Purpose: Small cell lung cancer (SCLC) is a highly malignant cancer for which there is no curable treatment. Novel therapies are therefore in great demand. In the present study we investigated the therapeutic effect of transcriptionally targeted suicide gene therapy for SCLC based on the yeast cytosine deaminase (YCD) gene alone or fused with the yeast uracil phosphoribosyl transferase (YUPRT) gene followed by administration of 5-fluorocytosine (5-FC) prodrug.
Experimental design: The YCD gene or the YCD-YUPRT gene was placed under regulation of the SCLC-specific promoter insulinoma-associated 1 (INSM1). Therapeutic effect was evaluated in vitro in SCLC cell lines and in vivo in SCLC xenografted nude mice using the nonviral nanoparticle DOTAP/cholesterol for transgene delivery.
Results: INSM1-YCD/5-FC and INSM1-YCD-YUPRT/5-FC therapy induced high cytotoxicity in a range of SCLC cell lines. The highest therapeutic effect was obtained from the YCD-YUPRT fusion gene strategy. No cytotoxicity was induced after treatment of cell lines of other origin than SCLC. In addition the INSM1-YCD-YUPRT/5-FC therapy was superior to an established suicide gene system consisting of the herpes simplex virus thymidine kinase (HSVTK) gene and the prodrug ganciclovir. The superior effect was in part due to massive bystander cytotoxicity of YCD-YUPRT-produced toxins. Finally, INSM1-YCD-YUPRT/5-FC therapy induced significant tumor growth delay in SCLC xenografts compared with control-treated xenografts.
Conclusions: The current study is the first to test cytosine deaminase-based suicide gene therapy for SCLC and the first to show an antitumor effect from the delivery of suicide gene therapeutics for SCLC in vivo. Clin Cancer Res; 16(8); 2308–19. ©2010 AACR.
Because current treatment protocols for the management of small cell lung cancer (SCLC) are inefficient, the advancement of novel therapies should have high priority. In transcriptionally targeted suicide gene therapy a cancer-specific promoter regulates the expression of a suicide gene, which promotes the efficient conversion of a nontoxic prodrug into a cell toxin. In the present study we evaluated the effect of the suicide gene yeast cytosine deaminase fused to uracil phosphoribosyltransferase under transcriptional control of the SCLC-specific insulinoma-associated 1 promoter. The strategy induces massive and specific cell death in SCLC cells lines and significant tumor growth delay upon in vivo delivery of suicide transgene using DOTAP/cholesterol liposome, an established delivery vehicle for lung cancer. The study shows that highly effective and specific therapy can be obtained in SCLC using transcriptionally targeted suicide gene therapy and suggests that initiative should be taken to implement this strategy to replace or supplement existing therapies for SCLC.
Small cell lung cancer (SCLC) represents nearly 20% of lung cancer cases and is characterized by a distinct neuroendocrine phenotype, aggressive progression, and early metastasis. Most newly diagnosed patients with SCLC respond to first-line chemotherapy and radiotherapy, but due to a high frequency of treatment-resistant relapse the 5-year survival rate is <10%. Despite many efforts to improve prognostics with presently available modalities, there have only been subtle improvements in the survival rate of SCLC patients for the last three decades (1, 2). Advancement of novel therapies to replace or complement existing treatment regimes is therefore of high priority for this malignancy.
Suicide gene therapy constitutes a novel promising treatment strategy for cancer, and is based on the introduction of a therapeutic gene encoding an enzyme capable of transforming a nontoxic prodrug into a cell toxin (3). The suicide gene–driven production of toxins obligates strict transgene regulation to obtain activation of prodrug exclusively in malignant cells and avoid toxic side effects. One particular attractive property of suicide gene therapeutics is the ability of produced toxins to spread to nearby cancer cells and induce bystander cytotoxicity. The bystander effect originates from either free diffusion of toxins across the cellular membrane or transport via gap junction intercellular communication (GJIC; refs. 4–6). The combined effect of cancer-specific transgene expression and the local spread of suicide toxins in the tumor environment can mediate efficient drug delivery to a large tumor cell population.
To develop transcriptionally regulated suicide gene therapy for SCLC we previously identified promoters that confer high and specific regulation of transgenes in SCLC cell lines (7–10). We (9, 10) and others (11–16) have previously reported the therapeutic significance of the suicide gene herpes simplex virus thymidine kinase (HSVTK) when combined with the prodrug ganciclovir for SCLC. HSVTK phosphorylates ganciclovir, forming ganciclovir monophosphate, whereas the subsequent formation of ganciclovir diphosphates and triphosphates are catalyzed by cellular kinases. Ganciclovir triphosphate gets incorporated into DNA, causing replication to be blocked (3).
The HSVTK/ganciclovir system has, however, major limitations compromising potential clinical use: (a) the highly charged ganciclovir metabolites can only spread to other cancer cells via GJIC, but because this intercellular communication is greatly compromised in many cancers (17, 18), bystander cytotoxicity may be limited accordingly; (b) the incorporation of ganciclovir triphosphates into DNA during replication causes only the actively dividing cell population to be sensitive to treatment (3); and finally (c) ganciclovir is considered a potential carcinogen with adverse systemic toxic effects even at subclinical concentrations (19).
