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.
REVIEWS EXPLOITING TUMOUR HYPOXIA IN CANCER TREATMENT J Martin Brown* and William R Wilson‡ 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 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% O2 in 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 NATURE REVIEWS | C ANCER The presence of HYPOXIA — 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 clinical trials with patients undergoing radiotherapy in hyperbaric oxygen chambers, to try to force more oxygen 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 molecules — 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 conducted, 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, particularly 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 hallmark of solid cancers and whether it affected the outcome of radiotherapy The situation changed in the 1990s with the introduction 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 concentrations in human tumours are highly heterogeneous with many regions at very low levels (less than mmHg partial pressure of oxygen (pO2); mmHg corresponds to approximately 0.7% O2 in the gas phase or µM in solution), with median values much lower than the normal tissues from which the tumours arose (TABLE 1) Several investigators have now demonstrated unequivocally 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 apoptosis, which might lessen sensitivity to some anticancer agents; fourth, the action of some anticancer agents (for example, bleomycin) resembles that of radiation in that VOLUME | JUNE 2004 | REVIEWS Box | Radiation resistance of hypoxic cells Ionizing radiation, such as that used in radiotherapy, kills cells by producing DNA damage, particularly DNA doublestrand 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 (pO2) that produces sensitivity midway between the oxic and hypoxic responses is approximately 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 pO2 less than 10 mmHg) are more radioresistant than the less hypoxic tumours ⋅ a Ionizing radiation b Restitution DNA-H Reduction in hypoxic conditions DNA-H Survivng fraction RSH DNA• 0.1 0.01 O2 Oxidation in aerobic conditions 0.001 DNA-OO• DNA breaks Cell death Damage fixation 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 438 | JUNE 2004 | VOLUME oxygen increases the cytotoxicity of the DNA lesions they cause11,12; fifth, hypoxia upregulates genes involved in drug resistance, including genes encoding 13,14 P-GLYCOPROTEIN These clear links between hypoxia and intrinsic resistance to chemotherapy provide the ‘smoking gun’, yet, surprisingly, clinical studies investigating the role of hypoxia in response to chemotherapy have not been reported Finally, in addition to its effect on response to cytotoxic 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 angiogenesis17 and tumour invasion18, and is associated with a more metastatic phenotype of human cancers19,20 By enhancing metastasis, hypoxia can compromise curability of tumours by surgery 21,22 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 not occur in normal tissues and so are potentially 10 20 30 40 Dose (Gy) Aerobic Hypoxic exploitable in cancer therapy Four general strategies are now being developed: PRODRUGS activated by hypoxia; hypoxia-selective gene therapy; targeting the hypoxiainducible factor (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 In essence, hypoxia-selective cytotoxicity requires one-electron reduction of a relatively 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 illustrated 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 www.nature.com/reviews/cancer REVIEWS 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 (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 examples39 TPZ was a significant advance over the previously known classes because its differential toxicity 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 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 benzotriazine 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 a Normal b Tumour Blind ends Temporary occlusion Hypoxia Break in vessel walls Red blood cells AV shunt Figure | 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 NATURE REVIEWS | C ANCER VOLUME | JUNE 2004 | REVIEWS Table | Oxygenation of tumours and the surrounding normal tissue Tumour type Median tumour pO2* (number of patients) Median normal pO2* (number of patients) References Glioblastoma 4.9 (10) 5.6 (14) ND ND 128 129 Head and neck carcinoma 12.2 (30) 14.7 (23) 14.6 (65) 40.0 (14) 43.8 (30) 51.2 (65) 130 