Acridine-modified, clamp-forming antisense oligonucleotides synergize with cisplatin to inhibit c-Myc expression and B16-F0 tumor progression

Abstract

The c-myc protooncogene plays a key role in the abnormal growth regulation of melanoma cells. We have targeted three polypurine sequences within the mouse myc mRNA with acridine-modified, clamp-forming antisense oligonucleotides (AS ODNs) in an effort to inhibit growth of murine melanoma cells. These ODNs are unique in that they hybridize to the target mRNA by both Watson–Crick and Hoogsteen hydrogen bond interactions, forming a triple-stranded structure. At a concentration of 3 µM E1C, E2C and E3C inhibit B16-F0 proliferation by 76, 66 and 78%, respectively. Both immunofluorescent staining and western blot analysis corroborate a proportional reduction in c-Myc expression by all three ODNs. There were clear distinctions in the ability of these ODNs to inhibit tumor progression in C57BL/6 mice as a function of Myc expression. There was no synergy demonstrated between ODN E1C with cisplatin (DDP), which inhibited tumor growth by 77% alone and 82% in combination. Although E2C inhibited growth by 54%, its effect was decreased to 32% with DDP, when compared with controls. E3C, on the other hand, demonstrated a synergistic effect with DDP, inhibiting growth by 72% in combination, but only by 1% as a single agent. Immunofluorescence analysis of tumors for each group revealed a concomitant reduction in c-Myc expression in tumors from mice treated with the most active clamp ODN alone (E1C) or clamp ODN + DDP (E1C/E3C + DDP). Western blot analysis confirmed this decrease in target protein expression. Our results document the growth-inhibitory activity of two myc-targeting antisense clamp ODNs; E1C, which has activity as a single agent, and E3C, which has in vivo synergy with DDP pretreatment. These data confirm the antiproliferative effects of these novel ODNs and document an interesting synergy with the chemotherapeutic agent DDP.

INTRODUCTION

Many studies have documented c-myc overexpression in a variety of cancer cell types (1–5), both at the mRNA and protein levels. Almost all of these studies found a correlation between the protein expression level and metastatic potential (2–5). Overexpression of the c-myc protooncogene has been thought to act as a causal factor in the aberrant proliferation of cancer cells (6,7). In addition to roles governing cellular growth and differentiation, myc also contributes to the signal transduction cascades that regulate apoptosis (8). It is classified as an ‘immediate early response’ gene, because it functions in several aspects of cell cycle control (8). Because of the large body of data implicating c-myc in abnormal cancer cell growth, we sought to utilize myc as a target in our efforts to revert the malignant melanoma phenotype using modified antisense oligonucleotides (AS ODNs) (9).

In part because of positive results in animal models, efforts continue to explore the therapeutic potential of AS ODNs in treating cancers, inflammatory diseases and viral infections (10–13). A number of these agents have entered clinical trials for antineoplastic and anti-HIV therapies (14). Indeed, over the past 10 years, AS ODNs have been used in a wide variety of tissue culture and in vivo models (2–7,15–19). Modified AS analogs, in particular phophorothioates (PS), the most frequently studied, in which one of the non-bridging oxygen atoms is substituted by a sulfur atom in the phosphate backbone (2,3,6,7,15–19), have shown the most promise at inhibiting gene expression (8–11,14). We utilized PS analogs in this study because they also have been demonstrated to possess increased serum stability and wide tissue distribution (20,21). Our goal was to mimic human melanoma in a mouse model, so we used an ODN prototype that had been studied extensively.

AS-mediated inhibition of gene expression depends on the ability of ODNs to specifically interact with target mRNAs by Watson–Crick base pairing (22). The resulting DNA–mRNA hybrid structure serves as a cleavage substrate for RNase H (23). We have added an additional ODN fragment in our AS design. This modification enables ODNs to target the mRNA by triplex-like interactions giving rise to an enhanced sequence-specific mechanism (6,9,23,24). Traditionally, triplex-forming ODNs have been designed to inhibit gene expression at the level of transcription (25). Unfortunately, their success has been hampered by the inaccessibility of target sequences within the chromatin structure of the nucleus (26). We have attempted to use this means of inhibiting gene expression at the translational level in which the target is single-stranded and more easily accessible (27). We also chose this approach because it is known that when targeting certain sequences, normal ODNs do not hybridize with adequate affinity to generate the desired biological effect of blocking the activity of ribosomes or polymerases (26). Several groups have studied triplex formation with single-stranded purine strands (28–33). We have used this approach to target three different polypurine sequences in the second and third exons of the mouse c-myc primary transcript with ‘clamp-forming’ polypyrimidine AS ODNs. The exon 2 target is an 11 bp sequence located at +2617–2627 from the translation initiation start site. The two exon 3 targets are much closer to the AUG codon, at +34–43 (target sequence of clamp ODN E3C) and +60–69 (target sequence of clamp ODN E1C). The first exon 3 target is of particular interest because it has identical sequence homology with the human c-myc mRNA. An important point is that triple helical structures formed between single-stranded purine RNA targets and pyrimidine clamp or circular ODNs are stable at neutral pH (31,33). We also chose clamp ODNs as opposed to standard Watson–Crick AS ODNs because of their higher binding affinities (23).

