Cytotoxicity of σ-Receptor Ligands Is Associated with Major Changes of Cellular Metabolism and Complete Occupancy of the σ-2 Subpopulation

Culture Medium and Drugs

Dulbecco's minimum essential medium (DMEM), fetal calf serum (FCS), and trypsin were products of Invitrogen. AC915, haloperidol, (+)-pentazocine, rimcazole, and trypan blue (0.4% solution in phosphate-buffered saline) were purchased from Sigma. AC915 was dissolved in water with slight heating, and haloperidol and rimcazole were dissolved in ethanol. Stock solutions of (+)-pentazocine were made in 0.1N hydrochloric acid.

Radiopharmaceuticals

¹¹C-methionine, ¹¹C-choline, ¹⁸F-FLT, and ¹⁸F-FDG were prepared by standard procedures reported in the literature (25). ¹¹C-SA4503 was prepared by reaction of ¹¹C-methyl iodide with the appropriate 4-O-methyl precursor (26). Specific radioactivity at the end of synthesis was greater than 22 TBq/mmol. All radiochemical purities were greater than 95%.

Cell Culture

C6 rat glioma cells obtained from the American Type Culture Collection were grown as monolayers in DMEM (high glucose) supplemented with 7.5% FCS in a humidified atmosphere of 5% CO₂/95% air at 37°C. Before each experiment, the cells were seeded in 12-well plates (Costar). An equal number of cells was dispensed in each well and was supplied with 1.1 mL of DMEM (high glucose) supplemented with 7.5% FCS.

Binding Studies

Binding studies were performed 48 h after seeding cells in 12-well plates, when confluency reached 80%−90%. Various concentrations of an unlabeled competitor (AC915, haloperidol, (+)-pentazocine, or rimcazole; range, 10⁻⁸ to 10⁻³ M, in triplicate) were dispensed to the culture medium in the wells. After 2 min, 4 MBq of ¹¹C-SA4503 in less than 30 μL of saline (containing 30% ethanol) were added to each well. After about 45 min of incubation, the medium was quickly removed, and the monolayer was washed 3 times with phosphate-buffered saline. Cells were then treated with 0.2 mL of trypsin. When the monolayer had detached from the bottom of the well, 1 mL of DMEM (high glucose) supplemented with 7.5% FCS was applied to stop the proteolytic action. Cell aggregates were resolved by repeated (at least 10-fold) pipetting of the trypsin/DMEM mixture. Radioactivity in the cell suspension (1.2 mL) was assessed using a γ-counter (Compugamma 1282 CS; LKB-Wallac). A sample of the suspension was mixed with trypan blue solution (1:1 v/v) and used for cell counting. Cell numbers were manually determined, using a phase-contrast microscope (Olympus), a Bürker bright-line chamber (depth, 0.1 mm; 0.0025 mm² squares), and a hand-tally counter.

Uptake of radioactivity normalized to the number of viable cells was plotted against the logarithm of the dose of the competing drug (in [M]). A 3-parameter curve (single-site competition model) was fitted to these data, using the following equation:where IC₅₀ is drug concentration resulting in 50% occupancy of the available binding sites. GraphPad Prism (GraphPad Software) was used for curve fitting and graphical presentation of data. Less than 6% (in most wells < < 5%) of the administered tracer dose was bound to the cells under these conditions.

Treatment with σ-Ligands and Scoring of Morphologic Effects

σ-Ligands were administered 24 h after seeding cells in 12-well plates when confluency had reached 40%−45%. Final concentrations in culture medium were 1, 3, 10, 30, and 100 (in some cases also 300) times the IC₅₀ value of the test drug for inhibition of ¹¹C-SA4503 binding to cellular σ-receptors (determined as described above).

After 20–24 h of incubation with σ-ligands, the cells were monitored by phase-contrast microscopy. The effect of each compound and concentration was scored relative to a control receiving only solvent vehicle. Scoring was performed as described previously (12). Experiments were performed in quadruplicate and repeated at least twice. Decreases of cell number after drug treatment were quantified by harvesting and resuspending cells in medium containing FCS, followed by manual counting as described above.