131 132 Lung cancer 7.5 (17) 38.5 (17) Q Le (personal communication) Breast cancer 10.0 (15) ND 133 Pancreatic cancer 2.7 (7) 51.6 (7) 134 Cervical cancer 5.0 (8) 5.0 (74) (86) 51 (8) ND ND 135 136 137 Prostate cancer 2.4 (59) 30.0 (59) 138 Soft-tissue sarcoma 6.2 (34) 18 (22) ND ND 139 140 *p02 measured 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 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 440 | JUNE 2004 | VOLUME the damaging species was the TPZ radical itself 25, 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 radical26,27 or 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 nonsmall-cell lung cancer and has been shown to be effective — the addition of TPZ to the standard cisplatin regimen doubled the overall response rate and significantly prolonged survival32 TPZ has also been tested in a randomized Phase II trial with cisplatin-based chemoradiotherapy of advanced head and neck cancer, and the preliminary results of this trial also show improved survival in the group treated with TPZ33 A Phase III study with cisplatin-based chemoradiotherapy 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 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 prodrugs in being activated by two-electron reduction, which is effected mainly by the CYP3A members of the cytochrome P450 family 35, which are strongly expressed in some human tumours35 Inhibition by oxygen results from competition between O2 and prodrug for binding at the reduced haem group in the enzyme active site, rather than from redox cycling36 Although AQ4 is selective 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 tumours38 and 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 developing 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 a Oxic cell 1e– reductases e– D–• D O2–• O2 b Hypoxic cell 1e– reductases D e– D–• Toxic drug Figure | 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 O2 is present b | Hypoxiaselective 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 www.nature.com/reviews/cancer REVIEWS O– N+ O– 1e– reductases O– e–, H + N+ + N N – H2O N • N NH BTZ• N • N+ NH2 N O– O2–• TPZ NH2 Topoisomerase ll poisoning and DNA doublestrand breaks OH• OH O2 TPZ• Figure | 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 drug126 The available evidence is that the doublestrand 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 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 culture and a favourable interaction with radiotherapy and chemotherapy in transplanted tumours47 The relatively low DNA-binding affinity of the chloroquinoline unit is probably important in allowing adequate penetration 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 O– a OH O OH N+ HN O HN N+ AQ4N O– b CH3 CH3 2e– reductases (CYP3A) N HN N NO2–• 1e– reductases N CH3 CH3 N CH3 CH3 NO N N DNA-targeted reactive electrophile N O2 NLCQ-1 Cl c O O2 O2–• N N HN Topoisomerase ll poisoning NO2 HN O OH CH3 CH3 OH NO2 –• NO2 CONH2 NH CONH2 CONH2 1e– reductases O2N O2N N Cl Cl O2–• O2 Cl DNA crosslinks O2N N N Cl Cl Cl SN 23862 Figure | 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 NATURE REVIEWS | C ANCER 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 benzotriazines to DNA intercalators has also recently been shown to greatly increase hypoxic cytotoxicity 49, 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 metabolized to reactive ELECTROPHILES or FREE RADICALS that 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 cytotoxins 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 triggering 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 EFFECT offers 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 heterogeneous 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 oxygen 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 compounds52,55 and transition metals56, which typically have KO2 values ≤ 0.1 µM The high KO2 value of TPZ is both a strength and a weakness On the one hand, it ensures activation in mildly hypoxic tumour cells at ‘intermediate’ O2 concentrations (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 O2 allows 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 contributes to the myelotoxicity of TPZ The newer generation of prodrugs offers a way out of this dilemma Using trigger units that are activated only at very low O2 concentrations 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 eliminate adjacent radioresistant cells at higher pO2 This more sophisticated approach effectively redefines the VOLUME | JUNE 2004 | 4 REVIEWS Box | 