Synthesis and Evaluation of the Novel Prostamide, 15-Deoxy, Δ^12,14-Prostamide J₂, as a Selective Antitumor Therapeutic

Skin cancer is the most common human neoplasm in America with one in every five individuals developing this malignancy in their lifetimes (1). Skin cancer is comprised of two major subtypes including nonmelanoma skin cancer (NMSC) and melanoma. NMSC consists of basal and squamous cell carcinoma and is less aggressive and metastatic compared with melanoma. The annual cost of treating all types of skin cancer is estimated to be $8.1 billion, representing a substantial financial burden on the United States health care system (2). Furthermore, studies indicate that the incidence of these cancers is on the rise primarily beacuse of increased exposure to ultraviolet radiation.

Recent data from our group identified a newly discovered prostaglandin-ethanolamide (prostamide), 15-deoxy, Δ^12,14-prostamide J₂ (15d-PMJ₂; refs. 3, 4). Specifically, the endocannabinoid, arachidonoyl ethanolamide (AEA) was metabolized by cyclooxygenase-2 (COX-2) to prostaglandin-ethanolamide D₂ [prostamide D2 (PMD₂)], which was converted to the terminal product, 15d-PMJ₂. Endocannabinoids are endogenously synthesized lipids that bind to and signal through cannabinoid receptors. These lipid messengers modulate neuronal signaling, inflammation, cardiovascular function and other physiologic processes. In addition, endocannabinoids and their receptors are reported to be involved in both the development and elimination of cancer, indicating that the pathophysiologic interaction between cancer and endocannabinoids requires further study (5–7). Several reports suggest that the antitumor activity of AEA occurred through cannabinoid receptor–dependent and -independent mechanisms including the induction of endoplasmic reticulum (ER) stress (8–12).

ER stress occurs when the protein folding load in a cell exceeds the protein folding capacity, causing an accumulation of unfolded proteins and eliciting what is known as the unfolded protein response (UPR; refs. 13, 14). Induction of the UPR activates HSP70/GRP78 (BiP), thereby releasing the UPR sensors, double-stranded RNA-activated protein kinase (PRK)-like endoplasmic reticulum kinase (PERK), activated transcription factor 6 (ATF6), and inositol requiring kinase-1 (IRE1α) to propagate their signals (15). Depending on the intensity and duration of the ER stress signal, tumor cells can initiate either prosurvival or proapoptotic pathways. In the survival pathway, IRE1, PERK, and ATF6 activate signal transduction cascades that culminate in increased peptide folding and protein degradation responses with the goal of resolving the accumulated unfolded proteins (16). In the death pathway, the UPR sensors upregulate the expression of transcriptional factor C/EBP homologous protein (CHOP10)/growth arrest and DNA damage–inducible 153 (GADD153), thereby triggering apoptosis (17). In addition, the cytotoxic ER stress pathway was also found to be regulated by cellular death inducer protein-1 (CDIP1) through its interaction with B-cell receptor–associated protein 31 (BAP31; ref. 18).

Previous studies by our group demonstrated that ER and oxidative stress were needed for AEA cytotoxicity, but that the cannabinoid receptors had little impact on its activity. The prevention of AEA-mediated cell death by blocking 15d-PMJ₂ production suggested that the antiproliferative action of AEA was caused by 15d-PMJ₂. Therefore, in the current study, the antitumor activity of 15d-PMJ₂ was directly examined by synthesizing the novel prostamide and examining its effect on ER stress and cell death in vitro and in vivo.

15-deoxy, Δ^12,14-PGJ₂ (J-series prostaglandin ELISA kits) was purchased from Cayman Chemical company. TBTU, ethanolamine, diispropylethylamine, acetonitrile, CDCL₃, and the antibody for β-actin were purchased from Sigma-Aldrich. Antibodies directed toward full-length (FL)/cleaved caspase-3, FL/cleaved PARP, FL/cleaved caspase-8, FL caspase-4, P-PERK and total-PERK were from Cell Signaling Technology. Cleaved caspase-4 was from Invitrogen; anti-CHOP10, and anti-BAP31 antibodies were from Santa Cruz Biotechnology, anti-GAPDH was from Millipore, and anti-CDIP1 was from Novus Biologicals. Anti-rabbit 800CW and anti-mouse 680RD secondary antibody IRDyes were from LI-COR Biosciences. Anti-rabbit Alexa Fluor 555 was from Invitrogen while anti-rat Alexa Fluor was from Jackson Immunoresearch. Caspase-Glo 3/7 assays and MTS reagent were from Promega Life Sciences.

