Targeting of suicide gene delivery in pancreatic cancer cells via FGF receptors

Abstract

Pancreatic ductal adenocarcinomas (PDACs) overexpress various cell-surface tyrosine kinase receptors, including the type I high-affinity fibroblast growth factor receptor (FGFR-1). The purpose of this study was to determine whether FGFR-targeted gene therapy is feasible in this disorder. Accordingly, the effects of a conjugate consisting of fibroblast growth factor (FGF)-2 linked to a Fab′ fragment against the adenovirus knob region were evaluated in human pancreatic cancer cell lines treated with an adenoviral vector containing the herpes simplex virus thymidine kinase (AdTK) gene. An adenoviral vector containing the firefly luciferase reporter gene (AdLuc) served to assess infection efficiency, and was initially tested in L6 rat myoblasts. In parental L6 cells that express exceedingly low levels of high-affinity FGFRs, transduction with AdLuc was enhanced 7- to 10-fold with the FGF2–Fab′ conjugate, whereas in L6 cells transfected to express FGFR-1, it was enhanced 39- to 52-fold. The pancreatic cancer cell lines expressed variable levels of the four high-affinity FGF receptors, and exhibited 2- to 34-fold increases in gene transduction in the presence of the FGF2–Fab′ conjugate. In the absence of FGF2–Fab′ there was no correlation between surface binding of FGF2 and AdLuc transduction efficiency, whereas in the presence of FGF2–Fab′, enhanced AdLuc transduction efficiency correlated with greater surface binding of FGF2. In the absence of AdTK, all the cell lines were insensitive to ganciclovir, whereas after AdTK transduction, only ASPC-1 and PANC-1 cells were resistant to ganciclovir even in the presence of FGF2–Fab′. Ganciclovir-mediated inhibition was dependent on the conjugate in CAPAN-1 and COLO-357 cells, but was independent of the conjugate in T3M4 and MIA-PaCa-2 cells. Real-time quantitative PCR of laser-captured cancer cells revealed high levels of various FGFR mRNA species in six of seven PDAC tumor samples. These findings indicate that transduction efficiency with FGF2–Fab′ in pancreatic cancer cells is independent of native adenoviral transduction efficiency and is greatest in cells that exhibit concomitant expression of various high-affinity FGFRs. In view of the overexpression of high-affinity FGFRs in the cancer cells in PDAC, our findings also suggest that the combined use of AdTK, ganciclovir, and FGF2–Fab′ may ultimately be a promising therapeutic approach in a subgroup of patients with PDAC.

Main

Pancreatic ductal adenocarcinoma (PDAC) is a devastating disease that is the fourth to fifth leading cause of cancer-related mortality in the Western world.1,2 The 1-year survival rate of patients with PDAC is estimated at only 12%, and less than 1% of the patients are still alive 5 years after diagnosis.1,2 Due to the relatively late diagnosis, curative resection is possible only in a small subgroup of patients. Nonsurgical treatment options such as chemotherapy have been disappointing because of the resistance of pancreatic cancer cells to cytotoxic agents.3,4,5 In addition, these cancers do not respond to radiotherapy4,5,6 and are resistant to alternative treatment options such as hormonal or immunotherapy.7 There is, therefore, a crucial and urgent need for more effective treatment modalities for this disease.

Gene therapy has the potential to become a promising tool in the treatment of certain neoplastic diseases. Its implementation, however, requires the efficient and specific delivery of potentially therapeutic genes into cancer cells. Introduction of the herpes simplex virus thymidine kinase (HSV-TK) gene into tumor cells by retroviral and adenoviral vectors confers sensitivity to the viral drug ganciclovir that, at therapeutic dosages, does not exert toxic effects on human cells.8,9,10 Recent studies have examined the ability to specifically express the HSV-TK gene in the target organ using tumor-specific and/or tissue-specific promoters.11,12,13,14,15 Other studies have focused on specifically retargeting the adenovirus via the fibroblast growth factor receptor (FGFR) pathway.16,17,18,19 For example, Kaposi sarcoma (KS) cells can be efficiently transduced by a modified adenovirus to specifically target the FGFRs on the surface of KS cells.17

Pancreatic cancer cells are susceptible to transduction of adenoviral-mediated delivery of the HSV-TK gene.10,13,16,20 In addition, PDACs overexpress the two Ig-like variants of the FGFR-1 and FGFR-221,22 and multiple FGFs.23 We therefore sought to determine whether an adenoviral-based gene delivery system targeting FGF receptors may be of potential benefit in PDAC, by studying six human pancreatic cancer cell lines and by characterizing FGFR expression in PDAC samples. We now report that FGFR-targeting enhances gene transduction by 2- to 34-fold in these cells, depending, to some extent, on the expression levels of the four high-affinity FGFRs. We also show that, following HSV-TK transduction, ganciclovir exerts cytotoxic effects in two of six pancreatic cancer cell lines and that FGFR-targeted delivery of the HSV-TK gene enhances the cytotoxic effects of ganciclovir in two additional cell lines. Using laser capture microdissection and real-time quantitative PCR, we demonstrate the presence of multiple FGFRs in the cancer cells within PDAC samples. Taken together, these observations suggest that ganciclovir, in combination with FGFR-targeted delivery of AdTK, may ultimately have a therapeutic role in PDAC.

