The data was processed
with DENZO and SCALEPACK. The structure of MEF2/DNA complex was solved by
molecular replacement method using MOLREP (version 9.4.09) from the CCP4
program suite (version 6.0). The crystal structure of MEF2A/DNA complex (pdb
code1EGW) was taken as a partial search model in this molecular replacement. The
C-terminal half of the MEF2 domain was built in O and the structure was refined
with CNS. NCS restraints were used to two copies of MEF2 homodimer throughout
the refinement. The DNA molecule in one MEF2 dimer was modeled as two equally
probable alternate conformations and, therefore, has very high B factors.

Accession Numbers
Coordinates and structure factors have been deposited in the Protein Data Bank
with accession number 3KMD.

46

Chapter 3
Anti-leukemic activity of MEF2
modulators (MEF2m) resulting from
the inhibition of the MEF2:class IIa
HDAC interaction

3.1 Introduction

Leukemia causes more deaths than any other cancer among children and
young adults under the age of twenty (http://www.leukemia-lymphoma.org).
Nearly 50,000 new cases of leukemia will be diagnosed in 2013, while an
anticipated 22,000 deaths of the nearly quarter of a million current leukemia
sufferers will occur. While chronic leukemias are about eleven percent more
common than acute leukemias, most cases occur in adults with a median range at
diagnosis of sixty-six years. However, the most common cancer in children ages one
through seven is an acute lymphocytic leukemia (ALL), that is destined to affect
almost 6 , 0 0 0 of this n at ion’s yo uth in t he ne xt twelv e months.
47
There are four types of leukemias, separated into acute and chronic forms:
acute and chronic myelogenous leukemias, and acute and chronic lymphocytic
(lymphoblastic) leukemias. Acute myelogenous leukemia (AML) and chronic
lymphocytic leukemia (CLL) together will affect over 28,000 American adults this
year alone (http://www.leukemia-lymphoma.org). With symptoms such as swollen
lymph nodes, fever, lethargy, and weight loss, leukemias can often go unnoticed or
be mistaken for the common cold or flu (http://www.medicinenet.com). Only a
doctor can correctly diagnose the disease, either by a blood test to check the levels
of white blood cells, or by way of a bone marrow biopsy. Although this blood
disorder has a five-year survival rate above fifty percent, it is still a devastating
disease and is of great interest to scientists around the globe.
Acute lymphoblastic leukemia (ALL) is a form of leukemia, or cancer of the
white blood cells, characterized by excess lymphoblasts, the immature cells which
typically differentiate to form mature lymphocytes (Harrison's Principles of Internal
Medicine textbook). While lymphoblasts are normally found in the bone marrow, in
acute lymphoblastic leukemia, lymphoblasts proliferate uncontrollably and can be
found in excess in the peripheral blood. ALL can be distinguished from other
malignant lymphoid disorders because they have a B- or T-precursor cell phenotype
(http://emedicine.medscape.com). Ionizing radiation, chemicals, viruses, and
chromosomal abnormalities are a few possible causes of some forms of cancer,
including leukemias.
Chromosomal translocations, a type of abnormality that is often associated
w it h spe cific su b type s o f leuk emia , r esult in ab e r r an t e xpr essi on of f ull -length
48

Figure 3.1. MEF2D/DAZAP1 is a more potent transcriptional activator than
MEF2D. Expression constructs and reporter genes were transiently transfected into
CV-1 cells. The relative luciferase levels were normalized to the levels of β-
galactosidase expression. The average fold inductions s.e. of the means of three
separate experiments are indicated (Prima et al., 2005).

49
proteins or the generation of chimeric proteins that can also contribute to
tumorigenesis. Many of these translocations involve the E2A-PBX1 protein fusion.
In fact, this chimeric oncoprotein can be found in up to ninety-five percent of all
acute lymphoblastic leukemias (Prima et al., 2005). However, in perhaps five to ten
percent of the cases, there is a different translocation of the same two
chromosomes, including the MEF2 genes.
In a study by Prima and Hunger published in 2007, the reciprocal chimeric
proteins created by a t(1;19)(q23;p13.3) were evaluated on their ability to
transform NIH 3T3 cells (mouse embryonic fibroblasts) into cancer-like
malignancies. These chimeric proteins include important functional domains of
DAZAP1 and MEF2D responsible for protein dimerization, DNA or RNA binding,
transcriptional activation and subcellular localization. The sheer breadth of cellular
processes in which these proteins play a part may possibly contribute to
leukemogenesis by altering signaling pathways that are usually mediated by wild-
type MEF2D and/or DAZAP1. In soft agar assays, each of the fusion proteins
transformed NIH 3T3 cells with a 20-fold increase in colony formation as compared
to empty vector or wild-type MEF2D or DAZAP1 proteins, whereas co-expression of
both DAZAP1/MEF2D and MEF2D/DAZAP1 led to a 60-fold increase in colony
formation. The soft agar assay for colony formation is an anchorage-independent
growth assay in soft agar, which is considered the most stringent assay for detecting
malignant transformation of cells. The data presented in this paper show that these
protein fusions can, in fact, transform cells into a cancer-like state, and gives support
to previous studies showing that HeLa cells showed increased colony formation
50

Figure 3.2. The self-renewal-associated signature is found in human MLL-
rearranged AML and activated as a hierarchy of gene expression.
a, GSEA demonstrated significant enrichment (P = 0.016) of the murine self-
renewal-associated signature in human MLL-AML compared to other AML with
defined translocations. b, The 420 probe sets from the self-renewal-associated
signature were ranked according to the distinction of high-level expression in MLL –
AF9 –GFP transduced GMP compared to MSCV –GFP transduced GMP 40 h after
incubation with retroviruses. Eleven genes had a t-test score >2.0 and are labeled to
the right. Genes with a score <2.0 are shaded. c, The 11-gene set was assessed in
MLL-rearranged human AMLs as in a. GSEA demonstrated significant overlap (P =
0.004) (Krivtsov et al., 2006).

51
when they overexpressed the DAZAP1/MEF2D or MEF2D/DAZAP1 fusions (Yuki et
al., 2004).
Prima and Hunger previously demonstrated that MEF2D/DAZAP1 is a
drastically more potent transcriptional activator than wild-type MEF2D, and found
in this study that MEF2D/DAZAP1, but not wild-type MEF2D, transformed NIH 3T3
cells (Figure 3.1) (Prima et al., 2005; Prima and Hunger, 2007). Data suggest that
MEF2D/DAZAP1 is a gain-of-function mutation that directly activates transcription
of genes critical for lymphocyte growth and/or survival, such as c-jun or interleukin-
2, a known transcriptional target of MEF2D in T-cells. With a gain-of-function
mutation, histone deacetylases (HDACs) would not be able to bind to MEF2 and
facilitate gene silencing. Over-active genes, especially those of oncogenes, can cause
a major problem for the health of the cell and of the individual.
Another MEF2 family member, MEF2C, has been implicated in the
establishment of mixed-lineage leukemia (MLL)-induced leukemia stem cells (LSCs).
Aberrations in the MLL gene often develop into a very aggressive type of leukemia
that has both myeloid and lymphoid lineages, and unlike other forms of leukemia,
usually has a very poor prognosis. This poor prognosis is due to the fact that
treatments that usually help in cases of either a myeloid or lymphoid type of
leukemia do not have the same effect on this mixed lineage disease. In late 2006,
Mef2c was identified as a direct target of the MLL-AF9 fusion protein encoded by the
t(9;11)(p22;q23) within the LSC compartment (Krivtsov et al., 2006).

Mef2c, along
with a handful of other genes that are highly expressed in normal hematopoietic
stem cells and involved in hematopoiesis, was shown to be reactivated in murine
52

Figure 3.3. Ectopic expression of Mef2c into Irf8
−/−
BM progenitors induces
acute myelomonocytic leukemia in recipient mice. Kaplan-Meier survival curves
of mice receiving Irf8
−/−
BM cells after infection with retroviral vectors carrying
Mef2c/GFP or GFP alone, or uninfected controls (Schwieger et al., 2009).

53
LSCs (Swanson et al., 1998; Krivtsov et al., 2006). The authors showed that LSCs
could maintain the identity of the progenitor from which they arose, while still
activating stem cell or cell-renewal-associated programs (Figure 3.2). Strikingly,
these LSCs were able to transfer leukemia to secondary recipient mice when only
four cells were transferred.
Acute myelogenous leukemia (AML) is driven by leukemic stem cells (LSCs)
generated by mutations that confer (or maintain) self-renewal potential coupled to
an aberrant differentiation program (Schwieger et al., 2009). With the purpose of
identifying novel genes that cooperate with other genetic mutations to generate
AML LSCs, a group in Germany published a study using the technique of retroviral
mutagenesis in a murine model. The Irf8-deficient mice used in the study display a
“ p r e- leuk emic” p henoty p e a nd hav e a d ecr ea sed sen sit ivity to ap op tot i c stimuli, which allows for a simple model where only a minor deregulation can cause major
visible effects. The mice were infected intraperitoneally with a virus encoding the
MLL/ENL fusion protein, which is a result of a common translocation involving the
MLL gene that was shown to immortalize myelomonocytic progenitors in vitro.
Fusion proteins generated from translocations involving the MLL gene on
chromosome 11q23, which are associated with mixed-lineage leukemia and adult
AML with poor prognosis, have been shown to generate LSCs from committed
myeloid progenitors (Schwieger et al., 2009). All mice receiving Irf8
-/-
bone marrow
cells expressing Mef2c died of AML within a median latency of 79 days after
transplantation (Figure 3.3). Conversely, mice receiving Irf8
+/+
bone marrow cells

54

Figure 3.4. Mapping of 565 unique retroviral insertion sites onto the mouse
genome assembly from Celera Genomics. Blue lines indicate single MoMuLV
insertions, green lines loci previously identified by retroviral insertional
mutagenesis, and red lines newly discovered cancer-related CIS loci, with candidate
gene names in red (Lund et al., 2002).

55
transduced with Mef2c showed no disease over a 9-month span. This shows a
synergy between high Mef2c expression and lack of Irf8 in the induction of invasive
AML and the authors concluded that Mef2c induces myelomonocytic leukemia in
cooperation with Irf8 deficiency (Schwieger et al., 2009).

While Mef2c is not
necessary for MLL/ENL transformation in vitro nor is it required for MLL/ENL-
induced leukemia, it does shorten the disease latency and increases its ability to
mobilize to different tissues. Mef2c allows for its normal homing to various
hematopoietic organs, especially the bone marrow and spleen, while a deficiency of
Mef2c in MLL/ENL-induced leukemia shows an increased latency most likely due to
this fact.
This Schwieger et al. study also took advantage of a database of 285 people
afflicted with AML. MEF2C transcript levels were analyzed, and a positive
correlation between MEF2C and IRF8 gene expression was found, unlike the inverse
relation found in the mouse models. This just seems to suggest alternative
mechanisms in humans, such as deregulation of the MLL gene. The high levels of
MEF2C expression in MLL-leukemia samples agree with the MLL-AF9 studies by
Krivtsov et al published three years earlier.

Mef2c is a cooperating oncogene in leukemogenesis; however, it seems unable
to induce leukemia when expressed alone. This is an important point to note that,
although MEF2 has been linked to several cancers, it is not sufficient to induce
disease. MEF2 might have a trans, versus cis, effect on another gene or pathway that
has not been proven to date. On the other hand, it may turn out that some other
gene may have to be dysregulated first, which then affects MEF2 and leads to onset
56
of cancer. Whether the dysregulation of the MEF2 gene that leads to leukemia is a
cause or effect of the misregulation of another gene is as yet unclear, but the fact
that MEF2 is involved in this disease onset is unquestionable.
MEF2 is dysregulated in several types of cancer, especially in subtypes of
leukemia. Two separate groups performed retroviral tagging of genes and found
common insertion sites (CISs) that are likely to encode cancer genes (Suzuki et al.,
2002; Lund et al., 2002). These CISs were defined by multiple integrations found in
the same locus in independent tumors. Two of the CISs observed are members of
the MEF2 family (Mef2c and Mef2d), linking these genes to cancer for the first time
(Figure 3.4). Mef2c expression was also found to be up-regulated in myeloid
leukemias infected with Sox4 retroviruses. These cells with Sox4 integrations in
Mef2c genes acted cooperatively and resulted in an accelerated form of the disease
(Du et al., 2005). According to the National Cancer Institute, Mef2c has also been
described as a CIS in myeloid leukemias and B-cell lymphomas in mice
(http://RTCGD.ncifcrf.gov).
The MEF2 family is not only implicated in the onset and progression of
leukemia when it is dysregulated or involved in a translocation. In fact, the MEF2
family of transcription factors can actually prevent or protect against the onset of
leukemia. To explain this idea, it must be noted that Mef2c seems to regulate the
fate choice between myeloid and lymphoid lineages, normally repressing the
myeloid state and expressing the lymphoid panel of genes (Stehling-Sun et al.,
2009). The very intriguing fact though, is as follows: MEF2D, conversely, has been
shown to be required in the induced differentiation of HL60 cells (acute
57

Figure 3.5. Structural characterization of the binding of BML-210 to MEF2A.
(A) Electron density (blue mesh) matching the shape of BML-210 (blue stick) was
identified in the hydrophobic pocket of MEF2 (red and yellow ribbon). (B) The
phenyl group of BML-210 is surrounded by a number of hydrophobic residues of
MEF2. (C) The 2-aminophenyl group of BML-210 interacts with a number of
residues on MEF2. (D) A surface representation showing that the methylene groups
of the octanediamide fit snugly between helix H2 of the two MEF2 monomers and
that BML-210 adopts an extended conformation to bind the surface groove of MEF2.
Positive and negative surface potentials are indicated by red and blue, respectively.
(E) Structural superposition using MEF2 as the reference showing that BML-210
and HDAC9 share the same binding site on MEF2 and that the synthetic compound
mimics some of the binding interactions of the natural ligands (Jayathilaka et al.,
2012).

