Introduction

G3139 (Genasense) is an 18mer phosphorothioate antisense oligonucleotide targeted to the initiation codon region of the Bcl-2 messenger RNA (mRNA). The Bcl-2 protein, which is strongly antiapoptotic, is expressed not only in normal melanocytes and benign nevi but also in primary melanoma and melanoma metastases.^{1, 2, 3} When overexpressed in a variety of models, Bcl-2 protein has been shown to decrease the apoptotic response to cytotoxic chemotherapy.^{4, 5, 6} G3139, by virtue of its Watson–Crick specificity, downregulates Bcl-2 protein and mRNA expression in melanoma and other cells in vitro.^{7, 8, 9, 10} This agent has completed worldwide randomized Phase III clinical trials in advanced melanoma in combination with dacarbazine.¹¹ Statistically significant differences in overall survival and time to progression were noted in patients with normal values of lactate dehydrogenase (a marker of tumor necrosis) at presentation versus patients who received dacarbazine alone (Genta Inc., presented at the first Oligonucleotide Therapeutics Society Meeting, NY, September 2005).

However, both the in vitro and in vivo mechanisms of action of G3139 are at the moment somewhat uncertain, and may be more complex than solely Bcl-2 downregulation and subsequent chemosensitization.¹² It is true that in vitro, G3139 downregulates the expression of both Bcl-2 mRNA and protein in many cell lines, and has been reported to produce sensitization to chemo-^{13, 14, 15} and radiotherapy.¹⁶ However, in 518A2 melanoma cells, we could not demonstrate any sensitization to a variety of cytotoxic drugs.⁹ Furthermore, temporary Bcl-2 silencing with a 21mer short-interfering RNA, or permanent Bcl-2 silencing via an anti-Bcl-2 single-stranded DNA or short hairpin RNA¹⁷ approach, also did not produce chemosensitization. Rather, after lipotransfection with G3139, melanoma cells undergo caspase-dependent apoptosis via the intrinsic pathway subsequent to cytochrome c release from mitochondria into the cytoplasm in the absence of Bcl-2 silencing.¹⁰ This cytochrome c release appears to be a consequence of a general interaction of phosphorothioate oligonucleotides of length 16mer with outer mitochondrial membrane proteins known as the voltage-dependent anion-selective channel (VDAC).¹⁸ This interaction causes VDAC closure (i.e., its inability to conduct a current in the presence of a potential difference). Closure of VDAC, by a mechanism that remains uncertain,¹⁹ then appears to lead to cytochrome c release from the mitochondria, and hence apoptosis.

However, it is by no means clear that this VDAC-related mechanism operates in vivo. Moreover, further complicating an already unclear mechanistic situation, G3139 contains two cytidine-phosphate-guanosine motifs, each of which can be immunomodulatory even in severe combined immunodeficiency (SCID) mice by virtue of their ability to bind to and activate TLR9-positive plasmacytoid dendritic cells.²⁰ On stimulation, these cells can produce cytotoxic cytokines that may lead to tumor growth suppression in experimental animals. However, the extent to which this mechanism is responsible for the anti-human tumor xenograft effects of G3139 in severely immunocompromised mice is controversial.^{20, 21}

The nucleosides of G3139 are based on the naturally occurring D-deoxyribose stereoisomer at the C1 carbon of the furanose ring (2-deoxy--D-erythropentofuranose). In an attempt to sort out the various potential in vitro and in vivo mechanisms of action of G3139, we have synthesized the phosphorothioate isosequential L-deoxyribose (2-deoxy--L-erythro-pentofuranose) enantiomer L-G3139. L-Deoxyribose oligonucleotide isoforms (also known as "Spiegelmers", from the German word for mirror) are non-natural and, unlike the natural D-deoxyribose oligonucleotides, are virtually impervious to nucleases.^{22, 23} Because D-G3139 and L-G3139 are enantiomeric pairs (by virtue of chirality at the C1 carbon of deoxyribose) and are both phosphorothioate polyanions, their "bulk chemical" properties, and hence non-sequence specificity, might be similar. However, their potential as antisense oligonucleotides is quite different, as L-oligonucleotides do not elicit RNase H activity. Thus, L-G3139, in contrast to D-G3139, should not downregulate Bcl-2 expression. This critical difference might provide a way of examining the role of Bcl-2 protein in melanoma cells in both the in vitro and in vivo settings.

Results

The protein binding and inhibition of net-forming abilities of D-G3139 and L-G3139 are independent of chirality at C1

Phosphorothioate oligonucleotides can bind with high affinity and relative sequence independence to heparin-binding proteins, such as basic fibroblast growth factor (bFGF), vascular endothelial growth factor, and platelet-derived growth factor,²⁴ as well as to fibronectin, laminin,²⁵ and various cell surface proteins.²⁶ This binding may account for at least some of the non-sequence specificity observed in in vitro experiments with phosphorothioate oligonucleotides.²⁷ Therefore, to determine to what extent L- and D-G3139 differ in their protein binding affinities, we used the heparin-binding protein bFGF as a model. Each phosphorothioate oligonucleotide was used as a competitor of the binding of the alkylating, radiolabeled probe, ClRNH³²P-OdT18, for which the K_d (0.5 M) of binding to bFGF had previously been determined.²⁴ The plots of band intensity versus log D- or L-G3139 concentration are shown in Figure 1, and are both linear (R²>0.96 for each). The values of K_c, as calculated by equation 1, are 40 nM (L-G3139) and 60 nM (D-G3139), which are equivalent to within the margin of error of the experiment.

Figure 1.