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 trial122 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 DTD124 NHOH NH2 CONH2 2e– reductases: NTR O2N DNA mono-adducts O2N N NO2 CONH2 N CONH2 O2N NO2 N NO2 CONH2 2e– reductases: NTR, DTD 1e– reductases: P450R 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(Xethyl)amine group, where X is an electrophile that can react with nucleophiles such as the N7 position of guanine 442 | JUNE 2004 | VOLUME HOHN CONH2 O Acetyl CoA N 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 MUSTARD analogue 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 prodrug 62 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 nitroreductase63 and recent studies confirm its bystander effect when activated by endogenous reductases under hypoxia (S M Pullen, A.V Patterson and W.R.W., unpublished observations) Importantly, activation of SN 23862 by endogenous one-electron reductases is readily inhibited by very low concentrations of oxygen An analogue development programme based on SN 23862 is well advanced; a watersoluble 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 cytotoxins on reduction in hypoxic cells include nitrobenzyl phosphoramidate mustards64, nitroheterocyclic methylquaternary salts 65, cobalt(III) complexes66 and indoloquinones67 O N H DNA crosslinks N 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 (e aq–), which is a much more powerful reductant than enzymes and is readily scavenged by O2 in 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 activation68 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 competition 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-substituted 5-fluorouracil derivatives71 — but none have yet provided convincing activity in transplanted www.nature.com/reviews/cancer REVIEWS a Oxygenated tissue Promoter Prodrug-metabolizing enzyme gene Hypoxiaresponsive elements Prodrug Toxic drug b Hypoxic tissue HIF-1 dimer mRNA Enzyme Prodrug Toxic drug Figure | Rationale for hypoxia-dependent gene therapy Shows how the hypoxic environment of tumours, which produces high levels of the hypoxia-inducible factor (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, HIF1 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 tumours Use of effectors with much greater cytotoxic potency than those that have been investigated so far will probably be needed Hypoxia-selective gene therapy 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 NATURE REVIEWS | C ANCER A key limitation of present day gene therapy of cancer 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 expression of a particular protein would be tumour specific One way to 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 tissues 72,73 HIF-1 comprises a dimer of HIF-1α and 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 conditions 17 HIF-1α expression is also associated with poor prognosis and resistance to therapy in head and neck cancer, ovarian cancer and 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 response 17 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 hypoxiaspecific 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 developed 74,75, and in vivo activity has been obtained in experimental tumour systems either by direct injection of adenoviral vectors containing the HRE promoters76 or using tumour cells stably transfected with HRE-regulated prodrug-activating enzymes 77,78 Unfortunately the latter systems, in which 100% of the tumour cells carry the hypoxiaresponsive gene, are not realistic to achieve in a clinical 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 socalled ‘bystander effect’ discussed above), efficient antitumour activity can be obtained with much lower percentages of transformed cells63,79 A further potential 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 expression to further enforce tumour selectivity78 A similar approach relies on hypoxia-selective metabolism of AQ4N by CYP3A4 as 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 VOLUME | JUNE 2004 | 4 REVIEWS HIF-1α staining Blood vessels Active drug Clostridia-filling necrosis Figure | 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-1α (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 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 utilization and tumour-cell survival85, its association with poor prognosis17, and the fact that growth of mouse xenografts is inhibited by loss of HIF-1 activity85 all 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 mutations in RAS, SRC or ERBB2 (also known as HER2/NEU) Therefore, targeting HIF-1 