D2000 Nature America, Inc. 0929-1903/00/$15.00/+0 www.nature.com/cgt

Heat-directed gene targeting of adenoviral vectors to tumor cells Anthony M. Brade,1,2 Duc Ngo,3 Paul Szmitko,3 Pei-Xiang Li,3 Fei-Fei Liu,1,2 and Henry J. Klamut3

1Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada M5G 2M9; 2Department of Radiation Oncology, University of Toronto, Toronto, Ontario, Canada M5G 2M9; and 3Division of Experimental Therapeutics, Ontario Cancer Institute, University Health Network, Toronto, Ontario, Canada M5G 2M9.

Targeting therapeutic gene expression to tumor cells represents a major challenge for cancer gene therapy. The strong transcriptional response exhibited by heat shock genes, along with the beneficial therapeutic effects of hyperthermia have led us to develop a heat- directed gene-targeting strategy for cancer treatment. Heat shock gene expression is mediated in large part by the interaction of heat shock factor 1 with specific binding sites (heat shock elements;HSE) found in the promoters of heat-inducible genes. Here we present a quantitative analysis of heat-inducible gene expression mediated by the wild-type hsp70b gene promoter, as well as a modified hsp70b promoter containing additional HSE sequences. -Galactosidase ( -gal) expression was induced between 50- and 800-fold in a panel of human breast cancer cell lines infected with an adenoviral vector containing the wild-type hsp70b promoter (Ad.70b. g) following treatment at 438C for 30 minutes. Infection with an adenoviral vector containing the modified hsp70b promoter (Ad.HSE.70b. g) resulted in a 200- to 950-fold increase in -gal expression under the same conditions, and also provided a 1±28C decrease in the threshold of activation. Significant increases in the heat responsiveness of the Ad.HSE.70b. g construct were observed in five of six tumor cell lines tested, as well as under thermotolerant conditions. Finally, we demonstrate that localized heating of a HeLa cell xenograft can effectively target -gal gene expression following intratumoral injection of Ad.70b. g. Adenoviral vectors incorporating heat-inducible therapeutic genes may provide useful adjuncts for clinical hyperthermia. Cancer Gene Therapy (2000) 7, 1566±1574

Key words: HSE; adenoviral vectors; -galactosidase.

ene therapy has the potential to have a significant native strategy is to use existing antineoplastic modalities to Gimpact on cancer treatment. However, the inability of target gene expression to the tumor volume. For example, current gene transfer vectors to target tumor cells directly radiation-responsive promoters that provide a 3- to 4-fold limits the usefulness of this approach due to normal tissue induction in gene expression have been used successfully in toxicity. One solution to this problem is to utilize genes that preclinical studies to enhance the effects of radiation by are less toxic but can sensitize tumor cells to existing simultaneously driving the expression of a toxic gene in the antineoplastic modalities. For example, the synergistic effect radiation field.7,8 Our laboratory is interested in examining of wild-type p53 gene transfer combined with radiation the possibility of targeting therapeutic gene expression to treatment has already been established in preclinical studies,1 tumor cells using a heat-inducible gene transfer vector. and similar strategies are currently being evaluated in clinical Hyperthermia uses a variety of techniques (e.g., ultra- trials.2 A second solution is to specifically target the sound, radio frequency or microwave radiation) to raise the expression of therapeutic genes to tumor cells. A number temperature of a localized tumor, typically by 5±78C. In of tissue- or tumor-specific promoters have already been preclinical studies, hyperthermia was shown to act as a tested in a variety of models with promising results. These sensitizing agent for radiation and chemotherapy.9 Several include the -fetoprotein promoter in hepatocellular carci- randomized clinical trials have shown that the combination noma,3 the tyrosinase promoter in melanoma,4 the CEA of hyperthermia and radiation therapy achieves a superior promoter in CEA positive lung carcinoma,5 and the stress- response rate over radiation alone in a number of tumor sites, inducible grp78/BiP promoter in fibrosarcoma.6 An alter- including breast carcinoma.10 Widespread acceptance of hyperthermia as a standard cancer treatment has been hampered, however, by difficulties in achieving a homo- Received June 7, 2000; accepted October 11, 2000. geneous temperature dose throughout the target tumor. Address correspondence and reprint requests to Dr. Henry J. Klamut, Temperature inhomogeneity has been shown to impact on Division of Experimental Therapeutics, Ontario Cancer Institute, Uni- versity Health Network, 610 University Avenue, Room 10-721, Toronto, outcome and this is due to technical challenges as well as the 11,12 Ontario, Canada M5G 2M9. E-mail address: [email protected] heat-sink effects caused by the tumor vasculature.