Apart from HSVTK/ganciclovir no other suicide gene therapeutics have been tested for SCLC, although many suicide gene systems with seemingly attractive clinical features have emerged. The suicide gene cytosine deaminase (CD) from either bacteria or fungi converts the clinically safe and nontoxic prodrug 5-fluorocytosin (5-FC; ref. 20) into 5-fluorouracil (5-FU), a clinically approved cytostatic. Cellular enzymes further metabolize 5-FU, leading to the formation of active toxic metabolites that block DNA and RNA synthesis by simultaneous inhibition of the thymidylate synthase and incorporation in DNA and RNA during replication and transcription (3). It has been proven that yeast cytosine deaminase (YCD) has superior catalytic activity compared with its bacterial (Escherichia coli) counterpart (21–23) in line with the fact that 5-FC in the clinic is used as an antifungal agent and is less efficient against bacterial infections (20).
To circumvent problems of 5-FU resistance due to increased degradation or poor activation of the toxin (24), introduction of the uracil phosphoribosyl transferase (UPRT) gene in combination with the CD gene has been studied. The UPRT enzyme is involved in 5-FU processing in bacteria and yeast and is not present in human cells where other pathways of 5-FU conversion exist. The concomitant expression of the UPRT gene along with CD has been shown to increase therapeutic efficacy markedly (23–27).
Apart from features of safe prodrug application and the ability to affect both the dividing and nondividing cancer cell population, CD-based suicide gene therapy has been shown to induce strong bystander cytotoxicity due to the free diffusion of 5-FU and downstream toxins across cellular membranes (3, 4).
In light of these features, we aimed to investigate the therapeutic significance of CD-based therapy for SCLC. The YCD gene alone or fused to the yeast UPRT (YUPRT) gene (23) was cloned for transcriptional regulation from the SCLC-specific insulinoma-associated 1 (INSM1) promoter (8, 9). We report that INSM1-regulated YCD-YUPRT/5-FC therapy shows superior cytotoxicity when compared with YCD/5-FC and in particular with HSVTK/ganciclovir therapy in a range of SCLC cell lines and induces massive bystander effects. The clinical relevance of INSM1-YCD-YUPRT/5-FC therapy was further established in vivo using a nonviral delivery vehicle DOTAP/cholesterol (DOTAP/Chol) currently in clinical testing in non-SCLC patients (28). We show that treatment of SCLC xenografted mice with DOTAP/Chol encapsulated INSM1-YCD-YUPRT concomitant with 5-FC administration, results in significant tumor growth delay compared with control-treated mice.
Materials and Methods
The origin, characterization, and propagation of all cell lines utilized in the study have been described in detail elsewhere (8, 9).
The following plasmid vectors were used: pGL3 basic (Promega); EGFP-N1 (Clontech); INSM1-HSVTK and INSM1-Luciferase (Luc), described in Pedersen et al. (9); INSM1-YCD and INSM1-YCD-YUPRT, with the YCD and YCD-YUPRT genes kindly provided by Ulrich M. Lauer (Department of Internal Medicine I, University Clinic Tübingen, Tübingen, Germany) in pLXSN-greeN-YCD and pUC19-SCD (super cytosine deaminase equals YCD-YUPRT) plasmids, respectively (23). The excised 0.5 kb YCD (EcoNI/StuI) and 1.1 kb YUPRT (NcoI/HindIII) fragments were inserted into an INSM1-promoter expression vector (all blunt-end ligations). The plasmid vector was prepared from the excision with BglII/XbaI of the Luc gene from INSM1-Luc (9). INSM1-YCD-YUPRT.FLAG: A FLAG (DYKDDDDK) tag was added to the COOH-terminal of the YCD-YUPRT gene reading frame by PCR-amplification using a primer pair amplifying a 500-bp fragment of the 3′ part of the coding sequence, where the FLAG tag and stop codon were included in the reverse primer. Reverse primer: '5-GGGCATGCTTACTTATCATCATCATCTTTG-TAGTCAACACAGTAGTATCTGTCACCAAA-3′; forward primer: '5-CAAAGGGACGAGGAGACTGC-3′. The amplified fragment was cut with AgeI and SphI and inserted into the INSM1-YCD-YUPRT (AgeI/SphI) vector.
For all cell lines, 2 × 106 cells were transfected with 3 μg plasmid using 12 μL Lipofectamine 2000 (LF) in Opti-MEM Reduced Serum Medium (Invitrogen) for 3 h. Adherent cells were plated 1 d prior to transfection. Transfection efficiency was estimated by the transfection of EGFP-N1 plasmid and manual scoring of the proportion of cells expressing EGFP with fluorescence microscopy.
Cell viability assays
Cell viability was measured using a MTT assay (Sigma) with the addition of 20 μL 5 mg/mL MTT solution (dissolved in sterile water) to each well and incubation for 4 h before addition of 100 μL solubilization buffer (10% SDS, 0.01 mol/L HCl). Absorption at 570 nm was measured the following day.