could potentially target the better oxygenated cells in the tumours 84 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 colleagues, 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 444 | JUNE 2004 | VOLUME inhibit HIF-1 transcription, and early reports have indicated that such compounds exist 87 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 VHLindependent 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 transplanted thymic lymphoma and to increase the efficacy of immunotherapy against larger tumours 88 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 vascularity, 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 signalling 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 compounds 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 target 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, sporeforming bacteria that become vegetative and grow only in the absence (or at very low levels) of oxygen Malmgren and Flanagan were the first to demonstrate this phenomenon by observing that tumourbearing mice died of tetanus within 48 hours of intravenous injection of C tetani spores, whereas non-tumour-bearing animals were unaffected 95 Möse and Möse96,97 later reported that a nonpathogenic clostridial strain, C butyricum M-55, localized and germinated in solid Ehrlich tumours in mice, causing extensive lysis without any concomitant www.nature.com/reviews/cancer REVIEWS CDEPT (Clostridial-dependent enzyme prodrug therapy) A cancer therapy using the nonpathogenic 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 prodrugactivating 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 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 nonpathogenic 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 similar approach using genetically modified, live attenuated Salmonella, which, although producing excellent colonization of transplanted tumours in mice103, produced only marginal colonization of human tumours in a Phase I clinical trial104 The reasons for the difference 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 convert the non-toxic 5-fluorocytosine to the toxic anticancer drug 5-fluorouracil Animal experiments have demonstrated the efficacy of this approach105 and clinical 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 efficiently106 and 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 tumours 107.) 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 advantages, not the least of which is its favorable intratumour distribution Because the prodrug-activating enzyme from clostridia will be at its highest concentration in areas adjacent to necrosis and far from blood vessels (FIG 6), this guarantees the highest activedrug concentrations in the distant cells and also minimizes 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 NATURE REVIEWS | C ANCER (5,6-dimethylxanthenone-4-acetic acid), which acts primarily by inducing expression of tumour-necrosis factor in tumours 109,110, 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 tumours 111,112 The ability of vascular-targeting agents to produce tumour necrosis has been shown to increase the colonization113 and antitumour activity114 of clostridial spores injected intravenously This is a significant development for CDEPT because even if only a modest increase in tumour necrosis occurs with the clinical 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 nonsmall-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 better: in particular, we know that the efficacy of tirapazamine and other hypoxic cytotoxins is reduced by their limited diffusion through tumour tissue to reach all of the hypoxic cells115 We now have tools for quantifying the ability of prodrugs, and their activated metabolites, 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 preclinical studies What is now needed is the development 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 unselected patients who have a mixture of hypoxic and better-oxygenated tumours runs the clear risk of rejecting a treatment that could be of significant benefit to a subset of patients The most appropriate means of assessing tumour hypoxia to perform such a selection is under active investigation 117 Clearly, there is much to be done to exploit the unique features of hypoxia, HIF-1α, other molecular targets upregulated under hypoxia and necrosis in human solid tumours, but the future is bright VOLUME | JUNE 2004 | 4 REVIEWS 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 446 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) 