Studies on the anti cancer potential of sesquiterpene lactone parthenolide

component, parthenolide, are potent inhibitors of macrophage production of release of a group of pro-inflammatory cytokines, including tumor necrosis factor TNF and interleukins IL Hwang

1 CHAPTER 1

CHAPTER 1 INTRODUCTION

1.1 Parthenolide

1.1.1 Introduction: feverfew and Parthenolide

Feverfew (Tanacetum parthenium) is an aromatic perennial herb grown

originally in Europe It has been traditionally used as an herbal medicine in treatment

of migraine, fever, arthritis and menstrual problems since ancient times In the last few decades, feverfew became especially popular in Britain, France and Canada as a phytomedicine Powder of dried leaves has been made into capsules or tablets for oral consumption The major pharmacological benefits of feverfew include: (i) prevention

of migraine, (ii) relief of fever, and (iii) treatment of rheumatoid arthritis and

inflammatory disease (Berry, 1984; Johnson et al., 1985)

The main chemical constituents of feverfew include sesquiterpene lactones, flavonoid glycosides and monoterpenes (Abad et al., 1995) The most predominant and well studied active component of feverfew is a group of sesquiterpene lactones such as parthenolide, costunolide and germacranolide, etc (Knight, 1995) These sesquiterpene lactones are enriched in the leaves and seeds of the feverfew which are believed to be produced by superficial leaf glands and act as general insect repellents for plants Among feverfew extracts, parthenolide is the principle sesquiterpene lactone responsible for most of the pharmacological effects of feverfew (Knight, 1995)

1.1.2 Chemical structure, metabolism and bioactivities of parthenolide

1.1.2.1 Chemistry: sesquiterpene lactones and parthenolide

Chemically, parthenolide belongs to the group of sesquiterpene lactone It was first isolated from feverfew in 1965 (Berry, 1984) Parthenolide is a 4,5-epoxyger-macra-(10), 11(13)-dien-12,6-lactone and its structure is shown in Fig 1.1

The molecular structure contains a terpene compounds with fifteen carbon atoms and an exocyclic methylene lactone group, an α-methylene-γ-lactone moiety, which has been shown to process an anti-inflammatory activity The α-methylene-γ-lactone group has an exceptional ability to react with nucleophiles by Michael type addition (Kupchan et al., 1970) Parthenolide has the ability to alkylate intracellular nucleophiles, such as L-cysteine, glutathione (GSH) and a number of thiol-bearing cellular proteins, to form adducts (Fig 1.2) which are believed to be responsible for its pharmacological effects (Hall et al., 1979; Groenewegen et al., 1986; Knight, 1995) Parthenolide derivative, 11β,13-dihydroparthenolide with saturated exocyclic

O

CH3

CH2O

methylene group completely loses its bioactivity suggesting the biological importance

of exocyclic methylene group (Marles et al., 1992) In addition, parthenolide possesses an epoxide moiety, another functional group with alkylating ability, which results in an enhanced bioactivity compared to other sesquiterpene lactones Recently,

it reported that the spatial arrangement of the terpenoid skeleton is more important than any other functional groups (Neukirch et al., 2003) which are generally believed

to be responsible for the bioactivities of parthenolide

Fig 1.2 Formation of parthenolide-thiol adducts

1.1.2.2 Transportation in cell system and bioavailability

After administration of parthenolide, it can be quickly transported and absorbed

by human cells Khan et al (2003) has reported that parthenolide is predominately transported into human intestinal cells (Caco-2) through passive diffusion which can not be prevented by an inhibitor of multidrug resistance transporter P-glycoprotein (MRP) This report provides the evidence of the bioavailability of the parthenolide in

cell culture system (Khan et al., 2003) In the in vivo animal model, although the

bioavailability of parthenolide has not been reported, recently work in our laboratory using UVB-induced skin cancer model suggested the bioavailability of parthenolide

as parthenolide-feeding showed a delayed onset of papilloma incidence and a significant reduction in papilloma multiplicity (Won et al., 2004)

1.1.2.3 Bioactivities of parthenolide

Thiol reactivity

As discussed earlier, the chemical reactions of parthenolide involve a covalent conjugation between α,β-unsaturated carbonyl structure of parthenolide with various nucleophilic sulphydryl residues (e.g thiols) resulting in alkylation through Michael type addition The binding of parthenolide with sulphydryl groups present in intracellular proteins may disrupt the normal cellular function of macromolecules The three-dimensional structure of both parthenolide and its potential target are the decisive factors for the steric accessibility of parthenolide to its target Appropriate three-dimensional structure is the essential premise for bioactivities of parthenolide and may provide some extent specificity (Yoshioka et al., 1973) However, the biological consequences and the importance of parthenolide’s thiol reactivity are poorly studied

Anti-inflammatory activity

Inflammation is an important biological process contributing to wound healing and pathological responses to infection A complex network of signaling factors involved in inflammatory response has been evolved (Coussens and Werb, 2002) Anti-inflammatory activity is one of the most prominent bioactivities of parthenolide Feverfew has been traditionally used as a herbal medicine by ancient Greeks and early Europeans for treatment of inflammatory diseases, such as fever and rheumatoid arthritis (Berry, 1984) In 1989, a double-blind study carried out in the UK, demonstrated the effectiveness of feverfew in treatment of the symptoms of rheumatoid arthritis (Pattrick et al., 1989) which may due to pharmacological inhibition of pro-inflammatory cytokine prostaglandin (PG) synthesis The following

in vitro studies further elucidated that feverfew as well as its major bioactive

component, parthenolide, are potent inhibitors of macrophage production of release of

a group of pro-inflammatory cytokines, including tumor necrosis factor (TNF) and interleukins (IL) (Hwang et al., 1996) and this inhibitory activity are mainly regulated

via disruption of nuclear factor - kappaB (NF-κB) signaling pathway (Hehner et al.,

1999) and signal transducers and activators of transcription (STAT) pathway (Sobota

et al., 2000)

Anti-cancer activity

It has been suggested that chronic inflammation promotes cancer development

in many types of cancers, such as breast, colorectal, and liver cancer (Coussens and Werb, 2002) The potent anti-inflammatory activity of parthenolide implies its potential anti-cancer property Parthenolide has been reported to interrupt cell cycle regulation and induce apoptosis in human cancer cells (Wen et al., 2002) In addition,

Patel et al (2000) and his colleagues also demonstrated that pre-treatment of

parthenolide greatly sensitizes the human breast cancer cells in response to cancer drug paclitaxel and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced apoptosis (Patel et al., 2000; Nakshatri et al., 2004) However, the anti-cancer potential of parthenolide is still largely unknown Further studies,

anti-including both in vitro mechanism studies on anti-cancer potential of parthenolide and

in vivo animal study, are needed

1.1.3 The molecular mechanisms involved in the bioactivities of parthenolide

1.1.3.1 Effects on NF-κB signaling

NF-κB is a highly inducible transcription factor in response to diverse stimulations such as TNFα, ultraviolet (UV), interleukins, endotoxins etc NF-κB activation plays a pivotal role in regulating the inflammatory responses, cell growth/differentiation and apoptosis (Li and Verma, 2002; Karin and Lin, 2002) In

the typical NF-κB signaling pathway, NF-κB is a heterodimer formed by RelA (p65) and p50 subunit which are sequestered as an inactive form in the cytoplasm by NF-κB inhibitory protein (IκB) Upon stimulation, the IκB protein is phosphorylated by IκB kinase (IKK), ubiquitinized and degraded by 26s proteosome This process results in unmasking the nuclear localization sequence (NLS) of NF-κB protein, NF-κB nuclear translocation and activation (Li and Verma, 2002) Bork et al (1997), for the first time, screened the inhibitory effects of 54 Mexican Indian medicinal plants on transcription factor NF-κB Their data suggested that parthenolide, the major bioactive sesquiterpene lactone extracted from feverfew, shows a potent inhibitory effect on NF-κB activation at a concentration as low as 5µM (Bork et al., 1997) Later

on, Hehner et al (1998) reported that parthenolide inhibits several stimulations (TPA, TNFα, ligation of T cell receptor and hydrogen peroxide)-induced NF-κB activation