Cells were replated 1 d after transfection at concentrations of 0.025 to 0.05 × 106 cells/well in 96-well plates and incubated for 7 d with either 100 μL growth medium or medium containing prodrug or drug: ganciclovir (196 mmol/L; Cymevene, Roche) dissolved in sterile water, 5-FC (50 mmol/L; Sigma,) dissolved in sterile water and 5-FU (50 mg/mL; Mayne Pharma).
In bystander mixed-cell assays, transfected cells were mixed with untransfected parental cells (1 d after transfection) in defined ratios, plated, and incubated for 7 d.
In bystander medium assays, untransfected cells were plated with conditioned medium from suicide gene–transfected/prodrug-treated cells. The conditioned medium was prepared as follows: cells were transiently transfected at day 1, prodrug was added at day 2, and the growth medium was harvested at day 4.
Western blot analysis
Whole-cell lysates of cells were prepared by sonication in ice-cold Tris-HCl 20 mmol/L (pH 7.5), Triton X-100 2% supplemented with protease and phosphatase inhibitor mixture II and III (Calbiochem). Protein concentrations were determined using the bicinchoninic acid protein assay (Pierce) according to the manufacturer's instructions. Western blot was done using 4-12% Bis-Tris gels, loading 20 μg protein, in the NuPAGE PreCast Gel System (Invitrogen). Primary antibodies were rabbit polyclonal anti-DDDDK (FLAG; 1:4,000, Abcam), rabbit monoclonal anti-tubulin (1:1,000; Cell Signaling), rabbit polyclonal anti-connexin 43 (1:2,000; Sigma). Secondary antibodies were swine anti-rabbit IgG (1:1,000; DAKO). Signals were detected utilizing SuperSignal West Dura extended Duration Substrate (Pierce) in the UVP Biospectrum AC imaging system (AH Diagnostics).
The SCLC xenograft model was established by the injection of 5 × 106 NCI-H69 cells per flank s.c. into 6- to 8-week-old male nude NMRI mice (Taconic). Tumors from injected mice (termed passage 0) were used for serial transplantation of mice that entered treatment protocols (passage 1) or used for serial transplantation of new animals (passage 2). Untreated tumors from each passage were subjected to pathologic analysis to evaluate the existence of clinically validated SCLC markers.
The following treatment setup was done in two independent studies with treatment of mice from passages 1 and 2. Approximately 2 wk after transplantation visible tumor nodules appeared and were measured every second day with caliper to confirm exponential tumor growth before treatment start. Tumor volume was calculated using the formula: π/6 × (d1 × d2)(3/2), where d1 and d2 are diameters at perpendicular angles (26) After 3 to 4 wk tumors had reached a size of 200 to 600 mm3, and mice with tumors undergoing sustained growth were randomized into three treatment groups. The first two groups received either intratumoral (i.t.) injections of 100 μL of INSM1-YCD-YUPRT.FLAG or INSM1-LUC (mock) vector encapsulated in DOTAP/Chol liposome (see below), whereas the third group received i.t. injection of 100 μL 5% glucose (D5W; Sigma)) used for DOTAP/Chol:DNA mixing. From day 1 of i.t. injections 500 mg/kg of 5-FC (Ancotil, Meda AS) were injected i.p. daily for 10 d. When tumors reached 1,000 mm3, the mice were sacrificed and tumor tissue and major organs were resected. For detection of transgene expression in tumor tissue, mice were treated 3 consecutive days with DOTAP/Chol encapsulated INSM1-YCD-YUPRT.FLAG or EGFP-N1, sacrificed at day 4, and tumor tissue was resected. All animal experiments were done in accordance with ethical guidelines under valid license from the Danish Animal Experimentation Board.
DOTAP/Chol:DNA lipoplex preparation for i.t. injections
The DOTAP/Chol reagent was prepared as follows: DOTAP (N-[1-2,3-dioleyl)propyl]-N,N,N-trimethylammonium chloride, Avanti Polar Lipids Inc.) and synthetic cholesterol (Sigma), 140 μmol each, were mixed with rotation and dissolved in a total volume of 7 mL chloroform in a glass flask and fitted in a rota-vaporator-R (Büchi, Roland Carlberg Processystem). The solvent was evaporated with rotation under a nitrogen gas stream at 30°C. The lipid film was dried by applying high vacuum drying for 45 min, followed by hydration of the lipid film in 7 mL of D5W resulting in a 40 mmol/L total lipid solution. The tubes were sealed and placed at 50°C for 30 to 60 min with repeated rotary moment to ensure complete hydration of the lipids on the glass surface and left overnight at room temperature. The next day liposome preparations were sonicated in a sonication water bath (Branson ultrasonic model B.V.) for 5 min at 50°C and then downsized using a small-scale extruder (Avanti Polar lipids Inc.) with polycarbonate nanopore filters (400 nm, 200 nm, and 100 nm; Whatman, Frisenette) at 50°C doing 11 passes for each filter size. Size and charge (ζ potential) of particles were assessed as described previously (29) and were found to be 100 ± 2 nm and +45 mV, respectively, before mixing with DNA. After mixing, lipoplex size and charge was typically 450 ± 2 nm and +40 mV. For mixing with DNA, 20 μL of DOTAP/Chol was diluted in D5W and mixed with DNA solutions containing INSM1-YCD-YUPRT.FLAG or INSM1-LUC (45 μg), respectively, by rapid pipetting up and down, yielding a total volume of 100 μL DOTAP/Chol:DNA lipoplex solution. Within 1 to 2 h of mixing the lipoplexes were injected i.t. into xenografted tumors.