Uptake Studies

Cellular uptake of metabolic PET tracers was determined after 24 h of treatment with either vehicle or a σ-ligand. Experiments were performed in quadruplicate and repeated at least twice. At time zero, a PET tracer (4 MBq of ¹¹C-methionine or ¹¹C-choline or 2 MBq of ¹⁸F-FDG or ¹⁸F-FLT in <20 μL of saline) was added to each well. After 45 min of incubation, cells were washed, and cellular radioactivity was assessed as described above. Less than 1% (¹¹C-methionine), 2% (¹⁸F-FDG), or 4% (¹¹C-choline and ¹⁸F-FLT) of the administered tracer dose was taken up by the cells under these conditions.

Statistics

Differences between groups were examined using 1-way ANOVA. A P value of less than 0.05 was considered statistically significant.

Receptor Occupancy

Curves describing competition between the test drugs and the radioligand ¹¹C-SA4503 for cellular σ-receptors are presented in Figure 1. The data were best fitted by a 1-site competition model. IC₅₀ values were either in the submicromolar range (rimcazole, 0.36 μM; haloperidol, 0.27 μM) or in the micromolar range ((+)-pentazocine, 6.5 μM; AC915, 7.4 μM).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Dose-dependent increase of cellular ¹¹C-SA4503 binding by (+)-pentazocine (▪), haloperidol (□), AC915 (▾), and rimcazole (▿). Error bars indicate SEM. Specific binding was calculated by subtracting nonspecific binding (i.e., residual binding of radioligand in presence of highest dose of competing drug) from total binding; it amounted to 70%−80% of total radioactivity uptake by cells.

Subtype Involvement

Table 1 lists literature values for the affinities (K_d) of the test compounds to σ-1 and σ-2 receptors, determined by in vitro binding assays using membrane fragments isolated from rodent tissues. We observed much higher (60- to 1,500-fold) IC₅₀ values in intact cells than known K_d values at σ-1 receptors, particularly in the case of AC915, (+)-pentazocine, and haloperidol. However, IC₅₀ values of all test compounds corresponded closely to their K_d values at σ-2 receptors. Here, only minor (<3-fold) differences were observed. Therefore, IC₅₀ values for inhibition of ¹¹C-SA4503 binding to intact cells corresponded to the σ-2 K_d rather than to the σ-1 K_d of the test compounds.

Cytotoxicity

When the σ-antagonists rimcazole, haloperidol, or AC915 were applied to C6 cells in culture, a dose-dependent growth inhibition was observed after 24 h (Fig. 2). Low doses of the test drugs did not affect cellular growth. Above a certain threshold concentration (rimcazole, 20 μM; haloperidol, 25 μM; and AC915, 220 μM), a reduction in cell number became apparent. EC₅₀ values of the test compounds (effective concentrations resulting in 50% cell loss) were somewhat higher than the threshold for cytotoxicity (rimcazole, 31 μM; haloperidol, 58 μM; and AC915, 819 μM). Virtually complete cell death was observed after 24 h of treatment with either 100 μM rimcazole or 150 μM haloperidol (Fig. 2). Even 200 μM σ-1 agonist (+)-pentazocine did not yet affect cell division or viability. Here, the EC₅₀ value for growth inhibition was 710 μM.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

Dose-dependent cell loss after 24 h of incubation with (+)-pentazocine (▪), haloperidol (□), AC915 (▾), and rimcazole (▿). Note that x-axis scales of Figures 1 and 2 are different. Error bars indicate SEM.

Receptor Occupancy at Cytotoxic Dose

By comparing EC₅₀ values of the test compounds with their IC₅₀ values for inhibition of ¹¹C-SA4503 binding in intact cells, σ-receptor occupancies at cytotoxic doses of the test drugs could be estimated, using the curves presented in Figure 1 or the formula for a single-site competition model (see “Materials and Methods”). Because EC₅₀ values were about 2 orders of magnitude higher than the corresponding IC₅₀ values of the test drugs, cell death appeared to be associated with virtually complete (i.e., 99%) occupancy of the σ-receptor population (Table 2).

Changes of Cellular Metabolism

The metabolic PET tracers that we used responded differently to treatment of C6 cells with σ-receptor ligands. Cytotoxic doses of (+)-pentazocine, haloperidol, and rimcazole were generally associated with increases of the uptake of ¹⁸F-FDG, decreases of the uptake of ¹⁸F-FLT and ¹¹C-choline, and little change of ¹¹C-methionine uptake per viable cell (Figs. 3–6⇓⇓⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3. 