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 and 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) Brown, J M Clinical trials of radiosensitizers: what should we expect? Int J Radiat Oncol Biol Phys 10, 425–429 (1984) Overgaard, J Clinical evaluation of nitroimidazoles as modifiers of hypoxia in solid tumors Oncol Res 6, 509–518 (1994) 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) 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) 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) Tannock, I F Conventional cancer therapy: promise broken or promise delayed? Lancet 351 (Suppl 2), 9–16 (1998) Durand, R E The influence of microenvironmental factors during cancer therapy In vivo 8, 691–702 (1994) 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) 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) Batchelder, R M., Wilson, W R., Hay, M P & Denny, W A Oxygen dependence of the cytotoxicity of the enediyne antitumour antibiotic esperamicin A1 Br J Cancer Suppl 27, S52–S56 (1996) Comerford, K M et al Hypoxia-inducible factor-1dependent regulation of the multidrug resistance (MDR1) gene Cancer Res 62, 3387–3394 (2002) 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) 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 Yuan, J & Glazer, P M Mutagenesis induced by the tumor microenvironment Mutat Res 400, 439–446 (1998) Harris, A L Hypoxia — a key regulatory factor in tumour growth Nature Rev Cancer 2, 38–47 (2002) Pennacchietti, S et al Hypoxia promotes invasive growth by transcriptional activation of the met protooncogene Cancer Cell 3, 347–361 (2003) Rofstad, E K Microenvironment-induced cancer metastasis Int J Radiat Biol 76, 589–605 (2000) Subarsky, P & Hill, R P The hypoxic tumour microenvironment and metastatic progression Clin Exp Metastasis 20, 237–250 (2003) Hockel, M et al Association between tumor hypoxia and malignant progression in advanced cancer of the uterine cervix Cancer Res 56, 4509–4515 (1996) 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) 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 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) Brown, J M SR 4233 (tirapazamine): a new anticancer drug exploiting hypoxia in solid tumours Br J Cancer 67, 1163–1170 (1993) | JUNE 2004 | VOLUME 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 1,4dioxide 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 hypoxiaactivated topoisomerase II poison Cancer Res 62, 5248–5253 (2002) 30 Dorie, M J & Brown, J M Tumor-specific, scheduledependent 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 hypoxiaactivated 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 DNAintercalating 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]acridine4-carboxamide (DACA): diffusion and metabolism in 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 multicellular layers of tumor cells J Pharmacol Exp Ther 297, 1088–1098 (2001) 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) Hay, M P et al DNA-targeted 1,2,4-benzotriazine 1,4dioxides as hypoxia–selective analogues of tirapazamine J Med Chem 47, 475–488 (2004) 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) Koch, C J Unusual oxygen concentration dependence of toxicity of SR-4233, a hypoxic cell toxin Cancer Res 53, 3992–3997 (1993) 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) 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) 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) 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) 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) Lee, A E & Wilson, W R Hypoxia-dependent retinal toxicity of bioreductive anticancer prodrugs in mice Toxicol Appl Pharmacol 163, 50–59 (2000) 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) Cipolleschi, M G., Dello Sbarba, P & Olivotto, M The role of hypoxia in the maintenance of hematopoietic stem cells Blood 82, 2031–2037 (1993) 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) Helsby, N A et al Effect of nitroreduction on the alkylating reactivity and cytotoxicity of the 2,4-dinitrobenzamide-5aziridine CB 1954 and the corresponding nitrogen mustard SN 23862: distinct mechanisms of bioreductive activation Chem Res Toxicol 16, 469–478 (2003) 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 Borch, R F et al Synthesis and evaluation of nitroheterocyclic phosphoramidates as hypoxia-selective alkylating agents J Med Chem 43, 2258–2265 (2000) 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) 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) 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) 