Sigma-2 Receptor Agonists Activate a Novel Apoptotic Pathway and Potentiate Antineoplastic Drugs in Breast Tumor Cell Lines¹

Sigma receptors are unique drug-binding proteins that are present in the central nervous system as well as in various peripheral tissues (1). They recognize a variety of psychoactive agents from various structural classes, including opioids, such as pentazocine, and exhibit high affinity binding of neuroleptic drugs, such as haloperidol. There is strong evidence for endogenous sigma receptor ligands, and although progesterone and other neurosteroids have been suggested as candidates (2), none has yet been identified conclusively. The functions of these receptors are not yet clearly defined. In the central nervous system, they have been shown to be involved in regulation of neurotransmitter release, modulation of neurotransmitter receptor function, learning and memory processes, and regulation of movement and posture (3). Although present in peripheral tissues such as liver, kidney, and endocrine organs, functions in these tissues is much less understood.

Two subtypes of sigma receptor have been identified, termed sigma-1 and sigma-2 (4, 5, 6). The subtypes are distinguishable pharmacologically, functionally, and by molecular size. Sigma-1 receptors have been cloned and shown to be distinct from any known receptor class (7). The sigma-1 receptor is a M_r 25,000, single polypeptide with one putative trans-membrane region. The sigma-2 receptor is a M_r 18,000–21,000 protein but has not yet been cloned (4, 6).

We have shown that both sigma receptor subtypes are highly expressed in tumor cell lines from various tissues (8). Interestingly, sigma receptors are more highly expressed in rapidly proliferating cells and are down-regulated when cells become quiescent (9). In the human breast, sigma receptors were virtually absent in normal tissue but were present in high density in breast tumor biopsy tissue (10). Their high density in various tumor cell types, and particularly in proliferating cells, makes sigma receptors potential targets for diagnostic imaging as well as therapeutic agents (9, 11, 12).

We have demonstrated previously that certain sigma receptor ligands cause morphological changes in human SK-N-SH neuroblastoma cells, rat C6 glioma cells, and in several other neuronal and nonneuronal cell lines that contain sigma-1 and sigma-2 receptors (13). These morphology changes include loss of processes, rounding, and detachment from the substratum. Continued exposure of cells to sigma receptor ligands results in cell death. Subsequent studies revealed that the mode of cell death induced by sigma ligands in human SK-N-SH neuroblastoma cells is apoptotic, and that the effect is mediated with a pharmacological profile consistent with specific activation of sigma-2 receptors (14, 15). For example, the sigma-2 subtype-selective ligands, ibogaine and CB-64D, induced apoptosis. However, the sigma-1-selective ligands, (+)-pentazocine and dextrallorphan, or ligands for other receptors such as dopamine and opioid receptors, had little or no effect (13, 14, 15). Furthermore, some sigma ligands were found to inhibit proliferation of mammary and colon carcinoma cell lines and to induce apoptosis in colon and mammary adenocarcinoma cells (16, 17). The induction of morphology changes and apoptosis may be linked to sigma-2 receptor-mediated rises in intracellular calcium levels, because these effects exhibit the same pharmacological profile (17, 18, 19).

Mutations in the tumor suppressor gene, p53, are the most frequently observed genetic aberrations in tumors, occurring in up to 50% of some tumor types. In tumor cells with p53 mutations, a diminished response to agents that induce apoptosis has been observed (20, 21), and these tumors may be clinically resistant to antineoplastic drugs that produce DNA damage (22).

Here we demonstrate that sigma-2 receptor agonists induce cell death in various breast tumor cell lines with features consistent with apoptosis. We also show here that unlike DNA-damaging agents, sigma-2 receptor agonists exhibit similar potency in tumors with wild-type or mutant p53. The mechanism of sigma-2 receptor-mediated apoptosis differs from that of agents that trigger DNA damage, based on observations with inhibitors of caspases. The involvement of distinct apoptotic pathways is further supported by the ability of sigma agonists to potentiate the cytotoxicity of DNA-damaging, antineoplastics in various tumor cell lines.