Dose-dependent increase of cellular ¹⁸F-FDG uptake by (+)-pentazocine (▪), haloperidol (□), AC915 (▾), and rimcazole (▿).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4. 

Dose-dependent inhibition of cellular ¹⁸F-FLT uptake by (+)-pentazocine (▪), haloperidol (□), AC915 (▾), and rimcazole (▿).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5. 

Dose-dependent inhibition of cellular ¹¹C-choline uptake by (+)-pentazocine (▪), haloperidol (□), AC915 (▾), and rimcazole (▿).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6. 

Dose-dependent inhibition of cellular ¹¹C-methionine uptake by (+)-pentazocine (▪), haloperidol (□), AC915 (▾), and rimcazole (▿). MET = methionine.

Significant increases of ¹⁸F-FDG uptake were observed only when a certain threshold concentration was exceeded. The threshold dose for haloperidol and rimcazole was 10 μM but in the case of (+)-pentazocine was 30 μM (Fig. 3). Maximal increases of ¹⁸F-FDG uptake (+111%, +150%, and +166% for haloperidol, rimcazole, and (+)-pentazocine, respectively; all statistically significant) occurred at doses of 100 μM for haloperidol, 112 μM for rimcazole, and 660 μM for (+)-pentazocine. Higher doses of test drug could not be examined, because these resulted in a complete loss of cells after 24 h. In contrast to the previously mentioned σ-ligands, AC915 in doses of up to 776 μM did not increase cellular ¹⁸F-FDG uptake. A slight but statistically significant decline (−21%) of ¹⁸F-FDG uptake was observed after incubation of cells with 220 μM AC915 (Fig. 3).

Treatment of C6 cells with rimcazole, (+)-pentazocine, and haloperidol resulted in a dose-dependent depression of ¹⁸F-FLT uptake when drug concentrations exceeded threshold values of 20, 50, and 80 μM, respectively (Fig. 4). Strong and statistically significant depression of ¹⁸F-FLT uptake was observed after a 24-h incubation of the cells with 50 (rimcazole), 100 (haloperidol), or 645 μM ((+)-pentazocine), to 6%, 17%, and 26% of the control, respectively. In contrast to the previously mentioned σ-ligands, AC915 did not depress cellular ¹⁸F-FLT uptake but caused a biphasic increase with a maximum occurring at a drug dose of 74 μM (Fig. 4).

Cellular uptake of ¹¹C-choline was depressed by all drug treatments (Fig. 5). A decline occurred when the ligand concentration of (+)-pentazocine and AC915 exceeded 70 μM (20 μM for haloperidol and 15 μM for rimcazole). At concentrations of 650 ((+)-pentazocine), 711 (AC915), or 45 μM (rimcazole), ¹¹C-choline uptake was reduced more than 80%, whereas a concentration of 60 μM for haloperidol resulted in a 55% decline (Fig. 5).

No statistically significant changes of ¹¹C-methionine uptake occurred after 24 h of treatment of C6 cells with (+)-pentazocine, haloperidol, or AC915 (Fig. 6). In contrast, rimcazole at concentrations greater than 10 μM depressed uptake of the amino acid up to 40% (Fig. 6).

Competition Assays

¹¹C-SA4503 has been reported to bind preferentially to σ-1 receptors (IC₅₀, 17.4 nM at the σ-1 and 1,784 nM at the σ-2 subtype (27); IC₅₀, 4.7 nM at σ-1 and 63.1 nM at σ-2 receptors in a later study (28)). Therefore, we expected that the potent σ-1 selective ligands (+)-pentazocine and AC915 would compete more efficiently (i.e., at lower IC₅₀ values) with cellular binding of ¹¹C-SA4503 than would the non–subtype-selective antagonists rimcazole and haloperidol, which bind less potently to σ-1 receptors. In reality, however, rimcazole and haloperidol inhibited ¹¹C-SA4503 binding more potently (IC₅₀ values, 0.36 and 0.27 μM, respectively) than did (+)-pentazocine and AC915 (IC₅₀ values, 6.5 and 7.4 μM, respectively; Fig. 1). The IC₅₀ values corresponded closely to the σ-2 K_d but not to the σ-1 K_d of the test compounds (Table 1).