Materials and methods

Materials

The following were purchased: ASPC-1, CAPAN-1, MIA PaCa-2, and PANC-1 human pancreatic cancer cells and 293 human embryonal kidney cells from American Type Culture Collection (ATCC, Rockville, MD); FBS, DMEM and RPMI medium, trypsin solution, and penicillin–streptomycin solution from Irvine Scientific (Santa Ana, CA); [α-32P]dCTP, [α-32P]CTP, and 125I-FGF2 from Amersham (Arlington Heights, IL); lipofectamine from Life Technologies (Grand Island, NY); luciferase assay reagent and reporter lysis buffer from Promega (Madison, WI); RPA II kit from Ambion (Austin, TX); ganciclovir from Roche Laboratories (Nutley, NJ); AmpliTaq DNA polymerase and all real-time quantitative PCR reagents from Applied Biosystems (Branchburg, NJ). All other chemicals were from Sigma (St. Louis, MO). COLO-357 and T3M4 human pancreatic cancer cell lines were a gift from RS Metzgar (Duke University, Durham, NC). FGF2 was a gift from Dr J Abraham at Scios Nova (Mountainview, CA). Parental L6 cells were a gift from Dr R Bradshaw (University of California, Irvine, CA).

Ribonuclease protection assay

Total RNA was extracted by the single-step acid–guanidine–thiocyanate phenol chloroform method, as previously described.21,22 RNA (10 μg per sample) was hybridized overnight at 42°C with [α-32P]CTP-labeled riboprobes, and single-stranded RNA was subsequently digested with RNAse A/T1 according to Ambion's RNase protection assay kit RPA II. Yeast tRNA (10 μg) was used as a negative control.21 Samples were then subjected to denaturing gel electrophoresis on 6% polyacrylamide/8 M urea gels. Gels were dried and exposed to Kodak (Rochester, New York) BioMax MS films at −80°C using intensifying screens. The following riboprobes were used: nucleotides 664–1023 of FGFR-1 (Genbank accession: X51803), nucleotides 2290–2515 of FGFR-2 (Genbank accession: M97193), nucleotides 1413–1734 of FGFR-3 (Genbank accession: M58051), and nucleotides 200–387 of FGFR-4 (Genbank accession: X57205).

Real-time quantitative PCR analysis

To determine the relative mRNA expression levels of FGFR-1, -2, -3, and -4, a real-time fluorescent detection method was performed as previously described.24 Total RNA was extracted as above and reverse transcribed using the Superscript cDNA amplification kit according to manufacturers' directions. Each specific cDNA of interest and the β-actin reference cDNA were PCR-amplified using a fluorescent oligonucleotide probe with a 5′ reporter dye (6FAM) and a downstream 3′ quencher dye (TAMRA). The following oligonucleotides were used. Forward FGFR-1 primer, 5′-GGGCGGCTCCCCATAC; reverse, 5′-TGACCTCCTTCAGCAGCTT; probe 6FAM-5′-CCGGTGTGCCTGTGGAGGAACTTTT-TAMRA. Forward FGFR-2 primer, 5′-CCCCTGGGAGAAGGTTGC; reverse, 5′-TGTCTTTGTCAATTCCCACTGC; probe 6FAM-5′-TGGGCAAGTGGTCATGGCGG-TAMRA. Forward FGFR-3 primer, 5′-GGCCATCGGCATTGACAA; reverse, 5′-GCATCGTCTTTCAGCATCTTCA; probe 6FAM-5′-CGCCAAGCCTGTCACCGTAGCC-TAMRA. Forward FGFR-4 primer, 5′-GTGCAGAAGCTCTCCCGCT; reverse, 5′-TGAGCTTGACTTGCCGGAA; probe 6FAM-5′-CCCTCTGGCCCGACAGTTCTCCC-TAMRA. Forward β-actin primer, 5′-TCACCCACACTGTGCCCATCTACGA; reverse, 5′-CAGCGGAACCGCTCATTGCCAATGG; probe 6FAM-5′-ATGCCCTCCCCCATGCCATCCTGCGT-TAMRA.

The PCR reaction mixture (50 μl) consisted of 100 nM of each primer, 20 nM probe, 1.25 U Ampli-Taq Gold, 200 μM each of dATP, dCTP, and dGTP, 200 μM dUTP, 3.5 mM MgCl2, 0.01 U AmpErase uracil N-glycosylase, and 1× TaqMan Buffer A. The following cycling conditions were used: 50°C for 2 minutes, 95°C for 10minutes, followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute, using the ABI Prism 7700 sequence detection system from Applied Biosystems. Real-time PCR data were expressed as a relative quantity based on the ratioof the fluorescent change observed with FGFR-1, -2, -3, and -4 relative to the fluorescent change observed with β-actin.

Recombinant adenovirus and FGF2–Fab′ conjugates

Recombinant E1A-deleted and CMV-driven adenovirus expressing firefly luciferase (AdLuc) or green fluorescent protein (AdGFP), an E1-deleted Ad5 vector expressing the CMV-driven HSV-TK gene (AdTK), and the FGF2–Fab′ conjugate were constructed as previously reported16,17 and provided by one of the authors on this study (Dr B Sosnowski). The FGF2–Fab′ conjugate consists of a modified recombinant FGF2 linked with the Fab′ fragment from a blocking monoclonal antibody generated against the adenovirus 5 knob region.16,17 The specificity of the interactions of FGF2–Fab′ with FGFRs was previously demonstrated by neutralizing its activities with an anti–FGFR-1 antibody.16

Cell culture and infection

Cells were grown in DME medium (COLO-357, MIA-PaCa-2, PANC-1, and L6 cells) or RPMI medium (ASPC-1, CAPAN-1, and T3M4) at 37°C in humidified air with 5% CO2. Media were supplemented with penicillin G (100 U/mL), streptomycin (100 μg/mL), and 8% FBS. Generation of L6 cells that are stably transfected with a full-length FGFR-1 cDNA (L6-FGFR-1) was previously reported.25 For adenoviral infections, cells were plated in 12-well plates or 10-cm dishes and incubated overnight. Infection complexes containing adenovirus (AdLuc or AdTK) and FGF2–Fab′ conjugates (at varying ratios of FGF2–Fab′ conjugate per viral knob particle) were mixed in a final volume of 100 μl HEPES buffer (20 mM HEPES, pH 7.6, 0.1 M NaCl), and incubated for 30 minutes at 23°C. For these experiments, the number of knob particles per virion was calculated on the basis that there are 12 fibers per virion, and that each functional fiber is a trimer with 3 knob domains, yielding 36 knob domains per virion. Thus, assuming that there is no steric interference, each of the 36 knob domains is capable of binding a Fab molecule.16,17,18,19 The mixtures were then diluted in serum-free medium (medium containing 0.1% BSA, 5 μg/mL transferrin, 5 ng/mL sodium selenite, and antibiotics) and added to cells. Subsequently, cells were incubated at 37°C in humidified air with 5% CO2 for 1 hour with gentle agitation every 15 minutes. After 1 hour, infection medium was aspirated and cells were incubated in complete medium. TK expression under these experimental conditions was previously confirmed by performing an in vitro TK enzyme assay in cell lysates prepared 48 hours after infection with AdTK.26

