Perillyl Alcohol Inhibits a Calcium-Dependent Constitutive Nuclear Factor-κB Pathway

The monoterpenes isolated from the essential oils of plants elicit multiple chemotherapeutic and chemopreventive effects in diverse models of cancer (1). Anticancer activities associated with regressing mammary tumors include the induction of cytostasis and apoptosis upon monoterpene application. This antitumor activity correlates with the differential expression of growth and apoptotic genes necessary for tumor proliferation (2). The sequence-specific DNA binding of transcription factors often dictates the expression of genes (3). However, the transcription factors that regulate the anticancer apoptotic response to perillyl alcohol treatment are not well understood.

The monoterpene menthol exhibits calcium-related molecular actions not associated with anticancer activity. Studies in situ have shown that menthol is a direct inhibitor of L-type calcium channels (LTCC) expressed on the plasma membrane of normal tissues (4). Interestingly, these studies mechanistically explain the mild to moderate gastroesophageal reflux or “heartburn” observed in clinical trials evaluating similarly structured monoterpenes for the treatment of different diseases, including cancer (5, 6). LTCC antagonists, such as the dihydropyridines, also cause the same disorder by inhibiting the calcium channel activity in smooth muscle, resulting in the relaxation of the esophageal stomach sphincter (7). LTCCs are specific receptors for the dihydropyridines and are also called dihydropyridine-sensitive calcium channels (8).

Lymphoma cells originate from cell types (9) that express dihydropyridine-sensitive calcium channels (10, 11). WEHI-231 B-lymphoma cells constitutively express calcium-dependent nuclear factor-κB (NF-κB) DNA-binding activity responsible for the transcription of antiapoptotic genes permissive for cell survival (12, 13). LTCCs regulate calcium-dependent cell survival pathways independent of NF-κB transcription in neuronal cells (14). However, an association between the regulation of dihydropyridine-sensitive calcium channels and constitutive NF-κB–mediated antiapoptotic pathways in these B-lymphoma cells are not well established. Therefore, we asked whether the NF-κB antiapoptotic pathway in the WEHI-231 cells responded to monoterpene treatment. Our findings suggest how a rapidly targeted calcium-dependent decrease of NF-κB in B-lymphoma cells due to monoterpene treatment can induce apoptosis. Consequently, we investigated whether perillyl alcohol treatment reduced constitutive NF-κB activity through a similar mechanism in certain estrogen receptor–negative/independent (ER−) breast cancer cell lines.

Chemicals and materials. Propidium iodide (PI), 2′-(4-ethoxyphenyl)-5-(4-methyl-1-piperazinyl)-2,5′-bi-1H-benzimidazole trihydrochloride (Hoechst 33342), S-(−)-1,4-dihydro-2,6-dimethyl-5-nitro-4-[2-(trifluoromethyl)phenyl]-3-pyridine carboxylic acid methyl ester [BayK8644(−)], and (R)-(+)-1,4-dihydro-2,6-dimethyl-5-nitro-4-[2-(trifluoromethyl)phenyl]-3-pyridine carboxylic acid methyl ester [BayK8644(+)] were purchased from Sigma-Aldrich (St. Louis, MO). Aldrich Chemical Company (Milwaukee, WI) supplied the monoterpenes except for perillic acid, which was synthesized at the University of Wisconsin-Madison. All other chemicals were acquired from Calbiochem (San Diego, CA), including 1-[2-amino-5-(6-carboxyindol-2-yl)phenoxy]-2-(2′-amino-5′-methylphenoxy)ethane-N,N,N′,N′-tretraacetic acid pentaacetoxymethyl ester (INDO-1/AM) and 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetra(acetoxymethyl) ester (BAPTA-AM). The CD40 monoclonal antibody was purchased from BD Biosciences PharMingen (San Diego, CA).

Statistical analysis. Sigma Stat 3.0 software (SYSTAT Software, Inc., Richmond, CA) was used for the statistical analysis.