1566 Cancer Gene Therapy, Vol 7, No 12, 2000: pp 1566±1574 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING 1567

The temperatures reached in hyperthermia treatment maintained in -MEM but supplemented with 20% FBS. activate the heat shock response in cells. Certain members SK.Br-3 cells (ATCC) were maintained in McCoy's (5A) of the heat shock gene family are quiescent in resting cells medium plus 10% FBS. T47D cells (ATCC) were cultured but strongly induced following heat damage. This expression in RPMI 1640 medium plus 10% FBS. MDA-MB-231 cells of heat shock proteins allows cells to repair damage and (ATCC) were grown in Leibowitz (L15) medium plus 10% provides transient protection from further heat stress, a FBS. The culture medium was supplemented with 100 U/ phenomenon known as thermotolerance. Heat shock gene mL penicillin and 100 g/mL streptomycin and cells were expression is mediated in large part at the transcriptional grown at 378Cin5%CO2. Experiments were conducted level through the interaction of activated trimers of heat using monolayers of cells at 70±80% confluency. For shock factor 1 (HSF-1) with specific DNA sequences hyperthermia experiments, flasks (Nunc, Rosilde, Denmark) known as heat shock elements (HSE) found in the promoters were sealed and cells heat shocked by immersion in a of heat-inducible genes.13 Incorporation of a heat shock calibrated water bath (Grant Inst., Cambridge, UK). The gene promoter into a gene transfer vector has the potential to desired temperature was maintained at ‹0.058C for the time provide a high level of control over therapeutic gene stipulated. Following heating, flasks were immediately expression in a locally heated tumor. returned to the 378C incubator under the conditions We propose that limitations associated with temperature described above. inhomogeneity and normal tissue toxicity can be addressed by combining hyperthermia treatment with a heat-inducible Adenoviral vector construction gene transfer vector. Heat-directed expression of a ther- Figure 1 depicts the adenoviral constructs schematically. apeutic gene with a substantial bystander effect14,15 could Briefly, a nuclear localizing form of the Escherichia coli significantly enhance the therapeutic benefit of hyperthermia -gal gene flanked 30 by SV40 polyadenylation sequences although at the same time limit nonspecific expression in was cloned into the adenoviral shuttle vector pÁE1sp1A other tissues. This concept has been demonstrated through (Microbix) as a HindIII/BamHI fragment (designated heat-inducible expression of cytosine deaminase and pNLS- gal/sp1A). A 461-bp BamHI/HindIII fragment of thymidine kinase genes in prostate cancer cells.16 Recently, the human hsp70b gene promoter containing most of the 50 heat-directed targeting of GFP17,18 and therapeutic genes untranslated region ( À352to +114 bp) was subcloned such as interleukin-2,19 interleukin-12, and tumor necrosis from plasmid pSPV-120 (Stressgen, Victoria, British factor 18 has shown good success in glioma and melanoma Columbia, Canada) into the adenoviral shuttle vector tumor models. Further evaluation of the clinical utility of this pÁE1sp1B. Subsequently, the promoter was excised from approach will require a detailed characterization of the this plasmid as a BglII/SalI fragment and subcloned expression of heat-inducible promoters in the context of specific gene transfer vectors and target tumor cells. Our laboratory is interested in developing a heat-directed gene- targeting strategy for breast and cervical cancers using adenoviral vectors. In this study, we describe the construc- tion and characterization of recombinant adenoviral vectors in which the expression of a -galactosidase ( -gal) reporter gene is under the control of either the wild-type human hsp70b promoter or a modified hsp70b promoter to which extra HSE sequences have been appended. The results demonstrate that the hsp70b promoter exhibits low levels of basal activity in the context of an adenoviral vector, and can be induced significantly in response to heat treatment of Figure 1. Schematic representation of recombinant adenoviruses human breast and cervical cancer models in vitro and in vivo. Ad.70b. g and Ad.HSE.70b. g. Recombinant adenoviruses contain- Furthermore, the addition of HSE sequences was seen to ing either the wild-type hsp70b promoter (Ad.70b. g)or the HSE- modified promoter (Ad.HSE.70b. g)were generated by homologous increase gene expression in both heated and thermotolerant recombination in 293 cells as described in Materials and Methods. cells. Ad.70b. g contains a nuclear localizing form of the E. coli - galactosidase gene (NLS-LacZ)flanked 3 0 by SV40 polyadenylation sequences cloned downstream of a 461-bp fragment corresponding MATERIALS AND METHODS to the human hsp70b gene promoter (Hsp70b). Ad.HSE.70b. gis identical to Ad.70b. g except that it contains an additional 65 bp of Cell lines and hyperthermia sequence corresponding to three consensus binding sequences for MCF-7, MDA-MB-468 (ATCC) and 293 cells (Microbix, the heat shock transcription factor 1 (HSF-1)cloned upstream of the Toronto, Ontario, Canada) were maintained in -minimal hsp70b promoter (HSEÂ3). Expression cassettes substitute for deleted sequences in the E1 region of the adenoviral genome essential medium ( -MEM) containing 10% vol/vol fetal (nucleotides 342±3523)and are oriented in the opposite direction bovine serum (FBS). BT-20 cells (ATCC) were also relative to wild-type E1 gene transcription.

Cancer Gene Therapy, Vol 7, No 12, 2000 1568 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING upstream of the -gal gene into NLS- gal/sp1A. This 0.02% NP-40), stained for 3 hours at 378C with a solution vector (p70b g) was then modified by cloning a blunted containing 1 mg/ml X-gal (Sigma, St. Louis, MO), 5 mM

