with Ionizing Radiation

5 Combinations of Hypoxia-Targeting Compounds and Radiation-Activated Prodrugs

with Ionizing Radiation

G-One Ahn and J. Martin Brown

G. Ahn, P hD J. M. Brown, P hD

Division of Radiation and Cancer Biology, Department of Radiation Oncology, Stanford School of Medicine, 269 Campus Drive, Center for Clinical Science and Research, Rm 1255, Stanford, CA 94305-5152, USA

5.1

Targeting Tumor Hypoxia

Tumor hypoxia was first postulated from histo- logical studies of human lung adenocarcinomas by Thomlinson and Gray (1955). They reasoned that, because of unrestrained growth, tumor cells are forced away from blood vessels beyond the effec- tive diffusion distance of oxygen (O

) in respiring tissues, hence becoming hypoxic and eventually necrotic (Fig. 5.1a). Given the typical values for intracapillary O

tensions and consumption rates, they calculated that O

diffusion distances would be approximately 150 Pm and this was consistent with their histological observations (Thomlinson and Gray 1955). This type of hypoxia has come to be termed “chronic,” or “diffusion-limited,” hypoxia.

Acute hypoxia also develops in tumors through temporal (reversible) cessation or reduction of tumor blood flow resulting from highly disorga- nized tumor vasculature (Fig. 5.1b; Brown 1979).

Definitive evidence for acute hypoxia and fluctu- ating blood flow has been demonstrated in trans- planted tumors in mice injected at some time apart with two different diffusion limited fluorescent dyes showing mismatch of labeled cells (Chaplin et al. 1986; Trotter et al. 1989); however, acute and chronic hypoxia are in fact the two ends of a con- tinuum with fluctuations in blood flow without total occlusion, which are common in both experimental (Kimura et al. 1996) and human tumors (Hill et al. 1996), producing a dynamic situation with fluc- tuating oxygen diffusion distances in many parts of tumors.

Tumor hypoxia is a major factor contributing to the failure of radiotherapy (Fig. 5.2). This is largely because DNA damage produced by ionizing radia- tion, which would otherwise become fixed and lethal to cells by reacting with O

under well oxygenated conditions, can be restored to its undamaged form under hypoxic conditions (Brown and Wilson 2004). Clinically hypoxia predicts poor local control and survival of patients undergoing radiotherapy

CONTENTS

5.1 Targeting Tumor Hypoxia 67 5.2 Oxygen-Level Enhancers 68

5.3 Hypoxia-Selective Radiosensitizers 69 5.3.1 Nitroimidazoles 69

5.3.2 Mixed-Function Radiosensitizers 69 5.3.3 DNA-Affinic Radiosensitizers 71 5.3.4 Limitations of HSR 71

5.4 Hypoxic Cytotoxins 72 5.4.1 Introduction 72

5.4.2 Combination of Hypoxic Cytotoxins with Ionizing Radiation 72

5.4.3 Nitroimidazoles 73 5.4.4 Other Nitroaromatics 73 5.4.4.1 CB 1954 73

5.4.4.2 SN 23862 and PR-104 74 5.4.4.3 RSU 1069 and RB 6145 74 5.4.5 Quinones 75

5.4.5.1 Mitomycin C 75 5.4.5.2 Porfiromycin 75 5.4.5.3 EO9 75

5.4.6 Benzotriene di-N-Oxides 76 5.4.7 Tertiary Amine N-Oxides 78 5.4.7.1 Nitracrine N-Oxides 78 5.4.7.2 AQ4N 79

5.5 Combination of Radiation-Activated Prodrugs with Ionizing Radiation 79

5.5.1 Concepts of RAP 79

5.5.2 Nitro(Hetero)Cyclic Methylquarternary Ammonium Salts 79

5.5.3 5-Fluorouracil (5-FU)-Releasing Prodrugs 81 5.5.4 Transition Metal Complexes 81

5.6 Other Hypoxia-Targeting Strategies 82 5.6.1 GDEPT Targeting Tumor Hypoxia 82

5.6.2 Clostridia-Directed Enzyme Prodrug Therapy 82 5.6.3 Targeting HIF-1 83

5.7 Conclusion 83

References 83

for carcinoma of the head and neck (Nordsmark et al. 1996; Brizel et al. 1997), and cancer of the cervix (Hockel et al. 1993; Fyles et al. 1998).

Hypoxia further complicates cancer manage- ment by limiting the access of conventional chemo- therapeutic drugs (Fig. 5.1a; Brown and Wilson 2004). Hypoxia also increases genomic instability by increasing mutation frequency (Reynolds et al.

1996) or selecting for cells expressing an anti-apop- totic phenotype such as mutated p53 (Graeber et al.

1996). This leads to a more metastatic phenotype as has been observed clinically (reviewed by Rofstad 2000). In addition, expression of proangiogenic pro-

teins, such as vascular endothelial growth factor (VEGF), are increased under hypoxic conditions, potentially resulting in increased tumor angiogene- sis (Shweiki et al. 1992; Fang et al. 2001). Thus, there is substantial evidence that hypoxia both interferes with the effective therapy of solid tumors and con- tributes to a more malignant phenotype; however, hypoxia may also prove to be a therapeutic advan- tage: because it is virtually unique to tumor cells, therapies that target hypoxic regions may have the potential to kill malignant cells while leaving non- malignant cells relatively untouched. This chapter discusses some examples of hypoxia targeting com- pounds and approaches for combination with ioniz- ing radiation in experimental or clinical settings.

5.2

Oxygen-Level Enhancers

One of the earliest attempts to overcome the prob- lem of the resistance of hypoxic cells in tumors to radiotherapy was to increase O

levels in the blood stream, thereby increasing the diffusion distance of O

. A number of trials were performed with patients breathing 100% O

at a pressure of 3 atmo- spheres, but the results were mixed (Watson et al.

1978; Dische et al. 1983; Henk 1986). One potential reason for such failures is that increasing the diffu- sion distance of O

would not be expected to reduce the levels of acute hypoxia. In some systems, the use of carbogen (95% O

/5% CO

) appears to have

Fig. 5.1. a A diagram of a tumor capillary and surrounding tumor cells at decreasing oxygen concentrations (in the direction of arrows). Cells become hypoxic (green) and eventually necrotic (blue); chronic hypoxia. Cellular proliferation and chemo- therapeutic drug concentration are also decreasing in the same direction, as a function of distance from the capillary. (From Brown 1999). b A diagram of normal (left) and tumor (right) blood vasculature. The tumor vasculature is highly disorganized resulting in acutely hypoxic regions in the tumor. (From Brown and Giaccia 1998)

Fig. 5.2. A Kaplan-Meier plot of overall survival of patients with head and neck carcinoma undergoing radiotherapy.

Well-oxygenated tumors (pO

>10 mmHg; dotted line) showed better prognosis than poorly oxygenated (pO

<10 mmHg;

solid line) tumors. (From Brizel et al. 1997)

a b

greater effect than 100% O

in increasing O

level in the blood stream (Rockwell 1997) possibly by pre- venting the vasoconstriction caused by high partial pressures of O

.

Other potential approaches to overcome hypoxia include the use of nicotinamide (Horsman et al.

1987; Chaplin et al. 1991; Kjellen et al. 1991) in conjunction with carbogen (Corry and Rischin 2004), agents to increase tumor blood flow such as flunarizine (Jirtle 1988), artificial blood substi- tutes carrying increased levels of O

(Teicher and Rose 1984; Rockwell et al. 1986), drugs to reduce the affinity of hemoglobin for O

(Hirst and Wood 1989), blood transfusion (Bush et al. 1978), and hyperthermia (Song et al. 2001). Recently, RSR13, a drug reducing hemoglobin O

-binding affinity, has been claim to benefit non-small cell lung cancer patients receiving radiotherapy in phase-II trials (Choy et al. 2005).

5.3

Hypoxia-Selective Radiosensitizers

In the 1960s Adams and Cooke (1969) proposed that electron-affinic drugs might act like O

, a potent electron-affinic molecule, to sensitize hypoxic tumor cells. These agents (hypoxia-selective radi- osensitizers (HSR) mimic O

by reacting with the short-lived DNA free radicals generated by ionizing radiation; however, unlike O

, HSR are not rapidly metabolized by the cells through which they pen- etrate and are thus able to reach areas beyond the O

diffusion distance. Some examples of HSR are discussed below.

5.3.1

Nitroimidazoles

Metromidazole and misonidazole (Fig. 5.3) are the prototype members of this class. They were admin- istered to patients in the 1970s. Disappointingly, they showed very little efficacy sensitizing tumors but a high incidence of peripheral neuropathy (Urtasun et al. 1976; Dische et al. 1977). With the limited dose of misonidazole that can be administered to patients, almost all of the clinical trials of radio- therapy combined with misonidazole were negative (Dische 1985); however, a meta-analysis of 50 ran- domized trials later showed a small but significant benefit of misonidazole and other hypoxic radio- sensitizers when added to radiotherapy in head and neck cancers (Overgaard 1994).