15-deoxy-Δ^12,14-prostaglandin J₂ and 9,10-dihydro-prostaglandin J₂ (Cayman Chemical) was dissolved in a small volume of acetonitrile, approximately 1.5 mole equivalent of O-(Benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (TBTU), and 2 mole equivalents of diisopropylethylamine were added. The solution was stirred for approximately 60 minutes and 2 mole equivalents of ethanolamine were added. The reaction solution was stirred for 24 hours, vacuum filtered, and the solvent removed by rotary evaporation to yield a yellow crude product. This product was redissolved into ether solvent and washed three times with water. The organic phase was collected and evaporated to yield a pure faint yellow product (yield 99%), which was characterized by ¹H NMR (400 MHz, CDCl₃): 0.921 (t, 3H); 1.340 (m, 4H); 1.482 (m, 2H); 1.706 (q, 2H); 2.053 (q, 3H); 2.177 (t, 2H); 2.258 (q, 2H); 2.372 (m, 1H); 2.592 (m, 1H); 3.439 (q, 2H); 3.634 (m, 1H); 3.744 (t, 2H); 5.365 (m, 1H); 5.472 (m, 1H); 5.953 (s, 1H); 6.311 (m, 3H); 6.965 (d, 1H); 7.510 (m, 1H); and ESI-MS: [M+H]⁺ 360.2270 (patent # 9328060).

Nontumorigenic HaCaT keratinocytes (purchased from Cell Line Service), human squamous carcinoma cell line A431 (purchased from ATCC), and mouse B16F10 melanoma cells were cultured in DMEM (Invitrogen) containing 10% heat-inactivated FBS, penicillin (100 mg/mL), streptomycin (100 mg/mL), sodium pyruvate, and glutamine. The murine melanocyte Melan-A cell line was purchased from the Bennett-Sviderskaya laboratory (Molecular Cell Sciences Research Centre, St. George's, University of London, United Kingdom) and cultured in RPMI1640 medium supplemented with 10% FCS, penicillin (100 mg/mL), streptomycin (100 mg/mL), glutamine, and tetradecanoylphorbol acetate (200 nmol/L). Patient melanoma tissue (acquired from Dr. Timothy Fitzgerald, Department of Surgery, East Carolina University, Greenville, NC) was washed and dissociated using collagenase type 4 (Worthington Biochemical Corporation) then cultured in Mel 2 medium containing MCDB 153/Leibovitz L-15 medium (80%/20%), 2% heat-inactivated FBS, insulin (5 μg/mL), calcium chloride (1. 68 mmol/L), and bovine pituitary extract (15 μg/mL). Primary melanoma cells were verified by measuring tyrosinase levels via confocal microscopy. The murine squamous carcinoma cell line JWF2 was cultured in Eagle's minimal essential medium (US Biological) containing 5% heat-inactivated FBS, penicillin (100 mg/mL), streptomycin (100 mg/mL), nonessential amino acids, and glutamine. JWF2 cells were a kind gift from Dr. Susan Fischer (University of Texas MD Anderson Cancer Center, Houston, TX).

A431, HaCaT, B16F10, Melan-A, or JWF2 cells were plated in 96-well dishes and cultured for 48 hours. Serum-free medium containing different concentrations of 15d-PMJ₂ or 15d-PGJ₂ was then added to the cells. Twenty microliters of MTS reagent were added to each well after 12–24 hours of incubation and absorbance was measured at 495 nm as directed by the manufacturer (Promega). LC₅₀ (concentration required to kill 50% of the cell population) values were determined using GraphPad software.

Cells were plated in white-walled 96-well plates and cultured for 48 hours. Medium containing the appropriate agents were added to the cells for the indicated amount of time. One-hundred microliters of Caspase-Glo 3/7 reagent was then added to each well as directed by the manufacturer and luminescence was measured using a luminometer.

Culture medium collected from JWF2 and HaCaT cells was assayed for J-series prostaglandins using ELISA kits according to the manufacturer's protocol as described previously (3).

