|Home | About | Journals | Submit | Contact Us | Français|
We have evaluated the efficacy of the multinuclear platinum chemotherapeutics BBR3464, BBR3571, and BBR3610 against glioma cells in culture and animal models and investigated their mechanism of action at the cellular level. In a clonogenic assay, BBR3610, the most potent compound, had an IC90 dose (achieving 90% colony formation inhibition) that was 250 times lower than that of cisplatin for both LNZ308 and LN443 glioma cells. In subcutaneous xenografts of U87MG glioma cells, BBR3610 approximately doubled the time it took for a tumor to reach a predetermined size and significantly extended survival when these cells were implanted intracranially. Analysis of apoptosis and cell cycle distribution showed that BBR compounds induced G2/M arrest in the absence of cell death, while cisplatin predominantly induced apoptosis. Interestingly, the BBR compounds and cisplatin both induced extracellular signal-regulated kinase 1/2 phosphorylation, and inhibition of this pathway at the level of MEK antagonized the induction of G2/M arrest or apoptosis, respectively. Analysis of Chk1 and Chk2 status did not show any differential effects of the drugs, and it is thus unlikely to underlie the difference in response. Similarly, the drugs did not differentially modulate survivin levels, and knockdown of survivin did not convert the response to BBR3610 to apoptosis. Together, these findings support continued development of BBR3610 for clinical use against glioma and provide a framework for future investigation of mechanism of action.
Gliomas are a significant health problem for which new therapies are needed. Currently, emphasis is being placed on promising new compounds that target individual signal transduction pathways. However, until such therapies result in significant clinical advances, the rationale for studying more broadly targeted drugs, such as platinum chemotherapeutics, remains strong. The effectiveness of this class of drugs is not predicated on the presence of a particular molecular target or on a specific mutation (Lynch et al., 2004; Paez et al., 2004). Conventional platinum compounds, such as cisplatin and oxaliplatin, have not found widespread use for glioblastoma because they lack efficacy and because of problems with delivery to the tumor. Here we investigate members of the polynuclear platinum compounds, a class of compounds that is structurally distinct from cisplatin and whose clinical profile and mechanism of action are different from those of the established platinum compounds (Farrell, 2004; Perego et al., 1999; Servidei et al., 2001).
The prototype of this class, the trinuclear BBR3464 (Fig. 1), was shown to be effective against cells with inherent or acquired resistance to cisplatin, including glioma cells (Pratesi et al., 1999; Servidei et al., 2001), which suggests important differences in their mode of action. Cells with acquired resistance to cisplatin showed no cross-resistance to BBR3464 or differences in its uptake and cellular metabolism (Perego et al., 1999; Roberts et al., 1999a; Servidei et al., 2001). In addition, DNA mismatch repair status, which has been implicated as a major determinant of sensitivity to cisplatin, played no role in the response to BBR3464 (Perego et al., 1999). Another indication that multinuclear platinum agents are biologically distinct comes from the analysis of a panel of 60 cell lines from NCI, which showed that BBR3464 was more potent than cisplatin and revealed that the pattern of response did not match that of any other tested drug, including other platinum compounds (Manzotti et al., 2000). Analysis of DNA adduct formation in vitro showed that BBR3464 induced interstrand cross-links in approximately 20% of the DNA adducts, or about three times as many as cisplatin (Brabec et al., 1999), and these interstrand cross-links are also structurally different (Hegmans et al., 2004; Kasparkova et al., 2002), occurred at lower levels, and were more persistent than those formed by cisplatin (Perego et al., 1999). Furthermore, the adducts formed by BBR3464 are not recognized by antibodies against cisplatin adducts (Brabec et al., 1999) or high-mobility-group proteins (Kasparkova et al., 2002; Zehnulova et al., 2001). Last, an important basis for the speculation that the polynuclear platinum complexes will be more effective than conventional platinum compounds derives from their amphiphilic nature, due to their lipophilic alkanediamine chains and hydrophilic platinum-amine coordination spheres. A priori, this feature is likely to enhance membrane permeability in comparison to that for cisplatin or oxaliplatin. In fact, when compared to conventional platinum chemotherapeutics, BBR compounds show increased rates of cellular uptake and more rapid formation of DNA adducts (Harris et al., 2005, 2006; Roberts et al., 1999b).