65-mer oligonucleotide containing three consensus bind- K3Fe(CN)6,5mMK4Fe(CN)6, 2mM MgCl 2, washed three ing sequences for the heat shock transcription factor 1 times with PBS -Ca±Mg, three times with water, and air dried. (HSF-1) into the blunted BglII site at the 50 end of the -Gal enzymatic activity was determined using the Galacto- hsp70b promoter. The sequence of this oligonucleotide Star chemiluminescent -gal assay system (Tropix, Bed- (50AGGCCTCGAGGTTCTAAGAACGTTCTAAGAACG- ford, MA) following the manufacturers directions for TAAGAACGTTCTAAGAACGTTCTAAGAACGTTCTG- mammalian cells. Chemiluminescence was measured using TCGACTAGT30 ) was based on data from work by Cunniff a Berthold (Germany) Lumat LB 9507 tube luminometer and Morgan20 and Wang and Morgan.21 The plasmid and quantitated against a -gal (Sigma) standard curve. bearing the modified promoter was dubbed pHSE70b g. Protein assays were carried out using the detergent Recombinant replication-defective (E1/E3 deleted) ade- compatible protein assay kit (Bio-Rad, Mississauga, ON, noviruses (Ad.70b. g and Ad.HSE.70b. g) were generated Canada) modified with iodoacetamide (Sigma) to correct for by homologous recombination of the aforementioned shuttle the presence of DTT and detergent in the lysis buffer (as per vectors with the E1/E3 deleted adenoviral genomic plasmid Bio-Rad US/EG bulletin 1909). Activities are expressed as pJM17 (Microbix) by cotransfection into 293 cells using the picograms -gal per microgram protein and are presented as calcium phosphate method. Purified viral plaques were the mean‹SD of at least three independent experiments. screened initially by PCR and subsequently by -gal Statistical significance for comparisons between two mea- staining/assay (see below) to identify recombinant clones. surements was determined by the unpaired two-tailed Propagation and titer of recombinant adenoviruses was Student's t test. Values of P<.05 were considered significant. performed using standard methods.22 Titers for these two viruses were obtained by serial dilution in parallel and HeLa cell/SCID mouse hyperthermia model repeated three times to ensure accuracy of experiments HeLa cells (5Â105 in 20 L -MEM+10% FBS total comparing the two in a head-to-head fashion. The absence volume) were injected subcutaneously (s.c.) in the dorsum of replication competent adenovirus was confirmed by of a rear hind foot pads of SCID mice, in accordance with a negative plaque assays on HeLa and MCF-7 cells. protocol approved by the animal ethics committee at the Ontario Cancer Institute. Using this method, tumors form In vitro adenoviral infection within 4±6 weeks in approximately 75% of the animals. All in vitro infections with the exception of the thermo- Once the tumors reached 125 mm3, they were injected tolerance experiment (see below) were carried out as with three single, intratumoral doses of 1Â108 pfu (in 10 follows: cesium gradient purified recombinant adenovirus L of 10 mM Tris pH 8.0) of Ad.70b. g on three stock was diluted in a total volume of 500 L of serum-free successive days. Twenty-four hours following the last -MEM/well. The diluted adenoviral stock was added injection the mice were anaesthetized with Innuvar-Vet IP directly to each well of a six-well dish containing the desired (0.08 mL of a 1:100 dilution) and ketamine IM (0.08 mL of cell line in 2.5 mL of -MEM+10% FBS. This was a 20 mg/mL stock). Mice were immobilized in a Plexiglas incubated at 378C for 24 hours at which time the dishes were chamber mounted on a pedestal in a calibrated Haake water heated as described above. Adenovirus concentrations were bath.23 The tumor-bearing foot was drawn through a precut adjusted to ensure 80±90% infection of each cell type based hole in the chamber such that only the tumor-bearing foot on previously determined infection efficiency. was submerged at either 37 or 448C. The tumor was then Thermotolerant cells were infected as follows. Thermo- treated for 30 minutes at the desired temperature. Rectal tolerance was induced by an initial exposure to 438C for 30 temperature was monitored throughout treatment. Twenty- minutes. At various time points thereafter, cells were four hours later, the mice were sacrificed, and the tumors reheated to 438C or control treated at 378C for 30 minutes. and livers were immediately frozen in Tissue Tek O.T.C. One hour before reheating, cells were infected with the compound (VWR Can-Lab, Toronto, ON, Canada) and specified amount of adenovirus diluted in 300 L of serum- stored at À708C. Six serial sections (5 m) were taken at free -MEM per well of a six-well dish at 378C. each of five random levels throughout the tissues sampled. Immediately before reheating, the virus-containing medium The sections were fixed and stained for -gal activity as was aspirated and replaced with fresh prewarmed media. described above. Sections were counterstained with Neutral Red (0.1%), dried and examined by light microscopy. -Gal enzymatic activity and staining To stain for -gal activity, cells were first washed three times -Ca±Mg with PBS (150 mM NaCl, 15 mM Na phosphate, pH RESULTS 7.3) then fixed with 2% formaldehyde, 0.2% glutaraldehyde in PBS -Ca±Mg for 5 minutes at 48C. After fixation, cells were Induction of -gal activity following heating of MCF-7 cells -Ca±Mg washed three times with PBS ( +2mM MgCl 2, infected with Ad.70b. g. Recombinant adenoviruses con-

Cancer Gene Therapy, Vol 7, No 12, 2000 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING 1569