All cell lines were incubated in medium containing the indicated agents. Plates were subsequently scraped and protein concentration of the cell lysates was determined with BCA reagents (Pierce). Equal concentrations of each sample were loaded onto SDS-PAGE gels and protein bands transferred to nitrocellulose membranes (GE Healthcare Life Sciences). Blocked membranes were incubated with Pierce Protein Free Blocking Buffer (Thermo Scientific) containing FL/cleaved caspase-3 (1:1,000), FL/cleaved caspase-4 (1:1,000), FL/cleaved caspase-8 (1:1,000), FL/cleaved PARP (1:1,000), anti-P-PERK (1:500), anti-t-PERK (1:500), anti-CHOP10 (1:1,000), anti-β-actin (1:5,000), anti-GAPDH (1:5,000), or anti-CDIP1 (1:1,000) antibodies. Protein bands were visualized using the LI-COR system and digitized images were quantified using ImageJ software.

A431 cells were grown on culture slides followed by incubation in medium containing the indicated agents. Cells were then fixed with methanol, incubated with permeabilization buffer (0.1% Triton X-100 in PBS) for 10 minutes, and blocked with blocking buffer (PBS + 3% FBS + 0.5% Tween20) for 1 hour. Blocked cells were then incubated with indicated primary antibodies and the appropriate immunofluorescence-tagged secondary antibodies. Images were acquired and analyzed by confocal laser microscopy (Zeiss LSM 700 confocal microscope system). Fluorescence intensity was quantified using Zen Blue software.

All experiments were approved and conducted in accordance with the East Carolina University Institutional Animal Care and Use Committee guidelines. C57BL/6 female mice (7-week-old animals) were purchased from Jackson Laboratories. The flank region was shaved and the tumors were established by inoculating 2 × 10⁵ B16F10 cells subcutaneously as described previously (19).When tumors became palpable (7–10 days postimplantation), mice were either given daily peritumoral injections (0.5 mg/kg 15d-PMJ₂ or PBS as vehicle) or left untreated as a control for 5 days. Animal body weights were measured and recorded daily. The length and width of the tumors were measured daily with calipers [daily tumor volume was calculated as: volume = length² × width ×(π/6)]. After mouse euthanasia, tumors were dissected, blotted dry, and weighed at the end of the study. Animal livers were tested for cell death by TUNEL assay.

Tumors and livers were formalin-fixed, paraffin-embedded, and sectioned at 5-μm thickness. Sections were deparaffinized and rehydrated with a series of ethanol solutions and water. Endogenous peroxidase activity was blocked with 3.0% hydrogen peroxide for 30 minutes at room temperature. Sections were then incubated overnight with primary antibody followed by secondary antibody for 30 minutes. Preliminary experiments were conducted to optimize conditions for primary and secondary antibody. Staining was visualized using the Superpicture 3rd GEN IHC Detection Kit (ThermoFisher) and then counterstained in Harris' hematoxylin. Terminal deoxynucleotidyl transferase deoxyuridine triphosphate nick-end labeling (TUNEL) staining of paraffin-embedded tumor and liver sections was performed according to the standard protocols provided by the manufacturer (Roche Diagnostics). Four fields were evaluated per section with a minimum of twelve sections per tumor.

Data are representative of three independent experiments unless otherwise indicated. Data are presented as mean ± SEM. Student t test, one- and two-way ANOVA followed by Tukey post hoc analysis were carried out using GraphPad Prism and Microsoft Excel.

Our previous studies determined that the cytotoxicity of AEA in tumorigenic keratinocytes was mediated by its novel metabolic product, 15-deoxy, Δ^12,14-prostamide J₂ (15d-PMJ₂; ref. 3). As such, 15d-PMJ₂ (Fig. 1A, bottom) was synthesized using 15d-PGJ₂ (Fig. 1A, top) as a substrate. Verification of the product was performed by both ¹H-NMR and ESI-MS. Comparison of the ¹H-NMR spectra of the synthetic substrate (15d-PGJ₂; Supplementary Fig. S1A) and product (15d-PMJ₂; Supplementary Fig. S1B) demonstrated that addition of the ethanolamide group occurred with an approximate 99% yield.