Attempts to produce superior drugs to misoni- dazole resulted in the development of etanidazole (Fig. 5.3; Brown et al. 1981), pimonidazole (Fig. 5.3;

Smithen et al. 1980), and nimorazole (Fig. 5.3;

Dische 1985); of these, nimorazole showed a signifi- cant benefit of loco-regional control when given in conjunction with radiotherapy to patients with inva- sive carcinoma of larynx and pharynx (Overgaard et al. 1998; Overgaard et al. 2005) and is now given as part of the standard of care for radiotherapy of head and neck cancer patients in Denmark (Table 5.1).

5.3.2

Mixed-Function Radiosensitizers

The neurotoxicity of misonidazole stimulated the search for drugs not only with less toxicity, but also with increased efficiency as radiosensitizers on a

Fig. 5.3. Examples of nitroimidazole compounds as hypoxia-selective radiosensitizers (HSR)

Metronidazole Misonidazole Etanidazole

Pimonidazole Nimorazole

T able 5.1. E xam ples o f nit ro imidazole radiosen sit izer s and r es ul ts f ro m the clinical t rial s N ame o f dr ug Dr ug eval ua- ti o n c at eg o ry Ty p e o f tr ial T y pe o f canc er T y pe o f r adio ther apy N o . o f pa ti en ts O utcom e R efe re n ce Mis o nidazole HS R Phase III L o cal ly ad va nc ed no n- small c ell l u ng c anc er C o n ven ti o n al radio therap y 239 N o si g n ifi can t b enefi t S im pson et al. (1989) Car cino ma o f the u te rine c er v ix C o n ven ti o n al radio therap y 331, 120 N o si g n ifi can t b enefi t O v e r g aar d et al. (1989a); G r igs b y et al. (1999) Car cino ma o f the ce rv ix Radical radio therap y 73 N o si g n ifi can t b enefi t C ha n et al. (2004) In vasi ve car cino ma o f lar y nx and phar y nx Sp lit-c o ur se radio therap y 626 N o si g n ifi can t b enefi t O v e r g aar d et al. (1989b) A d vanc ed head and neck canc er C o n ven ti o n al f rac ti o n at io n radio therap y 523 N o si g n ifi can t b enefi t V an D en B og a e r t et al. (1995) E tanidazole HS R Phase II L o cal ly ad va nc ed pr o st ate ca nce r E xt er n al b eam radia tio n 58, 39 N o si g n ifi can t b enefi t B ear d et al. (1994); L aw t o n et al. (1996) Smal l-c el l l u ng canc er C o n ven ti o n al radio therap y 30 N o si g n ifi can t b enefi t U rt a s u n et al. (1998) Phase III H ead and neck car cino ma C o n ven ti o n al radio therap y 521, 374 N o si g n ifi can t b enefi t L ee et al. (1995); E sch wege et al. (1997) P imo nidazole HS R Phase III Car cino ma o f the u te rine c er v ix C o n ven ti o n al radio therap y 183 N o si g n ifi can t b enefi t; un us ual ly go o d r es p o n se in the c o n tr o l ar m D ische et al. (1993) N imo razole HS R P hase II H ead and neck C o n tin uo us h y per frac tio n at ed ac celera te d radia tio n therap y (CHAR T )

61 P o si ti ve e ffe ct com p are d w it h p rev io us CHAR T stud y H enk et al. (2003) HS R P hase III H ead and neck C o n ven tio n al radio therap y 422 P osit iv e eff ec t (b ett er lo co- re g io n al c o n tr o l ra te and canc er - rela te d dea th)

O v e r g aar d et al. (1998) HS R , h y po xia-selec ti ve radiosen sit izer s

dosage basis. The latter strategy has been achieved by incorporating other functional moieties in the drug molecule.

CB 1954 (2,4-dinitro-5-aziridinylbenzamide;

Fig. 5.4) is an HSR-bearing DNA alkylating moiety

by cellular metabolism occurring at 37qC (Roberts et al. 1987) and poor extravascular penetration through tumor cell layers (Wilson et al. 1986).

2-NLP-3 (5-[3-(2-nitro-1-imidazolyl)-propyl]-phen- anthridinium bromide; Figure 5.5) and 2-NLP-4 (5-[3-(2-nitro-1-imidazolyl)-butyl]-phenanthri- dinium bromide; Fig. 5.5) are 2-nitroimidazoles attached to the DNA intercalator phenanthridine.

They were shown to be 10–100 times more effi- cient as HSR than misonidazole both in vitro and in vivo (Cowan et al. 1991). The structurally simi- lar compound, NLA-1 (9-[3-(2-nitro-1-imidazolyl) propylamino]acridine hydrochloride; Fig. 5.5) is a DNA-targeted acridine-linked 2-nitroimidazole ( Papadopoulou et al. 1992). Although NLA-1 is a potent HSR in vitro, it lacked in vivo activity (Denny et al. 1992).

5.3.4

Limitations of HSR

The major drawback of HSR is that the radiosensi- tization efficacy of HSR to tumors decreases at the radiation doses used in clinically relevant frac- tionated irradiation. Experimental results showed that 2–3 Gy fractionated irradiation resulted in lower enhancement ratios than single large doses for misonidazole (Denekamp and Stewart 1978; Hill and Bush 1978; Sheldon and Fowler (1978) or etanidazole (Brown and Yu 1984). One explanation for the lack of radiosensitization at fractionated low doses could be reoxygenation occurring between each fraction (Brown and Yu 1984); however, an equally likely reason is that it is not the cells at maximum radiation resis- tance, but those at intermediate oxygenation and intermediate radioresistance that dominate the response to fractionated irradiation (Wouters and Brown 1997). Though hypoxic cell radiosen- sitizers at clinically realistic doses can sensitize the maximally hypoxic cells to radiation killing, they have little effect on the radiosensitivity of the cells at intermediate hypoxia and radiosen-

Fig. 5.4. Examples of mixed-function compounds as HSR

of aziridine, which showed more efficient radiosen- sitization than misonidazole (Chapman et al. 1979).

Similarly, RSU 1069 (1(2-nitro-1-imidazolyl)3- aziridinyl-2-propanol; Fig. 5.4), the second genera- tion of CB 1954, showed in vitro radiosensitization potency tenfold greater than misonidazole at equi- molar doses (Stratford 1982; Adams et al. 1984);

however, it appears that the more effective sensitiz- ing ability of RSU 1069 is due to a greater degree of hypoxic cell cytotoxicity rather than to enhanced radiosensitizing ability (Hill et al. 1986). Both CB 1954 and RSU 1069 are discussed further as hypoxic cytotoxins in section 5.4.4.

5.3.3

DNA-Affinic Radiosensitizers

This class of HSR incorporates a DNA-affinic moiety thereby attracting the drug molecule closer to DNA where they exert their radiosensitizing effect. Nitra- crine, 1-nitroacridine derivative (1-nitro-9-(dime thylaminopropylamino)acridine; Fig. 5.5) showed 1700 times more efficient in vitro radiosensitization than misonidazole (Roberts et al. 1987); however the development of nitracrines as HSR was limited

Fig. 5.5. Examples of DNA affi nic compounds as HSR

CB 1954 RSU 1069

Nitracrine NLA-1

2-NLP-3 (n=3)

or 2-NLP-4 (n=4)

sitivity. This is a consequence of the logarithmic nature of the curve of radiosensitization vs dose:

as cell sensitivity increases it takes geometrically more drug to increase their radiosensitivity still further; thus, a hypoxic cell partially sensitized by some oxygen takes much more drug to increase its sensitivity by a given amount compared with a fully hypoxic cell.

It thus seems unlikely that hypoxic cell radio- sensitizers – even newer ones with greater potency – have a major influence on the outcome of frac- tionated radiotherapy. They may, however, play a role – or probably should play a role – in those situations where large, single doses of radiation are given such as with stereotactic radiotherapy of brain tumors.

5.4

Hypoxic Cytotoxins

5.4.1

Introduction

An alternative strategy to overcome the problem of hypoxic tumor cells is to selectively kill them by using hypoxic cytotoxins. Typically these are pro- drugs of very low cytotoxicity that are reduced by enzyme(s) in hypoxic tumor cells. This results in conversion to potent cytotoxins killing the activat- ing cell and, in some cases, the surrounding cells.

The development of hypoxic cytotoxins was origi- nally stimulated by the findings that nitroimid- azoles were more toxic to hypoxic than oxic tumor cells even without irradiation (Sutherland 1974;

Hall and Roizon-Towle 1975).