A third modification of the ODNs used in these experiments was covalent conjugation of an intercalating acridine (Acr) moiety to the 5′ end, which increases the irreversibility (23) of the interaction, augments protection against exonuclease degradation and adds binding energy to better stabilize the complex (34). Because we were studying the efficacy of these ODNs in a melanoma system, it was also rational to test for possible synergy using a ‘standard’ chemotherapy drug. Cisplatin (DDP) is used clinically as a chemotherapeutic agent to treat malignant melanoma (35) and was chosen for our study based on previous work documenting a promising cooperation between antisense agents and DDP (2,36–38).

Numerous studies have documented successful inhibition of tumor cell growth after using AS ODNs targeting c-myc, in both in vitro and in vivo models (39–41). Here, we report positive results, using clamp-forming AS ODNs, in combination with cisplatin, targeted to three polypurine sequences within the mouse c-myc mRNA in a cell culture and whole mouse model system. After treatment with either ODN alone (3 µM) or following treatment with DDP (10 µM), B16-F0 mouse melanoma cellular growth was inhibited by 66–78% with either ODN alone. In the animals, only ODN E1C significantly slowed tumor progression alone, while only E3C worked synergistically with DDP to inhibit growth. Immunofluorescent and western blot detection of Myc demonstrated a reduction in expression after both cell culture and tissue administration. Further analysis, by TdT-mediated dUTP nick end-labeling (TUNEL) assay and proliferating cell nuclear antigen (PCNA) detection, confirmed an increased incidence of apoptosis in tumors treated with active ODNs compared with those whose treatments had little or no effect on growth. To the best of our knowledge, this work represents the first report of using clamp-forming AS ODNs to target coding region sequences in the murine c-myc transcript. We have successfully demonstrated inhibition of cellular and tumor growth with a concomitant reduction in target protein, by a proposed apoptotic mechanism, in a syngeneic mouse model system.

MATERIALS AND METHODS

Cell culture

B16-F0 is an established murine melanoma cell line obtained from the American Type Culture Collection (CRL-6322_FL). Cells were routinely passaged in Dulbecco’s modified minimal essential medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum and 1% Antibiotic–Antimycotic (10 000 IU/ml penicillin G, 10 000 µg/ml streptomycin sulfate, 25 µg/ml amphotericin B in 0.85% saline) (Gibco BRL). Cells were grown at 37°C in a water-jacketed CO₂ incubator (Nuaire).

Oligonucleotide synthesis

Four Acr-modified, CT-rich phosphorothioate ODNs were prepared in an Applied Biosystems 394 automated DNA/RNA synthesizer using standard phosphoramidite chemistry and purchased from Oligos Etc. Three ODNs (E1C, E2C and E3C) targeted exons 2 (E2C) and 3 (E1C and E3C) of the c-myc mRNA, while the other ODN (SCR) served as a control. The sequences were as follows: E1C, X-CCTCTTCTCCTTTTCCTCTTCTCC; E2C, X-TTCTCTTCCTCTTTTCTCCTTCTCTT; E3C, X-TTTCTTCCTCTTTTCTCCTTCTTT; SCR, X-CTCTCTCTCTCTTTTTCTCTCTCTCTCT; where X is Acr [1-dimethoxytrityloxy-2-(N-acridinyl-4-aminobutyl)-propyl-3- O-(2-cyanoethyl)-(N,N-diisopropyl)-phosphoramidite].

MTT cellular proliferation analysis

B16-F0 cells were seeded at 10² cells/well in 96-well plates in triplicate (day 1) and incubated for 48 h at 37°C. On days 3, 6 and 9 cells were treated with either the combination of a 10 µM dose of DDP for 1 h then transiently transfected with 3 µM ODN or ODN alone in DMEM (Gibco BRL). On day 10, 10 µl of MTT reagent (Sigma) was added to each sample for 4 h, then 100 µl of solubilization solution (0.01 HCl in 10% SDS) was added for 18 h. On day 11, plates were analyzed on a microplate reader at 570 nm.

Immunofluorescent detection of cellular c-Myc expression

B16-F0 cells were seeded at a density of 10² cells/well in 48-well plates (day 1), in duplicate, and incubated for 48 h at 37°C. On day 3, cells were treated according to the previously outlined transfection procedure for the MTT assay, with one modification. On day 9, 60 min after final treatment, cells were fixed with cold 4% paraformaldehyde (PFA) for 20 min on ice. The cells were rinsed three times with phosphate-buffered saline (PBS) and incubated in PBS containing 0.2% Triton-X100 for 15 min. They were then blocked for 1 h at room temperature in 5% goat serum and immunostained with the mouse c-Myc c8 monoclonal antibody (Santa Cruz Biotechnology) (1:100) with 2% goat serum overnight at 4°C. After washing in 1× PBS, cells were incubated with an Alexa488 anti-mouse fluorescein (FITC)-conjugated secondary antibody (Molecular Probes) (1:500) for 45 min in the dark. Immunofluorescent staining was observed and photographed on an Olympus inverted microscope, with the proper fluorescence filters.