The cytotoxic activity of synthetic 15d-PMJ₂ was examined in skin cancer cells by conducting MTS assays. 15d-PMJ₂ caused a concentration-dependent decrease in viability in squamous carcinoma cell lines (A431 and JWF2), a melanoma cell line (B16F10), and primary patient melanoma cells (Fig. 1B–D; Supplementary Fig. S2A). Cell treatment with 5 μmol/L 15d-PMJ₂, (the concentration utilized in subsequent in vitro experiments) led to an 80%, 63%, 80%, and 77% decrease in A431, B16F10, primary melanoma, and JWF2 survival, respectively. 15-deoxy, Δ^12,14 prostaglandin J₂ (15d-PGJ₂) is a molecule that is distinct, yet structurally related to 15d-PMJ₂. 15d-PGJ₂ is synthesized by COX-2 from the substrate arachidonic acid rather than AEA. Because arachidonic acid–derived 15d-PGJ₂ induces tumor cell death (20), the cytotoxic activity of this molecule was compared with novel, 15d-PMJ₂. Endocannabinoid metabolite, 15dPMJ₂, displayed greater potency versus the arachidonic acid metabolite, 15d-PGJ₂, in A431 cells (LC₅₀ = 4.7 μmol/L vs. >20 μmol/L), B16F10 (LC₅₀ = 3.3 μmol/L vs. 15.5 μmol/L), primary melanoma (LC₅₀ = 3.0 μmol/L vs. 10.5 μmol/L), and JWF2 (LC₅₀ = 3.3 μmol/L vs. 11.4 μmol/L).

To examine whether 15d-PMJ₂ preferentially targets tumor cells, tumorigenic skin cells (A431 and B16F10), and nontumorigenic skin cells (HaCaT and Melan-A) were treated with 15d-PMJ₂ and viability was assessed by conducting MTS assays. Tumorigenic A431 keratinocytes were more susceptible to 15d-PMJ₂ cytotoxicity as reflected by a 53% and 83% reduction in viability at 5 μmol/L and 10 μmol/L, respectively, whereas a minimal reduction in viability (5% at 5 μmol/L and 8% at 10 μmol/L) was observed at the same concentration in HaCaT cells (Fig. 2A). Similarly, B16F10 melanoma cells were preferentially targeted by 15d-PMJ₂ with a 64% and 90% decrease in viability at 5 μmol/L and 10 μmol/L, respectively, compared with Melan-A cells whose survival increased at 5 μmol/L and decreased by 24% at the 10 μmol/L concentration (Fig. 2B).

The selective induction of apoptosis by 15d-PMJ₂ was also observed in skin cancer cells. Caspase-3 and PARP cleavage increased in a time-dependent manner in A431 cells treated with 15d-PMJ₂, whereas cleavage of the apoptotic markers was not detected in HaCaT cells (Fig. 2C). In addition, caspase-3 and PARP cleavage was observed in B16F10 melanoma cells but not in nontumorigenic Melan-A cells (Fig. 2D). Consistent with these findings, apoptosis was induced by 15dPMJ₂ in primary melanoma cells (Fig. 2F) and JWF2 (Supplementary Fig. S2B) skin cancer cells. However, caspase-3 activation was less robust for arachidonic acid–derived 15d-PGJ₂ compared with 15d-PMJ₂ (Fig. 2C--E; Supplementary Fig. S2B).

The selectivity of 15d-PMJ₂ toward tumor cells was further tested by examining the cytotoxic activity of its precursor molecule, prostamide D₂ (PMD₂). Similar to 15d-PMJ₂, PMD₂ significantly reduced tumorigenic cell viability compared with nontumorigenic cell viability. Exogenous administration of PMD₂ (20 μmol/L) elicited a 76% decrease in JWF2 cell viability, whereas PMD₂ had no effect on HaCaT cell survival (Supplementary Fig. S2C). Interestingly, comparable levels of J-series prostaglandins were present in both the tumorigenic and nontumorigenic cells (Supplementary Fig. S2D) suggesting a specific strategy for 15d-PMJ₂ in targeting tumor cells.

