EXPLOITING TUMOUR HYPOXIA CANCER TREATMENT

EXPLOITING TUMOUR HYPOXIA IN CANCER TREATMENT Solid tumours contain regions at very low oxygen concentrations (hypoxia), often surrounding areas of necrosis. The cells in these hypoxic regions are resistant to both radiotherapy and chemotherapy. However, the existence of hypoxia and necrosis also provides an opportunity for tumourselective therapy, including prodrugs activated by hypoxia, hypoxiaspecific gene therapy, targeting the hypoxiainducible factor 1 transcription factor, and recombinant anaerobic bacteria. These strategies could turn what is now an impediment into a significant advantage for cancer therapy.

The presence ofHYPOXIA— regions of low levels of oxygen

— in human tumours was postulated by Thomlinson and Gray some 50 years ago based on their observations

of the distribution of necrosis relative to blood vessels1 It was known at that time that hypoxic cells were resistant to killing by ionizing radiation2 (BOX 1), and this led to clini-cal trials with patients undergoing radiotherapy in hyperbaric oxygen chambers, to try to force more oxy-gen into the blood and into the tumour These trials were not particularly successful, in part because it was not known at that time that, in addition to the chronic (or diffusion-limited) hypoxia postulated by Thomlinson and Gray, acute (perfusion-limited) hypoxia could also occur by temporary obstruction, or variable blood flow, in tumour vessels (FIG 1)

During the 1970s, interest in overcoming the assumed problem of the radiation resistance of hypoxic cells in tumours was rekindled with the discovery of small mole-cules — nitroimidazoles — that could mimic the effects

of oxygen and thereby sensitize hypoxic cells to radiation

Clinical trials adding nitroimidazoles (metronidazole, misonidazole and etanidazole) to radiotherapy were con-ducted, but in general did not result in significant improvements over radiotherapy alone, mainly because the toxicities of the drugs prevented them from being given at high enough doses3 Subsequently, however, a meta-analysis of all of the trials has shown that these drugs did produce a small but significant improvement in local control, compared with radiotherapy alone, particu-larly for head and neck cancers4 Nonetheless, the fact that the high expectations for hypoxic radiosensitizers in

combination with radiotherapy were not realized led researchers to question whether hypoxia was a hall-mark of solid cancers and whether it affected the outcome of radiotherapy

The situation changed in the 1990s with the intro-duction of a commercially available oxygen electrode (the ‘Eppendorf’ electrode), which enabled investigators

to make accurate measurements of oxygen levels in human tumours5 We now know that oxygen concentra-tions in human tumours are highly heterogeneous with many regions at very low levels (less than 5 mmHg par-tial pressure of oxygen (pO2); 5 mmHg corresponds to approximately 0.7% O2in the gas phase or 7 µM in solution), with median values much lower than the nor-mal tissues from which the tumours arose (TABLE 1) Several investigators have now demonstrated unequivo-cally that the extent of tumour hypoxia has a negative impact on the ability of radiotherapy to locally control tumours, because of the resistance of hypoxic cells to killing by radiation6,7 (BOX 1)

Hypoxic cells are also considered to be resistant to most anticancer drugs for several reasons: first, hypoxic cells are distant from blood vessels and, as a result, are not adequately exposed to some types of anticancer drugs8,9; second, cellular proliferation decreases as a function of distance from blood vessels10, an effect that

is at least partially due to hypoxia; third, hypoxia selects for cells that have lost sensitivity to p53-mediated apop-tosis, which might lessen sensitivity to some anticancer agents; fourth, the action of some anticancer agents (for example, bleomycin) resembles that of radiation in that

EXPLOITING TUMOUR HYPOXIA

IN CANCER TREATMENT

Solid tumours contain regions at very low oxygen concentrations (hypoxia), often surrounding areas

of necrosis The cells in these hypoxic regions are resistant to both radiotherapy and chemotherapy However, the existence of hypoxia and necrosis also provides an opportunity for tumour-selective therapy, including prodrugs activated by hypoxia, hypoxia-specific gene therapy, targeting the hypoxia-inducible factor 1 transcription factor, and recombinant anaerobic bacteria These strategies could turn what is now an impediment into a significant advantage for cancer therapy.

HYPOXIA

A low oxygen level However,

this means different levels to

different investigators depending

on the phenomenon being

investigated For the radiation

biologist, hypoxia occurs at

levels that produce severe

radiation resistance or levels less

than 0.1% O2in the gas phase.

Other effects of hypoxia occur at

oxygen levels above and below

this value.

*Division of Radiation and

Cancer Biology, Department

of Radiation Oncology,

Stanford, California

94305, USA.

‡ Experimental Oncology

Group, Auckland Cancer

Society Research Centre,

Private Bag 92019,

Auckland, New Zealand.

Correspondence to J.M.B.

e-mail:

mbrown@stanford.edu

doi:10.1038/nrc1367

exploitable in cancer therapy Four general strategies are now being developed:PRODRUGSactivated by hypoxia; selective gene therapy; targeting the hypoxia-inducible factor 1 (HIF-1) transcription factor; and the use of recombinant obligate anaerobic bacteria

Hypoxia-activated prodrugs in clinical trials

A common mechanism by which a non-toxic prodrug can be activated to a toxic drug in a hypoxia-dependent manner is shown in FIG 2 In essence, hypoxia-selective cytotoxicity requires one-electron reduction of a rela-tively non-toxic prodrug to a radical that then becomes a substrate for back-oxidation by oxygen to the original compound If the so-formed radical or downstream products of the radical are much more toxic than the superoxide generated by redox cycling in oxic cells, hypoxia-dependent cytotoxicity arises Examples of hypoxia-activated prodrugs in clinical trials are illus-trated below and those in preclinical development are discussed in the subsequent section

Tirapazamine Brown and Lee discovered the

hypoxic cytotoxicity of tirapazamine (TP2;FIG.3) almost 20 years ago, and this is the first compound to

P-GLYCOPROTEIN

A protein localized to the cell

membrane that actively pumps

many drugs out of the cell High

levels of this protein lead to

resistance to many anticancer

drugs.

PRODRUG

A latent form of a drug that can

be activated by metabolism or

other chemical transformation

in the body.

oxygen increases the cytotoxicity of the DNA lesions they cause11,12; fifth, hypoxia upregulates genes involved in drug resistance, including genes encoding

P-GLYCOPROTEIN13,14 These clear links between hypoxia and intrinsic resistance to chemotherapy provide the

‘smoking gun’, yet, surprisingly, clinical studies investi-gating the role of hypoxia in response to chemotherapy have not been reported

Finally, in addition to its effect on response to cyto-toxic therapy, it has also been demonstrated that hypoxia in tumours tends to select for a more malignant phenotype15, increases mutation rates16, increases expression of genes associated with angiogenesis17and tumour invasion18, and is associated with a more metastatic phenotype of human cancers19,20 By enhanc-ing metastasis, hypoxia can compromise curability of tumours by surgery21,22

Therefore, hypoxia has a key negative role in tumour prognosis both because it causes resistance to standard therapies and because it promotes a more malignant phenotype17 However, the very low levels of oxygen and the presence of necrosis are unique features of solid tumours — under normal physiological conditions they

do not occur in normal tissues and so are potentially

Box 1 | Radiation resistance of hypoxic cells

Ionizing radiation, such as that used in radiotherapy, kills cells by producing DNA damage, particularly DNA double-strand breaks This damage results from ionizations in or very close to the DNA that produce a radical on the DNA (DNA⋅) This radical then enters into a competition for oxidation, primarily by oxygen (which fixes, or makes permanent, the damage), or reduction, primarily by –SH-containing compounds that can restore the DNA to its original form (see figure, part a ) Therefore, DNA damage is less in the absence of oxygen This effect of oxygen in sensitizing cells

to radiation is illustrated in the cell-survival curve (see figure, part b ) and is quantitated as the ratio of dose in the absence of oxygen to dose in the presence of oxygen needed to obtain the same surviving fraction of cells For mammalian cells, this ratio is usually 2.5–3.0 (see horizontal dotted line) The oxygen partial pressure (pO 2 ) that produces sensitivity midway between the oxic and hypoxic responses is approximately 3 mmHg Clinical trials, particularly with head and neck cancers for which control of the primary tumour is the main problem, have demonstrated that the more hypoxic tumours (typically those with a median pO 2 less than 10 mmHg) are more radioresistant than the less hypoxic tumours.