58
promyelocytic leukemia) by VitD 3 along the myeloid lineage, eventually
differentiating into non-proliferating monocytes that enter the peripheral blood
(Shin et al., 1999). This is important because, even though this was done in vitro,
this shows that MEF2D can actually snap a leukemia cell out of its normal state of
uncontrollable proliferation, and actually cause the cell to differentiate along natural
pathways, which then allows the cell to undergo apoptosis just like any other non-
cancerous cell. This differentiation process is actually better than the alternative
option of just slowing or stopping the growth of cancer cells, because a tumor
cannot always be eradicated since they sometimes take on a resistant state.
However, if a cell can be differentiated, it is much more susceptible to the
endogenous immune system cells, as well as therapeutic treatments.
While lymphoblasts normally mature in the bone marrow before being
released as lymphocytes, leukemias result in uncontrolled lymphoblast
proliferation. These immature and non-functional B cells are found in excess in the
peripheral blood and lead to a significantly impaired immune system. Expression of
MEF2C and MEF2D occurs in lymphocytes, but MEF2C is restricted to B-cells.
However, in a subset of T-ALL cell lines ectopic MEF2C expression is correlated with
a 5q14 deletion, encompassing non-coding proximal gene regions, stating the
importance of MEF2 in normal lymphocyte development (Nagel et al., 2008).
Furthermore, a MEF2 family member is among the highest expressed during the
reversion from a committed progenitor to leukemia stem cell, marking its
membership in a leukemia self-renewal signature group and confirming its
oncogenic properties (Krivtsov et al., 2006). Both the altered expression profiles
59

Figure 3.6. Development of BML-210 analogs. (A) Chemical structures of BML-
210 analogs. (B) Effect of BML-210 analogs on the HDAC4:MEF2-mediated
luciferase response. Response is indicated as mean percentage from the DMSO
luc ifer ase r esp onse ± S D ( n = 2 ) ( Ja y at hilaka et al. , 2 0 1 2 ).

60
and increased frequency of mutations in the MEF2 genes found in leukemias and
lymphomas seem to be more than a coincidence. Interestingly, many of these
mutations do not affect the binding of HDACs to MEF2, and some even enhance this
binding such as the MEF2D/DAZAP1 fusion in pre-B ALL (Yuki et al., 2004; Prima
and Hunger, 2007), therefore this protein-protein interaction can be explored
further in these cancerous cells (Lund et al., 2002).
The structural research presented in Chapter 2 helped influence and guide
our current thinking on the MEF2 family and, more importantly, on drug
development to manipulate the interaction between MEF2 and its cofactors. The
design and development of synthetic small molecules by members of my lab in
conjunction with the lab of Dr. Nicos Petasis, has been an integral part of my
personal research. Without these small molecules, I would not have had the tools
necessary to complete my research focusing on the interaction of MEF2 with class
IIa HDACs in the context of leukemia.
The importance of having a structural template for learning about protein-
binding mechanisms and the following downstream effects on other genes and
molecular processes cannot be understated. For example, a recent paper published
by Morin et al. in 2011 showed that a majority (over 80%) of coding single
nucleotide variants and somatic mutations found in the MEF2B gene in non-Hodgkin
lymphomas are located in the MADS-box/MEF2 domain. From the structures
published by Santelli and Richmond, Han et al., and Wu et al., we know that these
mutations affect the binding pocket where MEF2 interacts with at least the class IIa
HDACs, Cabin1, and p300. Being able to visualize the structural motifs of a protein
61

Figure 3.7. Luciferase assay using GFP-HDAC4 (3-209). A) Normalized luciferase
readings using the GAL4-MEF2 and HDAC4-VP16 constructs supplemented with
either empty pEGFP plasmid or the plasmid with HDAC4 (3-209) cloned after the
GFP sequence. B) Normalized luciferase readings using the GAL4-Luciferase and
GAL4-VP16 constructs supplemented with either empty pEGFP plasmid or the
plasmid with HDAC4 (3-209) cloned after the GFP sequence.

62
and possible important residues for interaction with other proteins or DNA is
invaluable to a researchers quest to understand the functions of specific proteins
and their role in a binary interaction, pathway, or signaling cascade.
Our lab has been interested in MEF2 for many years, however our initial
in te r est f ocu sed on M E F 2 ’s r ole in sk ele ta l mu scle d evelopmen t a nd car d ia c
hypertrophy. Through this fairly narrow perception, we searched for small
molecules that could bind MEF2 and modulate its complex formation with its
transcriptional regulators such as Cabin1, class IIa HDACs, and CBP/p300. This
modulation of complex formation was theorized to further modulate the
transcriptional activity associated with MEF2 target genes. Through literature
searches, we found a paper by Savickiene et al. that detailed a molecule called BML-
210. BML-210 is a small benzamide molecule that can decrease the viability of
certain cancer cells, including the HL-60 promyelocytic leukemia cell line
(Savickiene et al., 2006). We used this knowledge, in addition to the
aforementioned links of MEF2 to leukemia and our previous work showing that
MEF2 and HDAC bind each other (Han et al., 2003; Han et al., 2005), to question
whether BML-210 and its related benzamide compounds act through MEF2 to
achieve their anti-leukemic properties.
We performed a series of experiments to test whether this BML-210
compound could affect the binding of MEF2 to class IIa HDACs. Through several
luciferase and surface plasmon resonance (SPR) experiments, along with solving the
crystal structure of BML-210 bound in the hydrophobic pocket of MEF2 (Jayathilaka
et al., 2012), we determined that this compound had the capability to ablate the
63

64
Figure 3.8. Anti-cancer effects of the benzamide derivative NKL-30 in leukemic
and non-leukemic B cells. A) Treatment of cells with NKL-30 at 2µM dissolved in
100% DMSO. B) Treatment of cells with NKL-30 at 4µM dissolved in 100% DMSO.
Cells were treated for 72 hours, with readings taken at each 24-hour time point.
Data points are expressed as a quotient of NKL-30 treatment luminescent value over
DMSO treatment luminescent value. Each data point has an n>3. All three cell lines
starting with the GM prefix are non-leukemic B cells.

65
class IIa HDAC:MEF2 interaction (Figure 3.5). Several BML-210 derivatives were
then developed in our lab with the help of Dr. Nicos Petasis, including one that we
refer to as NKL-30. We have shown that this small molecule occupies the same
hydrophobic pocket on MEF2 as the conserved N-terminal helix found in class IIa
HDACs and disrupts the interaction of the two proteins, just as BML-210, but with
higher potency (Jayathilaka et al., 2012). At the time that many of these
experiments were performed, NKL-30 was the most active drug that we had
developed, however, since then many variations of the benzamide drugs have
shown greater potency to both ablate the HDAC:MEF2 interaction through our
luciferase assay, as well as a greater effect on cellular viability (Figure 3.6). I have
also confirmed, through the use of the HDAC4 helix that binds to the hydrophobic
pocket of MEF2, that the luciferase experiments completed by Jayathilaka et al.,
2012 are specific for the MEF2:HDAC4 interaction (Figure 3.7).
Unlike many HDAC inhibitors (HDACi) which target the catalytic domain of
the HDAC proteins, our lab has developed a series of benzamide derivatives which
preferentially bind to the hydrophobic pocket of MEF2 and, in turn, disallow the
binding of any other molecules to that same area. These small molecule inhibitors,
which we have called MEF2 modulators (MEF2m), are specific for MEF2 and seem
to have less off-target effects than traditional HDACi. Because of their target
specificity, these MEF2m can be used in situations where a cause of the cancer is
MEF2-related, and can therefore be considered a targeted therapy similar to
Imatinib Mesylate (Gleevec), but with an even greater application audience.

66
Here, we propose a new mechanism of action for these benzamide molecules
that is not HDAC-dependent, but rather is MEF2-dependent. In our model, we state
that the small molecule inhibitors discovered by our lab are binding to MEF2 at the
normal class IIa HDAC/Cabin1 interface pocket, which does not allow the binding of
class IIa HDACs. This molecule binding and, therefore, lack of co-repressor binding
then results in the derepression of the coinciding MEF2 target genes.

3.2 Results

Do the newly developed benzamide derivatives have anti-cancer activity in vitro?
To assess the anti-cancer activity of our newly developed drugs, I performed
several cellular viability experiments in seven different leukemia cell lines and three
non-leukemic B cell lines. In all ten cell lines tested (some data not shown), the
addition of NKL-30 decreased the viability of the cells greatly over the DMSO-
treated controls. Importantly, NKL-30 decreased the viability of the leukemia cells
significantly more than it did when tested on the non-leukemic control B-cell lines.
The average IC 50
72h
, defined as the concentration of a drug needed to reduce the
viable cellular population to 50% of matched controls at 72 hours post drug
addition, of NKL-30 on the leukemia cell lines was under 2µM, whereas the average
IC 50
72h
for the non-leukemia lines was roughly 4µM (Figure 3.8). While this
difference in IC 50
72h
does not seem like much, at 2µM the non-leukemia cell lines
were roughly 80% viable. This preferential targeting of leukemia cells offers a very

67

Figure 3.9. Anti-cancer effects of benzamide derivatives on leukemia. Multiple
derivatives of benzamide molecules were developed and then tested for their anti-
cancer efficacy in a single leukemia cell line. REH leukemia cells were treated for 72
hours, with readings taken each 24 hours. Data points are expressed as a quotient
of drug treatment luminescent value over DMSO treatment luminescent value. Each
data point has an n>2.

68
important therapeutic advantage of NKL-30 over many current HDACi or related
small molecule drugs available. Why this effect is seen still needs to be confirmed,
but one hypothesis is that the dysregulation of MEF2 and its association with class
IIa HDACs in leukemia offers a greater opportunity for a targeted small molecule to
return the cell to homeostasis.

Is there a range of anti-cancer activity seen within the benzamide derivatives in
leukemia cells?
We have investigated the effects of several other derivatives of BML-210,
both in their ability to decrease the transcriptional response in a MEF2D-HDAC4
luciferase assay, as well as their effect on cellular viability. I would like to point out
that NKL-22 is compound 4b from a previously published paper (Herman et al.,
2006). This compound only differs from our NKL-30 by a single bromine molecule,
which had been added to the right ring structure on our molecule for the purpose of
crystallization studies. Our results from the luciferase assay and cell viability assay
seem to have a positive correlation. A compound that decreases the luciferase
response will almost certainly decrease the viability of a leukemia cell line.
Compounds NKL-22, -30, -54, and -88 all decreased the luciferase response greatly,
and had a similarly dramatic effect on the viability of a representative leukemia cell
line, REH. Compound NKL-19 was shown to have a low level of activity in our
MEF2D-HDAC4 luciferase assay, and similarly in the viability assay in the REH cell
line (Figure 3.9).

69
A)

B)

Figure 3.10. Combination treatment with MEF2m and Imatinib Mesylate. A)
Control B-cells or B) leukemia cells were treated for 72 hours with either NKL-30,
Imatinib, or both, with viability readings taken each 24 hours. Data points are
expressed as a quotient of drug treatment luminescent value over DMSO treatment
luminescent value. Each data point has an n>2.

70
I wanted to compare our MEF2m to an already FDA approved drug and the
only one currently used as a targeted therapy for leukemia cells: Imatinib Mesylate
(Gleevec). Imatinib Mesylate is a tyrosine kinase inhibitor that is specific for
Phildelphia positive (Ph+) chronic myelogenous leukemia, meaning that there is a
breakpoint fusion in the Bcr-Abl genes that results in a constitutively active tyrosine
kinase. Gleevec has many side effects seen in clinics today including nausea,
vomiting, edema, muscle cramps, liver toxicity, fluid-retention syndromes,
neuropenia and thrombopenia, and many patients develop resistance to the drug in
a short period of time. Sine Gleevec is specific for Ph+ leukemia, it should have little
to no activity on cells that are Ph-. I performed viability experiments on several cell
lines and found, contrary to what had been published before, that the Imatinib did,
in fact, have more than a negligible effect on the viability of my Ph- leukemia cells
(data not shown).
Following this observation, I decided to test a theory based on combined
small molecule administration. I wanted to use a combination treatment consisting
of our MEF2m (NKL-30) and Imatinib Mesylate. The rationale behind my thinking
was that we may be able to piggyback on the success of Gleevec while decreasing it ’s
side effects in patients. I thought that getting our own drug through clinical trials
would be harder than using it in combination with one that is already FDA
approved. As can be seen in figure 3.10, the combination treatment consisting of
Imatinib at 10µM with NKL-30 at 1µM, which is well below the IC 50
72h
for both the
control and leukemic B-cells, acted combinatorially when used on GM12878 healthy
B-cells. However, what is even more striking is the synergistic effect seen on Nalm6
71

A)

B)

72
Figure 3.11. HDAC activity of common HDAC inhibitors and benzamide
derivatives. A) HDAC inhibition of benzamide derivatives. B) Comparison of two
benzamide compounds with similar HDAC inhibition levels that differ in their MEF2
inhibition, as measured by luciferase assay. HDAC inhibition assay fluorescently
measures the amount of acetylated lysine in a given sample. Values are %HDAC
activity remaining. Data points are expressed as drug treatment fluorescent value
over control fluorescent value, which does not contain any inhibitor. TSA-treated
sample is considered positive control. Each data point has an n>2.

73
leukemia cells under this treatment. While the treatment of 10µM Imatinib with
1µM NKL-30 decrease the viability of healthy B cells to ~60% after 72 hours, the
same treatment decrease the viability of a leukemia cell line to below 5%. Several
reasons can account for the effects seen on viability in these cell lines including, but
not limited to: kinase inhibition, off target effects of either Imatinib or NKL-30,
preferential effects on leukemia cells, and membrane permeabilization. The last
possibility is the most intriguing, noting that the increase in membrane permeability
by Imatinib Mesylate can result in the increased intracellular concentration of our
MEF2m in cells. Knowing that our MEF2m preferentially affect leukemia cells above
healthy cells, this would allow us to use a concentration of MEF2m that would
barely affect the viability of healthy B-cells in a patient, while still eradicating the
leukemia cells.