100 458C for 30 minutes). Promoter activity, expressed as fold-induction relative to control values measured at 378C, 45°C is presented in Figure 3. The addition of synthetic HSE 10 44°C 43°C sequences to the wild-type hsp70b promoter resulted in 42°C significant increases (P<.05) in the fold-induction of 41°C reporter gene expression levels in cells heated to 428C and -Gal Activity β 1 40°C above. The increase in -gal activity and fold-induction 37°C was particularly evident at higher temperatures, reaching levels that were up to 10-fold higher than the maximum 0.1 0 3 6 9 12 15 18 21 24 expression levels observed with the hsp70b promoter alone Time (Hours) (200- vs. 2000-fold basal expression at 458C). At 438C Figure 2. Heat-responsiveness of gene expression from Ad.70b. g. reporter gene expression from the HSE-modified promoter MCF-7 cells were infected with Ad.70b. g (10 pfu/cell)and after 24 was 3.5-fold greater than the wild-type hsp70b promoter hours cells were heated for 30 minutes at 37, 40, 41, 42, 43, 44, or (see also Fig 4). These results demonstrate that the activity 458C. Cell lysates prepared at 0, 3, 6, 12, and 24 hours postheat treatment were assayed for both -gal enzymatic activity and total of the hsp70b promoter at clinically achievable tempera- protein levels as described in Materials and Methods. Promoter tures can be enhanced by the addition of synthetic HSE. activities, expressed as picograms -gal per microgram protein, represent the mean‹SD of a minimum of three independent Variable levels of induction are observed in different tumor experiments. cell lines following heat treatment To evaluate the generality of the observations made in taining either the wild-type hsp70b promoter (Ad.70b. g) MCF-7 cells, these experiments were repeated with five or the HSE-modified promoter (Ad.HSE.70b. g) were additional human breast cancer cells lines. In these generated by homologous recombination between the shuttle experiments, cells were infected with either Ad.70b. gor vector and the pJM17 plasmid following cotransfection into the modified Ad.HSE.70b. g vector at MOIs that resulted 293 cells (Fig 1). To examine the heat-responsiveness of in infection efficiency of 80±90%. The cells were then gene expression from the Ad.70b. g construct, MCF-7 cells heated to 438C for 30 minutes and -gal activity was were infected with a multiplicity of infection (MOI) of 10 measured the following day. As shown in Figure 4a, pfu/cell (transduction efficiency of 80±90%) and -gal induction of reporter gene expression following heat activity were assessed at various time points following heat treatment was observed in all cell lines. However, some treatment. Cells were heated at temperatures ranging from 40 to 458C for 30 minutes. As shown in Figure 2, a temperature- 10000 dependent increase in -gal activity was observed for all treatment temperatures 418C. The maximum expression rate observed occurred between 0 and 3 hours postheating for 1000 Ad.HSE.70b.βg temperatures in the 41±448C range and between 3 and 6 Ad.70b.βg hours posttreatment in cells heated to 458C. Fold induction of

-Gal Activity 100 -gal activity reached near-maximal levels by 6 hours but β peaked at 12hours postheating for all groups tested. In the clinical target range of hyperthermic treatment (42±448C), 10 -gal activity was induced 10- to 100-fold above basal levels. Induction was highest at 458C, reaching a maximum 1 of 200-fold over basal levels. A 2-fold increase was Fold Induction of 40 41 42 43 44 45 Temperature (°C) observed following heating to 418C, and no induction was 0.1 evident at 408C. These results demonstrate that the hsp70b Figure 3. Additional HSE elements increase the responsiveness of promoter can provide high levels of induction of gene the hsp70 promoter to heat treatment. MCF-7 cells were infected with expression in MCF-7 cells in the target temperature range for either Ad.70b. g or Ad.HSE.70b. g (10 pfu/cell)and after 24 hours clinical hyperthermia. cells were heated for 30 minutes at 37, 40, 41, 42, 43, 44, or 458C. Cell lysates prepared at 0, 3, 6, 12, and 24 hours postheat treatment The addition of synthetic Heat Shock Elements can increase were assayed for both -gal enzymatic activity and total protein levels as described in Materials and Methods. Promoter activities (pg heat responsiveness in the clinical target temperature range -gal/g protein)in cells heated to 40±45 8C are expressed as the To determine whether the addition of synthetic HSE could fold-induction relative to control values measured in cells kept at increase the responsiveness of the hsp70b promoter to heat 378C. Solid bars: MCF-7 cells infected with Ad.HSE.70b. g. Open bars: MCF-7 cells infected with Ad.70b. g. Values represent the treatment, MCF-7 cells were infected with either mean‹SD of a minimum of five independent experiments. Differ- Ad.70b. g or the modified Ad.HSE.70b. g vector and ences in fold-induction were statistically significant (P<.05)in cells -gal activity was measured 24 hours postheating (40 to heated to 428C and above.

Cancer Gene Therapy, Vol 7, No 12, 2000 1570 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING

1000 a) Ad.70b.βg - 37°C MB-231 cells (Fig 4a). The benefit provided by the Ad.70b.βg - 43°C additional HSE elements, expressed as an enhancement β ° 100 Ad.HSE.70b. g - 37 C ratio (Ad.HSE.70b. g induced -gal activity/Ad.70b. g Ad.HSE.70b.βg - 43°C induced -gal activity), is presented in Figure 4b. Levels 10 of heat induction from the modified hsp70b promoter

1 were significantly (P<.05) greater than the wild-type promoter in five of the six cell lines studied, providing -Gal Activity β 0.1 corroborative evidence that the addition of HSE elements can enhance the heat response over that seen with the 0.01 hsp70b promoter alone. Collectively, the results presented in Figures 2±4 suggest that heat-directed gene targeting 0.001 can provide high levels of induction of a therapeutic 468 231 T47D SKBr.3 BT20 Cell Line transgene in human tumors. Heat-directed gene expression is diminished but not 6.5 eliminated in thermotolerant cells 6 b)

5.5 Following heat exposure, cells that survive manifest

5 thermotolerance, a transient resistance to subsequent heat g)