BBR3464 has been tested and shown promising results in phase 1 and 2 clinical trials for late-stage solid tumors: The phase 2 trials in cisplatin-resistant and cisplatin- refractory ovarian cancer showed a number of confirmed partial responses; in contrast, the drug was not well tolerated in gastric cancer trials, and no great efficacy was seen (Farrell, 2004; Gourley et al., 2004; Jodrell et al., 2004; Sessa et al., 2000). Rapid metabolic decomposition may explain the low therapeutic index of BBR3464, and full clinical use is thus likely to depend on development of a more suitable pharmacokinetic profile (Oehlsen et al., 2003). Here, we are predominantly investigating newer members of the polynuclear platinum family, which have not yet been tested in the clinic: dinuclear complexes with polyamine linkers based on spermidine (BBR3571) and spermine (BBR3610) (Fig. 1). These second-generation polynuclear compounds have the same essential impact at the cellular and molecular levels, with DNA-binding profiles and cellular responses that are very similar to those of BBR3464 (Farrell, 2000; McGregor et al., 2002; Roberts et al., 1999a), and show cross-resistance with BBR3464 (Roberts et al., 1999a), which suggests common elements in their mechanism of action. In the work presented here, we focus on BBR3610, which has the greatest potency against glioma cells in culture and in xenograft models, and we provide evidence that it is more effective in the treatment of glioma in preclinical models than cisplatin or BBR3464. In addition, we show that polynuclear platinum agents in general induce predominantly G2/M arrest in glioma cells, an effect that is mediated by extracellular signal-regulated kinase (ERK)3 activation, as is the induction of apoptosis by cisplatin. This is the first direct evidence that the cellular response to polynuclear platinum compounds is distinct from the response to cisplatin, although these drugs make use of some of the same signal transduction pathways. These data support the further development of BBR3610 as a novel and potent agent for the treatment of brain tumors.
LNZ308 glioma cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Grand Island, N.Y.) supplemented with 10% fetal bovine serum; LN443 cells were grown in DMEM with 5% fetal bovine serum; all cells were grown in a 7.5% CO2 humidified incubator. Cisplatin (Sigma, St. Louis, Mo.) was prepared fresh in dimethyl sulfoxide (DMSO) at a 1000× working solution and diluted in medium before use, the DMSO concentration being kept at 0.1%. The BBR compounds were dissolved in water and stored at −80°C until use. MAPK/ERK kinase (MEK) inhibitors UO126 and PD98059 (Promega, Madison, Wis.) were reconstituted in DMSO at 10 mM and 50 mM, respectively. For MEK inhibition studies, cells were first pretreated for 30 min with MEK inhibitor or DMSO control and then treated with platinum drugs in the presence of MEK inhibitor or DMSO control.
For clonogenic assays, glioma cells were plated at low density and allowed to attach overnight. Cells were treated with various concentrations of drug as indicated. After 24 h, the cells were given fresh media. Cells were allowed to grow for one to two weeks before being fixed with ice-cold methanol and stained with Giemsa’s stain. A cluster of at least 50 cells was scored as a colony. Surviving fraction was calculated by number of colonies/number of cells plated × CFE, where CFE is the colony-forming efficiency in the absence of drug, calculated from control untreated cultures. For water-soluble tetrazolium salt (WST) assays, cells were plated at 500 cells per well in a 96-well plate and allowed to attach overnight. Cells were treated with indicated doses for 24 h. After one week, cells were either analyzed with WST or replated for additional time. For analysis, 10 μl of WST-1 reagent (Roche, Indianapolis, Ind.) was added to each well, the cells were incubated for 4 h, and the absorbance was read at 450 nm in a plate reader. For replating assays, cells from the 96-well plate were instead trypsinized and replated in a 1:1 ratio into a 48-well plate and allowed to grow for an additional four days. Cells were then analyzed by adding 20 μl of WST-1 reagent per well and incubated for 4 h, and the absorbance was read at 450 nm in a plate reader.