Is HDAC inhibition and anti-cancer activity correlated?
Currently, a common theory in the fields of epigenetics and cancer biology is
that HDAC inhibition and anti-cancer activity are positively correlated. To address
this, I performed a series of HDAC inhibition assays using multiple HDAC inhibitors,
both commercially available and developed on our own (Figure 3.10A). The results
of this study did indeed agree with my previous experiments focused on the anti-
cancer activity of small molecule compounds, some of which are included in figures
3.8 and 3.9. However, since the BML-210 analogs that were developed with the help
of the Petasis lab are supposed to preferentially target the MEF2 protein and not

74

Figure 3.12. Treatment of a control B-cell line with HDAC inhibitors. Multiple
commercially available HDACi and derivatives of benzamide molecules were tested
for their effect on healthy B cells. GM12878 cells were treated for 72 hours, with
readings taken each 24 hours. Data points are expressed as a quotient of drug
treatment luminescent value over DMSO treatment luminescent value. Each data
point has an n>2.

75
HDAC proteins, we wanted to look into the structure activity relationship between
MEF2 inhibition and HDAC inhibition.
Because NKL30 and its analogs contain the 2-amino benzamide functional
group, which is the signature motif of the 2-aminobenzamide class of HDAC
inhibitor, we tested several compounds to see if their activities of MEF2 inhibition
follow their corresponding HDAC inhibition activities. The MEF2 inhibition assay
was carried out by Linlin Ma, via the same procedure I described for figure 3.7. The
results show that NKL-30 and its analogs indeed have HDAC inhibition activity and
that many potent MEF2 inhibitors are also strong HDAC inhibitors. However, the
activity of MEF2 inhibition does not always follow the activity of HDAC inhibition.
This can be seen in several pair-wise comparisons of compounds on their MEF2 and
HDAC inhibition. For example, NKL-30 inhibited the MEF2 reporter signal to 4.1%
and showed 50.1% of relative HDAC activity inhibition (Figure 3.10B). MGCD0103,
a well-established 2-aminobenzamide HDAC inhibitor known commercially as
Mocetinostat, showed slightly higher HDAC inhibition activity (51.6%) but a nearly
three-fold reduction in MEF2 inhibition (12.2%).
While anti-cancer activity, as measured by cellular viability assays, and total
HDAC inhibition often times seem to be positively correlated, the overall HDAC
inhibition also seems to be correlated with the viability of non-cancerous cells as
well. The addition of pan HDAC inhibitors such as TSA and SAHA, and the HDAC1, 2,
and 3 inhibitor MGCD0103, have a marked effect on the viability of previously
healthy B cells, whereas the MEF2 modulators developed in our lab have a much
lower effect on these same cells (Figure 3.11). This observation lends support to
76

Figure 3.13. NKL-30 dose-dependently induces protein expression in MEF2
target genes in a leukemia cell line. DMSO or NKL-30 at increasing concentrations
was added to Nalm6 leukemia cells for 24 hours, followed by whole cell lysis and
western blot analysis using an antibody for Nur77/NR4A1.

77
our model that MEF2 is highly involved in leukemia cells and its dysregulation can
be targeted specifically in those cells with our MEF2m. The MEF2m developed in
our lab do not solely target HDACs themselves, but rather MEF2, thus disallowing
HDAC-mediated repression of the cognate target genes.

Does the addition of benzamide derivatives induce MEF2 target genes in leukemia
cells?
To test whether our small molecule MEF2m can induce MEF2 target genes, I
focused on the nuclear orphan receptor Nur77/NR4A1. NR4A1 is involved in
apoptosis by translocating from the nucleus to the mitochondria and promoting the
release of cytochrome c (Li et al., 2000). When NR41A is up-regulated, it causes the
cell to undergo apoptosis by rearranging BCL from an anti-apoptotic to pro-
apoptotic state (Mullican et al., 2007). I treated leukemia cell lines with a range of
concentrations of NKL-30 overnight and observed that NR4A1 protein expression
increased greatly compared to matched controls (Figure 3.12). The induction
caused by the addition of the MEF2m, presumably by relieving the repression of
MEF2 genes on NR41A promoters by class IIa HDACs, may be the mechanism by
which the addition of NKL-30 results in the death of leukemia cells.

Does the addition of benzamide derivatives result in cytoplasmic shuttling of Class IIa
HDACs in leukemia cells?
Class IIa HDACs have been shown to shuttle between the nucleus and
cytoplasm. When class IIa HDACs become phosphorylated in the nucleus by
78
Ca
2+
/calmodulin-dependent kinase (CaMK), they interact with 14-3-3 and proceed
with CRM1-dependent nuclear export to the cytoplasm (Wang et al., 2000). Ser632
on HDAC4 is one of three Serines know to be phosphorylated by CaMK that
mediates 14-3-3 association and cytoplasmic shuttling. The mutation of all three
serine residues to alanines, which disallows phosphorylation, resulted in abrogation
of 14-3-3 binding. This lack of binding to 14-3-3 leads to nuclear localization of
HDAC4 and enhanced MEF2 repression (Grozinger and Schreiber 2000). However,
mutations made on HDAC4 that affect the binding to MEF2 impair the ability of
HDACs to localize in the nucleus, stating the importance of the MEF2:HDAC
interaction (Wang and Yang, 2001). So, although HDAC4 is predominantly nuclear
when it cannot bind 14-3-3, it is predominantly cytoplasmic when it cannot bind
MEF2, showing that the interaction with MEF2 is the driving force in the nuclear
localization of HDAC4. To test whether the addition of our MEF2m can result in the
shuttling of class IIa HDACs from the nucleus to the cytoplasm, I performed a
nuclear fractionation after treating cells with NKL-30. Leukemia cells treated with
the MEF2m for 24 hours showed an increase in phosphorylated HDAC4 in
cytoplasmic fractions, compared to the DMSO-treated control samples, while no
change in phosphorylated HDAC4 was seen in the nucleus (Figure 3.13). This
phenomenon was seen in two different types of leukemia (Nalm6 is pre-B ALL;
HL-60 is acute promyelocytic leukemia) and was also tested in
immunocytochemistry experiments with results that suggest the same.

79

Figure 3.14. NKL-30 treatment results in accumulation of phosphorylated
HDAC4 in the cytoplasm of leukemia cells. Addition of DMSO (lanes 1 and 2) or
NKL-30 (lanes 3 and 4) for 24 hours to A) Nalm6 or B) HL-60 leukemia cells,
followed by nuclear fractionation and western blot analysis with -pHDAC4.
Compare lanes 2 and 4 for phosphorylated cytoplasmic HDAC4. N=nuclear fraction.
C=cytoplasmic fraction.

80
Do benzamide derivatives show anti-cancer activity in vivo?
To assess the efficacy of NKL-30 on leukemia in vivo, a mammalian model had
to first be established. The following model and data are all courtesy of Dr. Meng
Xia. After my recommendation to proceed with a mouse model focusing on the
Nalm6 cell line due to its striking reaction to NKL-30 treatment, Dr. Xia went to
work with establishing a mouse model. This is in no way a trivial task, and it took
several months just to establish a working model of leukemia in the mice, even
before starting new rounds to establish drug toxicity and efficacy. Once the basic
model and toxicity rounds were completed, the next step was to inject these mice
with Nalm6 cells and start NKL-30 treatment. Athymic female Nu/nu mice, 4-5
weeks old, were first treated with cyclophosphamide (CP) via intraperitoneal (IP)
injection for two days to decrease their already compromised immune response,
which allows the resulting disease to take effect much more efficiently and rapidly
(Figure 3.14). Twenty-four hours after the last CP injection, 5 ×10
6
Nalm6 cells
were injected into the mice via their tail vein. Five days after the leukemia injection,
mice were weighed and then randomized into two groups of five for the
commencement of treatment with either NKL-30 at 50 mg/kg/d three times per
week, or a vehicle control administered following the same schedule and the body
weight of each mouse was recorded twice per week.
NKL-30 treatment alone did not have an adverse effect on the mice, and was
tolerated quite well compared with vehicle control before the onset of disease
(Figure 3.15). This is an extremely important observation, since a goal of many
cancer therapeutics is to kill cancer whether or not it also kills healthy cells. This
81
gives us great confidence that our small molecule is targeting leukemia cells over
healthy cells. NKL-30 showed remarkable efficacy to extend the survival time of
mice infected with Nalm6 leukemia cells. The mean survival time of the vehicle
control group was 36.6±3.91 days, whereas the NKL30 treated group was 48.8±9.65
days (Figure 3.16). An observed increase in mean survival time of roughly 25%
over vehicle control-treated mice (p-value<0.05) offered by NKL-30 treatment is a
great first step in the testing of our small molecule MEF2m on leukemias.

Is MEF2 constitutively nuclear?
One final, and seemingly unrelated, experimentally testable question
revolves around the localization of MEF2. Many agree that MEF2 is found in the
nucleus, constitutively bound to DNA. However, there are several groups who have
proven this otherwise. MEF2 can be found in the cytoplasm in a variety of cells and
situations, but with the shared caveat that the cell must not be terminally
differentiated. Not only can MEF2 be found in the cytoplasm, but two different
groups have provided evidence that MEF2 can actually shuttle into the nucleus (De
Angelis et al., 1998; Chen et al., 2001). The observation that MEF2 can be found in
the cytoplasm in some undifferentiated cells may provide some clarity to the data
showing that MEF2 is up-regulated in leukemia (Figure 3.17). Since leukemia cells
are usually blast-like and undifferentiated, these cells still can access their stem cell-
like gene expression signatures. The possibility that these signatures are a cause, or
more likely a result, of MEF2 localization has yet to be determined. However, it is
82

Figure 3.15. Nalm6 leukemia mouse model and in vivo study protocol. To
check the in vivo activity of NKL30, we used a mouse model of leukemia. Here
Nu/nu mice (4-5 weeks old, female) were first treated by IP injection of cyclo-
phosphamide for 2 days. After resting for 24 hours, 5 ×10
6
Nalm6 cells were
injected into the mice via the tail vein. The mice were randomized and grouped into
vehicle control [20%DMSO+ HOP- β-CD: (2-Hydroxypropyl)- β-cyclodextrin] and
treatment {vehicle plus NKL-30, 50mg/kg/d IP 5 times per week). The treatment
started 5 days after the inoculation of the Nalm6 cells. Body weight (B.W.) of the
mice was measured at least twice weekly and survival time was recorded.

Figure 3.16. NKL-30 treatment alone does not affect mouse body weight. The
drug seems to be tolerated well in vivo as evident by the stable body weight and the
little difference between the vehicle control and treatment groups.

Figure ## NKL30 is well tolerated in mice.

83

Figure 3.17. In vivo efficacy of NKL-30 in Nalm6 leukemia mouse model.
Survival plot showing the difference between groups of mice treated with either
NKL-30 or a vehicle control. Five mice were in each treatment group. The mean
survival time of the vehicle control group was 36.6±3.91 days, whereas the NKL30
treated group was 48.8±9.65 days (p-value=0.031). Data courtesy of Dr. Meng Xia.

84
intriguing to think that spatial localization, and not gene or protein expression, can
be a leading factor.

Does the NCI60 mRNA expression data lend support to an HDAC3-independent model?
Through the use of CellMiner (Genomics and Bioinformatics Group,
Laboratory of Molecular Pharmacology, Center for Cancer Research, National Cancer
Institute), I was able to tease out some very useful data from the NCI60 database. I
was able to find that in the six leukemia lines tested, there was a trend of the up-
regulation of CABIN1 gene expression and down-regulation of NR4A1 (Nur77) gene
expression (Figure 3.18). Cabin 1 protein represses MEF2-dependent NR4A1/Nur77
expression through the recruitment of HDAC1 and 2 after binding MEF2 (Youn and
Liu, 2000). This up-regulation of Cabin1, and its further association with MEF2 is
one way to explain an HDAC3-independent regulation of the MEF2 target gene
NR4A1, which is frequently up-regulated during apoptosis. The down-regulation of
NR4A1 in leukemia cells suggests a deregulation of normal cellular function leading
to uncontrolled proliferation without the checks and balances of self-induced cell
death. The addition of our small molecule MEF2 modulator (MEF2m) NKL-30 leads
to a decrease in cell viability as well as a prominent increase in NR4A1 protein
expression. At least a portion of this decrease in viability can be attributed to the
increase in NR4A1 protein and its role in apoptosis, and therefore is not dependent
upon HDAC3, which is the currently accepted method of action. Add to this that the
class I-specific HDACi MS275 was unable to induce apoptosis in T-ALL cells and this
fuels our argument that HDAC3, or even acetylation status, does not need to be at
85

Figure 3.18. MEF2 is found outside of the nucleus in HL-60 leukemia cells.
Nuclear fractionation of HL-60 cells followed by western blot analysis. Cells were
either treated with 0.05% DMSO or 1µM NKL-30. -HP1  antibody indicates nuclear
fraction. N=Nucleus; C=cytosol.

86
the forefront as the answer to our main question on the mechanism of action of our
benzamide analogs (Tsapis et al., 2007).

Can sumoylation account for the repression of MEF2?
HDAC3 has been shown to deacetylate MEF2 both in vitro and in vivo,
whereas HDAC4 on its own has not. However, class IIa HDACs can promote
sumoylation as a form of repression. This sumoylation on a lysine residue in the C-
terminal region of MEF2 not only acts to repress MEF2 on its cognate target genes,
but also prevents the binding of its coactivator CBP, and the further acetylation on
the same lysine of MEF2 (Zhao et al., 2005). In this acetylation-independent way,
class IIa HDACs repress MEF2 using their N-terminal domains, independent from
the catalytic domain, which have been shown to bind the MEF2 binding pocket, thus
occupying the same space as our MEF2m (Gregoire and Yang, 2005). The physical
binding of class IIa HDACs is required for the stimulation of MEF2 sumoylation;
therefore our MEF2m can block this epigenetic event from occurring. This lends
strong support to our model that the small molecule MEF2m ablates the physical
interaction between HDAC4 and MEF2, thus leading to derepression of the MEF2
proteins and their target genes. Further experimental work needs to be completed
in this area, as my data on this question has not been conclusive as of yet.