by blocking the degradation of phosphorylated IκB while the DNA binding activity of NF-κB has not been interfered (Hehner et al., 1998) In a subsequent study, they demonstrated that parthenolide directly inhibits the IKK activity induced by TNFα, while TNFα-inducible mitogen-activated protein kinase (MAPK) signaling pathways (p38 and JNK) are not affected by parthenolide (Hehner et al., 1999) Furthermore, parthenolide is capable of blocking NF-κB activation induced by overexpression of upstream activators of IKK, such as TNF receptor associated factor 2 (TRAF2) and mitogen-activated protein kinase 1 (MEKK1), which lead to a conclusion that parthenolide inhibits TNFα-induced NF-κB activation by targeting the IKK complex (IKC) (Hehner et al., 1999) At the same time, Rungeler et al (1999) screened the inhibitory activity on NF-κB of 28 sesquiterpene lactones including parthenolide Using a computer modeling, they proposed a possible molecular mechanism, which the inhibitory activity of parthenolide may due to the alkylation and cross-linking of

two cysteine residues (cys 38 and cys 120) located on the p65 subunit of NF-κB (Rungeler et al., 1999) Subsequently, two hypotheses emerged to explain the inhibitory effect of parthenolide on NF-κB activation First, parthenolide inhibits NF-

κB activity by direct targeting and inhibiting the IKKβ Single amino substitution in the activation loop (C179A) of IKKβ abolished the parthenolide’s IKKβ inhibitory activity (Kwok et al., 2001) Second, parthenolide suppresses NF-κB activation by

direct modification of p65 at cys38 via alkylation, which in turn prevents NF-κB

DNA binding (Garcia-Pineres et al., 2001) At present, the possible effects of parthenolide upstream of IKK activation have not been studied

1.1.3.2 Effects on inflammatory-related molecules

Cytokines

A big family of cytokines plays a pivotal role in inflammatory responses The cytokine signaling is highly complicated and the effect of parthenolide seems to be dependent on specific cell line and cellular contexture During inflammatory responses, NF-κB is the most important regulator of the gene expression of pro-inflammatory cytokines (Tak and Firestein, 2001) Inhibition of NF-κB pathway is believed to be one of the major mechanisms responsible for anti-inflammatory activity of parthenolide Meanwhile, the inflammatory cytokines, such as TNFα and IL-1 could also activate the NF-κB pathway which in turn promotes the inflammatory response through a positive feedback control (Lucey et al., 1996) The effects of parthenolide on expression of several cytokines have been reported Mazor et al (2000) first reported that parthenolide shows a potent inhibitory effect on IL-8 expression in human respiratory epithelium cells (Mazor et al., 2000) In human macrophages, Kang et al (2001) observed a similar inhibitory effect of parthenolide

on IL-12 production induced by LPS The p40 promoter activity of IL-12 which

contains a NF-κB binding sequence has been greatly suppressed by parthenolide (5µM) (Kang et al., 2001) As NF-κB regulatory elements have been found in the promoter sequences of many pro-inflammatory cytokines, the inhibitory effects of

parthenolide on other cytokine secretion have also been demonstrated In Li-Weber et

al (2002)’s report, IL-4, IL-2 and IFN-γ secretion from normal peripheral blood T

cells are also suppressed by parthenolide (Li-Weber et al., 2002) All these findings indicate that the parthenolide executes its anti-inflammatory effects by regulating the

secretion of pro-inflammatory cytokines at transcriptional level via inhibition of

NF-κB

5-hydroxytryptamine (5-HT) and anti-serotonergic activity

5-HT is a monoamine neurotransmitter secreted by central nervous system and platelets It is believed that 5-HT plays a central role in migraine pathophysiology (Peroutka, 1990) The anti-secretory activity of feverfew extracts, including parthenolide, was reported in the 1980s (Groenewegen et al., 1986) and parthenolide has been found to possess a potent inhibitory effect on 5-HT secretion and platelet aggregation induced by a number of stimulations (Groenewegen and Heptinstall, 1990) It has also been suggested that parthenolide may act as a low affinity antagonist at 5-HT receptors (Weber et al., 1997) Recently, the inhibitory effects of feverfew and parthenolide on 5-HT have been re-examined again (Mittra et al., 2000) However, the exact mechanisms are still not fully elucidated

Cell adhesion molecules (CAM)

In the inflammatory process, CAMs, which mediate the interaction between leukocyte and endothelium cells, play a key role to recruit and accumulate leukocytes

to the site of inflammation (Ulbrich et al., 2003) Many anti-inflammatory drugs process the inhibitory effect of the CAMs Since feverfew is a well documented anti-

inflammatory herbal medicine, the potential effects of its main active component, parthenolide, on CAMs have also been reported Piela-Smith and Liu (2001) first reported that both feverfew extracts and parthenolide greatly inhibited the intracellular cell adhesion molecule-1 (ICAM-1) expression induced by various cytokines stimulation, including IL-1, TNFα and IFNγ (Piela-Smith and Liu, 2001) As the ICAM-1 promoter sequence has a response element potentially regulated by NF-κB, the suppressed expression ICAM-1 is likely due to the inhibitory effect of parthenolide on NF-κB (Melotti et al., 2001) Besides the ICAM-1, the effect of parthenolide on expression of other CAMs has also been addressed Furthermore,

(VCAM-1)-induced by IL-4 via JAK2-STAT6 signaling was significantly suppressed

by parthenolide (Schnyder et al., 2002) It is interesting to note that this inhibitory effect seems converge to the suppression of STAT6 nuclear translocation and DNA binding by an unknown mechanism and the phosphorylation of STAT6 is not affected

by parthenolide (Schnyder et al., 2002) Currently, it is believed that suppression of

CAMs expression which alleviates the inflammatory response is one of the

mechanisms of anti-inflammatory activity of parthenolide

iNOS and NO production

Nitric oxide (•NO) is one of the important regulator molecules in inflammatory response The synthesis and release of the •NO are mediated by an inducible isoform

of nitric oxide synthase (iNOS) Parthenolide has been demonstrated to suppress the gene expression of iNOS in several cell lines under stimulation of LPS, interferon-γ

(INFγ) or 12-o-tetradecanoylphorbol-13-acetate (TPA) (Fukuda et al., 2000) In rat

aortic smooth muscle cells, parthenolide inhibits the iNOS gene transcription which is highly correlated with its suppressed activation of NK-κB, one of regulators of iNOS gene expression (Wong and Menendez, 1999) Similar effect of parthenolide on iNOS

promoter activity was observed in human monocyte cells in conditions with TPA or without TPA stimulation and the IC50 is approximately as low as 2µM (Fukuda et al.,

2000) Moreover, in an in vivo rodent’s endotoxic shock model, reduction of iNOS

mRNA and •NO level by parthenolide has also been reported (Sheehan et al., 2002) This beneficial effect during endotoxic shock also appears to be a direct result of NF-

κB inhibition by parthenolide However, the inhibition of LPS-induced iNOS expression and •NO synthesis in central nervous system seems not related to NF-κB but rather due to an inhibitory effect on LPS-induced p42/p42 MAPK activation (Fiebich et al., 2002) These studies imply an extensive role of parthenolide in regulation of iNOS/NO in response to different stimulations

COX-2 and prostaglandin production

Prostaglandin (PG) is another group of cytokines responsible for inflammatory response During PG synthesis, cyclooxygenase (COX) plays a main regulatory role since it catalyzes the first rate-limit step in converting the 20-carbon polyunsaturated fatty acid, such as arachidonic acid, to prostaglandin G2 and H2 (Smith et al., 2000; Dubois et al., 1998) COX-2 is the highly inducible isoform of COX which plays a central role in the inflammatory processes as it is induced by pro-inflammatory cytokines and tumor promoters (Dubois et al., 1998; Williams et al., 1999) As the COX-2 promoter sequence has NF-κB responsive elements, NF-κB is believed to be one of the most important regulators of COX-2 expression (Chun and Surh, 2004) Liang et al (1999) reported that the inhibitory effect of apigenin and related flavonoids on the inducible COX-2 and iNOS expression is due to suppressing lipopolysaccharide (LPS)-induced activation of NF-κB (Liang et al., 1999) More recently, Kojima et al (2000) demonstrated further evidence that COX-2 expression

is stimulated by LPS via NF-κB in human colorectal cancer cells (Kojima et al., 2000)