Western blot quantitation of cellular c-Myc and β-actin expression

B16-F0 cells were seeded at 10⁵ in 60 mm culture plates. Cells were treated according to the same transfection regimen, with one modification. Cell lysates were harvested 1 h after final treatment (day 9), with trypsin–EDTA (1:10 in Hank’s balanced salt solution), centrifuged and resuspended in 100 µl 1× lysis buffer (5X; Promega), incubated for 30 min, recentrifuged and the supernatant was transferred to a fresh tube. Total protein (25 µg) lysate for each treatment group was loaded with SDS gel loading buffer and fractionated by 8.5% SDS–PAGE and the gel was transferred to a PDVF membrane. Western blot analysis was performed according to the procedure provided with the same primary antibody. Briefly, the membrane was blocked at room temperature for 1 h in a 5% non-fat milk powder solution in PBS containing 0.05% Tween-20 (PBS-T) (Sigma) to reduce non-specific reactivity. The blot was then washed three times for 5 min again with PBS-T alone. It was then incubated for 1 h at room temperature in PBS-T containing a 1:500 dilution of mouse-reactive c-Myc c8 and β-actin C-2 monoclonal antibodies (Santa Cruz Biotechnology). After three additional 5 min washes, the blot was incubated for 45 min with a 1:5000 dilution of horseradish peroxidase (HRP)-conjugated goat anti-mouse secondary antibody (Santa Cruz Biotechnology). The blot was again washed three times in PBS-T then once in PBS alone for 5 min. Protein detection was performed using ECL chemiluminescent reagents (Amersham), followed by rapid autoradiography at room temperature. Densitometric analysis was used to determine the relative percentage of reduced c-Myc expression compared with the untreated control.

Tumor section preparation

B16-F0 cells were scaled up and harvested in 1× sterile PBS. On day 1, 2.5 × 10⁶ cells were injected into the right hind flank of 20–22 g C57BL/6 mice (total nine mice/group, five mice for E2C). The mice were treated on day 4 with either the combination of 3.3 mg/kg DDP i.p. (American Pharmaceutical Partners) for 30 min followed by a 3.3 mg/kg dose of each ODN/group (intra-tumoral) or with each ODN alone (E1C, E2C, E3C or SCR). These treatments were repeated on days 7 and 10 for a total of three treatments/mouse. On day 11, the experiment was ended and tumors from eight mice/group were excised for photography and western blot analysis. The other mouse in each group underwent full-body perfusion with 4% PFA and the tumor was removed and prepared for cryostat and immunofluorescence analysis. Tumors removed from perfused animals were immediately placed in 4% PFA and stored overnight at 4°C. The following day, they were washed twice with cold, sterile 1× PBS and transferred to a 15% sucrose solution for 24 h, again at 4°C. The tumors were prepared for sectioning by first being placed in an equal volume solution of 15% sucrose and OCT compound (Tissue-Tek) for 5 min, transferred to OCT alone for 5 min, then placed in fresh OCT and frozen on dry ice. Cryostat sections (30 µ) were made on a Leica CM3050, stabilized on glass slides and stored at –80°C for staining.

In vivo immunofluorescent detection of c-Myc expression

Slides containing tumor sections were thawed for 1 h prior to staining at room temperature. After one brief wash in 1× PBS, slides were blocked with 5% goat serum in 1× PBS for 1 h at room temperature. Slides were again quickly rinsed in 1× PBS, then incubated overnight at 4°C with polyclonal c-Myc (N-262) at 1:500 and 1.5% goat serum in a humidity chamber. The following day, the humidity chamber was allowed to sit at room temperature for 30 min before opening, then the slides were washed three times for 5 min in 1× PBS. Next, they were incubated for 45 min in the dark with an Alexa488 anti-rabbit FITC-conjugated secondary antibody at a 1:1000 dilution plus 1.5% goat serum. Following two final 5 min washes the slides were mounted in Mowiol 4 medium for photography.

In vivo western blot quantitation of c-Myc, α-actin and PCNA expression

For each treatment group, 0.1 g tumor tissue was weighed out and flash frozen in liquid nitrogen, then ground with a sterile mortar and pestle. Each resultant tissue powder was dissolved in 150 µl 1× lysis buffer, vigorously vortexed for 1 min, then incubated on ice for 20 min. After brief vortexing, the samples were centrifuged at 5000 g for 3 min at 4°C and the supernatants were transferred to fresh tubes. Protein quantification was performed according to the procedures outlined with the Bradford Reagent Kit (Bio-Rad), using bovine serum albumin (2 mg/ml) as the standard. Total protein (3.0 µg) from each sample was resolved by 8.5% SDS–PAGE and western blot analysis was carried out as previously outlined, with the exception of using the c-Myc (N-262; Santa Cruz Biotechnology) and α-actin (C-11; Sigma) rabbit polyclonal antibodies at 1:1500 and a goat anti-rabbit HRP-conjugated secondary antibody at 1:3000. Additionally, to confirm TUNEL assay results, the blot was stripped using commercially prepared stripping buffer (Pierce) for 15 min at room temperature and reincubated with PCNA (FL-261) rabbit polyclonal antibody (Santa Cruz Biotechnology) at the same dilutions. Data for the E2C-treated tumors are not shown in Figure 7 because they were considered to show an insignificant level of inhibition by antitumor efficacy and immunofluorescence analysis.

Figure 7.