Selectivity for hypoxia is usually a result of a futile redox cycling in which the presence of O

reoxidizes the one-electron reduced intermediate thereby regenerating the parent drug (Fig. 5.6a; Mason and Holtzman 1975). Superoxide anion (O

x-), a major by-product of this reaction, is capable of producing DNA strand breaks and other oxidative damage, but can be detoxified by superoxide dismutase and cata- lase (Biaglow et al. 1982). Hypoxia selective cyto- toxicity therefore occurs when the reduction of the hypoxic cytotoxin produces a more toxic interme- diate than the O

x- radical (Wardman et al. 1995;

Wardman 2001).

5.4.2

Combination of Hypoxic Cytotoxins with Ionizing Radiation

Hypoxic cytotoxins are the ideal drug to combine with ionizing radiation because they produce a pro- file of cytotoxicity as a function of distance from blood vessels in tumors that are the opposite to that produced by ionizing radiation (Fig. 5.6b; Brown 1999). In other words, hypoxic cytotoxins kill the tumor cells that are resistant to ionizing radiation.

In addition, hypoxic cytotoxins may release stable and diffusible cytotoxins capable of killing the sur- rounding tumor cells at relatively higher O

con- centrations, producing a so-called bystander effect (Denny and Wilson 1993). Unlike HSR, the efficacy of hypoxic cytotoxins is independent of the dose of radiation because there is no interaction between the two agents; hence, the benefit of adding a hypoxic cytotoxin to fractionated irradiation increases with the number of times the drug is administered and can produce greater benefit than if all of the tumor

Fig. 5.6. a The concept of hypoxic cytotoxins (indicated as D).

Hypoxia selectivity is achieved by the futile cycle by oxygen converting the reduced intermediate (D•-) back to its non- toxic prodrug form (D). b The prediction of tumor cell killing by a hypoxic cytotoxin or radiation and most anticancer drugs as a function of the distance from the capillary. Note the oppo- site profi le in cytotoxicity towards tumor cells; the combina- tion of the two should yield complementary cell killing and is shown as a dashed line. (From Brown 1999)

a

b

were fully oxygenated as shown in the theoretical study by Brown and Koong (1991).

5.4.3

Nitroimidazoles

A general scheme (Fig. 5.7) for nitroheterocycle reduction is that the addition of an odd number of electrons (1, 3, 5) leads to radical intermediates, while an even number of electrons (2, 4, 6) leads to the nitroso (R-NO), hydroxylamine (R-NHOH), and amine reductants (R-NH

), respectively (Mason and Holtzman 1975; Perez-Reyez et al. 1980). Oxygen is able to reverse or inhibit reduction at the one-elec- tron radical anion, although it could in principle act at various stages (Wardman and Clarke 1976). In the absence of O

, further reduction occurs primar- ily via disproportionation reactions of R-NO

x-, ulti- mately leading to the fragmentation of the imidazole ring (Varghese et al. 1976; Flockhart et al. 1978).

Two- and four-electron reduced derivatives may have different stabilities and reactivities depending upon the nature of the aromatic ring and its sub- stituents (McClelland et al. 1984).

Some examples of this class of hypoxic cytotoxins are metronidazole, misonidazole (Fig. 5.8; Rauth et al. 1984), bis-nitroimidazole (Fig. 5.8; Hay et al. 1994;

Moselen et al. 1995), and NLCQ-1 (Papadopoulou et al. 2000). NLCQ-1 is a nitroimidazole-linked chlo- roquinoline (4-[3-(2-nitro-1-imidazolyl)-propyl- amino]-7-chloroquinone hydrochloride; Fig. 5.8), a member of NLP-1/NLA-1/THNLA-1 series, which is also an efficient HSR ( Papadopoulou et al. 1994;

Papadopoulou et al. 1996). It shows time-dependent increase in hypoxic cytotoxicity in vitro and is cur- rently under preclinical evaluation ( Papadopoulou et al. 2000; Squillace et al. 2000).

In mammalian cells, aldehyde oxidase, DT- diaphorase, xanthine oxidase, NADPH:cytochrome P450 reductase, cytochrome b5 reductase, NADH- dehydrogenase, and succinate dehydrogenase ( Heimbrook and Sartorelli 1986; Walton and Workman 1987; Hodgkiss 1998) have been reported as nitroreductases.

5.4.4

Other Nitroaromatics

5.4.4.1 CB 1954

Though not strictly a hypoxic selective cytotoxin, CB1954 has been an important lead for the develop- ment of such agents. CB 1954 (Fig. 5.9) is a mono-

Fig. 5.7. The reduction of nitroimidazole hypoxic cytotoxins. R generalized nitroimidazole ring

Fig. 5.8. Examples of nitroimidazole compounds as hypoxic cytotoxins

Metronidazole Misonidazole

Bis-nitroimidazole NLCQ-1

functional alkylating agent that can be enzymatically activated to a potent difunctional alkylating agent which crosslinks DNA (Knox et al. 1988a; Knox et al. 1988b; Brown and Wilson 2004). Reduction of CB 1954 produces the potent cytotoxic metabolites, the 4-hydroxylamine, and 2-amine (Helsby et al.

2004). The activation of CB 1954 also occurs through a second activation step by a non-enzymatic reac- tion with a thioester (such as acetyl CoA) to form the final DNA reactive species, which is presumably 4-(N-acetoxy)-5-(aziridin-1-yl)-2-nitrobenzamide (Knox et al. 1991). The reduction of CB 1954 has been shown to proceed at about equal rates under aerobic and hypoxic conditions, presumably because the two-electron reduction by DT-diaphorase is not inhibited by O

. Currently, there is interest in CB 1954 in gene-directed enzyme-prodrug therapy (GDEPT) using Escherichia coli B. nitroreductase (Bridgewater et al. 1995; Green et al. 1997), and is in currently phase-I clinical trials (Chung-Faye et al. 2001; Dachs et al. 2005).

5.4.4.2

SN 23862 and PR-104

Structurally similar to CB 1954, a nitrogen mustard compound, SN 23862 (5-[N,N-bis(2- chloroethyl)amino]-2,4-dinitrobenzamide; Fig. 5.9) exploits the nitro-group as an electronic switch in that nitro-to-amine conversion shifts electron den- sity modifying the reactivity of a drug molecule (Siim et al. 1997; Helsby et al. 2003). The chemi- cal reduction of each of the 2-nitro and the 4-nitro group of SN 23862 showed an increase in cytotoxic- ity by 160- and 9-fold in AA8 cell line, and by 4400- and 83-fold in UV 4 cell line, respectively, and that the reduction of both nitro groups led to further increase in cytotoxicity (Palmer et al. 1995; Helsby

et al. 2003). Good bystander effect in both in vitro and in vivo systems together with substrate speci- ficity to Escherichia coli B nitroreductase warrants development of SN 23862 for GDEPT (Anlezark et al. 1995; Wilson et al. 2002).

PR-104 is a phosphate pre-prodrug of the dini- trobenzamide mustard PR-104A that is activated in hypoxia to become a potent bifunctional alkyl- ating agent producing DNA interstrand crosslinks ( Douglas et al. 2005). In addition, PR-104 has dem- onstrated substantial bystander effect killing aero- bic as well as hypoxic cells in solid tumors (Wilson et al. 2005). Because of this ability to kill both aero- bic and hypoxic cells in tumors PR-104 shows sig- nificant antitumor activity as a single agent alone.

It is also superior to tirapazamine in combination with fractionated irradiation in preclinical models (Dorie et al., unpublished data). PR-104 entered clinical phase-1 trials in December 2005.

5.4.4.3

RSU 1069 and RB 6145

RSU 1069 (Fig. 5.9) is bioreductively activated by NADPH-cytochrome P450 reductase forming a bifunctional crosslinking agent (Stratford et al.

1986; Whitmore and Gulyas 1986). In air, RSU 1069 functions as a monofunctional alkylating agent due to the presence of the aziridine group ( Stratford et al. 1986). It shows in vivo antitumor activity against KHT sarcoma and RIF-1 tumor when given before or after irradiation (Hill et al. 1986).

The bromoethylamine derivative RB 6145 (Fig. 5.9) was developed as a prodrug of RSU 1069 because of irreversible gastrotoxicity observed in early phase-I trials of RSU 1069 (Horwich et al.

1986). RB 6145 showed slightly less hypoxia selec- tive cytotoxicity than RSU 1069 in vitro (Jenkins

Fig. 5.9. Examples of nitroaromatic hypoxic cytotoxins

CB 1954

RSU 1069

SN 23862

RB 6145

et al. 1990) but was active against hypoxic cells in vivo with lowered toxicity compared with RSU 1069 ( Jenkins et al. 1990; Cole et al. 1990); however, fur- ther development of this compound was stopped because preclinical studies showed irreversible cytotoxicity toward retinal cells in mice (Parker et al. 1996; Breider et al. 1998).