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) 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) Ahn, G., Ware, D C., Denny, W A & Wilson, W R Optimization of the auxiliary ligand shell of cobalt(III)(8hydroxyquinoline) complexes as model hypoxia-selective radiation–activated prodrugs Radiat Res (in the press) Shibamoto, Y., Zhou, L., Hatta, H., Mori, M & Nishimoto, S A novel class of antitumor prodrug, 1-(2′-oxopropyl)-5fluorouracil (OFU001), that releases 5-fluorouracil upon hypoxic irradiation Jpn J Cancer Res 91, 433–438 (2000) www.nature.com/reviews/cancer REVIEWS 72 Zhong, H et al Overexpression of hypoxia-inducible factor 1α in common human cancers and their metastases Cancer Res 59, 5830–5835 (1999) 73 Talks, K L et al The expression and distribution of the hypoxia-inducible factors HIF-1α and HIF-2α in normal human tissues, cancers, and tumor-associated macrophages Am J Pathol 157, 411–421 (2000) 74 Shibata, T., Akiyama, N., Noda, M., Sasai, K & Hiraoka, M Enhancement of gene expression under hypoxic conditions using fragments of the human vascular endothelial growth factor and the erythropoietin genes Int J Radiat Oncol Biol Phys 42, 913–916 (1998) 75 Greco, O & Dachs, G U Gene directed enzyme/prodrug therapy of cancer: historical appraisal and future prospectives J Cell Physiol 187, 22–36 (2001) 76 Binley, K et al Hypoxia-mediated tumour targeting Gene Ther 10, 540–549 (2003) 77 Shibata, T., Giaccia, A J & Brown, J M Hypoxia-inducible regulation of a prodrug-activating enzyme for tumor-specific gene therapy Neoplasia 4, 40–48 (2002) 78 Patterson, A V et al Oxygen-sensitive enzyme-prodrug gene therapy for the eradication of radiation-resistant solid tumours Gene Ther 9, 946–954 (2002) 79 Trinh, Q T., Austin, E A., Murray, D M., Knick, V C & Huber, B E Enzyme/prodrug gene therapy: comparison of cytosine deaminase/5- fluorocytosine versus thymidine kinase/ganciclovir enzyme/prodrug systems in a human colorectal carcinoma cell line Cancer Res 55, 4808–4812 (1995) 80 McCarthy, H O et al Bioreductive GDEPT using cytochrome P450 3A4 in combination with AQ4N Cancer Gene Ther 10, 40–48 (2003) 81 Griffiths, L et al The macrophage: a novel system to deliver gene therapy to pathological hypoxia Gene Ther 7, 255–262 (2000) 82 Burke, B et al Expression of HIF-1α by human macrophages: implications for the use of macrophages in hypoxia-regulated cancer gene therapy J Pathol 196, 204–212 (2002) 83 Semenza, G L Targeting HIF-1 for cancer therapy Nature Rev Cancer 3, 721–732 (2003) 84 Giaccia, A., Siim, B G & Johnson, R S HIF-1 as a target for drug development Nature Rev Drug Discov 2, 803–811 (2003) 85 Semenza, G L Involvement of hypoxia-inducible factor in human cancer Intern Med 41, 79–83 (2002) 86 Kung, A L., Wang, S., Klco, J M., Kaelin, W G & Livingston, D M Suppression of tumor growth through disruption of hypoxia-inducible transcription Nature Med 6, 1335–1340 (2000) 87 Rapisarda, A et al Identification of small molecule inhibitors of hypoxia-inducible factor transcriptional activation pathway Cancer Res 62, 4316–4324 (2002) 88 Sun, X et al Gene transfer of antisense hypoxia inducible factor-1α enhances the therapeutic efficacy of cancer immunotherapy Gene Ther 8, 638–645 (2001) 89 Mabjeesh, N J et al Geldanamycin induces degradation of hypoxia-inducible factor 1α protein via the proteosome pathway in prostate cancer cells Cancer Res 62, 2478–2482 (2002) 90 Mabjeesh, N J et al 2ME2 inhibits tumor growth and angiogenesis by disrupting microtubules and dysregulating HIF Cancer Cell 3, 363–375 (2003) 91 Yeo, E J et al YC-1: a potential anticancer drug targeting hypoxia-inducible factor J Natl Cancer Inst 95, 516–525 (2003) 92 Lemmon, M J et al Anaerobic bacteria as a gene delivery system to tumors Proc Am Assoc Cancer Res 35, 374 (1994) 93 Fox, M E et al Anaerobic bacteria as a delivery system for cancer gene therapy: activation of 5-fluorocytosine by genetically engineered clostridia Gene Ther 3, 173–178 (1996) 94 Lemmon, M L et al Anaerobic bacteria as a gene delivery system that is controlled by the tumor microenvironment Gene Ther 4, 791–796 (1997) 95 Malmgren, R A & Flanigan, C C Localization of the vegetative form of Clostridium tetani in mouse tumors following intravenous spore administration Cancer Res 15, 473–478 (1955) 96 Möse, J R & Möse, G Onkolyseversuche mit apathogenen anaeroben Sporenbildern am Ehrlich Tumor des Maus Z Krebsforsch 63, 63–74 (1959) 97 Möse, J R & Möse, G Oncolysis by clostridia I Activity of Clostridium butyricum (M-55) and other nonpathogenic clostridia