Luciferase assay

Cells were seeded in 12-well plates at a density of 50,000 cells/well, infected with the AdLuc virus, and lysed after 48 hours in 200 μl Promega lysis buffer. Aliquots of each sample (10 μl) were then assayed using the MLX microtiter plate luminometer (Dynex Technologies, Chantilly, VA) 18, and the Promega assay reagent according to the manufacturers' instructions, as previously reported.27 Adenoviral infection of L6 cells was carried out with minor modifications. Thus, L6 cells were seeded in 12-well plates at a density of 25,000 cells/well. The AdLuc and FGF2–Fab′ conjugates were mixed in a final volume of 25 μl HEPES buffer, incubated for 30 minutes at 23°C, and then diluted in 300 μl DME medium containing 0.5% FBS and antibiotics. Cells were then infected with the AdLuc virus. After 1 hour, complete medium was added and cells were lysed after 48 hours in 100 μl Promega lysis buffer. Examination of all six pancreatic cancer cells by fluorescent microscopy following infection with Ad-GFP confirmed that virtually 100% of the cells were transduced. For all infection experiments, the recombinant adenoviral vectors were propagated on a permissive 293 cell line, followed by purification and determination of plaque-forming units (pfu), which were carried out according to standard techniques.28

Cell growth assays

To assess FGF2-AdTk- and AdTK-mediated killing, cells were seeded in 10-cm dishes at a density of 2×106 cells/dish and infected with virus for 24 hours. Cells were then trypsinized and seeded overnight in 96-well plates at a density of 5000 cells/well. Medium was then aspirated andthe cells were incubated in complete medium in the absence or presence of various concentrations of ganciclovir. After 3 days, cell numbers were determined by the 3-(4,5-methylthiazol-2-yl)-2,5-diaphenyltertrazolium bromide (MTT) assay as previously described.29 The assay was initiated by adding MTT solution to each well at a final concentration of 0.625 μg MTT/mL medium. After 4 hours the optical density was measured at 570 nm with an ELISA plate reader (Molecular Devices, Menlo Park, CA) following removal of the medium and dissolving the dye crystals in acidified isopropanol. In pancreatic cancer cells the results of the MTT assay correlate with results obtained by cell counting with a hemocytometer and by monitoring [3H]thymidine incorporation.29

Binding assay

The extent of surface FGF2 binding was next determined in three cell lines that exhibited varying levels of transduction efficiency. Accordingly, COLO-357, PANC-1, and T3M4 cells were grown to 75% confluency in 12-well plates (2 mL medium/well), rinsed with binding buffer (DMEM, 0.2% BSA, 25 mM HEPES, pH 7.4) and preincubated in binding buffer at 4°C for 20 minutes. 125I-FGF2 was then added for the indicated times (100,000 cpm/mL) in the absence or presence of 400 ng/mL unlabeled ligand. Cells were washed twice with cold PBS and twice with PBS/1.5 M NaCl. The cells were then lysed with 0.1 N NaOH/1% sodium dodecyl sulfate (SDS) and the radioactivity measured by gamma counting.25 Specific binding was determined by subtracting the nonspecific binding observed with samples that were incubated with the unlabeled ligand.

Patients and tissue sampling

PDAC tissues were obtained from three male and four female patients (median age: 62 years; range: 43–80 years). According to the Union Internationale Contre le Cancer (UICC) classification, there were one stage II, five stage III, and one stage IV duct cell adenocarcinomas. Freshly removed tissue samples were snap-frozen in liquid nitrogen immediately on surgical removal and maintained at −80°C until use. All studies were approved by the Human Subjects Committees at the University of California, Irvine, and the University of Heidelberg, Germany.

Laser capture microdissection

To assess the levels of expression of the various high-affinity FGFRs, cancer cells from seven PDAC samples were removed by laser capture microdissection,30 using a laser capture system (model PXL-100/ARC100) and CapSure LCM transfer film from Arcturus (Mountain View, CA). Total RNA was then extracted from these cells by using the single-step acid–guanidine–thiocyanate phenol chloroform method, as previously described,21,22 but with the addition of glycogen to enhance the quantitative precipitation of the RNA during the extraction procedure. RNA was then analyzed for FGFR mRNA expression levels using real-time quantitative PCR as described above.

Statistics

Whenever indicated, Student's t test was used for statistical analysis (two-sided), P<.05 was taken as the level of significance.

Results

AdLuc is specifically targeted to FGFR-1–expressing L6 myoblasts

The relationship between FGFR expression and FGFR-mediated targeted gene therapy was initially tested using parental L6 myoblasts that express very low levels of endogenous FGFRs and L6 cells transfected with FGFR-1. To confirm the low levels of expression of FGFRs in parental L6 myoblasts, RNase protection assays were carried out using specific probes for each major high-affinity FGFR. These assays did not detect the presence of FGFR-1, -2, -3, or -4 in parental L6 cells (data not shown). However, when a more sensitive real-time quantitative PCR assay was used, very low levels of FGFR-1 were detected in these cells (Fig1A). We previously reported that parental L6 cells do not bind radiolabeled FGF2, whereas L6 cells transfected with FGFR-1 exhibit approximately 20,000 cell-surface FGF receptors.25 Indeed, as determined by real-time quantitative PCR, transfected L6 cells expressed high levels of FGFR-1 (Fig 1B). Moreover, stable transfection of FGFR-1 was not associated with changes in expression of FGFR-2, -3, or -4 (Fig 1B).