All of the animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Henry Ford Hospital, where these studies were carried out. To determine the efficacy of the BBR compounds in vivo, tumors were implanted in nude mice and monitored for growth. Female nude mice (CD1 nu/nu; Charles River Laboratory, Wilmington, Mass.) were implanted subcutaneously with U87MG glioma cells at a density of 2 × 106 cells per flank in both flanks, in PBS mixed 1:1 with basement membrane matrix (Matrigel; BD Biosciences, San Diego, Calif.) in a final volume of 200 μl. Female nude mice were implanted with U87 cells expressing green fluorescent protein (GFP; a gift from Donna Senger and Peter Forsyth, University of Calgary) intracranially at a density of 3 × 105 cells per brain in 5 μl of sterile PBS. Mice were randomized into groups, and treatment was initiated one week after tumor implantation and continued once a week for a total of three treatments. Drugs were administered through injection of the tail vein. Cisplatin (at 1 mg/ml; American Pharmaceutical Partners, Inc., Schaumberg, Ill.) was administered at a dose of 6 mg/kg, the dose for BBR3464 was 0.35 mg/kg, for BBR3571 the dose was 0.2 mg/kg, and the dose for BBR3610 was 0.1 mg/kg (all in PBS). Subcutaneous tumors were measured regularly, and the volume was calculated by the formula volume (mm3) = a2 × b/2, where a is the shortest diameter and b the longest diameter, in millimeters. Data on tumors that could not be followed adequately because the contralateral tumor forced the sacrifice of the animal were eliminated from the analysis. Intracranial tumors were monitored by using blue light, and mice were sacrificed when they showed signs of tumor burden according to the IACUC-approved protocol, and the day was recorded. Mice that survived to the end of the experiment were euthanized, and their brains were dissected and examined under blue light, which is capable of revealing a few hundred viable GFP-expressing cells, to determine whether any viable tumor cells were present. In each of the treatment groups in the intracranial experiment, two mice failed to show any signs of tumor reflecting the take rate of approximately 75% that we routinely encounter in intracranial injections with this cell line, and these mice were removed from the experiment. Results were analyzed by using Prism software (GraphPad, San Diego, Calif.).
Cells were cultured and treated with drugs as indicated in the Results section and then harvested for analysis. Both floating cells in the medium and live cells on the plate were collected. Attached cells were trypsinized and combined with detached cells, pelleted by centrifugation at 208 g for 5 min, rinsed with PBS, and repelleted. Cells were fixed by adding 1 ml of ice-cold 70% ethanol dropwise, while vortexing, and incubated at 4°C for at least 1 h. Fixed cells were pelleted by centrifugation at 208 g for 5 min, rinsed with PBS, and repelleted. Cells were incubated with 100 μl RNase (1 mg/ml in PBS; Qiagen, Valencia, Calif.) at 37°C for 30 min, and 100 μl propidium iodide (PI) (50 μg/ml) was then added. Cells were analyzed by fluorescence-activated cell sorting (FACSCalibur, BD Biosciences) and CellQuest software (BD Biosciences) for apoptosis analyses, and with ModFit software (Verity, Sunnyvale, Calif.) for cell cycle analyses.
For analysis of apoptosis-related proteins, LNZ308 cells were treated at 10 × IC90, the concentration at which colony formation is inhibited at the 90% level, and lysates were collected at 18 h post-treatment. Cells were scraped into media that contained detached cells and collected by centrifugation (208 g for 5 min), rinsed with PBS, and recentrifuged. Cell pellets were resuspended with lysis buffer containing 30 mM tris (pH 7.5), 150 mM NaCl, 10% glycerol, and 1% Triton X-100, with 1X protease cocktail (Complete Cocktail, Roche) and phosphatase inhibitor cocktail II (Sigma) added before use. The cells were then lysed with an 18-gauge needle, incubated on ice for 15 min, and cleared by centrifugation. Laemmli sodium dodecyl sulfate sample buffer was added; samples were heated to 70°C for 5 min and stored at −20°C until use. Samples were run on NuPage BT gels (Invitrogen, Carlsbad, Calif.), transferred to polyvinylidene difluoride membranes (Pall, East Hills, N.Y.), blocked with 5% bovine serum albumin unless otherwise indicated below, and incubated in primary antibody at indicated dilutions overnight at 4°C. Membranes were washed with tris-buffered saline with 0.1% Tween three times for 5 min each. Secondary antibodies used were antimouse or antirabbit conjugated with horseradish peroxidase (1:3000), developed with