1

0.1

0.01

0.001

Dose (Gy)

Ionizing radiation

DNA • DNA-H

DNA-H

death Oxidation in aerobic conditions

Reduction in hypoxic conditions RSH

Restitution

O2

examples39 TPZ was a significant advance over the previously known classes because its differential toxic-ity towards hypoxic cells was larger23, and combination studies with fractionated radiation demonstrated its ability to kill hypoxic cells in transplanted tumours24 The mechanism for selective toxicity of TPZ to hypoxic cells follows the general scheme outlined in FIG 2 However, whereas it had previously been proposed that

be developed specifically as a hypoxic cytotoxin and for which antitumour activity has been demonstrated

in clinical trials Before the discovery of this

benzotri-azine di-N-oxide, two other classes of agents were

known that produced some selective killing of hypoxic cells: quinone-containing alkylating agents, of which mitomycin C is the prototype; and nitroaromatic compounds, of which misonidazole and RB 6145 are

Summary

• A characteristic feature of solid tumours is the presence of cells at very low oxygen tensions These hypoxic cells confer radiotherapy and chemotherapy resistance to the tumours, as well as selecting for a more malignant phenotype.

• These hypoxic cells, however, provide a tumour-specific targeting strategy for therapy, and four approaches are being investigated: prodrugs activated by hypoxia; hypoxia-selective gene therapy; targeting the hypoxia-inducible factor 1 (HIF-1) transcription factor; and the use of recombinant obligate anaerobic bacteria.

• Tirapazamine is the prototype hypoxia-activated prodrug Its toxic metabolite, a highly reactive radical that is present

at higher concentrations under hypoxia, selectively kills the resistant hypoxic cells in tumours This makes the tumours much more sensitive to treatment with conventional chemotherapy and radiotherapy.

• Several other hypoxia-activated prodrugs, including AQ4N, NLCQ-1 and dinitrobenzamide mustards, are in preclinical or early clinical development.

• Hypoxia-activated gene therapy using hypoxia-specific promoters provides selective transcription of enzymes that can convert prodrugs into toxic drugs The efficacy of this approach has been shown in animal models, but clinical testing must await better systemic delivery of vectors to hypoxic cells.

• Targeting HIF-1 is a third strategy This protein is stabilized under hypoxic conditions and promotes the survival of tumour cells under hypoxic conditions Several strategies to inactivate or to exploit this unique protein in tumours are being investigated at the preclinical level.

• Finally, using recombinant non-pathogenic clostridia — an obligate anaerobe that colonizes tumour necrosis after systemic administration — is another strategy to exploit the unique physiology of solid tumours This approach has demonstrated considerable preclinical efficacy.

AV shunt

Red blood cells

Break in vessel walls

Blind ends

Hypoxia

Temporary occlusion

Figure 1 | The vascular network of normal tissue versus tumour tissue Tumours contain regions of hypoxia and necrosis because their vasculature can not supply oxygen and other vital nutrients to all the cells Whereas normal vasculature (a) is

hierarchically organized, with vessels that are sufficiently close to ensure adequate nutrient and oxygen supply to all cells, tumour

vessels (b) are chaotic, dilated, tortuous and are often far apart and have sluggish blood flow As a consequence, areas of hypoxia

and necrosis often develop distant from blood vessels In addition to these regions of chronic (or diffusion-limited) hypoxia, areas of acute (or perfusion-limited) hypoxia can develop in tumours as a result of the temporary closure or reduced flow in certain vessels Adapted from REF 125 AV, arteriovenous.

AQ4N The only other hypoxia-activated prodrug now

in clinical trials — the anthraquinone AQ4N (FIG 4)— was designed specifically for this purpose It resembles

TPZ in being a di-N-oxide, but has a distinct mechanism

of activation and cytotoxicity AQ4N is a prodrug of a potent DNA intercalator/topoisomerase poison, AQ4, which is formed by reduction of the two tertiary amine

N-oxide groups that mask DNA binding in the prodrug

form34 AQ4N is unusual among hypoxia-activated pro-drugs in being activated by two-electron reduction, which is effected mainly by the CYP3A members of the cytochrome P450 family35, which are strongly expressed

in some human tumours35 Inhibition by oxygen results from competition between O2and prodrug for binding

at the reduced haem group in the enzyme active site, rather than from redox cycling36 Although AQ4 is selec-tive for cycling cells, its long residence time in tissue probably enables it to persist until hypoxic cells come into cycle37 AQ4N has substantial activity against hypoxic cells in various transplanted tumours38and has recently completed a Phase I clinical trial The results of further clinical evaluation are awaited with interest

Hypoxia prodrugs in preclinical development

Originally the province of academic groups with an interest in the radiation resistance of hypoxic cells39–46, several pharmaceutical companies are also now devel-oping prodrugs for exploiting hypoxia The strategies that are being pursued in the development of improved hypoxia prodrugs are outlined below

DNA targeting One approach is to try to increase

potency by linking DNA-targeting units to moieties known to damage DNA in hypoxic cells An example is

the damaging species was the TPZ radical itself25, it now seems that the toxic species is an oxidizing radical formed by spontaneous decay of the protonated TPZ radical; this ultimate cytotoxin has been indicated to be either the hydroxyl radical26,27or a benzotriazinyl (BTZ) radical formed by loss of H2O28 The oxidizing radical gives rise to cytotoxic DNA double-strand breaks through a TOPOISOMERASE-II-dependent process29 (FIG 3) TPZ potentiates the antitumour effect of radiation

by selectively killing the hypoxic cells in the tumours

As these are the most radiation-resistant cells in tumours, TPZ and radiation act as complementary cytotoxins, each one killing the cells resistant to the other, thereby potentiating the efficacy of radiation on the tumour TPZ is also very effective in enhancing the anticancer activity of the chemotherapeutic drug cisplatin30, an interaction that again depends on hypoxia31, but that results from an increase in cisplatin sensitivity in non-lethally-damaged TPZ-treated cells rather than from complementary killing of oxic and hypoxic cells by the two agents, as is the case with radiation The interaction with cisplatin has been tested in a Phase III clinical trial with advanced non-small-cell lung cancerand has been shown to be effec-tive — the addition of TPZ to the standard cisplatin regimen doubled the overall response rate and signifi-cantly prolonged survival32 TPZ has also been tested

in a randomized Phase II trial with cisplatin-based chemoradiotherapy of advanced head and neck can-cer, and the preliminary results of this trial also show improved survival in the group treated with TPZ33 A Phase III study with cisplatin-based chemoradiother-apy is now underway Although TPZ seems to have clinical activity, and therefore provides important proof of principle for this approach, the dose that can

be administered during chemoradiation is limited by neutropaenia and other toxicities by as yet unknown mechanisms So, there is a clear need for improved hypoxia-activated prodrugs

TOPOISOMERASE II

An enzyme that catalyses

changes in DNA topology by

transiently cleaving and

re-ligating both strands of the

double helix This enzyme

catalyses the passage of one

DNA double-stranded molecule

through another.