Figure 1
figure1

Expression of FGFRs in wild-type and transfected L6 cells. Real-time quantitative PCR was performed as described in the Materials and methods section for the four high-affinity FGF receptors in parental L6 myoblasts (A), and in L6 cells transfected with human FGFR-1 (B). The relative normalized fluorescence change (ΔRn) is plotted for each sample as a function of the PCR cycle number. The expression of high levels of FGFR-1 (denoted as R-1) is clearly evident in panel (B). R-2: FGFR-2; R-3: FGFR-3; R4: FGFR-4.

The efficiency of adenoviral transduction was initially assessed by infecting L6 cells with an adenoviral vector containing sequences encoding luciferase (AdLuc). At 5–50 pfu/cell, L6 cells were poorly transducible by AdLuc (data now shown). In contrast, at 100 pfu/cell, wild-type and transfected L6 cells yielded relative light units (RLUs) of 100 and 250, respectively (Fig 2). Cell detachment or cell death was not observed during any of these experiments. Next, studies were performed in the presence of ratios of 5:1, 10:1 and 50:1 of FGF2–Fab′ conjugate per viral knob particle (Fig 2), because this conjugate acts as a linker between the viral knob and the cellular FGFRs, thereby targeting the adenovirus to FGFRs. In the presence of the FGF2–Fab′ conjugate, transduction efficiency increased 7- to 10-fold in the parental L6 cells in a manner that was independent of the FGF2–Fab′ conjugate per viral knob particle (Fig 2 solid bars). In contrast, under the same conditions, transduction efficiency in the FGFR-1 transfected L6 cells increased 34-fold at a 5:1 ratio of FGF2–Fab′ conjugate per viral knob particle ratio, and 52-fold at a 50:1 ratio, yielding RLUs of 9000–13,000 (Fig 2 open bars).

Figure 2
figure2

Specificity of Adluc FGF2–Fab′ conjugate transduction. L6 cells and FGFR-1-L6 cells were infected with AdLuc (pfu of 100/cell) in the absence (0) or presence of the FGF2–Fab′ conjugate at ratios of 5, 10, and 50, as indicated. Data are the means±SD of three determinations from a representative of three experiments.

FGFR-targeted gene delivery enhances transduction efficiency in pancreatic cancer cells

Adenoviral infection efficiency in pancreatic cancer cells was determined by infecting each cell line with the AdLuc vector (100 pfu/cell) in the absence of the FGF2–Fab′ conjugate. CAPAN-1 and COLO-357 cells displayed a relatively low transduction efficiency, yielding RLUs of <10,000 per assay, whereas MIA-PaCa-2 cells yielded approximately 16,000 RLUs per assay. In contrast, ASPC-1, PANC-1, and T3M4 displayed moderate to high transduction efficiency ranging from 28,000 to approximately 100,000 RLUs per assay (Fig 3). Three cell lines were also tested for their ability to bind FGF2 at the cell surface by performing binding studies with 125I-FGF2 at 4°C. Analysis of binding data revealed that COLO-357 and T3M4 cells bound 8.5% and 9% of the added ligand, respectively, whereas PANC-1 cells bound 5.5% of the added ligand (Fig 4). Thus, as expected, in the absence of the FGF2–Fab′ conjugate there was no correlation between surface binding of FGF2 and AdLuc transduction efficiency.

Figure 3
figure3

Efficiency of adenoviral infection. Six pancreatic cancer cell lines were infected with the AdLuc vector, encoding the firefly luciferase gene (100 pfu/cell) in the absence of the FGF2–Fab′ conjugate. Luciferase gene expression was analyzed 48 hours after infection and expressed as relative light units. Data are the means±SD of three determinations per experiment from two separate experiments.

Figure 4
figure4

FGF2 binding. COLO-357, PANC-1 and T3M4 cells were grown to 75% confluency in 12-well plates (2 mL medium/well), rinsed with binding buffer (DMEM, 0.2% BSA, 25 mM HEPES, pH 7.4) and preincubated in binding buffer at 4°C for 20 minutes before the addition of 125I-FGF2 (4°C) for the indicated times (100,000 cpm/mL). Cells were washed and lysed, and the radioactivity in the lysates was measured by gamma counting as described in the Materials and methods section. Specific binding was determined by subtracting the values obtained in the presence of 400 ng/mL unlabeled FGF2. Data are the means±SE of three determinations per experiment from three separate experiments.

Next, cells were infected with the AdLuc vector in the absence or presence of the FGF2–Fab′ conjugate at ratios of 1, 5, and 10 conjugates per knob particle. It has been shown previously that the addition of the FGF2–Fab′ conjugate specifically targets the FGFRs via FGF2 because addition of FGF2 antiserum blocked its effects.17,31 In all six pancreatic cancer cell lines, there was an enhancement of AdLuc transduction, ranging from 2- to 34-fold (Fig 5). Addition of the FGF2–Fab′ conjugate resulted in a 2.5- to 4-fold increase in adenoviral transduction in PANC-1 and ASPC-1 cells, a 10- to 13-fold increase in CAPAN-1, MIA-PaCa-2, and T3M4 cells, and a 34-fold increase in COLO-357 cells (Fig 5). Increasing the ratio of FGF2–Fab′ conjugates per viral knob particle from 1 to 5 or 10 did not result in a significant increase in AdLuc transduction efficiency in any of the cell lines. Thus, when comparing cell lines for which FGF2 binding was tested, enhanced AdLuc transduction efficiency in the presence of the FGF2–Fab′ conjugate correlated with greater surface binding of FGF2.