Cell culture. The American Type Culture Collection (Manassas, VA) supplied the cell lines used in our studies. The WEHI Bcl-X_L cells have been described (15). All cell lines, including the WEHI-231, Ramos, MCF-7, T47D, MDA-MB 231, and MDA-MB 468 cells, were cultured in RPMI medium with 10% FCS (Sigma), 2 mmol/L l-glutamine, and penicillin/streptomycin. The WEHI-231 and WEHI Bcl-X_L cells were also grown with 50 μmol/L β-mercaptoethanol. Lymphocyte density medium was purchased from Mediatech (Herndon, VA). All cell culture materials unless specified were from Invitrogen (Carlsbad, CA). Solubilization of the solid monoterpenes perillic acid and menthol used ethanol as the vehicle. The final concentration of ethanol in the media for the solid monoterpenes was ≤0.07%. The dihydropyridines were suspended in 100% ethanol with a working stock concentration of 25 or 30 mmol/L. The concentration of ethanol in the media with 50 μmol/L of dihydropyridine was 0.2%.

Cell culture experiments. WEHI-231, WEHI Bcl-X_L, and Ramos cells were seeded at 1 × 10⁶ and 3 × 10⁵ to 5 × 10⁵ cells/mL for the 4- and 24-hour assays, respectively. The MDA-MB 468 cells were seeded at 2 × 10⁵ cells/well in a six-well plate (BD Bioscience, Bedford, MA) and allowed to attach overnight before treatment.

Apoptosis assays. The PI/viability assay with the WEHI-231, WEHI Bcl-X_L, and Ramos cells used the addition of 30 μg/mL of PI to the cells 5 minutes before analysis by FACScan flow cytometry (Becton Dickinson, San Jose, CA; ref. 16). Examination of the cells (17) was done with Cell Quest software (Becton Dickinson). The DNA laddering assay was carried out as previously described (18).

Calcium analysis. The calcium analysis used cells (5 × 10⁵/mL) preloaded with INDO-1/AM (2 μmol/L) in Buffer A [25 mmol/L HEPES, 5.4 mmol/L KCl, 0.8 mmol/L MgCl₂, 1.8 mmol/L CaCl₂, 121 mmol/L NaCl, 5.5 mmol/L glucose, 6 mmol/L NaHCO₃ (pH 7.3), 50 μmol/L β-mercaptoethanol, and 1% FCS] at 25°C for 30 minutes. Cells were isolated, resuspended in Buffer A with 10% FCS, and placed at 25°C for 20 minutes to allow deesterification of INDO-1/AM. PI (1 μg/mL) was added to the sample to gate live cells and the calcium analysis was carried out with flow cytometry at 37°C. A 325 to 360 nm wavelength excites the ratiometric INDO-1 fluorophore into emitting light at 405 nm when bound to calcium and 520 nm when free of calcium. To estimate cytoplasmic calcium concentrations, a procedure according to Eastman (19) was incorporated. Digitonin (20 μmol/L) was applied to the cells to examine the compartmentalization of INDO-1 into organelles. The application of ionomycin (20 μmol/L) and manganese (1 mmol/L) to the cells enabled the estimation of background unmetabolized INDO-1/AM contributing to the 405/520 reading (20).

RNA analysis. Total RNA was isolated with RNAzol B reagent (Tel-Test, Friendswood, TX). The RNase Protection Assay kit was used according to the instructions of the manufacturer (PharMingen). The normalizing control transcripts consisted of both glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and L32. L32 mRNA is a rRNA regulated posttranscriptionally (21).

Reverse transcription-PCR amplification of mouse L-type calcium channel α₁-subunit cDNA. First-strand cDNA synthesis from total RNA was done with oligo(dT)-primed reverse transcription using the Thermoscript reverse transcription-PCR (RT-PCR) kit (Invitrogen). The University of Wisconsin-Biotechnology Center synthesized oligonucleotide primers corresponding to conserved regions in the mouse LTCC α₁-subunit cDNA (Genbank accession number L01776; ref. 22). PCR amplification of the 920 bp fragment was done with the Herculase enzyme (Stratagene, La Jolla, CA) and gel electrophoresis was carried out to the isolate the PCR product. The Taq DNA polymerase–modified PCR product was subcloned into a pCRII TA vector (Invitrogen) and transformed into bacteria. Sequencing of multiple bacterial clones validated the expression of the Ca_v1.3 (α_1D) LTCC α₁-subunit in the WEHI-231 cells.

Reverse transcription-PCR amplification of human Ca_v1.3 (α_1D) L-type calcium channel α₁-subunit cDNA. The sequence of the primers and conditions used to amplify the LTCC gene from human breast cancer lines have been published (23, 24). The sequencing of the resulting PCR products was as described above.

