Abstract
Despite recent advances in the development of novel therapies against castration-resistant prostate cancer, the advanced form of the disease remains a major treatment challenge. Aberrant sphingolipid signaling through sphingosine kinases and their product, sphingosine-1-phosphate, can promote proliferation, drug resistance, angiogenesis, and inflammation. The sphingosine kinase 2 inhibitor ABC294640 is undergoing clinical testing in cancer patients, and in this study we investigated the effects this first-in-class inhibitor in castration-resistant prostate cancer. In vitro, ABC294640 decreased prostate cancer cell viability as well as the expression of c-Myc and the androgen receptor, while lysosomal acidification increased. ABC294640 also induced a greater than 3-fold increase in dihydroceramides that inversely correlated with inhibition of dihydroceramide desaturase (DEGS) activity. Expression of sphingosine kinase 2 was dispensable for the ABC294640-mediated increase in dihydroceramides. In vivo, ABC294640 diminished the growth rate of TRAMP-C2 xenografts in syngeneic hosts and elevated dihydroceramides within tumors as visualized by MALDI imaging mass spectroscopy. The plasma of ABC294640-treated mice contained significantly higher levels of C16- and C24:1-ceramides (but not dihydro-C16-ceramide) compared with vehicle-treated mice. In summary, our results suggest that ABC294640 may reduce the proliferative capacity of castration-resistant prostate cancer cells through inhibition of both sphingosine kinase 2 and dihydroceramide desaturase, thereby providing a foundation for future exploration of this small-molecule inhibitor for the treatment of advanced disease. Mol Cancer Ther; 14(12); 2744–52. ©2015 AACR.
Introduction
Prostate cancer is a major health issue among men worldwide. Currently, more than 200,000 new prostate cancer diagnoses are made annually and nearly 30,000 men die from the disease each year in the United States alone. One of the main challenges in prostate cancer is effective treatment of castration-resistant disease. Although the FDA recently approved several new therapies including Enzalutamide, Abiraterone, Provenge, and Xofigo, the survival advantage is limited to a few months (1). Thus, better options to more effectively treat advanced prostate cancer are needed.
A recent study has shown that deregulation of sphingolipid metabolism may contribute to the development of radiation resistance through upregulation of acid ceramidase (2). Acid ceramidase, an enzyme of the sphingolipid metabolic pathway, hydrolyzes the antiproliferative sphingolipid ceramide, to sphingosine, which is a substrate for sphingosine kinases (3). Through phosphorylation of sphingosine, sphingosine kinases produce sphingosine-1-phosphate (S1P), which promotes proliferation, drug resistance, angiogenesis, and inflammation. Consequently, sphingosine kinases as well as S1P itself have evolved as potential therapeutic targets (4–6).
The sphingosine kinase isozymes (SPHK1 and SPHK2) share significant homology but localize to different subcellular compartments and appear to have nonredundant biologic functions (7, 8). The recent development of a small-molecule inhibitor known as ABC294640 that specifically inhibits SPHK2 has led to some novel insights into the biology of this enzyme (9, 10). Pharmacologic inhibition of SPHK2 by ABC294640 results in antitumor effects through various mechanisms, including apoptosis or autophagic cell death (11), inhibition of NF-κB–mediated chemotherapy resistance (12), and synergy with other cancer therapeutics (13, 14). ABC294640 can also inhibit AKT and ERK signaling, elicit anti-estrogenic effects in breast cancer cells, and increase proteasomal degradation of c-Myc (14–16). On the basis of promising results in preclinical models of multiple diseases, including malignancies, ABC294640 has now entered the clinic for evaluation in solid and hematologic cancers (clinicaltrials.gov identifier: NCT01488513, NCT02229981). In this study, we investigated the effects of ABC294640 on castration-resistant prostate cancer cells. Our results suggest that ABC294640 reduces the proliferative capacity of castration-resistant prostate cancer cells, at least in part, through inhibition of dihydroceramide desaturase, which provides a foundation for future exploration of this small-molecule inhibitor for the treatment of advanced disease.
Materials and Methods
Reagents
ABC294640 was provided by Dr. Charles D. Smith (Apogee Biotechnology Corporation). A stock was prepared at 200 mmol/L in water and ethanol (1:1). Rapamycin was purchased from Calbiochem/EMD Millipore. Docetaxel was obtained from LC Laboratories. C8-cyclopropenylceramide (C8-CPPC) was obtained from Matreya LLC. The CellTiterBlue assay kit was purchased from Promega. Antibodies were from the following sources: Beclin, BD Transduction laboratories #B3522; LC3 I/II, Novus Biologicals #NB100-2220; androgen receptor, Abcam, ab74272, c-Myc, Abcam, ab32072; DEGS (17), and GAPDH, Santa Cruz Biotechnologies. HRP-conjugated secondary antibodies were also purchased from Santa Cruz Biotechnology. Male C57BL6J mice (5–6 weeks old) were purchased from Jackson laboratories. 2,5-Dihydroxybenzoic acid (DHB) and trifluoroacetic acid (TFA) were obtained from Sigma-Aldrich. HPLC-grade methanol (MeOH), ethanol (EtOH), and water were obtained from Fisher Scientific. Indium tin oxide (ITO) slides were purchased from Bruker for MALDI-IMS experiments.