Interaction of L- and D-G3139 with bFGF. (a) Competition by D-G3139 and L-G3139 for binding of ClRNH³²P-OdT₁₈ to bFGF. D-G3139 or L-G3139 at the stated concentrations were used as competitors of ClRNH³²P-OdT₁₈ (8 M) binding to bFGF (50 nM) as described in the Materials and Methods (12% PAGE). (b) Determination of the value of K_c for ClRNH³²P-OdT₁₈ by the Cheng–Prusoff equation (equation 1). The data from (a) were quantitated by laser densitometry. Shown is a plot of the normalized band intensity versus the log of D-G3139 or L-G3139 concentrations.

Full figure and legend (47K)

When 518A2 melanoma cells are plated on Matrigel (which is composed primarily of collagen and laminin),²⁸ they form net-like structures (Figure 2). A 24 h treatment with increasing and identical concentrations of either D- or L-G3139, in the absence of any transfecting lipidic agent, causes highly effective inhibition of net formation. Homotypic adhesion, however, is clearly preserved, and the cells remain viable in both cases as demonstrated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (not shown). Because no transfection agent was used, the inhibition of net-forming ability is probably due to a nonspecific anti-adhesive effect of phosphorothioate oligonucleotides, which has previously been described.²⁹

Figure 2.

Phase-contrast photomicrographs demonstrating the inhibitory effect of D-G3139 and L-G3139 on the migration of 518A2 melanoma cells on Matrigel. Cells (30 10⁴) in 0.2 mL without or with increasing concentrations of D-G3139 or L-G3139 were seeded in 24-well plates precoated with Matrigel and incubated at 37°C. After 48 h, the pictures were acquired at original magnification 4.

Full figure and legend (158K)

The uptake of D-G3139 and L-G3139 in melanoma cells is also independent of chirality at C1

The evaluation of the in vitro effects of D- and L-G3139 must be considered in the context of their relative uptake in 518A2 and A375 melanoma cells. Both 5'-fluorescein isothiocyanate-labeled oligomers (100 nM), with virtually identical fluorescent emissions at 524 nm as measured fluorimetrically, were mixed with 1.9 g/mL of lipofectamine 2000, and incubated with 518A2 or A375 cells. Five hours later, cells were analyzed by flow cytometry (Figure 3). As demonstrated, the uptake of fluoresceinated oligomer is actually somewhat higher for L-G3139/lipofectamine 2000 versus D-G3139/lipofectamine 2000. However, two cellular A375 populations seem to exist in an approximately 1:1 ratio, comprising cells that do and do not take up the lipid–oligonucleotide particles well under the conditions of the experiment.

Figure 3.

Uptake of 5'-fluorescein isothiocyanate-labeled oligonucleotides in 518A2 and A375 melanoma cells, as analyzed by flow cytometry. Cells were transfected with 100 nM of oligonucleotides complexed to 1.9 g/mL of lipofectamine 2000 for 5 h. Experiments were carried out in triplicate and data are presented as meanSD.

Full figure and legend (76K)

D- and L-G3139 cause the release of cytochrome c from isolated mitochondria and reduction of the MOM permeability

As described in the Materials and Methods, 518A2 melanoma cells were homogenized and mitochondria were isolated by differential centrifugation. The mitochondria were suspended in energizing buffer at 10°C and treated with increasing concentrations of either D- or L-G3139 (Figure 4a) for 2 h. For each oligonucleotide, the release of cytochrome c into the supernatant (as determined by Western blotting) was similar as a function of concentration, except in some experiments at the highest concentration (40 M).

Figure 4.

Effect of L- and D-G3139 on mitochondria. (a) Western blot of the concentration-dependent profile of cytochrome c release from isolated mitochondria by D-G3139 and L-G3139. Isolated mitochondria from 518A2 melanoma cells were treated with various concentrations of oligonucleotides in buffer B and the supernatants were collected and analyzed by Western blotting as described in the Materials and Methods. D- and L-G3139 have similar effects on the MOM permeability and VDAC conductance. (b) The concentration dependence of the reduction of outer membrane permeability to ADP by D- and L-G3139. The results were normalized to the permeability in the absence of G3139. (c) Fractional conductance remaining after treatment of isolated and reconstituted VDAC channels with 5 or 20 M D- or L-G3139. Each result is the average of three experiments and presented as the meanSE. Statistical tests indicate significance in the reduction of the conductance. ^*P<0.05, ^**P<0.01, ^***P<0.005.

Full figure and legend (64K)

L- and D-G3139 were also evaluated in isolated mitochondria from rat liver. The permeability of the outer membrane to adenosine diphosphate (ADP) was measured at room temperature as described previously.³⁰ D-G3139 has been shown to reduce this permeability with K_i values between 0.2 and 0.5 M.³¹ Figure 4b shows that the L- and D-enantiomers each reduce the outer membrane permeability with a similar concentration dependence. This inhibition occurs within 10 min. The reduction in mitochondria outer membrane permeability to ADP is attributed to VDAC closure, because VDAC is the main pathway for metabolite flux through the outer membrane.^{29, 32, 33}

As we have described previously, D-G3139 interacts with VDAC, leading to loss of ionic conductance through this mitochondrial outer membrane (MOM) channel. Maximal effects were observed at 40 M, but a significant loss of conductance was detected at 5 M D-G3139. Figure 4c demonstrates that both D- and L-G3139 have similar inhibitory effects at 5 M, indicating an indistinguishable potency for the L-G3139 Spiegelmer. Both inhibitors also caused the same flickering of VDAC channels (data not shown), indicating the same underlying inhibitory mechanism.

L-G3139 does not downregulate Bcl-2 protein expression but does produce apoptosis in melanoma cells

Because L-G3139 does not serve as a substrate for RNase H, the expression of Bcl-2 protein is not downregulated for up to 3 days after transfection into melanoma cells (Figure 5a). In contrast, as has been well described, after treatment of melanoma cells with D-G3139, Bcl-2 protein and mRNA expression continuously decline to a minimum level, which can be seen after 3 days.