Tissue preparation and analysis
Mice tissues were either fixed overnight in freshly prepared neutral 4% paraformaldehyde followed by incubation in 70% ethanol or freshly frozen in O.M.T. Tissue-Tech (Sakura Finetek) and liquid nitrogen. Paraformaldehyde-fixed tissue were embedded in paraffin and cut to 4-μm sections in a routine fashion on plus coated slides. The slides were deparaffinized, hydrated, and stained with H&E or using the Vectastain ABC kit (Vector Laboratories) according to the manufacturer's instructions except for previously described modifications (30, 31). Polyclonal rabbit anti-GFP (1.3,500; Abcam) antibody was used to detect EGFP expression and polyclonal rabbit anti-FLAG (DDDDK; 1:20,000; Abcam) was used to detect YCD-YUPRT.FLAG expression. Sections were counterstained with hematoxylin and mounted for microscope evaluations using an Olympus BX51 microscope (Olympus A/S). Frozen tissue sections (6 μm) were prepared using a cryostat microtome in a routine fashion. After thawing, sections were mounted and immediately analyzed by Olympus BX2 confocal laser scanning microscopy (emission 488 nm, exitation 530 nm) using Fluoview software version 2.1 (Olympus).
Software and statistics
Data from in vitro cytotoxicity assays were obtained from at least three independent experiments, each done in triplicate or quadruplicate. In vivo data were obtained from two independent studies. For all experiments, data were plotted as mean ± SE unless otherwise stated, and nonlinear regression analysis was done using GraphPad Prism 3.0 software (Inno-Max). All mean comparisons are based on the Student's t-test at indicated significance level.
INSM1 promoter–driven YCD-YUPRT/5-FC suicide gene therapy induces high and specific cytotoxicity superior to YCD/5-FC in SCLC cell lines
Three SCLC cell lines – GLC16, NCI-H69, and DMS53 – which were established at different laboratories and exhibit different growth characteristics (32–34), were selected to obtain a SCLC cell culture model for the evaluation of INSM1-driven suicide gene therapy. We had previously tested INSM1 promoter activity in a large number of SCLC cell lines showing that the promoter exhibit 2- to 10-fold higher activity than the constitutive active SV40 promoter. Among the number of cell lines tested, DMS53 showed relative low activity from the INSM1 promoter whereas NCI-H69 and GLC16 exhibited medium to high INSM1 promoter activity (8, 9). Different levels of INSM1-driven suicide gene expression could therefore be expected in these cell lines. Furthermore, transgene transfection efficiency varied among the cell lines, with 60 ± 3% efficiency in GLC16, 45 ± 2% in NCI-H69, and 25 ± 3% in DMS53.
To evaluate the effect of CD-based suicide gene therapy on SCLC cell viability, INSM1-YCD, INSM1-YCD-YUPRT, and pGL3 basic (mock) transfected cells were treated with increasing concentrations of the prodrug 5-FC. Cell cytotoxicity from treatment was measured, and inhibitory concentration causing 50% reduction in cell viability (IC50) was calculated (Fig. 1A). Significant cytotoxicity was observed in all cell lines after INSM1-YCD/5-FC or INSM1-YCD-YUPRT/5-FC treatment. Importantly, no cell death was observed in mock-transfected cells upon exposure to 5-FC. Highest sensitivity was in general obtained in GLC16 (1,000 μmol/L 5-FC, P < 0.0005), followed by NCI-H69 (1,000 μmol/L 5-FC, P < 0.005) and last by DMS53 (1,000 μmol/L 5-FC, P < 0.01), and in all cell lines the YCD-YUPRT fusion gene rendered cells significantly more sensitive to prodrug treatment than did the YCD gene alone. To investigate whether variations in the sensitivity towards 5-FU influenced the cytotoxic effect of the suicide gene therapy in the cell lines, untransfected GLC16, NCI-H69, and DMS53 cells were exposed to increasing 5-FU concentrations. No significant difference in IC50 values was observed (data not shown) excluding any differential 5-FU sensitivity to influence gene therapy response rates.
To clarify transgene product levels in the cells, the YCD-YUPRT fusion gene was tagged COOH-terminally with the FLAG (DYKDDDDK) epitope. Following transfection with INSM1-YCD-YUPRT.FLAG or mock vector, Western blot analysis was done on total protein lysates of transfected cells (Fig. 1B). In agreement with cytotoxicity data (Fig. 1A), the highest level of transgene expression was detected in GLC16 followed by NCI-H69, whereas only trace levels were detectable in DMS53 cells.
