Role of bcl 2 in metabolic and redox regulation via its effects on cytochrome c oxidase and mitochondrial functions in tumor cells

1
1. Introduction and literature review

1.1 Early discovery of Bcl-2 as an oncogene:
Bcl-2, which stands for B-cell Lymphoma/Leukemia-2 gene, was first discovered in
B-cell malignancies more than twenty years ago (Tsujimoto, Cossman et al. 1985). It
was identified through a set of chromosomal translocations that resulted in its
activation in the majority of non-Hodgkin’s B-cell and follicular lymphomas. More
specifically, bcl-2 was found to translocate from its usual 18q21 chromosomal
location to 14q32, where it fuses with the promoter and enhancer of the
immunoglobulin heavy chain gene to result in its excessive and deregulated
expression (Cleary, Smith et al. 1986). Also, this became known as the t(14,18)
breakpoint.

In terms of function, increased Bcl-2 expression has been demonstrated to confer a
survival advantage in B-cells, thus promoting tumorigenesis (Reed, Cuddy et al.
1988). In a pilot study, mice injected with NIH3T3 cells containing constructs of bcl-
2 gene developed a greater number of tumors than their negative control counterparts
(Reed, Cuddy et al. 1988). In separate studies, Bcl-2 transgenic mice demonstrated an
uncontrolled expansion of B-cell lymphocytes, leading to lymphadenopathy whereas
Bcl-2 knockout mice were more susceptible to irradiation-mediated apoptosis and
displayed lower T-lymphocyte survival rates (McDonnell, Deane et al. 1989;
Sentman, Shutter et al. 1991). These results point towards the ability of Bcl-2 to
protect cells from apoptosis and promote survival. The deregulation of cellular life

2
and death homeostasis is the key to the onset and maintenance of the transformed
phenotype.

1.2 Bcl-2 and Bcl-2 family proteins:

Since the discovery of Bcl-2, many other Bcl-2-like proteins were subsequently
discovered and documented. These Bcl-2 family proteins were generally classified
into two major groups, namely pro-apoptotic and anti-apoptotic. Some of the pro-
apoptotic proteins include Bax and Bcl-xs (Boise, Gonzalez-Garcia et al. 1993; Oltvai,
Milliman et al. 1993). An alternate form of Bcl-xs is Bcl-xL which exerts anti-
apoptotic characteristics (Boise, Gonzalez-Garcia et al. 1993). Sequence analysis of
the Bcl-2 family of proteins revealed strong homology in several regions, commonly
referred to as Bcl-2 homology (BH) domains. These domains were shown to be
important for the heterodimerization of the Bcl-2 family proteins, such as the BH1
and BH2 domains necessary for Bcl-2 and Bax interaction (Yin, Oltvai et al. 1994).
The ability of these proteins to heterodimerize suggests that their ratio in cellular
abundance is critical in determining the life and death outcome of the cell.

Currently, four BH domains have been elucidated and extensively studied. Today,
Bcl-2 family proteins are divided into three classifications based on these domains.
The first group of Bcl-2 family proteins is anti-apoptotic and contains BH1-4 domains.
These include Bcl-2, Bcl-xL, Bcl-w and Mcl-1 (Strasser 2005). The second group of
members is pro-apoptotic and contains BH1-3 domains. These include Bax, Bak and
Bok. Indeed, deletion of Bax and Bak impaired the apoptotic pathway through the

3
failure to induce mitochondrial outer membrane permeability, thus preventing the
release of essential apoptotic factors such as cytochrome c (Wei, Zong et al. 2001;
Kuwana and Newmeyer 2003). Furthermore, deletion of the BH3 domain obliterated
the pro-apoptotic activity of Bax and Bak by preventing the binding of these proteins
to anti-apoptotic Bcl-2, suggesting that these pro-apoptotic proteins kill by binding
and inhibiting their anti-apoptotic counterparts through the crucial BH3 motif
(Chittenden, Flemington et al. 1995; Sedlak, Oltvai et al. 1995). The third group of
proteins is also pro-apoptotic in nature and consist only the BH3 domain. They are
the BH3-only proteins and include Bad, Bid, Bim, Bmf, Noxa and PUMA (Youle and

Strasser 2008). These small proteins act through either the direct binding and
inhibition of anti-apoptotic Bcl-2 proteins or the direct activation of Bax and Bak.
They also exhibit varying specificities in their binding to other Bcl-2 family members
(Willis and Adams 2005).

Apart from their BH domains, Bcl-2 family proteins also consist of a carboxyl
terminal hydrophobic transmembrane domain, which is critical for membrane
localization and insertion (Goping, Gross et al. 1998). Through various imaging and
biochemical techniques, Bcl-2 was found localized to various sub-cellular
membranous compartments, namely the nuclear envelope, endoplasmic reticulum and
outer mitochondrial membrane (Krajewski, Tanaka et al. 1993). Interestingly,
structural studies of Bcl-xL revealed the importance of BH1-3 domains in defining
the top of the hydrophobic groove, which is part of an essential region that interacts
with pro-apoptotic members such as Bax and Bak (Muchmore, Sattler et al. 1996;

4
Sattler, Liang et al. 1997). Analogous observation was also made in Bcl-2, differing
only by amino acid sequences and size of the hydrophobic groove, possibly
accounting for the different binding affinities for pro-apoptotic proteins between Bcl-
2 and Bcl-xL.