Western blot analysis of B16-F0 excised tumors treated with clamp ODNs and DDP. Tumor lysates were prepared as described in Materials and Methods and 3 µg total protein was analyzed with the c-Myc and α-actin rabbit polyclonal antibodies (Santa Cruz Biotechnology and Sigma). Lanes are labeled with appropriate animal treatments (i.e. lane 1, CTL, is untreated). Western blot detection verified immunofluorescence results that in vivo E1C treatment almost completely inhibits Myc protein expression in the absence of DDP (D), by >90%, and still significantly with the drug, at 80%. Tumor lysates from mice treated with E3C alone or combined with drug again showed a reduced level of target protein expression, by 20 and 95%, respectively. There was no change in Myc expression after either SCR treatment or with DDP alone compared with the untreated control expression. Also, no significant changes in α-actin expression were observed. To confirm the TUNEL assay results, PCNA expression was evaluated by western blot analysis. Again, only the most active treatments significantly reduced PCNA; by 75% with E1C, 70% with E1C + DDP and 65% with E3C + DDP. No significant changes in PCNA expression resulted after DDP, E3C, SCR or SCR + DDP administration, as compared with the untreated CTL.

In vivo TUNEL detection

Cryopreserved tumor sections taken from the animal experiment were analyzed using a Fluorescein-based In Situ Cell Death Detection Kit (Roche) using the standard protocol for tissue prepared in this manner. Briefly, each slide was rinsed twice for 5 min with sterile 1× PBS. The TUNEL reagent was prepared as directed in the protocol and 50 µl was loaded onto each specimen and covered with a 5 × 5 cm square of parafilm. The slides were placed in a humidity box and incubated for 1 h in the dark at 37°C in a water-jacketed CO₂ incubator (Nuaire). Slides were quickly washed three additional times with PBS then mounted with Mowiol 4 mounting medium and observed for photography on an Olympus-BX60 epi-fluorescence microscope.

RESULTS

We have demonstrated previously the ability of clamp ODNs, with a 5′ covalently linked psoralen, to bind the intended c-myc mRNA target, by thermal denaturation studies and electrophoretic mobility shift assay, and to inhibit cellular growth and c-Myc expression in vitro (9). The proposed mechanism by which the ODNs interact with their target sequences is diagrammed in Figure 1. In order to translate this work to an in vivo system we have used ODNs containing a 5′-end conjugated Acr intercalator and have expanded the range of coding region targets to include an additional exon 3 sequence. This new sequence is located 16 bp downstream of the original E3C target. The clamp ODN complementary to this sequence has been designated E1C. Acr-modified clamp ODNs E2C and E3C target the same 11 bp exon 2 and 10 bp exon 3 sequences, respectively (Fig. 1) (9).

Figure 1.

c-myc transcript target sequences and clamp ODNs. This diagrams the clamp ODN interactions with their respective c-myc exon 2 and exon 3 targets. The bracket regions indicate the interaction of the bound ODN to the mRNA sequence. The target cDNA sequences are located directly above the ODN.

Acr-modified clamp ODN effects on B16-F0 proliferation

In order to evaluate the ability of these Acr-modified clamp ODNs to inhibit cellular proliferation, B16-F0 cells were plated in triplicate in 96-well plates (day 1) and incubated for 48 h at 37°C. On days 3, 6 and 9 cells were treated with 3 µM ODN (E1C, E2C or E3C) alone in supplemented DMEM (Gibco BRL). In other experiments, B16-F0 cells were treated for 1 h with 10 µM DDP then exposed to ODN (3 µM E1C, E2C or E3C). On day 11, cell viability was determined by the MTT assay.

Figure 2 demonstrates that E1C, E2C and E3C alone at a concentration of 3 µM inhibited cellular proliferation by 76, 66 and 78%, respectively. When the cells were pretreated with DDP, neither E1C nor E3C, both at 79%, demonstrated any synergistic effect on proliferation. On the other hand, E2C, the only ODN which targets exon 2, displayed increased inhibitory activity at 92% when used in combination with DDP. The growth of cells treated with DDP alone was inhibited <20% compared with the non-treated control cells, while the growth of cells treated with the control clamp ODN (SCR) was inhibited by only 11–15% in the absence or presence of DDP.

Figure 2.

B16-F0 murine melanoma growth inhibition by c-myc clamp-forming AS ODNs, as measured by the MTT assay. Cells were treated with each Acr-modified ODN at 3 µmol or in combination with 10 µmol DDP. Alone, E1C, E2C and E3C resulted in 76, 66 and 78% reductions in cellular proliferation, respectively. ODN treatments after DDP pretreatment resulted in 79, 82 and 79% growth inhibition. SCR treatment resulted in no significant effects, as compared with control growth. Mean ± SE, defined by bars, of triplicate wells for all data points, representative of four independent experiments, are given.

Immunofluorescence detection of c-Myc after B16-F0 ODN treatment

To determine whether the Acr-modified clamp ODNs reduce the expression of cellular c-Myc protein, we used a c-Myc murine monoclonal primary antibody (Santa Cruz Biotechnology) and an FITC-conjugated secondary anti-mouse HRP-conjugated antibody (Molecular Probes) in an immunofluorescence analysis. B16-F0 cells were seeded at a density of 10² cells/well in 48-well plates (day 1), in duplicate, and treated according to the previously outlined procedure for the MTT assay, with one modification. On day 9 of the experiment, 1 h after the third treatment, cells were fixed with cold 4% PFA for 20 min on ice and stained using a mouse-specific c-Myc antibody (Santa Cruz Biotechnology).