Figure 5
figure5

Effects of FGF2–Fab′ conjugate on Adluc transduction. Pancreatic cancer cells were infected with the AdLuc vector in the absence (0) or presence of the FGF2–Fab′ conjugate. The ratios of conjugates per knob particle were 1, 5, and 10. Data are expressed as percent of control (which is set at 100), and are the means±SD of three determinations per experiment from two separate experiments.

FGFR-targeted suicide gene therapy is efficient in a subgroup of pancreatic cancer cells

Next, we used an adenoviral vector encoding the conditionally toxic gene product HSV-TK (AdTK). Cells were infected with AdTK (100 pfu/cell) in the absence or presence of the FGF2–Fab′ conjugate at ratios of 1 and 10 FGF2–Fab′ conjugates per knob particle, and then incubated in the absence or presence of ganciclovir. Cell growth was then assessed by the MTT assay (Fig 6). In the absence of AdTK, ganciclovir did not alter the growth of any of the pancreatic cancer cell lines. Furthermore, following AdTk transduction in the absence of the FGF2–Fab′ conjugate, only MIA-PaCa-2 and T3M4 cells were responsive to ganciclovir-mediated growth inhibition. Thus, ganciclovir (50 μg/mL) inhibited the growth of MIA-PaCa-2 and T3M4 cells by 40% and 80%, respectively, and this effect was not significantly enhanced by the addition of the FGF2–Fab′ conjugate. In contrast, in CAPAN-1 and COLO-357 cells, the FGF2–Fab′ conjugate was necessary for AdTK-mediated growth inhibition. In COLO-357 cells, ganciclovir (50 μg/mL) inhibited growth by 60% (P<.001) in the presence of the FGF2–Fab′ conjugate. Ganciclovir (50 μg/mL) inhibited growth to a lesser extent in CAPAN-1 cells, with conjugate to knob ratios of 1:1 and 1:10 inhibiting growth by 30% (P<.01) and 22% (P<.05), respectively. However, the differences between these two inhibitory effects were not statistically significant.

Figure 6
figure6

Effects of ganciclovir and AdTK on pancreatic cancer cells. Cells were either treated with ganciclovir alone (solid circles) or infected (100 pfu/cell) with AdTK in the absence (solid triangle) or presence of the FGF2–Fab′ conjugate at ratios of 1 (open square) and 10 (open triangle) per knob particle. Cells were subsequently maintained in the absence or presence of various concentrations of ganciclovir, and MTT assays were performed as described in the Materials and methods section. Data are the means±SE of three determinations per experiment from three separate experiments.

Expression of FGFRs in pancreatic cancer cells and tumors

RNase protection assays (Fig 7) and real-time PCR analysis (Fig 8) were next carried out to compare the expression levels of the four high-affinity FGFRs in all six pancreatic cancer cells lines. As summarized in Table 1, both methods revealed relatively high levels of FGFR-1 and -4 in ASPC-1 cells, moderate levels of FGFR-4 in CAPAN-1 and Mia PaCa-2 cells, and relatively high levels of FGFR-1, moderate levels of FGFR-2 and -3, but low levels of FGFR-4 in COLO-357 cells. PANC-1 cells expressed relatively moderate levels of FGFR-1 and -4, very low levels of FGFR-3, and undetectable levels of FGFR-2 by RNase protection. In contrast, by real-time PCR, these cells expressed moderate levels of FGFR-3 and -4, low levels of FGFR-1, and very low levels of FGFR-2. Similarly, in T3M4 cells, RNase protection revealed the presence of moderate levels of FGFR-3, low levels of FGFR-4, and barely detectable levels of FGFR-2, whereas FGFR-1 was below the level of detection. In contrast, by real-time PCR, these cells expressed high levels of FGFR-3, low levels of FGFR-2 and -4, and very low levels of FGFR-1.

Figure 7
figure7

Expression of FGFRs in pancreatic cancer cell lines by ribonuclease protection assay. Total RNA (10 μg per sample) from the indicated cell lines was hybridized overnight at 42°C with [α-32P]CTP-labeled riboprobes (100,000 cpm/sample), as described in the Materials and methods section. Yeast tRNA (10 μg) was used as a negative control. Samples were subjected to denaturing gel electrophoresis on 6% polyacrylamide/8 M urea gels. The protected FGFR fragments are indicated on the left, and probe migration is indicated on the right.

Figure 8
figure8

Expression of FGFRs in pancreatic cancer cell lines by real-time quantitative PCR. mRNA levels encoding FGFR-1, -2, -3, and -4 were determined on RNA extracted from the indicated cell lines (MIA-PaCa-2: MIA), as described in the Materials and methods section, using real-time quantitative PCR.

Table 1 Expression of FGF receptors in pancreatic cancer cell lines

Using immunostaining and in situ hybridization techniques as well as RNase protection assays, we previously reported that FGFR-1, FGFR-2, and the related keratinocyte growth factor receptor (KGFR) are expressed at relatively high levels in pancreatic cancer cells within the pancreatic tumor mass.21,22 Using laser capture microdissection and real-time PCR, in the present study we were able to assess the expression levels of all four high-affinity FGFRs in the pancreatic cancer cells from seven different PDAC samples (Fig 9). Three samples exhibited relatively high levels of FGFR-1 (samples 1, 5, and 6), and two of these three samples also exhibited high levels of FGFR-2 and moderate levels of FGFR-3 and -4 (samples 1 and 6). In each sample, FGFR-1 was expressed at the highest levels by comparison with the other receptors, with the exception of sample 3, in which FGFR-2 was expressed at the highest levels. Only sample 7 exhibited low levels of all four receptors. Thus, there was concomitant, albeit variable, overexpression of mRNA moieties encoding the high-affinity FGFRs in the captured cancer cells from the majority of the tumors (Fig 9).