This outcome of our competition assays can be understood when the applied mass of ¹¹C-SA4503 is considered. To have acceptable count statistics even at high doses of a competing drug after 60 min of incubation, it was necessary to add 4 MBq of ¹¹C-SA4503. At a specific radioactivity of 22 TBq/mmol, this corresponds to a ligand mass of 0.2 nmol and a ligand concentration of 200 nM in each well. The receptor K_d reported by Matsuno et al. (27) predict that at this radioligand concentration, saturation of σ-1 receptors will occur and specific binding will be observed mainly at the σ-2 subtype.

Thus, the methods described in this article investigated the relationship between cytotoxicity and σ-2 receptor occupancy, using the relatively low affinity of ¹¹C-SA4503 for σ-2 receptors. During in vivo scans with ¹¹C-SA4503 in humans, ligand concentrations in the 10⁻⁸ or 10⁻⁷ M range will never be reached. Human SA4503-PET scans will, therefore, reflect the regional distribution of σ-1 rather than σ-receptors.

Cytotoxic Effects

According to the literature, relatively high doses of σ-antagonists (20–100 μM for rimcazole or haloperidol (12,13)) are required for the killing of glioma cells. Our data confirm these reports (EC₅₀, 31 μM for rimcazole and 58 μM for haloperidol; Fig. 2). Even higher doses of (+)-pentazocine and AC915 were necessary to induce cell death (EC₅₀ values, 710 and 819 μM, respectively; Fig. 2). Such high doses are not required because diffusion barriers limit drug access to intracellular sites, for half-maximal occupancy of the σ-receptor population is reached already at concentrations 100-fold lower than the EC₅₀ for cell killing (Table 2). Because we observed a decrease in cellular viability only at drug concentrations 2 orders of magnitude higher than the corresponding IC₅₀ values for inhibition of ¹¹C-SA4503 binding, our data indicate that cell death is associated with virtually complete (>98%) occupancy of the σ-receptor population. Because ¹¹C-SA4503 binding occurred mainly at σ-2 receptors under the conditions of our assay, the data also suggest that cell death is induced via the σ-2 rather than via the σ-1 subtype if the observed cytotoxic effects are really mediated through σ-receptors.

The curves describing cell loss after 24 h of treatment with rimcazole and (+)-pentazocine have a steep slope, indicating rapid cell killing when the drug dose exceeds a sharp threshold concentration (Fig. 2). In contrast, cell death induced by haloperidol and AC915 occurs more gradually (Fig. 2). These different curve shapes may signify that haloperidol and AC915 are partial agonists at σ-2 receptors, whereas rimcazole and (+)-pentazocine are full agonists. It is also possible that rimcazole and (+)-pentazocine induce cell death via an additional mechanism, which occurs in parallel and independently of occupancy of the σ-2 receptor population. σ-2 receptor–mediated cell death is suggested by the fact that losses of cell viability always occur at concentrations 2 orders of magnitude higher than the IC₅₀ of the 4 test compounds for inhibition of ¹¹C-SA4503 binding to cellular σ-2 receptors (Table 2). An additional mechanism of cell death is suggested by the different shapes of the cell-loss curves (Fig. 2) and by a cellular morphology that is different after treatment of the cells with different σ-ligands (data not shown).

On the basis of these data, rimcazole appears to be the most interesting compound for further study as an anticancer drug.

Metabolic Changes

In our in vitro model, we observed a large increase of ¹⁸F-FDG uptake after 24 h of incubation with σ-receptor ligands (Fig. 3). A similar increase of ¹⁸F-FDG uptake has been observed after some forms of chemo- and radiotherapy in which the increase of uptake preceded cell death (29,30). This increase may be a “metabolic flare” due to increased energy consumption as a consequence of DNA and membrane repair (29). Glucose transporters (glucose transporter 1 and sodium-dependent glucose transporter 1) are known to be overexpressed after various cellular stresses, including hypoxia and heat shock (31). Rimcazole and (+)-pentazocine interfered strongly with cellular ¹⁸F-FDG uptake, whereas haloperidol and AC915 had smaller effects on glucose metabolism (Fig. 3).