Super-Signal West Pico substrate (Pierce, Rockford, Ill.), and exposed to film. The following antibodies were used: PKCδ, (C-20, 1:1000 blocked with 5% nonfat dry milk [Santa Cruz Biotechnology, Santa Cruz, Calif.]); PARP (clone 42, 1:500 [BD Biosciences]); caspase 3 (clone 19, 1:1000 [BD Biosciences]); actin (AC-15, 1:5000 [Sigma]); Bcl-2 (clone 7, 1:500 [BD Biosciences]); Bad (clone 48, 1:500 [BD Biosciences]); Bax (clone 3, 1:250 [BD Biosciences]); Chk1, Chk2, and phospho-Chk1 (Ser 317, rabbit, 1:500, 5% nonfat milk [Cell Signaling, Beverly, Mass.]); survivin (clone 6E4, 1:1000, 5% milk [Cell Signaling]); and phospho-survivin (Thr 34, rabbit, 1:200, 5% milk [Santa Cruz Biotechnology]). For analysis of ERK, LNZ308 cells were treated with IC90 and 10-fold IC90 concentrations of drug for various lengths of time, from 15 min to 6 h. Plates were rinsed with PBS and cell lysates collected into lysis buffer as above. Lysates were resolved on NuPage BT gels as above, blotted to polyvinylidene difluoride membranes, and blocked with 5% milk. Primary antibodies used were anti-p44/p42 MAPK (clone 3A7) and anti-phospho p44/p42 MAPK (clone E10, both from Cell Signaling, used at 1:2000). Secondary antibody used was antimouse-HRP (1:3000), and blots were developed as above.
Cells were plated into six-well dishes and allowed to attach overnight. Cell density was at 50% to 60%, and 100 nM survivin or control siRNA (Cell Signaling) was transfected with Lipofectamine 2000 (Invitrogen). Forty-eight hours after transfection, cells were lysed for Western blot analysis or treated with platinum drugs as indicated. Twenty-four hours after drug treatment, cells were collected and stained with PI for FACS analysis as described above.
To determine the potency of BBR compounds relative to cisplatin and to one another, glioma cells were plated at clonal density, exposed to drug for 24 h, and cultured for up to two weeks. Colonies of more than 50 cells were counted, and the surviving fraction was calculated relative to untreated cells (Fig. 2). LNZ308 and LN443 glioma cells showed similar levels of sensitivity to drug and were more sensitive to BBR compounds than to cisplatin (compare Fig. 2A with Fig. 2B and C; please note differences in scale). Responses to BBR3464 and BBR3571 were similar (Fig. 2B), and sensitivities to BBR3610 were particularly pronounced (Fig. 2C). From the curves in Fig. 2, we estimated the concentration at which colony formation was inhibited at the 90% level (IC90) to be 2 μM for cisplatin, 0.1 μM for BBR3464 and BBR3571, and 8 nM for BBR3610. These concentrations were used as the basis for other experiments because clonogenic assays, which measure both cell death and growth arrest, provide the most comprehensive indication of a drug’s impact on the cell. On the basis of these data, BBR3464 and BBR3571 can be said to be 20 times more effective than cisplatin at inhibiting colony formation at this threshold, and BBR3610, 250 times more effective.
Activity in vivo is a prerequisite for the evaluation of the clinical potential of chemotherapeutics, and we therefore next tested the ability of BBR compounds to impact the growth of human gliomas in nude mice. In these experiments, we used U87MG glioma cells because these had a higher efficiency for generating tumors than did LNZ308 or LN443 cells. (U87MG cells were not used for clonogenicity assays because their highly motile phenotype prevents them from forming readily distinguishable colonies.) First, we tested the effect of drug treatment on the growth of subcutaneous tumors because they allowed the monitoring of tumor growth during the experiment. Treatment was initiated one week after tumor cells were implanted. The dosages and regimen were based on previous work with three treatments being given intravenously one week apart at the following dosages: cisplatin, 6 mg/kg; BBR3464, 0.35 mg/kg; BBR3571, 0.2 mg/kg; and BBR3610, 0.1 mg/kg (Manzotti et al., 2000; Pratesi et al., 1999; Riccardi et al., 2001). In the case of BBR3571 and BBR3610, previous animal experiments had shown that these dosages were tolerated (N.P.F., unpublished data). Tumor growth was then monitored, and the day that tumors exceeded a fixed size was determined. This approach was used because tumors treated with the BBR compounds would often show ulceration, perhaps due to drug treatment, necessitating the euthanasia of the animal at earlier times than controls. The results showed that the BBR compounds were effective at delaying tumor growth and that BBR3610 was more effective than cisplatin or BBR3464, nearly doubling the time it took for tumors to reach the threshold (Fig. 3A).