87

88

Figure 3.19. NCI60 mRNA expression data lends support to an HDAC3-
independent mechanism of action for benzamide derivatives. Gene expression
profile of A) NR4A1/Nur77 and B) Cabin1 from the NCI60 cancer cell lines, courtesy
of C ellM in er. T he six leuk emia cell line s are liste d in lig ht g r ee n with the p r efix “ L E ” .

89
3.3 Discussion

The small molecule benzamide MEF2 modulator (MEF2m) NKL-30, which is
a derivative of BML-210, was used as a tool to investigate the interaction of HDAC
and MEF2 in leukemia. This compound allowed us to visualize the importance of
this protein-protein interaction in an appropriate cellular setting. We found that,
while the interaction of these two proteins in healthy cells is necessary and required
for normal function and development, the ablation of this MEF2:HDAC interaction in
leukemic cells leads to growth inhibition and cell death. Importantly, we also
present a new mechanism of benzamide compounds that is not HDAC3-dependent.
We provide evidence that our small molecule compounds specifically bind MEF2,
and that this results in preferential growth inhibition and cell death of leukemia
cells over healthy control B cells.
We further show that the ablation of the MEF2:HDAC interaction by NKL-30
results in HDAC4 cytoplasmic shuttling. This shuttling, preceded by HDAC4
phosphorylation, offers an explanation about the mechanism of derepression of
MEF2 target genes such as NR4A1/Nur77 (Figure 3.19). Induction of the nuclear
orphan receptor Nur77 protein after the addition of NKL-30 provides some insight
into how this small molecule causes cell death, since Nur77 is commonly up
regulated before the onset of apoptosis (Youn and Liu, 2000).
These new benzamide derivatives should be examined further from a clinical
standpoint as a possible new therapy with less severe side effects than currently
available drugs, owing to their specificity for MEF2. Also, further efforts should
90
f ocu s on the t arg et in g o f sp eci f ic M E F 2 an d H DA C f amily membe r s an d the dr ug ’s
effects on other types of cancers or lymphomas. Creating a comprehensive list of
MEF2:HDAC binding interactions on genes and promoters via ChIP-seq, as well as
determining the gene transcriptional changes induced by NKL-30 are also of
importance in the near future to fully understand how this drug works and the exact
mechanism of the MEF2:HDAC interaction in leukemia.
While MEF2 is up-regulated in many types of cancers, unpublished work
from our lab shows that HDAC9 gene expression is also up-regulated in leukemia
cells when compared to control B cells. This could help elucidate the answer as to
why MEF2 modulators (MEF2m) cause cell death in leukemia even though they
result in the derepression of MEF2. Basal levels of class IIa HDACs in leukemia cells
could still saturate MEF2 on crucial target genes, regardless of the basal increase
seen in MEF2 in leukemia shown in previous publications. Therefore, the addition
of MEF2m would ablate the MEF2:HDAC interaction, allowing derepression of MEF2
and subsequent transcription of its target genes, which can lead to a halt in growth
progression, and even apoptosis.
As can be envisioned, MEF2 is an oncogene that seems to play a role in
leukemogenesis. And, not only does it play a role in the general disease, but it has
been implicated in multiple families of this blood disease. From ALL to AML and
beyond, there is increasing evidence that the MEF2 family of transcription factors
affects the mobility, homing, aggressiveness, and latency of leukemia. The fact that
various members of the MEF2 family affect each type of leukemia in a different way
only proves that this family of transcription factors is implicated in more than just
91

92

Figure 3.20. Proposed model of the mechanism of MEF2m NKL-30. A) In the
absence of the MEF2m NKL-30, class IIa HDACs use their N-terminal region to bind
the MADS/MEF2 domain hydrophobic pocket of MEF2 proteins in the nucleus. B)
After NKL-30 is added, it permeates the cell and nuclear membranes where it binds
the MADS/MEF2 domain hydrophobic pocket of MEF2 proteins. This binding
disallows further binding of class IIa HDACs and results in their phosphorylation by
CaMK. This phosphorylation event leads to association with 14-3-3 protein and
active shuttling to the cytoplasm, thus resulting in derepression of MEF2 and its
target genes.

93
myogenesis. Only further research, preferably in humans, or at least more in vivo
work in other mammals , w ill help to e luc id at e M E F 2 ’s tr ue r ole in the pathogenesis
of these blood cancers.
Along with further research on the mechanisms of action and delineating the
pathway that leads from MEF2 dysregulation to disease, there is also a chance to
work on therapeutic approaches to combat this disease using all of the readily
available information gathered on MEF2 already. Possible future experiments
include using drugs similar to the NKL molecules that are targeted at full-length wild
type MEF2, especially at the sites by which it interacts with other proteins. If the
drugs were able to interfere with the interaction, it could be envisioned that the
gene expression could be altered to revert the cell back into a healthy state, or to
keep it from making the transition into a cancer cell.
Future studies may not have to focus on MEF2 itself, but could instead focus on
proteins that regulate this transcription factor. One of the protein families that has
been shown to directly interact with, and regulate, MEF2 is the histone deacetylase
family (HDAC). HDACs regulate MEF2 in a repressive manner, and therefore inhibit
certain genes that may keep the cell safe from moving into a leukemia-like state.
HDAC inhibitors have been in the spotlight for pharmaceutical companies and there
have been studies that report that certain drugs can induce apoptosis and
differentiation and also have a growth inhibitory effect on human leukemia cell lines
(Savickiene et al., 2006). However, there are eleven different HDACs, split into three
different families, so a more targeted approach similar to the one detailed here is
needed over the broad-spectrum inhibition that is provided by most drugs.
94
A good portion of research in academia is aimed at furthering scientific
knowledge. However, in my personal opinion, it is very special when research can
be used to improve human health. While this is never guaranteed, the research that
I have carried out in my time at USC may benefit people afflicted with leukemia. I do
not want to overstate the importance or implications of my research, as I know that
many before me have had more successful results that never transitioned to clinical
use. However, the results that have been presented here, subject to further
experimental validation and mechanistic analysis, suggest that we have found a
group of small molecules that can specifically target leukemia cells for death over
healthy cells. This allows us to add yet another bullet into the chamber to combat
the enemy of disease.

3.4 Materials and Methods

Cell Culture
All B-cell lines were maintained in RPMI 1640 media supplemented with 2mM L-
glutamine and 10% FBS and cultured in an incubator at 37 C with 5% CO 2. Cell lines
GM12878, GM15850, and GM15851 were purchased from the National Institute of
General Medical Sciences Human Cell Repository at the Coriell Institute. The
remaining cell lines were a gift from Dr. Markus Muschen (USC, CHLA).

95
Cell Viability Assay
The cell viability assay was performed according to the Promega CellTiter Glo
protocol. Briefly, cells were seeded in 24 well plates at a density of 0.25x10
6
cells/well. Cells were treated with the appropriate drug and concentration and
incubated for 72 hours, taking samples at each 24-hour interval. Samples were read
using a Berthold Lumat LB 9507 luminometer. Values were normalized over DMSO
treated control wells. All experiments were done in duplicate.

Western Blotting
All protein analyzed was endogenous, therefore no transfections or tagged proteins
were required. -Nur77 (SCBT 1:1000), -pHDAC4 (SCBT 1:750), -MEF2(C-21)
(SCBT 1:200), -MEF2(B-4) (SCBT 1:200), and -HP1  (Sigma 1:5000) primary
antibodies were used overnight at 4 C, followed by the appropriate secondary
antibodies from GE Healthcare at 1:10,000.

NR4A1/Nur77 protein expression
Cells were seeded in 6-well plates at a density of 1x10
6
cells/well. Cells were
treated with the appropriate drug and concentration and incubated at 37 C and 5%
CO 2 overnight. Cells were then harvested, pelleted, and then lysed in SDS sample
buffer, followed by membrane disruption using a cell sonicator. Protein
concentrations were determined via BCA assay and 10µg of total protein was loaded
onto a 10% SDS gel and further processed via western blotting technique. Santa
96
Cruz Biotech anti-Nur77 (1:1000) and Sigma-Aldrich anti-Beta-Tubulin (1:2000)
were used as primary antibodies.

Nuclear Fractionation
Nuclear fractionation was carried out according to previously published procedures
(Wu, Trievel, and Rice, 2007). Briefly, cells were grown in a T25 flask and treated
with the indicated concentration of small molecule drug for 24 hours. Cells were
harvested and then washed with PBS. Pellet was resuspended in Nuclear Isolation
Buffer (NIB) and incubated on ice for 10 minutes. Solution was centrifuged at 600 x
g for 5 minutes at 4 C to isolate nuclei.

Histone deacetylase inhibition assay
HDAC inhibition was measured using an HDAC Inhibitor Drug Screening Kit
(BioVision; K340- 1 0 0 ), a cco r d in g to man uf act u r er’ s p r otocol , w it hou t an y deviations.

Mammalian Model
Athymic female Nu/nu mice, 4-5 weeks old (purchased from Simonsen
Laboratories, Gilroy, CA), were first treated with cyclophosphamide (CP) via
intraperitoneal (IP) injection for two days. Twenty-four hours after the last CP
injection, 5 ×10
6
Nalm6 cells were injected into the mice via the tail vein. Five days
after the leukemia injection, mice were weighed and then randomized into two
groups of five for the start of treatment with either vehicle control [20%DMSO+
97
HOP- β-CD: (2-Hydroxypropyl)- β-cyclodextrin] or NKL-30 (vehicle plus NKL-30 at a
concentration of 50mg/kg/d IP, five times per week). The body weight of each
mouse was recorded twice per week, and survival time was recorded.

Luciferase Assay
The luciferase assay was performed according to the manufacturer's protocol
(Promega). The luciferase response was normalized against the Renilla Luciferase
as an internal control. The data is presented as normalized HDAC4:MEF2 luciferase
response against the normalized response values for GAL4-VP16 for each condition
to correct for non-specific inhibition of the luciferase signal. HeLa cells were
transfected using the calcium phosphate precipitation method.

Immunocytochemistry
Nalm6 and RS4;11 leukemia cells were cultured in 6 well tissue culture plates on
glass cover slips coated with Poly-D-lysine. Cells were allowed 24 hours to attach to
the surface of the cover slips at 37 C and 5% CO 2 before being treated with either
DMSO or 1µM NKL-30 overnight. Cells were fixed with 4% PFA for 5 minutes are
drug treatment and immunocytochemistry was performed according to Lewis, Jr. et
al., 2009. A blocking solution containing 3% bovine serum albumin and 0.1% Triton
X-100 in PBS was in place of the published blocking solution. Primary antibodies
were diluted in blocking solution and incubated at room temperature on the glass
coverslips for 45 minutes. The primary antibody mouse anti-MEF2 (B4) (Santa Cruz
Biotechnology) was used at 1:50 and rabbit anti-HDAC4 (Coriell Institute) was used
98
at 1:500. Secondary goat antibodies were conjugated to fluorophores Alexa 488 or
Alexa 594 (Invitrogen) and were used at 1:1000 for 30 minutes in the dark at room
temperature.

Image Analysis
Imaging was performed on an Olympus BX51 upright fluorescence microscope
using a 40X and 100X objective. SPOT Advanced software was used to visualize the
fluorescent images and perform the image merges.

99

Chapter 4
Genome-wide association studies on
the effect of MEF2m in B cells

4.1 Introduction

Throughout this dissertation, others and myself have performed many
different types of experiments or assays that have focused on visible responses at
the protein, overall cellular, or organismal level. However, I wanted to see what
happens to cells that are treated with NKL-30 at gene level. Although changes at
the gene level do not always translate into changes at the protein level, we cannot
discount these in our quest to understand the mechanism of our MEF2m, and the
MEF2:class IIa HDAC interaction. Over the past few years, I have completed several
small qPCR studies on a few handful of genes, but with the price of DNA sequencing
plummeting as new sequencing platforms and protocols have been developed, it
finally became feasible to try a more high-throughput approach. qPCR is a small-
100
scale approach at understanding gene expression through the isolation of cellular
RNA, mRNA enrichment, and cDNA synthesis. The final cDNA products, which are
DNA versions of messenger RNA molecules that are often the precursors to proteins,
can be amplified with primers designed for specific genes in an organism, with the
final goal of total expression level counts. RNA-seq (mRNA-seq in this case),
however, follows a fairly similar protocol, but allows for the final cDNA product to
be sequenced in extreme depth on next generation sequencers such as those made
by 454, ABI, or Illumina, on which I will focus. The technology can offer insights into
pathways affected under different conditions and provides an unbiased look at the
inner workings of the cell. I wanted to see any possible changes in global gene
expression and transcript usage, possibly owing to differential splicing, and RNA-
seq gave me the best chance to get all of this data in a single experiment without
biases associated with my theories on which proteins or gene products may be
affected.
After the cDNA products are run on the Illumina HiSeq Genetic Analyzer
machine, sequencing read files, stored in fastq format, are compiled. A great way to
analyze this data is by using the Tuxedo protocol (Figure 4.1). From the fastq basic
DNA read files, the first step is to align sequencing reads to the reference genome.
TopHat is a program that aligns RNA-seq reads to the genome using Bowtie, which
is a general purpose short read aligner. After a proper alignment, the next step is to
run the Cufflinks package which is comprised of four separate programs that
generate a transcriptome assembly. The Cufflinks program is the first part of the
Cufflinks package and it assembles the transcripts for each condition. After
101

102
Figure 4.1. An overview of the Tuxedo protocol. In an experiment involving two
conditions, reads are first mapped to the genome with TopHat. The reads for each
biological replicate are mapped independently. These mapped reads are provided as
input to Cufflinks, which produces one file of assembled transfrags for each
replicate. The assembly files are merged with the reference transcriptome
annotation into a unified annotation for further analysis. This merged annotation is
quantified in each condition by Cuffdiff, which produces expression data in a set of
tabular files. These files are indexed and visualized with CummeRbund to facilitate
exploration of genes identified by Cuffdiff as differentially expressed, spliced, or
transcriptionally regulated genes. FPKM, fragments per kilobase of transcript per
million fragments mapped (Trapnell et al., 2012).