β 4.5 treatments. If heat-inducible vectors are to be used in 4 conjunction with hyperthermic treatment, an understanding 3.5 g:Ad.70b. of the impact of thermotolerance on gene expression from β 3 these vectors is essential. This issue was addressed by 2.5 measuring the inducibility of the modified and unmodified 2

(Ad.HSE.70b. hsp70b promoters in MCF-7 cells at various time points Induction Enhancement Ratio 1.5 following the induction of thermotolerance. Thermotoler- 1 468 231 T47D SKBr.3 BT20 MCF-7 0.5 ance was induced by heating MCF-7 cells to 438C for 30 Cell Line 0 minutes. Thermotolerant cells were then infected with Figure 4. Basal and induced hsp70b promoter activities in different human breast tumor cell lines. A panel of human breast cancer cell lines were infected with Ad.70b. g or Ad.HSE.70b. g at a MOI that 100 provided equivalent levels of -gal staining (80±90% positive)in Ad.70b.βg - 37°C control cultures infected with an adenoviral vector expressing the - β ° gal gene from a CMV promoter. After 24 hours cells were incubated Ad.70b. g - 43 C Ad.HSE.70b.βg - 37°C for 30 minutes at either 37 or 438C. Cells lysates were prepared 24 10 hours postheat treatment and assayed for both -gal activity and total Ad.HSE.70b.βg - 43°C protein. Promoter activities were expressed as pg -gal/g protein. a: The results of a representative experiment showing the relative levels of basal (378C)and induced (43 8C)hsp70b promoter activity -Gal Activity β 1 following infection with either Ad.70b. g (HSP)or Ad.HSE.70b. g (HRE). b: Induction enhancement conferred by the presence of additional HSE elements upstream of the hsp70b promoter. The induction enhancement ratio was calculated by dividing the fold 0.1 induction in -gal activity (from 37 to 438C)provided by Ad.- 8h 14h 30h Control HSE.70b. g by that provided by Ad.70b. g. Values represent the Hours post induction of TT mean‹SD of a minimum of three independent experiments in each cell line tested. The increase in fold-induction provided by Ad.- Figure 5. Effects of thermotolerance on hsp70b promoter induction HSE.70b. g was statistically significant ( P<.05)in all cell lines with following heat treatment. MCF-7 cells were heated to 438C for 30 the exception of MDA-MB-231 cells. minutes to induce thermotolerance. Cells were subsequently infected with either Ad.70b. g or Ad.HSE.70b. g (MOI=10)and again subjected to heat treatment (37 or 438C for 30 minutes)at 8, 14, or variability was observed in both the absolute levels of - 30 hours following the induction of thermotolerance. Control cultures were not subjected to the initial 438C heat dose but otherwise were gal expressed at both 37 and 438C and the maximum levels treated in the same manner. Cells lysates were prepared after a of induction in different cell lines. For example, levels of further 24-hour incubation at 378C and assayed for -gal activity and reporter gene induction following Ad.70b. g infection total protein levels. Promoter activities measured in cells infected with ranged from 55-fold in MCF-7 cells to 800-fold in either Ad.70b. g or Ad.HSE.70b. g are expressed as picograms - MDA-MB-231 cells following heating to 438C. Similarly, gal per microgram protein and represent the mean‹SD of a minimum infection of cells with the modified Ad.HSE.70b. g of three independent experiments at each time point. Differences in promoter activities in cells infected with either Ad.HSE.70b. gor construct resulted in maximum induction levels ranging Ad.70b. g and heated to 438C were statistically significant ( P<.05) from 200-fold in MCF-7 cells to 950-fold in MDA- at all time points.

Cancer Gene Therapy, Vol 7, No 12, 2000 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING 1571

Ad.70b. g or Ad.HSE.70b. g 1 hour before reheating. At 8, thermotolerance, which increased over time as thermotoler- 14, and 30 hours postinduction of thermotolerance, cells ance decayed. In addition, -gal gene expression levels from were reheated to 438C for 30 minutes to stimulate -gal the Ad.HSE.70b. g vector were significantly (P<.05) expression, and -gal activity was measured 24 hours later higher at all time points relative to Ad.70b. g. These as in previous experiments. Induction of thermotolerance observations are consistent with the activity of this vector in had no effect on infectivity of MCF-7 cells at the MOI levels nonthermotolerant cells and further emphasize the enhanced used for these experiments (data not shown). As shown in activity of the modified promoter. Figure 5, levels of heat-inducible gene expression were decreased in thermotolerant cells compared to control- Heat-directed gene targeting of an adenoviral vector in a treated cells at all time points studied. However, thermo- HeLa xenograft model tolerant cells infected with either Ad.70b. gor To examine whether this therapeutic strategy could be Ad.HSE.70b. g still expressed significant levels of -gal extended to a tumor model, HeLa cell tumors were grown activity as early as 8 hours following the induction of s.c. in the hind limb footpads of SCID mice. The footpad

Figure 6. Heat-directed gene targeting to a xenograft tumor model. HeLa cell tumors were grown s.c. in the hind limb footpad of SCID mice. When a tumor volume of 125 mm3 was reached, each tumor was injected with Ad.70b. g(1Â108 pfu)once daily for 3 days. After 24 hours injected tumors were heated to either 37 or 448C for 30 minutes as described in Materials and Methods, and 24 hours later animals were sacrificed and tumors and livers removed, frozen, and sectioned. Individual sections were stained for -gal activity and counter-stained with Neutral Red. Representative sections are shown for tumors injected with Ad.70b. g and heated to 378C (panels a and c)or 44 8C (panel b and d). Panels c and d are higher magnifications of boxed areas indicated in panels a and b. Under these experimental conditions -gal staining was patchy and consistently represented 10±15% of sections taken from tumors heated to 448C. Also shown is a mock-infected tumor heated to 448C (panel e) and a liver section from an animal injected with Ad.70b. g and heated to 448C (panel f). Bars=100 m in panels a, b, e, and f;20m in panels c and d.