Figure 5.

(a) Downregulation of Bcl-2 protein levels by D-G3139, D-G4126, and L-G3139 after 3 days of treatment in both 518A2 and A375 melanoma cells. Apoptotic induction by D-G3139, D-G4126, and L-G3139 in (b) 518A2 melanoma and (c) A375 melanoma cells. Cells were treated as described previously for 9.5 and 24 h and poly ADP-ribose polymerase-1 and caspase-3 cleavage were analyzed by Western blotting.

Full figure and legend (92K)

Also as described previously, when 518A2 and A375 melanoma cells are treated with D-G3139/lipofectamine 2000 complexes (100 nM; Figure 5b and c, respectively), the cleavage of poly ADP-ribose polymerase-1 can be observed as little as 9.5 h after the initiation of the 5 h transfection. Also at this time, the cleavage of procaspase-3 to caspase-3 can be observed, consistent with the activation of the intrinsic apoptotic pathway. (As we have previously shown,¹⁰ the appearance of procaspase-3 cleavage products by Western blotting is directly correlated with increased caspase-3 activity, as measured by the cleavage of fluorogenic peptide substrates.) In contrast, treatment with L-G3139, which does not downregulate Bcl-2 protein expression, does not produce poly ADP-ribose polymerase-1 or procaspase-3 cleavage at 9.5 h, and only very minimal cleavage of these two proteins can be observed by 24 h after the transfection. In fact, approximately 48 h is required for the L-G3139-induced activation of the intrinsic pathway to approximate the apoptotic effects observed with D-G3139.

The activities of L- and D-G3139 in A375 melanoma human tumor xenografts in SCID mice are chirality dependent

As is clearly demonstrated in Figure 6, A375 melanoma cells readily form tumors in SCID mice, and the growth of these tumors is not inhibited at all by L-G3139 at either dose schedule employed. In sharp contrast, treatment of the animals with D-G3139, either at 10 mg/kg every day for 14 days or at 20 mg/kg every other day for 14 days, led to substantial inhibition of the rate of tumor growth.

Figure 6.

The growth of A375 melanoma human tumor xenografts in SCID mice is inhibited substantially by the treatment of D-G3139 either at 10 mg/kg every day for 14 days, or at 20 mg/kg every other day for 14 days. However, the rate of tumor growth is not inhibited at all by L-G3139 at either dose schedule.

Full figure and legend (62K)

L-G3139 is not immuno-stimulatory, in contrast to D-G3139

As immuno-stimulatory phenomena could possibly be the cause of the observed xenograft tumor responses, we examined D- and L-G3139 for chirality-dependent differences. Accordingly, BALB/c mouse splenocytes (5 10⁵ cells/assay) were treated with increasing concentrations of either L-G3139 or D-G3139. Control cells were treated with either 10 g/mL concanavalin A or lipopolysaccharide. At the end of either 24 or 48 h, the supernatants were assayed by enzyme-linked immunosorbent assay (ELISA) for the presence of interleukin (IL)-6, IL-12, or IP10. As shown in Figure 7, a dose-dependent increase in the production of both IL-6 and IL-12 was observed in cells treated with D-G3139, but there was no increase in the production of these cytokines after treatment of the splenocytes with L-G3139. Neither oligonucleotide produced an increase in IP10.

Figure 7.

Native BALB/c mouse splenocytes were treated with increasing concentrations (0.156–40 M) of either L-G3139 or D-G3139. Control cells were treated with either 10 g/mL concanavalin A or lipopolysaccharide. At the end of either 24 or 48 h, the supernatants were assayed by ELISA for the presence of IL-6, IL-12, or IP10, according to the manufacturer's instruction. Dose-dependent increases in the production of both IL-6 and IL-12 were observed in cells treated with D-G3139 after 24 and 48 h. However, there was no increase in production of these cytokines after treatment of the splenocytes with L-G3139.

Full figure and legend (124K)

In another set of experiments, eight Fox-Chase™ SCID mice were injected with D-G3139, L-G3139, or saline control at a dose of 20 mg/kg 7 every other day for 14 days. At the end of this time, serum samples were obtained and assayed by ELISA for IL-12 and interferon (IFN)- levels. The results are shown in Figure 8; for D-G3139, the levels of these cytokines, respectively, were 16245 and 6514 pg/mL, versus 09 and 04 for L-G3139. Cytosine C5-methylation of the two G3139 cytidine-phosphate-guanosine motifs (producing G4232) caused, as expected, a dramatic diminution in the production of IFN- (134 pg/mL), and a not so substantial decrease in the production of IL-12 (8762 pg/mL). G4126 (G3139 without cytidine-phosphate-guanosine motifs) also produced negligible levels of either cytokine. We then evaluated the ability of IL-12 (25, 100, 10³, and 10⁴ pg/mL) or IFN- (600 and 6,000 pg/mL) to inhibit the growth of A375 and 518A2 cells in tissue culture after 3 days, as evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. In each case, cytokine concentrations significantly higher than the in vivo concentrations achieved after D-G3139 injection were evaluated. However, in no case was there any inhibition of cell growth whatsoever, as determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (not shown). Therefore, in this model system, the production of these two cytokines alone cannot explain the observed in vivo antitumor effect.

Figure 8.

ELISA assay demonstrating the increase of production of (a) IL-6 or (b) IFN- after treatment of mice with D-G3139 or L-G3139. Eight Fox-Chase SCID mice were injected with D-3139, L-G3139, or saline at a dose of 20 mg/kg 7 every other day for 14 days. At the end of this time, serum samples were assayed by ELISA according to the manufacturer's protocol.