Electrophoretic mobility shift assay. The electrophoretic mobility shift assay (EMSA) analysis was done as described (25). Briefly, nuclear protein extracts from a single reaction were divided into duplicate samples and incubated with a ³²P-radiolabeled oligonucleotide containing either the consensus NF-κB or Oct-1 (5′-TGTCGAATGCAAATCACTAGAA-3′; Promega, Madison, WI) binding sequence. The paired samples were then run on the same 4% nondenaturing polyacrylamide gel and analyzed by a PhosphorImager (Amersham Biosciences, Piscataway, NJ). Densitometry values of NF-κB (c-rel/p50) were normalized to paired Oct-1 and Oct-2 values from the same sample and then to the nontreated control values at the indicated time point. Supershift analysis was done with Oct-1 and Oct-2 antibodies from Santa Cruz Biotechnology (Santa Cruz, CA). The breast cancer studies used total protein extracts (6 μg) for EMSA analysis (26).

Perillyl alcohol induces apoptosis of WEHI-231 B-lymphoma cells. The monoterpenes are known inducers of apoptosis in vivo (2); however, their induction of programmed cell death in cell culture is not well characterized. To determine the sensitivity of the WEHI-231 cells to monoterpene treatment for 24 hours, PI staining was used to examine cell death. This method readily detects apoptosis in nonfixed cells with flow cytometric analysis (16). Perillyl alcohol–treated Ramos cells (another B-lymphoma cell line) helped to validate the incidence of apoptosis in B-lymphoma cells. The sorted Ramos cells showed cell populations with normal (R1), apoptotic (R2), and secondary necrotic (R3) morphologies (17) after monoterpene treatment (Fig. 1A). The perillyl alcohol–treated WEHI-231 cells also showed the fragmented nuclear morphology characteristic of apoptosis (data not shown). Furthermore, the isolated R3 (secondary necrotic gate) population of treated WEHI-231 cells displayed oligonucleosomal laddering indicative of apoptosis (Fig. 1B; ref. 18). Further examination of the WEHI-231 cells treated with perillyl alcohol for 24 hours showed a marked reduction in viability with increasing concentrations of perillyl alcohol. Collectively, these results show that the B-lymphoma cell lines undergo apoptosis upon perillyl alcohol treatment. The WEHI Bcl-X_L cells, which highly express an exogenous antiapoptotic Bcl-X_L gene, did not show a decrease in cell viability with perillyl alcohol treatment (Fig. 1C).

Studies in WEHI-231 cells showed varying cell death–inducing activities with monoterpene treatment for 24 hours (Fig. 1D). Perillyl alcohol application induced a >40-fold increase in cell death, whereas limonene and perillic acid applications elicited a minimum increase in cell death at ∼2.5- and 3.0-fold, respectively. Menthol did not induce cell death in these studies. Therefore, perillyl alcohol is the most potent, whereas limonene and perillic acid are the least potent, inducers of cell death of the monoterpenes tested in our studies.

Perillyl alcohol inhibits constitutive nuclear factor-κB activity in WEHI-231 cells. As part of the regulation of transcriptional factor activity by the 10-carbon cyclic monoterpenes (Fig. 2A), we ask the question if treatment would rapidly reduce the calcium-mediated constitutive NF-κB DNA-binding activity in the WEHI-231 cells. The NF-κB c-rel/p50 heterodimer, generally regarded as a transcriptional activator, has a transactivation domain that is absent in the transcriptionally repressive p50 homodimer complex (27). The WEHI-231 cells were treated with perillyl alcohol in 4-hour time course studies because the cells did not exhibit apoptotic morphology at this early time point with PI viability analysis (data not shown). Nuclear proteins were isolated for EMSA using a consensus κB oligomer. Perillyl alcohol application to the WEHI-231 cells showed a marked down-regulation of NF-κB DNA-binding activity at 4 hours as detected by the reduction of the NF-κB band (c-rel/p50) intensity compared with controls (Fig. 2B). The multiple bands of the Oct-1 probe were composed of the constitutive Oct-1 and Oct-2 transcriptional factors (data not shown; ref. 28) and used as a calcium-independent control with agent treatment. Further EMSA studies included the analysis of the different monoterpenes with varying anticancer activities. Menthol and limonene moderately decreased NF-κB DNA-binding activity, whereas perillic acid treatment was not effective (Fig. 2C).