1.3 Role of Bcl-2 in non-apoptotic cell death:
Oncogenesis is typically characterized by an imbalance between life and death,
whereby an excessive signal for proliferation is further aggravated by an inability to
respond to physiological death triggers, eventually leading to a buildup of cell mass.
Thus, the ability to avoid various forms of cell death must certainly be a hallmark of
cancer. Cell death is classified into programmed and non-programmed. Programmed
cell death consists of apoptosis and autophagy, which are organized and sequential
processes involved in the orderly removal of unwanted cells. In contrast, non-
programmed cell death consists of a series of random events that lead to the

disorderly disruption of cellular components, often leading to inflammation. This is
known as necrotic cell death. Necrosis-associated loss of mitochondrial functions
resulting in ROS formation and leakage, leading to downstream deleterious events
can be modulated and altered by the action of Bcl-2 at the outer mitochondrial
membrane, regulating the organelle’s membrane integrity and permeability (Kane,
Ord et al. 1995; Bredesen, Rao et al. 2006).

In normal cells, the physiological function of autophagy seems to promote survival in
order to protect cells from starvation and nutrient-deprived conditions (Levine and

5
Klionsky 2004). However, in tumor cells, excessive breakdown of cellular
components may lead to cell death (Otsuka and Moskowitz 1978; Kisen, Tessitore et
al. 1993). In this respect, nutrient-deprived cancer cells often generate a lower
autophagic response than normal cells. This protective down-regulation may perhaps
be associated with Bcl-2. Indeed, a key autophagic and tumor suppressive protein
known as Beclin 1, was shown to physically interact with Bcl-2 and Bcl-xL using its
BH3 domain, thus neutralizing its autophagic activity (Shimizu, Kanaseki et al. 2004;
Pattingre, Tassa et al. 2005; Maiuri, Le Toumelin et al. 2007). Disruption of this
interaction restored the autophagic function of Beclin 1, suggesting an anti-
autophagic role for Bcl-2 and Bcl-xL (Maiuri, Le Toumelin et al. 2007).

1.4 Classical mechanisms of Bcl-2 in apoptotic cell death:
Although apoptosis was discovered in 1972, the first detailed illustration of the
apoptotic cell death pathway was elegantly conducted in by following the
development of Caenorhabditis elegans (Kerr, Wyllie et al. 1972; Sulston and
Brenner 1974). In mammals, apoptosis can be separated into two forms, the extrinsic
and intrinsic pathways (Danial and Korsmeyer 2004). Both pathways lead to the
downstream processing of unique proteases, known as initiator and executioner
caspases. The extrinsic pathway is signaled through the activation of a surface

receptor such as Fas receptor, leading to the activation of initiator caspase 8,
triggering the cleavage and activation of downstream effector caspases such as
caspase 3 (Hengartner 2000). Cells that are deficient in the extrinsic pathway are
often compensated by a robust intrinsic pathway, where the mitochondria play a

6
central role in the induction of apoptosis. The intrinsic pathway usually involves the
translocation of cleaved Bid to the mitochondria, which in turn drives the activation
of Bax to induce cytochrome c release via the disruption of the mitochondrial outer
membrane permeability, leading to downstream events including the formation of the
apoptosome, activation of caspase 9, cleavage of caspase 3 and the downstream
degradation of cellular components such as lamin and PARP (Hengartner 2000).

Indeed, overexpression of Bcl-2 in Caenorhabditis elegans was shown to rescue the
cells from programmed cell death (Vaux, Weissman et al. 1992). Furthering this,
many other studies went on to demonstrate the involvement of various other Bcl-2
family proteins in the regulation of apoptosis (Horvitz 1999). With respect to Bcl-2,
given its localization to the outer mitochondrial membrane, overexpression of Bcl-2
would block the intrinsic apoptotic pathway and not the extrinsic pathway (Krajewski,
Tanaka et al. 1993; Nguyen, Millar et al. 1993).

Mitochondria, the powerhouse of the cell, essential for providing the main source
energy, is also a crucial regulator of the intrinsic apoptotic pathway as it contains a
plethora of apoptogenic factors that can trigger apoptosis upon release (Green and
Reed 1998; Kroemer, Dallaporta et al. 1998). Death-inducing stimuli such as
irradiation, cytokine deprivation and chemotherapeutic compounds can all trigger
mitochondrial-dependent apoptosis, characterized by the depolarization of
mitochondrial transmembrane potential leading to the permeabilization of the
mitochondrial outer membrane (MOMP) (Hail 2005).