103
transcripts are assembled, the Cuffcompare program can be started, which
compares transcript assemblies to the annotated reference genome used in TopHat.
Following the annotation comparison, Cuffmerge merges multiple transcript
assemblies to provide a uniform basis for calculating gene and transcript expression
in each condition. Using these merged assemblies and sequencing reads, Cuffdiff
can then calculate differentially expressed genes and transcripts, while also
detecting differential splicing and promoter usage between conditions. Cuffdiff also
tests the statistical significance of any observed change. The output files from the
Cufflinks package can then be run in the R environment using the CummeRbund
plotting tool. CummeRbund allows for many types of graphical data visualization
including volcano, scatter, and box plots (Trapnell et al., 2012). Finally a program
like Ingenuity IPA can be used for complex network and pathway analysis.
The number of RNA-seq reads generated from a transcript is directly
p r op or ti ona l to that tra nscr ip t’s r ela ti v e a b und an ce i n a samp le (T r ap ne ll et a l. , 2012). Cufflinks and Cuffdiff implement a linear statistical model to estimate an
assignment of abundance to each transcript that explains the observed reads with
the maximum likelihood. The commonly used fragments per kilobase of transcript
per million mapped fragments (FPKM) incorporates two normalization steps to
ensure that expression levels for different genes and transcripts can be compared
between multiple conditions and multiple sequencing runs. This FPKM is directly
proportional to abundance, and explains how to calculate expression levels from
read counts. This calculated expression level can then be manipulated and analyzed
104
to compare two conditions or simply shed light on the transcriptional profile of a
previously unannotated organism.
No matter the desired outcome, RNA-seq offers the opportunity to compile
large amounts of data that can be used in a variety of ways. The large amount of
data received from a single experiment, once published, can then even be
reanalyzed by other scientists with a different focus to provide more opportunities
for learning. For these reasons, and more, RNA-seq is an invaluable tool that can be
used to further scientific knowledge, even by those who did not perform the
experiments themselves.
Several studies coming out of the Gottesfeld laboratory at The Scripps
Research Institute in La Jolla, CA have centered on HDAC inhibitors in the treatment
of Friedre ich’s A ta xia (F RD A ) and H unti ngton’s Disea se (H D) . FRDA is caused by a
defect in transcription resulting from hyperexpansion of GAA-TTC triplet repeats in
the first intron of a nuclear gene that encodes the essential mitochondrial protein
frataxin (FXN) (Campuzano et al., 1996; Pandolfo, 2003). Gene silencing at
expanded FXN alleles is accompanied by hypoacetylation of histones H3 and H4 and
trimethylation of histone H3 at Lys9, which is consistent with a heterochromatin-
mediated repression mechanism (Herman et al., 2006). However, there is no
apparent correlation between total HDAC inhibition activity and the ability of
compounds to activate transcription of FXN in live cells.
A substantial increase in frataxin expression in lymphoblastoid cells and in
primary, non-replicating lymphocytes from FRDA patients was only obtained when
using BML-210 analogs, and not with other pan- or class-specific HDAC inhibitors.
105
Compound 4b(N1-(2-aminophenyl)-N7-phenylheptanediamide) and related
compounds raised frataxin levels to and above those of carriers, without apparent
toxicity. The increase in FXN transcription was accompanied by increased
acetylation of histone H3 at lysine14 (H3K14), as well as at H4K5 and H4K12 near
the GAA repeat (Herman et al., 2006).
This group followed their biochemical results with papers focusing on
mammalian models of FRDA (Rai et al., 2008) and H unti ngton’s Disease (Thomas et
al., 2008). In the HD studies, they used R6/2 transgenic mice, which is the most
widely used model for preclinical trials, and demonstrated therapeutic efficacy in
preventing motor deficits and neurodegenerative processes. It was determined that
HDACi 4b treatment ameliorated gene expression abnormalities, detected by
microarray analysis in these mice (Thomas et al., 2008). In the FRDA studies, they
generated homozyogous fxn
(GAA)230/(GAA)230
(KIKI) mice that have mildly but
significantly lower frataxin levels than wildtype animals with the same strain
background (C57Bl/6). These KIKI mice were treated with a BML-210 analog called
compound 106, which substantially increased frataxin mRNA levels in cells from
Friedreich’s A taxia individuals. Treatment increased histone H3 and H4 acetylation
in chromatin near the GAA repeat and restored wild-type frataxin levels in the
nervous system and heart and it was observed that most of the differentially
expressed genes in KIKI mice reverted towards wild-type levels, as determined by
quantitative RT-PCR and semi-quantitative western blot analysis (Rai et al., 2008).
While these aforementioned publications focused on BML-210 analogs and
their effect on viability and gene expression, the Gottesfeld group maintains an
106
HDAC3-centric theory and “slow-on/slow-off ” k in et ics e xplan at ion on the
mechanism of action of these drugs (Soragni et al., 2012). This is in disagreement
with our MEF2-centric theory, however we do not rule out off-target effects on
HDAC proteins, or even other proteins. To help understand how our theory may
affect gene expression, we did extensive literature research and came across a paper
published by Flavell et al. in 2008 which made use of MEF2 loss-of-function and
gain-of-function protein constructs. The goal of this research was to determine
which activity-regulated genes might be targets of MEF2 that control synapse
development. MEF2 expression in neurons was reduced by introducing into the
neurons MEF2A- and MEF2D-specific shRNAs using lentiviral vectors. By
comparing the microarray results between the control (scrambled shRNA
expression) and MEF2 RNAi conditions both before and after membrane
depolarization, a large set of genes whose expression is altered as a result of MEF2
knockdown was identified (1365 probe sets).
It is possible, however, that particular activity-regulated MEF2 target genes
were not identified in the MEF2 loss-of-function experiments because another
transcription factor compensated for the loss of MEF2 at the promoters of these
genes. To address this, a MEF2 gain-of function experiment was performed in
hippocampal neurons at the time that synapses are forming and maturing. An
inducible form of MEF2 was used, in which the ligand binding domain of the
estrogen receptor (ER) is fused to the C-terminus of the constitutively active MEF2-
VP16 fusion protein to generate MEF2-VP16-ER. A total of 251 probe sets showed
107
an increase in expression, with 55 of those showing an overlap with the MEF2 RNAi
data.
Taken together, the studies by Gottesfeld show that BML-210 analogs alter
gene expression in a variety of cells in vitro and in vivo, and the genome-wide
association studies by Flavell et al. show that MEF2 knockdown or upregulation
alters activity-dependent transcriptional programs in neurons. These data helped
set the groundwork for our thinking on how MEF2 may be affected by BML-210
analogs, and allowed us to design an experiment to test our theory of MEF2 ’s r ole in leukemia.

4.2 Results

First of all, without the tutelage of Zach Frazier, a graduate student in Dr.
Frank Alber ’s lab , I w ou ld not ha v e b ee n a b le t o an aly z e a ny o f the se q u encing data
myself. I owe him a great amount of thanks for his time which started with teaching
me the very basic lessons of Linux, R, and simple script writing and ended with a
finished product that allowed me to analyze differential expression between eight
different conditions in two different cell lines. While I completed the computational
data analysis on my own, I cannot thank Zach enough for all of the time he spent
teaching me the basics and then helping me troubleshoot as problems arose.
The rationale for this mRNA-seq experiment was to not only see the
differences between two cell lines that I have worked with over the past several
108

109

Figure 4.2. Metabolic analysis of up-regulated genes in GM12878 cells treated
with NKL-30 2μM for 24 hours. Grey shading indicates genes shown to be
differentially expressed between the two samples. Solid lines denote direct
interaction, while dotted lines indicate indirect interaction. The legend is applicable
to this figure and the following figures.

110
years, but also to see how our MEF2 modulator (MEF2m) actually affected these cell
lines in the most unbiased way possible. GM12878 cells are healthy control B cells
and Nalm6 cells are a representative leukemia cell line with high sensitivity to our
MEF2m, NKL-30. Both of these cell lines have been studied in depth by me and were
presented in the previous chapter. The treatment of NKL-30 2 μM is the average
IC 50
72h
for the group of leukemia cell lines I tested, with it decreasing the viability of
Nalm6 cells to ~60% after 24 hours and ~20% after 48 hours. My decision to have
samples at two time points, 24 and 48 hours after drug addition, was centered on
my desire to examine early and late gene expression changes to observe the
pathways affected that lead to a decrease in viability. With all of these thoughts in
mind, the experiment was completed and analyzed.
The analysis of GM12878 cells treated for 24 hours with DMSO or NKL-30 at
2 μM (GM24D-GM24NKL) showed an up-regulation of genes associated with cellular
assembly and organization (Figure 4.2). In fact, only two genes were found to be
significantly differentially expressed in a down-regulated manner in this pairwise
examination (FAM156A, which codes for a trans-membrane protein found in blood
plasma, and an unannotated gene). FOXP1, a gene familiar to my lab that encodes a
forkhead box family transcription factor that is an essential regulator of B cell
development, is one of the sixty-three total genes that are up-regulated after the
addition of NKL-30. RAS GTPase-activating proteins, p53-induced proteins, and
many proteins associated with the matrix and cytoskeleton account for the bulk of
the observed up-regulated genes.

111

Figure 4.3. Metabolic analysis of up-regulated genes in Nalm6 leukemia cells
treated with NKL-30 2μM for 24 hours.
112
Nalm6 cells treated for 24 hours with DMSO or NKL- 3 0 at 2 μ M (N 6 2 4 D -N624NKL)
showed an up-regulation of genes associated with cellular movement, connective
tissue development and function, and cellular assembly and organization networks
(Figure 4.3). Up-regulated genes include cytochrome genes involved in drug
metabolism, MAP kinases, calcium channels, proteases, stress-related chaperones
and genes involved with development, differentiation, and death. The down-
regulated genes are associated with cellular, tissue, and organismal development, as
well as hematopoiesis.
Three of the five annotated down-regulated genes (IGLL1, MYC, and DNTT)
are directly regulated by the transcription factor IKZF1 (Ikaros), which is involved
in chromatin remodeling and B cell differentiation, amongst other processes (Figure
4.4). IGLL1 codes for a pre-B cell receptor that is found on the surface of pro-B and
pre-B cells, where it is involved in transduction of signals for cellular proliferation.
The MYC gene encodes a transcription factor that plays a role in cell cycle
progression, apoptosis and cellular transformation. Mutations, overexpression,
rearrangement and translocation the MYC gene have been associated with a variety
of hematopoietic tumors, leukemias and lymphomas. DNTT is a template-
independent DNA polymerase that is expressed in a restricted population of normal
and malignant pre-B and pre-T lymphocytes during early differentiation, alluding to
the theory that NKL-30 induces differentiation as part of its mechanism that leads to
cell death. Another down-regulated gene, ZNF415, normally suppresses the
transcriptional activity of AP-1 and p53/TP53, both of which are involved in tumor
suppression and apoptosis.
113

Figure 4.4. Ikaros interaction with down-regulated genes. The transcription
factor Ikaros (IKZF1) regulates three of the five annotated genes that are down-
regulated in Nalm6 cells treated with NKL-30 at 2 μM for 24 hours.

114
Nalm6 cells treated for 48 hours with DMSO or NKL- 3 0 at 2 μ M (N 6 4 8 D -
N648NKL) showed an up-regulation of genes associated with cancer, the cell cycle,
and cellular growth and proliferation networks. The four annotated genes that were
found to be down-regulated (COL21A1, SERPINB10/SERPINB2, IGLL1, and, DOCK9-
AS2 [anti-sense RNA]) are associated with diseases such as those of the
gastrointestinal tract, connective tissue, and the skin. SERPINB10 is a protease
inhibitor that may play a role in the regulation of protease activities during
hematopoiesis and apoptosis induced by TNF (tumor necrosis factor). Up-regulated
genes found in the N648D-N648NKL comparison code for proteins that may
regulate apoptosis, cell proliferation, and differentiation, as well as those associated
with the regulation of pre-mRNA splicing and editing. Of the 86 total genes that
were observed to be differentially expressed in this comparison, only 17% (15/86)
are annotated, offering little information on what is occurring in the cell, but
allowing for a great opportunity to investigate novel genes.
Comparing the expression levels of protein coding genes in healthy B cell and
pre-B leukemia cell lines is somewhat of a task, however, it is great to look at this
through general characteristics. Gene sets that have lower expression levels in
Nalm6 cells compared with GM12878 are associated with networks of cellular
movement, hematological system development and function, and tissue morphology
(Figure 4.5). The genes include those related to B cell development, targets of TNF
and retinoic acid, drug metabolism, signal transduction, and multiple interferon
regulatory factors, amongst others. In fact, the three largest expression differences
115

Figure 4.5. Metabolic analysis of genes with lower expression in Nalm6
leukemia cells compared with healthy GM12878 B cells.
differentiation antigen protein, and a protein involved in the Cullin3-based E3

116
in the negative direction toward the leukemia cells is for a long non-coding RNA
(lncRNA) involved in chromosome X-inactivation, a myeloid cell nuclear
ubiquitin ligase complex. So, as can be seen, genes across the board have been
observed to be differentially expressed between these two cell lines, making it
extremely hard to pinpoint the causes that result in the change from healthy to
leukemic B cell.
Genes that are expressed higher in Nalm6 cells than GM12878 cells are
associated with a network of cancer, hematological disease, and cardiovascular
development and function (Figure 4.6). Several other genes with higher expression
in Nalm6 cells fall under such networks as cellular development, hematological
system development and function, and cell-mediated immune responses (Figure
4.7). As can be seen in the figures, most of the genes are colored in grey, meaning
that these genes are differentially expressed according to my experimental results.
This is contrary to many of the previous figures, where the majority of genes listed
in the pathways are white, meaning that they were not found to be differentially
expressed. This gives more confidence in the analysis of the data, and it can be
owed, in part, to the sheer number of genes found to be differentially expressed.
Interestingly, the two networks of genes show some cross-talk and share
either direct or indirect interactions between each other (Figure 4.8). This cross-
talk between pathways or networks of gene expression is shown via the orange
coloring of the lines in the figure, noting either direct (solid) or indirect (dotted)
interactions. While this data is intriguing, this alone cannot show the difference
between a healthy cell and one that is classified as leukemia. Several other cell lines,
117

Figure 4.6. Network 1 of genes with higher expression in Nalm6 leukemia
cells compared with healthy GM12878 B cells. The network involved in cancer,
hematological disease, and cardiovascular development and function shown.
118

Figure 4.7. Network 2 of genes with higher expression in Nalm6 leukemia
cells compared with healthy GM12878 B cells. The network involved in cellular
development, hematological system development and function, and cell-mediated
immune responses shown.
119
or primary cells, of each healthy and leukemia cells would need to be investigated to
get a clearer picture of the common gene expression signatures adopted by each
group. Until then, this can serve as preliminary data that shows that while there is
an observable difference in overall gene expression, we cannot point out what
exactly causes a general B cell to make the transition into a leukemia cell.