Human breast tumor cell lines (MCF-7, T47D, SkBr3, and MCF-7/Adr−) were cultured in DMEM containing 3.7 g/liter NaHCO₃, 10% fetal bovine serum, and insulin (10 mg/liter). For cytotoxicity assays, cells were transferred to DMEM + Ham’s nutrient mixture F-12 (without phenol red) with 1.2 g/liter Na₂HCO₃. Cells were seeded at 10⁵ cells/well (1 ml/well). All cell culture processes were carried out in a humidified atmosphere of 5% CO₂/95% air at 37°C.

Cell death was assessed by release of LDH³ into the culture medium using the CytoTox 96 kit from Promega Corp. (Madison, WI). The method was performed as specified by the manufacturer, with minor modifications. After treatment of cells with test compound for 24 h, the medium from the wells was transferred to microcentrifuge tubes and centrifuged (10,000 rpm for 5 min) to remove floating cells and cell debris. After sedimentation, 50 μl of supernatant from samples were transferred to 96-well, flat-bottomed plates (Costar, Cambridge, MA), to which was then added 50 μl of substrate mix in assay buffer. Plates were kept protected from light for 30 min at room temperature, at which time 50 μl of stop solution were added to each well. Any air bubbles present were removed using a syringe, and production of formazan was monitored at 490 nm in a plate reader (Ceres UV 900; Bio-Tec Instruments, Winoorsi, VT). The percentage of cytotoxicity produced by drugs was calculated relative to absorbance values for no-drug controls and values resulting from total lysis of cells by Triton X-100 (100% cell kill) according to the following formula:

Cytotoxicity in drug-treated culture wells was expressed as a percentage of total cell kill in untreated cells, uniformly for all drug concentrations. Cytotoxicity ED₅₀s were determined from dose-response curves analyzed using GraphPad Prism (San Diego, CA).

Apoptosis is characterized by fragmentation of nuclear DNA by endonucleases activated during the process. DNA fragmentation occurring during apoptosis can be detected by incorporating fluorescein-12-dUTP at the 3′-OH DNA ends using the enzyme terminal deoxynucleotidyl transferase (23). TUNEL was performed using the Apoptosis Detection System, Fluorescein kit (Promega Corp., Madison, WI) as described in the manufacturer’s Technical Bulletin. After treatment of cells with sigma ligands at the concentration and times specified, cells were labeled. Attached cells were labeled in the chambers. Detached (floating) cells were carefully harvested by centrifugation and reattached to gelatin-covered glass slides before labeling. In brief, cells were fixed in 4% formaldehyde in PBS for 25–30 min at 4°C. After washing with PBS at room temperature, cells were permeabilized with 0.2% Triton X-100 solution for 5 min on ice. Cells were then washed with PBS. To each chamber or glass slide was added 50–100 μl of equilibration buffer. Glass slides were covered with plastic coverslips, and chambers and slides were equilibrated with equilibration buffer for 10 min at room temperature. After removing plastic coverslips and excess liquid, 50 μl of incubation buffer (45 μl equilibration buffer, 5 μl nucleotide mix, and 1 μl terminal deoxynucleotidyl transferase enzyme) were added to each sample. They were covered with plastic coverslips or with chambered coverslip caps and placed in an incubator under a humidified atmosphere at 37°C for 60 min. Slides were then dipped in stop solution, or an equal volume of 2× stop solution was added to chambered coverglasses, and samples were incubated 15–20 min at room temperature. After washing samples with PBS at room temperature, cells were stained with propidium iodide (1 μg/ml) for 15 min at room temperature in the dark. The preparations were then washed in distilled water. After adding 1 drop of Anti-Fade Solution (Molecular Probes, Eugene, OR), slides were sealed with glass coverslips and clear nail polish. Cells on chambered coverglasses were occasionally covered with a glass coverslip for temporary preservation.