We next tested the efficacy of the BBR compounds by using a more relevant intracranial model. Because of its effectiveness in the subcutaneous model, we focused on BBR3610. U87MG cells that express GFP were used to enable tumor monitoring during and after the experiment. As in the subcutaneous model, treatment was initiated one week after implantation, and the doses used were 6 mg/kg for cisplatin and 0.1 mg/kg for BBR3610. Mice were sacrificed when tumor burden induced morbidity, or at 72 days. The brains were dissected and analyzed for GFP-expressing cells to confirm the presence of viable tumor cells. Although there was no significant difference between control mice and those treated with cisplatin, BBR3610 significantly prolonged median survival from 23 to 41.5 days (Fig. 3B). These experiments indicate that the BBR compounds, particularly BBR3610, are more effective than cisplatin both in vitro and in vivo and therefore have potential as a new therapy in the treatment of glioma. Furthermore, the experiments with intracranial xenografts suggest that BBR3610 may be able to cross the tumor-associated blood-brain barrier and so be useful in treating brain tumors. To understand the mechanism of action of the BBR compounds, we examined the cellular signaling responses of the BBR compounds in comparison with those of cisplatin.
A more rapid and high-throughput method than the clonogenic assay was needed for the measurement of cell survival in these experiments to determine the basis for differences in cellular response to different platinum compounds. Therefore, we established the WST-1 assay, which is similar to the MTT cell-proliferation assay in that it measures the production of chromogenic products from tetrazolium salts, but these are water soluble (Ishiyama et al., 1996). We observed good concordance between the data from WST and clonogenic assays for cisplatin, but not for BBR compounds. Our initial experiments used the IC90 concentrations from the clonogenic assays as the midpoint of a five-dose range, at 10-fold intervals (Fig. 4A). Treatment of cells with cisplatin at the IC90 resulted in approximately 20% of the control signal, while essentially no signal was obtained at 10-fold or 100-fold IC90 (Fig. 4A). The results for the BBR compounds were significantly different. Although the signal obtained at the IC90 was statistically different from the untreated controls, they were at most 50% lower (Fig. 4A). Furthermore, there was no additional decline in signal when the drug concentration was increased to 10-fold or 100-fold IC90. As the WST assay measures viable cells but does not distinguish between growth-arrested, nonclonogenic cells and proliferating, clonogenic cells, we hypothesized that the cells remaining after BBR compound treatment belonged to the former category. To test this, we passaged cells after drug treatment and gave them the opportunity to grow after replating, a measure of their clonogenic potential at the population level. This experiment showed that there was now no difference between cell numbers obtained after treatment with cisplatin or BBR compounds at isotoxic drug concentrations and that cells behaved the same in the WST and clonogenic assays (Fig. 4B). Furthermore, these findings suggested that the impact of BBR compounds and that of cisplatin at the level of the cell were very different.
One possible explanation for the differences observed in the WST assay is a difference in the induction of apoptosis by cisplatin and BBR compounds. Apoptosis manifests itself at the level of DNA content. Thus, to test whether BBR compounds induced apoptosis, we performed FACS analysis of PI-stained cells after drug treatment. Exposure of LNZ308 glioma cells to cisplatin caused the appearance of a population in the sub-G1 region of the profile, where apoptotic cells are found (Fig. 5A). In contrast, no increase in the sub-G1 proportion of the population was found when cells were exposed to isotoxic doses of BBR compounds (summarized in Fig. 5B). To confirm that the increase in the sub-G1 region of the FACS profile was due to apoptosis, we examined the response of several proteins known to be proteolytically cleaved by the activation of the caspase cascade. Exposure of LNZ308 glioma cells to cisplatin for 18 h resulted in the cleavage of caspase 3, PARP, and PKCδ (Okhrimenko et al., 2005) (Fig. 5C). In contrast, isotoxic doses of the BBR compounds had no effect on these proteins, which suggests that the caspase cascade was not activated (Fig. 5C). A slight reduction in the level of the antiapoptotic Bcl-2 in these experiments was also observed in cisplatin-treated cells, but no elevation was observed in BBR-treated cells. No changes in the expression levels of the pro-apoptotic proteins Bad or Bax were seen after drug treatment (data not shown).