The ER stress pathway has been identified as a chemotherapeutic target that can confer selective toxicity (21). During ER stress, PERK becomes activated by undergoing homodimerization and autophosphorylation. Furthermore, insurmountable or prolonged ER stress increases the expression of proapoptotic CHOP10. As such, PERK phosphorylation and CHOP10 expression were examined as indicators of ER stress. ER stress was preferentially activated by 15d-PMJ₂ in tumorigenic skin cells as demonstrated by the notable increase in phospho-PERK levels in tumorigenic A431, B16F10, and JWF2 cells (Fig. 3A and B; Supplementary Fig. S2E). However, ER stress protein expression did not increase in nontumorigenic HaCaT and Melan-A cells. Moreover, expression of the cytotoxic ER stress protein, CHOP10, was increased in tumorigenic but not nontumorigenic skin cells (Fig. 3A). Arachidonic acid–derived 15d-PGJ₂ also selectively upregulated both ER stress proteins in the tumorigenic cell lines, although 15d-PMJ₂ elicited greater induction of cytotoxic CHOP10.

To test whether ER stress was required for 15d-PMJ₂–mediated apoptosis, two distinct ER stress pathways were blocked using the pharmacologic ER stress inhibitors, salubrinal and 4-phenylbutrate (4-PBA). Salubrinal inhibits PP1/GADD34, which prevents the dephosphorylation of phospho-eIF2α, thereby averting propagation of the cellular ER stress signal. On the other hand, PBA inhibits ER stress by acting as a chemical chaperone that enhances resolution of unfolded proteins (22, 23). Blockade of ER stress with 4-PBA or salubrinal prevented 15d-PMJ₂–mediated apoptosis in each of the skin cancer cell lines (Fig. 3C and D; Supplementary Fig. S2F), consistent with our previous evidence that ER stress is needed for AEA-induced apoptosis (3).

The biochemical mechanism by which 15d-PMJ₂ elicits ER stress was examined by investigating the role of the double bond contained within the cyclopentenone ring of the molecule. It has been demonstrated that in arachidonic acid–derived 15d-PGJ₂, the double bond promotes oxidative stress through covalent interactions with cysteine-containing proteins in an ordered fashion (24). The role of the cyclopentenone double bond in ER stress, however, has not been investigated. As such, a neutral analogue of 15d-PMJ₂ (neutral 15d-PMJ₂; Fig. 4A) lacking the reactive double bond was synthesized (Materials and Methods section). Unlike 15d-PMJ₂, neutral 15d-PMJ₂ did not elicit cell death in tumorigenic skin cells (Fig. 4B). Furthermore, the neutral molecule was unable to activate ER stress pathway proteins as notable increases in PERK phosphorylation or CHOP10 expression were not observed (Fig. 4B).

To further characterize the mechanism of 15d-PMJ₂–mediated ER stress apoptosis, additional modulators of the cytotoxic pathway were examined. CDIP1 has been identified as a p53 transcriptional target that is a key transducer of ER stress apoptosis signals (25). In the presence of ER stress, CDIP1 expression is upregulated and the protein dimerizes with constitutively expressed BAP31. BAP31–CDIP complexes are necessary for BAX oligomerization and apoptosis via caspase-8 (18). In addition, caspase-4 is known to be selectively cleaved as a consequence of ER stress–initiated apoptosis (26). 15d-PMJ₂, but not neutral 15d-PMJ₂, significantly upregulated the expression of CDIP1 and cleavage of caspase-4 (Fig. 4D). Furthermore, CDIP1 was observed to colocalize with Bap31, whereas the neutral molecule failed to promote colocalization (Fig. 4E). Moreover, the cleaved form of caspase-8 was selectively-induced by 15d-PMJ₂ rather than its neutral analogue. These results provide the first evidence that the cyclopentenone double bond is required for the generation and propagation of ER stress apoptotic signals.

The effect of 15d-PMJ₂ on tumor growth was examined in the well-established C57BL/6 mouse, B16F10 melanoma tumor model. Mice were dosed with 0.5 mg/kg 15d-PMJ₂, vehicle, or dosing was omitted. 15d-PMJ₂ significantly reduced B16F10 solid tumor growth as animals dosed with 0.5 mg/kg exhibited a mean tumor volume of 20.0 mm³ compared with 85.3 and 115.6 mm³ in untreated and vehicle animals, respectively (Fig. 5A). In addition, mean tumor weights were dramatically reduced by 15d-PMJ₂ (Fig. 5B). Of importance, animals treated with 15d-PMJ₂ exhibited greater TUNEL-positive tumor cells than vehicle or untreated mice indicating a prominent induction in cell death by this agent (Fig. 5C).