7

In this respect, overexpression of Bcl-2 in tumor cells can inhibit MOMP and bring
about chemoresistance (Vander Heiden and Thompson 1999). Upon exposure to
apoptotic triggers, MOMP is induced by pro-apoptotic cytosolic Bid and Bax, which
undergo a conformational change caused by mechanisms such as dephosphorylation
and proteolytic cleavage in order to expose the pro-apoptotic BH3 domain of these
proteins (Zha, Harada et al. 1996; Desagher, Osen-Sand et al. 1999; Li, Boehm et al.
2007). This conformational change brings about the translocation of these pro-
apoptotic members to the mitochondria. Upon translocation, these pro-apoptotic
members such as Bax and Bak have been postulated to oligomerize and form pore-
like channels to permeabilize the outer mitochondrial membrane or regulate
mitochondrial membrane channels such as ANT and VDAC in a fashion that causes
mitochondrial matrix swelling and outer membrane disruption, with MOMP being the
end result (Brenner, Cadiou et al. 2000; Wei, Zong et al. 2001; Zamzami and
Kroemer 2001).

The onset of MOMP leads to the release of several apoptogenic factors resident
within the mitochondrial intermembrane space and these include cytochrome c and
Apoptosis Inducing Factor (AIF). Cytochrome c released into the cytosol is a pre-
condition for the downstream induction of Apaf-1 oligomerization as well as
activation of caspase 9. These components associate together to form a complex
called the apoptosome that triggers the activation of executioner caspases 3 and 7,

8
leading to protein degradation and overall breakdown of the cell (Gross, McDonnell
et al. 1999; Slee, Harte et al. 1999; Hengartner 2000).

Contrary to the actions of Bax and Bak, Bcl-2 and Bcl-xL are able to inhibit MOMP
through the direct interaction with the outer mitochondrial membrane channel, VDAC,

preventing its closure induced by Bax and Bak (Shimizu, Narita et al. 1999; Vander
Heiden, Li et al. 2001; Shi, Chen et al. 2003). On the other hand, Bcl-2 has also been
proposed to function as an ionophore to dissipate the transmembrane potential that is
responsible for the closure of VDAC (Vander Heiden and Thompson 1999).
Nonetheless, both mechanisms of action result in the maintenance of the ATP/ADP
exchange and prevent hyperpolarization of the mitochondrial transmembrane
potential, leading to organelle swelling, rupture and eventual collapse of the
transmembrane potential.

1.5 Bcl-2 and its network of interacting partners:
It is well-established that Bcl-2 is able to recognize and bind to their pro-apoptotic
counterparts, thus leading to their sequestration and inability to carry out their pro-
apoptotic function. The ‘addiction’ of Bcl-2 family proteins to seek out and bind to
one another in tumor cells suggests that the ratio of proteins from the various classes
of Bcl-2 family can tilt the cell either towards life or death. This implicates a major
chemotherapeutic advantage considering that tumor cells often overexpress anti-
apoptotic Bcl-2 and introducing pro-apoptotic Bcl-2 family mimetics can specifically
target and neutralize Bcl-2 in tumor cells, without affecting or killing normal cells.

9

More importantly, given that Bcl-2 has also been shown to localize to the nuclear
envelope and endoplasmic reticulum, many studies have demonstrated the ability of
Bcl-2 to bind and interact with proteins outside of the Bcl-2 family as well as beyond
the mitochondria. The interactions with these non-homologous proteins bear
significance in the capability of Bcl-2 to integrate into a larger signaling network,
incorporating components and organelles outside of the mitochondria to govern cell
death.

Recently, p53 was shown to be able to localize to the mitochondria and directly

induce apoptosis by inducing mitochondrial permeabilization and cytochrome c
release (Marchenko, Zaika et al. 2000). Upon apoptotic stimuli such as irradiation, the
ability of p53 to directly induce apoptosis via the mitochondrial-dependent pathway
was attributed to its direct binding of Bcl-2 and Bcl-xL, displacing sequestered Bax
and triggering the downstream oligomerization of Bax, leading to cytochrome c
release (Mihara, Erster et al. 2003). Interestingly, this was achieved in the absence of
a BH3 domain in p53, instead p53 binds to Bcl-2 using its proline-rich domain
(Mihara, Erster et al. 2003). The results of these studies suggest that an
overexpression of Bcl-2 could inhibit the transcriptional-independent, death-inducing
role of p53 through the direct binding and sequestration of p53.

Apart from p53, Bcl-2 can also bind to oncogenic Ras and orphan nuclear receptor
Nur77 (Fernandez-Sarabia and Bischoff 1993; Lin, Kolluri et al. 2004). In the former

10
interaction, although Ras is usually known to promote survival in tumor cells through
the PI3-kinase/Akt pathway, it has also been demonstrated to possess pro-apoptotic
activity by up-regulating Fas ligand and bringing about Fas receptor-mediated
apoptosis. In this aspect, overexpression of Bcl-2 rescued cells from Fas-mediated
apoptosis by interacting and blocking the apoptotic activity of mitochondrial Ras
(Downward 1998; Denis, Yu et al. 2003). With regard to Bcl-2 interaction with
Nur77, a highly novel function of Bcl-2 was reported. Interaction of Bcl-2 with
Nur77 led to a conformational change in Bcl-2, exposing its BH3 domain, converting
Bcl-2 from anti-apoptotic to pro-apoptotic (Lin, Kolluri et al. 2004).