Full figure and legend (45K)

Discussion

In this work, we have compared the activity of D-G3139 and its deoxyribose enantiomer pair L-G3139 against melanoma cells, both in vitro and in vivo. To a significant extent, the in vitro activities are quite similar to each other and to other non-Bcl-2 targeting phosphorothioate oligonucleotides.²⁴ For example, both enantiomers bind nonspecifically to the heparin-binding protein bFGF with approximately the same K_c. However, this does not imply that the D- and L-enantiomers will bind to all heparin-binding proteins similarly. Clearly, the binding to each protein must be individually evaluated. Lipid-mediated cellular uptake for both enantiomers in both 518A2 and A375 melanoma cells actually favors the L-enantiomer, for unclear reasons that may relate to particle size, rate of particle uptake, and/or relative rate of dissociation of the oligonucleotide from the lipid carrier.

At the level of the isolated mitochondria, the L- and D-enantiomers also behave similarly. This is reflected in the diminution in mitochondrial permeability to ADP, which is transported through open VDAC channels;^{29, 32, 33} this observation is consistent with the equivalent ability of each enantiomer to bind to the category of heparin-binding proteins, of which VDAC is a member. We have previously shown that the interaction of D-G3139 with VDAC leads to VDAC closure (i.e., reduction in conductance). It has been postulated that the opening of another channel in the MOM then occurs, through which cytochrome c and other proapoptotic proteins can be released.¹⁹ It is possible that this channel is formed by oligomerized Bax³⁴ or by ceramide,¹⁹ but its true nature is still controversial. When VDAC is incorporated into planar membranes, the conductance loss observed after treatment with L-G3139 is also quite similar to that observed with D-G3139.

Despite the similarities in behavior of this enantiomeric pair in relatively simple systems, there are striking differences in the behavior observed between the two enantiomers both in intact cells and especially in the in vivo setting. In melanoma (and other) cells in tissue culture, D-G3139 activates the intrinsic apoptotic pathway (activation of caspase-3, cleavage of poly ADP-ribose polymerase-1) subsequent to release of cytochrome c from the mitochondria, as little as 9.5 h after the initiation of the transfection.¹⁰ It is striking that a similar initiation of activation of apoptosis by L-G3139 takes as long as 24 h after the beginning of the transfection. However, once the intrinsic pathway is activated by either enantiomer, apoptosis will proceed over the next 2 days to produce massive cellular death.

One possible explanation for the enantiomeric difference in the rate of progression of apoptosis may lie in the inability of L-G3139 to downregulate Bcl-2 protein expression via an antisense mechanism in vitro. When melanoma cells are treated with D-G3139, Bcl-2 protein expression will fall continuously over the ensuing 24 h.⁹ Thus, Bcl-2 protein in these cells may function more as a modulator of the rate of apoptosis progression than as an absolute block. The ultimate endpoint, cell death, still seems nevertheless fixed and determined on initiation of treatment in these cells with these drugs.

The differences between the D- and L-enantiomers in an A375 melanoma xenograft model in immunodeficient mice lacking B-, T-, and natural killer cells are striking, as L-G3139, in contrast to D-G3139, does not appear to induce any change in the rate of tumor growth versus phosphate-buffered saline control alone. The reasons for this difference may be quite complicated, and may relate to recent evidence demonstrating that Bcl-2 is a strongly proangiogenic protein in vivo. This, in turn, is probably due to its ability to increase the production of proangiogenic cytokines and chemokines via modulation of the nuclear factor-kappa B pathway^{35, 36, 37, 38} Furthermore, it has been recently demonstrated that 518A2 melanoma cells with permanently silenced Bcl-2 do not growth as human tumor xenografts in SCID mice.³⁹ If D-G3139 can indeed downregulate Bcl-2 expression in melanoma tumors in vivo, which has previously been suggested,⁸ this could account for the enantiomeric difference in the rate of tumor growth in vivo.

However, it is unclear to what, if any, extent the mitochondrial mechanism of antitumor activity of G3139¹⁸ can be operative in vivo, given that both the D- and L-enantiomers have equivalent cytochrome c-releasing effects on isolated mitochondria in vitro. Nevertheless, as there also appear to be Bcl-2-dependent effects on the rate of progression of apoptosis in vitro, these may be appearing in vivo as well.

Finally, to further complicate mechanistic understanding, D-G3139 is immune-stimulating to the TLR9-positive plasmacytoid dendritic cells, myeloid dendritic cells, and macrophages^{40, 41, 42} which are present even in immunodeficient mice, whereas L-G3139 completely lacks this activity. Significantly elevated levels of IL-12 and IFN- were found in mice after 3 weeks of G3139 treatment, but in vitro, significantly higher levels of these cytokines were not toxic. These experiments do not, however, rule out (or for that matter rule in) other immune-mediated mechanisms for the observed tumoricidal effects of D-G3139 (e.g., tumor necrosis factor- or tumor necrosis factor-related apoptosis-inducing ligand). Furthermore, in nude mouse xenograft models, which lack T cells, substantial antitumor effects with sequence optimized cytidine-phosphate-guanosine-containing oligonucleotides have been observed only in combination with cytotoxic chemotherapy.^{43, 44} In our study, antitumor effects were observed in SCID mice with single-agent D-G3139 in the absence of any cytotoxic chemotherapy.

In conclusion, although there are significant similarities in the activities of L- and D-G3139, there are also profound differences, especially in the in vivo setting, where L-G3139 is inactive. However, because of the mechanistic complexity of D-G3139 in vivo, it is likely that no single mechanism of action can completely account for all pertinent observations. Rather, we propose that a multiplicity of antitumor mechanisms are involved, some likely Bcl-2 dependent, others probably not. We propose that it is their combination that has produced some of the dramatic antitumor effects observed in vivo with this drug; this interpretation suggests that directed chemical improvements in D-G3139 may increase clinical efficacy as well.

