Replicase-based plasmid DNA shows anti-tumor activity

Plasmid pCMV-β was from the American Type Culture Collection (ATCC, Manassas, VA). The pSIN-β plasmid was constructed following a previously described method [23]. The pSIN-β-Δnsp was constructed in two steps. First, the pSIN-β was digested with Pst I (Invitrogen, Carlsbad, CA), and the resultant fragment was gel extracted and purified using a PureLink Gel Extraction kit (Invitrogen). The DNA fragment was further digested with Hind III (Invitrogen). The correct fragment was gel extracted, and the adhesive ends were ligated using T4 DNA ligase (Invitrogen). All plasmids were amplified in E. coli DH5α under selective growth conditions.

Plasmid DNA was methylated at CpG sites with CpG methyl transferase (M.SssI) (New England BioLabs, Beverly, MA). The M.SssI methylates at the carbon position 5 of cytosine residues within double stranded recognition sequence. Methylation reaction containing 2 U of methylase per μg of DNA was incubated at 37°C for at least 3 h. The extent of methylation by the M.SssI was determined using a BstU I endonuclease assay (Invitrogen). Plasmid was purified from bacteria using a QIAGEN midiprep kit (Valencia, CA). Large scale plasmid preparation was performed by GenScript (Piscataway, NJ).

Mouse lung tumor cells (TC-1, ATCC, CRL-2785) and mouse melanoma cells (B16-F10, ATCC, CRL-6475) were cultured in RPMI 1640 medium (Invitrogen) and DMEM medium (Invitrogen), respectively. The media were supplemented with 10% fetal bovine serum (FBS, Invitrogen), 100 U/ml of penicillin (Invitrogen), and 100 μg/ml of streptomycin (Invitrogen).

The ovalbumin (OVA)-expressing B16-OVA cell line was generously provided by Dr. Edith M. Lord and Dr. John Frelinger (University of Rochester Medical Center, Rochester, NY) [36]. B16-OVA cells were cultured in RPMI 1640 medium supplemented with 5% FBS and 400 μg/ml of G418 (Sigma).

TC-1 cells were seeded in 24 or 48 well plates (20 000 cells/well) and incubated at 37°C, 5% CO₂ for 24 h or until 60% confluency followed by transfection using plasmid DNA (0.15 or 0.40 μg as where mentioned) complexed with Lipofectamine^® (Invitrogen) following the manufacturer's instruction. The transfection medium was replaced with fresh medium 3 h later.

Total RNA was isolated from TC-1 cells (1 × 10⁷) transfected with plasmid using a QIAGEN RNeasy mini kit. On-column DNase digestion was performed using RNase-free DNase set (QIAGEN) to eliminate DNA contamination. The RNA quality was assessed using the OD260/OD280 ratio.

The presence of dsRNA in TC-1 cells (n = 3) transfected with the plasmid was confirmed using ELISA as previously described with modification [37]. Briefly, 96-well plates were coated at 4°C overnight with 1 μg total RNA diluted in PBS. Plates were washed with PBS/Tween 20 (10 mM, pH 7.4, 0.05% Tween 20, Sigma-Aldrich, St. Louis, MO) and blocked with 4% (w/v) bovine serum albumin (BSA, Sigma-Aldrich) in PBS/Tween 20 for 1 h at 37°C. Plates were washed again with PBS/Tween 20. Monoclonal anti-dsRNA antibody J2 (English & Scientific Consulting Bt. Szirák, Hungary) was added to each well following the removal of the blocking solution. The plates were incubated for an additional 3 h at 37°C. Horseradish peroxidase (HRP) labeled goat anti-mouse IgG2a (5 000-fold dilution Southern Biotechnology Associates, Birmingham, AL) was added to the wells, followed by 1 h of incubation at 37°C. The presence of bound secondary antibody was detected after a 30 min incubation with 3,3',5,5'-tetramethylbenzidine substrate (TMB) (Sigma-Aldrich). The reaction was stopped by the addition of sulfuric acid (0.2 M, Sigma).

