Carbonic anhydrase IX is a predictive marker of doxorubicin resistance in early-stage breast cancer independent of HER2 and TOP2A amplification.

BACKGROUND: In early-stage breast cancer, adjuvant chemotherapy is associated with significant systemic toxicity with only a modest survival benefit. Therefore, there is considerable interest in identifying predictive markers of response to therapy. Doxorubicin, one of the most common drugs used to treat breast cancer, is an anthracycline chemotherapeutic agent, a class of drugs known to be affected by hypoxia. Accordingly, we examined whether expression of the endogenous hypoxia marker carbonic anhydrase IX (CA IX) is predictive of outcome in early-stage breast cancer patients treated with doxorubicin. METHODS: We obtained 209 early-stage pre-treatment surgically-resected breast tumours from patients, who received doxorubicin in their chemotherapeutic regimen and had >10 years of follow-up. Immunohistochemistry was used to detect CA IX, and we used fluorescence in situ hybridisation to detect both human epidermal growth factor receptor (HER2) and DNA topoisomerase II-alpha (TOP2A) gene amplification. RESULTS: Carbonic anhydrase IX intensity was significantly correlated with progression-free survival (PFS) and overall survival (OS) in patients receiving 300 mg m(-2) of doxorubicin (HR=1.82 and 3.77; P=0.0014 and 0.010, respectively). There was a significant, inverse correlation between CA IX score and oestrogen receptor expression, but no significant correlations were seen with either HER2 or TOP2A ratio. CONCLUSION: We demonstrate that CA IX expression is correlated with worse PFS and OS for breast cancer patients treated with doxorubicin, independent of HER2 or TOP2A gene amplification. This study provides evidence that using CA IX to detect hypoxia in surgically-resected breast tumours may be of clinical use in choosing an appropriate chemotherapy regimen.

Potential for a novel manganese porphyrin compound as adjuvant canine lymphoma therapy.

PURPOSE: Manganese porphyrins are redox-active drugs and superoxide dismutase mimics, which have been shown to chemosensitize lymphoma, a cancer which frequently occurs in dogs. This study aimed to identify critical information regarding the pharmacokinetics and toxicity of Mn(III) meso-tetrakis (N-n-butoxyetylpyridium-2-yl) porphyrin, (MnTnBuOE-2-PyP5+, MnBuOE) in dogs as a prelude to a clinical trial in canine lymphoma patients. METHODS: A single-dose pharmacokinetic (PK) study in normal dogs was performed to determine the plasma half-life (t 1/2) of MnBuOE. A dose reduction study was performed to establish the maximum tolerated dose (MTD) of MnBuOE. The safety and PK of a multi-dosing protocol was assessed. RESULTS: Peak plasma drug concentration occurred 30 min post-injection. The t 1/2 was defined as 7 h. MnBuOE induced an anaphylactic reaction and prolonged tachycardia. The MTD was defined as 0.25 mg/kg. The dogs were given MTD 3×/week for 2-3 weeks. The highest recorded tissue drug levels were in the lymph nodes (4-6 µM), followed by kidney and liver (2.5, 2.0 uM, respectively). CONCLUSIONS: We obtained critical information regarding the PK and toxicity of MnBuOE in dogs. The acute drug reaction and tachycardia post-injection have not been described in other species and may be specific to canines. The high tissue drug levels in lymph nodes have not been previously reported. MnBuOE accumulation in lymph nodes has important implications for the utility of adjuvant MnBuOE to treat lymphoma. With MnBuOE lymph node accumulation, reduction in the dose and/or administration frequency could be possible, leading to reduced toxicity.

Initial stages of tumor cell-induced angiogenesis: evaluation via skin window chambers in rodent models.