To explore the ability of the YUPRT part of the YCD-YUPRT fusion gene to increase 5-FU toxicity, mock- INSM1-YCD–, and INSM1-YCD-YUPRT–transfected cells were exposed to 5-FU and cell viability evaluated. In NCI-H69 and DMS53 cells, 5-FU treatment was equally toxic to mock, YCD-, and YCD-YUPRT–transfected cells (data not shown). However, in YCD-YUPRT–transfected GLC16 cells, significantly higher cytotoxicity (P < 0.05) was observed upon 5-FU exposure compared with YCD- and mock-transfected cells (Fig. 1C). This shows the capability of the transfected YUPRT transgene to enhance 5-FU toxicity in this cell line.
Although the specificity of the INSM1 promoter previously has been thoroughly investigated (9, 10), the strong effects of YCD-YUPRT/5-FC therapy in SCLC prompted us to confirm that no trace transgene expression in cells of origin other than SCLC would allow for cytotoxicity with the potent YCD-YUPRT/5-FC system. For that purpose, the non-SCLC cell line H1299 and the glioblastoma multiforme cell line U87MG were transiently transfected with the INSM1-YCD-YUPRT vector and exposed to increasing concentrations of 5-FC and 5-FU (Fig. 1D). Both cell lines exhibited transfection efficiencies (60-80%; data not shown) comparably higher than the three SCLC cell lines. Both suicide gene– and mock-transfected cells were highly sensitive to 5-FU treatment (P < 0.01), confirming sensitivity to the toxic end product. However, no cytotoxicity was observed from INSM1-YCD-YUPRT/5-FC therapy, manifesting the high specificity of the gene therapeutic strategy.
INSM1-YCD-YUPRT/5-FC therapy is superior to INSM1-HSVTK/ganciclovir in SCLC cell lines
To assess whether INSM1-YCD-YUPRT/5-FC therapy is superior to INSM1-HSVTK/ganciclovir therapy, a direct comparison of the systems was made in GLC16, NCI-H69, and DMS53 cells (Fig. 2A). The maximum tolerated dose of 5-FC for mock-transfected cells was 1,000 μmol/L (Fig. 1A), whereas a maximum tolerated dose of 10 μmol/L ganciclovir was established forHSVTK therapy (data not shown). YCD-YUPRT/5-FC therapy induced significantly higher cytotoxicity in all tested SCLC cell lines compared with HSVTK/ganciclovir therapy (P < 0.05 or P < 0.005 as indicated). Interestingly, with INSM1-HSVTK/ganciclovir therapy the highest efficacy was obtained in NCI-H69 followed by GLC16, whereas no cytotoxicity could be detected in the DMS53 cell line. To improve HSVTK-based treatment another prodrug analogue, penciclovir, was tested. Penciclovir has been suggested as a more suitable prodrug candidate for HSVTK suicide gene therapy due to lower intrinsic activity and a pronounced safer clinical profile (19, 35). Penciclovir could be applied at a dose of up to 100 μmol/L without affecting mock-transfected cells (results not shown), but did at this concentration cause significantly less cytotoxicity than ganciclovir (P < 0.05) in HSVTK-transfected GLC16 and NCI-H69 cells (Fig. 2A). Clearly, no advantagefrom replacing ganciclovir with penciclovir could be obtained.
We went on to investigate whether cytotoxic bystander effects from HSVTK/ganciclovir treatment was present in treated SCLC cells. When cultured medium from INSM1-HSVTK/ganciclovir-treated NCI-H69 or GLC16 cells was transferred to untransfected parental cells, no reduction in cell viability could be observed in accordance with the notion that phosphorylated ganciclovir compounds are trapped in the producer cells (ref. 6; data not shown). To further investigate GJIC-dependent bystander effects, we investigated the gap junctional status in the cell lines. Gap junctions are made up of connexin proteins in which connexin 43 is the most widely expressed subform (36) and therefore also an established marker for functional GJIC (17). We therefore investigated connexin 43 levels by Western blot analysis carried out on total protein lysates from the cell lines and two diploid nonimmortalized lung fibroblast cell lines, CCD32Lu and CCD19Lu (Fig. 2B). The lung fibroblasts were included to represent normal healthy lung cells, which are known to express significant levels of connexin 43 (37) as also shown in the Western blot analysis. Of the cancer cells, only the NCI-H69 cell line expressed detectable levels of connexin 43, correlating with the fact that highest effect of HSVTK/ganciclovir treatment is obtained in this cell line (Fig. 2A).
INSM1-YCD-YUPRT/5-FC therapy causes a pronounced bystander effect in SCLC cells
In contrast to phosphorylated ganciclovir compounds, 5-FU and downstream toxins are not trapped in the cell and are able to diffuse freely over the cellular membrane. To explore the bystander effect of INSM1-YCD-YUPRT/5-FC gene therapy, cultured medium from treated GLC16, NCI-H69, and DMS53 cells was transferred to untransfected parental cells. Cell viability data showed that cultured medium caused massive cell death to untreated cells (Fig. 3A). To further quantify the bystander effect, INSM1-YCD-YUPRT–transfected cells were mixed with untransfected parental cells in different ratios and exposed to 5-FC (Fig. 3B). Significant bystander cytotoxicity was observed in all cell lines, with the most pronounced effects in GLC16 and NCI-H69. Bystander cytotoxicity increased with increasing 5-FC doses rendering 70% to 100% of GLC16 cells and 30% to 80% of NCI-H69 cells at all mixing ratios sensitive to treatment with 500 μmol/L 5-FC. More modest bystander effects were observed in DMS53; however, when transfected GLC16 cells were mixed with untransfected DMS53 at different ratios, a similar reduction of cell viability was observed in this cell line (Fig. 3C).