The number of viable TC-1 cells was determined using a 3-(4,5-dimethylthiazol)-2-,5-diphenyltetrazolium bromide (MTT) kit (Sigma-Aldrich) 24, 48, and 72 h after the initiation of the transfection (n = 4) [12]. Cells treated with sterile PBS were used as a control. Formula used to calculate the relative cell number (%) was: Relative cell number = 100 × number of live cells transfected with pCMV-β (pSIN-β, or pSIN-β-Δnsp)/number of live cells treated with sterile PBS.

Cationic liposomes were prepared using cholesterol (Sigma-Aldrich), egg phosphatidylcholine (Avanti Polar Lipids, Inc, Alabaster, AL), and 1,2,-dioleoyl-3-trimethylamonium-propane (DOTAP, Avanti) at a molar ratio of 4.6:10.8:12.9 by thin film hydration method followed by membrane extrusion (1, 0.4, and 0.1 μm, sequentially) [38]. The final concentration of DOTAP in the liposome was 10 mg/ml. The plasmid-liposome lipoplexes were prepared by mixing equal volumes of plasmid DNA (25 μg in 25 μl) solution and liposome suspension containing 50 μg of DOTAP liposomes. The mixture was allowed to stay at room temperature for at least 15 min before further use. Particle size was measured using a Malvern Zetasizer Nano ZS (Worcestershire, United Kingdom). The size of the liposomes was 110 ± 0.6 nm with a polydispersity index (PI) of 0.121. The pSIN-β-liposome lipoplexes were 255 ± 31 nm (PI, 0.177). The pCMV-β-liposome lipoplexes were 249 ± 33 nm (PI, 0.183). The sizes of the two lipoplexes were not statistically different (p = 0.83, t-test, n = 3).

All animal studies were carried out following the National Institutes of Health animal use and care guidelines. Animal protocol was approved by the Institutional Animal Care and Use Committee at the University of Texas at Austin. Female C57BL/6 mice (6-8 weeks) were from Simonsen Laboratories (Gilroy, CA) or Charles River laboratories, Inc. (Wilmington, MA). Female athymic nude mice (6-8 weeks) were from Charles River laboratories. Mice were subcutaneously injected with TC-1, B16/F10, or B16-OVA cells (5 × 10⁵) in the right flank. When tumors reached an average diameter of 3-4 mm, the plasmid DNA-liposome lipoplexes were injected subcutaneously peritumorally (s.c., p.t.) for 5 or 10 consecutive days [12, 15, 19]. The dose of the plasmid DNA was 25 μg DNA per mouse per injection. Tumor size was measured using a digital caliper and calculated using the following equation [39]: tumor diameter = (Length + Width)/2. To examine whether the nsp genes were expressed in vivo, pCMV-β, pSIN-β, or pSIN-β-Δnsp (25 μg) was injected into the gastrocnemius muscles in the hind legs of mice (n = 2). After 24 h, the injected muscle tissues were collected and homogenized using TRIzol reagent (Invitrogen) to isolate total RNA. RT-PCR was performed to amplify nsp4 gene or β-gal gene using the nsp4-1 primers or the β-gal primers (5'-GACGTCTCGTTGCTGCATAA-3'; 5'-CAGCAGCAGACCATTTTCAA-3').

TC-1 tumors in mice that were treated for 6 consecutive days with plasmids were collected, fixed in formaldehyde, embedded in paraffin, and sectioned. Immunohistochemistry was performed to detect apoptosis using the anti-ACTIVE caspase-3 antibody (Promega, Madison, WI) according to manufacturer protocol. Fifteen random fields per sample at 40 × magnification were scored for cleaved caspase-3. Apoptotic index was determined based on the % of cleaved caspase-3 positive cells found within total cells counted [12].

Mice were subcutaneously injected with 125 μg of plasmid DNA in lipoplexes (DNA/liposomes, 1:2, w/w). Ten h later, serum was collected, and the concentration of IFN-α was determined using a mouse IFN-α (Mu-IFN-α) ELISA kit (PBL Biomedical Laboratories, Piscataway, NJ).

Statistical analyses were completed using ANOVA followed by Fisher's protected least significant difference procedure. A p-value of < 0.05 (2-tail) was considered statistically significant.