Cancer Gene Therapy, Vol 7, No 12, 2000 1572 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING model for tumor growth was chosen because it provides a Our data demonstrate that the hsp70b promoter can drive relatively straightforward model for tumor heating and the high-level transgene expression in the context of an maximum intratumoral temperature that can be reached is adenoviral vector while maintaining relatively low levels well defined. Several human breast and cervical tumor cell of basal expression. Kinetic studies in MCF-7 cells lines were tested for both adenoviral infection efficiency and following heat shock suggest that gene expression is their growth characteristics in the footpad model. The HeLa activated within 30 minutes, peaks between 3 and 6 hours, cell line derived from a human adenocarcinoma of the cervix and returns to baseline levels by 12to 24hours. This result is was chosen because of its high adenoviral transduction in keeping with previous studies16,24 and suggests that the efficiency, ease of growth in the mouse footpad, and its well- hsp70b promoter functions normally within the context of characterized response to hyperthermia treatment. Further- the vector sequences. more, in vitro studies confirmed that HeLa cells infected with High levels of induction were also observed in five other Ad.70b. g exhibit comparable kinetics, basal levels, and human breast cancer cell lines ranging from 90- to 850-fold maximum heat-inducible levels of -gal gene expression to above basal levels following treatment for 30 minutes at MCF-7 cells (data not shown). When the HeLa cell tumors 438C. In an attempt to enhance this response further, a reached 125 mm3, they were injected intratumorally with second adenovirus was constructed containing an additional Ad.70b. g(1Â108 pfu per dayÂ3 days). Twenty-four hours synthetic HSE cassette upstream of the hsp70b promoter. following the last injection, tumors were heated to either 37 This modification significantly increased the heat respon- or 448C for 30 minutes by water-bath immersion. A further siveness of the promoter in five of the six cell lines tested. 24 hours following heating, each mouse was sacrificed and When tested in MCF-7 cells over a range of temperatures, its tumor and liver were removed for assessment. As shown significant increases in expression were apparent at 428C and in Figure 6 (panels a and c), no appreciable -gal staining above. These results raise the possibility that the hsp70b was observed in tumors treated at 378C. In contrast, patches promoter can be further modified to decrease its activation of intense, nuclear-localized staining were observed in threshold to lower temperatures. However, care must be tumors heated to 448C at all levels sampled (panels b and d), taken to retain the minimal basal levels of activity found in and staining patterns were consistent between serial sections. all cell lines tested (ranging from one to four times Under these experimental conditions 10±15% of the total background, data not shown). This suggests that the hsp70b area of tumor sections stained positive for -gal expression. promoter (or a modified version of it) can provide a No evidence of -gal staining was observed in either mock- relatively broad therapeutic index and may be useful for infected tumors (panel e) or in the livers (panel f) of either driving expression of highly toxic genes. Radiation respon- control or heat-treated mice. These results demonstrate the sive promoters have been used in an analogous manner to feasibility of heat-directed gene expression in an in vivo drive expression of tumor necrosis factor . This approach system. has yielded promising results using clinically relevant radiation dosing.8,25 At these dose levels, induction of gene expression is in the 2- to 4-fold range 7 further emphasizing DISCUSSION the potential offered by heat-responsive promoters. Both basal and maximal levels of gene expression, as well Heat-inducible gene transfer vectors may allow therapeutic as the degree of enhancement conferred by the additional genes to be targeted to locally heated tumors, thereby HSEs, displayed some intercell line variation. Similar reducing nonspecific systemic gene expression. Similarly, differences in the response of the hsp70b promoter have hyperthermic outcomes may be improved by expressing a been noted in other cell lines.26,27 Hsp70b promoter transgene with a significant bystander effect to compensate activation is highly dependent on HSE/HSF-1 and because for areas of the tumor that are under-heated due to vascular activation of HSF-1 is highly regulated, there are many or technical factors. In a previous study,16 heat-directed gene factors that could contribute to this variability. HSF-1 expression was demonstrated in prostate cancer cells heated converts from a serine phosphorylated, inactive monomer to to 418C in vitro. The goal of this study was to examine a hyperphosphorylated, activated trimer localized to discrete quantitatively several clinically relevant parameters of heat- nuclear granules following disruption of protein homeostasis inducible gene expression using this type of vector. These (e.g., heat shock, amino acid analogues13). This multistep included threshold of induction, basal level of expression, process is regulated on several levels by various chaperones maximal induction levels, dose response to temperature, and regulatory factors, (e.g., Hsp90, Hsp70, Hdj-1, and effect of thermotolerance, and the variability of heat HSBP-1;28 ±30) and through alterations in certain oncogenes induction among different tumor cell lines. Synthetic HSE and cell signaling pathways. For example, activation of the sequences were added to the hsp70b promoter to assess MAP kinase pathway has been shown to repress HSF-1 and whether these characteristics could be improved. In addition, the heat shock response.31 ±33 Frequently mutated tumor- we examined whether this approach can be used successfully suppressor genes such as p53 and WT1 can induce gene in an in vivo setting. expression in an HSE dependent manner, although the