No further cytotoxicity is obtained from combining INSM1 promoter–driven HSVTK/ganciclovir and YCD-YUPRT/5-FC therapy in SCLC cells
Although INSM1-HSVTK/ganciclovir therapy clearly is less effective than INSM1-YCD-YUPRT/5-FC therapy in SCLC cell lines, we investigated the combined therapeutic effect of the systems, because others previously have shown synergistic interactions between HSVTK/ganciclovir and CD/5-FC therapy (38, 39). For that purpose NCI-H69 cells were transfected with equal amounts of INSM1-HSVTK and INSM1-YCD-YUPRT vectors followed by concomitant exposure to 5-FC and ganciclovir (Fig. 4). With the use of a fixed ganciclovir concentration of 10 μmol/L (maximum tolerated) it was observed that significant enhancement of cytotoxicity was achieved at 5-FC doses ≤100 μmol/L (P < 0.05). However, no additional cytotoxicity from HSVTK/ganciclovir therapy was obtained by applying higher 5-FC doses. Hence, the advantage from combining HSVTK/ganciclovir and YCD-YUPRT/5-FC treatments was neutralized at optimal 5-FC doses.
INSM1-YCD-YUPRT/5-FC treatment of SCLC xenografts induces significant tumor growth delay
Having observed specific and high therapeutic activity of INSM1-YCD-YUPRT/5-FC therapy in vitro, we proceeded to evaluate the system in vivo on SCLC xenografted tumors. For gene delivery, the liposome-based nanoparticle DOTAP/Chol, which is an established nonviral delivery system in clinical trials for the treatment of non-SCLC (28, 40, 41), was used. For the xenograft model we chose to use the NCI-H69 cell line, because this cell line shows much better growth and propagation properties than GLC16 and DMS53 (data not shown), and hence represents a better model of aggressively growing SCLC. NCI-H69 cells were xenotransplanted in nude mice and treatment was initiated when tumor size had reached 200 to 600 mm3 (Fig. 5A, right) to allow for sustained and very aggressive tumor growth (data not shown) at the initiation of therapy.
Tumor-bearing mice were randomized into three treatment groups, of which the first two groups received DOTAP/Chol encapsulated INSM1-YCD-YUPRT.FLAG or INSM1-LUC (mock control) vector injected i.t. once daily for three consecutive days. The third group was injected i.t. with D5W. From day 1 of treatment, 500 mg/kg of 5-FC was administered once daily i.p. for 10 days to all groups and tumor size was measured daily by caliper. As shown in Fig. 5A and Table 1, significant tumor growth delay of the mice treated with INSM1-YCD-YUPRT/5-FC was observed (P < 0.001) compared with mock- and D5W-treated mice as reflected by a significant increase in tumor doubling time (Td; Table 1). No systemic toxicity was induced from treatments as evaluated by histology of resected major mice organs and mice weight data (results not shown). Tumor growth restraint of the mock-treated tumors was observed compared with the D5W-treated tumors presumably due to an unspecific antitumor effect of lipoplexes as previously reported by others (41–43). Also, mice treated i.t with DOTAP/Chol:INSM1-YCD-YUPRT without administration of 5-FC showed similar tumor growth delay to DOTAP/Chol:INSM1-LUC/5-FC–treated mice (data not shown), emphasizing that the unspecific effect was indeed caused by the lipoplexes and not due to an unspecific effect of the mock plasmid. After eight days, all (7 of 7) D5W-treated mice were withdrawn from treatment due to maximal tumor sizes (1,000 mm3) whereas 3 of 11 (27%) mock-treated mice and 6 of 12 (50%) suicide gene–treated mice survived >10 days. To relate in vivo efficacy to transfection efficacy of DOTAP/Chol:DNA lipoplexes of SCLC tumors, EGFP-N1 or INSM1-YCD-YUPRT.FLAG vectors were encapsulated in DOTAP/Chol and injected i.t. for three consecutive days, and tumor tissue was resected at day 4 for immunohistochemical and fluorescence microscopy analysis of EGFP and YCD-YUPRT (FLAG) expression (Fig. 5B). Distinct areas of EGFP and FLAG expression were found distributed throughout the tumor, confirming the transfection capability of the lipoplex formulation.
|Treatment protocol .||Tumor doubling time Td (d) mean (95% Cl) .||Survivors after 10 d .|
|DOTAP/Chol:INSM1-YCD-YUPRT/5-FC||8.5 (7.8-9.4)||6/12 (50%)|
|DOTAP/Chol:INSM1-LUC/5-FC||6.3 (6-6.6)||3/11 (27%)|
|D5W||4.8 (4.6-5)||0/7 (0%)|
|Treatment protocol .||Tumor doubling time Td (d) mean (95% Cl) .||Survivors after 10 d .|
|DOTAP/Chol:INSM1-YCD-YUPRT/5-FC||8.5 (7.8-9.4)||6/12 (50%)|
|DOTAP/Chol:INSM1-LUC/5-FC||6.3 (6-6.6)||3/11 (27%)|
|D5W||4.8 (4.6-5)||0/7 (0%)|
NOTE: Tumor doubling time (Td) in days as mean (95% confidence interval) calculated from nonlinear regression analysis and number of surviving mice at day 10. There were less mice than tumors because animals were xenotransplanted on both flanks and, therefore, depending on tumor take and growth, had either one or two tumors which entered treatment.