against the Ehrlich carcinoma Cancer Res 24, 212–216 (1964) 98 Thiele, E H., Arison, R N & Boxer, G E Oncolysis by clostridia III Effects of clostridia and chemotherapeutic agents on rodent tumors Cancer Res 24 (1964) NATURE REVIEWS | C ANCER 99 Engelbart, K & Gericke, D Oncolysis by clostridia V Transplanted tumors of the hamster Cancer Res 24, 239–243 (1964) 100 Carey, R W., Holland, J F., Whang, H Y., Neter, E & Bryant, B Clostridial oncolysis in man Europ J Cancer 3, 37–46 (1967) 101 Heppner, F & Mose, J R The liquefaction (oncolysis) of malignant gliomas by a non pathogenic clostridium Acta Neuro 12, 123–125 (1978) 102 Heppner, F., Mose, J., Ascher, P W & Walter, G Oncolysis of malignant gliomas of the brain 13th Int Cong Chemother 226, 38–45 (1983) 103 Pawelek, J M., Low, K B & Bermudes, D Tumor-targeted Salmonella as a novel anticancer vector Cancer Res 57, 4537–4544 (1997) 104 Toso, J F et al Phase I study of the intravenous administration of attenuated Salmonella typhimurium to patients with metastatic melanoma J Clin Oncol 20, 142–152 (2002) 105 Liu, S C., Minton, N P., Giaccia, A J & Brown, J M Anticancer efficacy of systemically delivered anaerobic bacteria as gene therapy vectors targeting tumor hypoxia/necrosis Gene Ther 9, 291–296 (2002) First data demonstrating in vivo efficacy of CDEPT 106 Bridgewater, J A et al Expression of the bacterial nitroreductase enzyme in mammalian cells renders them selectively sensitive to killing by the prodrug CB1954 Eur J Cancer 31A, 2362–2370 (1995) 107 Minton, N P Clostridia in cancer therapy Nature Rev Microbiol (2003) 108 Martin, J et al Antibody-directed enzyme prodrug therapy: pharmacokinetics and plasma levels of prodrug and drug in a phase I clinical trial Cancer Chemother Pharmacol 40, 189–201 (1997) 109 Joseph, W R et al Stimulation of tumors to synthesize tumor necrosis factor-α in situ using 5,6dimethylxanthenone-4-acetic acid: a novel approach to cancer therapy Cancer Res 59, 633–638 (1999) 110 Zhao, L., Ching, L M., Kestell, P & Baguley, B C The antitumour activity of 5,6-dimethylxanthenone-4-acetic acid (DMXAA) in TNF receptor-1 knockout mice Br J Cancer 87, 465–470 (2002) 111 Ching, L M et al Induction of endothelial cell apoptosis by the antivascular agent 5,6-Dimethylxanthenone-4-acetic acid Br J Cancer 86, 1937–1942 (2002) 112 Galbraith, S M et al Effects of 5,6-dimethylxanthenone-4acetic acid on human tumor microcirculation assessed by dynamic contrast-enhanced magnetic resonance imaging J Clin Oncol 20, 3826–3840 (2002) 113 Theys, J et al Improvement of Clostridium tumour targeting vectors evaluated in rat rhabdomyosarcomas FEMS Immunol Med Microbiol 30, 37–41 (2001) 114 Dang, L H., Bettegowda, C., Huso, D L., Kinzler, K W & Vogelstein, B Combination bacteriolytic therapy for the treatment of experimental tumors Proc Natl Acad Sci USA 98, 15155–15160 (2001) 115 Hicks, K O., Pruijn, F B., Sturman, J R., Denny, W A & Wilson, W R Multicellular resistance to tirapazamine is due to restricted extravascular transport in HT29 multicellular layer cultures: A pharmacokinetic/pharmacodynamic study Cancer Res 63, 5970–5977 (2003) 116 Helsby, N A., Ferry, D M., Patterson, A V., Pullen, S M & Wilson, W R 2–amino metabolites are key mediatiors of CB 1954 and SN 23862 bystander effects in nitroreductase GDEPT Br J.Cancer 90, 1084–1092 (2004) 117 Bussink, J., Kaanders, J H A M & van der Kogel, A J Tumor hypoxia at the micro-regional level: clinical relevance and predictive value of exdogenous and endogenous hypoxic cell markers Radiother Oncol 67, 3–15 (2003) 118 Cobb, L M et al 2,4-Dinitro-5-ethyleneiminobenzamide (CB 1954): A potent and selective inhibitor of the growth of the Walker carcinoma 256 Biochem Pharmacol 18, 1519–1527 (1969) 119 Knox, R J., Friedlos, F., Jarman, M & Roberts, J J A new cytotoxic, DNA interstrand crosslinking agent, 5-(Aziridin-1YL)-4-hydroxylamino-2-nitrobenzamide, is formed from 5(Aziridin-1-YL)-2,4-dinitrobenzamide (CB 1954) by a nitroreductase enzyme in Walker carcinoma cells Biochem Pharm 37, 4661–4669 (1988) 120 Knox, R J et al The nitroreductase enzyme in Walker cells that activates 5-(Aziridin-1-YL)-2,4-dinitrobenzamide (CB 1954) to 5-(Aziridin-1-YL)-4-hydroxylamino-2nitrobenzamide is a form of NAD(P)H dehydrogenase (quinone) (EC 99 2) Biochem Pharmacol 37, 4671–4677 (1988) 121 Anlezark, G M et al The bioactivation of 5-(Aziridin-1-YL)2,4-dinitrobenzamide (CB 1954)-I Purification and properties of a nitroreductase enzyme from Escherichia Coli: A potential enzyme for antibody-directed enzyme prodrug therapy (ADEPT) Biochem