5.4.5 Quinones

5.4.5.1 Mitomycin C

Mitomycin C (Fig. 5.10), isolated from Streptomyces caespitosus, was introduced into the clinic in 1958, and was subsequently shown to have a moderate in vitro hypoxia selective cytotoxicity [hypoxic cyto- toxicity ratio (concentration of drug in air divided by concentration of drug in hypoxia to produce the same level of cell killing; Brown 1993) of about 1–5;

Rockwell et al. 1982; Fracasso and Sartorelli 1986]. The cytotoxicity of mitomycin C is associ- ated with formation of monofunctional alkylation and more potently with intra- and inter-strand crosslinks of DNA, all of which require bioreduc- tive activation (Iyer and Szybalski 1963; Tomasz et al. 1987; Volpato et al. 2005).

The one-electron reduction of mitomycin C results in a semiquinone, which under hypoxic conditions activates the aziridine ring and results in binding of the drug to DNA (Pan et al. 1984). Following the initial covalent attachment of mitomycin C to DNA, the drug can undergo further reductive activation to form a second alkylating site (Pan et al. 1984). The one-electron reduction pathway can be catalysed by any of several enzymes including NADPH:cyto- chrome P450 reductase (Pan et al. 1984; Keyes et al. 1984) and xanthine oxidase (Pan et al. 1984), in a process that can be reversed by O

(Bachur et al.

1978; Bachur et al. 1979; Pritsos and Sartorelli

1986). Mitomycin C can also be reductively activated via two-electron reducing DT-diaphorase generat- ing an O

-insensitive hydroquinone (Iyanagi and Yamazaki 1970; Keyes et al. 1984).

Despite variable results in hypoxia selective cyto- toxicity observed in preclinical studies ( Rockwell and Kennedy 1979; Rockwell et al. 1982; Keyes et al. 1984), clinical trials have reported that mito- mycin C in combination with radiation shows a sig- nificant benefit in local regional control rates for patients with head and neck cancer (Weissberg et al. 1989; Haffty et al. 1993; Haffty et al. 1997), squamous cell carcinoma of the cervix (Roberts et al. 2000), and laryngeal and hypopharyngeal cancer (Saarilahti et al. 2004) (Table 5.2).

5.4.5.2 Porfiromycin

Porfiromycin (Fig. 5.10), a second-generation version of mitomycin C, showed superior in vitro hypoxic selectivity over mitomycin C as a result of lowered aerobic cytotoxicity (Fracasso and Sartorelli 1986; Rockwell et al. 1988) and an improved in vivo therapeutic index as a result of higher LD

(the dosage required to kill 50% of the treated popula- tion) in mice (Keyes et al. 1985). Although a phase-I trial showed an acceptable toxicity profile, a recent phase-III randomized trial showed that porfiromy- cin was inferior to mitomycin C as an adjunct to radiotherapy in squamous cell cancer of the head and neck (Haffty et al. 2005). For this and other reasons (Brown (1999), it therefore seems unlikely that the clinical activity of mitomycin C is the result of its (modest) selectivity as a hypoxic cytotoxin.

5.4.5.3 EO9

EO9 ([3-hydroxymethyl-5-aziridinyl-2-methyl-2- (H-indole-4,7-indione)-propenol]; Fig. 5.10) was originally developed as a synthetic analog of mito-

Fig. 5.10. Examples of quinones

Mitomycin C Porfi romycin EO9

mycin C (Oostveen and Speckamp 1987). The reductive bioactivation of EO9 occurs in a similar fashion to mitomycin C such that one- and two- electron reduction processes yield the correspond- ing semiquinone and hydroquinone, respectively (Maliepaard et al. 1995). The semiquinone is believed to be more cytotoxic than the hydroqui- none based on its ability to produce DNA interstrand crosslinks and strand breaks (Bailey et al. 1993;

Plumb et al. 1994; Maliepaard et al. 1995).

Although partial responses were observed in a small number of patients in phase-I trials ( Schellens et al. 1994), no apparent antitumor activity by EO9 alone was demonstrated in phase- II studies in patients with advanced breast, gastric, pancreatic, and colorectal carcinoma (Dirix et al.

1996); however, some concerns about the design of the trials were raised in that enzymatic activ- ity in patients’ tumors was not measured routinely ( Phillips 1996) and that hypoxic cytotoxins, such as EO9, should be combined with other treatment modalities, such as radiotherapy or chemother- apy, to demonstrate detectable clinical responses ( Workman and Stratford 1993).

5.4.6

Benzotriene di-N-Oxides

Tirapazamine (SR 4233; Fig. 5.11a) is the prototype of this class of hypoxic cytotoxins. Under hypoxic conditions, tirapazamine is reduced to an O

-sensi- tive tirapazamine radical (Lloyd et al. 1991). The tirapazamine radical then eliminates a water mol- ecule to form a nitrogen-centered oxidant, benzo-

triazinyl radical, which can mediate the initial cyto- toxic process by abstracting a hydrogen atom from the deoxyribose backbone of DNA (Anderson et al.

2003). The tirapazamine radical may also lead to the formation of the two-electron reduced product, SR 4317, a major non-toxic metabolite detected in cul- tured hypoxic cells (Baker et al. 1988; Hicks et al.

2003), in the mouse (Walton and Workman (1993), and in humans (Grahams et al. 1997). Formation of SR 4317 can occur by a number of different routes, including the direct two-electron reduction of tira- pazamine by DT-diaphorase (Riley and Workman 1992), by radical disproportionation reaction, or by hydrogen abstraction from macromolecules other than DNA (Brown and Wang 1998).

Tirapazamine has a unique O

concentration dependency such that its cytotoxicity does not level off at high concentrations, but gradually decreases as the O

concentration increases (Koch 1993; Hicks et al. 2004). Tirapazamine shows hypoxic cytotoxicity ratios of up to 200 in murine and 50 in human cell lines (Zeman et al. 1986).

The hypoxic cytotoxicity of tirapazamine is due to the formation of DNA strand breaks resulting in chromosome aberrations (Beidermann et al. 1991;

Wang et al. 1992; Siim et al. 1996). The chromosome breaks produced by tirapazamine were shown to be less easily repaired than those produced by X-rays, and this has been suggested to be a result of prob- able metabolism by reductases located close to DNA (Wang et al. 1992). Although a large proportion of tirapazamine is metabolized in the cytoplasm by enzymes, such as cytochrome P450 (Walton et al.

1992; Wang et al. 1993; Riley et al. 1993), NADPH cytochrome P450 reductase (Cahill and White

Fig. 5.11. a Chemical structure of tirapazamine. b Synergistic activity of tirapazamine in killing SCCVII tumors in vivo when combined with fractionated irradiation (fi lled circles), deter- mined by clonogenic cell survival. Tirapazamine alone and radiation alone (2.5 Gy per fraction) are shown as open circles

and dots, respectively. From Brown and Lemmon (1990) Number of fractions

Relativ e clonogenic c ells/tumor

a

b

T able 5.2. P o te n tial candida te s o f h y po xic c y to to xin s and their r ec en t clinical/p re clinical ev al ua tio n da ta N ame o f dr ug Dr ug eval ua ti o n ca te g o ry E val ua ti o n T y pe o f t rial/ ev a lua ti on syst em T y pe o f canc er/ tumo r eval ua te d N o . o f p at ien ts and/o r t y pe of R T Ou tc o m e R ef er enc e Mit o m ycin C BEP C linical Phase II A d vanc ed lar y n geal and h y po phar y n geal cance r

21 p at ien ts w ith ac celera te d h y per frac ti o n at ed R T Po si ti ve e ff ec t Saar ilah t i et al. (2004) A d vanc ed head and neck canc er Po si ti ve e ff ec t W idder et al. (2004) Phase III A d va nc ed sq ua- mo us-c el l car cino ma o f the c er v ix

160 p at ien ts w ith radical R T P osit iv e eff ec t (4-y ear disease-f re e sur v iv al) R obe r ts et al. (2000) S q uamo us c el l car cino ma o f the head and neck 120 p at ien ts w ith co n ven ti o n al R T P osit iv e eff ec t (5-y ear disease-f re e su rv iv al and lo cal r ecur renc e f ree sur v iv al)

W ei ss ber g et al. (1989) 182 p at ien ts w ith co n ven ti o n al R T H aff t y et al. (1993) A Q4N BEP P reclinical In v iv o T50/80 m ur ine mammar y car ci- no ma

12-G y sing le-dose R T A ddit iv e eff ec t M c K eo w n et al. (1995) F rac ti o n at ed R T (5 u 3 G y) A ddit iv e eff ec t RS U-1069 HS R and BEP P reclinical In v iv o FSaIIC m ur ine fi b rosar co ma F rac ti o n at ed R T (5 u 3 G y) A ddit iv e eff ec t T ei c h er et al. (1991) KHT sar co ma 10-G y sing le-dose R T A ddit iv e eff ec t C ole et al. (1990) RB-6145 HS R and BEP P reclinical In v iv o KHT sar co ma 10-G y sing le-dose R T A ddit iv e eff ec t C ole et al. (1990) NL CQ-1 HS R and BEP P reclinical In v iv o U251 h uman g lio ma xe nog raf t 5-G y sing le-dose R T A ddit iv e eff ec t P ap ad opoul ou et al. (2005) F rac ti o n at ed R T (4 u 1 G y) N o si g n ifi can t b enefi t T irap aza- mine BEP C linical Phase I L o cal ly ad vanc ed head and neck cance r 16 p at ien ts w ith c o n ven ti o n - al ly f rac ti o n at ed R T Po si ti ve e ff ec t R ischin et al. (2001) Phase-II rando mized t rial 122 p at ien ts, defi nit iv e R T P osit iv e eff ec t (3-y ear f ail ur e- fr ee s ur v iv al ra te s; 3-y ear lo co- re g io n al f ail ur e-f re e ra te s)

R ischin (2005) RT , radio therap y

1990; Lloyd et al. 1991; Walton et al. 1992; Wang et al. 1993; Patterson et al. 1995; Patterson et al.