Abbreviation: 95% Cl, 95% confidence interval.
In this study we showed that INSM1 promoter-driven (8, 9) YCD-YUPRT/5-FC therapy induces significant cytotoxicity in SCLC cell lines superior to YCD/5-FC (Fig. 1B) and HSVTK/ganciclovir (Fig. 2A), which until the present study was the only suicide gene strategy tested for SCLC (9–16). Importantly, the INSM1-YCD-YUPRT/5-FC strategy was also investigated in vivo, showing significant tumor growth delay compared with control treatment. This is the first study to describe INSM1-driven YCD-YUPRT/5-FC therapy and the first to investigate the therapeutic potential of delivery of suicide gene therapeutics for SCLC in vivo. In the only previous in vivo study investigating gene therapy for SCLC, cancer cells were adenovirally transduced ex vivo before the injection in mice for xenograft establishment (15).
Because INSM1 promoter activity is increased in all tumors of neuroendocrine origin it is of interest to test the promoter in suicide gene therapy for neuroendocrine tumors other than SCLC. In one very recent study the potential of the INSM1 promoter for regulated HSVTK/ganciclovir therapy for primitive neuroectodermal tumors was shown in vitro and in vivo (44).
The majority of studies describing the CD-UPRT fusion strategy have utilized the E. coli ortholog of CD-UPRT (25–27, 45), although the superiority of the yeast CD compared with the E. coli version has previously been established (21–23). Additionally, only very few studies have focused on the regulated expression of CD-UPRT from cancer-relevant regulatory elements. By far the majority of studies have shown CD-UPRT/5-FC efficacy driven from the constitutive active cytomegalovirus promoter (23–27, 45), although this strategy is not feasible for systemic treatment due to the constitutive activity of the cytomegalovirus promoter in normal tissues. The few attempts to include cancer-regulated expression is represented by a study of paclitaxel-resistant ovarian cancer (45) and a study of prostata cancer using the MDR1 and prostate-specific membrane antigen (PSMA) promoter (46), respectively, for the regulated expression of the E. coli CD-UPRT. However, both the MDR1 and the PMSA gene are expressed in a range of normal healthy tissues (47, 48), which puts into question the safety of exploiting these promoters for suicide gene therapy.
In the present study we show that INSM1-regulated YCD-YUPRT/5-FC therapy is convincingly more effective than HSVTK/ganciclovir therapy in SCLC cell lines (Fig. 2A). In an attempt to improve the HSVTK/ganciclovir effect, we replaced ganciclovir with the prodrug penciclovir, which has been suggested as a more suitable prodrug candidate for HSVTK therapy. Although penciclovir could be applied at higher doses than ganciclovir without affecting mock-treated cells (data not shown), the prodrug induced significantly less cytotoxicity to HSVTK-positive cells compared with ganciclovir (Fig. 2A). The compromised effect of HSVTK/ganciclovir treatment prompted us to investigate if GJIC-dependent bystander effects were compromised in the SCLC cells. Western blot analysis failed to detect expression of the GJIC marker connexin 43 in GLC16 and DMS53 cells, whereas detectable levels of the marker was observed in NCI-H69 (Fig. 2B). The expression of connexin 43 in NCI-H69 correlated with the fact that this cell line showed the highest sensitivity to HSVTK/ganciclovir treatment. To further investigate the expression of connexin 43 and GJIC in SCLC, we tested a large number of SCLC cell lines and xenografts for connexin 43 expression (results not shown). Overall no difference in connexin 43 levels was observed between cell lines and established xenografts, excluding any in vivo related events to influence connexin 43 expression. Only few of the SCLC cell lines and corresponding xenografts had detectable connexin 43 and in these, as for NCI-H69, expression was subtle compared with the diploid lung fibroblast cell lines CCD32Lu and CCD19Lu. The combined results indicate that the potential of HSVTK/ganciclovir therapy may be limited in SCLC due to compromised GJIC-dependent bystander effect in this malignancy.
Because others have shown a synergistic interaction between CD/5-FC and HSVTK/ganciclovir suicide gene therapy (38, 39) we explored the possibility of combining HSVTK/ganciclovir and YCD-YUPRT/5-FC therapy to achieve increased therapeutic effect. Because the NCI-H69 cells showed highest sensitivity towards HSVTK/ganciclovir treatment we initially explored the combined effect in this cell line. However, no advantage from combining systems was achieved at optimal 5-FC doses (Fig. 4). The combination of HSVTK and YCD-YUPRT fusion gene has not previously been tested, but clearly the potent effect of YCD-YUPRT therapy does not benefit from this combination. However, the combination of YCD-YUPRT with other potential suicide gene systems should be investigated to explore synergistic possibilities.