An explanation for the observation that BBR compounds reduced clonogenic potential but did not induce apoptosis could be that they mediate cell cycle arrest. To test this hypothesis, we analyzed cell cycle distribution following drug treatment by FACS. We found that BBR compounds caused a significant increase in the proportion of cells that were in the G2/M phase of the cell cycle (Fig. 6A and B), which persisted to 48 h (data not shown). Therefore, BBR compounds induce cell cycle arrest rather than apoptosis, which was induced by cisplatin at isotoxic doses. These data confirm the results seen with the WST-1 assay, in which the cells were alive but not clonogenic, as seen by their failure to proliferate upon replating.
One possible basis for the different cellular responses to cisplatin and BBR compounds could be that the adducts of these compounds activate different signaling events in the cell. Therefore, we next studied the activation of kinases that respond to cellular stress. Examination of the induction of JNK phosphorylation showed that even challenging cells with cisplatin or BBR compounds at concentrations that were 200-fold higher than the IC90 did not appreciably elevate phosphorylation, nor were there significant differences in levels of p38 phosphorylation (data not shown), which suggested that this pathway was unlikely to be an important component of the cellular response to drug in vivo. This is in agreement with the work of others, for example, the demonstration that 1000 μM cisplatin, or 500-fold our IC90, is required to observe phosphorylation of JNK in glioma cells (Potapova et al., 1997). We also did not see any changes in the level of Jun protein expression after treatment with cisplatin or multinuclear platinum compounds (data not shown). Analysis of ERKs, in contrast, revealed measurable activation at the IC90 or 10-fold IC90 concentrations of platinum compounds (Fig. 7A). An increase in ERK phosphorylation was seen as early as 30 min and at 6 h, which suggests that it is a rapid and prolonged response to treatment with these compounds. Quantification of the phospho-ERK (pERK) signal relative to ERK levels showed that the induction by isotoxic doses of BBR3571 and BBR3610 appeared to be higher than that by cisplatin or BBR3464 (Fig. 7B). These two drugs may share this aspect of their mechanism of action because they are chemically similar, both having polyamine linkers between the two platinum groups. Interestingly, at the 10-fold IC90 dose, both BBR3571 and BBR3610 caused an initially strong response at 30 min, which declined markedly, whereas at the IC90 dose, they induced a more sustained ERK activation. Overall, these data showed that induction of ERK activation was common to the cellular response to cisplatin and the cellular response to BBR compounds.
Cisplatin activation of ERKs has been shown to be important for its impact on cells. Thus, to test the role of this pathway in the action of BBR compounds, we used MEK inhibitors to selectively inhibit ERK activation. We found that the inhibitor UO126 was more effective at suppressing ERK activation than was PD98059 (Fig. 8A) and therefore used it in subsequent experiments. FACS analysis of cells after treatment of the cells with MEK inhibitors and platinum compounds was performed, and as before, the proportion of cells in each phase of the cell cycle was measured. As expected, the presence of the MEK inhibitor significantly reduced the levels of apoptosis induced by cisplatin (Fig. 8B and C). However, MEK inhibition also significantly reduced the proportion of cells in G2/M phase after exposure to BBR3610 (Fig. 8B and D). Therefore, MEK inhibitors antagonized the impact of both cisplatin and multinuclear platinum compounds, although ERK pathway activation was associated with very different effects at the cellular level in each case.
One potential reason for the differences in response to classes of platinum compounds could be that they differentially activate members of the Chk family of signaling molecules, which are known to mediate G2/M arrest by inhibiting the cdc25 phosphatase in response to genomic stress (Bartek and Lukas, 2003; Hirose et al., 2004, 2005). All platinum compounds induced phosphorylation of Chk1 at Ser 317 and Chk2 at Thr 68 in LNZ308 cells exposed to 10-fold IC90 concentrations of platinum drugs for 18 h (data not shown), which suggests that differential activation of Chk proteins is not the cause for the observed difference in cellular response to platinum drugs.