1.6 Non-canonical role of Bcl-2 in redox regulation:
Just as p53 has been portrayed to display a non-conventional transcriptional-
independent role in cell death regulation, the role of onco-protein Bcl-2 in promoting
tumor cell survival has been designated for further investigation from another
perspective, that of ROS and mitochondrial bioenergetics. Given the mitochondrial

localization of Bcl-2, can Bcl-2 possibly preserve or optimize oxidative
phosphorylation to tailor to the survival instincts of the tumor cell from a ROS
perspective? Traditionally, Bcl-2 has been portrayed as an anti-oxidant due to its
ability to suppress oxidative stress-induced lipid peroxidation when overexpressed in
murine lymphoma cells (Hockenbery, Oltvai et al. 1993). Many other studies went on
to confirm this finding (Tyurina, Tyurin et al. 1997). In addition, Bcl-2 was also
shown to reduce NO
2
-
production in response to oxidative stress and in contrast, mice
lacking Bcl-2 were more susceptible to oxidative stress-mediated damage (Hochman,

11
Sternin et al. 1998; Lee, Hyun et al. 2001). Moreover, it was reported that the anti-
oxidative property of Bcl-2 was attributed to its ability to up-regulate cellular anti-
oxidant defense mechanisms such as Cu/Zn SOD, catalases, glutathione peroxidases
and GSH levels in tumor cells (Ellerby, Ellerby et al. 1996; Jang and Surh 2003;
Rudin, Yang et al. 2003; Zimmermann, Loucks et al. 2007).

In spite of conventional acceptance of Bcl-2 as an anti-oxidant, another body of
evidence has challenged this notion by demonstrating that under normal physiological
conditions, overexpression of Bcl-2 in cells did not result in an initial anti-oxidative
intracellular milieu but instead brought about increased oxidative damage (Steinman
1995). The consequential up-regulation of intracellular anti-oxidant defenses was
postulated to be a compensatory response to the initial pro-oxidant activity of Bcl-2
(Steinman 1995). Still, other models employing mouse and bacteria corroborated our
results demonstrating Bcl-2 as a pro-oxidant protein (Adams, Pierce et al. 2001).
Various experimental models established the anti-oxidant property of Bcl-2 by
triggering cells with death-inducing stimuli or directly overwhelming the cells with
oxidative stress before accruing the resultant anti-oxidant response to Bcl-2

expression. At best, these studies accredit a redox regulatory role for Bcl-2 in
countering oxidative stress, but do not suffice to confirm Bcl-2 as having innate anti-
oxidant characteristics. True to this aspect, pure Bcl-2 has been shown to be devoid of
intrinsic anti-oxidant activity (Lee, Hyun et al. 2001). Hence, it was suggested that
the anti-oxidant function of Bcl-2 previously reported could be accrued to the
physiological response of Bcl-2 as a modulator of ROS.

12

1.7 Types of ROS:
Reactive oxygen species (ROS) refer to a set of molecules derived from molecular
oxygen and consist of a combination of oxygen radicals including superoxide (O
2
-
),
and hydroxyl (OH
-
) as well as non-radical derivatives of molecular oxygen such as
hydrogen peroxide (H
2
O
2
). Physiological processes such as the mitochondrial
electron transport chain activities and membrane-bound enzymes such as NADPH
oxidase, typically found in phagocytic cells for its microbicidal function, represent
the two main producers of intracellular ROS (Bergstrand 1990; Nohl, Gille et al.
2005). The leakage of electrons to spontaneously combine with molecular oxygen in
the mitochondria and the enzymatic reduction of molecular oxygen with a single
electron by NADPH oxidase forms O
2

-
.

Despite the notion that O
2
-
and H
2
O
2
are not highly reactive with other intracellular
constituents, the end product of a reaction between these two species is able to elicit
extensive detrimental effects within a cell and accounts for most of the intracellular
oxidative damage. This end product is known as the hydroxyl radical (Fridovich
1978). Hydroxyl radical interacts and damages intracellular components such as
DNA, lipids and proteins in a non-specific fashion, eventually leading to necrotic cell
death. ROS is also widely implicated in a variety of oxidative stress-related clinical
manifestations and applications such as neurodegeneration, inflammation, aging,
atherosclerosis, ischemia-reperfusion injury in cardiac tissues and chemotherapeutic
effects (Waris and Ahsan 2006). Involvement of ROS in an extensive network of

13
signaling pathways, governing a multitude of cellular processes that eventually
determine cellular functions, survival and death necessitates and justifies the amount
of past and ongoing work that has been conducted in the field.

1.8 Regulation of ROS – Producers of ROS and anti-oxidant mechanisms:
As mentioned in the previous section, the mitochondria constitute one of the largest
ROS-producing organelles. Being the powerhouse of the cell in generating valuable
energy for various cellular processes and functions, the mitochondrial electron

transport chain naturally and inevitably becomes one of the greatest sources of ROS.
This is due to the fact that the electron transport is not an entirely efficient process as
up to 1-3% of molecular oxygen can be converted to ROS in the mitochondria due to
the leakage of electrons (Boveris and Chance 1973). Mitochondrial respiration results
in a continuous stream of electrons being transferred from one complex enzyme to
another before eventually reducing molecular oxygen to produce water at the terminal
enzyme, COX. Some of these electrons are inevitably lost during electron transport
and react with the surrounding oxygen molecules to give rise to O
2
-
within the
organelle.