EMSA analysis was also done after 24 hours to examine if the NF-κB DNA-binding levels correlated with the apoptosis observed after 24 hours. These data suggested that over time, limonene, perillic acid, and menthol actually increased nuclear NF-κB levels compared with Oct-1 levels in the WEHI-231 cells (data not shown). The apoptosis-resistant Bcl-X_L cells were next treated for 24 hours with perillyl alcohol to ask if the monoterpene prolonged the decrease of NF-κB activity without the interference from apoptotic cell death. The NF-κB DNA-binding activity in these cells is considered to be regulated similarly to the parental WEHI-231 cells (15). Perillyl alcohol (0.7 mmol/L) administration maintained a marked reduction of NF-κB levels at 4 and 24 hours in the WEHI-231 cells (Fig. 2B and D) and Bcl-X_L cells at 24 hours. Thus, perillyl alcohol treatment of the WEHI-231 cells for 24 hours show a persistent decrease of the prosurvival NF-κB DNA-binding activity.

Expression of certain nuclear factor-κB target genes is repressed in perillyl alcohol–treated WEHI-231 cells. To determine if NF-κB DNA-binding activity leads to a reduction in target gene expression, RNase protection analysis examined whether a decrease in the mRNA levels of candidate genes known to be positively regulated by NF-κB correlated with the reduction of the DNA-binding activity of the transcriptional factor following monoterpene treatment (Fig. 2). Known NF-κB transcriptionally regulated genes, such as notch, jagged, and bcl-2, were analyzed because they are associated with neoplasia (29, 30). However, these mRNAs were not highly expressed and probably not mediated by the high constitutive NF-κB nuclear levels in the WEHI-231 cells (Fig. 3A and B). Perillic acid was included as the no-effect control because this monoterpene did not alter NF-κB DNA-binding levels compared with controls (Fig. 2B). Moreover, perillic acid treatment did not consistently reduce mRNA levels either (Fig. 3A and B). Bax was analyzed because steady-state mRNA levels of this cell death transcript increased with perillyl alcohol treatment of regressing mammary tumors (2). However, Bax mRNA levels seemed to be relatively unaltered (Fig. 3A and B), whereas IκBα mRNA expression decreased below control levels with perillyl alcohol treatment. In accord with this outcome, NF-κB is a known positive regulator of the IκBα gene in the WEHI-231 cells (31). Furthermore, perillyl alcohol (0.7 mmol/L) application induced a reduction of the Bcl-2 gene family mRNA transcript, Bfl-1/A-1, to ∼50% of controls levels. Thus, a rapid decrease of NF-κB nuclear levels at 4 hours was associated with a drop in the known NF-κB–regulated antiapoptotic Bfl-l/A1 (30) and IκBα genes in WEHI-231 cells with perillyl alcohol treatment.

Forced activation of nuclear factor-κB can reverse perillyl alcohol–induced apoptosis of WEHI-231 cells. The activated CD40 receptor induces a tumor necrosis–associated factor–dependent NF-κB–mediated antiapoptotic pathway in the WEHI-231 cells (32, 33). Therefore, CD40 stimulation was used to validate that the persistent reduction of the constitutive calcium-dependent NF-κB antiapoptotic pathway is a plausible mechanism of cell death with perillyl alcohol treatment (Fig. 1). Cells were initially exposed to transient CD40 antibody pretreatment followed by perillyl alcohol application for 18 hours. A density gradient was used to isolate viable cells for gel shift analysis. EMSA analysis showed that perillyl alcohol treatment did not inhibit the NF-κB DNA-binding activity from the CD40 antibody pretreated cells compared with controls (P = 0.760; Fig. 4A). However, consistent with our previous results (Fig. 2A and D), perillyl alcohol application induced a statistically significant (P < 0.05) reduction in the NF-κB DNA-binding activity compared with controls (Fig. 4B). Cellular viability analysis showed a statistically significant (P < 0.05) protective effect with CD40 stimulation followed by perillyl alcohol application compared with the perillyl alcohol–only treated WEHI-231 cells at 24 hours (Fig. 4C). We infer that the prolonged reduction of constitutive NF-κB DNA-binding activity (Fig. 2A and D) is responsible for a significant percentage (∼80%) of the cell death in the WEHI-231 cells (Fig. 1C and D) with perillyl alcohol treatment. These results show the specificity of the targeted reduction of a calcium-dependent NF-κB cell survival pathway without interfering with a calcium-independent CD40 induced NF-κB cell survival pathway with perillyl alcohol treatment.