4.3 Discussion

NKL-30 treatment usually results in in an up-regulation of gene expression.
Between GM24D and GM24NKL 97% (63/65) of genes with a significant expression
change were up-regulated. Between N624D and N624NKL, that number was 80%
(24/30). And between N648D and N648NKL, the change of 64% (55/86) of total
significant gene expression was in a positive direction. Regardless of whether the
cell was healthy or cancerous, NKL-30 seemed to have an activation effect on gene
expression, possibly by relieving normal gene repression on MEF2 and other
transcription factors.
Focusing on the comparison between Nalm6 cells treated for 48 hours with
DMSO or NKL-30 at 2 μM, it is interesting to note that forty of the fifty-five (73%)
total genes that are up-regulated and twenty-seven of thirty-one (87%) are
unannotated in the most recent Ensembl human genome annotation release 71 from
April 2013. This is intriguing to think that our MEF2m is causing the expression of
genes that are either extremely rare or completely absent from previously examined
120

Figure 4.8. Cross-talk of networks in a comparison of GM12878 and Nalm6
gene expression. Merge of figures 4.6 and 4.7 from GM24D-N624D of more highly
expressed genes. Orange lines represent interactions between the different
networks. Grey lines represent interactions within each network.

121
human genomes. Many new genes can be characterized using this mRNA-seq data,
and this may shed more light on the mechanism of the benzamide class of MEF2m.
Comparison of the GM12878 and Nalm6 cell lines yields 350 total genes that
are significantly differentially expressed. Fifty-three percent (185/350) of these
genes are expressed at a higher value in Nalm6 leukemia cells than the healthy
GM12878 cells. This does not necessarily mean this rings true in all leukemia cells,
so more sequencing data covering multiple leukemia and healthy B cells needs to be
evaluated. Also, interestingly, MEF2C is expressed 8-fold lower in these Nalm6 cells
than in the GM12878 cells. This is speculation, but could lend some evidence to the
theory that HDACs can saturate the MEF2 proteins in a leukemia cell, and therefore
the addition of our MEF2m can correct that bias and allow the cell to go back into a
more natural state, with a better MEF2:HDAC ratio.
Another interesting po i nt to not e i s the fee li ng o f “ cor r ect ion” that ou r MEF2m seems to make to bring a leukemia cells gene signature back to normal.
This is a small sample set, but two examples stood out to me during the analysis of
this data. The expression of DNTT is undetected in GM12878 cells, and expressed
extremely highly in Nalm6 cells. After treatment of Nalm6 cells with NKL-30 for 24
hours, DNTT expression decreased by 3- f old . T he other example of th is “ cor r ect ion” is the gene LGALS1, which codes for a galectin that is involved in differentiation and
apoptosis, that is expressed almost 7-fold lower in Nalm6 than GM12878 cells.
However, after 48 hours of treatment with our MEF2m, Nalm6 cells experience a 3-
fold increase in LGALS1 expression. While these observations are not a part of the
majority, it is intriguing data that can lend guidance to future experimental design.
122
There are, however, drawbacks and limitations to RNA-seq experiments in
general, and mine specifically. To start, of the twelve possible relevant pairwise
analyses from my eight samples, three of the analyses did not yield any significant
gene changes: Nalm6 cells treated with NKL-30 at 2 μM for 24 hours versus 48 hours
(N624NKL-N648NKL), the same cells treated with DMSO for 24 hours versus 48
hours (N624D-N648D), and GM12878 cells treated with either DMSO or NKL-30 at
2 μM for 48 hours (GM48D-GM48NKL). One possible reason for this is that I ran all
of my experiments in singlet, without any replicates. I ran eight human samples on
a single lane; therefore I do not have the total sequencing depth that is desired. I
also only ran one sequencing experiment. This experiment should be completed
again, with each sample having both a biological and technical replicate, as well as
using multiple sequencing lanes so that depth of coverage increases. This will help
increase the number of significant gene expression calls, as well as decrease error
bars, and increase confidence in the observed data.
Genes coding for all types of molecules, and those found in all locations of the
cell, can be seen to be changed after drug treatment by NKL-30 in both cell lines
tested. This could be due to the fact that, although we believe that the main
interaction is with MEF2, these small molecule compounds interact with other
molecules throughout the cell and cause signaling pathways to be altered. The goal
to achieve 100% specificity of a man-made molecule for MEF2 family members is
unreasonable, so we expect to see off-target effects, however, the higher the
specificity the better for drug development. This is a reason why our lab has spent
so much time on developing more and more derivatives of the original BML-210
123
compound, and why we believe we have a leg up on many of the commercially
available HDAC inhibitors.

4.4 Materials and Methods

Library Preparation
Nalm6 or GM12878 cells were cultured in T75 flasks as mentioned in the previous
chapter and 1x10
7
cells were treated with either DMSO, 1µM NKL-30 or 2µM NKL-
30 for either 24 or 48 hours. Cells were harvested, and washed with PBS before
total RNA was extracted using a Direct-zol RNA miniprep kit (Zymo Research)
accor d in g to man uf act u r er’ s p r otocol . mR NA li b r aries wer e p r ep ared f or I llu mina sequencing using previously published methods (Chang et al., 2011). Only one
deviation from the protocol was made, by using AmPure XP beads for size selection
instead of agarose gel electrophoresis. Samples were loaded on an Illumina HiSeq
Genetic Analyzer in a 50-cycle paired-end (50PE) run.

Gene expression analysis
Bowtie2, TopHat2, Cufflinks, Cuffmerge, and Cuffdiff were on a CentOS Linux cluster.
R (CummeRbund) and Ingenuity IPA were used for graphical analysis and altered
pathway generation.

124

Chapter 5
Conclusion

Since 1989, when MEF2 was first discovered as a regulator of gene
expression in vertebrate muscle cells, scientists have been studying the role of MEF2
in development, adaptive responses, and survival/apoptosis. The first physical
structures of MEF2 were available eleven years later in 2000, and this set forth the
groundwork for a good chunk of the research focus in my P.I., Dr. Lin Chen ’s
laboratory. Crystal structures of Cabin1 bound to MEF2B (2003) and HDAC9 bound
to MEF2B (2005) solved by Dr. Aidong Han, a previous member of my lab, were
accepted by the scientific community based upon extremely accurate and
convincing X-ray diffraction data, analyses, and structural modeling. In these
studies, mutational analyses were performed to pinpoint important residues on
MEF2B, Cabin1 and HDAC4 proteins at their binding interface. These mutations
were able to decrease or ablate the protein binding altogether. However, many
people still thought that the hydrophobic binding pocket of MEF2, which interacts
with the α helices of Cabin1 and the class IIa HDACs, only formed with the
125
interaction of a binding partner. Our lab believed that the binding pocket was
always present as seen in the cofactor-bound structures, so we set out to solve the
stru ctur e of ME F 2 A w it hou t a b in d in g p artne r , called the “ ap o” stru ctur e. M E F 2 A can be used without hesitance here because the binding pocket is in N-terminal
region that is highly conserved in all MEF2 proteins.
After solving the crystal structure and proving that the MEF2 binding pocket
was not only present while bound to another protein, nor was a result of crystal
packing, our lab moved forward to investigate whether the MEF2:HDAC interaction
could be affected by small molecules for the purpose of treating cardiac
hypertrophy. While performing literature research and virtual screening of the
ZINC database of commercially available compounds, it was found that a small
molecule compound in the pimeloylanilide o-aminoanilide (PAOA) class known as
N-(2-Aminophenyl)- N’ p henyl octan ol d ia mine mig ht be ab le t o bi nd M E F 2 in the
same hydrophobic pocket necessary for Cabin1 and class IIa HDAC binding. This 2-
amino benzamide molecule, commercially known as BML-210, was then found to be
effective in decreasing the viability of leukemia cells.
This, in turn, led our lab to investigate whether its effects were due to its
interaction with MEF2. A former graduate student in our lab, Dr. Nimanthi
Jayathilaka, was able to solve the crystal structure of BML-210 bound in the
hydrophobic pocket of MEF2. This then lead to our lab, in collaboration with a lab
run by Dr. Nicos Petasis, to develop derivatives to this drug that would lead to more
eff ect ive an d sp eci f ic b i nd in g to M E F 2 . T hr oug h Nim an thi’s w or k , she sho w ed that these compounds do, in fact, negatively affect the binding of class IIa HDACs to
126
MEF2, and that it is dose-dependent and specific to this protein-protein interaction
through the use of luciferase and surface plasmon resonance assays.
Following these studies, we decided to move into the MEF2:HDAC interaction
in leukemia after noticing that MEF2 is often dysregulated in many different types of
leukemia. Often times MEF2 is overexpressed, and HDACs have long been found to
be up-regulated in many cancers as well. HDAC inhibitors (HDACi) are used to treat
many cancers, however a majority of currently used HDACi target the catalytic
domain of the HDAC proteins, and many are not class-specific. The benzamide BML-
210, of which our lab has made many derivatives we titled NKL, does not seem to
target the HDAC protein itself, but rather the MEF2:HDAC interaction. These
molecules bind in the same hydrophobic pocket of MEF2 proteins as class IIa HDACs
and therefore disallow the HDAC protein to bind to MEF2, which in turn relieves the
repression on the coinciding MEF2 target gene. Through cell culture experiments, it
is clear that the addition of the NKL drugs decreases the viability of many types of B
cells. However, the intriguing part is that the drugs seem to have a preference
toward the killing of leukemia cells. These drugs, with a range of efficacy based
upon slight chemical or structural deviations, greatly decrease the viability of
leukemia cells, while only slightly affecting healthy control B cells in culture.
After accumulating this data, a research scientist in our lab, Dr. Meng Xia, was
able to establish a mammalian model in nude mice using Nalm6 leukemia cells. The
Nalm6 leukemia cells were selected based on the extreme effect of NKL-30 I saw on
Nalm6 cells while performing cellular viability experiments in culture. After several
months of work to establish the Nalm6 mouse model, Dr. Xia was able to start
127
testing the effect of NKL-30 drug treatment on these mice. Her results indicated that
the nude mice that showed signs of leukemia and were treated with NKL-30 lived
much longer those treated with the vehicle alone. This then lead to more in vitro
experiments to help us understand the mechanism of action of these NKL drugs.
Our main theory behind the mechanism of the NKL drugs is that they bind to
MEF2 and ablate the interaction of MEF2:class IIa HDACs. When the class IIa
HDAC:MEF2 interaction is broken, these HDACs become phosphorylated, and are
then actively shuttled to the cytoplasm after association with the 14-3-3 proteins.
To test this, I performed an experiment to separate the nucleus from the cytoplasm
of a cell, called a nuclear fractionation. After treating Nalm6 and RS4;11 pre-B
leukemia cells with either DMSO or NKL-30, cells were harvested and separated into
their nuclear and cytoplasmic components. These samples were then run on an SDS
gel and tested for phosphorylated HDAC4 protein levels. After treatment with
NKL-30 for 24 hours, an increase in phosphorylated HDAC4 protein was observed in
the cytoplasm compared to DMSO treated samples, supporting our theory.
To further implicate the action of NKL-30 on MEF2 proteins, I checked the
expression of a well-known MEF2 target gene involved in apoptosis, NR4A1/Nur77.
Normally this gene is not expressed, or expressed at very low levels in the cell. This
expression is held in check by the action of HDACs keeping MEF2 proteins
deacetylated at the NR4A1/Nur77 gene promoter. I found that treating leukemia
cells with NKL-30 for 24 hours resulted in the increase of protein expression of
NR4A1/Nur77 in a dose-dependent manner. The increase in protein expression is
likely a result of NKL-30 breaking the MEF2:class IIa HDAC interaction, which
128
culminates in the derepression of MEF2 on the NR4A1/Nur77 promoter. This is one
possible reason that treating cells with our MEF2m NKL-30 causes a decrease in
viability.
Finally, I performed an mRNA-seq experiment in order to see any global gene
expression or transcript splicing variation in NKL-30 treated samples compared to
control samples, both in a control B-cell line (GM12878) and a pre-B cell leukemia
line (Nalm6). While the experiments need to be reproduced for further confidence
and comparisons, the preliminary data suggests that our MEF2m often results in the
up-regulation of gene expression. Also, the treatment seems to express many genes
in the Nalm6 cell line that are unaccounted for in current annotated genome
assemblies provided by Ensembl, leading to the intriguing thought that we have
only seen the tip of the iceberg of possible data and analyses.