1997), xanthine oxidase (Laderoute et al. 1988), and DT-diaphorase (Riley and Workman 1992;

Wang et al. 1993), it is the nuclear metabolism of tirapazamine (ca. 20% of the overall cellular metab- olism) that accounts for essentially all of the tira- pazamine-induced DNA damage under hypoxic conditions (Evans et al. 1998; Delahoussaye et al.

2001). Recently, tirapazamine in hypoxic conditions has also shown to markedly inhibit DNA replication (Peters et al. 2001) and to poison topoisomerase-II activity (Peters and Brown 2002) in vitro leading to the suggestion that the DNA double-strand breaks produced by tirapazamine result from its poisoning topoisomerase II (Peters and Brown 2002).

Tirapazamine potentiates cell killing with frac- tionated irradiation in mouse tumors (Fig. 5.11b;

Brown and Lemmon 1990), and with cisplatin in a highly schedule-dependent manner (Dorie and Brown 1993). Tirapazamine is currently in phase- III evaluation and appears particularly effective in combination with cisplatin (Rodriguez et al. 1996;

Miller 1997; Treat 1998; von Powel 2000) or as an adjunct to cisplatin/radiotherapy treatment (Rischin et al. 2001; Rischin et al. 2005).

5.4.7

Tertiary Amine N-Oxides

5.4.7.1

Nitracrine N-Oxides

Nitracrine N-oxide (Fig. 5.12a) was developed as a prodrug of nitracrine, a hypoxia cytotoxin whose cytotoxicity is due to nitroreduction that is inhibited by O

(Wilson et al. 1986). While nitracrine showed the degree of hypoxic selectivity similar to misoni- dazole (ca. tenfold) in vitro, its metabolism was too rapid to provide selective killing of hypoxic cells in vivo (Wilson et al. 1986). Derivatization with ter- tiary amine N-oxide lowered DNA binding (15-fold), cell uptake, and aerobic cytotoxicity compared with nitracrine, while its hypoxic selectivity was greatly increased (1000- to 1500-fold; Wilson et al. 1992).

The very high hypoxia selectivity was suggested to be due to a requirement for both the nitro and N- oxide moieties for full activation in an O

inhibitable manner (Wilson et al. 1992). Despite some improve- ment of extravascular diffusion properties observed in vitro (Wilson et al. 1992), in vivo activity against KHT tumors was only observed at doses lethal to the host (Lee et al. 1996).

Fig. 5.12. a Examples of tertiary amine N-oxide compounds. b Tumor growth delay of T50/80 in BDF mice treated with AQ4N.

Combination with a single dose radiation (fi lled squares) pro- duced a signifi cantly greater antitumor activity compared with AQ4N alone (open squares). (From McKeown et al. 1996) a

b

Nitracrine N-oxide

AQ4N

5.4.7.2 AQ4N

AQ4N (1,4-bis-{[2-(dimethylamino-N-oxide)ethyl]

a mi no}5, 8 -d i hyd rox ya nt hracene-9,10 -d ione;

Fig. 5.12a) is a prodrug designed to be prevented from binding to DNA (Patterson 1993; Smith et al. 1997b) until metabolized in hypoxic cells to give AQ4, a stable, oxygen insensitive metabolite (Smith et al. 1997a). AQ4 is a DNA intercalator and potent inhibitor of DNA topoisomerase II ( Patterson 1993; Patterson et al. 1994; Smith et al. 1997b).

Early development of AQ4N has been confounded by the lack of hypoxia selective activity in several cell lines; however, a large (>100-fold) increase in cytotoxicity of AQ4N was observed when cells were incubated under hypoxic conditions with liver microsomes (Patterson 1993). It was later shown that AQ4N is not activated by NAD(P)H-dependent cytochrome P450 reductase per se (Patterson et al. 1999), but by cytochrome P450 (CYP) isoforms.

Several studies have shown that the metabolism of AQ4N correlates with levels of CYP1A1, 2B6, and 3A in humans (Patterson and McKeown 2000), 3A in mice (Patterson et al. 2000), and 2B and 2E in rats (Raleigh et al. 1999). Reduction of the two N-oxide functionalities of AQ4N has been suggested to be involved in an oxygen atom transfer from the N-oxide side chains in a process that is O

sensitive (Patterson and McKeown 2000). Moreover, heme- containing systems other than CYP, such as nitric oxide synthase, have also shown to mediate AQ4N reduction (Raleigh et al. 1998). AQ4N has shown in vivo antitumor activity when combined with radia- tion (Fig. 5.12b; McKeown et al. 1995), McKeown et al. 1996) or when combined with a range of methods inducing additional tumor hypoxia, e.g., hydralazine (Patterson et al. 2000), dimethylxanthenone acetic acid (Wilson et al. 1996), or clamping ( Patterson et al. 2000). It is currently in phase-I clinical trials.

5.5

Combination of Radiation-Activated Prodrugs with Ionizing Radiation

5.5.1

Concepts of RAP

Another way to target hypoxic tumor cells is to use ionizing radiation, rather than enzymes, to effect the reduction of the prodrug. Radiation-activated

prodrugs (RAP) are reduced by e

- (aquated/

hydrated electron) generated from the radiolysis of water, offering two distinct mechanisms strongly inhibited by O

. These mechanisms involve either back-oxidization of the one-electron reduction intermediate of RAP (RAPx-) or to scavenge e

- at diffusion controlled rate (Fig. 5.13; Other primary radicals formed by radiolysis of water (OHx and Hx) are expected to be scavenged at very high rates by other biomolecules. Some of the resulting organic radicals also have reducing properties so might make a further contribution to prodrug reduction.

The RAP approach offers a number of potential advantages such as selective activation within the radiation field (tumor bearing volume); exploita- tion of necrotic hypoxic regions lacking reductase activity through release of a stable cytotoxin with a good bystander effect, no requirement for expres- sion of specific reductase(s); lack of two-electron activation; and improved extravascular transport by allowing design of prodrugs that are not sub- strates for metabolism in tumors (Wilson et al.

1998). Some of the main classes of the compounds that have been considered as candidates for RAP are discussed below.

5.5.2

Nitro(Hetero)Cyclic Methylquarternary Ammonium Salts

NMQ ammonium salts such as 4-nitroimidazole (SN 25341; N,N-bis(2-chloroethyl)-N-methyl-N-[(1- methyl-4-nitro-5-imidazolyl)methyl]ammonium chloride; 4-NIQ-HN2; Fig. 5.14a) and 5-nitropyrrole (N,N-bis(2-chloroethyl)-N-methyl-N-[(1-methyl- 5-nitro-1-pyrrolyl)methyl]ammonium chloride; 5- NPQ-HN2; Fig. 5.14a) that were originally developed as hypoxic cytotoxins (Tercel et al. 2001), have also been reported as RAP that are reduced with

Fig. 5.13. The concept of radiation-activated prodrugs (RAP).

Hypoxia selectivity occurs by the futile cycling of oxygen

reverting the reduced intermediate of RAP (RAP•-) or scav-

enging hydrated electron (e

-) generated from the radiolysis

of water

one-electron stoichiometry by pulse and steady- state radiolysis (Anderson et al. 1997; Wilson et al. 1998). Irradiating 4-NIQ-HN2 in anoxic human plasma, followed by exposure to UV4 cells showed that IC

(concentration of a compound giving 50%

growth inhibition relative to the control cells) was markedly decreased with radiation, approaching a value equal to that of the cytotoxin, mechloretha- mine (HN2), itself after ca. 20 Gy (Fig. 5.14b; Wilson et al. 1998); however, a relatively high one-electron

reduction potential of the prodrug (therefore sus- ceptible for enzymatic reduction) and only modest cytotoxicity of HN2 limited their utility to that of only model compounds.