Samples were analyzed using a Nikon Diaphot inverted fluorescence microscope and a dual filter set for FITC/rhodamine (excitation, 450–490 nm/barrier, 520 nm) at ×40. Propidium iodide stains the DNA in both apoptotic and nonapoptotic cells with red to orange color. Apoptosis was indicated by the presence of green or yellow-green fluorescence within the nucleus of cells as confirmation of fluorescein-12-dUTP incorporation at 3′-OH ends of fragmented DNA.

The early stages of apoptosis are characterized by translocation of PS from the inner surface of the plasma membrane to the outer surface of the membrane (24). Externalized PS can then be detected using Annexin V, a protein with high affinity for PS. This was carried out using the ApoAlert Annexin V Apoptosis kit (Clontech, Palo Alto, CA) according to the manufacturer’s specifications.

All experiments were performed on attached cells in four- or eight-well chambered coverglasses (Nalge Nunc International, Naperville, IL). After incubation with test compounds at concentrations and times suspected to produce early stages of apoptosis, cells were stained as described in the ApoAlert Annexin V Technical Bulletin, with some modifications. In brief, cells were rinsed in binding buffer. They were then covered with 100–200 μl of binding buffer to which was added 5 μl of Annexin V (20 μg/ml in Tris-NaCl buffer) and 10 μl propidium iodide (1 μg/ml). Samples were incubated for 10–15 min at room temperature in the dark and then analyzed.

Samples were analyzed using a Nikon Diaphot inverted fluorescence microscope with a dual filter set for FITC/rhodamine (excitation, 450–490 nm/barrier, 520 nm). Apoptosis was confirmed by green staining in the plasma membrane, indicating Annexin V binding to PS. Red or orange staining is indication of loss of membrane integrity that occurs during later stages of the process and results from failure to exclude propidium iodide and staining of DNA. Red or orange staining can occur in both late-stage necrotic and apoptotic cells. However, green staining will only occur in early-stage apoptotic cells.

Cells were cultured to 90% confluency as described above in 175-cm² flasks (Costar). Membranes were prepared essentially as described previously (8). Medium was decanted, and cells were rinsed in cold PBS, detached, and pelleted. The pellet was resuspended in ice-cold 10 mm Tris-HCl (pH 7.4), containing 0.32 m sucrose (0.5 ml/flask), and homogenized with seven to nine strokes in a Potter-Elvehjem homogenizer (Teflon pestle). The homogenate was centrifuged at 31,000 × g for 15 min at 4°C, and the supernatant was discarded. The final pellet was resuspended in ice-cold 10 mm Tris-HCl (pH 7.4) to a protein concentration of 15–20 mg/ml, and the crude membrane preparation was stored at −80°C until use. Protein was determined using the BCA assay (Pierce).

Sigma-2 receptors were labeled using [³H]DTG (28.1 Ci/mmol) in the presence of 1 μm dextrallorphan to mask the sigma-1 sites (6, 8). Nonspecific binding was determined in the presence of 10 μm haloperidol. Various radioligand concentrations were prepared and incubated in 50 mm Tris-HCl (pH 8.0), with 200 μg of membrane protein in a total volume of 500 μl for 120 min at 37°C. Binding assays were terminated by the addition of 5 ml of ice-cold 10 mm Tris-HCl (pH 7.4) and filtration through polyethyleneimine (0.5%)-soaked glass fiber filters using a Brandel cell harvesting apparatus (Gaithersburg, MD). Filters were washed twice with 5 ml of ice-cold buffer. Radioactivity on filter-trapped membranes was quantified by scintillation counting using Cytoscint (ICN, Costa Mesa, CA).

Doxorubicin and caspase inhibitors (ZVAD-FMK, YVAD-CHO, and DEVD-CHO) were obtained from Calbiochem (San Diego, CA). Actinomycin D was purchased from Sigma Chemical Co. (St. Louis, MO). Haloperidol and reduced haloperidol are subtype-nonselective sigma ligands (5, 6, 25) and were obtained from Research Biochemicals, Inc. (Natick, MA). The 5-phenylmorphans CB-64D and CB-184 are sigma-2 subtype-selective agonists (19, 26), synthesized in the Laboratory of Medicinal Chemistry/National Institute of Diabetes and Digestive and Kidney Diseases (Dr. C. Bertha).