Survivin, a protein that is widely expressed in cancers and exerts an antiapoptotic function (Altieri, 2001, 2003), was examined as another potential underlying cause for the lack of apoptosis observed in the response to multinuclear platinum drugs. However, no difference in the levels or phosphorylation state of survivin in glioma cells treated with different platinum compounds was observed (data not shown).
Furthermore, siRNA mediated knockdown of survivin in LNZ308 glioma cells, which effectively reduced the expression of this protein at 48 h and increased the level of apoptosis in the culture, but did not convert the response to BBR3610 to apoptosis, at the 10-fold IC90, for another 24 h (data not shown). In contrast, cisplatin induced apoptosis on its own, and when combined with survivin siRNA, the level of apoptosis was higher than and statistically different from that obtained with either cisplatin or survivin siRNA alone (data not shown). Therefore, the absence of survivin potentiated the induction of apoptosis by cisplatin, but did not alter the response to BBR3610.
The need for new therapies for glioma is evident, and in the present study we have investigated polynuclear platinum compounds, and in particular BBR3610, for their ability to suppress glioma cell growth in culture and in xenografts. In a clonogenic assay, BBR3610 was 250 times as potent as cisplatin, and more than 10 times as potent as BBR3464. The response of glioma cells in these experiments was similar to that observed previously for cisplatin and BBR3464, taking cell line differences into account: In our experiments using LNZ308 and LN443 cells, the IC90 concentrations for these compounds were 2 μM and 0.1 μM, respectively, while the IC50 concentrations previously reported for these compounds in U87MG cells were 3.8 μM and 0.08 μM, respectively (Servidei et al., 2001). We did not use U87MG for clonogenic assays because, in our experience, their highly motile behavior in culture prevents the formation of isolated colonies that could be accurately scored. We did use U87MG cells in our xenograft model, and here we observed that BBR3610 induced a significant delay in tumor growth in a subcutaneous model and extended survival in an intracranial model. The animal experiments were pursued with a dose and regimen adapted from previous work on cisplatin and BBR3464 (Manzotti et al., 2000; Pratesi et al., 1999; Riccardi et al., 2001; White et al., 1995). Therefore, it is quite possible that the maximum dose and ideal schedule of administration have not yet been identified, and this warrants further investigation as an important step toward the clinical development of BBR3610 for glioma chemotherapy.
Apoptosis is a common response of cells to platinum compounds (Sorenson et al., 1990), and in agreement with this, we observed apoptosis as the primary effect of cisplatin on glioma cells in our experiments. However, none of the three BBR compounds elicited measurable apoptosis in the first 48 h after treatment. This was observed by direct examination of DNA content by FACS and by Western blotting of proteins that are cleaved by caspases. Analysis of cell number in WST-1 assays further demonstrated that apoptosis did not set in within a week of drug treatment with BBR compounds. Our finding is in agreement with results from previous reports, which showed that glioma cells do not have a robust apoptotic response to BBR3464, while neuroblastoma cells can be induced to undergo apoptosis at higher drug concentrations (Servidei et al., 2001). Instead of apoptosis, multinuclear platinum agents induced a G2/M cell cycle arrest, as measured by FACS analysis of DNA content and supported by the determination by WST-1 assay that there was no loss of cell viability after drug treatment. Again, this is in agreement with the observation that BBR3464 induces G2/M phase arrest in U87MG cells (Servidei et al., 2001). With these data taken together, the finding that BBR treatment induces G2/M arrest, but little apoptosis, is extended to BBR3571 and more significantly to BBR3610, which is the most promising of the BBR compounds for continued clinical development. It had previously been suggested in a study of p53 wild-type U87MG cells that differential induction of p53 underlies the difference in the response to BBR3464 and cisplatin: The former caused a modest induction of p53 with upregulation of p21 but not Bax, whereas the latter gave a robust induction of p53 as well as increased expression of Bax (Servidei et al., 2001). In our studies, we used LNZ308 cells, which are p53 null, and observed a similar difference in cell response to cisplatin and BBR compounds, suggesting that p53 is not directly responsible for this difference in response. No differences in Bax, Bad, or Bcl-2 expression were seen in our experiments. The results are in contrast to the caspase-induced apoptosis initiated by BBR3464 in primary and tumor mast cells (Farrell, 2000), which suggests that the extent and nature of the apoptotic response may be cell specific or tissue specific. Likewise, p53 response to cisplatin varies, depending on cell type (Siddik, 2003)