Perillyl alcohol treatment reduces the estimated steady-state calcium levels compared with controls. To answer the question of why the monoterpenes decreased the calcium-dependent NF-κB pathway in the WEHI-231 cells, steady-state calcium concentrations were estimated at relatively early time points after monoterpene treatment. Perillic acid application did not reduce calcium levels within 60 minutes relative to controls (Fig. 5A). Treatment with the monoterpenes (perillyl alcohol, limonene, menthol) seemed to decrease the calcium concentrations at 30 minutes compared with control levels, whereas the transient limonene– and menthol-mediated decreases approached controls at 60 minutes (Fig. 5B). However, perillyl alcohol administration elicited a consistent and statistically significant (P < 0.05) reduction of the estimated calcium concentrations at both time points relative to the controls (Fig. 5C).

L-type calcium channel antagonist decreases the calcium-dependent nuclear factor-κB pathway. Further studies characterized the LTCC channel expression in the WEHI-231 cells as a potential mechanism of how the monoterpenes reduced intracellular calcium levels in the WEHI-231 cells. The characterization of the LTCC α₁-subunit or calcium pore is indicative of a dihydropyridine-sensitive calcium channel (34). To confirm the expression of a dihydropyridine-sensitive calcium channel in the WEHI-231 cells, RNA was isolated to screen for multiple LTCC gene families by semiquantitative PCR. The PCR analysis showed the expected product, which when cloned and sequenced showed 100% homology to the Ca_v1.3 or α_1D gene family of LTCCs (Fig. 6A,, lane 2; ref. 22). An increase of calcium in cells treated with the LTCC agonist BayK8644(−) functionally shows a constitutively opened LTCC (34). BayK8644(−) treatment of INDO-1/AM-loaded WEHI-231 cells caused an increase in calcium levels from ∼70 to ∼120 nmol/L over 4 minutes using flow cytometric analysis (Fig. 6B). However, application of the BayK8644(+) LTCC antagonist did not seem to alter calcium levels over time. Therefore, these results suggest that the WEHI-231 cells express a constitutively active dihydropyridine-sensitive calcium channel that can directly influence steady-state calcium levels.

The dihydropyridine antagonists nitrendipine and nifedipine were used to test whether targeting the dihydropyridine-sensitive calcium channel would initiate the reduction of the calcium-mediated constitutive NF-κB DNA-binding activity in the WEHI-231 cells. Nitrendipine decreased NF-κB DNA-binding activity at 4 hours (Fig. 6C), whereas nifedipine did not (Fig. 6D). The prolonged treatment of nitrendipine in the Bcl-X_L cells persistently decreased NF-κB DNA-binding levels, whereas Oct-1 remained relatively unchanged (Fig. 6E). Nifedipine treatment was not effective in this assay. Furthermore, nitrendipine treatment also induced cell death in the WEHI-231 cells over 24 hours, whereas nifedipine was not effective (Fig. 6F). Neither dihydropyridine compound induced apoptosis in the Bcl-X_L cells with 24 hours of treatment. Thus, the reduction of the constitutive antiapoptotic NF-κB levels for 24 hours is associated with apoptosis at 24 hours with the dihydropyridine antagonist compound nitrendipine. These results suggest that perillyl alcohol maybe similar to nitrendipine in its mechanism of action.