However, other transcription factor responsive elements are also present on the

COX-2 promoter sequence including cyclic AMP response element (CRE) and NF-IL-6 element (Kosaka et al., 1994) and it is also reported that COX-2 expression can be up-regulated by different simulations through CRE or IL-6 regulatory mechanisms (Xie

et al., 1994; Han et al., 2003)

The effects of parthenolide on prostaglandin synthesis have been demonstrated

Collier et al (1980) first showed that feverfew aqueous extracts can inhibit

arachidonic acid metabolism and prostaglandins production without blocking the

cyclooxygenase expression in vitro (Collier et al., 1980) However, subsequent studies

by other researchers found that parthenolide inactivates the prostaglandin mediated pathway by interacting with sulphydryl amino groups of various enzymes (Pugh and Sambo, 1988) Further studies demonstrated that feverfew and parthenolide

synthetase-irreversibly suppress arachidonic acid metabolism via inhibition of cyclooxygenase

and 5-lipoxygenase activities in a leukocyte cell model (Sumner et al., 1992) Moreover, Hwang et al (1996) showed that parthenolide inhibits the inducible COX-

2 in LPS-stimulated murine macrophages cell line (RAW 264.7) Nevertheless, the molecular mechanisms involved in parthenolide’s inhibitory effect on COX-2 remain

to be further addressed

1.1.3.3 Effects on Mitogen-activated protein kinase (MAPK) pathway

The inhibition of kinase phosphorylation by parthenolide was first observed by Hwang et al (1996) in LPS-induced macrophage (RAW 264.7) cells Cells exposed to parthenolide resulted in a non-specific inhibition of MAPKs including extracellular regulated protein kinase (ERK), p38 and c-Jun N-terminal kinase (JNK) (Hwang et al., 1996) This observation was further supported by Fiebich et al (2002) who detected a similar inhibitory effect on LPS-induced ERK activation by parthenolide However,

the phosphorylation of p38 induced by LPS was not affected by parthenolide (Fiebich

et al., 2002) Meanwhile, another study conducted by Uchi et al showed that parthenolide suppresses LPS-induced, not TNFα-induced, p38 activation in dendritic cells which prevents the dendritic cells reaching maturation (Uchi et al., 2002) Moreover, they also observed that the phosphorylation of upstream kinases of p38 signaling pathway, such as MAPK kinase (MKK) 3 and MKK6, induced by LPS has also been abolished by parthenolide, indicating the possible upstream targets of parthenolide in p38 signaling pathway (Uchi et al., 2002) However, current knowledge about the potential effects of parthenolide on MAPKs is rather limited The exact role of MAPK pathway in the bioactivity of parthenolide remains to be further investigated

1.1.3.4 Effects on Janus Kinase (JAK)-Signal Transducers and Activators of

Transcription (STAT) pathway and cytokine signaling

As discussed above, NF-κB-regulated cytokines play a critical role in inflammatory responses In addition to NF-κB, the JAK-STAT signaling pathway also plays an important role in cytokine signaling and inflammatory response (Pfitzner et al., 2004) The IL-1 type cytokines (IL-1, TNFα etc.) are the central regulators in the early stage of the inflammatory process The activation of NF-κB by these cytokines up-regulates the expression of other cytokines, such as IL-6-type cytokines, which further participate in the progression of the inflammatory This process involves the JAK-STAT pathway (Heinrich et al., 1998; Hanada and Yoshimura, 2002) The possible effect of parthenolide on JAK-STATs pathway has also been addressed Sobota et al reported that pretreatment with parthenolide not only blocked the expression of IL-6 secretion but also suppressed the IL-6 signaling through inhibition

of phosphorylation of tyrosine 705 on STAT3 protein (Sobota et al., 2000) They also

postulated that the conjugation with the sulphydryl (-SH) groups located on JAKs may be responsible for the mechanism of inhibition Schnyder et al (2000) demonstrated that the nuclear translocation of the STAT6 induced by IL-4 is prevented by parthenolide However, in their human endothelial cell model, the phosphorylation of both upstream JAKs and STAT6 were not affected by parthenolide (Schnyder et al., 2002) On the other hand, it is interesting to note that parthenolide can increase the tyrosine phosphorylation of STAT5 protein and up-regulate the STAT5 regulated genes, an effect probably due to the negative cross-talk between NF-κB and STAT5 signaling pathway (Geymayer and Doppler, 2000)

1.1.3.5 Effects on cell proliferation and induction of apoptosis

The effects of parthenolide on cell proliferation and viability are less reported Hoffman et al (1977) first showed the cytotoxicity of parthenolide to human epidermoid carcinoma cell culture (Hoffmann et al., 1977) Further study was conducted by Ross and her colleagues They examined the inhibitory effects of

parthenolide on growth of two tumor cell lines in vitro The low concentration of

parthenolide can reversibly inhibit the growth of tumor cell lines in a cytostatic fashion (Ross et al., 1999)

Although the reports of cytotoxity of parthenolide can be traced back to the 1980s, the characteristics and mechanisms of parthenolide-induced cell death have not been investigated, until recently, when Wen et al (2002) reported that parthenolide could trigger apoptosis in invasive sarcomatoid hepatocellular carcinoma cells Their study provided evidence of involvement of oxidative stress, caspase activation and mitochondrial changes in parthenolide-induced apoptosis (Wen et al., 2002) They also demonstrated that the stress sensitive gene GADD153 expression was induced in parthenolide-induced apoptosis On the other hand, parthenolide induced cell death

was found to be atypical Pozarowski et al (2003b) reported, in HL-60 cells, parthenolide concurrently induces atypical apoptosis and necrosis by unknown mechanisms (Pozarowski et al., 2003b) Beside the direct apoptosis-inducing activity

of parthenolide, it has also been used in combination with other chemotherapeutic drugs In breast cancer cells (MDA-MB-231 and HBL100) which have constitutively activated NF-κB, parthenolide pretreatment significantly sensitized the breast cancer cells to chemotherapeutic drug paclitaxel-induced apoptosis (Patel et al., 2000) A recent report indicated that parthenolide also sensitizes the breast cancer cells to

TRAIL-induced apoptosis via inducing a sustained JNK activation (Nakshatri et al.,

2004) These observations suggest that parthenolide may execute its anti-cancer

property via increasing the sensitivity of cancer chemotherapy As apoptosis-inducing

ability of parthenolide is closely related to the potential of anti-cancer property, it is important to further investigate the mechanisms of parthenolide-induced apoptosis and the molecular targets of parthenolide

1.1.3.6 Effects on cell cycle regulation

Ross et al (1999) first reported that low concentration of parthenolide (5µM)

could inhibit cancer cell growth in a cytostatic fashion in vitro, implying certain

effects of parthenolide on cell cycle regulation (Ross et al., 1999) In HL-60 cells, Pozarowski et al (2003) observed a transient cell cycle arrest in G2/M phase followed

by an atypical apoptosis under a narrow concentration range (2-10µM) of parthenolide treatment This cell cycle arrest was not believed to be related to the inhibitory effects of parthenolide on NF-κB (Pozarowski et al., 2003a) A similar

G2/M arrest can also be found in hepatoma cells treated with non-cytotoxic concentration (1 to 3µM) of parthenolide (Wen et al., 2002) All these studies

indicated that, at sub-lethal concentrations, parthenolide could inhibit cell growth by inducing cell cycle arrest at G2/M