Shown in Figure 1 are maps of the plasmids used. In the pCMV-β, the β-galactosidase gene is driven by the CMV promoter. In the pSIN-β, the nsp1-4 genes are driven by a CMV promoter, while the β-galactosidase gene is driven by a sindbis viral subgenomic promoter. The pSIN-β-Δnsp was constructed by deleting the nsp1-3 and part of the nsp4 genes from the pSIN-β. To confirm that the pSIN-β plasmid can produce dsRNA when delivered into cells, TC-1 cells were transfected with pSIN-β, and the total RNA was extracted from the cells 24 h later. The total RNA was reverse transcribed with either oligo dT primers or nsp4 gene specific primers (p4-F or p4-R). The cDNA was then amplified with nsp4-specific primers. A 501 bp nsp4 gene fragment was observed in all samples transfected with pSIN-β, but not in cells transfected with the pCMV-β (Figure 2A), indicating the presence of both sense and anti-sense RNA of the nsp4 gene in cells transfected with the pSIN-β. The production of dsRNA in cells transfected with the pSIN-β was further confirmed using ELISA. As shown in Figure 2B, TC-1 cells transfected with pSIN-β had an elevated level of dsRNA compared to untransfected cells, whereas cells transfected with pCMV-β and the untransfected cells had a similar level of dsRNA. The dsRNA generated within cells transfected with the pSIN-β seemed to be functional because the number of live cells in samples transfected with pSIN-β decreased gradually with the increase in incubation time, in contrast to the increase in the number of live cells in samples transfected with the pCMV-β (Figure 2C). As expected, the pSIN-β-Δnsp no longer caused a decrease in the number of live cells when transfected into the TC-1 tumor cells (Figure 2C), indicating the significance of the nsp1-4 genes for the pSIN-β plasmid to be functional. This observation is in agreement with the finding by Leitner et al. (2004), who showed that the survival of BHK-21 cells transfected with a replicase-based plasmid was significantly lower than cells transfected with a conventional CMV promoter-driven plasmid [40]. The cell death after transfection with the pSIN-β was likely caused by the pro-apoptotic dsRNA produced by the sindbis RNA replicase complex [40]

Prior to carrying out tumor treatment studies, the ability of the pSIN-β to express the nsp1-4 genes in vivo was examined. As shown in Figure 3 using RT-PCR, nsp4 RNA expression was detected only in mouse muscle tissues injected with the pSIN-β, not in the ones injected with the pCMV-β or the pSIN-β-Δnsp, demonstrating that only the pSIN-β was capable of expressing the nsp genes in vivo. The β-gal mRNA was present in all the mouse muscle tissues since the β-gal gene is endogenous (Figure 3).

To evaluate the extent to which the pSIN-β can control the growth of tumors pre-established in mice, mouse TC-1 lung cancer cells were seeded in mice. When tumors reached 3-4 mm in diameter, mice were treated with pSIN-β or pCMV-β daily for 10 days. Mice in the negative control group were not treated. As shown in Figure 4A, TC-1 tumors grew significantly slower in mice that received the pSIN-β plasmid than in mice that received the pCMV-β plasmid. In fact, 25 days after cell seeding, only 20% or 40% of tumor-bearing mice that were left untreated or received the pCMV-β plasmid, respectively, were alive, but all mice that received the pSIN-β were still alive (Table 1). Moreover, on day 25, there was only one mouse in the group that received the pSIN-β plasmid had a tumor of 3.1 mm in diameter, which completely regressed on day 37 (Table 1). Clearly, the pSIN-β plasmid was more effective in controlling the growth of the TC-1 tumors than the pCMV-β plasmid. Finally, mice in the negative control group were left untreated because it was shown that repeated peritumoral injection of sterile PBS or the liposomes did not have any effect on the growth of the TC-1 tumors as compared to mice left untreated (Figure 4B), demonstrating that potential inflammations caused by the liposomes alone or by the peritumoral injections per se were not responsible for the anti-tumor activity observed in Figure 4A.