129
BIBLIOGRAPHY
Allfrey VG, Faulkner R, and Mirsky AE. Acetylation and methylation of histones and
their possible role in the regulation of RNA synthesis. Proc. Natl Acad. Sci. 1964;51:
786–794.

Bolden JE, Peart MJ, and Johnstone RW. Anticancer activities of histone deacetylase
inhibitors. Nat Rev Drug Discov. 2006;5:769–84.

Borutinskaite VV, Navakauskiene R, and Magnusson KE. Retinoic Acid and Histone
Deacetylase Inhibitor BML-210 Inhibit Proliferation of Human Cervical Cancer HeLa
Cells. Ann. N.Y. Acad. Sci. 2006;1091:346–355.

Borutinskaite VV, Savickiene J, and Navakauskiene R. Apoptotic effects of the novel
histone deacetylase inhibitor BML-210 on HeLa cells. BIOLOGIJA. 2008;54:217–220.

Bour BA, O'Brien MA, Lockwood WL, Goldstein ES, Bodmer R and Taghert PH.
Drosophila MEF2, a transcription factor that is essential for myogenesis. Genes Dev.
1995;9:730–741.

Butler KV and Kozikowski AP. Chemical origins of isoform selectivity in histone
deacetylase inhibitors. Curr. Pharm. Des. 2008;14:505–528.

Campuzano V, Montermini L, Moltò MD, Pianese L, Cossée M, Cavalcanti F, Monros E,
and Pandolfo, M. Friedreich's ataxia: autosomal recessive disease caused by an
intronic GAA triplet repeat expansion. Science 1996;271:1423-1427.

Chan JK, Sun L, Yang XJ, Zhu G, and Wu Z. Functional characterization of an amino-
terminal region of HDAC4 that possesses MEF2 binding and transcriptional repressive
activity. J. Biol. Chem. 2003;278:23515-23521.

Chang PL, Dunham JP, Nuzhdin SV, and Arbeitman MN. Somatic sex-specific
transcriptome differences in Drosophila revealed by whole transcriptome sequencing.
BMC Genomics 2011;12:364.

Chen L, Glover JN, Hogan PG, Rao A, and Harrison SC. Structure of the DNA-binding
domains from NFAT, Fos and Jun bound specifically to DNA. Nature 1998;392:42-48.

Chen SL, Wang SCM, Hosking B, and Muscat, GE. Subcellular localization of the
steroid receptor coactivators (SRCs) and MEF2 in muscle and rhabdomyosarcoma
cells. Molecular Endocrinology 2001;15:783-796.

Chuang DM, Leng Y, Marinova Z, Kim HJ, and Chiu CT. Multiple roles of HDAC
inhibition in neurodegenerative conditions. Trends Neurosci. 2009;32:591-601.

130
Clocchiatti A, Di Giorgio E, Demarchi F, and Brancolini C. Beside the MEF2 axis:
Unconventional functions of HDAC4. Cellular Signalling 2013;25: 269–276.
De Angelis L, Borghi S, Melchionna R, Berghella L, Baccarani-Contri M, Parise F.,
Ferrari S, and Cossu G. Inhibition of myogenesis by transforming growth factor β is
density-dependent and related to the translocation of transcription factor MEF2 to the
cytoplasm. Proceedings of the National Academy of Sciences 1998;95:12358-12363.

de Ruijter AJ, van Gennip AH, Caron HN, Kemp S, and van Kuilenburg AB. Histone
deacetylases (HDACs): characterization of the classical HDAC family. J. Biochem.
2003;370:737-749.

Dodou E and Treisman R. The Saccharomyces cerevisiae MADS-box transcription
factor Rlm1 is a target for the Mpk1 mitogen-activated protein kinase pathway. Mol.
Cell. Biol. 1997;17:1848–1859.

Du Y, Spence SE, Jenkins NA, and Copeland NG. Cooperating cancer-gene
identification through oncogenic-retrovirus-induced insertional mutatgenesis. Blood
2005;106: 2498-2505.

Finnin MS, Donigian JR, Cohen A, Richon VM, Rifkind RA, Marks PA, Breslow R, and
Pavletich NP. Structures of a histone deacetylase homologue bound to the TSA and
SAHA inhibitors. Nature 1999;401:188-193.

Fischle W, Emiliani S, Hendzel MJ, Nagase T, Nomure N, Voelter W, and Verdin E. A
new family of human histone deacetylases related to Saccharomyces cerevisiae HDA1p
J. Biol. Chem., 1999;274:11713–11720.

Flavell SW, Kim TK, Gray JM, Harmin DA, Hemberg M, Hong EJ, Markenscoff-
Papadimitriou E, and Greenberg ME. Genome-wide analysis of MEF2 transcriptional
program reveals synaptic target genes and neuronal activity-dependent
polyadenylation site selection. Neuron, 2008;60:1022-1038.

Gregoire S, Tremblay AM, Xiao L, Yang Q, Ma K, Nie, J, Mao Z, Wu Z, Giguere V, and
Yang, XJ. Control of MEF2 transcriptional activity by coordinated phosphorylation and
sumoylation. J. Biol. Chem. 2006;,281:4423-4433.

Gregoire S and Yang XJ. Association with Class IIa histone Deacetylases upregulates
the sumoylation of MEF2 transcription factors. Mol. Cell. Biol. 2005;25:2273-2287.

Gossett LA, Kelvin DJ, Sternberg EA and Olson EN. A new myocyte-specific enhancer-
binding factor that recognizes a conserved element associated with multiple muscle-
specific genes. Mol. Cell. Biol. 1989;9:5022–5033.

131
Grozinger CM, Hassig CA, and Schreiber SL. Three proteins define a class of human
histone deacetylases related to yeast Hda1p Proc. Natl. Acad. Sci. U. S. A.,
1999;96:4868–4873.
Grozinger CM and Schreiber SL. Regulation of histone deacetylase 4 and 5 and
transcriptional activity by 14-3-3-dependent cellular localization. PNAS 2000;97:
7835-7840.

Han A, He J, Wu Y, Liu JO, and Chen L. Mechanism of recruitment of class II histone
deacetylases by myocyte enhancer factor-2. J. Mol. Biol. 2005;345:91-102.

Han A, Pan F, Stroud JC, Youn HD, Liu JO, and Chen L. Sequence-specific recruitment
of transcriptional co-repressor Cabin1 by myocyte enhancer factor-2. Nature
2003;422:730-734.

He J, Ye J, Cai Y, Riquelme C, Liu JO, Liu X, Han A, and Chen L. Structure of p300 bound
to MEF2 on DNA reveals a mechanism of enhanceosome assembly. Nucleic Acids Res.
2011;39:4464-4474.

Herman D, Jenssen K, Burnett R, Soragni E, Perlman SL, and Gottesfeld, JM. Histone
deacetylase inhibitors reverse gene silencing in Friedreich's ataxia. Nature chemical
biology, 2006;2:551-558.

Huang K, Louis JM, Donaldson L, Lim F, Sharrocks AD, and Clore GM. Solution
structure of the MEF2A–DNA complex: structural basis for the modulation of DNA
bending and specificity by MADS-box transcription factors. EMBO J. 2000;19:2615-
2628.
Humeniuk R, Mishra PJ, Bertino JR, and Banerjee D. Molecular targets for epigenetic
therapy of cancer. Curr. Pharm. Biotechnol., 2009;10:161–165.

Ishikawa F, Miyoshi H, Nose K, and Shibanuma M. Transcriptional induction of MMP-
10 by TGF-b, mediated by activation of MEF2A and downregulation of class IIa HDACs.
Oncogene, 2010;29:909–919.

Jayathilaka N, Han A, Gaffney KJ, Dey R, Jarusiewicz JA, Noridomi K, Philips MA, Lei X,
He J, Ye J, Gao T, Petasis NA, and Chen L. Inhibition of the function of class IIa HDACs
by blocking their interaction with MEF2. Nucleic Acids Research. 2012;40:5378–5388.

Kim Y, Phan D, van Rooij E, Wang DZ, McAnally J and Qi X. The MEF2D transcription
factor mediates stress-dependent cardiac remodeling in mice. J. Clin. Invest. 2008;
118:124–132.

Kouzarides T. Chromatin modifications and their function. Cell. 2007;128:693-705.

132
Krivtsov AV, Twomey D, Feng Z, Stubbs MC, Wang Y, Faber J, Levine JE, Wang J, Hahn
WC, Gilliland DG, Golub TR, and Armstrong SA. Transformation from committed
progenitor to leukaemia stem cell initiated by MLL–AF9. Nature Letters 2006;442:818-
822.

Lewis, Jr. TL, Mao T, Svoboda K, and Arnold DB. Myosin-dependent targeting of
transmembrane proteins to neuronal dendrites. Nat. Neurosci. 2009;12:568-576.

Lund AH, Turner G, Trubetskoy A, Verhoeven E, Wientjens E, Hulsman D, Russell R,
DePinho RA, Lenz J, and van Lohuizen M. Genome-wide retroviral insertional tagging
of genes involved in cancer in Cdkn2a-deficient mice. Nat. Genet. 2002;32:160-165.
Mao Z, Bonni A, Xia F, Nadal-Vicens M and Greenberg ME. Neuronal activity-
dependent cell survival mediated by transcription factor MEF2. Science 1999;
286:785–790.
Marks PA, Richon VM, and Rifkind RA. Histone Deacetylase Inhibitors: Inducers of
Differentiation or Apoptosis of Transformed Cells. J. Natl Cancer Inst 2000;92:1210-
1216.
McKinsey TA, Zhang CL, and Olson EN. Activation of the myocyte enhancer factor-2
transcription factor by calcium/calmodulin dependent protein kinase-stimulated
binding of 14-3-3 to histone deacetylase 5. PNAS 2000;97:14400-14405.
McKinsey TA, Zhang CL, and Olson EN. Identification of a signal-responsive nuclear
export sequence in class II histone deacetylases. Mol. Cell. Biol. 2001;21:6312-6321.
Minucci S, and Pelicci PG. Histone deacetylase inhibitors and the promise of epigenetic
(and more) treatments for cancer. Nat Rev Cancer 2006;6:38–51.

Miska EA, Karlsson C, Langle E, Nielsen SJ, Pines J, and Kouzarides T. HDAC4
deacetylase associates with and represses the MEF2 transcription factor. EMBO J.
1999;18:5099-5107.

Molkentin JD, Black BL, Martin JF, and Olson EN. Mutational analysis of the DNA
binding, dimerization, and transcriptional activation domains of MEF2C. Mol. Cell Biol.
1996;16:2627-2636.

Mullican SE, Zhang S, Konopleva M, Ruvolo V, Andreeff M, Milbrandt J, and Conneely
OM. Abrogation of nuclear receptors Nr4a3 andNr4a1 leads to development of acute
myeloid leukemia. Nat. Med. 2007;13:730-735.

Mullighan CG, Zhang J, Kasper LH, Lerach S, Payne-Turner D, Phillips LA, Heatley SL,
Holmfeldt L, Collins-Underwood JR, and Ma J. CREBBP mutations in relapsed acute
lymphoblastic leukaemia. Nature 2011;471:235-239.

133
Nagel S, Meyer C, Quentmeier H, Kaufmann M, Drexler HG, and MacLeod RAF.
MEF2C is activated by multiple mechanisms in a subset of T-acute lymphoblastic
leukemia cell lines. Leukemia 2008;22:600–607.
Nurrish SJ, and Treisman R. DNA binding specificity determinants in MADS-box
transcription factors. Mol. Cell Biol. 1995;15:4076-4085.
Pan F, Ye Z, Cheng L, and Liu JO. Myocyte enhancer factor 2 mediates calcium-
dependent transcription of the interleukin-2 gene in T lymphocytes: a calcium
signaling module that is distinct from but collaborates with the nuclear factor of
activated T cells (NFAT). J. Biol. Chem. 2004;279:14477–14480.
Pandolfo, M. "Friedreich ataxia." In Seminars in pediatric neurology. WB Saunders,
2003;10:163-172.
Paris M, Porcelloni M, Binaschi M, and Fattori D. Histone deacetylase inhibitors: from
bench to clinic. J. Med. Chem., 2008;51:1505-1529.
Potthoff MJ and Olson EN. MEF2: a central regulator of diverse developmental
programs. Development 2007;134:4131–4140.
Prima V, Gore L, Caires A, Boomer T, Yoshinari M, Imaizumi M, Varella-Garcia M, and
Hunger SP. Cloning and functional characterization of MEF2D/DAZAP1 and
DAZAP1/MEF2D fusion proteins created by a variant t(1;19)(q23;p13.3) in acute
lymphoblastic leukemia. Leukemia 2005;19:806-813.
Prima V and Hunger SP. Cooperative transformation by MEF2D/DAZAP1 and
DAZAP1/MEF2D fusion proteins generated by the variant t(1;19) in acute
lymphoblastic leukemia. Leukemia 2007;21:2470-2475.

Rai M, Soragni E, Jenssen K, Burnett R, Herman D, Coppola G, Geschwind DH,
Gottesfeld JM, and Pandolfo M. HDAC inhibitors correct frataxin deficiency in a
Friedreich ataxia mouse model. PLoS One 2008;3:1958.

Romanski A, Bacic B, Bug G, Pfeifer H, Gul H, Remiszewski S, Hoelzer D, Atadja P,
Ruthardt M, and Ottmann OG. Use of a novel histone deacetylase inhibitor to induce
apoptosis in cell lines of acute lymphoblastic leukemia. Haematologica 2004;89:419-
426.