Materials and Methods

Synthesis of L- and D-G3139. The phosphorothioate L-G3139 with sequence 5'-L-d(TCT CCC AGC GTG CGC CAT)-3' was synthesized DMT-ON using standard solid-phase DNA phosphoramidite chemistry on an ÄktaPilot100 synthesizer (GE Healthcare, Munich, Germany). L-DNA-phosphoramidites and 500 Å L-dT cytidine-phosphate-guanosine support were from ChemGenes Corp. (Wilmington, MA). Thioation was achieved by using 0.2 M phenylacetyl disulfides (PADS). The solid support was treated for 10 min with diethylamine/acetonitrile (v/v 1:9), followed by deprotection with 40% aqueous methylamine (40 min, 60°C). After cooling to 20°C the support was removed by filtration and washed with 50% aqueous ethanol. The combined filtrates were concentrated at 30°C/40 mbar, and the product was purified by IEX-high-performance liquid chromatography (Source15Q (GE Healthcare)). Buffer A: 1 mM EDTA, 25 mM Tris, 10 mM NaClO₄ in water/acetonitrile (v/v 7:3). Buffer B: 1 mM EDTA, 25 mM Tris, 500 mM NaClO₄ in water/acetonitrile (v/v 7:3); gradient: 20% B 100% B in 12CV; product elutes at 60% buffer B. Fractions containing >90% of product were combined, and the pH was adjusted to 5 by the addition of acetic acid. Detritylation was monitored by analytical high-performance liquid chromatography. After 30 min the solution was desalted by ultrafiltration through a 5 kDa regenerated cellulose membrane (Millipore Corp., Bedford, MA) and the product was lyophilized. Fluorescein isothiocyanate-L-G3139 was prepared at NOXXON Pharma AG.

D-G3139 and D-G4126 were synthesized by standard methods and provided by Genta Incorporated. D-G4126 is a two-based mismatched control oligonucleotide of D-G3139. 5'-FAM-G3139 was synthesized by standard methods and purchased from TriLink Biotechnology (San Diego, CA).

Cell culture and oligonucleotide transfections. The mycoplasma-free human melanoma cell line 518A2 was a kind gift of Dr Volker Wacheck (University of Vienna, Austria). Cells were grown in Dulbcco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum and 100 U/mL penicillin G sodium and 100 g/mL streptomycin sulfate. A375 cells were purchased from ATCC (Rockville, MD) and were cultured in RPMI with the identical supplements+1% non-essential amino acids and 1% pyruvate. Stock cultures of all cells were maintained at 37°C in a humidified 5% CO₂ incubator.

Cells were seeded the day before the experiment in six-well plates at a density of 15 10⁴ cells per well, to be 60–70% confluent on the day of the experiment. All transfections were performed in Opti-MEM medium plus complete media without antibiotics, as previously described.⁴⁵ The incubation time for oligonucleotide/lipofectamine 2000 complexes was 5 h. The total incubation time before cell lysis and protein isolation was 24–72 h at 37°C.

Subcellular fractionation and oligonucleotide treatment of isolated mitochondria. Cells were harvested by trypsinization and were washed with cold phosphate-buffered saline. Cell pellets were re-suspended in 300 L of buffer A (250 mM sucrose, 10 mM Tris-HCl, pH 7.4, 1 mM EGTA, 50 g/mL Pefabloc and 15 g/mL leupeptin, aprotinin and pepstatin). Cells were then homogenized on ice in a dounce homogenizer until 90% of cells were disrupted, as judged by Trypan blue staining. Crude lysates were centrifuged at 1,000 g for 10 min at 4°C twice to remove nuclei and unbroken cells. The supernatant was collected and subjected to a 10,000 g centrifugation for 30 min at 4°C The supernatant was collected as the cytosolic fraction and the mitochondrial pellets were re-suspended in 20 L of buffer B (=buffer A+10 mM KCl, 3 mM KH₂PO₄, 5 mM succinate, and 100 M ADP).

Western blot analysis. Cells treated with oligonucleotide–lipid complexes were extracted in lysis buffer at 4°C for 1 h. Aliquots of cell extracts, containing 15–25 g of protein, were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to Hybond ECL filter paper (Amersham, Arlington Heights, IL). After treatment with appropriate primary and secondary antibody, ECL was performed, as previously described.⁴⁵ The typical margin of error for a Western blotting is at least 20–25%.

Uptake of fluorescein isothiocyanate-labeled oligonucleotides. 518A2 and A375 melanoma cells were treated with fluorescein isothiocyanate-L-G3139 or -D-G3139 (100 nM) complexed to lipofectamine 2000 (1.9 g/mL) for 5 h at 37°C and the complexes were replaced by 1 M of a 28mer oligodeoxyphosphorothioate homopolymer of cytidine (SdC28) for blocking for 10 min. Cells were then trypsinized and washed twice in phosphate-buffered saline and the intensity of the cellular fluorescence was analyzed by a flow cytometer. A minimum of 10,000 cells/sample was acquired using a log scale and data were analyzed using CELLQuest software (Becton Dickinson).

Synthesis of the probe, alkylating oligodeoxynucleotide ClRNH³²P-OdT₁₈. In brief (Yakubov et al.⁴⁶), 10 optical density units of an 18mer phosphodiester homopolymer of deoxythymidine (OdT18) were 5'-labeled with [³²P]-phosphate by reaction with 5'-polynucleotide kinase. Excess ATP was separated from reaction product by Sephadex G25 chromatography in 0.1 M aqueous lithium perchlorate. The oligonucleotide was then precipitated by the addition of 2% LiClO₄/acetone and dissolved in water at a concentration of 200 optical density units per microliter. The oligonucleotide was then precipitated by the addition of an 8% aqueous solution of cetyltrimethylammonium bromide solution and dried. To this was added 6.5 mg of p-(benzylamino)-N-chloroethyl-N-methylamine (ClRNH₂) in 20 L of dimethylformamide, followed by 8 mg of dipyridyl disulfide and 9.5 mg of triphenylphosphine. After 2 h, the oligonucleotide was precipitated by the addition of 2% LiClO₄/acetone, dissolved in 25 L of 1 M NaCl, precipitated with ethanol, and dried. The final product was redissolved in water and stored at -80°C.