The cell lines examined in this study are presented in Table 1, along with the p53 genotype and the density of sigma-2 receptors. All of the cell lines express high levels of sigma-2 receptors. Although MCF-7 cells express wild-type p53 protein, MCF-7/Adr−, SKBr3, and T47D cells all have mutations in p53 and do not express active protein (27, 28, 29, 30, 31).

Various lines of cultured cells were incubated with sigma receptor ligands or antineoplastic agents, and the cells were assayed by the TUNEL method to assess drug-induced effects and potential mechanisms. Positive TUNEL staining would be indicative of the DNA fragmentation that is characteristic of apoptosis (23). Results for T47D and MCF-7 cells are shown in Fig. 1.

Apoptosis was induced by doxorubicin (Fig. 1, C and D), as well as by other antineoplastic drugs (actinomycin D and cyclophosphamide) in T47D and SKBr3 cell lines (not shown). CB-64D and its 3,4-dichloro derivative, CB-184, are novel 5-phenylmorphans with high affinity and 185- and 554-fold selectivity for sigma-2 receptors, respectively, over sigma-1 receptors (26). Both compounds are sigma-2 receptor agonists (19). CB-64D at a concentration of 100 μm produced extensive apoptotic changes in T47D and MCF-7 cells by 48 h (Fig. 1, E and F). In some experiments, particularly in MCF-7 cells, up to 100% of visualized cells displayed apoptotic nuclei. Haloperidol and reduced haloperidol bind with high affinity to both sigma-1 and sigma-2 sites and are agonists at sigma-2 receptors (19, 25). Haloperidol (Fig. 1, G and H) and reduced haloperidol (Fig. 1, I and J) also produced apoptotic nuclei, but less extensively than the more potent CB-64D.

Thus, antineoplastic agents and representative sigma receptor ligands including haloperidol, reduced haloperidol (25), and CB-64D (26), induced apoptosis in MCF-7 and T47D breast tumor cell lines. Although T47D cells express both sigma-1 and sigma-2 receptors, MCF-7 cells do not express active sigma-1 receptors (8). Taken together with the sigma-2 subtype selectivity of CB-64D, the results support the notion that the apoptotic effect of the sigma ligands is mediated by specific interaction with sigma-2 receptors.

The presence of p53 mutations renders some of the cell lines shown in Table 1 resistant to certain antineoplastic agents (27, 28, 29, 30, 31). To quantify cell death induced by doxorubicin, MCF-7 and MCF-7/Adr− cells were incubated for 48 h in the presence of various concentrations of doxorubicin and cytotoxicity measured by the release of LDH into the culture medium. The results are shown in Fig. 2. The MCF-7/Adr− cell line with mutant p53 displayed markedly diminished sensitivity to doxorubicin compared with MCF-7 cells, which express wild-type p53.

To examine the effect of sigma-2 receptor agonists, cells were incubated in the presence of various concentrations of CB-64D or CB-184, at different time intervals, and cytotoxicity was quantified by LDH release. Fig. 3 shows a concentration-dependent, cytotoxic effect of the compounds in two cell lines, SKBr3 and T47D, at 24 h.

The potencies (EC₅₀s) of CB-64D and CB-184 at inducing cell death in the four cell lines shown in Table 1 are compared in Table 2 for a 48-h exposure to drug. Table 2 shows that CB-184 exhibited greater potency than CB-64D in all four cell lines. The potency of CB-184 was similar across the cell lines, regardless of p53 status. There was somewhat more variation across cell lines with CB-64D. However, compared with MCF-7 cells with wild-type p53 protein, the potency was either greater (MCF-7/Adr−), the same (SKBr3), or only 2-fold lower (T47D) in cells with p53 mutations. Thus, the presence of p53 mutations did not significantly decrease the cytotoxic potency of these sigma-2 receptor agonists.