Sandmann T, Jensen LJ, Jakobsen JS, Karzynski MM, Eichenlaub MP, Bork P and
Furlong EE. A temporal map of transcription factor activity: Mef2 directly regulates
target genes at all stages of muscle development. Dev. Cell . 2006;10:797–807.
Santelli E, and Richmond TJ. Crystal structure of MEF2A core bound to DNA at 1.5 Å
resolution. J Mol Biol. 2000;297:437-449.

134
Sasaki K, Yamagata T, Mitani K. Histone deacetylase inhibitors trichostatin A and
valproic acid circumvent apoptosis in human leukemic cells expressing the RUNX1
chimera. Cancer Sci. 2008;99:414-422.

Savickiene J, Borutinskaite VV, Treigyte G, Magnusson KE, and Navakauskiene R. The
novel histone deacetylase inhibitor BML-210 exerts growth inhibitory, proapoptotic
and differentiation stimulating effects on the human leukemia cell lines. Eur. J.
Pharmacol. 2006;549:9-18.

Schwieger M, Schüler A, Forster M, Engelmann A, Arnold MA, Delwel R, Valk PJ,
Löhler J, Slany RK, Olson EN, and Stocking C. Homing and invasiveness of MLL/ENL
leukemic cells is regulated by MEF2C. Blood. 2009;114:2476-2488.

Shalizi A, Gaudilliere B, Yuan Z, Stegmuller J, Shirogane T and Ge Q. A calcium-
regulated MEF2 sumoylation switch controls postsynaptic differentiation. Science.
2006;311:1012–1017.
Shin HH, Seoh JY, Chung HY, Choi SJ, Hahn MJ, Kang JS, Choi MS, Han TH.
Requirement of MEF2D in the induced differentiation of HL60 promyeloid cells. Mol
Immunol. 1999;36:1209-1214.

Shore P and Sharrocks AD. The MADS-box family of transcription factors. Eur. J.
Biochem. 1995;229:1–13.
Soragni E, Xu C, Plasterer HL, Jacques V, Rusche JR, and Gottesfeld JM. Rationale for
the development of 2-aminobenzamide histone deacetylase inhibitors as therapeutics
for Friedreich Ataxia. Journal of Child Neurology, 2012;27:1164-1173.

Sparrow DB, Miska EA, Langley E, Reynaud-Deonauth S, Kotecha S, Towers N, Spohr
G, Kouzarides T and Mohun TJ. MEF-2 function is modified by a novel co-repressor,
MITR. EMBO J. 1999;18:5085-5098.

Stehling-Sun S, Dade J, Nutt SL, DeKoter RP, and Camargo FD. Regulation of lymphoid
versus myeloid fate ‘choice’ by the transcription factor Mef2c. Nature Immunology.
2009;10:289-296.

Suzuki T, Shen H, Akagi K, Morse III HC, Malley JD, Naiman DQ, Jenkins NA, and
Copeland NG. New genes involved in cancer identified by retroviral tagging. Nat.
Genet. 2002;32: 166-174.
Swanson BJ, Jack HM, and Lyons GE. Characterization of Myocyte enhancer factor 2
(MEF2) expression in B and T cells: MEF2C is a B cell-restricted transcription factor in
lymphocytes. Mol. Immunol. 1998;35:445-458.

135
Thomas EA, Coppola G, Desplats PA, Tang B, Soragni E, Burnett R, Gao F, and
Gottesfeld JM. The HDAC inhibitor 4b ameliorates the disease phenotype and
transcriptional abnormalities in Huntington's disease transgenic mice. Proceedings of
the National Academy of Sciences, 2008;105:15564-15569.

Tsapis M, Lieb M, Manzo F, Shankaranarayanan P, Herbrecht R, Lutz P, and
Gronemeyer H. HDAC inhibitors induce apoptosis in glucocorticoid-resistant acute
lymphatic leukemia cells despite a switch from the extrinsic to the intrinsic death
pathway. The Int’l J. of Biochem and Cell Biol. 2007; 39:1500-1509.
Ververis K, Hiong A, Karagiannis TC, and Licciardi PV. Histone deacetylase inhibitors
(HDACIs): multitargeted anticancer agents. Biologics 2013;7:47-60.
Ververis K and Karagiannis TC. Overview of the classical histone deacetylase enzymes
and histone deacetylase inhibitors. ISRN Cell Biology 2012;2012:1-12.

Wang AH, Kruhlak MJ, Wu J, Bertos NR, Vezmar M, Posner BI, Bazett-Jones DP, and
Yang XJ. HDAC4, a human histone deacetylase related to yeast HDA1, is a
transcriptional co-repressor Mol. Cell. Biol., 1999;19:7816–7827.
Wang AH, Kruhlak MJ, Wu J, Bertos NR, Vezmar M, Posner BI, Bazett-Jones DP, and
Yang XJ. Regulation of histone deacetylase 4 by binding of 14-3-3 proteins. Mol. Cell
Biol. 2000;20:6904-6912.
Wang AH and Yang XJ. Histone deacetylase 4 possesses intrinsic nuclear import and
export signals. Mol. Cell. Biol. 2001;21:5992-6005.
West AG, Shore P, and Sharrocks AD. DNA binding by MADS-box transcription factors:
a molecular mechanism for differential DNA bending. Mol. Cell Biol. 1997;17:2876-
2887.
Wu S, Trievel RC, and Rice JC. Human SFMBT is a transcriptional repressor protein
that selectively binds the N-terminal tail of histone H3. FEBS letters. 2007;581:3289-
3296.

Wu Y, Dey R, Han A, Jayathilaka N, Philips M, Ye J, and Chen L. Structure of the MADS-
box/MEF2 domain of MEF2A bound to DNA and its implication for myocardin
recruitment. J. Mol. Biol. 2010;397:520-533.

Xu C, Soragni E, Chou CJ, Herman D, Plasterer HL, Rusche JR, and Gottesfeld JM.
Chemical probes identify a role for histone deacetylase 3 in Friedreich's ataxia gene
silencing. Chem. Biol. 2009;16:980-989.

Yang XJ and Seto E. The Rpd3/Hda1 family of lysine deacetylases: from bacteria and
yeast to mice and men. Nat. Rev. Mol. Cell Biol. 2008;9:206–218.

136
Yang Q, She H, Gearing M, Colla E, Lee M, Shacka JJ and Mao Z. Regulation of neuronal
survival factor MEF2D by chaperone-mediated autophagy. Science. 2009;323:124–
127.

Youn HD and Liu JO. Cabin1 represses MEF2-dependent Nur77 expression and T cell
apoptosis by controlling association of histone Deacetylases and acetylases with MEF2.
Immunity 2000;13:85-94.

Yuki Y, Imoto I, Imaizumi M, Hibi S, Kaneko Y, Amagasa T, and Inazawa J.
Identification of a novel fusion gene in a pre-B acute lymphoblastic leukemia with
t(1;19)(q23;p13). Cancer Sci. 2004;95:503-507.

Zhao X, Sternsdorf T, Bolger TA, Evans RM, and Yao TP. Regulation of MEF2 by
histone deacetylase 4- and SIRT1 deacetylase-mediated lysine modifications. Mol. Cell.
Biol. 2005;25:8456-8464.
Harrison's Principles of Internal Medicine, 16th Edition, Chapter 97. Malignancies of
Lymphoid Cells. Clinical Features, Treatment, and Prognosis of Specific Lymphoid
Malignancies.

National Cancer Institute. http://RTCGD.ncifcrf.gov
MedicineNet, Inc. http://www.medicinenet.com
Medscape Reference: Drugs, Diseases & Procedures. http://emedicine.medscape.com
The Leukemia and Lymphoma Society. http://www.leukemia-lymphoma.org

Abstract (if available)

Abstract The Myocyte Enhancer Factor 2 (MEF2) family of transcription factors are DNA-bound proteins that regulate gene expression via their interaction with co-regulatory proteins. This interaction between MEF2 and its co-factors is necessary for the development and proper function of many mammalian cell types, and is a major factor in the onset and progression of many cancers, specifically leukemia. Understanding the MEF2:co-factor interaction in leukemia, which is often times an epigenetic event, is of utmost importance to those searching for therapies that have high specificity and low chance of side effects. ❧ In the first part of this paper, the structural characteristics of the MEF2 family that allow its interaction with co-regulators is examined. The study revealed that the structure of MEF2 proteins are not majorly altered by the binding of a co-regulator, such as class IIa histone deacetylases (HDACs), Cabin1, or p300, at the protein-protein interface, and are inherently able to be bound while interacting with DNA. This overruled previous notions that the hydrophobic groove, or "binding pocket", of MEF2 was not present unless it was interacting with a co-regulator. ❧ The second, and most involved, study of this paper examines the interaction of MEF2 and class IIa HDACs in leukemia. This study was aided by the availability of a small molecule benzamide compound NKL-30, which we call a MEF2 modulator (MEF2m), that was used as a tool to observe the consequences of inhibiting the protein-protein interaction both in vitro and in vivo. Our MEF2m were able to preferentially decrease the viability of leukemia cells over non-cancerous control B cells. We were also able to show that our compound was able to ablate the MEF2:class IIa HDAC interaction and result in the cytoplasmic shuttling of these co-repressors. Additionally, we were able to show that the expression of the protein NR4A1/Nur77, a MEF2 target gene associated with apoptosis, increased after treatment with NKL-30, likely because of the derepression of MEF2 afforded by the exodus of class IIa HDACs from the nucleus. Finally, we were able to show in a mammalian leukemia model utilizing Nalm6 cells, that mice treated with NKL-30 enjoy an increased lifespan compared to those receiving vehicle control injections. ❧ The third and final part of this work involves the gene expression analysis of B cells that have been treated with our MEF2m. Both healthy and leukemic B cells were treated with our small molecule compound or a solvent control and then data was collected and analyzed using mRNA-seq. We have shown that our MEF2m often preferentially activates gene expression, possibly due to the derepression seen on MEF2 and additional off-target effects. While these differentially regulated genes seen after drug treatment do not fit a single profile or pathway, we do know that the treatment of leukemia cells with our MEF2m usually results in death. Therefore, this data is crucial in our quest for understanding this phenomenon. This work is still ongoing and can be massaged for more data and relationships with the help of future replication and analyses. ❧ In summation, the work presented in this dissertation addresses the possibility of a small molecule compound to specifically target leukemia cells for death over healthy cells due to its ability to bind to MEF2 proteins, disallowing the binding of co-repressors, and resulting in an epigenetic alteration event. The mechanism and consequence of this event, while not yet completely understood, is detailed within the following chapters and the data suggest that we are on the right track toward understanding a way to develop targeted therapies for previously untreatable or high mortality-rate diseases.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The development of targeted transcription factor transposition and understanding chromatin dynamics in hypertrophic cardiomyopathy

PDF

Multiple functions of the PR-Set7 histone methyltransferase: from transcription to the cell cycle

PDF

The role of survivin in drug resistant pediatric acute lymphoblastic leukemia

PDF

Improved methods for the quantification of transcription factor binding using SELEX-seq

PDF

Forkhead transcription factors regulate replication origin firing through dimerization and cell cycle-dependent chromatin binding in S. cerevisiae

PDF

Genetics and the environment: evaluating the role of noncoding RNA in autism spectrum disorder

PDF

Role of integrin α4 in drug resistant acute lymphoblastic leukemia

PDF

Study of the role of bone morphogenetic proteins in prostate cancer progression

PDF

Modulation of transcription and receptor function with synthetic small molecules and multi-finctional integrin ligands

PDF

The transcriptional regulation of the pro-survival protein Grp78 by activating transcription factors and chromatin-modifying enzymes: its upregulation in response to HDAC-inhibitor treatment and ...

PDF

Exploration of the roles of cancer stem cells and survivin in the pathogenesis and progression of prostate cancer

PDF

The function of Rpd3 in balancing the replicaton initiation of different genomic regions

PDF

The role of fibroblast growth factor signaling on postnatal hepatic progenitor cell expansion

PDF

The role of adipocytes in acute lymphoblastic leukemia cell migration and survival against daunorubicin

PDF

The essential role of histone H2A deubiquitinase MYSM1 in natural killer cell maturation and HSC homeostasis

PDF

Novel roles for Maf1 in embryonic stem cell differentiation and adipogenesis

PDF

Placenta growth factor-miRNAs-lncRNAs axis in the regulation of ET-1 gene involved in pulmonary hypertension in sickle cell disease

PDF

Runx2 interactions with the osteoblast genome

PDF

The role of Ryk and Smek in neurogenesis; Mechanisms of CBP/β-catenin signaling inhibitor and IL-6 mediators in head and neck cancer

PDF

Glucocorticoids and Runx2 synergistically stimulate Wif1 and compromise pre-osteoblats in vitro

Asset Metadata

Creator Philips, Michael Albert (author)

Core Title Using novel small molecule modulators as a tool to elucidate the role of the Myocyte Enhancer Factor 2 (MEF2) family of transcription factors in leukemia

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Publication Date 08/22/2013

Defense Date 07/03/2013

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

Tag 2-aminobenzamide,acute lymphoblastic leukemia,B cells,BML-210,crystallography,epigenetics,Gleevec,GM12878,HDAC,HDACi,histone deacetylase,histone deacetylase inhibitor,Imatinib Mesylate,leukemia,MEF2,mRNA-seq,Myocyte Enhancer Factor 2,Nalm6,NKL-30,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Chen, Lin (committee chair), Frenkel, Baruch (committee member), Tower, John G. (committee member)

Creator Email micphilips@gmail.com,mphilips@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-323542

Unique identifier UC11294756

Identifier etd-PhilipsMic-2017.pdf (filename),usctheses-c3-323542 (legacy record id)

Legacy Identifier etd-PhilipsMic-2017.pdf

Dmrecord 323542

Document Type Dissertation

Format application/pdf (imt)

Rights Philips, Michael Albert

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