BACKGROUND: There is a paucity of information about events that follow immediately after tumor cells are triggered to initiate the process of angiogenesis (the formation of new blood vessels). Such information is relevant to the issue of when micrometastases vascularize and has implications for the accessibility of micrometastases to various treatments. In this study, we attempted to monitor events at the initiation of angiogenesis at the earliest possible stage of tumor growth in vivo. METHODS: Two different rodent mammary tumor cell lines, R3230Ac from the Fischer 344 rat and 4T1 from the BALB/c mouse, were stably transfected with a gene that encodes an enhanced version of green fluorescence protein (GFP). GFP-labeled R3230Ac or 4T1 cells (about 20-50 cells) were implanted into dorsal skinfold window chambers of Fischer 344 rats or BALB/c mice, respectively. Tumor angiogenesis was then monitored serially and noninvasively for up to 4 weeks. RESULTS: Clear evidence of modification of the host vasculature was observed when tumor mass reached approximately 60-80 cells, and functional new blood vessels were seen when tumor mass reached roughly 100-300 cells. Individual tumor cells exhibited a chemotaxis-like growth pattern toward the pre-existing host vasculature. When ex-flk1 (a soluble, truncated vascular endothelial cell growth factor receptor protein known to be antiangiogenic) was injected with the tumor cells, the initial angiogenic and tumor growth activities were inhibited considerably, indicating that angiogenesis inhibitors may halt tumor growth even before the onset of angiogenesis. CONCLUSION: Angiogenesis induced by tumor cells after implantation in the host begins at a very early stage, i.e., when the tumor mass contains roughly 100-300 cells. Identification of chemotactic signals that initiate tumor cell migration toward the existing vasculature may provide valuable targets for preventing tumor progression and/or metastases.

Transcriptional response to hypoxia in human tumors.

BACKGROUND: The presence of hypoxic regions within solid tumors is associated with a more malignant tumor phenotype and worse prognosis. To obtain a blood supply and protect against cellular damage and death, oxygen-deprived cells in tumors alter gene expression, resulting in resistance to therapy. To investigate the mechanisms by which cancer cells adapt to hypoxia, we looked for novel hypoxia-induced genes. METHODS: The transcriptional response to hypoxia in human glioblastoma cells was quantified with the use of serial analysis of gene expression. The time course of gene expression in response to hypoxia in a panel of various human tumor cell lines was measured by real-time polymerase chain reaction. Hypoxic regions of human carcinomas were chemically marked with pimonidazole. Immunohistochemistry and in situ hybridization were used to examine gene expression in the tumor's hypoxic regions. RESULTS: From the 24 504 unique transcripts expressed, 10 new hypoxia-regulated genes were detected-all induced, to a greater extent than vascular endothelial growth factor, a hypoxia-induced mitogen that promotes blood vessel growth. These genes also responded to hypoxia in breast and colon cancer cells and were activated by hypoxia-inducible factor 1, a key regulator of hypoxic responses. In tumors, gene expression was limited to hypoxic regions. Induced genes included hexabrachion (an extracellular matrix glycoprotein), stanniocalcin 1 (a calcium homeostasis protein), and an angiopoietin-related gene. CONCLUSIONS: We have identified the genes that are transcriptionally activated within hypoxic malignant cells, a crucial first step in understanding the complex interactions driving hypoxia response. Within our catalogue of hypoxia-responsive genes are novel candidates for hypoxia-driven angiogenesis.

The utility of thermal dose as a predictor of tumor and normal tissue responses to combined radiation and hyperthermia.

A total of 236 dogs and cats with a variety of cancers were randomized to receive radiation (XRT) or heat plus XRT. In those tumors which were heated, thermal gradients developed which varied in temperature minima and maxima. The influence of the thermal gradient characteristics on tumor and normal tissue responses was examined by correlation of response with the magnitude of gradient minima and maxima. Using multivariate analysis, the influence of other factors such as tumor histology, volume, site, heat treatment method, and number of heat fractions on tumor response was examined. Of all factors examined, tumor volume and non-site-specific average minimum equivalent min at 43 degrees emerged as consistent predictors of both complete response rate (p less than 0.001) and duration response (p less than 0.05). No significant enhancement of moist desquamation or late fibrosis was seen for heat + XRT versus XRT alone. The incidence of direct thermal injury to skin was positively correlated with maximum intratumoral equivalent min at 43 degrees. These results indicate that a therapeutic gain is achievable with heat + XRT, but successful application of the therapy is dependent on achieving high tumor thermal gradient minima and low maxima.

Hyperthermia enables tumor-specific nanoparticle delivery: effect of particle size.