Cancer Gene Therapy, Vol 7, No 12, 2000 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING 1573 mechanism underlying this process is not clear.34,35 It is Therapy and the Canadian Breast Cancer Foundation (to therefore probable that the oncogenic profile of a tumor cell H.J.K.), and from the Medical Research Council of Canada line can influence basal levels of hsp70 expression as well as (to F.F.L. and H.J.K.). A. M. B. is the recipient of a Jessie fold induction following heat treatment. Although studies Davidson Medical Research Council of Canada Fellowship such as these illustrate the complexity of stress signaling and wishes to thank Drs. B. Cummings and P. Catton and the through HSF-1 leading to HSE-dependent induction of gene DRO for their continued support. expression, they also raise the possibility that heat-inducible promoters are amenable to manipulations that could improve their performance in gene therapy applications. REFERENCES Induction of thermotolerance has a significant impact on clinical hyperthermia and treatments are generally given 1. Li JH, Lax SA, Kim J, et al. The effects of combining ionizing radiation only once per week to allow for its decay. Understanding and adenoviral p53 therapy in nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys. 1999;43:607±616. the impact of thermotolerance on gene therapy strategies 2. Roth JA, Swisher SG, Meyn RE. p53 Tumor suppressor gene therapy for involving hyperthermia is therefore, of critical importance. cancer. Oncology (Huntington's). 1999;13:148±154. Hsp70a gene expression is attenuated under thermotole- 3. Kanai F, Shiratori Y, Yoshida Y, et al. Gene therapy for alpha- rance36 and our data suggest that the hsp70b promoter is fetoprotein±producing human hepatoma cells by adenovirus-mediated also less responsive to heat in thermotolerant cells. The transfer of the herpes simplex virus thymidine kinase gene. Hepatology. addition of HSE sequences was also shown to result in 1996;23:1359±1368. 4. Vile RG, Hart IR. In vitro and in vivo targeting of gene expression to enhanced gene expression under thermotolerant conditions melanoma cells. Cancer Res. 1993;53:962±967. suggesting that additional manipulation of the hsp70b 5. Osaki T, Tanio Y, Tachibana I, et al. Gene therapy for carcinoem- promoter sequence may further improve therapeutic gene- bryonic antigen-producing human lung cancer cells by cell type- expression levels in thermotolerant cells. specific expression of herpes simplex virus thymidine kinase gene. Cancer Res. 1994;54:5258±5261. As with any novel approach to cancer therapy, clinical 6. Gazit G, Kane SE, Nichols P, et al. Use of the stress-inducible grp78/ application of this strategy will require extensive preclinical BiP promoter in targeting high level gene expression in fibrosarcoma in testing in in vivo tumor models. Our results demonstrate that vivo. Cancer Res. 1995;55:1660±1663. intratumoral administration of a heat-inducible adenoviral 7. Manome Y, Kunieda T, Wen PY, et al. Transgene expression in vector can induce gene expression following local tumor malignant glioma using a replication-defective adenoviral vector containing the Egr-1 promoter: activation by ionizing radiation or heating in a HeLa xenograft model. These results demon- uptake of radioactive iododeoxyuridine [see comments]. Hum Gene strate that heat-directed gene therapy may be a feasible Ther. 1998;9:1409±1417. approach to targeting gene expression to locally heated 8. Hallahan DE, Mauceri HJ, Seung LP, et al. Spatial and temporal control tumors using clinical hyperthermia. The patchy distribution of gene therapy using ionizing radiation. Nat Med. 1995;1:786±791. of gene expression can be attributed to nonuniform 9. Dewhirst MW, Prosnitz L, Thrall D, et al. Hyperthermic treatment of 37,38 malignant diseases: current status and a view toward the future. Semin distribution of the recombinant adenovirus as well as Oncol. 1997;24:616±625. 23 inhomogeneous tumor heating. 10. Vernon CC, Hand JW, Field SB, et al. Radiotherapy with or without In summary, we have demonstrated heat-targeted gene hyperthermia in the treatment of superficial localized breast cancer: expression from two novel adenoviral vectors in vitro and in results from five randomized controlled trials. International Collabora- vivo. Low levels of basal expression were observed coupled tive Hyperthermia Group [see comments]. Int J Radiat Oncol Biol Phys. 1996;35:731±744. with high levels of heat-induced expression. The observa- 11. Li X, Brown SL, Hill RP.Factors influencing the thermosensitivity of two tions were consistent in a variety of human breast cancer cell rodent tumors. Radiat Res. 1992;130:211±219. lines. Appending additional HSEs to the hsp70b promoter 12. Sherar M, Liu FF, Pintilie M, et al. Relationship between thermal dose improved inducibility and response under both naive and and outcome in thermoradiotherapy treatments for superficial recur- thermotolerant conditions. Testing of one of the vectors in a rences of breast cancer: data from a phase III trial. Int J Radiat Oncol human tumor xenograft model demonstrated proof of Biol Phys. 1997;39:371±380. 13. Morimoto RI. Regulation of the heat shock transcriptional response: principle that heat can be used to target gene expression in cross-talk between a family of heat shock factors, molecular vivo. Given the very low level of basal expression observed, chaperones, and negative regulators. Genes Dev. 1998;12:3788±3796. one can envisage construction of viruses bearing highly toxic 14. Rizk NP, Chang MY, El Kouri C, et al. The evaluation of adenoviral genes that could be targeted to tumors using currently p53-mediated bystander effect in gene therapy of cancer. Cancer Gene Ther. 1999;6:291±301. available hyperthermia technology. 15. Ishii-Morita H, Agbaria R, Mullen CA, et al. Mechanism of ``bystander effect'' killing in the herpes simplex thymidine kinase gene therapy model of cancer treatment. Gene Ther. 1997;4:244±251. ACKNOWLEDGMENTS 16. Blackburn RV, Galoforo SS, Corry PM, et al. Adenoviral-mediated transfer of a heat-inducible double suicide gene into prostate carcinoma cells. Cancer Res. 1998;58:1358±1362. We thank our colleagues, in particular M. Rauth and M. 17. Vekris A, Maurange C, Moonen C, et al. Control of transgene Sherar, for stimulating discussions. This work was supported expression using local hyperthermia in combination with a heat- in part by grants from the Foundation for Gene and Cell sensitive promoter. J Gene Med. 2000;2:89±96.