¹⁸F-FLT traces the salvage pathway of DNA synthesis. The retention of this nucleoside is a measure of thymidine kinase 1 activity within the cells (32,33). After treatment of C6 cells with σ-receptor ligands, we observed striking decreases of ¹⁸F-FLT uptake. (+)-Pentazocine, rimcazole, and haloperidol induced more than a 70% decline of cellular ¹⁸F-FLT uptake at a dose 100 times greater than the IC₅₀ for inhibition of cellular ¹¹C-SA4503 binding (Fig. 4). However, AC915 caused a less marked inhibition of ¹⁸F-FLT uptake, which is consistent with a smaller reduction of cellular proliferation (Fig. 4). The rank order of drug effects on ¹⁸F-FLT uptake was the same as for ¹⁸F-FDG. In the range of 3–30 times IC₅₀ for inhibition of cellular ¹¹C-SA4503 binding, AC915 induced a small increase of ¹⁸F-FLT uptake, which may reflect activation of repair processes within the cells.

We tested ¹¹C-choline uptake after treatment of C6 cells with σ-receptor ligands because total choline levels and phosphatidylcholine synthesis are known to be elevated in cancers of different origin, as compared with healthy tissue, and are related to cellular proliferation. We observed the expected strong decrease of ¹¹C-choline uptake when cells were treated with σ-ligands (Fig. 5). Above the threshold rimcazole concentration, the uptake of ¹¹C-choline was reduced rapidly, whereas (+)-pentazocine, AC915, and haloperidol resulted in slower but progressive inhibition of phospholipid metabolism. These results are consistent with the fact that ¹¹C-choline uptake by cancer cells is related to the activity of the enzyme choline kinase (34), which is generally directly proportional to the proliferative status of the cell.

In contrast to ¹⁸F-FLT and ¹¹C-choline, ¹¹C-methionine did not behave as a proliferation marker, and ¹¹C-methionine uptake per viable cell showed hardly any change (Fig. 6). Thus, ¹¹C-methionine appeared to trace cell number rather than cellular viability after σ-ligand treatment. The methyl group of methionine, which contains the ¹¹C label, is implicated in many metabolic processes such as protein synthesis, transmethylation, and polyamine synthesis. Therefore, ¹¹C-methionine has a complex metabolism, and uptake of ¹¹C-methionine by tumor cells may be a reflection of several processes, which are affected differently by anticancer treatment, some being increased and others showing a decrease (35). The lack of change in ¹¹C-methionine uptake may also indicate unaltered transport of this amino acid through the cell membrane and slow transmethylation processes (31).

Our data indicate that ¹⁸F-FLT and ¹¹C-choline are the tracers of choice for evaluation of the effects of σ-ligands on cellular proliferation. ¹⁸F-FDG may also provide useful information, but ¹¹C-methionine uptake is a reflection of the cell number rather than the proliferation state.

Growth reduction of glioma cells and changes of cellular metabolism were observed only at ligand concentrations more than 30-fold higher than those required for 50% occupancy of the σ-2 receptor population. Strong decreases of the number of viable cells occurred at drug concentrations greater than or equal to 100-fold higher than the IC₅₀ for inhibition of ¹¹C-SA4503 binding. Thus, cell death was associated with virtually complete (99%) occupancy of the intracellular σ-2 receptor pool.

Although high levels of receptor occupancy and drug doses in the 10⁻⁵ M range are required for rapid cell killing, σ-ligands appear to have potential as cancer therapeutics for the following reasons. First, preliminary studies, both from our own laboratory and from other institutions (18), have demonstrated that cell death is limited to tumor cells and occurs only at much higher concentrations in healthy normal cells. Adverse effects could be further diminished by local application or selective drug targeting. Second, significant inhibition of in vivo tumor growth has been observed in a variety of tumor models in response to σ-ligands given systemically (17,18). Third, because σ-2 ligands trigger cell death both by caspase-dependent and caspase-independent mechanisms, such compounds may be suitable for treatment of a wide variety of tumors, including caspase-deficient tumor types (36). Last, subtoxic doses of σ-ligands can be applied in combination therapy and can potentiate the antitumor effects of “classic” cytostatic agents (36,37). Because of the therapeutic potential of σ-ligands, radiopharmaceuticals for σ-receptors may be used not only as diagnostic imaging agents but also as tools to assess σ-receptor occupancy during experimental tumor therapy.