Perillyl alcohol and different regulators of calcium decrease constitutive nuclear factor-κB levels in a human breast cancer cell line. We have previously characterized perillyl alcohol–induced cytostasis (0.3-1.0 mmol/L) and the induction of apoptosis (1.0 mmol/L) in human breast cancer cell lines with perillyl alcohol treatment for 3 days (35). However, the effects of perillyl alcohol on NF-κB DNA-binding activity are unknown in these breast cancer cell lines. Generally, ER− breast cancer cell lines (MDA-MB 468, MDA-MB 231) exhibit higher levels of the NF-κB DNA-binding complex consisting of the p65/p50 dimer than the estrogen receptor–positive/dependent (ER+) breast cancer cell lines (MCF-7, T47D; ref. 36). Because our current study suggested that perillyl alcohol treatment reduced calcium-dependent NF-κB levels in cells that express the Ca_v1.3 α₁-subunit (Fig. 6A), we first assayed for the expression of this human LTCC subunit. RT-PCR analysis of total mRNA and sequencing of the resulting gel-purified product showed that the MCF-7, T47D, and MDA-MB 468 cells express the Ca_v1.3 gene (Fig. 7A,, lanes 2, 3, and 5, arrow; refs. 23, 24). The MDA-MB 231 cells (Fig. 7A , lane 4) also generated an uncharacterized faint band migrating at a similar molecular weight as the other Ca_v1.3 gene bands.

To characterize the potential calcium-dependent regulation of constitutive NF-κB levels in human breast cancer cells expressing the Ca_v1.3 gene family, the calcium chelator BAPTA-AM was applied in short-term studies (2 hours). We did not detect constitutive NF-κB levels in the ER+ cell lines and thus focused on the regulation of the high NF-κB levels in the MDA-MB 468 cells (data not shown). BAPTA-AM treatment at both concentrations (10 and 30 μmol/L) significantly reduced NF-κB levels in the MDA-MB 468 cells (P < 0.001; Fig. 7B). BAPTA-AM treatment did not affect Oct-1 DNA binding. Thus, the MDA-MB 468 cells seem to express calcium-dependent NF-κB activity.

Because perillyl alcohol and nitrendipine treatment persistently reduced long-term NF-κB levels in the WEHI-231 cells (Figs. 2 and 6), we applied these same agents to the MDA-MB 468 cells. Perillyl alcohol treatment elicited a significant decrease in the NF-κB DNA-binding levels in the MDA-MB 468 cells at 24 (P < 0.05) and 48 hours (P < 0.005; Fig. 7C). Oct-1 levels were not significantly affected. Application of nitrendipine also significantly decreased NF-κB levels at 24 hours (Fig. 7D) whereas nifedipine did not (P < 0.01). Nitrendipine treatment significantly increased Oct-1 DNA binding in the MDA-MB 468 cells, suggesting that Oct-1 maybe regulated differently in these cells with nitrendipine (P < 0.001; Fig. 7D). Moreover, vehicle and nifedipine treatments were not significantly different from control values with either NF-κB or Oct-1 in these studies. Therefore, perillyl alcohol and nitrendipine treatment elicited a significant decrease in a plausible calcium-dependent NF-κB activity in an ER− breast cancer cell line.

Diverse groups of chemicals with varying molecular structures bind directly to LTCCs (37). However, we observed widely varied outcomes in the persistent decrease of constitutive calcium-dependent NF-κB DNA-binding activity using multiple compounds with known LTCC antagonistic activity in the WEHI-231 cells. In agreement with studies characterizing menthol as an LTCC antagonist (4), the monoterpenes (perillyl alcohol, menthol, limonene) exhibited at least a temporary reduction in steady-state calcium levels in our studies in the WEHI-231 cells. These results suggest that the dihydropyridine-sensitive calcium channel is a plausible monoterpene-mediated target in certain cancer cells expressing calcium-dependent NF-κB levels. In support of this conclusion, perillyl alcohol mediated a significant reduction of calcium-sensitive NF-κB levels in the MDA-MB 468 cell line shown to express a LTCC.

The calcium reducing activity of monoterpene treatment did not generally correlate with the most lipophillic monoterpene applied in our studies. Limonene, the most lipophillic monoterpene used (38), did not reduce calcium levels as markedly or persistently as perillyl alcohol relative to controls with short-term treatment. The control dihydropyridines (nitrendipine, nifedipine) applied in our studies bind to the dihydropyridine receptor with approximately similar affinities (39). Reasonably increasing the concentration (30-50 μmol/L) of the lower affinity dihydropyridine (nifedipine) in the WEHI-231 cells did not have an effect on NF-κB DNA-binding activity or apoptosis. Therefore, the efficacy of nitrendipine treatment in reducing constitutive calcium-dependent NF-κB activity was plausibly dependent on the binding and subsequent regulation of the dihydropyridine-sensitive calcium channel. Alternatively, other sites of action could be involved in eliciting this unreported anticancer activity of the well-studied dihydropyridines. However, perillyl alcohol treatment seemed to reduce calcium levels longer than the other monoterpenes. Thus, perillyl alcohol may regulate the dihydropyridine-sensitive channel differently from the other monoterpenes. Furthermore, the calcium-independent constitutive DNA-binding activity of the Oct transcriptional factors in the WEHI-231 cells was unaffected with agent treatment, showing relative specificity toward calcium-dependent NF-κB activity. Our data supports the hypothesis that the different outcomes with monoterpene and dihydropyridine treatments are due to the agent-specific regulation of the dihydropyridine-sensitive calcium channel in the WEHI 231 cells.