1.1.4 in vivo study of parthenolide

The pharmacological effects of parthenolide have also been studied in several in

vivo animal models In an albino mice model, both parthenolide and feverfew extracts

show an anti-nociceptive effect which significantly increases the pain threshold of mice In addition, the anti-inflammatory activity of parthenolide was also observed (Bejar, 1996) Parthenolide has also been reported to help rodents (rats and mice) overcome the endotoxin-induced shock Both pretreatment and post-treatment reduce the plasma nitrate/nitrite and lung neutrophil infiltration Moreover, the nitrotyrosine formation, iNOS mRNA content and apoptosis in thoracic aortas of rats are also abolished by parthenolide treatment Accompanied with these effects, parthenolide suppresses the NF-κB DNA binding in the lung tissue This study strongly supported

the anti-inflammatory activity of parthenolide in vivo (Sheehan et al., 2002) Another

study from the same group investigated the effects of parthenolide on induced myocardial damage in rats Intra-peritoneal administration of the parthenolide has been found to greatly ameliorate the myocardial injury while the activation the

reperfusion-NF-κB activation is concurrently inhibited in vivo (Beranek, 2002) Since parthenolide shows a potent inhibitory effect on pro-inflammatory cytokines in vitro, its effects have also been examined in in vivo animal model Smolinski and Pestka

(2003) investigated inhibitory effects of oral administration of parthenolide on cytokines release in LPS-stimulated mice Although parthenolide significantly

suppresses the LPS-induced IL-6 and TNFα secretion in vitro, the IL-6 and TNFα

level in the blood of LPS-stimulated mice are below the detection limit (Smolinski and Pestka, 2003) A more recent study further addressed this issue In a

polymicrobial sepsis rat model, parthenolide treatment significantly reduces the plasma concentrations of several pro-inflammatory cytokines, including TNFα, IL-6 and IL-10 (Sheehan et al., 2003) In short, the current knowledge of parthenolide’s

biological activities, especially the anti-cancer potential, in in vivo animal model is

rather limited and needs to be further studied

1.1.5 Toxicity and adverse side effects

Currently, the toxicity and side effects studies only carried out on Feverfew Generally, the Feverfew treatment is relatively safe In animal studies of Feverfew, no significant toxicity has been observed at doses up to 100 times higher than usual human dose No toxic effects on DNA stability or adverse effects to promote carcinogenesis have been reported by taking Feverfew (Newall et al., 1996) Some reports have suggested Feverfew may cause allergic response and contact dermatitis However, these side effects may due to potential contamination of herbal products (Paulsen et al., 1998) So far, the side effects of parthenolide have not been reported

1.2 Oxidative stress, biothiols and intracellular redox balance

1.2.1 Reactive Oxygen Species

1.2.1.1 Definition

Reactive oxygen species are known to be involved in many biological processes and in many human diseases Physically, the atoms are surrounded by multiple orbits which contain paired electrons spinning in opposite direction The free radical is defined as any species capable of independent existence that contains one or more unpaired electrons The species with unpaired electrons are highly unstable and can react with other molecules by means of oxidation or reduction reactions (Halliwell, 1991; Gilbert, 1994) In biological systems, the most important oxygen-based free

radicals are generally termed as reactive oxygen species (ROS), including superoxide (O2·-) and hydroxyl radical (·OH) (Bergendi et al., 1999) In addition, other forms of oxygen-based molecules also functionally act as free radicals in many biological

oxidative processes, including singlet oxygen (1O2) and hydrogen peroxide (H2O2) Excessive cellular free radicals and their derivatives may cause oxidative damage due

to their high reactivity towards a variety of molecular targets (Halliwell, 1999; Marnett, 2000; Cooke et al., 2003) In the biological systems, most ROS are produced as O2· - which is instantly converted by superoxide dismutase (SOD) into

H2O2 , a relatively stable form of ROS (Halliwell et al., 2000) However, H2O2 can be further converted to hydroxyl radical (·OH), the most reactive free radical responsible

for many oxidative damages, via Fenton or Haber-Weiss reaction in vivo (Fig 1.3)

(Frenkel, 1992; Halliwell et al., 2000) The potent oxidative reactivity of hydroxyl radical triggers a set of chain reactions resulting in protein oxidation, DNA oxidative damage (Cooke et al., 2003; Marnett, 2000) and lipid peroxidation (Halliwell and Chirico, 1993)

Fig 1.3 Fenton and metal catalyzed Haber-Weiss reaction

Reaction) Weiss

(Haber O

OH OH O

H O

Reaction) (Fenton

Fe OH OH O

H Fe

O Fe O

Fe

2

Fe 2 2 2

3 2

2 2

2

2 2

3

− +

+

⎯→

⎯ +

+ +

→ +

+

→ +

+

−

• +

Net:

1.2.1.2 Sources of ROS

The intracellular ROS are produced via a variety of pathways Mitochondrion

and endoplasmic reticulum are the major organelles where ROS are produced In the aerobic eukaryotic cells, mitochondrion is the central place where the oxygen is metabolized through the electron transport chain (ETC) to generate adenosine

triphosphate (ATP) There are five enzymatic complexes, which are arranged by their oxidative/reductive (Redox) potentials, located in the inner membrane of mitochondrion In the aerobic respiration, a series of oxidative/reductive reactions are catalyzed by using nicotinamide-adenine dinucleotide phosphate (reduced) (NADPH) and shuttle molecules (such as cytochrome c and coenzyme Q) as substrate to produce ATP (Forman and Azzi, 1997) Although these enzyme-catalyzed processes are highly efficient, there are still a fraction of electrons leakage resulting in generation of

O2· - (Yu, 1993) The O2·- is quickly converted by superoxide dismutase (SOD) present

in the mitochondria and converted into H2O2 which could be further converted into highly reactive ·OH Another source of intracellular ROS is endoplasmic reticulum where ROS is generated through NADP/NADPH dependent enzymatic reactions using both endogenous and exogenous substrates (Trush and Kensler, 1991; Ames et al., 1993)

1.2.2 Biothiols

In the live organism, ROS are constantly produced during the aerobic respiration To avoid potential oxidative damage, a highly regulated antioxidant defense system, which consists of a series of the enzymatic proteins and non-enzymatic molecules, has been evolved In this complicated system, the thiol buffering systems plays a pivotal role in defense of oxidative stress

1.2.2.1 Definition

Thiols are a group of organic mercaptans (R-SH) that are characterized by the presence of sulphydryl residues at their active site In cell system, the biological thiols (or biothiols) are often referred to biological mercaptans present in the cell In

general, biothiols can be classified into low molecular weight free thiols, such as glutathione (GSH), and high molecular weight protein thiols (Packer, 1995)

1.2.2.2 Biological properties and metabolism

The active sites of most biologically important thiols are located at the functional cysteinyl residues on the molecular side chain A thiolates (RS-) is produced when a thiol (R-SH) loses the H atom from the thiol group or sulphur loses

an electron followed by a proton Under physiological pH conditions, thiolates are unstable and may be crosslinked to form a disulphide linkage (-S-S-) between two –

SH group The dynamic transition of thiol/disulphide depends on cellular reduction/oxidation status and plays an important role in determination of protein structure

All the thiols also act as reducing agents ROS generated during aerobic metabolism has a strong tendency to transfer electrons to other molecules and oxidize the latter Reducing agents, such as thiols, have negative reductive potentials and thus act as prompt electron acceptors A depiction of these reactions is shown in Fig 1.4 The ROS can actively interact with thiols and the oxidant-thiol interaction neutralizes the harmful oxidants into a relatively less toxic by-product and, meanwhile, the biothiols themselves are oxidized to a disulphide (R-S-S-R)

+

− + +

Fig 1.4 Thiol and disulfides

In biological systems, there are a variety of reductases responsible for recycling disulphides back to reduced thiols by using cellular reducing equivalents such as NADPH or NADH (Paget and Buttner, 2003) The general equilibrium reaction

between thiols and disulphide is depicted in Fig 1.5 This cycling metabolism plays a central role in maintaining a favorable oxidative/reductive (Redox) milieu of thiols containing protein cysteine residues and other small molecules, such as cysteine/cystine and glutathione (GSH)/glutathione disulphide (GSSG) (Gilbert, 1995)

1.2.3 Anti-oxidant defense system

In order to avoid the oxidative stress-induced biological damage, the aerobic organisms evolved a highly regulated anti-oxidant defense system Glutathione is the most important low molecular biothiol that maintains redox balance within the cell GSH is actively synthesized in the cytoplasm by γ-glutamyl cysteine synthestase (γ-

GCS) or GSH synthetase as shown in Fig 1.6

Fig 1.5 Cycling of biothiols

GSH/TRX

Reduced form of glutathione acts as an antioxidant neutralizing the excessive ROS And the oxidative form of glutathione (GSSG) will be either transferred back to the reduced form by glutathione reductase (GRase) using NADPH or exported by cells to keep an intracellular high ratio of GSH/GSSG (Griffith, 1999) Furthermore,