The main O
2
-
-producing complexes are complex I and III, where it was further shown
that complex III predominantly produces O
2
-
to the cytoplasmic side of the inner
mitochondrial membrane, suggesting that O
2
-
from complex III preferentially
accumulates at the mitochondrial intermembrane space (Grigolava, Ksenzenko et al.

14
1980; Turrens and Boveris 1980; Turrens, Alexandre et al. 1985; Muller, Liu et al.
2004).

In addition to the mitochondrial electron transport chain, NADPH oxidase complex is
another major source of ROS production. The enzyme consists of subunits gp91phox,
p22phox, p47phox, p67phox, p40phox (Babior 1999). Association of the small GTP-
binding protein Rac1 is necessary for the membrane complex to be functional in order
to carry out its microbicidal role through the massive production of an oxidative burst
of O
2
-
, typically observed in phagocytic cells (Babior 1999). Recent evidence has
demonstrated that the role of NADPH oxidase observed in phagocytic cells extends
into transformed cells as well and contributes towards oncogenesis. Indeed, various
forms of NADPH oxidases (Nox) have been reported in different cancers such as
Nox4 in pancreatic cancer and Nox5 in melanoma and prostate cancer (Brar, Corbin
et al. 2003; Mochizuki, Furuta et al. 2006). In these studies, ROS produced from
these Nox proteins contribute towards the survival signaling pathways observed in
these cancers.

Apart from NADPH oxidases, other ROS-producing enzymes include xanthine
oxidase, flavoprotein dehydrogenase, aldehyde oxidase, tryptophan dioxygenase and
dihydroorotate dehydrogenase (Freeman and Crapo 1982). Out of these, xanthine
oxidase remains the most widely studied oxidase and its O
2
-
-producing function has
been exploited for in vitro studies to understand the effect of ROS on a variety of
cellular processes. Over and above the enzymatic production of ROS, organelles such

15
as endoplasmic reticulum and nuclear membrane has also been reported to participate

in ROS production. More specifically, endoplasmic reticulum-derived O
2
-
, though yet
to be implicated in growth and survival signaling, has nonetheless been postulated to
play a role in the some of the organelle’s main functions such as protein folding and
secretion (Bauskin, Alkalay et al. 1991; Hwang, Sinskey et al. 1992; Bader, Muse et
al. 1999). In relation to nuclear membrane, studies have demonstrated the existence of
cytochrome oxidases and electron transport systems that may contribute to the
formation of ROS through the leakage of electrons from these systems, leading to
DNA damage, even though the exact function of these systems remain unclear (Droge
2002). Moreover, nuclear localization of Nox4 also implies redox regulation of
nuclear-related or even cellular processes (Kuroda, Nakagawa et al. 2005; Ushio-
Fukai 2006).

With the presence of ROS-producing entities and the escalating threat of oxidative
stress-induced damage if uncontrolled, the cell has evolved a set of anti-oxidant
protective mechanisms to maintain a level of ROS that is sufficient for intracellular
signaling functions but does not suffice to produce deleterious effects. The anti-
oxidant defense machinery comprises of several enzymes and non-enzymatic ROS
scavengers. These enzymes include MnSOD, Cu/Zn SOD, catalase, GSH peroxidase
and GSH reductase. ROS-scavenging compounds include alpha-tocopherol (vitamin
E), β-carotene, ascorbate (vitamin C), glutathione and some free amino acids
(Fridovich 1986; Halliwell 1999).


16
The function of MnSOD and Cu/Zn SOD is to dismutate radical O
2
-

to non-radical
H
2
O
2
(Fridovich 1986). In spite of H
2
O
2
not being highly reactive with other cellular
components, the iron-dependent, reductive homolytic cleavage of the compound is
able to generate a more potently cytotoxic compound called hydroxyl radical in a
reaction known as the Fenton reaction (Andreyev, Kushnareva et al. 2005). The
hydroxyl radical is responsible for the majority of oxidative damage leading up to
necrotic cell death. In order to prevent this detrimental route of H
2
O
2
breakdown, the
presence of catalase converts two molecules of H
2
O
2
to form molecular oxygen and
water. Apart from catalase, which is a form of peroxidase, glutathione peroxidase also
helps to transform H
2
O
2
to molecular oxygen and water, albeit with the requirement

of the intracellular reducing agent, glutathione (Fridovich 1978; Furtmuller,
Zederbauer et al. 2006).