Insight was gained on potential mechanisms of tumor cell death by evaluating ER stress levels in the tumors. 15d-PMJ₂ promoted ER stress as demonstrated by significant increases both phospho-PERK and CHOP10 content in the tumors (Fig. 6A and B). As anticipated, none of the treatments altered the levels of total PERK expression in the tumors. Of note, significant differences in body weight and liver cytotoxicity (as measured by TUNEL staining; data not shown) between the experimental groups was not observed suggesting that peritumoral installation of 15d-PMJ₂ did not produce overt toxicity.

In the current study, our goal was to synthesize the novel molecule, 15d-PMJ₂, and evaluate its activity and mechanism of cytotoxicity in skin cancer cells. A strong cytotoxic effect against NMSC and melanoma cells was exhibited by 15d-PMJ₂ consistent with observations that its precursor molecule, AEA, causes cell death in different cancer cell types (12). We also determined that tumorigenic skin cancer cells were more sensitive to the cytotoxic action of 15d-PMJ₂ than nontumorigenic skin cells suggesting that this agent exhibits selective toxicity. It was further demonstrated that the ER stress pathway, an important determinant of cell death and survival (27), played a critical role in the initiation of death by 15d-PMJ₂. Moreover, the cyclopentenone double bond in 15d-PMJ₂ was required for ER stress–mediated apoptosis. Furthermore, ER stress and cell death were detected in tumors of 15d-PMJ₂–treated animals whose tumors were significantly reduced in size compared with control group animals. These results support our in vitro findings that 15d-PMJ₂ is a potent and selective ER stress apoptosis inducer in tumor cells.

15d-PMJ₂ exhibited striking cytotoxicity against skin cancer cell lines, primary patient melanoma cells, as well as xenograft melanoma tumors suggesting a potential role for this agent in cancer chemotherapeutic regimens. The IC₅₀ values of 15d-PMJ₂–treated tumorigenic cells were in the low micromolar range and tumors isolated from animals treated with 15d-PMJ₂ contained high levels of cell death, which likely caused the lesions to be smaller in size than control group mice. Other research showed that arachidonic acid–derived 15d-PGJ₂, reduced tumor growth and increased cell death in leukemia, melanoma, and colorectal carcinoma xenografts (28, 29). Cholangiocarcinoma xenograft tumor growth was also decreased by AEA, a precursor of 15d-PMJ₂ (30, 31). In light of our finding that the cytotoxic effect of AEA was primarily mediated by its metabolic product, 15d-PMJ₂ (3), this agent may prove to be useful against different cancer cell types and may also curb tumor growth in humans.

ER stress plays a key role in initiating apoptosis, and agents that target this pathway can inhibit proliferation and induce death selectively in cancer cells. ER stress is heightened in cancer cells as a result of the increased protein folding demand that is needed to drive uncontrolled cell proliferation (16). In addition, hypoxic conditions within most tumors impairs ATP production leading to an accumulation of unfolded proteins (16). As such, agents that increase ER stress cause the cytotoxic threshold to be reached more readily in tumor cells than in nontumor cells whose endogenous stress levels are lower. The data shown here indicate that ER stress was preferentially activated in tumor cells and that this pathway was required for 15d-PMJ₂–mediated cell death. ER stress was also found to be upregulated in tumors of animals treated with 15d-PMJ₂. In alignment with these results, arachidonic acid–derived J-series prostaglandins induced cytotoxic ER stress in tumor cell lines, although the selectivity of ER stress induction in tumor cells was not examined (32). Antineoplastic agents including bortezomib are being utilized clinically to modulate the ER stress pathway as a therapeutic strategy (33). Hence, agents such as 15d-PMJ_2, which promote cytotoxic ER stress, may provide an effective approach for treating cancers in which ER stress levels can be exploited.