GSSG/TRX-SS

Glutathione peroxidaseThioredoxin peroxidase

Oxidant

Thioredoxin reductase

Glutathione peroxidase

NADPH

GSH is also participating in the synthesis of other small molecular antioxidants such

as ascorbic acid and tocopherols (Meister, 1994)

Pi MgADP cysteine

L Glutamyl γ

L

MgATP cysteine

L Glutamate

+ +

− +

Fig 1.6 Glutathione (GSH) synthesis

In addition to GSH, there are a number of protein thiols exist in the cells The most important antioxidant non-enzymatic protein thiol system is thioredoxin (TRX) Thioredoxin is a pleiotropic NADPH-dependent disulphide oxidoreductase which catalyzes the reduction of protein disulphide bond (-S-S-) The two cysteine residues

of thioredoxin can also be reversibly oxidized into disulphide serving as antioxidant to remove free radicals and protect cell from oxidative stress Similar to glutathione, the oxidized thioredoxin will be recycled through thioredoxin reductase/peroxidase pathway (Nishinaka et al., 2001; Sen, 2000) Thioredoxin system has also been shown

to be involved in restoring ascorbate and maintaining other forms of antioxidants Thioredoxin has been shown to execute its antioxidant capability especially under the GSH-depleted condition (Iwata et al., 1997)

A group of the anti-oxidant enzymes also plays a critical role in defending oxidative stress, which includes: (i) SOD which converts superoxide radical to less harmful H2O2; (ii) catalase which further metabolize H2O2 into oxygen and water; (iii) glutathione peroxidase which is involved in GSH-GSSG cycling and removal of H2O2 (Michiels et al., 1994)

1.2.4 Redox balance

The intracellular reduction/oxidation status (Redox) is a precise balance between levels of ROS generation and endogenous thiol buffers exist in the cell

(Davis, Jr et al., 2001) Cellular redox status is important in determination of protein structure, regulation of enzyme activity and control of transcription activity In the living cells, aerobic metabolism continually produces a large number of reactive oxygen species (ROS) which lead to an oxidative stress and a potential harmful modification of functional cellular proteins On the other hand a group of low molecular thiol buffers (GSH) and a variety of reductive enzymatic pathways, such as ubiquitous disulphide reductase, thioredoxin reductase, contribute to preventing the oxidation of important proteins thiols by ROS and maintaining a relatively reduced condition status in living cells (Paget and Buttner, 2003) However, in some situations, the excessive oxidative stress induced by extracellular stimulations will disrupt the intracellular redox balance and lead the oxidative damage of many cellular targets

1.2.5 Biological consequences of redox imbalance

1.2.5.1 Lipid peroxidation

Lipid peroxidation is a free radical-initiated self propagating chain reaction It is one of the most well studied biological effects of oxidative damage Many types of free radicals can trigger the chain reaction of lipid peroxidation Firstly, the excessive free radicals attack the double bond of polyunsaturated fatty acid (PUFA) to generate

a lipid radical L· The lipid radical L· then directly reacts with O2 which, in turn, form

a lipid peroxy radical LOO· and another series of L· triggering the chain reaction of lipid peroxidation (Horton and Fairhurst, 1987; Girotti, 1985) As the cell membranes are enriched with PUFAs, it is believed to be the main target of lipid peroxidation The severe oxidation of the phospholipids in the membrane results in a series of biological effects including re-arrangement of the membrane structure, changing the permeability of the membrane, disruption of the normal function of membrane-

binding proteins and simultaneous production of carcinogenic byproducts such as malondialdehyde (MDA) (Esterbauer et al., 1990)

1.2.5.2 DNA damage

During oxidative stress, DNA can be attacked by many kinds of free radicals which results in ROS-oxidized DNA modifications, including DNA base modifications, DNA-protein cross-linking, DNA adducts formation and DNA strand breaks (Breimer, 1990; Cadet et al., 1999; Dizdaroglu, 1992) Although under physiological condition, DNA is well protected by nuclear membrane and nuclear proteins, H2O2 can diffuse freely into nuclei It can be further concerted to an extremely reactive ·OH radicals immediately adjacent to the nucleic acid molecules when iron or copper ions are presented ·OH radicals can directly attack the DNA bases and extract hydrogen atoms in genome DNA which result in an OH- form DNA modification Among these types of oxidative modifications, 8-hydroxy-2’- deoxyguanosine (8-OHdG), which has OH form guanine at C8 position of DNA, is the most well studied one The ROS oxidization and modification of DNA has been suggested to cause gene mutation and is closely related to carcinogenesis (Cooke et al., 2003; Kamiya et al., 1992)

1.2.5.3 Signal transduction

Signal transduction mainly refers to a series of phosphorylation and dephosphorylation actions of tyrosine and serine/threonine on the cellular proteins These cell signaling proteins are known to be sensitive to redox changes It has been demonstrated that the several protein tyrosine kinases can be phosphorylated by oxidative stress For instance, the Src-family protein tyrosine kinases have been found

to be activated by ROS (Vasant et al., 2003) and similar phosphorylation /activation

of PKC protein family members have also been reported (Shibukawa et al., 2003) The MAPK protein kinase family is also affected by redox changes especially the stress-activated protein kinase JNK (Inanami et al., 1999) On the other hand, redox also regulates the protein-DNA binding of transcription factors and then affects gene transcription It has been suggested that the cysteine residue in the molecule of NF-κB

is critical for maintaining its spatial structure and DNA binding capability The conserved cysteine residue in the RxxRxRxxC motif in the N-terminal of NF-κB Rel protein is required for optimal DNA-protein interactions and is regulated by redox changes (Kumar et al., 1992; Toledano et al., 1993; Hayashi et al., 1993) Furthermore, the unbalanced redox status could also disrupt intracellular calcium homeostasis and calcium signaling resulting in a variety of cellular function changes (Donoso et al., 1997)

1.2.5.4 Apoptosis

The involvement of redox in regulation of apoptosis has been extensively studied (Hampton et al., 1998; Cai and Jones, 1999) which will be discussed in detail

in the following sections

1.3 Apoptosis and Cancer

1.3.1 Introduction

Apoptosis is a description of a highly regulated and conserved program cell death in eukaryotic cells which is morphologically and biochemically distinct from necrotic cell death It is a genetic controlled process that is involved in embryological development and maintenance of tissue homeostasis in eukaryotic organisms Dysregulation of apoptosis process has been shown to be implicated in tumor development, chemoresistance, autoimmune syndromes and many other human

diseases (Hengartner, 2000; Kaufmann and Earnshaw, 2000; Kaufmann and Hengartner, 2001)

During apoptosis, the cells show a series of distinct morphological changes including plasma membrane blebbing, nuclear chromatin condensation, and finally degradation of apoptotic cell and formation of small membrane-enclosed apoptotic bodies Simultaneously, apoptotic cell is also under a chain reaction of biochemical changes, including flipping of phosphatidylserine to outer cell membrane, cleavage/activation of caspases, dysfunction of the mitochondrial and specific cleavage at DNA histone octomers which results in DNA ladder formation (Kaufmann and Hengartner, 2001; Hengartner, 2000) Different from necrotic cell death, all the apoptotic cell death processes do not trigger any inflammatory response

in vivo

There are generally three phases in apoptosis process: induction/initiation phase, execution phase and degradation phase In the initiation stage, the changes of cellular environment trigger the initiation of apoptosis cascade There are two classical pathways that lead to activation of caspase: the death receptor pathway and mitochondrial pathway The death receptor pathway starts from a group of death receptors (DR) Ligation with their respective ligands leads to recruitment of other adaptor proteins and formation of death-inducing signaling complex (DISC), which results in cleavage and activation of initiator caspase thereby transducing the death signaling to effectors In the mitochondrial-dependent pathway, disruption of mitochondrial normal function, e.g loss of the mitochondrial potential and release the mitochondrial pro-apoptotic proteins, also leads to execution phase of apoptosis In addition, there is a cross-talk between death receptor pathway and mitochondrial pathway, which the initiator caspase (caspase 8) cleaves Bcl-2 protein Bid and the

truncated Bid (tBid) translocate to mitochondria (Luo et al., 1998; Li et al., 1998) Death receptor pathway and mitochondrial pathway converge at effector caspases (e.g caspase 3) and it is believed that once the cell enters the execution phase, the cell death process becomes irreversible In the final degradation phase, the activated effectors facilitate the apoptotic degradation and by targeting/activating of more than