The individual response rates in the cell lines upon INSM1-YCD-YUPRT/5-FC therapy was found to correlate to transgene expression levels (Fig. 1A and B). In GLC16, all cells died from the administered suicide gene therapy. Additionally, increased sensitivity towards 5-FU in YCD-YUPRT–transfected cells was observed in GLC16 (Fig. 1C) but not in NCI-H69 and DMS53 (not shown), presumably due to the higher levels of YCD-YUPRT fusion enzyme in this cell line. Importantly, although only very low transgene expression levels could be detected in DMS53, >50% of cells succumbed to YCD-YUPRT/5-FC therapy in this cell line (Fig. 1B). Because DMS53 was equally sensitive to 5-FU treatment as GLC16 and NCI-H69 (data not shown), the lack of response in this cell line was directly correlated to low transgene expression level and not related to treatment resistance. Furthermore, we showed that bystander effects from treated GLC16 cells effectively targeted DMS53 cells, showing that low-transgene-expressing cells can be targeted by exposure to high doses of bystander toxins from high-transgene-expressing cells (Fig. 3A-C). Taken together these data show that highly effective therapy can be obtained with INSM1-driven YCD-YUPRT/5-FC therapy in SCLC. Importantly, effective therapy can be obtained in spite of putative variations in the transcriptional activity from the INSM1 promoter, which could be expected to exist in a heterogeneous cancer cell population.
The clinical relevance of the strategy was confirmed in vivo (Fig. 5A). Despite very aggressive tumor growth, DOTAP/Chol:INSM1-YCD-YUPRT/5-FC treatment caused significant tumor growth attenuation compared with control-treated animals, showing proof of concept that a specific antitumor effect can be obtained by delivery of suicide gene therapeutics to SCLC in vivo. The DOTAP/Chol:DNA lipoplexes were shown to convey small foci of transgene expression throughout the tumor (Fig. 5B). However, the distinct areas and low staining intensity of transgene expression in the tumor tissue, as especially shown from FLAG detection as seen in Fig. 5B from INSM1-YCD-YUPRT.FLAG treatment, strongly indicate that a higher treatment effect would be obtained by increasing delivery efficiency and thereby transgene expression. Significant antitumor effects of DOTAP/Chol-mediated gene delivery have been shown in non-SCLC preclinical models (40, 41, 43) and patients (28); however, no studies other than the present have tested the DOTAP/Chol formulation for SCLC, and other delivery formulations might be more suitable for this malignancy.
It is likely that bystander effects influence the positive treatment response when considering the in vitro results (Fig. 3) and the low level and spread of transgene expression (Fig. 5B). However, this remains speculative. To further quantify this effect we attempted to compare suicide transgene expression with cell death using terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) and FLAG staining on consecutive sections of tumor tissue. Due to the aggressively growing properties of tumors and the fact that treatment was done with direct i.t. injections of DOTAP/Chol:DNA lipoplexes causing general toxicity, resected SCLC tumors had large regions of necrosis and apoptosis as evaluated by H&E and TUNEL staining (results not shown), which hindered a thorough evaluation of bystander effects. However, a quantitative bystander model could be made by mixing stable INSM1-YCD-YUPRT.FLAG–expressing cells with untransfected parental cells in different ratios for xenograft establishment. Treatment of small-size mixed xenografts with 5-FC followed by evaluation of FLAG and TUNEL staining on resected tissue might clarify the level of bystander influence on antitumor response in vivo.
It would be highly relevant to further test the presented treatment strategy in SCLC preclinical models that mimic the metastatic nature of SCLC and hereby develop systemic treatment formulations and schedules which could be adapted into the clinic. Orthotopic disseminated models have successfully been developed for non-SCLC (40, 41, 43), and treatment with DOTAP/Chol delivered gene therapeutics in such models have resulted in dramatic antitumor responses due to the distribution of the DOTAP/Chol lipoplexes in particular to intrathoraic organ sites. Kuo et al. (49) have shown seeding of SCLC cells in the lung, heart, and liver after i.v. injection of SCLC cells. However, in a later study of Moreira et al. (50) no seeding of SCLC cells could be detected in lung or other organs even after several attempts of orthotopic reconstitution, and to our knowledge no other studies have been carried out with orthotopic SCLC models. Thus, further clarification of the establishment of orthotopic disseminated models for SCLC is needed.
The results of the present study yield promise for the future development of YCD-YUPRT/5-FC suicide gene therapy towards the design of a clinical protocol for the treatment of SCLC.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
We thank Pia Pedersen for technical assistance and Birgit Guldhammer Skov for consulting in pathological examination of tissues from in vivo experiments.
Grant Support: Grants from the University of Copenhagen, the Danish Cancer Society, the Novo Nordisk Foundation, Aase and Ejnar Danielsens Foundation and VFK Krebsforschung gGmbH and the National Institute of Health (NIH P50 CA 70907 SPORE in Lung Cancer).