Pharmacol 44, 2289–2295 (1992) 122 Chung-Faye, G et al Virus-directed, enzyme prodrug therapy with nitroimidazole reductase: a phase I and pharmacokinetic study of its prodrug, CB1954 Clin Cancer Res 7, 2662–2668 (2001) 123 Stratford, I J., Williamson, C., Hoe, S & Adams, G E Radiosensitizing and cytotoxicity studies with CB 1954 (2,4-dinitro-5-aziridinylbenzamide) Radiat Res 88, 502–509 (1981) 124 Palmer, B D., Wilson, W R., Cliffe, S & Denny, W A Hypoxiaselective antitumor agents Synthesis of water-soluble nitroaniline mustards with selective cytotoxicity for hypoxic mammalian cells J Med Chem 35, 3214–3222 (1992) 125 Brown, J M & Giaccia, A J The unique physiology of solid tumors: Opportunities (and problems) for cancer therapy Cancer Res 58, 1408–1416 (1998) 126 Wouters, B G et al Mitochondrial dysfunction after aerobic exposure to the hypoxic cytotoxin tirapazamine Cancer Res 61, 145–152 (2001) 127 Aebersold, D M et al Expression of hypoxia-inducible factor-1α: a novel predictive and prognostic parameter in the radiotherapy of oropharyngeal cancer Cancer Res 61, 2911–2916 (2001) Shows that HIF-1 can be used as an endogenous marker of tumour hypoxia to predict response to radiotherapy 128 Rampling, R., Cruickshank, G., Lewis, A D., Fitzsimmons, S A & Workman, P Direct measurement of pO2 distribution and bioreductive enzymes in human malignant brain tumors Int J Radiat Oncol Biol Phys 29, 427–431 (1994) 129 Collingridge, D R., Piepmeier, J M., Rockwell, S & Knisely, J P Polarographic measurements of oxygen tension in human glioma and surrounding peritumoural brain tissue Radiother Oncol 53, 127–131 (1999) 130 Nordsmark, M., Bentzen, S M & Overgaard, J Measurement of human tumour oxygenation status by a polarographic needle electrode Acta Oncol 33, 383–389 (1994) 131 Becker, A et al Oxygenation of squamous cell carcinoma of the head and neck: comparison of primary tumors, neck node metastases, and normal tissue Int J Radiat Oncol Biol Phys 42, 35–41 (1998) 132 Le, Q T et al Comparison of the comet assay and the oxygen microelectrode for measuring tumor oxygenation in head-and-neck cancer patients Int J Radiat Oncol Biol Phys 56, 375–383 (2003) 133 Vaupel, P., Briest, S & Hockel, M Hypoxia in breast cancer: pathogenesis, characterization and biological/therapeutic implications Wien Med Wochenschr 152, 334–342 (2002) 134 Koong, A C et al Pancreatic tumors show high levels of hypoxia Int J Radiat Oncol Biol Phys 48, 919–922 (2000) 135 Lyng, H., Sundfor, K & Rofstad, E K Oxygen tension in human tumours measured with polarographic needle electrodes and its relationship to vascular density, necrosis and hypoxia Radiother Oncol 44, 163–169 (1997) 136 Fyles, A W et al Oxygenation predicts radiation response and survival in patients with cervix cancer Radiother Oncol 48, 149–156 (1998) 137 Nordsmark, M et al Measurements of hypoxia using pimonidazole and polarographic oxygen-sensitive electrodes in human cervix carcinomas Radiother Oncol 67, 35–44 (2003) 138 Movsas, B et al Hypoxia in human prostate carcinoma: an Eppendorf PO2 study Am J Clin Oncol 24, 458–461 (2001) 139 Brizel, D M et al Radiation therapy and hyperthermia improve the oxygenation of human soft tissue sarcomas Cancer Res 56, 5347–5350 (1996) 140 Nordsmark, M et al The relationship between tumor oxygenation and cell proliferation in human soft tissue sarcomas Int J Radiat Oncol Biol Phys 35, 701–708 (1996) Acknowledgements The authors work is funded by grants from the United States National Institutes of Health (J.M.B and W.R.W.) and the Health Research Council of New Zealand (W.R.W.) Competing interests statement The authors declare competing financial interests: see web version for details Online links DATABASES The following terms in this article are linked online to: Cancer.gov: http://cancer.gov/ head and neck cancer | non-small-cell lung cancer | oesophageal cancer | ovarian cancer Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene CYP3A4 | ERBB2 | HIF-1α | HIF-1β | P450R | p53 | SRC Access to this interactive links box is free online VOLUME | JUNE 2004 | 4 ... fractionated irradiation of mouse tumors Cancer Res 50, 7745–7749 (1990) Brown, J M SR 4233 (tirapazamine): a new anticancer drug exploiting hypoxia in solid tumours Br J Cancer 67, 1163–1170 (1993) | JUNE... terms in this article are linked online to: Cancer. gov: http:/ /cancer. gov/ head and neck cancer | non-small-cell lung cancer | oesophageal cancer | ovarian cancer Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene... 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,