100 cellular substrates, including DNase II, caspase-activated deoxyribonuclease (CAD) and cytoskeletal proteins which responsible for the distinct morphological changes (Hengartner, 2000) Finally, cells undergoing apoptosis are marked themselves with certain kinds of “eat me” signals, such as exposure of phosphatidylserine, to help phagocytes to identify and remove them by engulfment and degradation (Savill et al., 1993; Savill and Fadok, 2000)

1.3.2 Cell death receptors

Cell death receptors includes a subset of the tumor necrosis factor (TNF) receptor super family, comprising of TNF receptor 1(TNFR1/CD120a), Fas (CD95/APO-1), death receptor 3 (DR3/APO-3/LARD/TRAMP/WSL1), death receptor 6 (DR6) and the two TRAIL (TNF-related apoptosis-inducing ligand) receptors: TRAIL receptor 1 (DR4/APO-2) and TRAIL receptor 2 (DR5/KILLER/TRICK2) They are characterized by a 2-5 cysteine-rich extracellular repeats and an intracellular death domain (DD) which is essential for transducing the apoptotic signal (Schmitz et al., 2000) The death receptors are activated through binding of their native ligands Upon receptor ligation, the death domain acts as a docking site for interaction with death domain-containing cytoplasmic proteins such

as FADD (Fas-associated death domain protein, also named MORT1) (Boldin et al., 1995; Chinnaiyan et al., 1995) and TRADD (TNF-R1-associated death domain protein) (Hsu et al., 1995) FADD serves as an adapter protein that binds directly to

Fas as well as other death receptors known as TNF-R1, DR3, DR4, DR5 and possibly

other related receptors via TRADD (Hsu et al., 1996; Chinnaiyan et al., 1996) The

death effector domain at the N-terminal end of FADD then interacts with corresponding death motif in the prodomain of pro-caspase 8 or caspase 10 to form the death-inducing signaling complex (DISC) (Muzio et al., 1996; Peter and Krammer, 2003; Curtin and Cotter, 2003) The oligomerization of caspases 8 on the DISC help

to promote the autoproteolytic activation and the activated caspase 8 subunit is then released into cytoplasm (Martin et al., 1998; Salvesen and Dixit, 1999)

Caspases are a group of highly conserved cysteine proteases that are homogenous to each other sharing the property of cleavable aspartate residues at the carboxyl side (Strasser et al., 2000; Earnshaw et al., 1999) Caspases are synthesized

as inactive zymogens containing an N-terminal prodomain, a large subunit and a small subunit Each caspase has a substrate specific sequence of four residues (Asp-

Xxx) It becomes activated by transactivation or cleavage via upstream proteases in an

intracellular cascade and transformed into tetramer comprising two identical large subunits and another two identical small subunits (Strasser et al., 2000) Among all the caspases, caspase 2, 3, 6, 7, 8, 9, 10 and 12 have been shown to be involved in

Fig 1.7 Mitochondria and Bcl-2 family: the central point of apoptosis signaling

B ax

B cl- 2

Death receptor independent stimuli

The activity of caspase is regulated by a group of the inhibitors of apoptosis (IAPs) On the death receptor pathway, the activation of initiator caspases can be regulated by polypeptides termed FLICE inhibitory protein (FLIP) which inhibits the recruitment and activation of pro-caspase 8 (Kataoka et al., 1998; Scaffidi et al., 1999) On the other hand, the activation of caspase 3 and 7 can be negatively

regulated by binding of IAPs such as cIAP1/2, XIAP and survivin (Deveraux and Reed, 1999) The final caspase activity is under a precise regulation balanced by pro-apoptosis signals and apoptosis inhibitory proteins When the effector caspases have been activated, they execute their death mission by cleaving extensive cellular substrates, such as cytoskeleton protein, to disassemble the cell structure In addition, activation of endonuclease leads to cleavage of the chromosome and internucleosomal spacers, resulting in the nuclear degradation (Strasser et al., 2000)

1.3.4 Bcl-2 protein family and mitochondria

The Bcl-2 family members are characterized with Bcl-2 homology (BH) domains (BH1-BH4), a series of homologous α-helical amino acid stretches on the protein (Gross et al., 1999; Curtin and Cotter, 2003) Based on their cellular function and structural properties, they can be classified into two groups: (i) anti-apoptotic proteins which specifically contain BH4 domain, including Bcl-2, Bcl-XL, Bcl-w, Mcl-1 Structurally, these anti-apoptotic Bcl-2 proteins also contain a c-terminal hydrophobic tail which facilitates the attachment of these proteins to intracellular membranes, such as mitochondrial membrane, and antagonizing the pro-apoptotic Bcl-2 protein by forming a heterodimer; (ii) pro-apoptotic proteins which contain an essential pro-apoptotic BH3 domain including Bax, Bak but lack of BH4 domain In addition, there is a subgroup of pro-apoptotic proteins which contains a single BH3 domain, such as Bid, Bad, Bim Currently, Bcl-2 proteins have been considered as the main regulators in controlling the release of cytochrome c and other apoptosis promoting proteins from mitochondria (Gross et al., 1999; Hengartner, 2000; Curtin and Cotter, 2003)

The mitochondrion is the central organelle responsible for the aerobic ATP production in mammalian cells It was first observed that the mitochondria are

required in apoptosis process in isolated nuclei (Newmeyer et al., 1994) and the involvement of mitochondrial cytochrome c (cyto c) in caspase 9 activation along with ATP and apoptosis-activating factor 1 (Apaf-1) was further revealed (Li et al., 1997; Zou et al., 1997) Disruption of normal mitochondria function may result in apoptosis through three mechanisms: (i) disruption of electron transportation and energy metabolism (ATP production); (ii) generation of excessive ROS and disruption

of redox balance; (iii) dissipation of mitochondrial membrane potential (Δψm) and release of pro-apoptotic mitochondrial proteins including cytochrome c, apoptosis-inducing factor (AIF), second mitochondrial activator of caspases (Smac) and serine protease Omi/HtrA2 (Kroemer and Reed, 2000; Desagher and Martinou, 2000; Martins et al., 2002) Among all the mechanisms, the loss of mitochondrial inner trans-membrane potential is a decisive step to drive the cell into an irreversible apoptotic cell death

Multiple lines of evidence have implied the two major players in mitochondrial release of pro-apoptotic factors: the mitochondrial permeability transition (PT) pore and the Bcl-2 family members The structure of PT pore is still elusive However, it is generally believed that the PT pore contains: (i) voltage-dependent anion channel (VDAC) which is located on the mitochondrial outer membrane; (ii) adenylate translocator (ANT) which resides in the inner membrane and (iii) cyclophilin D, a water soluble matrix protein (Kroemer and Reed, 2000; Desagher and Martinou, 2000) Under apoptosis stimulation, the opening of this non-specific PT pore results in

an immediate dissipation of H+ gradient across the mitochondrial inner membrane which leads to an uncoupling of the electro transportation chain and a progressive osmotic swelling of the matrix This swelling will ultimately cause the rupture of outer mitochondrial membrane and release of pro-apoptotic proteins from

intermembrane space to cytosol (Green and Reed, 1998; Desagher and Martinou, 2000) The theory of this outer-mitochondrial membrane rupture has been supported

by the observation that the specific PT pore inhibitors, e.g bongkrekic (an antagonizing ligand of ANT) or cyclosporine A (a ligand of cyclophilin D), can significantly inhibit the apoptosis by several apoptosis inducers (Walter et al., 1998; Sugano et al., 1999; Budihardjo et al., 1999) However, a couple of studies provided the evidence that the PT pore inhibitors can not completely prevent apoptotic cell death and the cytochrome c release happens before or independent of Δψm dissipation which suggests a possible different regulatory mechanism (Goldstein et al., 2000; Yang et al., 1997b; Bossy-Wetzel et al., 1998)