The expression of c-Myc is easily detected by bright green nuclear and diffuse cytoplasmic staining in untreated cells (Fig. 3b). Treatment with either of the active clamp ODNs alone at 3 µM moderately reduced this expression (Fig. 3d, f and h), while a similar reduction occured with SCR treatment (Fig. 3j). Treatment with 10 µM DDP alone had no inhibitory effect on Myc expression (Fig. 3c). However, when cells were treated with ODN in combination with DDP there was a significant reduction in c-Myc expression. Unlike the results from the proliferation experiment, there is clear synergy of DDP with E1C and E3C in their effect on expression of the target oncogene. In the cells treated with both agents, Myc staining is no longer detectable (Fig. 3e and i). E2C plus DDP treatment also synergistically inhibited c-Myc, by almost 90% compared with the controls. Finally, there was no enhanced reduction of Myc staining in the cells treated with DDP and SCR, the control ODN (Fig. 3k). There was also no staining for c-Myc detected in the control cells serving as a background for immunostaining (Fig. 3a). At this point, we see synergy of DDP with the two exon 3-targeting clamp ODNs (E1C and E3C) in inhibiting c-Myc expression, while there does not appear to be synergism of growth inhibition at the concentrations tested. In contrast, E2C demonstrates high dependence on DDP to exert significant inhibition of either proliferation or myc expression. This suggests that the ODN may be less effective at inhibiting translation of the target mRNA or that the exon 2 sequence is not as important to biological activity as was previously thought (9).

Figure 3.

Photographs of B16-F0 cells after ODN treatment and immunofluorescence analysis with the c8 c-Myc monoclonal antibody. Cells were treated on days 3, 6 and 9 with either the combination of a 10 µmol dose of DDP for 1 h then transiently transfected with 3 µmol ODN (e, g, i and k) or ODN alone (d, f, h and j) in DMEM. Endogenous c-Myc expression is clearly detected in the control untreated cells (b). The cells in (a) were treated with medium alone and not stained with primary Myc antibody, thus serving as a background control for antibody reactivity. Cells treated with DDP (c) alone, SCR (j) or SCR + DDP (k) show immunoreactivity consistent with the untreated control cells (b). In contrast, only treatment with E1C (e), E2C (g) or E3C (i) after DDP administration exhibit a drastic reduction in cellular Myc staining. These results further substantiate that the active clamp ODNs can specifically target the c-myc mRNA to reduce cellular c-Myc expression.

Western blot analysis of cellular c-Myc expression after ODN treatment

Figure 4 is a western analysis showing the degree to which the active Acr-modified clamp ODNs inhibit Myc expression, either alone or in combination with DDP. Cells were treated according to the same protocol used for immunofluorescence detection and they were harvested 60 min after the last treatment. Western blot analysis was performed, as described in Materials and Methods, using the c-Myc antibody, as well as a monoclonal antibody for β-actin to serve as an internal control. Treatment with active clamp ODNs E1C, E2C and E3C alone resulted in 45, 69 and 60% reductions in Myc expression, respectively, while their treatment combined with DDP decreased cellular c-Myc by 66, 68 and 50%, respectively, compared with the non-treated or DDP alone-treated controls. Thus, there again appears to be no synergistic benefit of treating with DDP prior to that of E1C or E3C, also shown by MTT assay. In contrast, the synergy that was demonstrated previously between the drug and E2C is not mimicked by western analysis, at the in vitro level. There was no significant change in Myc expression after treatment with SCR alone (23%) or combined with DDP, and β-actin expression was not affected, supporting the argument that these antisense clamp ODNs act by specifically targeting the primary c-myc oncogene transcript.

Figure 4.

Western blot analysis of B16-F0 cells treated with clamp ODNs and drug. After three 3 µmol ODN or 10 µmol DDP treatments, total cellular lysates were harvested and 25 µg total protein was resolved by 8.5% SDS–PAGE and evaluated using the c8 c-Myc and C-2 β-actin mouse monoclonal antibodies (Santa Cruz Biotechnology). The blot was exposed to ECL reagents (Amersham) for c-Myc and actin protein detection. The lanes are labeled with the treatments (i.e. lane 1, CTL, is the untreated control for B16-F0 cellular c-Myc expression). E1C treatment reduced Myc protein expression by 45% alone and by 66% with DDP. E2C treatment resulted in 68 and 69% reductions in Myc expression, alone and combined with drug. Cells treated with E3C alone or combined with drug reduced expression by 60 and 50%, respectively. There was no significant change in Myc expression after either SCR treatment or with DDP alone compared with the untreated control expression. Also, no significant changes in β-actin expression were observed.

In vivo antitumor efficacy of Acr-modified clamp ODNs with DDP

The ability of Acr-modified clamp ODNs to inhibit B16-F0 tumor growth was evaluated in weanling age C57BL/6 mice (nine per group, except E2C, with five per group). On day 1 of each study, each mouse was injected with 2.5 × 10⁶ B16-F0 mouse melanoma cells in sterile 1× PBS. Prior to tumor challenge with each Acr-modified clamp ODN alone or in combination with DDP, each mouse was observed for a visible, palpable tumor mass. On days 4, 7 and 10 all tumor-bearing mice, except those in the control (untreated) group, received an intra-tumoral ODN injection at 3.3 mg/kg/day. The animals in the combination groups were treated by i.p. injection with a 3.3 mg/kg/day dose of DDP, 30 min before ODN administration. At the end of the 10 day regimen, each mouse had received a total dose of 10 mg/kg of either ODN (E1C, E2C, E3C or SCR) alone or in combination with a total dose of 10 mg/kg DDP. Twenty-four hours later (day 11), the single agent and the combined treatment responses were evaluated as a function of average tumor weight, c-Myc expression and apoptosis induction.