15d-PMJ₂ contains an electrophilic α,β-unsaturated carbonyl group on its cyclopentenone ring that readily reacts by Michael addition with free sulfhydryls of cellular proteins (34–37). The same electrophilic double bond is present in all cyclopentenone prostaglandins including the J- and A-series prostaglandins and prostamides. Generation of oxidative stress by arachidonic acid–derived 15d-PGJ₂ has been attributed to interactions with antioxidants including glutathione and thioredoxin reductase (38, 39). These covalent interactions neutralize the antioxidants thereby allowing unchecked oxidant activity. However, it was unclear whether the cyclopentenone double bond also regulated ER stress. Elimination of the reactive group in 15d-PMJ₂ prevented the initiation of ER stress and hence apoptosis, thereby identifying the double bond as a critical moiety for ER stress initiation. Studies are currently underway by our group to identify ER stress–modulating proteins that form covalent interactions with 15d-PMJ₂.

Prostaglandins are formed as a result of the metabolism of AA by COX-2, whereas prostaglandin-ethanolamides (prostamides) are produced from the metabolism of AEA by COX-2. However, little is known about the biological differences between these structurally related molecules. Kozak and colleagues found that prostamides possessed significantly longer plasma half-lives and were less subject to metabolic degradation (40). Matias and colleagues established that prostamides of the E, F, and D-series possessed unique pharmacologic and biological activity compared with their prostaglandin counterparts as they did not bind to traditional prostaglandin receptors (41). As a result of these findings, considerable effort is focused on identifying cellular receptors for the prostamides (42, 43). We determined that the activity of 15d-PMJ₂, was similar to 15d-PGJ₂, but that 15d-PMJ₂ was a more potent inducer of ER stress, cell death, and apoptosis. As such, 15d-PMJ₂ may interact with distinct cellular targets or have robust stability compared with 15d-PGJ₂ providing a potential explanation for the differences in activity shown here.

Adverse effects associated with cancer chemotherapeutic agents limit their use and are commonly attributed their cytotoxic effect on noncancerous cells. Hence, an important goal of this study was to determine whether the tumorigenic cells were more sensitive to 15d-PMJ₂ than nontumorigenic cells. 15d-PMJ₂ induced nearly complete cell death in tumorigenic keratinocytes and melanocytes, whereas partial cell death occurred in counterpart nontumorigenic cells only at concentrations several fold higher than the effective concentration used in this study. Indeed, cleavage of the apoptotic proteins, caspase-3, and PARP, was initiated solely in the tumor cells. Moreover, tumorigenic keratinocytes treated with exogenous PMD₂ (which is the upstream metabolic precursor of 15d-PMJ₂) produced similar levels of 15d-PMJ₂ as nontumorigenic keratinocytes yet the tumorigenic cells were more prone to death indicating that a selective tumor cell target for 15d-PMJ₂ likely exists. These findings agree with our previous report that AEA preferentially eliminated tumorigenic keratinocytes (12). Tumorigenic T cells were also more susceptible to cell death caused by the structurally related molecule, 15d-PGJ₂, than nontumor T-cells (44). As overt toxicity was not observed in this or other studies utilizing J-series prostaglandins or cannabinoids (29, 30), 15d-PMJ₂ may be a safe and effective agent for cancer treatment.

D.A. Ladin, C. Burns, and R. Van Dross have ownership interest (including patent) in 15d-PMJ₂. No potential conflicts of interest were disclosed by the other authors.

Conception and design: D.A. Ladin, T.L. Fitzgerald, R. Van Dross

Development of methodology: L.V. Yang, C. Burns, R. Van Dross

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.A. Ladin, E. Soliman, R. Escobedo, L.V. Yang, C. Burns, R. Van Dross

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.A. Ladin, E. Soliman, R. Escobedo, T.L. Fitzgerald, R. Van Dross

Writing, review, and/or revision of the manuscript: D.A. Ladin, E. Soliman, T.L. Fitzgerald, L.V. Yang, C. Burns, R. Van Dross

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.A. Ladin, T.L. Fitzgerald

Study supervision: R. Van Dross

Other (chemical synthesis): D.A. Ladin

We thank Dr. William Allen and David Ferral for their assistance with organic synthesis. We also thank Eddie Sanderlin and Reuben Chemmanam for their help with the in vivo study and Joani Zary for her assistance with tissue processing.

This work was supported by grants from Golfers Against Cancer (to R. Van Dross), the Brody Brother Endowment (to R. Van Dross and E. Soliman), and the Division of Research, Internal Seed Grant (R. Van Dross and C. Burns).

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.