Extensive studies have suggested a critical role the Bcl-2 family member in regulation of mitochondrial apoptosis pathway The anti-apoptotic Bcl-2 proteins, such as Bcl-2 and Bcl-XL, exist predominantly on the mitochondrial membrane On the contrary, other pro-apoptotic Bcl-2 proteins (Bax, Bak, Bid, Bim, Bad) appear to

be cytosolic under normal conditions Upon apoptotic stimulation, these proteins will undergo a variety of post-transcriptional modifications and translocate to mitochondrial membrane, facilitating death signaling transduction from cytosol to mitochondria and triggering the release of cytochrome c (Gross et al., 1999; Adams and Cory, 2001) Several post-transcriptional modification models have been elucidated: (i) dimerization; Bax is a pro-apoptotic Bcl-2 protein which exists monomerically in the cytosol or loosely attached to the mitochondrial outer membrane in the normal physiological conditions Various death stimulations, especially the truncated form of Bid (tBid), can trigger an N-terminal conformational change of Bax and cause a consequent Bax translocation to mitochondrial membrane The translocated Bax is dimerized into homodimer and inserts into mitochondrial

outer membrane to form a pore leading to a cytochrome c release (Gross et al., 1998; Epand et al., 2002; Desagher et al., 1999; Eskes et al., 2000) This Bax-induced cytochrome c release can be independent of mitochondrial PT pore (Eskes et al., 1998) Another pro-apoptotic Bcl-2 protein Bak, which resides on the mitochondria, has the similar function as Bax With presence of tBid, it is dimerized into homodimer to promote release of cytochrome c (Griffiths et al., 1999; Wei et al., 2000) Or it can also interact with Bax, acting as a redundant apoptosis promoter (Wei

et al., 2001; Mikhailov et al., 2003) While the pro-apoptotic Bax and Bak promote

cell death via dimerization and pore formation, the membrane-integrating

anti-apoptotic proteins Bcl-2 and Bcl-XL antagonize their pro-apoptotic effects by forming

of heterodimers with Bax or Bak It has been reported that over expression of the

Bcl-2 protein or Bcl-XL significantly inhibits the Bax N-terminal conformational changes and the consequent Bax mitochondrial translocation and cytochrome c release (Murphy et al., 1999; Finucane et al., 1999; Murphy et al., 2000b; Murphy et al., 2000a; Zamzami et al., 2000; Kim et al., 2000) (ii) Cleavage; In the cell-death receptors induced apoptosis pathway, Bid is cleaved into a p15 C-terminal fragment

by the activated caspase 8 (Li et al., 1998; Luo et al., 1998) As discussed above, the truncated Bid then translocates to mitochondria and induces cytochrome c release by either tBid direct insertion (Luo et al., 1998) or cooperation with Bax/Bak (Korsmeyer

et al., 2000) However, the possible interaction between Bid and proposed mitochondrial PT pore is still elusive (Kim et al., 2000; Li et al., 1998) (iii) Phosphorylation; Bad, a pro-apoptotic BH-3 domain only Bcl-2 protein, is regulated

by phosphorylation (Bassik et al., 2004) With presence of survival signals, Bad is phosphorylated at three serine sites (Ser-112, Ser-136 and Ser 155) and sequestered

by cytosolic protein 14-3-3 Upon the death signals, Bad is phosphorylated and

de-associated from 14-3-3 and then translocates to mitochondria and competitively interacts with Bax which diminishes the inhibitory effects of the Bcl-2-Bax interaction (Zha et al., 1996; Zha et al., 1997; Klumpp and Krieglstein, 2002) The forming of Bad-Bax heterodimers further promotes cytochrome c release In conclusion, although the exact mechanism of regulation of cytochrome c release during apoptosis is very complex, it is well established that Bcl-2 family proteins have

an important regulatory role in mitochondrial apoptotic pathway and the mitochondrial dysfunction is the decisive step, so called “a step of no return”, in apoptosis process induced by a variety of stimulations The major Bcl-2 members involved in the mitochondrial apoptotic pathway are shown in Fig 1.7

1.3.5 Other important regulators in apoptosis

1.3.5.1 Thiols and intracellular redox balance in apoptosis

As discussed above, in aerobic mammalian cells, the mitochondrion is the central organelle which constantly produces ATP using oxygen and generates free radicals Free radicals generated from aerobic respiration can be neutralized by cellular antioxidant buffering system including SOD, GSH and protein thiols to maintain the redox balance (Curtin et al., 2002) In the mitochondrion, GSH is an important thiol buffer The majority of mitochondrial GSH is transported from cytoplasm to maintain a similar content in the matrix as that in cytoplasm (Smith et al., 1996; Lash and Jones, 1983) Compared to GSH in cytoplasm, the mitochondrial GSH is more resistant to exogenous treatment which causes GSH depletion (Meister, 1995; Fernandez-Checa et al., 1998) However, the overwhelmed ROS generation, caused by either dysfunction of mitochondria or severe depletion of cytosolic thiol buffering system, will lead to a disruption of redox balance and oxidative stress The thiol depletion-induced oxidative stress enforces the mitochondrial permeability

transition and cytochrome c release which eventually result in an apoptotic cell death (Meister, 1995; Smith et al., 1996; Armstrong and Jones, 2002; McStay et al., 2002) The cellular redox status also affects other apoptosis regulators such as Bcl-2 protein (Voehringer, 1999), caspases (Hampton et al., 1998) and anti-apoptotic protein NF-κB (Sen et al., 1997), in regulating apoptosis

1.3.5.2 Endoplasmic reticulum (ER) stress and Ca2+ in apoptosis

The ubiquitous cellular Ca2+ acts as a second messenger responsible for a wide range of cellular biochemical processes including muscle contraction, chemotaxis and energy metabolism (Ermak and Davies, 2002) The intracellular Ca2+ homeostasis is tightly controlled by cytosolic Ca2+ stores and Ca2+ pumps (Berridge, 1997) The Intracellular redox balance largely affects the cellular calcium signaling Severe oxidative stress leads to a prolonged ER stress and a rapid Ca2+ influx into cytoplasm from both the biggest intracellular Ca2+ store ER and the extracellular environment The cytosolic Ca2+ burst enforces a further influx of Ca2+ into mitochondrion and nuclei In mitochondrion, the severe oxidative stress-induced aberrant Ca2+ level switches the mitochondrial Ca2+ storage from a physiologically beneficial process to a cell death signal by disrupting the respiratory metabolism and normal function of mitochondria In nuclei, excessive Ca2+ promotes apoptosis by modulating the gene transcription and nucleases activity (McConkey and Orrenius, 1997; Ermak and Davies, 2002) Meanwhile, the prolonged ER stress also triggers the cleavage and activation of the effector caspase 12, which is localized on the ER membrane Once it has been activated, it can work coordinately with calpains, a group of Ca2+ dependent cysteine proteases, to induce apoptosis (Nakagawa et al., 2000; Nakagawa and Yuan, 2000) It is interesting to note that the Bcl-2 family members, especially Bcl-2 protein, also subcellularly distributed in ER and nuclear membrane which act as a regulator of

Ca2+ transportation and inhibits ER stress-induced apoptosis (Distelhorst et al., 1996; Bassik et al., 2004) These data further imply the importance of Ca2+ signaling in regulation of apoptosis