Modification of heparin-binding proteins by ClRNH³²P-OdT₁₈. This was accomplished by the method of Guvakova et al.²⁴ Initially, bFGF (50 nM) was incubated in 0.1 M Tris-HCl, pH 7.4, containing 10–20 M ClRNH³²P-OdT₁₈. L-G3139 and D-G3139 were used at increasing concentrations as competitors of the binding of the probe phosphodiester oligonucleotide to the proteins. After 1 h at 37°C, one volume of a buffer containing 10% glycerol, 4% 2-mercaptoethanol, 4% sodium dodecyl sulfate, and 0.2% bromophenol blue was added, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed. The gels were dried and allowed to expose Kodak X-ray film until bands were visualized. The film was developed, and band densities were quantitated by laser densitometry.

Formation of tube-like structures on Matrigel. The formation of capillary-like structures on Matrigel was assessed as described previously. Frozen Matrigel was thawed on ice and 250 L aliquots were poured into each well of a 24-well plate. The plate was incubated for 2 h at 37°C to allow the Matrigel to gel. Then, 400 L of a cellular suspension containing 5 10⁴ 518A2 cells with or without L-G3139 or D-G3139 at increasing concentrations was plated onto the layer of Matrigel. Tube-like structure formation was evaluated after 24–48 h incubation in 5% CO₂ at 37°C using an inverted phase contrast microscope.

Measurement of MOM permeability and intactness. Mitochondria were isolated from rat (Sprague–Dawley) liver.^{47, 48} A portion of the mitochondrial suspension (containing 1 mg mitochondrial protein) was diluted into 3 ml of respiration buffer (0.3 M mannitol, 10 mM NaH₂PO₄, 5 mM MgCl₂, and 10 mM KCl, pH 7.2). The mitochondrial respiration was measured according to the method of Lee et al.⁴⁹ Briefly, 5 mM succinate was added into the respiration buffer with mitochondria present followed by ADP addition (90 M). The oxygen consumption was measured by using a Clark oxygen electrode and the MOM permeability was obtained by fitting the data to a theoretical model.⁴⁹

Assay of protein content. Mitochondrial protein was measured using the bicinchoninic acid method following addition of Triton X-100 to the mitochondrial suspension (1% w/v final). Bovine serum albumin was the standard.

Planar phospholipid membrane studies. The planar phospholipid membranes were generated according to standard methods.⁵⁰ The membrane was formed from phospholipid monolayers consisting of diphytanoylphosphatidylcholine, asolectin (polar extract of soybean phospholipids), and cholesterol (1:1:0.05 mass ratio).

VDAC was purified from mitochondria isolated from rat liver.^{47, 48} A 1–3 L aliquot of the VDAC-containing solution (2.5% Triton 100, 50 mM KCl, 10 mM Tris, 1 mM EDTA, 15% dimethyl sulfoxide, pH 7.0) was stirred into 4–6 mL of 1.0 M KCl, 5 mM CaCl₂, 1 mM EDTA, and 5 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (pH 7.2) on one side of the chamber. The opposite side, containing the same aqueous solution, was held at virtual ground by the voltage clamp. All experiments were performed at approximately 23°C.

Immunostimulatory assay. Native BALB/c mouse splenocytes (5 10⁵ cells/assay) were treated with increasing concentrations (0.156–40 M) of either L-G3139 or D-G3139 at 37°C with 5% CO₂. Control cells were treated with either10 g/mL concanavalin A or lipopolysaccharide. At the end of either 24 or 48 h, the supernatants were assayed by ELISA for the presence of IL-6, IL-12, or IP10, according to the manufacturer's instruction.

Animal studies. A375 melanoma cells (9 10⁶) were injected subcutaneously into the flanks of Fox-Chase SCID mice. These mice lack B and T cells, and are also V(D)J recombinant deficient. Tumors were allowed to grow to approximately 100 mm³, at which point either D-G3139 or L-G3139 was injected IP (10 animals/group, either 10 mg/kg every day for 14 days, or 20 mg/kg 7 every other day for 14 days in 100 L normal saline). Normal saline was also used as a control injection. Tumors were measured twice weekly in two dimensions, and tumor volumes were calculated by the formula ((width)² length)/2. Mice were observed daily, and signs of stress were assessed by loss of body weight (no group had >10% body weight loss), lethargy, and ruffled coat. Mice were sacrificed when tumor volumes reached 1500 mm³. (Data are presented as the mean tumor volumeSD, as determined by a two-sided Student's t-test with unequal variances.)

In another set of experiments, eight Fox-Chase SCID mice were injected with either D-G3139, L-G3139, or saline control at a dose of 20 mg/kg 7 every other day for 14 days. At the end of this time, serum samples were obtained and assayed by ELISA for IL-12 and IFN- levels according to the manufacturer's protocol.