The efficacy of novel cancer therapeutics has been hampered by the ability to deliver these agents to the tumor at effective concentrations. Liposomes have been used as a method to overcome some delivery issues and, in combination with hyperthermia, have been shown to increase drug delivery to tumors. Particle size has been shown to affect the delivery of liposomes, but it is not known how hyperthermia affects size dependence. This study investigates the effect of hyperthermia (42 degrees C) on the extravasation of different sized nanoparticles (albumin; 100-, 200-, and 400-nm liposomes) from tumor microvasculature in a human tumor (SKOV-3 ovarian carcinoma) xenograft grown in mouse window chambers. In this model (at 34 degrees C), no liposomes were able to extravasate into the tumor interstitium. Hyperthermia enabled liposome extravasation of all sizes. The magnitude of hyperthermia-induced extravasation was inversely proportional to particle size. Thus, at normothermia (34 degrees C), the pore cutoff size for this model was between 7 and 100 nm (e.g., liposomes did not extravasate). At 42 degrees C, the pore cutoff size was increased to >400 nm, allowing all nanoparticles tested to be delivered to the tumor interstitium to some degree. With hyperthermia, the 100-nm liposome experienced the largest relative increase in extravasation from tumor vasculature. Hyperthermia did not enable extravasation of 100-nm liposomes from normal vasculature, potentially allowing for tumor-specific delivery. These experiments indicate that hyperthermia can enable and augment liposomal drug delivery to tumors and potentially help target liposomes specifically to tumors.

A new temperature-sensitive liposome for use with mild hyperthermia: characterization and testing in a human tumor xenograft model.

The single biggest challenge now facing drug delivery (for liposomes and indeed other carriers) is to initiate and produce release of the encapsulated drug only at the diseased site and at controllable rates. Our efforts have focused on developing a new thermal-sensitive drug delivery system, specifically for the local control of solid tumors. We describe here a new lipid formulation containing doxorubicin that has been optimized for both mild hyperthermic temperatures (39 degrees C to 40 degrees C) that are readily achievable in the clinic and rapid release times of drug (tens of seconds). This new liposome, in combination with mild hyperthermia, was found to be significantly more effective than free drug or current liposome formulations at reducing tumor growth in a human squamous cell carcinoma xenograft line (FaDu), producing 11 of 11 complete regressions lasting up to 60 days posttreatment.

Available volume fraction of macromolecules in the extravascular space of a fibrosarcoma: implications for drug delivery.

Steric exclusion of molecules in the extravascular space of tissues can be quantified by the available volume fraction (K(AV)). Despite its clinical importance, however, there is a paucity of data in the literature regarding the available volume fraction of macromolecules in the extravascular space of tumor tissues. In this study, we quantified K(AV) of inulin, BSA, and dextran molecules of Mr 10,000-2,000,000 in polymer gels and fibrosarcoma tissues. The measurement involved: (a) sectioning of gels or tumor tissues into thin slices (approximately 600 microm) using a Vibratome, (b) ex vivo incubation of the slices in solutions containing fluorescently labeled tracers, and (c) quantification of the equilibrium tracer concentrations in both slices and solutions. We found that K(AV) in gels decreased monotonically when the Mr of dextran was increased from Mr 10,000 to 2,000,000. However, K(AV) in tumor tissues was insensitive to the molecular weight of dextran in the range between Mr 10,000 and 40,000. There was a sharp decrease in K(AV) from 0.28 +/- 0.14 to 0.10 +/- 0.06 when the molecular weight was increased from Mr 40,000 to 70,000. In addition to the molecular weight dependence, K(AV) was heterogeneous in tumors, with intertumoral difference being greater than intratumoral variation. The interstitial fluid space, which was quantified by K(AV) of inulin, was 50% of the total tissue volume. These data indicate that the fraction of the extravascular volume in tumors that is accessible to large therapeutic agents is heterogeneous and depends on the size of agents.

Characterization of the effect of hyperthermia on nanoparticle extravasation from tumor vasculature.

The efficacy of novel cancer therapeutics can be hampered by inefficient delivery of agents to the tumor at effective concentrations. Liposomes have been used as a method to overcome some delivery issues and, in combination with hyperthermia, have been shown to increase drug delivery to tumors. This study investigates the effects of a range of temperatures (34-42 degrees C) and hyperthermia treatment scheduling (time between hyperthermia and drug administration as well as between consecutive hyperthermia treatments) on the extravasation of nanoparticles (100-nm liposomes) from tumor microvasculature in a human tumor (SKOV-3 ovarian carcinoma) xenograft grown in athymic nude mouse window chambers. Under normothermic conditions (34 degrees C) and at 39 degrees C, nanoparticles were unable to extravasate into the tumor interstitium. From 40 to 42 degrees C, nanoparticle extravasation increased with temperature, reaching maximal extravasation at 42 degrees C. Temperatures higher than 42 degrees C led to hemorrhage and stasis in tumor vessels. Enhanced nanoparticle extravasation was observed several hours after heating, decaying back to baseline at 6 h postheating. Reheating (42 degrees C for 1 h) 8 h after an initial heating (42 degrees C for 1 h) did not result in any increased nanoparticle extravasation, indicating development of vascular thermotolerance. The results of this study have implications for the application and scheduling of hyperthermia combined with other therapeutics (e.g., liposomes, antibodies, and viral vectors) for the treatment of cancer.