Figure 5 displays a graphical representation of tumors excised from mice in this experiment. Compared with the untreated control group (CTL), only treatment with E1C alone caused a marked decrease in tumor growth, an average of 77% inhibition (Table 1). Combined with DDP pretreatment, it inhibited growth by an average of 82%. In contrast, when comparing the ODN dose alone versus the ODN plus drug, only E3C demonstrated a synergistic advantage for inhibiting growth; an average of 71% better with DDP than alone (1%). The in vivo effectiveness of E2C treatment was less than expected based on its in vitro antiproliferative activity. Treatment with E2C alone resulted in moderate inhibition of growth at 54%, but decreased to only 32% with DDP, clearly not anticipated considering the level of antiproliferative activity of these two agents in cell culture. The averaged results from this experiment reiterate the ability of E1C to exert its inhibitory effects in the presence or absence of drug and support the conclusion that E3C plus DDP work better together than either agent alone. These data are in agreement with other findings that c-myc AS ODNs increase melanoma sensitivity to DDP treatment (42).

Figure 5.

Effects of c-myc clamp ODNs alone or combined with DDP on B16-F0 melanoma tumor growth in C57BL/6 mice. Three treatment cycles were administered of ODNs alone or subsequent to DDP, as described in Materials and Methods. No control treatments (DDP alone, SCR or SCR + DDP) affected tumor growth, whereas E1C treatment alone or plus DDP significantly inhibited progression by 77 and 82%. E2C affected growth less dramatically, at 54%, and only 32% when combined with DDP (results from four mice/group only). A cooperative effect was demonstrated with E3C + DDP treatment, inhibiting growth by 72%, while there was no inhibitory effect alone (1%). For E1C, E3C and SCR treatment graphical data represent treatment of six mice/group.

Table 1. In vivo efficacy of c-myc clamp ODNs and DDP on B16-F0 cells in C57BL/6 mice.

ODN/drug treatments	TTWA (g)a	PTGI (%)b
Control	0.235	0
DDP (D)	0.230	1
E1C	0.065	77
E1C + D	0.045	82
E2C	0.110	54
E2C + D	0.160	32
E3C	0.225	1
E3C + D	0.075	72
SCR	0.290	0
SCR + D	0.280	0

^aTotal tumor weight averages (TTWA) represent combined averages from the measured weight of each excised tumor, added then divided by the appropriate number of tumors weighed.

^bPercent tumor growth inhibition (PTGI) calculated using the TTWA number, to determine efficacy.

Some toxicity was observed in mice treated with E3C alone, DDP alone and E1C + DDP. The overall death rate was 11%, because one mouse died in each of these three treatment groups. The deaths in both groups treated with DDP are believed to be associated with well-documented drug toxicity (42), following the second (one mouse in DDP group) or third (one mouse in E1C + DDP group) drug administrations and prior to ODN treatment. Death of the mouse in the E3C group after the full three treatments is believed to be tumor related, due to evidence of metastatic disease following animal necropsy. On average, no growth inhibitory effects were found after treatment with DDP alone, SCR or SCR + DDP (Fig. 5 and Table 1).

The percentages of tumor growth inhibition are summarized in Table 1. Our data indicate that the most effective therapeutic doses were treatment with Acr-modified clamp ODN E1C alone or in combination with DDP and E3C treatment with DDP, based on the concentrations tested. On day 11, one animal per group was perfused with 4% PFA, prior to tumor excision, to evaluate the expression of c-Myc protein and the level of apoptosis in each tumor section.

In vivo immunofluorescence detection of c-Myc after ODN and DDP treatment

Tumors removed from perfused animals were stained for c-Myc expression as outlined in Materials and Methods. As seen in Figure 6, the sections stained with a rabbit polyclonal (N-262) anti-c-Myc antibody show distinct nuclear (brighter) versus cytoplasmic detection of Myc, especially in all the control tumors, CTL (Fig. 6a), DDP (Fig. 6b), SCR (Fig. 6c) and SCR + DDP (Fig. 6d). As predicted by the in vitro reduction in Myc after treatment with E1C (Fig. 6e) or combined with DDP (Fig. 6f), there is a drastically lower level of myc protein in these tumors. The same is true for the tumors treated with E3C (Fig. 6i) or E3C + DDP (Fig. 6j). However, E1C alone appears to be more effective at inhibiting tumor c-Myc expression compared with a combination of this ODN and DDP, while the E3C–DDP joint treatment is most effective. Once more, immunofluorescence results are reproduced in the tumors treated with E2C based on the cellular efficacy of E2C to reduce Myc expression alone or with DDP pretreatment; where alone (Fig. 6g) there is no effect, but the addition of DDP (Fig. 6h) leads to a 50% reduction compared with the controls. At this point the data suggest that the ability to generate sustained growth inhibition of this melanoma is definitely connected to the expression level of the c-myc oncogene. After these positive results, it was necessary to corroborate the data with a more quantitative assay.