In order to incorporate a more potent cyto- toxin within the prodrug system, 4-NBQ-AMAC (Fig. 5.14a), 4-NIQ-AMAC (Fig. 5.14a), and 5-NPQ- AMAC (Fig. 5.14a) were synthesized. These prodrugs were two orders of magnitude less cytotoxic than the released cytotoxin (AMAC) against a panel of tumor

Fig. 5.14. a Examples of NMQ compounds as RAP. b Radiolytic acti- vation of 4-NIQ-HN2 in human plasma determined as apparent IC

against UV4 cells. Oxic (95% O

and 5% CO

) and anoxic (95% N

and 5% CO

) are shown in open and fi lled circles, respectively. The IC

for authentic HN2 is also shown (fi lled square) with an arrow

SN 25341 5-NPQ-HN2

5-NPQ-AMAC 4-NIQ-AMAC

4-NBQ-AMAC

a

b

cell lines (Wilson et al. 1998). Irradiation in anoxic culture medium followed by exposure to UV4 cells showed significant activation of 4-NIQ-AMAC and 5-NPQ-AMAC at 2 Gy, although the yield of AMAC was lower than that for the HN2 analogs (Wilson et al. 1998). The AMAC prodrugs showed appre- ciable enzymatic reduction in hypoxic cell cultures, and convulsion in mice by a mechanism unrelated to the cytotoxin release (Wilson et al. 1998), sug- gesting that they are not likely to be useful as RAP prodrugs.

5.5.3

5-Fluorouracil (5-FU)-Releasing Prodrugs

Nishimoto et al. (1992) first reported that the N(1)-C(5)-linked dimer of 5-FU (1-(5’-fluoro-6’- hydroxy-5’,6’-dihydrouracil-5’-yl)-5-f luorouracil) releases 5-FU with ca. three-electron stoichiometry by irradiation in anoxic aqueous buffer and some indication of antitumor effect in combination with radiation in mice bearing SCCVII tumors. Recently,

OFU001 (1-(2’-oxopropyl)-5-fluorouracil; Fig. 5.15) and OFU101 (Fig. 5.15) have been reported to release 5-FU and 5-fluoro-2’deoxyuridine (FdUrd), respectively, with higher efficiency (two-electron stoichiometry) in anoxic buffer upon irradiation ( Shibamoto et al. 2000; Shibamoto et al. 2004).

The mechanism proposed is that attachment of e

-, generated by radiolysis of water, weakens the N(1)- C(1’) bond that links 5-FU or FdUrd to the oxoalkyl side chain. Although these compounds showed cell killing after hypoxic irradiation, they failed to dem- onstrate any significant antitumor activity in vivo (Shibamoto et al. 2000, 2001, 2004).

5.5.4

Transition Metal Complexes

A number of transition metal complexes, such as those of cobalt (III) (Co(III) (Fig. 5.16; Wilson et al. 1994; Ahn et al. 2004b) and copper (II) (Cu(II) (Fig. 5.16; Parker et al. 2004; Torre et al. 2005), have been reported as hypoxia-selective prodrugs.

In particular, Co(III) complexes are kinetically inert when not reduced; however, when reduced radio- lytically by e

-, the inert d

electron spin state of Co(III) becomes labile d

, which dramatically weak- ens the coordination bond hence releasing the cyto- toxin (Fig. 5.17; Ware et al. 1993). A Co(III) complex bearing a very potent DNA minor groove alkylator cyclopropylindoline (Fig. 5.17) has demonstrated a number of attractive features as a RAP such as: a good masking of the cytotoxicity of cyclopropyl- indoline in the intact Co(III) complex; an efficient release of the cytotoxin with clinically relevant radiation dose of 2 Gy in anoxic human plasma

Fig. 5.15. Chemical structures of 5-FU prodrugs of RAP

Fig. 5.16. Chemical structures of Co(III) and Cu(II) complexes of hypoxic prodrugs

OFU001 OFU001

Co(III) complex Cu(II) complex

or buffer; cell exclusion property of the complex enhancing extravascular transport through multi- cellular tumor layers; a HCR of up to 20 in human colorectal HT29 cells (Ahn et al. 2002; Ahn 2003);

however, further development of this Co(III) com- plex as a RAP has failed due to the lack of efficacy in RIF-1 bearing mice when combined with ionizing radiation (Ahn 2003).

5.6

Other Hypoxia-Targeting Strategies

5.6.1

GDEPT Targeting Tumor Hypoxia

As discussed in section 5.1, many genes supporting anaerobic metabolism and angiogenesis are upregu- lated under hypoxic conditions by the transcription factor, hypoxia-inducible factor-1 (HIF-1). Specifi- cally, HIF-1, once it heterodimerizes with D and E subunits, binds to a common hypoxia-responsive element (HRE) found in the enhancer region of all HIF-1 responsive genes (Semenza 2001). One GDEPT approach targeting tumor hypoxia aims to express an exogenous therapeutic gene by introduc-

ing an HRE sequence into an appropriate expres- sion cassette (Dachs et al. 1997). With this system, studies have shown that in transfected tumor cells, NADPH:cytochrome P450 reductase enhances the antitumor efficacy of RSU 1069 or tirapazamine in combination with ionizing radiation (Patterson et al. 2002; Cowen et al. 2004). Another approach has shown antitumor activity with HRE driven nitrore- ductase enhancing the activity of CB1954 (Shibata et al. 2002).

Dual responsive promoters have also been devel- oped whose expression is driven not only by HRE but also by radiation-responsive promoter sequence such as the early growth response-1 (Egr-1) to enhance the reporter gene expression in response to hypoxia in conjunction with low dose of radiation (Greco et al. 2002; Chadderton et al. 2005).

5.6.2

Clostridia-Directed Enzyme Prodrug Therapy The above-mentioned GDEPT systems have the major drawback of utilizing ex vivo manipulated (trans- fected) tumor cell lines implying that their clini- cal applicability may be difficult. To overcome this limitation Brown and colleagues have exploited the

Fig. 5.17. A Co(III) complex of cyclopropylindoline releasing its cytotoxin upon ionizing radiation (IR)

DNA alkylation Inactive IR

e _aq ^–

+

necrotic regions as a target for cancer therapy using a non-pathogenic strain of the bacterial genus Clos- tridium, an obligate anaerobe (Lemmon et al. 1994;

Fox et al. 1996). This Gram-positive, spore-forming bacteria becomes vegetative and grows only in the absence or at very low levels of oxygen (Brown and Wilson 2004). Clinical studies in 1970s already have proven the safety of Clostridium and its tumor-spe- cific germination (Carey et al. 1967; Heppner and Mose 1978; Heppner et al. 1983). Recently, geneti- cally engineered Clostridium expressing E. coli enzyme cytosine deaminase showed tumor-specific delivery of the enzyme and the consequent antitu- mor activity upon 5-fluorocytosine (5-FC) prodrug administration (Liu et al. 2002; Ahn (2004a). The combination of CDEPT with ionizing radiation is expected to result in enhanced antitumor activity because of ionizing radiation killing well-oxygen- ated tumor cells and/or the radiosensitizing abilities of 5-FC and 5-FU.

5.6.3

Targeting HIF-1

In addition to its role in hypoxic conditions, HIF-1 can be upregulated in the tumor in an O

-inde- pendent manner. Oncogenic mutations and loss- of-function mutations in tumor-suppressor genes have been shown to be associated with the increased activity of HIF-1 (Semenza 2003; Li et al. 2005).

Evidence for HIF-1 serving as a potential target for cancer therapy is substantial and extensively reviewed elsewhere (Semenza 2003; Giaccia et al.

2003). For example, HIF-1D (the O

sensitive HIF-1 subunit) expression has been shown to be a poor prognostic factor in a number of cancers including bladder cancer (Nakanishi et al. 2005; Theodo- ropoulos et al. 2004), pancreatic cancer (Shibaji et al. 2003), and breast carcinoma (Bos et al. 2003).

HIF-1 is also reported to promote tumor growth in many studies (Ryan et al. 1998; Carmeliet et al.

1998; Chen et al. 2003; Stoeltzing et al. 2004).

Some recent approaches targeting HIF-1 include using small molecule inhibitors blocking transcrip- tional activity of HIF-1 (Kung 2004; Kong et al.

2005) or those blocking HIF-1D protein synthesis (Tan et al. 2005), inhibitors of HIF-1 downstream target (Majumder et al. 2004; Zhong et al. 2004;

Fang et al. 2005), enhancement of HIF-1D protein degradation (Li et al. 2004), and antisense (Sun et al. 2005) or RNA interference against HIF-1D (Li et al. 2005).

Interestingly, ionizing radiation has been recently shown to upregulate HIF-1 activity in tumors ( Moeller et al. 2004); however, the effect of HIF-1 on the radiosensitivity of the tumor is controversial.

Moeller et al. (2005) showed that functional HIF-1 enhances tumor radiosensitivity by promoting ATP metabolism, cell proliferation, and p53 activation, whereas others have reported that it is the defi- ciency in HIF-1 that increases tumor radiosensitiv- ity, independent of p53 function (Unruh et al. 2003;

Williams et al. 2005). The reason for this discrep- ancy is yet to be determined, but the difference in experiment systems (RNA interference vs cell lines deficient for either HIF-1D or -E) could have contrib- uted.