It should be noted that the percentage of cytotoxicity was calculated relative to the LDH released from total lysis (by Triton X-100) of cells in the untreated control, as described by the equation in “Materials and Methods.” It is assumed here that the drug-treated wells and the control wells contain the same total number of cells (dead plus alive) at the end of the treatment period. However, if CB-64D and CB-184 have an inhibitory effect on cell proliferation superimposed on the cell killing effect, this will lead to an overall decrease in the total number of cells in the treated condition relative to control, resulting in an underestimation of the true cytotoxicity. This would be especially problematic at longer treatment times.

To address this issue for the various cell lines, Triton X-100 lysis solution was added to CB-184-treated wells at the end of the 48-h treatment period to assess total cells in the treated condition (absorbance_treated, lysed). This was compared with the absorbance produced in untreated wells after lysis (absorbance untreated, lysed; equivalent to “absorbance _{total cells}” of the equation in “Materials and Methods”). Using the equation (absorbance _treated, lysed/absorbance_untreated, lysed) × 100, it was found that wells treated with 30 and 100 μm CB-184 generally had 10–15% fewer cells and 30–35% fewer cells, respectively, compared with the untreated, lysed condition. This reflects the growth-inhibitory effect of the sigma ligand at these higher doses. At the lower doses, the total number of cells was found to be comparable in treated and untreated wells, although more of the cells were dead in the wells containing CB-184.

The adjusted cytotoxic potency against the number of cells found in drug-containing wells can be determined by the equation: (absorbance_treated/absorbance_treated, lysed) × 100. When this is done, the apparent percentage of cytotoxicity is greater than that obtained when absorbance_untreated, lysed is used in the denominator. Thus, when the growth-inhibitory effect of the sigma-2 agonists is taken into account, EC₅₀s shown in Table 2 may be reduced by 31–48%.

Caspases are a family of cysteine-aspartyl proteases that are the executioners of apoptotic signals from diverse stimuli, including receptor activation (e.g., Fas ligand and TNF-α), DNA-damaging agents, hypoxia, growth factor deprivation, or ionizing radiation (32). The targets of caspases include a vast array of key proteins that are necessary for cell survival. Some include cytoskeletal proteins, cell cycle regulatory proteins, and nuclear matrix proteins, such that the proteolytic cleavage of these targets is consistent with the morphological and biochemical alterations characteristic of apoptosis. Both selective and nonselective inhibitors of caspases have been developed as biochemical tools to help dissect the pathways by which an apoptotic signal is transmitted (33).

An early step in the induction of apoptosis is the inversion of PS, which can be detected by the binding of Annexin V to the cell surface (24). The ability of different caspase inhibitors to abrogate apoptosis induced by sigma-2 receptor ligands and some antineoplastic drugs was compared after 48 h of exposure using Annexin V binding. Results are shown in Fig. 4. Both CB-64D (100 μm; Fig. 4,C) and doxorubicin (100 μm; Fig. 4,D) induced apoptosis in MCF-7 cells as indicated by increased Annexin V binding (green staining) relative to untreated control cells (Fig. 4,A). Neither ZVAD-FMK (50 μm), an inhibitor of all known caspases (Fig. 4,B), nor YVAD-CHO (50 μm), a potent inhibitor of caspase-1 (not shown), had any effect when incubated alone with MCF-7 cells. When either ZVAD-FMK or YVAD-CHO were combined with doxorubicin, the appearance of apoptotic cells was blocked (Fig. 4, F and H). The persistence of orange staining in the presence of these caspase inhibitors indicated that some cell death still occurred, but not by an apoptotic mechanism. By contrast, ZVAD-FMK and YVAD-CHO failed to inhibit apoptosis induced by the sigma-2 agonist, CB-64D (Fig. 4, E and G).