Diminished leukocyte-endothelium interaction in tumor microvessels.

Leukocyte-endothelium interaction in vivo consists of the rolling of leukocytes along the vascular wall and, under certain conditions, their adherence to endothelial cells. In a rat tumor microcirculation model (mammary adenocarcinoma implanted in rat skinfold window chamber), we demonstrated that this interaction, measured as flux of rolling leukocytes and density of adhering leukocytes, was significantly reduced in tumor microvessels compared to normal microvessels, both under control conditions and during inflammation induced by N-formylmethionylleucylphenylalanine (1 microM), bacterial lipopolysaccharide (1 microgram/ml), or tumor necrosis factor alpha (500 units/ml). We also measured the blood flow shear rate in the tumor and normal microvessels and found that the difference in shear rate between the two types of microvessels could not account for the differences in leukocyte-endothelium interaction. The diminished leukocyte-endothelium interaction in tumors under various stimulated conditions suggests that a number of adhesion molecules may not be expressed properly on tumor endothelial cells.

Importance of minimum tumor temperature in determining early and long-term responses of spontaneous canine and feline tumors to heat and radiation.

A total of 130 dogs and cats with squamous cell carcinomas, melanomas, fibrosarcomas, mammary adenocarcinomas, or mast cell sarcomas were randomized to receive radiation (XRT) or heat plus XRT. Time-temperature data for each monitored tumor location were converted to degree-minutes or equivalent min at 43 degrees (Eq43). Response rates and durations of response were compared for subgroups of histology, volume, site, and heat treatment method. Thermal gradients existed in all heated tumors. The influence of these gradients on tumor response was examined by correlation of response with degree-minutes and Eq43 minima, maxima, averages, and ranges. A pattern emerged from these analyses linking dose minima, maxima, and ranges with prognostic subgroups as classified by volume, site, or treatment method. The data indicated that the coolest part of the tumor governed the biological response to combined heat + XRT. Tumors which received a minimum of 35 Eq43 had significantly longer durations of response than did those receiving XRT alone or less than 3 Eq43 (p less than or equal to 0.006 and 0.014, respectively; log-rank test). Furthermore, broad temperature ranges were associated with power-limiting "hot spots" and invariably led to underheating in other areas of tumor. Multivariate analysis found minimum Eq43 on the first treatment to be the best predictor of long-term response (p less than 0.05). Other biological covariates of site, volume, and histology contributed strength to the model, which was independent of Eq43 (p less than 0.05).

Persistent genetic instability in cancer cells induced by non-DNA-damaging stress exposures.

A hallmark of cancer cells is their pronounced genetic instability, which has been implicated in both tumor development and negative treatment outcomes. Recently, it has been reported that ionizing radiation may induce a persistent state of hypermutability in mammalian cells that lasts for many (>30) cell divisions. In this study, we examined whether other stress signals (both DNA-damaging non-DNA-damaging) can initiate a similar process. We show that persistent genetic instability was induced by nongenotoxic stress exposures such as heat treatment, serum starvation, or the tumor microenvironment, as well as genotoxic stresses such as ionizing radiation and exposure to hydrogen peroxide. Progeny of 10-20% of surviving cells exhibited persistent and pronounced genetic instability at both an artificially transfected gene and a genomic minisatellite locus 23 cell divisions after the initial exposure. Stress-induced persistent genetic instability may be a general response of tumor cells to a wide range of genotoxic or nongenotoxic stress conditions.

Determination of local oxygen consumption rates in tumors.

At any location in a respiring tissue, partial pressure of oxygen (PO2) is influenced by the local oxygen consumption rate. Consumption rates in vascular tumor tissues have previously been estimated for macroscopic regions. Using oxygen electrodes, we measured PO2 profiles across microregions (87 microns to 286 microns wide) of tumors (R3230AC mammary adenocarcinoma) in a rat dorsal skin flap preparation and mapped adjacent microvessels. By comparing measured PO2 values with theoretical simulations, we deduced local consumption rates. Results for six profiles ranged from 0.83 to 2.22 cm3 O2/100 g/min. The mean (+/- SD) was 1.52 +/- 0.51 cm3 O2/100 g/min. This technique permits investigation of variations in consumption at a microregional level.