1.3.5.3 MAPK in apoptosis

The mitogen-activated protein kinase (MAPK) family is an important group of proline-directed serine/threonine kinases responsible for a wide range of cellular responses The MAPK mainly includes p42/44 extracellular signal-related kinases (ERK 1/2), p38 MAP kinase and c-Jun N-terminal protein kinase (JNK/SAPK) (Seger and Krebs, 1995) Due to their extensive cellular effects, the role of MAPK in regulation of apoptosis has triggered a great interest in recent years Although the role

of MAPK in apoptosis is highly context-dependent, it is generally believed that ERK tends to play an anti-apoptotic role while the JNK and p38 MAPK are promoting apoptosis (Xia et al., 1995) The ERK activation plays an important role in cell proliferation and apoptosis Several lines of evidence indicate that ERK activation, induced by a broad range of stimulations such as growth factor withdraw, oxidative stress and chemotherapeutic drugs, inhibits apoptosis (Anderson and Tolkovsky, 1999; Wang et al., 1998b; Erhardt et al., 1999) As for p38 MAPK, exact mechanism of its role in apoptosis is still unknown However, the p38 specific inhibitors can block apoptosis induced by different stimuli in various cell lines suggesting the anti-apoptotic role of p38 (Ono and Han, 2000; Kuhn et al., 2003; Koul, 2003; Tikhomirov and Carpenter, 2004) JNK is the key regulator of transcription factor c-Jun activity The involvement of JNK in apoptosis seems to be more complex and highly controversial Depending on cell type, property of stimulus and the duration of JNK activation, JNK can play anti-apoptotic, pro-apoptotic or no roles in apoptosis (Lin, 2003) The pro-apoptotic role of JNK was first suggested in neuronal cells (Kuranaga

and Miura, 2002; Yang et al., 1997a) and studies on JNK1 and JNK2 double out mouse embryonic fibroblasts (MEFs) It provides convincing evidence that JNK activation plays a pro-apoptotic role in UV-induced apoptosis (Tournier et al., 2000)

knock-As a protein kinase, JNK has also been found to phosphorylate Bcl-2 family proteins (Bim and Bif) to induce a Bax-dependent apoptosis (Lei and Davis, 2003) and Deng

et al (2003) provided further evidence of the essential role of JNK in TNF-induced apoptosis (Deng et al., 2003) On the other hand, in certain circumstance, the JNK activation can serve as an anti-apoptotic regulator by phosphorylating Bad (Yu et al., 2004) Recently, it has been shown that the duration of JNK activation is a decisive factor for the role of JNK in TNFα-induced apoptosis A prolonged JNK activation is more likely to promote apoptosis when compared to the transient activation (Tang et al., 2001; Lin, 2003) In summary, current evidence tends to suggest that JNK is a conditional regulator of apoptosis but not an intrinsic pro- or anti- apoptosis protein Further studies are needed to elucidate the precise role of JNK in regulation of apoptosis

1.3.6 Dysregulated apoptosis in cancer

Cancer is a general term for the disease characterized by unscheduled and uncontrolled cellular proliferation Tumorigenesis or carcinogenesis is a multi-step process that consists of a series of mutations in key cellular genes which allow the cancer cells to grow uncontrollably by evading growth-inhibitory signals and host immune responses The aberrant ability of cancer cells to replicate indefinitely, exhaust oxygen and nutrition supply, and invade adjacent and distant tissues make cancer one of the most lethal diseases (Evan and Vousden, 2001) And most importantly, the dysregulation of apoptosis is an essential hallmark of cancer which leads to disruption of cellular homeostasis

Cellular homeostasis is regulated by an intrinsic apoptosis pathway which implicates a number of pro-apoptotic and anti-apoptotic proteins The disruption of apoptosis pathway is a common event observed in most cancers First of all, the anti-apoptotic protein Bcl-2 has been found to be over expressed in many cancers (Reed, 1999) On the other hand, the pro-apoptotic Bax and Bak proteins have been found to

be down-regulated or mutated (Rampino et al., 1997; Kondo et al., 2000) In addition

to Bcl-2 protein family, other apoptosis regulators have also been dysregulated during carcinogenesis The FLIPs have been found to be up-regulated in some cancers which inhibits DISC formation and caspase 8 activation conferring chemoresistance of cancer cells to anti-cancer drugs (Tepper and Seldin, 1999; Wajant, 2003) The mutation of tumor suppressor gene PTEN results in a persist activation of protein kinase B (also known as Akt) which phosphorylates pro-apoptotic Bad protein and leads to Bad sequestration in cytosol and thus inhibits the pro-apoptotic property of Bad (Di Cristofano and Pandolfi, 2000) Moreover, the IAPs, such as XIAP, are also frequently up-regulated in many tumors which prevents the apoptotic process downstream of mitochondrial changes (Deveraux and Reed, 1999) In addition, the tumor suppressor gene p53 is an important gene closely involved in regulation of

apoptosis via transcription-dependent or transcription-independent pathway Many of

pro-apoptotic Bcl-2 family proteins, such as Bax, Bak, are transcriptionally activated

by p53 whereas the anti-apoptotic protein (Bcl-2, Bcl-XL, IAPs) are inhibited by p53 (Ryan et al., 2001; Wu et al., 2001) Moreover, p53 activation also enhances the transcription of cell death receptors, e.g CD95 and DR5, to sensitize the chemotherapeutic-induced apoptosis in cancer cells (Herr and Debatin, 2001; Ryan et al., 2001) However, the fact that mutation-induced loss of p53 function which occurs

in more than half of human cancers implicates the importance of p53 in apoptosis within cancer cells (Hofseth et al., 2004)

Due to the importance of apoptosis regulation in carcinogenesis, numerous chemopreventive and chemotherapeutic approaches have been investigated by focusing on the regulation of apoptosis in tumors These approaches can be classified into: (i) induction of apoptosis within tumors by either cell death receptors ligands TRAIL/CD95L; (ii) introducing chemotherapeutic drugs by synthetic activation of caspases with chimeric inducible caspases or modulation mitochondria function by pro-apoptotic Bcl-2 family proteins (Ferreira et al., 2002); (iii) induction of apoptosis

by apoptosis-permissive approaches or chemosensitization approaches such as treatment of chemopreventive drugs which inhibit anti-apoptotic pathways (NF-κB, Akt etc.) to sensitize cancer cells to apoptosis-inducing agents (Makin and Dive, 2001; Ferreira et al., 2002) Understanding the in-depth mechanism of drug-induced apoptosis will definitely help in exploring more effective chemotherapeutic targets and to overcome chemoresistance

co-1.4 TNF and NF-κB activation

Tumor necrosis factor (TNF) is one of the earliest cytokines that has been found

to possess anti-cancer bioactivity (Beutler and Cerami, 1992) A large number of studies are focusing on the ability of TNF-induced apoptosis which is believed to be the major mechanism for its anti-cancer property However, TNF signaling is also a double-edged sword Simultaneous activation of anti-apoptotic NF-κB pathway impedes the clinical application of TNF (Aggarwal, 2003)

1.4.1 TNF superfamily and TNF-induced apoptosis

Two TNF isoforms were first isolated in the 1980s (Aggarwal et al., 1985) and the extensive follow-up studies elucidated that TNF superfamily is a big group of proteins with 19 members which execute their cellular functions through corresponding 29 cell membrane receptors (Aggarwal, 2003) The TNF superfamily members have been identified to be the essential cytokines in regulation of a great number of cellular events such as immune responses, cell differentiation, cell proliferation and cell death (Gaur and Aggarwal, 2003) Among all the TNF family members, the TNFα, CD95 ligand (CD95L) and TRAIL are the factors closely related

to death receptor-mediated apoptosis TNFα has been shown to induce apoptosis via

binding to its cell membrane-located receptor TNF receptor 1(TNFR1) in certain cell types (Chopra et al., 2004; Neu et al., 2003) TNFR1 contains a similar 2-5 cysteine-rich extracellular repeats and an intracellular death domain (DD) which is responsible for transducing the apoptotic signal (Schmitz et al., 2000) The ligation of TNFα to TNFR1 triggers the oligomerization of membrane associated protein TNF-R1-associated death domain protein (TRADD) which serves as a docking site for interaction with death domain-containing cytoplasmic proteins such as FADD (Fas-

associated death domain protein, also named MORT1) via its death domain (DD)

(Boldin et al., 1995; Chinnaiyan et al., 1995) FADD acts as an adapter protein that is

recruited to TNFR1 via TRADD (Hsu et al., 1996; Chinnaiyan et al., 1996) The death

effector domain at the N-terminal end of FADD then interacts with corresponding death motif on the prodomain of pro-caspase 8 or caspase10 to form the death-inducing signaling complex (DISC) (Muzio et al., 1996; Peter and Krammer, 2003; Curtin and Cotter, 2003) The oligomerization of caspases 8 on the DISC results in an autoproteolytic activation and release of the activated caspase 8 subunit, to cytoplasm