Figure 9
figure9

Expression of FGFRs in laser-captured pancreatic cancer cells within the tumor mass. mRNA levels encoding FGFR-1, -2, -3, and -4 were determined by real-time quantitative PCR, using RNA extracted from cancer cells that were captured from seven different tumor samples, as described in the Materials and methods section.

Discussion

PDAC remains one of the most difficult malignancies to treat, exhibiting a poor response to standard chemotherapy and radiotherapy, as well as to alternative treatment options such as immunotherapy or antihormonal therapy. These types of therapy either fail or offer only limited benefits in terms of survival and quality of life for patients with PDAC. Nonetheless, the standard of care for these patients often includes chemotherapy and/or radiotherapy. Therefore, novel, experimental therapies are needed for this deadly disease.

Suicide gene therapy, such as HSV-TK–based suicide gene therapy, offers a unique approach for suppressing cancer growth.9,10,11,12,13,14 Transduction of tumor cells with a virus bearing the HSV-TK gene leads to its expression in these cells. HSV-TK–expressing cells can then be targeted for killing by the nucleoside analog ganciclovir, which is phosphorylated by HSV-TK.32 Cellular kinases in turn further phosphorylate ganciclovir, which then incorporates into DNA thereby inhibiting and terminating its synthesis and causing subsequent cell death.9,10,11,12,13,14,15,16,32 This anticancer suicide gene therapy has been successfully employed in a number of in vitro and in vivo studies and has found its way into clinical trials. Its potential usefulness in pancreatic cancer cells in vitro,13,16 as well as in animal tumor models for pancreatic cancer,20,33,34,35 has been previously demonstrated. However, the specificity of this approach in relation to the expression profile of FGFRs in PDAC has not been yet been clearly documented.

It is possible to specifically target adenoviral vectors to FGFR expressing cells by using a conjugate protein that consists of FGF2 linked to the Fab′ fragment of an anti-adenoviral knob antibody. In theory, this approach has the potential to decrease inappropriate vector targeting and enhancing transduction of target cells that express high levels of FGFR. In the present study, we confirmed that the FGF2–Fab′ approach was targeting FGFR expressing cells, using parental L6 rat myoblasts, which express very low levels of FGFR-1, and L6 cells transfected to overexpress FGFR-1. Thus, addition of the FGF2–Fab′ conjugate markedly enhanced transduction in the FGFR-1 transfected L6 cells. However, transduction efficiency was also slightly enhanced by FGFR targeting in parental L6 cells, indicating that the FGF2–Fab′ conjugate can act even when low levels of FGFRs are present. This observation is in agreement with results obtained by Doukas et al.16

Both by RNase protection and by real-time quantitative PCR analysis, all six pancreatic cancer cell lines tested in the present study expressed high-affinity FGFRs. As expected, there was no correlation between transduction efficiency and FGFR expression or FGF2 binding when the cells were infected with the AdLuc vector in absence of the FGF2–Fab′ conjugate. Furthermore, addition of the FGF2–Fab′ conjugate only resulted in a modest increase in adenoviral transduction in ASPC-1 (4-fold) and PANC-1 (2.5-fold) cells. In contrast, in the presence of FGF2–Fab′, there was considerable enhancement of AdLuc transduction in CAPAN-1 (13-fold), COLO-357 (34-fold), MIA-PaCa-2 (10-fold), and T3M4 (10-fold) cells. Taken together, these observations suggest that the concomitant expression of various high-affinity FGFRs leads to improved transduction efficiency of pancreatic cancer cells following FGF2–Fab′ addition. It is also possible, however, that endogenous production of FGFs by the cancer cells and/or differences in the levels of cell-surface FGFRs and the corresponding mRNA species also dictate the ability of the FGF2–Fab′ conjugate to improve transduction efficiency.

Not surprisingly, ganciclovir alone did not alter the growth of any of the pancreatic cancer cell lines tested in the present study. Furthermore, ASPC-1 and PANC-1 cells, which exhibited the smallest increase in transduction efficiency in the presence of the FGF2–Fab′ conjugate, were completely resistant to ganciclovir-mediated growth inhibition after AdTk transduction, irrespective of the absence or presence of the FGF2–Fab′ conjugate. In contrast, ganciclovir inhibited in a dose-dependent manner the growth of MIA-PaCa-2 and T3M4 cells infected with AdTK, and this effect was especially marked in T3M4 cells. However, it was not enhanced by the FGF2–Fab′ conjugate, despite the conjugate's ability to improve transduction efficiency in both cell lines. In the case of T3M4 cells, this lack of requirement for the FGF2–Fab′ conjugate was most likely due to the high efficiency of transduction by the adenoviral vector. Only CAPAN-1 and COLO-357 cells exhibited a significant increase in ganciclovir susceptibility after addition of the FGF2–Fab′ conjugate, and this effect was most pronounced in COLO-357 cells. Altogether, ganciclovir was an effective growth inhibitor in four of six pancreatic cancer cell lines, and this effect was enhanced by FGFR targeting in the two cell lines (CAPAN-1 and COLO-357) that exhibited the greatest increase in transduction efficiency in the presence of the FGF2–Fab′ conjugate.