References

1. Saenz-Santamaria, MC, Reed, JA, McNutt, NS and Shea, CR 1994. Immunohistochemical expression of BCL-2 in melanomas and intradermal nevi. J Cutan Pathol 21: 393–397. | Article | PubMed | ChemPort |
2. Plettenberg, A et al. 1995. Human melanocytes and melanoma cells constitutively express the Bcl-2 proto-oncogene in situ and in cell culture. Am J Pathol 146: 651–659. | PubMed | ISI | ChemPort |
3. Cerroni, L, Soyer, HP and Kerl, H 1995. bcl-2 protein expression in cutaneous malignant melanoma and benign melanocytic nevi. Am J Dermatopathol 17: 7–11. | PubMed | ISI | ChemPort |
4. Miyashita, T and Reed, JC 1993. Bcl-2 oncoprotein blocks chemotherapy-induced apoptosis in a human leukemia cell line. Blood 81: 151–157. | PubMed | ISI | ChemPort |
5. Kamesaki, S et al. 1993. bcl-2 protein inhibits etoposide-induced apoptosis through its effects on events subsequent to topoisomerase II-induced DNA strand breaks and their repair. Cancer Res 53: 4251–4256. | PubMed | ISI | ChemPort |
6. Teixeira, C, Reed, JC and Pratt, MA 1995. Estrogen promotes chemotherapeutic drug resistance by a mechanism involving Bcl-2 proto-oncogene expression in human breast cancer cells. Cancer Res 55: 3902–3907. | PubMed | ISI | ChemPort |
7. Jansen, B et al. 1998. bcl-2 antisense therapy chemosensitizes human melanoma in SCID mice. Nat Med 4: 232–234. | Article | PubMed | ISI | ChemPort |
8. Jansen, B et al. 2000. Chemosensitisation of malignant melanoma by BCL2 antisense therapy. Lancet 356: 1728–1733. | Article | PubMed | ISI | ChemPort |
9. Benimetskaya, L et al. 2004. Relative Bcl-2 independence of drug-induced cytotoxicity and resistance in 518A2 melanoma cells. Clin Cancer Res 10: 8371–8379. | Article | PubMed | ChemPort |
10. Lai, JC et al. 2005. Phosphorothioate oligodeoxynucleotides and G3139 induce apoptosis in 518A2 melanoma cells. Mol Cancer Ther 4: 305–315. | PubMed | ChemPort |
11. Kirkwood, J, Bedikian, A and Millward, M 2005. Long term sruvival results of a randomized multinational phase 3 trial of dacarbazine (DTIC) with or without Bcl-2 antisense (oblimersen sodium) in patients (pts) with advanced malignant melanoma. Proc Am Soc Clin Oncol 23: 711.
12. Benimetskaya, L, Miller, P and Stein, CA 2005. 518A2 melanoma cells are protected by G3139 and other antineoplastic agents against the cytotoxic effects of DTIC. Oligonucleotides 15: 206–214. | Article | PubMed | ChemPort |
13. Hu, Y et al. 2004. Antitumor efficacy of oblimersen Bcl-2 antisense oligonucleotide alone and in combination with vinorelbine in xenograft models of human non-small cell lung cancer. Clin Cancer Res 10: 7662–7670. | Article | PubMed | ChemPort |
14. Ramanarayanan, J, Hernandez-Ilizaliturri, FJ, Chanan-Khan, A and Czuczman, MS 2004. Pro-apoptotic therapy with the oligonucleotide genasense (oblimersen sodium) targeting Bcl-2 protein expression enhances the biological anti-tumour activity of rituximab. Br J Haematol 127: 519–530. | Article | PubMed | ChemPort |
15. Tauchi, T et al. 2003. BCL-2 antisense oligonucleotide genasense is active against imatinib-resistant BCR-ABL-positive cells. Clin Cancer Res 9: 4267–4273. | PubMed | ChemPort |
16. Mu, Z, Hachem, P and Pollack, A 2005. Antisense Bcl-2 sensitizes prostate cancer cells to radiation. Prostate 65: 331–340. | Article | PubMed | ChemPort |
17. Castanotto, D and Scherer, L 2005. Targeting cellular genes with PCR cassettes expressing short interfering RNAs. Methods Enzymol 392: 173–185. | PubMed | ChemPort |
18. Lai, JC et al. 2006. A pharmacologic target of G3139 in melanoma cells may be the mitochondrial VDAC. Proc Natl Acad Sci USA 103: 7494–7499. | Article | PubMed | ChemPort |
19. Siskind, LJ, Kolesnick, RN and Colombini, M 2002. Ceramide channels increase the permeability of the mitochondrial outer membrane to small proteins. J Biol Chem 277: 26796–26803. | Article | PubMed | ISI | ChemPort |
20. Gekeler, V et al. 2006. G3139 and other CpG-containing immunostimulatory phosphorothioate oligodeoxynucleotides are potent suppressors of the growth of human tumor xenografts in nude mice. Oligonucleotides 16: 83–93. | Article | PubMed | ChemPort |
21. Wacheck, V et al. 2002. Antitumor effect of G3139 Bcl-2 antisense oligonucleotide is independent of its immune stimulation by CpG motifs in SCID mice. Antisense Nucleic Acid Drug Dev 12: 359–367. | Article | PubMed | ChemPort |
22. Eulberg, D and Klussmann, S 2003. Spiegelmers: biostable aptamers. Chembiochem 4: 979–983. | Article | PubMed | ISI | ChemPort |
23. Vater, A and Klussmann, S 2003. Toward third-generation aptamers: Spiegelmers and their therapeutic prospects. Curr Opin Drug Discov Devel 6: 253–261. | PubMed | ChemPort |
24. Guvakova, MA et al. 1995. Phosphorothioate oligodeoxynucleotides bind to basic fibroblast growth factor, inhibit its binding to cell surface receptors, and remove it from low affinity binding sites on extracellular matrix. J Biol Chem 270: 2620–2627. | Article | PubMed | ISI | ChemPort |