Table 1 | Oxygenation of tumours and the surrounding normal tissue

(number of patients) (number of patients)

communication)

*p02measured in mmHg Measurements were made using a commercially available oxygen

electrode (the ‘Eppendorf’ electrode) The values shown are the median of the median values for

each patient ND, not determined; pO2, oxygen partial pressure.

Toxic drug

O2–• O2 1e – reductases

1e – reductases

e –

e –

a Oxic cell

b Hypoxic cell

Figure 2 | The usual mechanism by which prodrugs act as

hypoxia-selective cytotoxins The non-toxic prodrug (D) must

be a substrate for intracellular one-electron (1e – ) reductases, such as cytochrome P450 reductase, which add an electron to

the prodrug and therefore convert it to a free radical a | In oxic

cells, the unpaired electron in the prodrug radical is rapidly transferred to molecular oxygen, forming superoxide and regenerating the initial prodrug This futile redox cycle prevents build-up of the prodrug radical when O2is present b |

Hypoxia-selective cell killing is achieved if the prodrug radical that accumulates in hypoxic cells is more cytotoxic than the superoxide formed in oxic cells In principle, the prodrug radical could itself be the cytotoxin, but more commonly it undergoes further reactions to form the ultimate toxic species.

whether the tumour activity of NLCQ-1 is primarily because of radiosensitization or hypoxic cytotoxicity, but its toxicology is now being evaluated in anticipation

of a Phase I clinical trial The linkage of TPZ-like ben-zotriazines to DNA intercalators has also recently been shown to greatly increase hypoxic cytotoxicity49, and analogues with a range of DNA-binding affinities are under investigation50

Prodrugs of relatively stable cytotoxins that provide bystander effects Most of the first-generation hypoxic

cytotoxins (AQ4N is an exception) were quinones, nitro

compounds or aromatic N-oxides, which are

metabo-lized to reactive ELECTROPHILESor FREE RADICALSthat are not able to escape from the hypoxic cells in which they are generated Much of the recent research is focused on the development of prodrugs that release more stable cyto-toxins on reduction These prodrugs can be considered

to comprise three modular domains: a ‘trigger’ unit that

is reduced selectively under hypoxia; an ‘effector’, which

is the drug moiety responsible for cell killing (or other desired effect); and a ‘linker’, which transmits the trigger-ing event to the effector (for example, by fragmentation

or through an electronic change)43,51 The release of an active drug that can diffuse from the cell of origin to generate a BYSTANDER EFFECToffers a way of killing more than just the hypoxic subpopulation

in tumours42, and by partially decoupling activation from killing an important problem in exploiting hypoxia as a tumour-selective target is solved Although

it is certainly a valid generalization that hypoxia is more severe in tumours than in normal tissues, there is a het-erogeneous distribution of oxygen in many tissues The normal tissues that are known or suspected to include regions of mild (physiological) hypoxia include liver, bone marrow, skin, testis, retina and cartilage The oxy-gen concentration required to inhibit the activation of TPZ by 50% (KO2) is approximately 1–3 µM52,53, which

is considerably higher than for most quinones54, nitro compounds52,55and transition metals56, which typically have KO2values ≤ 0.1 µM The high KO2value of TPZ is both a strength and a weakness On the one hand, it ensures activation in mildly hypoxic tumour cells at

‘intermediate’ O2concentrations (1–25 µM) that are considered to be the most important in limiting the response to radiation therapy57 But on the other hand, this relative insensitivity to O2allows some activation in normal tissues, as illustrated by the irreversible toxicity

of TPZ to physiologically hypoxic photoreceptor cells in the retinal tissue of mice58 Retinal damage has not been

a problem in humans, but it is possible that hypoxia in the stem-cell compartment of bone marrow59,60 con-tributes to the myelotoxicity of TPZ The newer genera-tion of prodrugs offers a way out of this dilemma Using trigger units that are activated only at very low O2 con-centrations should make it possible to confine activation

to regions of severe hypoxia, which are essentially unique to tumours, whereas release of an effector that can cause bystander killing makes it possible to elimi-nate adjacent radioresistant cells at higher pO2 This more sophisticated approach effectively redefines the

the linkage of a chloroquinoline DNA-targeting unit to 2-nitroimidazole, as in the prodrug NLCQ-1 (FIG 4), which shows hypoxia-selective cytotoxicity in cell cul-ture and a favourable interaction with radiotherapy and chemotherapy in transplanted tumours47 The rela-tively low DNA-binding affinity of the chloroquinoline unit is probably important in allowing adequate pene-tration through tumour tissue, which can be severely constrained by avid DNA binding48 It is not yet clear

ELECTROPHILE

A chemical group that reacts

with electron-rich centres in

molecules.

FREE RADICAL

A compound with an unpaired

electron and that is usually very

reactive because of this feature.

N +

N

N +

O –

O –

N

N +

O –

OH

NH 2

– H 2 O

e–, H +

•

N N

N +

O –

NH

•

OH •

BTZ•

Topoisomerase ll poisoning and DNA double-strand breaks 1e – reductases

Figure 3 | The mechanism by which tirapazamine selectively kills hypoxic cells.

Tirapazamine (TPZ) is a substrate for one-electron (1e – ) reductases The resulting free radical

(TPZ • ) undergoes spontaneous decay to an oxidizing hydroxyl radical (OH • ) or an oxidizing

benzotriazinyl radical (BTZ • ) When oxygen is present, the TPZ radical is back-oxidized to the

parent compound, producing a superoxide radical (O2– • ) which might be responsible for the

muscle cramps seen in patients given the drug 126 The available evidence is that the

double-strand breaks are not caused directly by the oxidizing radical (OH • or BTZ • ), but, at least in part,

through poisoning of topoisomerase II29 This could be the result of radical damage directly to

the topoisomerase II enzyme, therefore poisoning it midway through its catalytic cycle and

producing a double-strand break in much the same way as etoposide; or the radical damage

to DNA could act as a substrate for topoisomerase II, so producing double-strand breaks

OH

HN

N + CH3

O – CH 3

N +

CH 3

O –

CH3

OH

HN

N CH3

CH 3

N

CH 3

CH 3

O2

N

N

NO 2

Cl

O 2–• O 2

N N

NO 2–•

N N NO

NO2

CONH2

O2N

N

Cl

Cl

2

4

NO 2 CONH 2

O 2 N N

Cl Cl

O 2–• O 2

NH2 CONH2

O2N N

Cl Cl

–•

Topoisomerase ll poisoning

DNA-targeted reactive electrophile

DNA crosslinks

2e – reductases (CYP3A)

a

b

c

1e – reductases

1e – reductases

AQ4N

NLCQ-1

SN 23862

Figure 4 | Mechanisms of activation under hypoxia of prodrugs a | AQ4N is reduced by a

2-electron (2e – ) process to form the potent DNA intercalator and topoisomerase inhibitor AQ4.

b | With NLCQ-1, DNA binding of the chloroquinoline unit is considered to target reactive

species, arising from reduction of the 2-nitroimidazole moiety to DNA c | SN 23862 is a latent

nitrogen mustard in which reduction (predominantly at the 2-nitro group) to the corresponding

amine under hypoxic conditions greatly increases the reactivity of the mustard moiety CYP3A,

cytochrome P450 3A.