We next evaluated the effect of different agents on cytotoxicity (LDH release) and their sensitivity to caspase inhibitors. Results are shown in Fig. 5. Doxorubicin induced a dose-dependent increase in LDH release in MCF-7 cells during 48 h, with ZVAD-FMK (50 μm) abolishing the effect of lower doses and attenuating the effect of higher doses (Fig. 5A). Actinomycin D has been shown to induce apoptosis by either inhibition of nucleic acid synthesis or by intercalation with DNA base pairs, leading to the induction of p53 (34). MCF-7 cells were treated with various concentrations of either actinomycin D for 48 h (Fig. 5,B) or the sigma-2 agonist CB-184 for 72 h (Fig. 5,C) in the absence or presence of DEVD-CHO (50 μm), a potent inhibitor of caspase-3 (33). DEVD-CHO abrogated actinomycin D-induced cytotoxicity, whereas in cells treated with CB-184, DEVD-CHO had no statistically significant effect on cytotoxicity. These observations in the LDH release assay are consistent with the lack of effect of caspase inhibitors on the activity of CB-64D in the Annexin V binding experiments described above (Fig. 4). The differential effect of caspase inhibitors on apoptotic cell death induced by DNA-damaging agents and sigma-2 receptor agonists suggests induction of apoptosis by caspase-dependent and caspase-independent mechanisms, respectively.

The interaction of sigma-2 receptor agonists with DNA-damaging antineoplastic agents was investigated. The results are shown in Figs. 6 and 7. The combination of the sigma-2 agonist with antineoplastic agents, at concentrations of the individual agents producing modest to no effects on cell killing, resulted in substantial potentiation of cytotoxicity. Fig. 6 shows that a clear synergistic effect was observed in MCF-7 cells at 24 h when CB-184 (1 μm) was combined with 10 μm doxorubicin (Fig. 6,A). Synergy was also apparent at 48 h, but less pronounced because of significant toxicity of each agent alone at this time point (Fig. 6,B). Fig. 7 shows that in MCF-7/Adr− cells, the cytotoxicity of 1.0 μg/ml actinomycin D at 24 h (Fig. 7,A) or 0.1 μg/ml actinomycin D at 48 h (Fig. 7 B) was potentiated by CB-184 (1 μm).

Several agents that possess sigma-2 receptor agonist activity are currently in clinical use. To determine whether clinically available sigma agonists displayed a similar effect as CB-184, the antipsychotic haloperidol and the analgesic (±)-pentazocine (Talwin) were combined with various concentrations of doxorubicin. The results are shown in Fig. 8. In MCF-7/Adr− cells, haloperidol (25 μm) markedly potentiated the cytotoxicity of doxorubicin. (±)-Pentazocine (35 μm) potentiated at only the highest dose of doxorubicin. Little or no potentiation was observed by these agents when combined with doxorubicin in MCF-7 cells (data not shown).

We show here that various sigma receptor ligands, particularly selective sigma-2 receptor agonists, induce apoptosis in breast tumor cell lines as indicated by DNA fragmentation (TUNEL staining) and inversion of phosphatidyl serine (Annexin V binding). Sigma-1 receptors do not appear to contribute to the apoptotic effect, because sigma-2 subtype-selective ligands with low to negligible affinity for sigma-1 sites (CB-64D and CB-184) induced apoptosis. Furthermore, sigma-2 receptor agonists induced apoptosis in MCF-7 cells, which do not express active sigma-1 receptors (8).

In some experiments, we have used LDH release as a method to quantify cell death. However, this method does not distinguish apoptotic cell death from necrotic cell death. At the appropriate sigma ligand concentration and duration, up to 100% cell killing as measured by LDH release can be observed with selective sigma-2 receptor agonists. Under the same conditions, we can observe 100% of cells in the visualized field undergoing apoptosis when analyzed by both the TUNEL assay and Annexin V binding, although the optimal treatment duration may vary between the two assays. These findings suggest that the LDH release we observe in sigma-2 receptor agonist-treated cells is attributable to apoptotic and not necrotic cell death.