It has also been noted that the ability of free amino acids to act as ROS scavengers is
more enhanced than proteins. Thus, as far as oxidative proteolysis is concerned, the
ability of free amino acids to scavenge ROS provides the basis for self-regulation and
homeostasis of redox-dependent protein breakdown (Droge 2002). Thus, anti-
oxidative defense mechanisms are essential to regulate the cellular redox state
without incurring excessive and overwhelming levels of ROS that can bring about
irreversible cellular damage. However, such protective, self-regulatory mechanisms
are only possible and functional under transient increases in ROS whereas,
overwhelming levels of ROS would culminate in cell death and tissue damage,
regardless whether the mechanisms are elicited. On the other hand, deregulation in

17
ROS production and/or its regulatory machinery can lead to abnormal accumulation
of ROS, leading to species-dependent conditions that are favorable for the onset,
maintenance and progression of cancer.

1.9 ROS in cell death and survival:
Conventional dogma has long established ROS as agents of detriment. Indeed, a great
number of studies have documented the role of ROS as mediators of damage to
cellular structures, nucleic acids, proteins and lipids. Modification of the DNA
molecule by ROS represents the first step of mutagenesis and if left unchecked,
carcinogenesis ensues. Lipid peroxidation and protein modification of Bax monomer
to promote oligomerization of Bax are common features of ROS-mediated damage,
culminating in the compromise of the mitochondrial outer membrane integrity and the
subsequent release of cytochrome c to initiate the mitochondrial death pathway
(Buccellato, Tso et al. 2004). These form the premise for the use of ROS-based
chemotherapeutics in cancer intervention and management.

While it is true that overwhelming ROS are harmful to cells, an emergent growing
body of evidence indicates the importance of low levels intracellular ROS in
physiological signaling, tumor promotion/ initiation and its subsequent maintenance
and progression. There is sufficient evidence to strongly support a paradigm shift
from the convention of ROS as only a mediator of cell damage/death to that of
survival/proliferation. Mild elevation in O
2
-
or H
2
O
2
has been demonstrated to
promote growth responses in a variety of cell types through activation of growth-

18
related genes such as c-fos and c-jun, alterations in protein kinase activities, oxidative
modifications to phosphatases and activation of transcription factors (Burdon 1995;
Sauer, Wartenberg et al. 2001). Some of these include ROS stimulatory effect on the
PI3-kinase/AKT survival pathway through the oxidative inactivation of PTEN, redox
regulation of MAPKs such as c-Jun N-terminal kinase, p38MAPK and ERK, as well
as activation of transcription factors such as AP-1 and NF-κB (Heffetz, Bushkin et al.
1990; Droge 2002; Leslie, Bennett et al. 2003; Qin and Chock 2003; Bubici, Papa et
al. 2006). More importantly, NADPH-dependent generation of ROS has been
reported upon growth factor stimulation or cytokine receptor activation such as
PDGF, TNF-α, IL6, IL3, FGF-2, TGF-α and insulin, implicating ROS as secondary
messengers of survival and proliferative signaling (Sauer, Wartenberg et al. 2001;
Droge 2002).

Intracellular ROS functions in a diverse fashion, implicating different cellular
components, to promote cell growth and survival. Indeed, ROS has been shown to
positively modulate a variety of ion channels such as IP3 receptor-dependent Ca
2+

transport, Na
+
/Ca
2+
and Na
+
/H
+
exchangers at the plasma membrane, leading to
growth stimulation (Burdon 1995; Sauer, Wartenberg et al. 2001). One of these
studies recently implicate ROS in the species-dependent activation and inactivation of
Na
+
/H
+
exchanger to promote cell division or death via the creation of an alkaline or
acidic intracellular environment respectively (Shibanuma, Kuroki et al. 1988; Akram,
Teong et al. 2006). Moreover, species-specific regulation of various processes in
different lineages of cells has also been widely investigated. It has been shown that

19
O
2
-
in the presence of an intracellular reduction in H

2
O
2
results in an enhanced T-cell
activation during an immune response via an increase in the promoter activity,
transcription and expression of IL2 and its receptor (Droge, Eck et al. 1992; Droge
2002).

The imbalance between ROS production and breakdown has been postulated to be
responsible for various disorders, more prominently, carcinogenesis. The increased
metabolic rate of tumor cells, coupled with the enhanced production and reduced
removal of ROS to create a pro-oxidant intracellular milieu, has been linked to
promote the survival of cancer cells (Cerutti 1985; Burdon, Gill et al. 1989; Burdon,
Gill et al. 1990). This causal effect of ROS to promote cellular transformation is
backed by evidence provided from various studies such as the ability of O
2
-
to cause
DNA oxidation and promote the transformed cell type, which can be reversed by the
expression of MnSOD. Expectedly, a pro-oxidant milieu is also able to suppress
MnSOD expression and activity, leading to further accumulation of the species. This
not only emphasizes the role of MnSOD as a tumor suppressor but also the
importance of O
2
-
in tumorigenesis (Church, Grant et al. 1993; St Clair, Oberley et al.
1994; Oberley 2001). In addition, oncoprotein p21Ras activation of Rac1 led to an
increase in cell proliferation via a concomitant increase in levels of O
2
-

(Irani, Xia et
al. 1997). This was corroborated by the ability of constitutively active Ras to maintain
an elevated level of O
2
-
, contributing to the resistance upon drug-induced apoptosis.
Conversely, the expression of a dominant-negative form of Rac1 reduced the levels of

20
O
2
-
and abolished the resistance conferred against apoptosis (Irani and Goldschmidt-
Clermont 1998; Pervaiz, Cao et al. 2001).