Radiation-activated prodrugs One of the limitations

in restricting prodrug activation to severely hypoxic tissue is that a large proportion of such tissue is necrotic, lacking the enzymes and cofactors needed to reduce prodrugs It would therefore be very attractive

to activate prodrugs under hypoxia by reducing them with ionizing radiation (which is widely used to treat tumours despite the presence of significant hypoxia) rather than enzymes Radiolysis of water generates the aquated electron (eaq–), which is a much more powerful reductant than enzymes and is readily scav-enged by O2in oxic cells (to form superoxide) Such radiation-activated prodrugs offer several theoretical advantages in addition to expanding the extent of the hypoxic zone that can be used for prodrug activa-tion68 In particular, the fact that radiotherapy focuses the radiation field on the tumour and a small volume

of surrounding tissue provides tumour specificity additional to hypoxia alone In addition, the lack of requirement of enzymatic activation makes the approach independent of the reductase expression profile in the target tumour However the low yield of

eaq–during radiotherapy (approximately 20 µmol/kg over a typical course of 70 Gy), coupled with compe-tition with endogenous electron acceptors, will require release of very potent cytotoxins if this theoretically attractive approach is to be realized Three prodrug systems have been described that are efficiently activated by ionizing radiation under hypoxia — nitrobenzyl quaternary ammonium salts69, cobalt(III) complexes70, and oxypropyl-substi-tuted 5-fluorouracil derivatives71— but none have yet provided convincing activity in transplanted

therapeutic target as not simply hypoxia in any cell, but

as cells adjacent to regions of severe (pathological) hypoxia, therefore sparing physiologically hypoxic normal tissues

A class of hypoxia-activated prodrugs that fits this profile is the dinitrobenzamide mustards, illustrated by

SN 23862 (FIG 4) This is a NITROGEN MUSTARDanalogue of

an aziridine prodrug, CB 1954, which has long fascinated experimental oncologists (BOX 2) Reduction of either nitro group of SN 23862 acts as an electronic switch, redistributing electron density in the aromatic ring (linker) to activate the nitrogen mustard61 The 2-amine reduction product is a key metabolite under hypoxia, and shows a 2000-fold increase in alkylating reactivity and cytotoxic potency relative to the parent prodrug62 This metabolite is known to provide an efficient bystander effect in three-dimensional cultures of WiDr cells when

SN 23862 is metabolized by Escherichia coli

nitroreduc-tase63and recent studies confirm its bystander effect when activated by endogenous reductases under hypoxia (S M

Pullen, A.V Patterson and W.R.W., unpublished observa-tions) Importantly, activation of SN 23862 by endoge-nous one-electron reductases is readily inhibited by very low concentrations of oxygen An analogue development programme based on SN 23862 is well advanced; a water-soluble derivative (SN 28343) with excellent activity against hypoxic cells in transplanted murine and human tumours is in preclinical development

Other prodrugs that can release well-defined cyto-toxins on reduction in hypoxic cells include nitroben-zyl phosphoramidate mustards64, nitroheterocyclic methylquaternary salts65, cobalt(III) complexes66 and indoloquinones67

BYSTANDER EFFECT

Influence of a drug on

untargeted cells, in the present

context by diffusion of an

activated cytotoxin from

hypoxic cells to surrounding

cells at higher oxygen

concentrations.

NITROGEN MUSTARD

DNA-crosslinking alkylating

agents containing a

bis(X-ethyl)amine group, where X is an

electrophile that can react with

nucleophiles such as the N7

position of guanine.

Box 2 | CB 1954: prodrug extraordinaire

CB 1954 first came to attention because of its marked curative activity against the Walker rat tumour118 It was subsequently shown to be a bioreductive prodrug, activated within the tumours by rat DT-diaphorase (DTD), which reduces its 4-nitro group to the corresponding hydroxylamine119,120; following reaction of the hydroxylamine with acetyl CoA, the latter becomes a very potent DNA-crosslinking agent119 CB 1954 is even more efficiently activated by an Escherichia coli nitroreductase (NTR, the product of the nfsB gene)121, and has recently entered a clinical trial 122 as gene-dependent enzyme prodrug therapy (GDEPT) using a non-replicating adenoviral vector that expresses NTR NTR reduces either (but not both) nitro groups of CB 1954, and recent data indicate that the 2-amino reduction product might be the key metabolite responsible for bystander effects in NTR-GDEPT116 CB 1954 is also activated by one-electron (1e – ) reductases, such as cytochrome P450 reductase (P540R), selectively under hypoxia123, but the related dinitrobenzamide mustards (see main text) seem to have greater selectivity for hypoxia because of their lack of sensitivity to activation by DTD 124

NO 2 CONH 2

O 2 N N

2 4

NHOH CONH 2

O 2 N N

NO 2 CONH 2

HOHN N

NH 2 CONH 2

O 2 N N

NO 2 CONH 2

N N O

O

DNA crosslinks

DNA mono-adducts

2e – reductases:

NTR

Acetyl CoA

2e – reductases:

NTR, DTD 1e – reductases:

P450R

expression of a particular protein would be tumour specific One way to do this is to use the fact that the transcription factor HIF-1 is expressed at high levels

in most tumours, but not generally in normal tis-sues72,73 HIF-1 comprises a dimer ofHIF-1αand

HIF-1β, and it is the former that is increased in tumour cells both by increased transcription by transformed cells and by stabilization of the protein under hypoxic conditions17 HIF-1α expression is also associated with poor prognosis and resistance to therapy in head and neck cancer,ovarian cancerand

oesophageal cancer72,73 HIF-1α stimulates the transcription of a large number of genes involved in such processes as oxygen transport, angiogenesis, glycolysis and stress response17 Transcription of all of these genes is effected by the binding of the HIF-1 dimer to sequences known as hypoxia-responsive elements (HREs) in the promoter regions of the target genes Therefore, the strategy indicated to obtain hypoxia-specific transcription of a therapeutic gene would be

to develop a promoter that is highly responsive to HIF-1 that would therefore drive the expression of the therapeutic gene specifically in tumours (FIG 5) Expression of an enzyme that is not normally found

in the human body (for example, cytosine deaminase

derived from E coli) could, under the control of a

hypoxia-responsive promoter, convert a non-toxic prodrug into a toxic drug in the tumour Promoters using HREs from hypoxia-responsive genes have been developed74,75, and in vivo activity has been

obtained in experimental tumour systems either by direct injection of adenoviral vectors containing the HRE promoters76or using tumour cells stably trans-fected with HRE-regulated prodrug-activating enzymes77,78 Unfortunately the latter systems, in which 100% of the tumour cells carry the hypoxia-responsive gene, are not realistic to achieve in a clini-cal situation However, this might not be a crucial limitation: various investigators have shown that, provided the active drug can diffuse from the cell in which it is generated to kill surrounding cells (the so-called ‘bystander effect’ discussed above), efficient antitumour activity can be obtained with much lower percentages of transformed cells63,79 A further poten-tial strategy for exploiting hypoxia in gene therapy is

to deliver the gene encoding a one-electron reductase such as cytochrome P450 reductase (P450R) as the prodrug-activating therapeutic gene, so confining prodrug activation to hypoxic regions; this can be combined with HIF-1 regulation of P450R expres-sion to further enforce tumour selectivity78 A similar approach relies on hypoxia-selective metabolism of AQ4N by CYP3A4as an enzyme–prodrug system for gene-directed enzyme prodrug therapy (GDEPT)80