Another immediate consequence of treating cells with chemotherapeutics is activation of stress-related signaling pathways, and their importance to the outcome at the cellular level is increasingly recognized. Cisplatin can induce signaling pathways, such as the MAPK pathway, and the ERK1/2 members of the MAPK family in particular are activated after exposure to cisplatin and other toxic agents at doses relevant to their impact at the cellular level. However, there are conflicting reports of the consequences of ERK activation. Some studies show that activation of the ERK pathway is a mechanism of cell survival (Persons et al., 1999), whereas other reports show that ERK activation is involved in apoptosis following chemotoxic exposure (Bacus et al., 2001; Tang et al., 2002; Wang et al., 2000). Specifically for cisplatin, ERK has been shown to mediate either cell survival (Hayakawa et al., 1999; Persons et al., 2000) or cell death (Wang et al., 2000; Woessmann et al., 2002). This divergence in effect may be related to cell or cancer type or to the other oncogenic alterations that the cell has undergone, and it was for this reason that we characterized the response to multinuclear platinum compounds. Our data show that BBR compounds are as capable of inducing ERK activation as cisplatin and, in the case of BBR3571 and BBR3610, may even induce a stronger response. Using the MEK inhibitor UO126, we demonstrated that the activation of ERK1/2 was involved in mediating apoptosis in response to cisplatin and in mediating G2/M arrest in response to BBR compounds. These results demonstrate that the activation of the same pathway by two different drugs in the same cell type can have very distinct consequences. Furthermore, these results provide evidence that other, as yet undescribed signaling events, which are unique to cisplatin or the BBR family, may affect the outcome of ERK activation. We investigated the Chk1 and Chk2 kinases in this context because they have been implicated in cell cycle arrest downstream of drug treatment, but found no evidence for differential activation. We found that differential induction of the apoptosis regulators Bcl-2, Bad, and Bax was also not responsible. Similarly, no difference in the level or activity of survivin was observed following treatment with different drugs, and knocking down survivin expression did not convert the response of cells to BBR3610 to an apoptotic one. Therefore, the molecular differences that underlie the nature of the response to cisplatin and BBR compounds in glioma cells remain to be defined. A recent comparison by cDNA array of the response to BBR3464 of A431 cells and of a platinum-resistant variant showed that the genes induced were quite different: The parental cells, which underwent primarily G2/M arrest, upregulated genes involved in cell cycle regulation, while the platinum-resistant sub-line, which preferentially underwent apoptosis, favored genes related to the apoptosis pathway (Gatti et al., 2004). This indicates that the analogous comparison of cells exposed to cisplatin or BBR3610 may also show molecular signatures consistent with their response.
In summary, we have shown that BBR compounds are more effective than cisplatin against glioma cells in culture and in vivo, including in an intracranial model of glioma. While treatment with cisplatin induced apoptosis, BBR compounds caused the cells to undergo a G2/M arrest; interestingly, in both instances, these responses appeared to be mediated, at least in part, by activation of ERK1/2. Therefore, although the BBRs share some molecular details of their mechanism of action with cisplatin, the response at the cellular level is very distinct, and at the level of a tumor in an animal is even better defined.
We gratefully acknowledge Donna Senger and Peter Forsyth for the U87-GFP cells used in this study, and Chaya Brodie for helpful discussions.
1This work was supported by CA-108499 (O.B.) and CA-78754 (N.F.) from the National Cancer Institute and the generosity of the Hermelin Brain Tumor Center donors, and particular thanks are extended to William and Karen Davidson (O.B.).
3Abbreviations used are as follows: DMEM, Dulbecco’s modified Eagle’s medium, DMSO, dimethyl sulfoxide; ERK, extracellular signal-regulated kinase; FACS, fluorescence-activated cell sorting; GFP, green fluorescent protein; IACUC, Institutional Animal Care and Use Committee; IC90, the concentration at which colony formation is inhibited at the 90% level; MAPK, mitogen-activated protein kinase; MEK, MAPK/ERK kinase; pERK, phospho-ERK; PI, propidium iodide; WST, water-soluble tetrazolium salt.