Further studies have gone on to show that the intricate balance between the different
species of free oxygen radicals is important in determining the decision of the cell
fate. A tilt in favor of O
2
-
levels such as the inhibition of Cu/Zn SOD using DDC,
with no appreciable increase in H
2
O
2
renders the cancer cell refractory to death
execution and preserves viability, irrespective of the trigger (Clement, Ponton et al.
1998; Clement and Pervaiz 1999; Pervaiz, Ramalingam et al. 1999; Pervaiz, Seyed et
al. 1999; Clement and Pervaiz 2001). In contrast, increased levels of H
2

O
2
and a
corresponding drop in O
2
-
levels via an overexpression of Cu/Zn SOD create a
reduced and acidified environment conducive for the apoptotic signal to filter through
(Clement, Ponton et al. 1998; Pervaiz, Seyed et al. 1999). This notion was further
reinforced by a particular study showing a direct determination of cell death or
survival by non-necrotic and non-overwhelming fluctuation in levels of O
2
-
(Lin,
Pasinelli et al. 1999). Correspondence between increased activity of O
2
-
producing
systems and proliferation networks such as the capability of Nox 1 to produce O
2
-
and
induce cell growth

as well as the fact that several anti-cancer drugs exert their effects
via

H
2
O

2
mediated killing further lends weight to the pro-oxidant theory

of
carcinogenesis, prevalent in many tumor types, depending on the

intracellular levels
and species involved.


21
On the other hand, the requirement of a reduced intracellular environment for pro-
apoptotic conditions confounds the essential role of O
2
-
in apoptotic signaling, thus
throwing the involvement of H
2
O
2
into the spotlight (Halliwell and Gutteridge 1990;
Jacobson, Burne et al. 1993). Despite being relatively non-reactive and a poor redox
agent, H
2
O
2
has been shown to trigger apoptosis or necrosis depending on its
concentration (Hampton and Orrenius 1997; Clement, Hirpara et al. 1998). H
2
O

2
-
mediated apoptosis is typified by an environment permissive for death signaling
through a corresponding drop in O
2
-
levels and intracellular pH whereas H
2
O
2
-
induced necrosis brings about characteristics reminiscent of oxidative stress and
damage (Clement, Ponton et al. 1998; Ahmad, Clement et al. 2004; Ahmad, Iskandar
et al. 2004). Generally, treatment with anti-cancer therapeutics triggers the
intracellular production of H
2
O
2
, leading to MOMP and release of apoptogenic
factors such as cytochrome c and AIF (Reed and Kroemer 2000). Nonetheless, of
interest to this thesis is to pinpoint the mechanisms leading to the emergence of the
early O
2
-
that is generated in creating the pro-oxidant intracellular state, favorable for
the survival of tumor cells, with respect to Bcl-2.

1.10 Paradigm shift on ROS implicating a pro-oxidant role of Bcl-2:
Recent work has unraveled ROS as being more than just indiscriminate mediators of
cell and tissue damage through oxidative stress. Although it may be true that

overwhelming levels of ROS may induce non-specific and random deleterious events
detrimental to the cell, moderate levels of ROS may determine specific signaling
pathways governing cell death and survival that are both species- and threshold-

22
dependent. Specifically, when the ratio of O
2
-
over H
2
O
2
is increased, survival is
favored through the enhancement of proliferative signals (Clement, Hirpara et al.
1998; Pervaiz, Cao et al. 2001). Conversely, when the concentration of H
2
O
2
is
greater than O
2
-
, the intracellular microenvironment becomes conducive for the
apoptotic machinery to function through an acidification of the cytosolic
compartment (Hirpara, Clement et al. 2001; Pervaiz and Clement 2002; Ahmad,
Iskandar et al. 2004; Akram, Teong et al. 2006). Thus, low levels of ROS are
essential for the homeostatic regulation of various cellular processes by acting as
signaling messengers, thus ensuring a normal cellular turnover, necessary for tissue
survival (Sauer, Wartenberg et al. 2001). In the tumor context, moderate levels of
ROS are tightly regulated by both ROS-producing systems and anti-oxidant defense

mechanisms, maintaining a species-specific preference for cancer cell survival and
death pathways. The deregulation of mechanisms controlling ROS production and
turnover can lead to serious consequences in upsetting the balance in concentration
between the different species of ROS, leading to a more pronounced and aggressive
malignancy or a regression in tumorigenesis.