Figure 6.

Photographs of B16-F0 tumor sections after DDP and ODN treatment, immunofluorescently analyzed with the N-262 c-Myc rabbit polyclonal antibody. C57BL/7 mice were treated as previously described in Materials and Methods. Myc expression is clearly detected in all tissues [(a) CTL; (b) DDP; (c) SCR; (d) SCR + DDP; (g) E2C; (h) E2C + DDP] except for those treated with the most effective ODN or drug + ODN combinations on tumor growth [(e) E1C; (f) E1C + DDP; (i) E3C; (j) E3C + DDP]. These results represent staining from one set of almost adjacent analyzed tumor sections. The other tumor tissues, located within 10 cryostat sections of those shown, produced nearly identical c-Myc staining patterns (data not shown).

In vivo western blot analysis of c-Myc expression after ODN and DDP treatment

In order to confirm the decrease in Myc expression observed with immunohistochemistry in tumors treated with the antiproliferative ODNs, protein extracts were prepared as previously outlined in Materials and Methods. After being resolved by 8.5% SDS–PAGE and transferred to PDVF membrane, western blot analysis was performed using the same Myc (N-262) antibody (Santa Cruz Biotechnology) used for immunofluorescence detection. Figure 7 demonstrates a relative decrease in the expressed 67 kDa protein after E1C treatment, by >90% alone and by 80% with DDP pretreatment. The expression of Myc after either E2C treatment alone (50%) or combined with DDP (54%) nearly mimics the results seen in both the in vitro western blot (Fig. 4) and in vivo immunofluorescence detection (Fig. 6 and data not shown). It is now more clear that the reduced tumor efficacy of this ODN is a consequence of its inability to target the myc mRNA as effectively as the other two Acr-modified clamp ODNs. E3C treatment alone was also unable to reduce tumor Myc expression, but the E3C + DDP regimen resulted in ∼95% inhibition. No changes in tissue Myc expression were observed after DDP, SCR or SCR + DDP treatment and there were also no changes in α-actin expression. We have concluded that the inability of E2C to significantly inhibit tumor growth in vivo is related to its poor activity as an antisense molecule against the c-myc message. On the other hand, we also suggest that the growth inhibitory efficacy of E1C and E3C with DDP are based on their ability to inhibit c-Myc expression. The exact mechanism by which DDP is implementing synergistic effects with the E3C clamp ODN is still not clear. More than likely, it is increasing the induction of apoptosis (2,36,37,42), which is also a consequence of myc down-regulation (8) and generated by the AS clamp ODN. It is also feasible to propose that E1C alone reduces Myc enough to irreversibly commit cells to undergo death.

Apoptosis evaluation of Acr-modified clamp ODN-treated tumors by TUNEL and PCNA western blot analysis

To verify whether or not there was an increase in apoptotic events in tumors treated with E1C and E3C alone or combined with DDP, we first performed the TUNEL assay using a Fluorescein In Situ Cell Death Detection Kit (Roche). The slides containing cryosectioned tumor specimens were assayed according to the standard protocol, outlined in Materials and Methods. Figure 8 confirms that the increased antiproliferative activity of E1C in cultured B16-F0 cells and tumors is related to an increased induction of apoptosis. The level of apoptotic positivity in the tumors treated with E1C alone (Fig. 8e) is pronounced to the same extent as that in the tumors treated with both ODN and drug (Fig. 8f). Interestingly, induction of apoptosis by E2C and E3C correlates with in vivo activity; where E2C treatment alone induced some apoptosis (Fig. 8g) and inhibited tumor growth by 54% (see Fig. 5), when combined with DDP, antiproliferative activity as well as apoptosis (Fig. 8h) both decreased in what appears to be an antagonistic manner. E3C treatment alone did not significantly affect tumor growth (see Fig. 5) or activate the programmed death pathway (Fig. 8i), whereas following DDP treatment, the synergy is clearly activating apoptosis (Fig. 8j) to kill the tumor cells with >70% efficiency (Table 1). Although there is some detection of apoptosis in the CTL (untreated) (Fig. 8a) and DDP-treated (Fig. 8b) controls, it is not like the distinct nuclear staining indicative of DNA fragmentation, characteristic in this assay, as clearly seen in the tumor sections treated with E1C or E3C + DDP (Fig. 8e, f and j). Treatments with SCR or SCR + DDP demonstrated no induction of apoptosis (Fig. 8c and d).

Figure 8.

Detection of apoptosis by TUNEL assay in B16-F0 mouse melanomas harvested from C57BL/6 mice after DDP and AS clamp ODN treatment. Apoptotic tissue is represented by bright green fluorescence, where the brightest staining is indicative of fragmented DNA. Only E1C (e), E1C + DDP (f) and E3C + DDP (j) treatment significantly induce apoptosis in this tumor model. Compared with the untreated CTL (a) or DDP-treated control (b), SCR (c), SCR + DDP (d), E2C + DDP (h) and E3C (i) treatments are all unable to drive tumor cells into apoptosis. E2C treatment alone (g) generates a moderate apoptotic signal.