Most apoptotic signals initiate a cascade of sequential caspase activation, causing degradation of specific proteins and resulting in cell death (32). In assays where cell death was quantified by LDH release (Fig. 5) or apoptosis was monitored by Annexin V binding (Fig. 4), caspase inhibitors decreased the cytotoxicity of the DNA-damaging agents doxorubicin or actinomycin D, as has been observed by others (35, 36). However, caspase inhibitors failed to block the cytotoxicity of sigma-2 receptor agonists in the tumor cell lines we have studied. For doxorubicin, caspase inhibitors abolished Annexin V binding (Fig. 4, F and H) and shifted the dose curve for LDH release to the right (Fig. 5,A) but did not totally block cell death. This suggests that doxorubicin kills MCF-7 cells by other mechanisms, in addition to apoptosis. The cytotoxic effect of actinomycin D in MCF-7 cells was completely abolished by DEVD-CHO (Fig. 5,B). In contrast, there was no change in the phenotype of cell death (Fig. 4, C, E, and G) or shift in the potency for LDH release by caspase inhibitors (Fig. 5 C) in MCF-7 cells treated with sigma-2 receptor agonists. We have similarly observed that in human SK-N-SH neuroblastoma cells, characteristic apoptotic morphological changes induced by sigma-2 receptor agonists (14, 18) were unaffected by ZVAD-FMK treatment.⁴

It should be pointed out that DEVD-CHO is generally considered to be a selective inhibitor of caspase-3, because it exhibits its highest potency against this caspase (33). However, in addition to caspase-3, DEVD-CHO is also a relatively potent inhibitor of caspase-1, caspase-7, caspase-6, caspase-8, caspase-9, and caspase-10 (33). This accounts for the observed inhibitory effect of DEVD-CHO against actinomycin D in MCF-7 cells, which are known to lack active caspase-3 because of aberrant mRNA splicing (37, 38). This observation suggests that actinomycin D-induced apoptosis in MCF-7 cells involves activation of caspases other than caspase-3 and is also consistent with the ability of DEVD-based inhibitors to block apoptosis in MCF-7 cells induced by other stimuli (39). Doxorubicin also uses caspases other than caspase-3 in MCF-7 cells because ZVAD-FMK, a potent inhibitor of caspase-1 through caspase-10 (33), blocked apoptosis induced by this agent. The inability of YVAD-FMK, ZVAD-FMK, and DEVD-CHO to inhibit apoptosis induced by CB-64D or CB-184 indicates that sigma-2 receptor-mediated apoptosis does not involve caspases commonly activated by many other apoptotic stimuli and may be caspase independent.

Mutations in p53 often confer resistance to DNA-damaging agents that induce apoptosis (27, 28, 29, 30, 31), although some studies show increased sensitivity in p53 mutants (27). The p53 mutant cell lines that we have examined displayed chemoresistance to certain agents (see Fig. 2) consistent with previous reports [MCF-7/Adr− to doxorubicin (28) and TNF-α (29); SKBr3 to several agents (27)]. However, when the cytotoxicity of selective sigma-2 agonists was examined across various breast tumor cell lines, their potency was not generally affected by the status (wild-type or mutant) of p53 (Table 2). These results are consistent with sigma-2 receptor-mediated apoptosis via a mechanism that is independent of p53.

Whether apoptosis can proceed in the absence of caspase activation and p53 involvement is an important question. Alternative forms of cell death that differ in certain morphological and biochemical features from apoptosis and necrosis have been described. For example, paraptosis lacks nuclear fragmentation, apoptotic body formation, and chromatin condensation and instead, presents with cytoplasmic vacuolation and mitochondrial swelling (40). Paraptosis is insensitive to the caspase inhibitors ZVAD-FMK, BAF, p35, and X-chromosome-linked inhibitor of apoptosis but does involve an alternative form of caspase-9. Because nuclear fragmentation does not occur, paraptosis is TUNEL negative (40). As shown here, sigma-2 receptor-mediated cell death is characterized by TUNEL-positive staining. Furthermore, we have shown previously that sigma-2 receptor agonists induce chromatin condensation in SK-N-SH neuroblastoma cells as assessed by bisbenzimide (Hoechst 33258) DNA staining (14). Thus, sigma-2 receptor-mediated cell death appears to be distinct from paraptosis.