Determination of continuous atracurium infusion rate in dogs undergoing whole-body hyperthermia.

Infusion rates for atracurium were calculated from multiple bolus injection data for normothermic (38 degrees C; n = 4) and hyperthermic (42 degrees C; n = 14) dogs anesthetized with thiopental and oxymorphone while undergoing whole-body hyperthermia treatment. The calculated infusion rate for atracurium at 38 degrees C was 6.2 +/- 0.3 micrograms/kg/min and the calculated infusion rate at 42 degrees C was 8.5 +/- 0.4 micrograms/kg/min. Infusion of atracurium at the calculated infusion rate of 8.5 micrograms/kg/min produced an estimated 90-100% neuromuscular blockade during heating from 38-42 degrees C and at 42 degrees C. Following discontinuation of the infusion and cooling to 38 degrees C, neuromuscular function returned to normal within 20 min with no evidence of recurarization. Atracurium infusion rates appear to be linear and related to body temperature from 26-42 degrees C. Clinically useful neuromuscular blockade in dogs may be obtained during whole-body hyperthermia by utilizing the 42 degrees C atracurium infusion rate throughout the 38-42 degrees C heating phase.

Increased microvascular permeability contributes to preferential accumulation of Stealth liposomes in tumor tissue.

Stealth liposomes have recently emerged as a promising antitumor drug delivery system, yet no studies have been reported to examine their dynamic behavior at the microcirculatory level. In this investigation, we have used in vivo fluorescence videomicroscopy to study the decay in plasma concentration and the interstitial accumulation of Stealth and conventional liposomes in tumor and granulating tissue microcirculatory preparations. Fluorescently labeled Stealth or conventional liposomes were injected i.v. into rats bearing dorsal skinflap window chambers, some of which contained a vascularized mammary adenocarcinoma. After injection, fluorescent light intensities arising from liposomes within blood vessels and the interstitium were measured over time. These measurements were used to derive plasma pharmacokinetics and vascular permeability coefficients for each liposome species in both tumor and granulating normal tissues. Within the first 90 min after injection, Stealth liposome accumulation in the tumor interstitium was 3-4-fold that for conventional liposomes. The percentage of administered liposomes remaining in the circulation at the end of 90 min was 60.2% for Stealth and 20.4% for conventional liposomes. Tumor vascular permeability was 3.42 +/- 0.78 x 10(-7)cm/s for Stealth and 1.75 x 0.38 x 10(-7)cm/s for conventional liposomes. In normal granulating tissues permeability for the 2 constructs was equivalent at 0.8-0.9 x 10(-7)cm/s. In conclusion, preferential accumulation of Stealth liposomes in tumors was attributable to a combination of slower plasma clearance and higher vascular permeability relative to conventional liposomes. Our method of combining in vivo microscopy with a tumor microcirculatory model provides a unique approach to study quantitatively the delivery of liposomes to tumor tissues, since it can be used to study the process in real time at the microcirculatory level.

Correlation between initial and long-term responses of spontaneous pet animal tumors to heat and radiation or radiation alone.

Most early-phase testing of new therapeutic modalities involves analysis of initial tumor response as opposed to estimation of long-term response. In this study, the validity of initial response rates to predict long-term responses was examined for tumors treated with radiotherapy alone compared with heat combined with radiotherapy. A total of 130 pet animals with either squamous cell carcinomas, melanomas, fibrosarcomas, mammary adenocarcinomas, or mast cell sarcomas were randomized to receive either radiation alone (XRT) or heat + radiation (delta + XRT). Responses to treatment were evaluated by response rates and response duration. The complete response (CR) rates were consistently higher for delta + XRT than for XRT across different histology groups. The combined therapy led to prolonged tumor response in all histological subgroups except melanomas, which had a longer response duration when treated with XRT alone (p = 0.043). This was in spite of a relatively high CR rate in that group (100% versus 12.5% for delta + XRT and XRT, respectively). In contrast, while no significant improvement in CR rate was observed for dermal squamous cell carcinomas treated with delta + XRT (XRT = 52.9%; delta + XRT = 68.8%), a significant improvement in response duration was noted (p = 0.002). These are two examples where CR rate did not predict long-term response. When all histological subgroups were combined (except melanomas), the CR rate was higher (p less than 0.001), and response duration was prolonged (p = 0.031) for delta + XRT compared to XRT alone.