The concentrations of the dihydropyridines used in our studies are equal to or less than the concentrations to examine LTCC-mediated differentiation and signal transduction pathways (40, 41). Thus, the relative comparisons made between the monoterpenes and the dihydropyridines with the concentrations used in the current study are rationally based. Furthermore, the monoterpenes (perillyl alcohol, limonene, perillic acid) elicit cytostatic activities in mammalian fibroblast and carcinoma cells in vitro with the concentrations used in the present paper (<1.0 mmol/L; refs. 42–44). The apoptotic activity of monoterpene (limonene and perillyl alcohol) treatment in vivo in regressing rat mammary tumors is associated with high serum concentrations (∼0.8 mmol/L) of metabolites, including perillic acid (1, 2, 45). In the present study, however, perillic acid and limonene were noneffective inducers of apoptosis in the WEHI-231 cells compared with perillyl alcohol. Therefore, the molecular mechanisms associated with the monoterpene-induced apoptosis in mammary cancer either do not involve the reduction of NF-κB antiapoptotic activity or are not manifested within the duration of our experiments in the WEHI-231 cells.

A constitutive calcium-dependent process targets the proteasome-independent degradation of IκBα, releasing NF-κB to the nucleus in the WEHI-231 cells (12). This unique degradation pathway seems to be saturated because calcium ionophore (46) and BayK8644(−) treatment of the WEHI-231 cells did not augment NF-κB DNA-binding activity (data not shown). Furthermore, stimulation of calcium-independent NF-κB DNA-binding activity through the CD40 receptor abrogated the effects of perillyl alcohol treatment. Our previous work implicated the involvement of the calcium effector protein calmodulin in the constitutive pathway (15). Calmodulin inhibition transiently decreased IκBα degradation, resulting in the concomitant reduction of NF-κB DNA-binding activity. The modulation of the dihydropyridine-sensitive calcium channel upon administration of several monoterpenes in the present study resulted in kinetically similar transient reductions of NF-κB DNA-binding activity. Previous studies with calmodulin inhibitors indicate that a prolonged reduction in constitutive NF-κB activity is required to cause apoptosis in these cells (15). Similarly, in the current study, the induction of apoptosis with perillyl alcohol and nitrendipine treatments associated with a prolonged decrease of NF-κB activity.

Agents targeting dihydropyridine-sensitive calcium channels are effective for treatment of cardiovascular diseases (37). However, a knowledge gap exists on whether the rapidly stimulated reduction of steady-state calcium levels in neoplastic tissues is a desirable end point in chemotherapy and/or chemoprevention. Interestingly, this premise was useful in elucidating a novel calcium-dependent proteasome inhibitor–resistant (PIR) pathway of constitutive NF-κB activity (12, 47). Studies in certain mammary tumor models suggest that HER-2/neu overexpression can induce proteasome-independent NF-κB activity (48). However, the relationship between HER-2/neu overexpression and PIR constitutive NF-κB activation remains undefined. Nevertheless, our study shows that certain compounds (perillyl alcohol and nitrendipine) from two different families of agents known to bind to LTCCs can significantly reduce NF-κB in a widely used ER− human breast cancer cell line. The outcomes of the inhibition of NF-κB activity probably varies in different cancer cells because this transcription factor regulates multiple target genes ranging from those affecting proliferation and tumor promotion to cell survival (49). We conclude that the dihydropyridine-sensitive calcium channel is a candidate receptor for monoterpene regulation of NF-κB levels in certain cancers.

Grant support: NIH grants CA38128 (M.N. Gould) and CA81065 (S. Miyamoto).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank J. Haag and Dr. Laurie Shepel for reading the manuscript and Kathy Schell (University of Wisconsin Flow Cytometry Facility) for support.