25. Benimetskaya, L et al. 1995. Binding of phosphorothioate oligodeoxynucleotides to basic fibroblast growth factor, recombinant soluble CD4, laminin and fibronectin is P-chirality independent. Nucleic Acids Res 23: 4239–4245. | Article | PubMed | ChemPort |
26. Rockwell, P et al. 1997. Cell-surface perturbations of the epidermal growth factor and vascular endothelial growth factor receptors by phosphorothioate oligodeoxynucleotides. Proc Natl Acad Sci USA 94: 6523–6528. | Article | PubMed | ChemPort |
27. Lebedeva, I and Stein, CA 2001. Antisense oligonucleotides: promise and reality. Annu Rev Pharmacol Toxicol 41: 403–419. | Article | PubMed | ISI | ChemPort |
28. Kleinman, HK et al. 1982. Isolation and characterization of type IV procollagen, laminin, and heparan sulfate proteoglycan from the EHS sarcoma. Biochemistry 21: 6188–6193. | Article | PubMed | ISI | ChemPort |
29. Hodge, T and Colombini, M 1997. Regulation of metabolite flux through voltage-gating of VDAC channels. J Membr Biol 157: 271–279. | Article | PubMed | ISI | ChemPort |
30. Lee, AC, Zizi, M and Colombini, M 1994. Beta-NADH decreases the permeability of the mitochondrial outer membrane to ADP by a factor of 6. J Biol Chem 269: 30974–30980. | PubMed | ISI | ChemPort |
31. Tan, W, Lai, JC, Miller, P, Stein, CA and Colombini, M 2006. Phosphorothioate oligonucleotides reduce mitochondrial outer membrane permeability to ADP. Submitted.
32. Rostovtseva, T and Colombini, M 1997. VDAC channels mediate and gate the flow of ATP: implications for the regulation of mitochondrial function. Biophys J 72: 1954–1962. | PubMed | ISI | ChemPort |
33. Colombini, M 2004. VDAC: the channel at the interface between mitochondria and the cytosol. Mol Cell Biochem 256–257: 107–115.
34. Dejean, L et al. 2005. Oligomeric bax is a component of the putative cytochrome c release channel MAC, mitochondrial apoptosis-induced channel. Mol Biol Cell 16: 2424–2432. | Article | PubMed | ISI | ChemPort |
35. Lervolino, A, Trisciuoglio, D and Ribatti, D 2002. Bcl-2 overexpression in human melanoma cells increases angiogenesis through VEGF mRNA stabilization and HIF-1 mediated transcriptional activity. FASEB J 16: 1453–1456. | PubMed | ChemPort |
36. Biroccio, A, Candiloro, A and Mottolese, M 2000. Bcl-2 overexpression and hypoxia synergistically act to modulate vascular endothelial growth factor expression and in vivo angiogenesis in a breast carcinoma line. FASEB J 14: 652–660. | PubMed | ISI | ChemPort |
37. Trisciuoglio, D, Desideri, M and Ciuffreda, L 2005. Bcl-2 overexpression in melanoma cells increases tumor progression-associated properties and in vivo tumor growth. J Cell Physiol 205: 414–421. | Article | PubMed | ChemPort |
38. Karl, EWK and Zeitlin, B 2005. Bcl-2 acts in a proangiogenic signaling pathway through nuclear factor-kB and CXC chemokines. Cancer Res 65: 5063–5069. | Article | PubMed | ChemPort |
39. Benimetskaya, L et al. 2006. Bcl-2 protein in 518A2 melanoma cells in vivo and in vitro. Clin Cancer Res 12: 4940–4948. | Article | PubMed | ChemPort |
40. Krieg, A 2001. From bugs to drugs: therapeutic immunomodulation with oligodeoxynucleotides containing CpG sequences from bacterial DNA. Antisense Nucleic Acid Drug Dev 11: 181–188. | Article | PubMed | ChemPort |
41. Akira, S 2003. Mammalian toll-like receptors. Curr Opin Immunol 15: 5–11. | Article | PubMed | ISI | ChemPort |
42. Hornung, V et al. 2002. Quantitative expression of toll-like receptor 1–10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligonucleotides. J Immunol 168: 4531–4537. | PubMed | ISI | ChemPort |
43. Balsari, A et al. 2004. Combination of a CpG-oligonucleotide and a tpoisomerase I inhibitor in the therapy of human tumor xenografts. Eur J Cancer 40: 1275–1281. | Article | PubMed | ChemPort |
44. Pratesi, G et al. 2005. Therapeutic synergism of gemcitabine and CpG-oligonucleotides in an orthotopic human pancreatic carcinoma xenograft. Cancer Res 65: 6388–6393. | Article | PubMed | ChemPort |
45. Lai, J, Benimetskaya, L, Santella, RM, Wang, Q, Miller, PS and Stein, CA 2003. G3139 (oblimersen) may inhibit prostate cancer cell growth in a partially bis-CpG-dependent non-antisense manner. Mol Cancer Ther 2: 1031–1043. | PubMed | ISI | ChemPort |
46. Yakubov, L et al. 1993. Oligodeoxynucleotides interact with recombinant CD4 at multiple sites. J Biol Chem 268: 18818–18823. | PubMed | ISI | ChemPort |
47. Freitag, H, Benz, R and Neupert, W 1983. Isolation and properties of the porin of the outer mitochondrial membrane from Neurospora crassa. Methods Enzymol 97: 286–294. | PubMed | ChemPort |
48. Blachly-Dyson, E, Peng, S, Colombini, M and Forte, M 1990. Selectivity changes in site-directed mutants of the VDAC ion channel: structural implications. Science 247: 1233–1236. | Article | PubMed | ChemPort |
49. Lee, AC, Xu, X and Colombini, M 1996. The role of pyridine dinucleotides in regulating the permeability of the mitochondrial outer membrane. J Biol Chem 271: 26724–26731. | PubMed | ChemPort |
50. Colombini, M 1987. Characterization of channels isolated from plant mitochondria. Methods Enzymol 148: 465–475. | PubMed | ChemPort |