The reasons why some pancreatic cancer cell lines are resistant to ganciclovir are not known. Resistance to ganciclovir could not be explained by poor transduction efficiency, inasmuch as ASPC-1 and PANC-1 cells were transducible as determined by AdLuc infection, but were completely resistant to ganciclovir. Instead, it is likely that resistance of HSV-TK–transduced tumor cells to ganciclovir is due to other mechanisms. Thus, the HSV-TK gene may be either partially or completely deleted by the cells following retroviral infection.35 Alternatively, some HSV-TK–transduced cells exhibit an attenuated capacity to transfer the toxic phosphorylated ganciclovir via intercellular gap junctions to nontransduced tumor cells, perhaps as a consequence of decreased gap junction expression and/or function.10,36,37,38 In the absence of such an efficient transfer, a bystander effect cannot occur, and cells that are not directly transduced with the suicide gene are not able to respond to ganciclovir. In addition, cells that harbor a mutated p53 tumor suppressor gene have been reported to be resistant to ganciclovir.39 Although pancreatic cancer cells frequently harbor p53 mutations, they can be rendered sensitive to ganciclovir.13,16,33,34,35,40 Because ganciclovir is a cell cycle active agent, it is also conceivable that differences in cell growth rates may explain some of the differences in sensitivity to this agent.

Several studies have previously demonstrated that PDACs overexpress FGFRs.20,21 In addition, in the present study, analysis of laser-captured cancer cells from seven PDAC samples revealed the concomitant presence of high levels of FGFRs in the majority of the tumors. Indeed, following normalization to β-actin, FGFR levels in the PDAC samples were approximately 10-fold higher than in the cultured pancreatic cancer cell lines. These high expression levels of FGFRs are consistent with the strong intensity of the in situ hybridization signals observed in the pancreatic cancer cells within the tumor mass in our previous studies.20,21 This observation suggests that FGFR-targeted suicide gene therapy may ultimately prove to be beneficial in various subgroups of patients with PDAC. The advantages of targeted therapy include higher specificity, reduced virus load, higher transduction rates, and reduced systemic toxicity.41,42 In this context, analysis of tumor biopsy samples by the highly reproducible and sensitive quantitative PCR method employed in the present study may also provide information as to which patients are most likely to benefit from this type of therapy.

References

  1. 1

    Warshaw AL, Fernández del Castillo C . Pancreatic carcinoma N Engl J Med 1992 326: 455–465

  2. 2

    Parker SL, Tong T, Bolden S, Wingo PA . Cancer statistics, 1997 CA-Cancer J Clin 1997 47: 5–27

  3. 3

    Link KH, Gansauge F, Pillasch J, Beger HG . Multimodal therapies in ductal pancreatic cancer. The future Int J Pancreatol 1997 21: 71–83

  4. 4

    Poen JC, Ford JM, Niederhuber JE . Chemoradiotherapy in the management of localized tumors of the pancreas Ann Surg Oncol 1999 6: 117–122

  5. 5

    Brammer R, Buckels J, Bramhall S . Advances in non-operative therapy in pancreatic cancer Int J Clin Pract 2000 54: 373–381

  6. 6

    Dobelbower RR Jr . Current radiotherapeutic approaches to pancreatic cancer Cancer 1981 47: 1729–1733

  7. 7

    Büchler M, Friess H, Schultheiss KH et al. A randomized controlled trial of adjuvant immunotherapy (murine monoclonal antibody 494/32) in resectable pancreatic cancer Cancer 1991 68: 1507–1512

  8. 8

    Moolten FL . Tumor chemosensitivity conferred by inserted herpes thymidine kinase genes: paradigm for a prospective cancer control strategy Cancer Res 1986 46: 5276–5281

  9. 9

    Barba D, Hardin J, Ray J, Gage FH . Thymidine kinase–mediated killing of rat brain tumors J Neurosurg 1993 79: 729–735

  10. 10

    Yang L, Chiang Y, Lenz HJ et al. Intercellular communication mediates the bystander effect during herpes simplex thymidine kinase/ganciclovir-based gene therapy of human gastrointestinal tumor cells Hum Gene Ther 1998 9: 719–728

  11. 11

    Ido A, Nakata K, Kato Y et al. Gene therapy for hepatoma cells using a retrovirus vector carrying herpes simplex virus thymidine kinase gene under the control of human alpha-fetoprotein gene promoter Cancer Res 1995 55: 3105–3109

  12. 12

    Garver RI Jr, Goldsmith KT, Rodu B, Hu PC, Sorscher EJ, Curiel DT . Strategy for achieving selective killing of carcinomas Gene Ther 1994 1: 46–50

  13. 13

    Rosenfeld ME, Vickers SM, Raben D et al. Pancreatic carcinoma cell killing via adenoviral mediated delivery of the herpes simplex virus thymidine kinase gene Ann Surg 1997 225: 609–618

  14. 14

    Robertson MW III, Wang M, Siegal GP et al. Use of a tissue-specific promoter for targeted expression of the herpes simplex virus thymidine kinase gene in cervical carcinoma cells Cancer Gene Ther 1998 5: 331–336

  15. 15

    Siders WM, Halloran PJ, Fenton RG . Melanoma-specific cytotoxicity induced by a tyrosinase promoter-enhancer/herpes simplex virus thymidine kinase adenovirus Cancer Gene Ther 1998 5: 281–291

  16. 16

    Doukas J, Hoganson DK, Ong M et al. Retargeted delivery of adenoviral vectors through fibroblast growth factor receptors involves unique cellular pathways FASEB J 1999 13: 1459–1466

  17. 17

    Goldman CK, Rogers BE, Douglas JT et al. Targeted gene delivery to Kaposi's sarcoma cells via the fibroblast growth factor receptor Cancer Res 1997 57: 1447–1451

  18. 18

    Hoganson DK, Chandler LA, Fleurbaaij GA et al. Targeted delivery of DNA encoding cytotoxic proteins through high-affinity fibroblast growth factor receptors Hum Gene Ther 1998 9: 2565–2575

  19. 19

    Gu DL, Gonzalez AM, Printz MA et al. Fibroblast growth factor 2 retargeted adenovirus has redirected cellular tropism: evidence for reduced toxicity and enhanced antitumor activity in mice Cancer Res 1999 59: 2608–2614