5.7

Conclusion

The potential clinical benefit of exploiting tumor hypoxia by combining a hypoxia activated drug with conventional cancer therapy has yet to be realized in routine clinical practice. Despite this, the positive clinical results with the combination of the hypoxic cytotoxin tirapazamine with cisplatin in advanced non-small cell lung cancer and with chemoradio- therapy with advanced head and neck cancer demon- strate the potential of this approach. There is a good reason to expect that future drugs or strategies will do better: indeed advances made in experimental models identifying the determinants of the efficacy of these hypoxia-targeting compounds, together with other strategies to exploit tumor hypoxia, auger well for the future of this field.

References

Adams GE, Cooke MS (1969) Electron-affinic sensitization. I.

A structural basis for chemical radiosensitizers in bacteria.

Int J Radiat Biol Relat Stud Phys Chem Med 15:457–471 Adams GE, Ahmed I, Sheldon PW, Stratford IJ (1984) Radia-

tion sensitization and chemopotentiation: RSU 1069, a compound more efficient than misonidazole in vitro and in vivo. Br J Cancer 49:571–577

Ahn G (2003) Investigation of an aza-chloromethylbenzindo- line cobalt(III) complex as a hypoxia-activated prodrug for cancer therapy. Pathology Auckland, University of Auck- land, New Zealand

Ahn G, Ware DC, Botting KJ, Kriste AG, Denny WA, Wilson WR

(2002) A novel Co(III) complex of the DNA minor-groove

binder aza-seco-CBI-TMI as a hypoxia-selective prodrug

that is activated by enzymatic or radiolytic reduction. Proc Am Assoc Cancer Res 43:656

Ahn G, Liu SC, Menke D, Dorie MJ, Brown JM (2004a) Enhanced antitumor activity of 5-FC by genetically engineered anaer- obic bacteria C. sporogenes expressing cytosine deaminase when combined with a vascular targeting agent, ZD6126 or DMXAA. Second International Conference on Vascular Targeting, Miami, Florida

Ahn GO, Ware DC, Denny WA, Wilson WR (2004b) Optimiza- tion of the auxiliary ligand shell of cobalt(III)(8-hydroxy- quinoline) complexes as model hypoxia-selective radia- tion-activated prodrugs. Radiat Res 162:315–325

Anderson RF, Denny WA, LI W, Packer JE, Tercel M, Wilson WR (1997) Pulse radiolysis studies on the fragmentation of arylmethyl quaternary nitrogen mustards by one-elec- tron reduction in aqueous solution. J Phys Chem 101:9704–

9709

Anderson RF, Shinde SS, Hay MP, Gamage SA, Denny WA (2003) Activation of 3-amino-1,2,4-benzotriazine 1,4-diox- ide antitumor agents to oxidizing species following their one-electron reduction. J Am Chem Soc 125 748–756 Anlezark GM, Melton RG, Sherwood RF, Wilson WR, Denny

WA, Palmer BD, Knox RJ, Friedlos F, Williams A (1995) Bio- activation of dinitrobenzamide mustards by an E. coli B nitroreductase. Biochem Pharmacol 50:609–618

Bachur NR, Gordon SL, Gee MV (1978) A general mechanism for microsomal activation of quinone anticancer agents to free radicals. Cancer Res 38:1745–1750

Bachur NR, Gordon SL, Gee MV, Kon H (1979) NADPH-cyto- chrome P-450 reductase activation of quinone anticancer agents to free radicals. Proc Natl Acad Sci USA 76:954–957 Bailey SM, Lewis AD, Patterson LH, Fischer GR, Workman P

(1993) Free radical generation following reduction of EO9:

involvement in cytotoxicity. Br J Cancer 67:8

Baker MA, Zeman EM, Hirst VK, Brown JM (1988) Metabolism of SR 4233 by Chinese hamster ovary cells: basis of selective hypoxic cytotoxicity. Cancer Res 48:5947–5952

Beard C, Buswell L, Rose MA, Noql L, Johnson D, Coleman CN (1994) Phase-II trial of external beam radiation with etanidazole (SR 2508) for the treatment of locally advanced prostate cancer. Int J Radiat Oncol Biol Phys 29:611–616 Beidermann KA, Wang J, Graham RP, Brown JM (1991) SR 4233

cytotoxicity and metabolism in DNA repair-competent and repair-deficient cell cultures. Br J Cancer 63:358–362 Biaglow JE, Varnes ME, Koch CJ, Sridhar R (1982) Metabolic

activation of carcinogenic nitro-compounds to oxygen reactive intermediates. In: Floyd RA (ed) Free radicals and cancer. Marcel Dekker, New York

Bos R, Van der Groep P, Greijer AE, Shvarts A, Meijer S, Pinedo HM, Semenza GL, Van Diest PJ, Van der Wall E (2003) Levels of hypoxia-inducible factor-1D independently pre- dict prognosis in patients with lymph node negative breast carcinoma. Cancer 97:1573–1581

Breider MA, Pilcher GD, Graziano MJ, Gough AW (1998) Reti- nal degeneration in rats induced by CI-1010, a 2-nitroimi- dazole radiosensitizer. Toxicol Pathol 26:234–239

Bridgewater JA, Springer CJ, Knox RJ, Minton NP, Michael NP, Collins MK (1995) Expression of the bacterial nitroreduct- ase enzyme in mammalian cells renders them selectively sensitive to killing by the prodrug CB1954. Eur J Cancer 31A:2362–2370

Brizel DM, Sibley GS, Prosnitz LR, Scher RL, Dewhirst MW (1997) Tumor hypoxia adversely affects the prognosis of

carcinoma of the head and neck. Int J Radiat Oncol Biol Phys 38:285–289

Brown JM (1979) Evidence for acutely hypoxic cells in mouse tumours, and a possible mechanism of reoxygenation. Br J Radiol 52:650–656

Brown JM (1993) SR 4233 (tirapazamine): a new anticancer drug exploiting hypoxia in solid tumours. Br J Cancer 67:1163–1170

Brown JM (1999) The hypoxic cell: a target for selective cancer therapy. Eighteenth Bruce F. Cain Memorial Award Lecture.

Cancer Res 59:5863–5870

Brown JM, Giaccia AJ (1998) The unique physiology of solid tumors: opportunities (and problems) for cancer therapy.

Cancer Res 58:1408–1416

Brown JM, Koong A (1991) Therapeutic advantage of hypoxic cells in tumors: a theoretical study. J Natl Cancer Inst 83:178–185

Brown JM, Lemmon MJ (1990) Potentiation by the hypoxic cytotoxin SR 4233 of cell killing produced by fractionated irradiation of mouse tumors. Cancer Res 50:7745–7749 Brown JM, Wang LH (1998) Tirapazamine: laboratory data rel-

evant to clinical activity. Anticancer Drug Des 13:529–539 Brown JM, Wilson WR (2004) Exploiting tumour hypoxia in

cancer treatment. Nat Rev Cancer 4:437–447

Brown JM, Yu NY (1984) Radiosensitization of hypoxic cells in vivo by SR 2508 at low radiation doses: a preliminary report. Int J Radiat Oncol Biol Phys 10:1207–1212

Brown JM, Yu NY, Brown DM, Lee WW (1981) SR-2508: A 2- nitroimidazole amide which should be superior to miso- nidazole as a radiosensitizer for clinical use. Int J Radiat Oncol Biol Phys 7:695–703

Bush RS, Jenkin RDT, Allt WEC, Beale FA, Bean H, Dembo AJ, Pringle JF (1978) Definitive evidence for hypoxic cells influ- encing cure in cancer therapy. Br J Cancer 37:302–306 Cahill A, White IN (1990) Reductive metabolism of 3-amino-

1,2,4-benzotriazine-1,4-dioxide (SR 4233) and the induc- tion of unscheduled DNA synthesis in rat and human derived cell lines. Carcinogenesis 11:1407–1411

Carey RW, Holland JF, Whang HY, Neter E, Bryant B (1967) Clostridial oncolysis in man. Eur J Cancer 3:37–46 Carmeliet P, Dor Y, Herbert JM, Fukumura D, Brusselmans K,

Dewerchin M, Neeman M, Bono F, Abramovitch R, Maxwell P, Koch CJ, Ratcliffe P, Moons L, Jain RK, Collen D, Keshert E (1998) Role of HIF-1 alpha in hypoxia-mediated apop- tosis, cell proliferation and tumour angiogenesis. Nature 394:485–490

Chadderton N, Cowen RL, Sheppard FCD, Robinson S, Greco O, Scott SD, Stratford IJ, Patterson AV, Williams KJ (2005) Dual responsive promoters to target therapeutic gene expression to radiation-resistant hypoxic tumor cells. Int J Radiat Oncol Biol Phys 62:213–222

Chan P, Milosevic M, Fyles AW, Carson J, Pintilie M, Rauth M, Thomas G (2004) A phase III randomized study of misoni- dazole plus radiation vs radiation alone for cervix cancer.