A challenge with these approaches will be achieving efficient systemic delivery of vectors to cells expressing HIF-1 and/or hypoxic cells, which are generally found

in regions distant from blood vessels One possibility for delivery of the HRE-driven therapeutic protein to tumours would be to take advantage of the fact that

tumours Use of effectors with much greater cyto-toxic potency than those that have been investigated

so far will probably be needed

Hypoxia-selective gene therapy

A key limitation of present day gene therapy of can-cer is the lack of specificity of the gene-delivery system Accordingly, essentially all of the protocols now being investigated in cancer gene therapy involve local administration of the delivery vectors directly into the tumour, usually by needle injection

Although this might be useful in some cases, it has limited applicability to cancer in general because metastases from the primary tumour are usually too numerous, inaccessible or undetected to allow for direct injection An alternative to direct targeting of tumours is to have the therapeutic gene transcribed

or translated by a tumour-specific property so that

GDEPT

(Gene-directed enzyme prodrug

therapy) A cancer treatment

strategy that aims to deliver a

prodrug-activating enzyme

specifically to tumour cells using

gene therapy The anticancer

effect would be achieved by

subsequent systemic

administration of the non-toxic

prodrug, which would be

converted to a toxic drug

preferentially in the tumour cells.

Promoter Prodrug-metabolizing enzyme gene

Hypoxia-responsive elements

a Oxygenated tissue

b Hypoxic tissue

Prodrug

Prodrug Toxic drug

Toxic drug

mRNA

Enzyme

HIF-1 dimer

Figure 5 | Rationale for hypoxia-dependent gene

therapy Shows how the hypoxic environment of tumours,

which produces high levels of the hypoxia-inducible factor 1 (HIF-1) transcription factor, could be used in gene therapy to produce tumour-specific expression of an enzyme that can metabolize a non-toxic prodrug into a toxic drug selectively in

the tumour a | In oxygenated tissue, there is little or no HIF-1

transcription factor Also, oxygen inhibits HIF-1 transactivation Consequently, no prodrug-metabolizing enzyme is produced, so little or none of the prodrug is

converted to the toxic drug b | In hypoxic tumour tissue,

HIF-1 is produced and downstream genes are transcribed following binding of HIF-1 to the hypoxia-responsive elements (HREs) in the promoter region of the genes.

Therefore, the prodrug-activating enzyme with HREs in its promoter will be transcribed and, after translation, activate the prodrug to the toxic drug selectively in the tumour.

inhibit HIF-1 transcription, and early reports have indicated that such compounds exist87 However, anticancer effects directly attributable to inhibition of HIF-1 transactivation have yet to be reported

A second approach is to suppress HIF-1 protein levels, either by destabilizing the protein or inhibiting its production The heat-shock protein 90 inhibitor geldanamycin has been shown to reduce HIF-1 protein levels by promoting its oxygen and VHL-independent degradation through the proteasome88,89 However, it has yet to be demonstrated that this

occurs in vivo or that the antitumour activity of this

compound is a direct result of reduced levels of HIF-1,

as many other proteins are also affected Targeting of HIF-1 by direct injection of an antisense construct to HIF-1α has been shown to eradicate a small trans-planted thymic lymphoma and to increase the efficacy

of immunotherapy against larger tumours88 However, small-molecule inhibitors of HIF-1 would

be preferable, and two groups have reported success Mabjeesh and colleagues reported that microtubule inhibitors such as 2-methoxyestradiol, vincristine and paclitaxel reduce HIF-1α levels in vitro apparently by

inhibiting translation of HIF-1α mRNA90 These compounds can also reduce tumour growth and vas-cularity, but whether this is an effect of reduced levels

of HIF-1 or a direct effect on microtubules is not known The second small molecule that has been reported to reduce HIF-1α levels and inhibit tumour growth is the soluble guanylyl cyclase stimulator YC-1 (REF 91) Soluble guanylyl cyclase is the receptor for nitric oxide (NO) — a molecule involved in many sig-nalling pathways, including those regulating vascular tone and platelet function However, the authors attribute the antitumour and anti-angiogenic effects

of YC-1 to a reduction in HIF-1α protein levels (by an unknown post-translational effect) rather than to an effect on NO signalling

A third approach would be to screen for com-pounds that are preferentially toxic to cells expressing HIF-1α At present this is a theoretical possibility with no published data demonstrating its efficacy

Recombinant anaerobic bacteria

Brown and colleagues first indicated that the necrotic regions in human solid tumours could be used to tar-get cancer therapy to tumours using a genetically engineered non-pathogenic strain of the bacterial

genus Clostridium92–94 This genus comprises a large and heterogeneous group of Gram-positive, spore-forming bacteria that become vegetative and grow only in the absence (or at very low levels) of oxygen Malmgren and Flanagan were the first to demon-strate this phenomenon by observing that tumour-bearing mice died of tetanus within 48 hours of

intravenous injection of C tetani spores, whereas

non-tumour-bearing animals were unaffected95 Möse and Möse96,97 later reported that a

non-pathogenic clostridial strain, C butyricum M-55,

localized and germinated in solid Ehrlich tumours in mice, causing extensive lysis without any concomitant

macrophages are often recruited to tumours and that such macrophages show increased levels of HIF-1α in various human tumours81,82

Targeting HIF-1

The characteristics, functions and possibilities for targeting HIF-1 in cancer therapy have been recently reviewed83,84 Its role in angiogenesis, glucose utiliza-tion and tumour-cell survival85, its association with poor prognosis17, and the fact that growth of mouse xenografts is inhibited by loss of HIF-1 activity85all make it a potentially attractive tumour-specific target It should be noted that the expression of HIF-1 is not restricted to hypoxic cells alone in many tumours, but is also upregulated by oncogenic muta-tions in RAS, SRC or ERBB2 (also known as HER2/NEU) Therefore, targeting HIF-1 could potentially target the better oxygenated cells in the tumours84 Three general approaches could be used

to exploit the high levels of HIF-1α in cancers

First, inhibition of transactivation of HIF-1 target genes (such as the angiogenesis inducer vascular endothelial growth factor) would be expected to have

an antitumour effect Proof of principle of this approach comes from studies by Kung and col-leagues, who showed that tumour cells infected with

a polypeptide that disrupted the binding of HIF-1α

to its transcriptional coactivators p300/CREB, thereby inhibiting hypoxia-induced transcription, markedly reduced the growth of these cells when transplanted into nude mice86 These data have led investigators to screen for small molecules that

HIF-1 α staining

Blood vessels Active drug Clostridia-filling necrosis

Figure 6 | Clostridial-dependent enzyme prodrug

therapy—simulation of how it might work

A photomicrograph of a section of a human head and neck cancer biopsy immunostained for the hypoxia-inducible factor-1 α (HIF-1α) The likely distribution of viable anaerobic bacteria (clostridia) and the concentration of active drug formed from the reaction of the prodrug with the enzyme expressed by the recombinant clostridia in the tumour have been simulated by drawing over the necrotic regions The concentrations of the active drug will probably be highest next to areas of necrosis and far from blood vessels.