To confirm the levels of apoptosis observed by TUNEL staining, western blot detection for PCNA was conducted. We used a 1:1500 dilution of rabbit polyclonal PCNA FL-261 (Santa Cruz Biotechnology) to establish whether the apoptotic program was possibly being driven by G₁ or S phase arrest, as a result of decreased c-Myc expression and DDP-generated DNA damage. The lower panel in Figure 7 shows a reduced level of PCNA expression in the tumor lysates corresponding to the same tumor sections with increased apoptotic signaling (Fig. 8). Specifically, PCNA is reduced in both E1C- and E1C + DDP-treated tumors (Fig. 7), while TUNEL staining is increased in both of these tumor sections as well (Fig. 8e and f). Similarly, treatment with E3C + DDP also reduces PCNA with increased TUNEL detection (Fig. 8j). As was the case with TUNEL staining, no other ODN or ODN plus drug treatment significantly affected PCNA expression (either E2C treatment >40%; data not shown). It is important to note that the levels of PCNA closely correlate with the expression profile of c-Myc in these tumors, as well as the antiproliferative activity of each treatment. Taken together, these data strengthen the conclusion that these modified AS clamp ODNs can inhibit murine melanoma growth and induce apoptosis, either alone (E1C) or more effectively when combined with DDP (E3C).

DISCUSSION

The ability of AS ODNs to inhibit the expression of oncogenes such as bcl-2, bcr/abl, c-myb, c-myc and ras in a variety of tumors has been repeatedly demonstrated (36,42–44). Despite the considerable potential of this approach, major problems of uptake and stability must be addressed. Multiple ODN analogs have been designed to resist nuclease degradation. Currently, one PS AS ODN, Vitravene (Isis Pharmaceuticals), is an approved drug for the treatment of CMV-induced retinitis in AIDs patients, while several others are in clinical trials (45). Obvious key antisense targets should be cell cycle genes regulating growth, differentiation and death (46).

The dysregulation of c-myc and mutation of the ras gene family members have been the focus of many studies. With one or both commonly found in nearly all human tumors, these two genes clearly cooperate in malignant transformation (47). In melanoma, c-Myc overexpression has been associated with poor outcome and decreased disease-free interval (48). c-myc is therefore important to study in this tumor type. Also, two recent studies have suggested that activated N-ras, which is expressed in 12–15% of all melanomas (15,16), significantly contributes to chemoresistance (42).

In agreement with our findings, a majority of previous studies using AS ODNs targeted to c-myc or other oncogenes in combination with DDP (2,22,36–45,47–49) also demonstrated an increased induction of apoptosis at various stages of the cell cycle (42). Unfortunately, in spite of the success in these animal studies, which basically all focused on reversing the dysregulated actions of one oncogene or another, eventually all test subjects died of their ensuing disease (36,42). The strategy of targeting major genetic contributors simultaneously with chemotherapeutic agents already proven to have some efficacy could prove deadly for cancers and other diseases.

In terms of specifically inhibiting gene expression by the AS method, clamp ODNs have an advantage in tested systems (9,23). Because triplex formation is highly sequence sensitive and any mismatch in the Hoogsteen fragment can prevent complex formation (50), this design should be considered along with other topological modifications, such as backbone analogs and end conjugates. It might be expected that triplex formation on single-stranded targets will more effectively inhibit gene expression than is the case with duplexes formed with normal AS ODNs (23,24). Previous work in our laboratory, using two of the clamp ODNs presented here (E2C and E3C), but substituted with a 5′ psoralen moiety, did not inhibit tumor growth (D.A.Steward and D.M.Miller, unpublished results). Based on those results we have demonstrated that Acr and possibly other naturally reactive conjugates, like chlorambucil, may be more effective than psoralen as end-conjugates for these ODNs.

We report here the identification of three sequences within the coding region of the mouse c-myc mRNA and have demonstrated that targeting two of them results in biologically significant inhibition of mouse melanoma proliferation. The reduced in vivo efficacy of E2C may be attributed to the fact that the exon 2 target sequence is more than 75 bases downstream of the translation initiation start site. Thus, while clamp ODNs increase the spectrum of available AS targets, from mRNA regulatory regions to include the coding regions, there may be a limit to how far within the coding sequences the target can be and still remain relevant. Our observed effects reiterate that targeting c-myc with appropriately modified AS ODNs also contributes to DDP efficacy. Again, in this system the dominant mechanism of action appears to be activation of the apoptotic program. Considering the PCNA data, no definitive judgment can be made at this time using only the PCNA expression profile as to whether treatment with the ODNs alone or combined with cisplatin induces apoptosis after a G₁ or S phase arrest, primarily because PCNA is expressed during both of these phases, and our main goal in evaluating PCNA expression was to better confirm the levels of apoptosis observed in the TUNEL assay. To determine whether the pre-apoptotic cells arrest in G₁ versus S phase will require further analysis; for example, p53 and Ki67 antigen evaluation, in which an increase in p53 and decrease in Ki67 expression would suggest G₁ arrest, or monitoring Cdk2, whereby a decrease in activity accompanied by an increase in p21 expression would favor S phase arrest. Although the exact mechanism through which apoptosis is being driven after clamp ODN treatment is yet to be elucidated, based on other ODN studies in our laboratory we propose that the cells are arresting in S phase prior to the activation of the death program.