Radiother Oncol 70:295–299

Chaplin DJ, Durand RE, Olive PL (1986) Acute hypoxia in tumors: implications for modifiers of radiation effects. Int J Radiat Oncol Biol Phys 12:1279–1282

Chaplin DJ, Horsman MR, Aoki DS (1991) Nicotinamide, Fluo- sol DA and carbogen: a strategy to oxygenate acutely and chronically hypoxic cells in vivo. Br J Cancer 63:109–113 Chapman JD, Raleigh JA, Pederson L, Ngan J, Shum FY, Meeker

BE, Urtasun RC (1979) Potentially three distinct roles for

hypoxic cell radiosensitizers in the clinic. Proc 6th Int Congr Radiat Res, Tokyo

Chen J, Zhao S, Nakada K, Kuge Y, Tamaki N, Okada F, Wang J, Shindo M, Higashiro F, Takeda K, Asaka M, Katoh H, Sugi- yama T, Hosokawa M, Kobayashi M (2003) Dominant-nega- tive hypoxia-inducible factor-1D reduces tumorigenicity of pancreatic cancer cells through the suppression of glucose metabolism. Am J Pathol 162:1283–1291

Choy H, Nabid A, Stea B, Scott C, Roa W, Kleinberg L, Ayoub J, Smith C, Souhami L, Hamburg S, Spanos W, Kreisman H, Boyd AP, Cagnoni PJ, Curran WJ (2005) Phase II multi- center study of induction chemotherapy followed by con- current efaproxiral (RSR13) and thoracic radiotherapy for patients with locally advanced non-small-cell lung cancer.

J Clin Oncol 23:5918–5928

Chung-Faye G, Palmer D, Anderson D, Clark J, Downes M, Bad- deley J, Hussain S, Murray PI, Searle P, Seymour L, Harris PA, Ferry D, Kerr DJ (2001) Virus-directed, enzyme pro- drug therapy with nitroimidazole reductase: a phase I and pharmacokinetic study of its prodrug, CB1954. Clin Cancer Res 7:2662–2668

Cole S, Stratford IJ, Adams GE, Fielden EM, Jenkins TC (1990) Dual-function 2-nitroimidazoles as hypoxic cell radiosen- sitizers and bioreductive cytotoxins: in vivo evaluation in KHT murine sarcomas. Radiat Res 124:S38–S43

Corry J, Rischin D (2004) Strategies to overcome accelerated repopulation and hypoxia: What have we learned from clinical trials? Semin Oncol 31:802–808

Cowan DS, Panicucci R, Mcclelland RA, Rauth AM (1991) Tar- geting radiosensitizers to DNA by attachment of an inter- calating group: nitroimidazole-linked phenanthridines.

Radiat Res 127:81–89

Cowen RL, Williams KJ, Chinje EC, Jaffar M, Sheppard FCD, Telfer BA, Wind NS, Stratford IJ (2004) Hypoxia targeted gene therapy to increase the efficacy of tirapazamine as an adjuvant to radiotherapy: reversing tumor resistance and effecting cure. Cancer Res 64:1396–1402

Dachs GU, Patterson AV, Firth JD, Ratcliffe PJ, Townsend KMS, Stratford IJ, Harris AL (1997) Targeting gene expression to hypoxic tumor cells. Nat Med 3:515–520

Dachs GU, Tupper J, Tozer GM (2005) From bench to bedside for gene-directed enzyme prodrug therapy of cancer. Anti Cancer Drugs 16:349–359

Delahoussaye YM, Evans JW, Brown JM (2001) Metabolism of tirapazamine by multiple reductases in the nucleus. Bio- chem Pharmacol 62:1201–1209

Denekamp J, Stewart FA (1978) Sensitization of mouse tumours using fractionated X-irradiation. Br J Cancer 37:259–263 Denny WA, Wilson WR (1993) Bioreducible mustards: a para-

digm for hypoxia-selective prodrugs of diffusible cytotox- ins (HPDCs). Cancer Met Rev 12:135–151

Denny WA, Roberts PB, Anderson RF, Brown JM, Phil D, Wilson WR (1992) NLA-1: A 2-nitroimidazole radiosensitizer tar- geted to DNA by intercalation. Int J Radiat Oncol Biol Phys 22:553–556

Dirix LY, Tonnesen F, Cassidy J, Epelbaum R, Bokkel Huinink WW, Pavlidis N, Sorio R, Gamucci T, Wolff I, Te VA, Lan J, Verweij J (1996) EO9 Phase II study in advanced breast, gastric, pancreatic and colorectal carcinoma by the EORTC Early Clinical Studies Group. Eur J Cancer 32A:2019–2022 Dische S (1985) Chemical sensitisers for hypoxic cells: a

decade of experience in clinical radiotherapy. Radiother Oncol 3:97–115

Dische S, Anderson PJ, Sealy R, Watson ER (1983) Carcinoma of the cervix: anaemia, radiotherapy and hyperbaric oxygen.

Br J Radiol 56:251–255

Dische S, Chassagne D, Hope-Stone HF, Dawes PJDK, Roberts JT, Yosef H, Bey P, Horiot J-C, Jacobson A, Frankendal B, Gonzales Gonzales D, Nguyen TD, Daly NJ, Le Floch O, Newman H, Vieiro E, Bennett MH, Bichel P, Duvillard P, Jacobson A, Cook PA, Everett V, Machin D (1993) A trial of Ro 03-8799 (pimonidazole) in carcinoma of the uterine cervix: an interim report from the Medical Research Coun- cil Working Party on advanced carcinoma of the cervix.

Radiother Oncol 26:93–103

Dische S, Saunders MI, Lee ME, Adams GE, Flockhart IR (1977) Clinical testing of the radiosensitizer Ro 07-0582: experi- ence with multiple doses. Br J Cancer 35:567–579 Dorie MJ, Brown JM (1993) Tumor-specific, schedule-depend-

ent interaction between tirapazamine (SR 4233) and cispla- tin. Personal communication

Douglas RS, Patterson AV, Yang S, Wilson WR, Siim BG (2005) Induction of DNA damage and repair responses by the hypoxia-activated prodrug PR-104. AACR-NCI-EORTC Conference on Molecular Targets and Cancer Therapeutics:

Discovery, Biology, and Clinical Applications. Philadelphia Eschwege F, Sancho-Garnier H, Chassagne D, Brisgand D,

Guerra M, Malaise EP, Bey P, Busutti L, Cionini L, N’guyen T, Romanini A, Chavaudra J, Hill C (1997) Results of an European randomized trial of etanidazole combined with radiotherapy in head and neck carcinomas. Int J Radiat Oncol Biol Phys 39:275–281

Evans JW, Yudoh K, DeLahoussaye YM, Brown JM (1998) Tira- pazamine is metabolized to its DNA-damaging radical by intranuclear enzymes. Cancer Res 58:2098–2101

Fang J, Yan L, Shing Y, Moses MA (2001) HIF-1 alpha-mediated up-regulation of vascular endothelial growth factor, inde- pendent of basic fibroblast growth factor, is important in the switch to the angiogenic phenotype during early tum- origenesis. Cancer Res 61:5731–5735

Fang J, Xia C, Cao Z, Zheng JZ, Reed E, Jiang B-H (2005) Api- genin inhibits VEGF and HIF-1 expression via PI3K/AKT/

p70S6K1 and HDM2/p53 pathways. FASEB J 19:342–353 Flockhart IR, Large P, Troup D, Malcolm SH, Marten TR (1978)

Pharmacokinetic and metabolic studies of the hypoxic cell radiosensitizer misonidazole. Xenobiotica 8:97–105 Fox ME, Lemmon MJ, Mauchline ML, Davis TO, Giaccia AJ,

Minton NP, Brown JM (1996) Anaerobic bacteria as a delivery system for cancer gene therapy: activation of 5- fluorocytosine by genetically engineered clostridia. Gene Therapy 3:173–178

Fracasso PM, Sartorelli AC (1986) Cytotoxicity and DNA lesions produced by mitomycin C and porfiromycin in hypoxic and aerobic EMT6 and Chinese hamster ovary cells. Cancer Res 46:3939–3944

Fyles AW, Milosevic M, Pintilie M, Hill RP (1998) Cervix cancer oxygenation measured following external radiation ther- apy. Int J Radiat Oncol Biol Phys 42:751–753

Giaccia A, Siim BG, Johnson RS (2003) HIF-1 as a target for drug development. Nat Rev Drug Discovery 2:1–9 Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Lowe

SW, Giaccia AJ (1996) Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours.

Nature 379:88–91

Grahams MA, Senan S, Robin HJ, Eckhardt N, Lendrem D,

Hinks J, Greenslade D, Rampling R, Kaye SB, Roemeling R