Photomicrograph modified with permission from REF 127

© (2001) American Association for Cancer Research

(5,6-dimethylxanthenone-4-acetic acid), which acts primarily by inducing expression of tumour-necrosis factor in tumours109,110, and the tubulin-binding agent combretastatin 4A and its analogue ZD 6126 These agents produce a rapid and selective occlusion

of tumour blood vessels, leading to necrosis within 16–24 hours of administration Phase I clinical trials

of DMXAA, combretastatin 4A and ZD 6126 have recently been completed and have demonstrated reduced blood flow in human tumours111,112 The ability of vascular-targeting agents to produce tumour necrosis has been shown to increase the colo-nization113and antitumour activity114of clostridial spores injected intravenously This is a significant development for CDEPT because even if only a mod-est increase in tumour necrosis occurs with the clini-cal use of vascular-targeting agents, this could be exploited to great advantage with clostridial spores that grow exclusively in necrotic areas

Future directions

The benefits of exploiting tumour hypoxia have yet to

be fully realized Despite this, the positive clinical results with the combination of the hypoxic cytotoxin tirapazamine with cisplatin to treat advanced non-small-cell lung cancer and with chemoradiotherapy to treat advanced head and neck cancer demonstrate the potential of this approach There is good reason to expect that future drugs or strategies will do better: in particular, we know that the efficacy of tirapazamine and other hypoxic cytotoxins is reduced by their lim-ited diffusion through tumour tissue to reach all of the hypoxic cells115 We now have tools for quantifying the ability of prodrugs, and their activated metabo-lites, to diffuse in tumour tissue48,63,115,116, so designing second-generation prodrugs with properly optimized micropharmacokinetic properties is a clear possibility for the future

The other strategies discussed above — hypoxia activated gene therapy, targeting HIF-1 and the use of recombinant clostridia — are certainly promising and have demonstrated antitumour efficacy in pre-clinical studies What is now needed is the develop-ment of the optimum drug or vector combination for each strategy and their clinical testing Relevant to the latter is the appropriate selection of patients with whom to use hypoxia-directed treatments As with any targeted anticancer strategy, hypoxia-directed therapy can only be effective on those tumours expressing the target or, in this case, with sufficient levels of hypoxia Performing clinical trials on unse-lected patients who have a mixture of hypoxic and better-oxygenated tumours runs the clear risk of rejecting a treatment that could be of significant ben-efit to a subset of patients The most appropriate means of assessing tumour hypoxia to perform such

a selection is under active investigation117 Clearly, there is much to be done to exploit the unique fea-tures of hypoxia, HIF-1α, other molecular targets upregulated under hypoxia and necrosis in human solid tumours, but the future is bright

effect on normal tissues Such observations were soon confirmed and extended by several investigators using tumours in mice, rats, hamsters and rabbits98,99, and were followed by clinical studies with patients with cancer100–102 Although the anaerobic bacteria did not significantly alter tumour control or eradication, these clinical reports demonstrated that spores of non-pathogenic strains of clostridia could be given safely, that the spores germinate in the necrotic regions of tumours, and that lysis in these tumour regions can occur This is an important distinction over the simi-lar approach using genetically modified, live

attenu-ated Salmonella, which, although producing excellent

colonization of transplanted tumours in mice103, pro-duced only marginal colonization of human tumours

in a Phase I clinical trial104 The reasons for the differ-ence between the rodent and human tumours in

colonization by Salmonella are unknown However,

colonization by clostridia is different from that of

Salmonella in being dependent on hypoxic necrotic

regions, which are equally common in human and rodent tumours In addition, as noted above, excellent colonization of human tumours has been reported following intravenous injection of clostridial spores

The Clostridium used in the clinical studies was a strain of C sporogenes, renamed C oncolyticum to

reflect the lysis that was produced in human tumours

This strain has been genetically modified to express

the E coli enzyme cytosine deaminase, which can

con-vert the non-toxic 5-fluorocytosine to the toxic anti-cancer drug 5-fluorouracil Animal experiments have demonstrated the efficacy of this approach105and clin-ical studies are planned In addition, other enzyme–prodrug systems for arming clostridia are in development, including CB 1954 (BOX 2),which, when

activated by E coli nitroreductase, kills non-cycling

cells efficiently106and is therefore expected to have greater activity against cells in hypoxic regions (See

a recent review for further details of possible enzyme–prodrug combinations that can be used with clostridial targeting of tumours107.) Although clostridial-dependent enzyme prodrug therapy (CDEPT)

is similar to the strategy of antibody-dependent enzyme prodrug therapy (ADEPT), which is now under clinical evaluation, it has several significant advan-tages, not the least of which is its favorable intratu-mour distribution Because the prodrug-activating enzyme from clostridia will be at its highest concen-tration in areas adjacent to necrosis and far from blood vessels (FIG 6),this guarantees the highest active-drug concentrations in the distant cells and also mini-mizes the problem of leakage of activated drug back into the blood vessels, which has been reported to be a problem for ADEPT108

Improving CDEPT with vascular targeting drugs A

way in which targeting of clostridia to solid tumours might be improved still further, and perhaps extended to very small tumours that have not yet developed necrotic areas, is by the addition of

VASCULAR-TARGETING AGENTS Such agents include DMXAA

CDEPT

(Clostridial-dependent enzyme

prodrug therapy) A cancer

therapy using the

non-pathogenic species of the

obligate anaerobe genus

clostridia that have been

genetically engineered to express

a prodrug-activating enzyme.

This is used to activate a prodrug

within the hypoxic/necrotic

regions that are colononized by

the bacterium.

ADEPT

(Antibody-directed enzyme

prodrug therapy) A cancer

treatment strategy that involves

conjugation of a

prodrug-activating enzyme (such as

cytosine deaminase, which

converts the non-toxic prodrug

5-fluorocytosine to the

anticancer drug 5-fluorouracil)

to a tumour-targeting antibody.

VASCULAR-TARGETING AGENT

Drugs that damage existing

blood vessels and therefore

interfere with blood flow in

tumours.

1 Thomlinson, R H & Gray, L H The histological structure

of some human lung cancers and the possible

implications for radiotherapy Br J Cancer 9, 539–549

(1955).

2 Gray, L H., Conger, A D., Ebert, M., Hornsey, S & Scott, O C.

Concentration of oxygen dissolved in tissues at the time of

irradiation as a factor in radiotherapy Br J Radiol 26,

638–648 (1953).

References 1 and 2 are classic papers describing both

the general nature of the oxygen effect in reducing

radiation sensitivity (reference 2) and the fact that the

hypoxic cells almost certainly are present in human

tumours (reference 1).

3 Brown, J M Clinical trials of radiosensitizers: what should we

expect? Int J Radiat Oncol Biol Phys 10, 425–429 (1984).

4 Overgaard, J Clinical evaluation of nitroimidazoles as

modifiers of hypoxia in solid tumors Oncol Res 6, 509–518

(1994).