Cancer Gene Therapy, Vol 7, No 12, 2000 1574 BRADE, NGO, SZMITKO, ET AL: HEAT-DIRECTED GENE TARGETING

18. Huang Q, Hu JK, Lohr F, et al. Heat-induced gene expression as a 29. Satyal SH, Chen D, Fox SG, et al. Negative regulation of the heat novel targeted cancer gene therapy strategy. Cancer Res. 2000; shock transcriptional response by HSBP1. Genes Dev. 1998;12:1962± 60:3435±3439. 1974. 19. Gerner EW, Hersh EM, Pennington M, et al. Heat-inducible vectors for 30. Shi Y, Mosser DD, Morimoto RI. Molecular chaperones as HSF1- use in gene therapy. Int J Hyperthermia. 2000;16:171±181. specific transcriptional repressors. Genes Dev. 1998; 12:654±666. 20. Cunniff NF, Morgan WD. Analysis of heat shock element recognition 31. Chu B, Soncin F, Price BD, et al. Sequential phosphorylation by by saturation mutagenesis of the human HSP70.1 gene promoter. J Biol mitogen-activated protein kinase and glycogen synthase kinase 3 Chem. 1993;268:8317±8324. represses transcriptional activation by heat shock factor-1. J Biol 21. Wang Y, Morgan WD. Cooperative interaction of human HSF1 heat Chem. 1996;271:30847±30857. shock transcription factor with promoter DNA. Nucleic Acids Res. 32. Chu B, Zhong R, Soncin F, et al. Transcriptional activity of heat 1994;22:3113±3118. shock factor 1 at 37 degrees C is repressed through phosphorylation 22. Graham FL, Prevec L. Adenovirus-based expression vectors and on two distinct serine residues by glycogen synthase kinase 3 and recombinant vaccines [Review]. Biotechnology. 1992;20: protein kinases Calpha and Czeta. J Biol Chem. 1998;273:18640± 363±390. 18646. 23. Brown SL, Hunt JW, Hill RP. Differential thermal sensitivity of tumor 33. Kline MP, Morimoto RI. Repression of the heat shock factor 1 and normal tissue microvascular response during hyperthermia. Int J transcriptional activation domain is modulated by constitutive phos- Hyperthermia. 1992;8:501±514. phorylation. Mol Cell Biol. 1997;17:2107±2115. 24. Arai Y, Kubo T, Kobayashi K, et al. Control of delivered gene 34. Tsutsumi-Ishii Y, Tadokoro K, Hanaoka F, et al. Response of heat expression in chondrocytes using heat shock protein 70B promoter. J shock element within the human HSP70 promoter to mutated p53 Rheumatol. 1999;26:1769±1774. genes. Cell Growth Differ. 1995;6:1±8. 25. Chung TD, Mauceri HJ, Hallahan DE, et al. Tumor necrosis factor- 35. Maheswaran S, Englert C, Zheng G, et al. Inhibition of cellular alpha-based gene therapy enhances radiation cytotoxicity in human proliferation by the Wilms tumor suppressor WT1 requires association prostate cancer. Cancer Gene Ther. 1998;5:344±349. with the inducible chaperone Hsp70. Genes Dev. 1998;12:1108±1120. 26. Dressel R, Johnson JP, Gunther E. Heterogeneous patterns of 36. Theodorakis NG, Drujan D, De Maio A. Thermotolerant cells show an constitutive and heat shock induced expression of HLA- linked attenuated expression of Hsp70 after heat shock. J Biol Chem. HSP70-1 and HSP70-2heat shock genes in human melanoma cell 1999;274:12081±12086. lines. Melanoma Res. 1998;8:482±492. 37. Toloza EM, Hunt K, Swisher S, et al. In vivo cancer gene therapy with a 27. Muramatsu T, Kobayashi N, Tada H, et al. Induction of the 72- recombinant interleukin-2adenovirus vector. Cancer Gene Ther. kDa heat shock protein in human skin melanoma and squamous 1996;3:11±17. cell carcinoma cell lines. J Dermatol. 1995;22:907±912. 38. Mauceri HJ, Seung LP, Grdina WL, et al. Increased injection number 28. Zou J, Guo Y, Guettouche T, et al. Repression of heat shock enhances adenoviral genetic radiotherapy. Radiat Oncol Invest. transcription factor HSF1 activation by HSP90 (HSP90 complex) 1997;5:220±226. that forms a stress-sensitive complex with HSF1. Cell. 1998;94: 471±480.

Cancer Gene Therapy, Vol 7, No 12, 2000