This paradigm shift in the perspective on ROS was illustrated by a growing body of
evidence implicating Bcl-2 and the production of O
2
-
to result in a pro-oxidant state
that is responsible in conferring a survival advantage in tumor cells. Recent work
indicate that Bcl-2 did not operate as an anti-oxidant on its own but rather, its
expression levels was directly associated with a pro-oxidant intracellular milieu that
triggered the reinforcement of the endogenous anti-oxidant defense machinery

23
(Pervaiz and Clement 2007). In turn, this mild pro-oxidant state was connected to the
death inhibitory activity of Bcl-2 as shown in leukemia cells (Clement, Hirpara et al.
2003). Moreover, inhibition of NADPH oxidase activity by DPI and dominant-
negative form of Rac1 decreased intracellular O
2
−
and rendered Bcl-2 overexpressing
cells more sensitive to apoptosis, suggesting specificity for O
2
−
in Bcl-2 mediated
pro-oxidant state (Clement, Hirpara et al. 2003). These reports firmly established the
significance of a mild pro-oxidant milieu in tumor progression/initiation through

enhanced survival, with Bcl-2 at the heart of this phenomenon. This slight pro-
oxidant state has been shown to favor a cascade of survival signaling pathways in
cancer cells (Steinman 1995; Clement and Stamenkovic 1996; Ahmad, Clement et al.
2003; Clement, Hirpara et al. 2003; Pervaiz and Clement 2007).

Thus, mitochondria being a major site of O
2
-
production through

its electron transport
chain activities are crucial organelles for the

study on the potential impact of onco-
proteins such as Bcl-2 on its

physiology in order to validate the pro-oxidant state,
necessary for the

transformed phenotype. The pro-oxidant role for mitochondria is
further supported by evidence demonstrating that upstream events such as increased
mitochondrial biogenesis or mutant mitochondrial proteins can lead to increased
production of ROS via increased mitochondrial respiration (Zhang, Gao et al. 2007;
Dasgupta, Hoque et al. 2008). Hence, increased activity of the terminal, rate-limiting
COX enzyme will inadvertently enhance the overall rate of electron transport across
the mitochondrial respiratory chain and increase the propensity for leakage of
electrons along the chain to form O
2
-
upon reaction with molecular oxygen,

24
particularly at complex I and III. However, one confounding factor is the

notion that
cancer cells generally exhibit reduced oxidative phosphorylation

and the next section
seeks to address this issue.

1.11 Altered tumor metabolism and cell fate:
Early studies on tumor metabolism proposed a unique, signature characteristic in the
way these rogue cells obtained their source of ATP to meet the rigorous demands of
proliferation, invasion and adaptations to the harsh tumor microenvironment. Back in
1924, Otto Warburg proposed that cancer cells preferentially utilize the glycolytic
pathway over oxidative phosphorylation to provide for the majority of the energy
supply. This is in stark contrast to normal cells where the reliance on oxidative
phosphorylation is far greater than glycolysis for the generation of ATP. Warburg
attributed this phenomenon to the dysfunction of the mitochondria whereby these
metabolic differences were regarded as an adaptation to the hypoxic environment
within the solid tumor (Gatenby and Gillies 2004).

More than 80 years on, the Warburg effect is widely applied as the de rigueur
metabolic phenomenon to distinguish cancer cells from non-cancerous ones. The idea
that cancer cells predominantly utilize glycolysis for energy production is being
employed as a parameter in positron emission tomography to assess the prevalence of
tumors in the clinical setting. Indeed, extensive studies have shown that fast-growing
and highly de-differentiated cancer cell types demonstrated highly modified
metabolic patterns compared to their normal counterparts (Pedersen 1978; Dastidar

25
and Sharma 1989; Mazurek, Michel et al. 1997; Rodriguez-Enriquez, Torres-Marquez
et al. 2000; Ziegler, von Kienlin et al. 2001; Griguer, Oliva et al. 2005). In agreement
with these observations, a large body of evidence has emerged suggesting significant
up-regulation in the expression of glycolytic genes in aggressive malignancies
(Bustamante and Pedersen 1977; Oskam, Rijksen et al. 1985; Vora, Halper et al.
1985; Nakashima, Paggi et al. 1988; Atsumi, Chesney et al. 2002; Medina and Owen
2002; Wood and Trayhurn 2003; Macheda, Rogers et al. 2005; Marin-Hernandez,
Rodriguez-Enriquez et al. 2006).

With these compelling evidence, the concept that enhanced glycolysis is always
induced or accompanied by near-defunct oxidative phosphorylation has been
ubiquitously and indiscriminately applied to all types of cancer. The universal
acceptance of the Warburg effect thus formed the central metabolic dogma that come
to characterize all cancer cells. However, it is important to note that fundamental
genetic, biochemical and morphological heterogeneity of tumor cells may render the
convenient application of the Warburg effect to define all types of cancer as
generalization. Tumor cell types such as glioblastoma multiforme, astrocytoma,
MCF7 and certain forms of hepatoma utilize both glycolysis and oxidative
phosphorylation to an equal extent for energy production (Elwood, Lin et al. 1963;
Lowry, Berger et al. 1983; Liu, Hu et al. 2001; Zu and Guppy 2004). More
importantly, tumors such as bone sarcoma, lung carcinoma, breast cancer, skin
melanoma, cervical, ovarian and uterus carcinomas all primarily make use of
oxidative phosphorylation for the generation of ATP (Elwood, Lin et al. 1963;