5 Vaupel, P., Schlenger, K., Knoop, C & Hockel, M.

Oxygenation of human tumors: evaluation of tissue

oxygen distribution in breast cancers by computerized O2

tension measurements Cancer Res 51, 3316–3322

(1991).

6 Nordsmark, M., Overgaard, M & Overgaard, J.

Pretreatment oxygenation predicts radiation response in

advanced squamous cell carcinoma of the head and neck.

Radiother Oncol 41, 31–40 (1996).

7 Brizel, D M., Dodge, R K., Clough, R W & Dewhirst, M W.

Oxygenation of head and neck cancer: changes during

radiotherapy and impact on treatment outcome Radiother.

Oncol 53, 113–117 (1999).

8 Tannock, I F Conventional cancer therapy: promise

broken or promise delayed? Lancet 351 (Suppl 2), 9–16

(1998).

9 Durand, R E The influence of microenvironmental

factors during cancer therapy In vivo 8, 691–702 (1994).

10 Tannock, I F The relation between cell proliferation and the

vascular system in a transplanted mouse mammary tumour.

Br J Cancer 22, 258–273 (1968).

11 Teicher, B A., Lazo, J S & Sartorelli, A C Classification of

antineoplastic agents by their selective toxicities toward

oxygenated and hypoxic tumor cells Cancer Res 41,

73–81 (1981).

12 Batchelder, R M., Wilson, W R., Hay, M P & Denny, W A.

Oxygen dependence of the cytotoxicity of the enediyne

anti-tumour antibiotic esperamicin A1 Br J Cancer Suppl 27,

S52–S56 (1996).

13 Comerford, K M et al Hypoxia-inducible

factor-1-dependent regulation of the multidrug resistance (MDR1)

gene Cancer Res 62, 3387–3394 (2002).

14 Wartenberg, M et al Regulation of the multidrug resistance

transporter P-glycoprotein in multicellular tumor spheroids

by hypoxia-inducible factor (HIF-1) and reactive oxygen

species FASEB J 17, 503–505 (2003).

15 Graeber, T G et al Hypoxia-mediated selection of cells with

diminished apoptotic potential in solid tumours Nature 379,

88–91 (1996).

Describes how hypoxia in tumours selects against

wild-type p53 by causing apoptosis in these cells.

16 Yuan, J & Glazer, P M Mutagenesis induced by the tumor

microenvironment Mutat Res 400, 439–446 (1998).

17 Harris, A L Hypoxia — a key regulatory factor in tumour

growth Nature Rev Cancer 2, 38–47 (2002).

18 Pennacchietti, S et al Hypoxia promotes invasive growth by

transcriptional activation of the met protooncogene Cancer

Cell 3, 347–361 (2003).

19 Rofstad, E K Microenvironment-induced cancer

metastasis Int J Radiat Biol 76, 589–605 (2000).

20 Subarsky, P & Hill, R P The hypoxic tumour

microenvironment and metastatic progression Clin Exp.

Metastasis 20, 237–250 (2003).

21 Hockel, M et al Association between tumor hypoxia and

malignant progression in advanced cancer of the uterine

cervix Cancer Res 56, 4509–4515 (1996).

22 Brizel, D M et al Tumor oxygenation predicts for the

likelihood of distant metastases in human soft tissue

sarcoma Cancer Res 56, 941–943 (1996).

23 Zeman, E M., Brown, J M., Lemmon, M J., Hirst, V K &

Lee, W W SR-4233: a new bioreductive agent with high

selective toxicity for hypoxic mammalian cells Int J Radiat.

Oncol Biol Phys 12, 1239–1242 (1986).

The initial paper describing the hypoxic selectivity of

tirapazamine for cell killing.

24 Brown, J M & Lemmon, M J Potentiation by the hypoxic

cytotoxin SR 4233 of cell killing produced by fractionated

irradiation of mouse tumors Cancer Res 50, 7745–7749

(1990).

25 Brown, J M SR 4233 (tirapazamine): a new anticancer drug

exploiting hypoxia in solid tumours Br J Cancer 67,

1163–1170 (1993).

26 Daniels, J S & Gates, K S DNA cleavage by the antitumor agent 3-amino-1,2,4-benzotriazine 1,4-dioxide (SR4233):

Evidence for involvement of hydroxyl radical J Am Chem.

Soc 118, 3380–3385 (1996).

27 Zagorevskii, D et al A mass spectrometry study of

tirapazamine and its metabolites insights into the mechanism of metabolic transformations and the

characterization of reaction intermediates J Am Soc Mass

Spectrom 14, 881–892 (2003).

28 Anderson, R F., Shinde, S S., Hay, M P., Gamage, S A &

Denny, W A Activation of 3-amino-1,2,4-benzotriazine 1,4-dioxide antitumor agents to oxidizing species following their

one-electron reduction J Am Chem Soc 125, 748–756

(2003).

29 Peters, K B & Brown, J M Tirapazamine: a

hypoxia-activated topoisomerase II poison Cancer Res 62,

5248–5253 (2002).

30 Dorie, M J & Brown, J M Tumor-specific, schedule-dependent interaction between tirapazamine (SR 4233) and

cisplatin Cancer Res 53, 4633–4636 (1993).

In vivo data showing that tirapazamine and cisplatin

have a marked hypoxia and schedule-dependent synergism.

31 Kovacs, M S et al Cisplatin anti-tumour potentiation by

tirapazamine results from a hypoxia-dependent cellular

sensitization to cisplatin Br J Cancer 80, 1245–1251

(1999).

32 von Pawel, J et al Tirapazamine plus cisplatin versus

cisplatin in advanced non-small-cell lung cancer: A report of

the international CATAPULT I study group J Clin Oncol 18,

1351–1359 (2000).

Clinical data showing the efficacy of tirapazamine in combination with cisplatin in a randomized multicentre Phase III trial.

33 Rischin, D et al Preliminary results of TROG 98.02—a

randomized phase II study of 5-fluorouracil, cisplatin and radiation versus tirapazamine, cisplatin and radiation for advanced squamous cell carcinoma of the head and neck.

Proc Am Soc Clin Oncol 22, A1992 (2003).

34 Patterson, L H Rationale for the use of aliphatic N-oxides of cytotoxic anthraquinones as prodrug DNA binding agents: a

new class of bioreductive agent Cancer Metastasis Rev 12,

119–134 (1993).

35 Patterson, L H Bioreductively activated antitumor N-oxides: the case of AQ4N, a unique approach to

hypoxia-activated cancer chemotherapy Drug Metab Rev 34,

581–592 (2002).

36 Patterson, L H & Murray, G I Tumour cytochrome P450

and drug activation Curr Pharm Des 8, 1335–1347

(2002).

37 Patterson, L H & McKeown, S R AQ4N: a new approach

to hypoxia-activated cancer chemotherapy Br J Cancer

83, 1589–1593 (2000).

38 Patterson, L H et al Enhancement of chemotherapy and

radiotherapy of murine tumours by AQ4N, a bioreductively

activated anti-tumour agent Br J Cancer 82, 1984–1990

(2000).

Demonstration of potentiation of radiation and chemotherapy by the bioreductive agent AQ4N.

39 Adams, G E & Stratford, I J Bioreductive drugs for cancer

therapy: the search for tumor specificity Int J Radiat.

Oncol Biol Phys 29, 231–238 (1994).