To understand whether the sindbis replicase genes, nsp1-4, were related to the anti-tumor activity of the pSIN-β plasmid, the anti-tumor activity of the pSIN-β-Δnsp was compared to that of the pSIN-β. When used to treat the TC-1 tumors in mice, the pSIN-β-Δnsp was significantly less effective in controlling the growth of the tumors than the pSIN-β in the beginning (Figure 5), demonstrating that the nsp1-4 genes, which were responsible for the dsRNA production, played a significant role in the anti-tumor activity of the pSIN-β plasmid. It needs to be noted that the TC-1 tumor cells are strongly immunogenic, and it was expected the peritumoral injection of pSIN-β-Δnsp to show anti-tumor activity because the plasmid, with CpG motifs, can activate innate anti-tumor immune responses [41].

As shown in Figure 6 more cells in tumors treated with the pSIN-β plasmid underwent apoptosis than in tumors treated with the pCMV-β plasmid. We suspect that the increased apoptosis in tumors that received the pSIN-β plasmid was related to the plasmid's ability to produce dsRNA in transfected cells. However, it is unclear to what extent the apoptosis was caused directly by dsRNA produced by the pSIN-β plasmid. Double stranded RNA is pro-apoptotic [2], but the type I IFNs induced by dsRNA are pro-apoptotic as well [2]. Moreover, the unmethylated CpG motifs on the plasmid, the dsRNA per se, and type I IFN are all known to be able to activate innate immunity such as natural killer (NK) cells, which can cause tumor death [42]. In fact, it was shown that subcutaneous injection of the pSIN-β plasmid induced an elevated level of IFN-α in mouse serum samples (Figure 7).

TC-1 tumor cells are highly immunogenic in C57BL/6 mice due to the human papillomavirus E6 an E7 genes in the TC-1 cells [43]. Data from several recent studies have shown that tumor cells with intracellular dsRNA were more immunogenic than tumor cells physically mixed with dsRNA [19‐22]. Therefore, it was expected that adaptive immune responses have contributed, to a certain extent, to the anti-tumor activity from the pSIN-β plasmid. To test this hypothesis, the same TC-1 tumors established in athymic mice were treated with pSIN-β or pCMV-β. As shown in Figure 8 the pSIN-β was no longer more effective than the pCMV-β in controlling the growth of the TC-1 cells, indicating that adaptive immunity contributed to the anti-tumor activity from the pSIN-β plasmid. However, the adaptive immunity was not absolutely required for the pSIN-β to have anti-tumor activity because recent preliminary data in our lab showed that in athymic mice, the pSIN-β caused total regression of pre-established model human tumors when targeted into the tumor cells using a tumor-specific ligand (Rodriguez and Cui, unpublished data).

The anti-tumor effect from the pCMV-β plasmid was likely due to the unmethylated CpG motifs present on the plasmid [41]. As shown in Figure 9 methylation of the pCMV-β depleted the plasmid's ability to inhibit the growth of the TC-1 tumor cells in mice. This is in agreement with a previous report showing that plasmid DNA itself had anti-tumor activity because the unmethylated CpG motifs on the plasmid can activate innate immunity [41]. Therefore, it is possible that both dsRNA produced by the RNA replicase complex encoded by the nsp1-4 genes and the unmethylated CpG motifs on the pSIN-β plasmid may have contributed to the anti-tumor activity from the pSIN-β.

To test whether the pSIN-β was effective against tumors other than the TC-1, mice with pre-established B16-F10 or B16-OVA tumors were treated similarly. As shown in Figure 10 the pSIN-β plasmid significantly controlled the growth of both poorly immunogenic B16-F10 tumors (Figure 10A) and the more immunogenic B16-OVA tumors (Figure 10B), indicating that the approach of controlling tumor growth with the replicase-based plasmid was not limited to the TC-1 tumors and likely not limited to highly immunogenic tumors as well. Again, it is not surprising that the pCMV-β also showed anti-tumor activity against the B16 melanoma. In a previous study, McCray et al. (2006) showed that intratumoral injection of an empty pcDNA3.1 delayed the growth of B16 tumors, as compared to the injection of saline [44]. It was also shown that intratumoral injection of the pcDNA3.1 followed by in vivo electroporation further improved the anti-tumor activity [44]. In the present study, the repeated peritumoral injection of the pCMV-β complexed with cationic liposomes may have improved the non-specific anti-tumor activity from the plasmid.