  20. 20

    Carrio M, Romagosa A, Mercade E et al. Enhanced pancreatic tumor regression by a combination of adenovirus and retrovirus-mediated delivery of the herpes simplex virus thymidine kinase gene Gene Ther 1999 6: 547–553

  21. 21

    Kobrin MS, Yamanaka Y, Friess H, Lopez ME, Korc M . Aberrant expression of type I fibroblast growth factor receptor in human pancreatic adenocarcinomas Cancer Res 1993 53: 4741–4744

  22. 22

    Ishiwata T, Friess H, Büchler MW, Lopez ME, Korc M . Characterization of keratinocyte growth factor and receptor expression in human pancreatic cancer Am J Pathol 1998 153: 213–222

  23. 23

    Kornmann M, Beger HG, Korc M . Role of fibroblast growth factors and their receptors in pancreatic cancer and chronic pancreatitis Pancreas 1998 17: 169–175

  24. 24

    Kornmann M, Danenberg KD, Arber N, Beger HG, Danenberg PV, Korc M . Inhibition of cyclin D1 expression in human pancreatic cancer cells is associated with increased chemosensitivity and decreased expression of multiple chemoresistance genes Cancer Res 1999 59: 3505–3511

  25. 25

    Lopez ME, Korc M . A novel type I fibroblast growth factor receptor activates mitogenic signaling in the absence of detectable tyrosine phosphorylation of FRS2 J Biol Chem 2000 275: 15933–15939

  26. 26

    Arner ES, Spasokoukotskaja T, Eriksson S . Selective assays for thymidine kinase 1 and 2 and deoxycytidine kinase and their activities in extracts from human cells and tissues Biochem Biophys Res Commun 1992 188: 712–718

  27. 27

    Kleeff J, Ishiwata T, Maruyama H et al. The TGF-beta signaling inhibitor Smad7 enhances tumorigenicity in pancreatic cancer Oncogene 1999 18: 5363–5372

  28. 28

    Mittereder N, March KL, Trapnell BC . Evaluation of the concentration and bioactivity of adenovirus vectors for gene therapy J Virol 1996 70: 7498–7509

  29. 29

    Raitano AB, Korc M . Tumor necrosis factor up-regulates gamma-interferon binding in a human carcinoma cell line JBiol Chem 1990 265: 10466–10472

  30. 30

    Emmert-Buck MR, Bonner RF, Smith PD et al. Laser capture microdissection Science 1996 274: 998–1001

  31. 31

    Rogers BE, Douglas JT, Ahlem C, Buchsbaum DJ, Frincke J, Curiel DT . Use of a novel cross-linking method to modify adenovirus tropism Gene Ther 1997 4: 1387–1392

  32. 32

    Elion GB . The biochemistry and mechanism of action of acyclovir J Antimicrob Chemother 1983 12: Suppl B 9–17

  33. 33

    Makinen K, Loimas S, Wahlfors J, Alhava E, Janne J . Evaluation of herpes simplex thymidine kinase mediated gene therapy in experimental pancreatic cancer J Gene Med 2000 2: 361–367

  34. 34

    Block A, Chen SH, Kosai K, Finegold M, Woo SL . Adenoviral-mediated herpes simplex virus thymidine kinase gene transfer: regression of hepatic metastasis of pancreatic tumors Pancreas 1997 15: 25–34

  35. 35

    Yang L, Hwang R, Chiang Y, Gordon EM, Anderson WF, Parekh D . Mechanisms for ganciclovir resistance in gastrointestinal tumor cells transduced with a retroviral vector containing the herpes simplex virus thymidine kinase gene Clin Cancer Res 1998 4: 731–741

  36. 36

    Chen CY, Chang YN, Ryan P, Linscott M, McGarrity GJ, Chiang YL . Effect of herpes simplex virus thymidine kinase expression levels on ganciclovir-mediated cytotoxicity and the “bystander effect” Hum Gene Ther 1995 6: 1467–1476

  37. 37

    Fick J, Barker FG II, Dazin P, Westphale EM, Beyer EC, Israel MA . The extent of heterocellular communication mediated by gap junctions is predictive of bystander tumor cytotoxicity in vitro Proc Natl Acad Sci USA 1995 92: 11071–11075

  38. 38

    Pitts JD . Cancer gene therapy: a bystander effect using the gap junctional pathway Mol Carcinog 1994 11: 127–130

  39. 39

    Matsubara H, Kawamura K, Sugaya M et al. Differential efficacy of suicide gene therapy by herpes simplex virus–thymidine kinase gene reflects the status of p53 gene in human esophageal cancer cells Anticancer Res 1999 19: 4157–4160

  40. 40

    Howard BD, Boenicke L, Schniewind B, Henne-Bruns D, Kalthoff H . Transduction of human pancreatic tumor cells with vesicular stomatitis virus G-pseudotyped retroviral vectors containing a herpes simplex virus thymidine kinase mutant gene enhances bystander effects and sensitivity to ganciclovir Cancer Gene Ther 2000 7: 927–938

  41. 41

    Douglas JT, Rogers BE, Rosenfeld ME, Michael SI, Feng M, Curiel DT . Targeted gene delivery by tropism-modified adenoviral vectors Nat Biotechnol 1996 14: 1574–1578

  42. 42

    Printz MA, Gonzalez AM, Cunningham M et al. Fibroblast growth factor 2–retargeted adenoviral vectors exhibit a modified biolocalization pattern and display reduced toxicity relative to native adenoviral vectors Hum Gene Ther 2000 11: 191–204

Download references

Acknowledgements

This work was supported by Public Health Service Grant CA-40162 awarded by the National Cancer Institute to MK. JK was the recipient of a fellowship award from the University of California Research and Education Grant on Gene Therapy for Cancer. KF was funded by an award from the Technology Development Center Panasonic, Cypress, CA.

Author information

Correspondence to Murray Korc.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • pancreatic cancer
  • FGF receptor
  • adenovirus
  • gene delivery

Further reading