Enhanced delivery of a monoclonal antibody F(ab')2 fragment to subcutaneous human glioma xenografts using local hyperthermia.

The purpose of this study was to investigate the effects of tumor-localized hyperthermia at 42 degrees C on the tissue distribution of radioiodinated monoclonal antibody F(ab')2 fragments. Paired-label biodistribution measurements were performed in athymic mice bearing D-54 MG human glioma xenografts on one leg. Mice received both the 131I-labeled F(ab')2 fragment of Mel-14, reactive with human gliomas and melanomas, and nonspecific 125I-labeled RPC 5 F(ab')2. Tumor-bearing legs were placed in a 42 degrees C water bath or a 37 degrees C water bath (control) for 2 or 4 h. In mice sacrificed immediately after 2 h of heating, no hyperthermia-induced differences in the distribution of either fragment were observed. In the 4-h groups, tumor uptake of Mel-14 F(ab')2 increased from 7.04 +/- 1.59% injected dose (ID)/g at 37 degrees C to 20.65 +/- 4.53% ID/g at 42 degrees C (P less than 0.0001), and tumor localization of the control fragment rose from 5.23 +/- 1.35% ID/g to 14.51 +/- 1.37% ID/g (P less than 0.0001). In another experiment, F(ab')2 fragments were injected, tumors were heated for 4 h, and groups were sacrificed at 4, 8, and 16 h after injection. Statistically significant 2- to 3-fold higher uptake of both fragments in tumor were observed at all time points. Hyperthermic conditions also resulted in higher tumor:tissue ratios for both fragments. These results suggest that it may be possible to use tumor-localized hyperthermia to increase the therapeutic utility of radiolabeled monoclonal antibodies, particularly when labeled with short lived nuclides such as the 7.2-h alpha-emitter 211At.

Targeting a genetically engineered elastin-like polypeptide to solid tumors by local hyperthermia.

Elastin-like polypeptides (ELPs) are biopolymers of the pentapeptide repeat Val-Pro-Gly-Xaa-Gly that undergo an inverse temperature phase transition. They are soluble in aqueous solutions below their transition temperature (T1) but hydrophobically collapse and aggregate at temperatures greater than T1. We hypothesized that ELPs conjugated to drugs would enable thermally targeted drug delivery to solid tumors if their T1 were between body temperature and the temperature in a locally heated region. To test this hypothesis, we synthesized a thermally responsive ELP with a T1 of 41 degrees C and a thermally unresponsive control ELP in Escherichia coli using recombinant DNA techniques. In vivo fluorescence videomicroscopy and radiolabel distribution studies of ELP delivery to human tumors (SKOV-3 ovarian carcinoma and D-54MG glioma) implanted in nude mice demonstrated that hyperthermic targeting of the thermally responsive ELP for 1 h provides a approximately 2-fold increase in tumor localization compared to the same polypeptide without hyperthermia. We observed aggregates of the thermally responsive ELP by fluorescence videomicroscopy within the heated tumor microvasculature but not in control experiments, which demonstrates that the phase transition of the thermally responsive ELP carrier can be engineered to occur in vivo at a specified temperature. By exploiting the phase transition-induced aggregation of these polypeptides, this method provides a new way to thermally target polymer-drug conjugates to solid tumors.

Heat-induced gene expression as a novel targeted cancer gene therapy strategy.

One of the main advantages of gene therapy over traditional therapy is the potential to target the expression of therapeutic genes in desired cells or tissues. To achieve targeted gene expression, we experimented with a new approach based on the long-established phenomenon of the heat shock response. By using the green fluorescence protein as a reporter gene, it was demonstrated that expression of a heterologous gene with a heat shock protein 70 promoter could be elevated to 500-1000-fold over background by moderate hyperthermia (39 degrees C to 43 degrees C) in tissue cultured cells. The heat-induced green fluorescence protein expression was first detectable at 3 h after heating and reached a maximum at 18-24 h. The expression dropped back to baseline within 72 h. In addition, when cells were infected with adenovirus vectors containing the heat-inducible interleukin 12 or tumor necrosis factor alpha genes and then heated (42 degrees C, 30 min), expression was at least 13,600- or 6.8 x 10(5)-fold over background, respectively. Intralesion injection of the interleukin-12-carrying adenovirus vector in a mouse melanoma tumor model caused significant tumor growth delay only with hyperthermia treatment. Our results therefore support heat-induced gene expression as a feasible approach for targeted cancer gene therapy.