40 Brown, J M & Siim, B G Hypoxia-specific cytotoxins in

cancer therapy Semin Rad Onc 6, 22–36 (1996).

41 de Groot, F M., Damen, E W & Scheeren, H W Anticancer prodrugs for application in monotherapy: targeting hypoxia,

tumor-associated enzymes, and receptors Curr Med.

Chem 8, 1093–1122 (2001).

42 Denny, W A & Wilson, W R Bioreducible mustards: a paradigm for hypoxia-selective prodrugs of diffusible cytotoxins

(HPDCs) Cancer Metastasis Rev 12, 135–151 (1993).

43 Naylor, M A & Thomson, P Recent advances in

bioreductive drug targeting Mini Rev Med Chem 1, 17–29

(2001).

44 Rauth, A M., Melo, T & Misra, V Bioreductive therapies: an

overview of drugs and their mechanisms of action Int J.

Radiat Oncol Biol Phys 42, 755–762 (1998).

45 Rockwell, S Use of hypoxia-directed drugs in the therapy of

solid tumors Semin Oncol 19, 29–40 (1992).

46 Workman, P & Stratford, I J The experimental development of bioreductive drugs and their role in cancer

therapy Cancer Metastasis Rev 12, 73–82 (1993).

47 Papadopoulou, M V & Bloomer, W D NLCQ-1 (NSC 709257): exploiting hypoxia with a weak

DNA-intercalating bioreductive drug Clin Cancer Res 9,

5714–5720 (2003).

48 Hicks, K O., Pruijn, F B., Baguley, B C & Wilson, W R.

Extravascular transport of the DNA intercalator and topoisomerase poison N-[2-(Dimethylamino)ethyl]acridine-4-carboxamide (DACA): diffusion and metabolism in

multicellular layers of tumor cells J Pharmacol Exp Ther.

297, 1088–1098 (2001).

49 Delahoussaye, Y M., Hay, M P., Pruijn, F B., Denny, W A & Brown, J M Improved potency of the hypoxic cytotoxin

tirapazamine by DNA-targeting Biochem Pharmacol 65,

1807–1815 (2003).

50 Hay, M P et al DNA-targeted 1,2,4-benzotriazine

1,4-dioxides as hypoxia–selective analogues of tirapazamine

J Med Chem 47, 475–488 (2004).

51 Denny, W A., Wilson, W R & Hay, M P Recent

developments in the design of bioreductive drugs Br J.

Cancer Suppl 27, S32–S38 (1996).

52 Koch, C J Unusual oxygen concentration dependence of

toxicity of SR-4233, a hypoxic cell toxin Cancer Res 53,

3992–3997 (1993).

53 Hicks, K O., Siim, B G., Pruijn, F B & Wilson, W R Oxygen dependence of the metabolic activation and cytotoxicity of tirapazamine: Implications for exptravascular

transport and activity in tumors Radiat Res (in the press).

54 Marshall, R S & Rauth, A M Oxygen and exposure kinetics as factors influencing the cytotoxicity of porfiromycin, a mitomycin C analogue, in Chinese hamster

ovary cells Cancer Res 48, 5655–5659 (1988).

55 Siim, B G., Atwell, G J & Wilson, W R Oxygen dependence of the cytotoxicity and metabolic activation of

4-alkylamino-5-nitroquinoline bioreductive drugs Br J.

Cancer 70, 596–603 (1994).

56 Wilson, W R., Moselen, J W., Cliffe, S., Denny, W A & Ware, D C Exploiting tumor hypoxia through bioreductive release of diffusible cytotoxins: the cobalt(III)-nitrogen

mustard complex SN 24771 Int J Radiat Oncol Biol.

Phys 29, 323–327 (1994).

57 Wouters, B G & Brown, J M Cells at intermediate oxygen levels can be more important than the ‘hypoxic fraction’ in determining tumor response to fractionated radiotherapy.

Radiat Res 147, 541–550 (1997).

58 Lee, A E & Wilson, W R Hypoxia-dependent retinal toxicity

of bioreductive anticancer prodrugs in mice Toxicol Appl.

Pharmacol 163, 50–59 (2000).

59 Allalunis, M J., Chapman, J D & Turner, A R Identification

of a hypoxic population of bone marrow cells Int J Radiat.

Oncol Biol Phys 9, 227–232 (1983).

60 Cipolleschi, M G., Dello Sbarba, P & Olivotto, M The role of hypoxia in the maintenance of hematopoietic stem cells.

Blood 82, 2031–2037 (1993).

61 Siim, B G., Denny, W A & Wilson, W R Nitro reduction as

an electronic switch for bioreductive drug activation Oncol.

Res 9, 357–369 (1997).

62 Helsby, N A et al Effect of nitroreduction on the alkylating

reactivity and cytotoxicity of the 2,4-dinitrobenzamide-5-aziridine CB 1954 and the corresponding nitrogen mustard

SN 23862: distinct mechanisms of bioreductive activation.

Chem Res Toxicol 16, 469–478 (2003).

63 Wilson, W R et al Quantitation of bystander effects in

nitroreductase suicide gene therapy using three-dimensional

cell cultures Cancer Res 62, 1425–1432 (2002).

Demonstrates the use of three-dimensional cell cultures to show bystander effects from GDEPT.

64 Borch, R F et al Synthesis and evaluation of

nitroheterocyclic phosphoramidates as hypoxia-selective

alkylating agents J Med Chem 43, 2258–2265 (2000).

65 Tercel, M et al Hypoxia-selective antitumor agents 16.

Nitroarylmethyl quaternary salts as bioreductive prodrugs of

the alkylating agent mechlorethamine J Med Chem 44,

3511–3522 (2001).

66 Wilson, W R., Moselen, J W., Cliffe, S., Denny, W A & Ware, D C Exploiting tumor hypoxia through bioreductive release of diffusible cytotoxins: the cobalt(III)-nitrogen

mustard complex SN 24771 Int J Radiat Oncol Biol.

Phys 29, 323–327 (1994).

67 Everett, S A et al Modifying rates of reductive elimination of

leaving groups from indolequinone prodrugs: a key factor in

controlling hypoxia-selective drug release Biochem.

Pharmacol 63, 1629–1639 (2002).

68 Wilson, W R., Tercel, M., Anderson, R F & Denny, W A Radiation-activated prodrugs as hypoxia-selective cytotoxins: model studies with nitroarylmethyl quaternary

salts Anticancer Drug Des 13, 663–685 (1998).

69 Kriste, A G., Tercel, M., Anderson, R F., Ferry, D M & Wilson, W R Pathways of reductive fragmentation of heterocyclic nitroarylmethyl quaternary ammonium prodrugs

of mechlorethamine Radiat Res 158, 753–762 (2002).

70 Ahn, G., Ware, D C., Denny, W A & Wilson, W R Optimization of the auxiliary ligand shell of cobalt(III)(8-hydroxyquinoline) complexes as model hypoxia-selective

radiation–activated prodrugs Radiat Res (in the press).

71 Shibamoto, Y., Zhou, L., Hatta, H., Mori, M & Nishimoto, S.

A novel class of antitumor prodrug, 1-(2 ′-oxopropyl)-5-fluorouracil (OFU001), that releases 5-′-oxopropyl)-5-fluorouracil upon

hypoxic irradiation Jpn J Cancer Res 91, 433–438 (2000).