A number of investigators report that inhibition of caspases fails to abolish morphological/biochemical features associated with apoptosis induced by various agents. These include ceramide (41), vitamin D (which modulates ceramide levels; Ref. 42), ganglioside G_D3 (43), calcineurin (44, 45), PML matrix-associated nuclear bodies (46), apoptosis-inducing factor (47), Bax (48), and the tumor suppressor Bin1, which also acts independently of p53 (49). Mitochondrial changes characteristic of apoptosis occur in yeasts, which lack caspases (50). The current results indicate that sigma-2 receptors may use a novel pathway to apoptosis. Investigation of the mechanisms surrounding sigma-2 receptor-mediated apoptosis could shed light on alternative forms of programmed cell death.

The observation that sigma-2 agonists are nearly equipotent in killing cells with mutant and wild-type p53 genes and can potentiate antineoplastic drug effects in breast tumor cells has tremendous implications for clinical practice. Although many tissues normally express high densities of the receptor (6, 51), we show here that subtoxic doses of potent sigma-2 agonists potentiate the cytotoxicity of antineoplastic drugs, which may already possess some limited selectivity for tumors. This phenomenon may also result in reversal of drug resistance in tumors at concentrations of the antineoplastic drugs that reduce the very severe adverse effects. For example, sigma-2 agonists could enhance the effect of doxorubicin in the high-dose CAF regimen [cyclophosphamide, Adriamycin (doxorubicin), 5-fluorouracil], used against aggressive breast tumors that overexpress erbB-2 and have p53 mutations, such as SKBr3 (52, 53).

We also show here that two clinically available drugs that have sigma-2 agonist activity, haloperidol and (±)-pentazocine (Talwin), display the potentiation phenomenon. Because (−)-pentazocine has higher sigma-2 receptor affinity than (+)-pentazocine (4, 5, 6), it is likely that the (−)-enantiomer of racemic pentazocine is responsible for the antineoplastic potentiating activity. Furthermore, the lower potency of racemic pentazocine compared with haloperidol is consistent with action at sigma-2 receptors, because (−)-pentazocine has lower affinity at sigma-2 sites than haloperidol (6, 8). The butyrophenones, droperidol and haloperidol, have efficacy against antineoplastic-induced emesis of moderate severity, although haloperidol is seldom used for this indication. Racemic pentazocine (Talwin) can be used in the management of moderate pain caused by metastatic tumors. Therefore, these agents that can potentiate antineoplastic activity have other actions useful in cancer patients.

MCF-7/Adr− cells with mutant p53 also overexpress the MDR gene product, P-glycoprotein (28). P-glycoprotein enhances the efflux of hydrophobic compounds that are often toxic to cells. Doxorubicin is a substrate for P-glycoprotein, and overexpression results in phenotypic resistance to doxorubicin, even in the presence of wild-type p53 (54). The sigma-2 receptor agonists CB-64D and BD737 have been shown to reduce the expression of the MDR gene in human SK-N-SH neuroblastoma and rat C6 glioma cells (55). However, this is not likely to be the predominant mechanism for the potentiation we have observed, because it occurs in “wild-type” MCF-7 cells that are sensitive to doxorubicin and which do not overexpress MDR.

In summary, sigma-2 receptor agonists induced apoptosis in various breast tumor cell lines in a manner apparently independent of both p53 and caspase activation. Also, sigma-2 agonists at doses that were not cytotoxic potentiated the action of DNA-damaging agents. This suggests that sigma-2 receptors use an apoptotic pathway distinct from those used by DNA-damaging agents and other apoptotic stimuli. The observed synergism between DNA-damaging antineoplastic agents and sigma-2 receptor agonists could result from simultaneous activation of these distinct apoptotic programs. Sigma-2 receptors may represent novel targets for the development of antineoplastic agents.

We thank Dr. Mikhail V. Blagosklonny (Medicine Branch, NCI, NIH) for SKBr3 cells. We thank Glinda Kohlhagen and Dr. Yves Pommier (Division of Basic Science, National Cancer Institute, NIH) for MCF-7/Adr− cells.