Cell lines and culture
TRAMP-C2 cells were obtained in 2004 from Dr. Norman Greenberg (Fred Hutchinson Cancer Research Center, Seattle, WA; ref. 18). C4-2 cells were obtained in 2010 from Urocor. Authentication was not performed for TRAMP-C2 or C4-2 but cells were not used for more than 30 passages after being thawed from stock. 22Rv1 cells modified with luciferase were obtained from Dr. Michael Henry (Holden Comprehensive Cancer Center, University of Iowa, Iowa City, IA) in 2006 (19). This cell line was authenticated by DNA profiling in 2015 (Biosynthesis DNA identity testing center, Louisville, TX). Cells were maintained in RPMI-1640 (22Rv1, C4-2) or DMEM containing 4.5 g/L glucose (TRAMP-C2) supplemented with 10% heat-inactivated FBS (Hyclone), 0.01% antibiotic-antimycotic solution, and 0.001% gentamycin. All cultures were maintained at 37°C in a 5% CO2 atmosphere and were periodically verified to be mycoplasma free. Mouse embryo fibroblasts (MEF) from wild-type and SPHK2-deficient mice were kindly provided by the Animal Core of the COBRE in Lipidomics and Pathobiology.
Viability assays
Cells were seeded into 96-well plates at densities between 2,000 and 5,000 cells/well. The next day varying concentrations of ABC294640 were added, and cancer cell viability was determined at 24 to 72 hours of treatment using the CellTiterBlue assay. Viability of MEFs was determined at 72 and 120 hours. Fluorescence was quantified using a FLUOstar Optima plate reader. All assays were performed in triplicate.
Flow cytometry
TRAMP-C2 cells (6 × 104) were plated into 6-well dishes and allowed to adhere overnight. Following treatment, nonadherent cells were collected and combined with the adherent fraction, which was removed from the dishes via trypsinization. Cells were washed and pelleted. For cell-cycle analysis, cells were fixed in 70% ethanol overnight, washed, treated with RNase, and then stained with propidium iodide. Staining was analyzed using a FACSCalibur flow cytometer (BD) and FlowJo software. To determine lysosomal acidification, 600 nmol/L LysoTracker solution was added to each well during the last hour of the assay. Cells were harvested as above but were not fixed. Instead, cell pellets were resuspended in PBS and lysotracker staining was analyzed using a Fortessa flow cytometer (BD) and FlowJo software.
Western blotting
Cells were plated and allowed to adhere overnight before treatment with ABC294640. After treatment, cells were collected as described for flow cytometry and Western blot analysis performed as previously described (20). Primary antibodies were used at a dilution of 1:1,000, except GAPDH, which was used at 1:6,000. The concentration of secondary antibody was 1:50,000 for anti-rabbit or 1:10,000 for anti-mouse. Band intensity was quantified using ImageJ software and values expressed relative to GAPDH.
Sphingolipid analysis
At 72 hours (TRAMP-C2) or 120 hours (MEFs) after treatment with ABC294640, cells were collected as described for flow cytometry and washed pellets were stored at −80°C until processing for sphingolipid analysis. High-performance liquid chromatography/mass spectrometry (LC/MS-MS) was performed at the MUSC Lipidomics Facility as previously described (21). Results are expressed as pmol sphingolipid/nmol lipid phosphate (22) or, for plasma, as pmol sphingolipid/mL.
Dihydroceramide desaturase enzyme assay
TRAMP-C2 cells (2 × 106) were seeded onto 150 mm plates and treated with increasing concentrations of ABC29640 for 72 hours. Cells were collected, pelleted, and assayed for dihydroceramide desaturase activity as previously described (23).
Animal studies
All animal experiments were performed with approval by the Institutional Animal Care and Use Committee at the Medical University of South Carolina (Charleston, SC). Mice were injected subcutaneously with 1 million TRAMP-C2 cells and 3 days later randomly assigned to vehicle or ABC294640 treatment groups. ABC294640 was diluted in 50% water/50% PEG200, and 50 mg/kg was injected intraperitoneally for 5 days per week. Vehicle-treated mice received water and PEG200 only. Once tumors developed, their size was measured using digital calipers and tumor volume was calculated using the formula: width × length × height × 0.52. After 4 weeks of treatment, mice were sacrificed 2 hours after final treatment and tumors as well as plasma were collected for further analyses. Tumor weight was recorded, and then the tissues were wrapped in aluminum, frozen in the liquid nitrogen vapor phase, and stored at −80°C until further analysis. Plasma was frozen at −80°C for subsequent analysis by LC/MS-MS.
MALDI imaging mass spectrometry
MALDI imaging mass spectrometry was performed as described previously (24). Briefly, mouse tumor tissue sections were sectioned at 10 μm and thaw mounted onto an indium tin oxide-coated slide. Slides were briefly dessicated before matrix addition using an ImagePrep (Bruker Daltonics). DHB matrix was added at a concentration of 30 mg/mL. Spectra were acquired across the entire tissue section on a Solarix dual source 7T FITCR mass spectrometer (Bruker Daltonics) to detect lipid species of interest (m/z 200—1,000) with a SmartBeam II laser operating at 1000 Hz, a laser spot size of 25 μm and a raster width of 100 μm. For analysis in positive ion mode, the instrument was calibrated using a mix of lipid standards. For each laser spot 400 spectra were collected. Following MS analysis, data were loaded into FlexImaging software focusing on the m/z mass range of 200 to 1,000 and reduced to an ion cyclotron resonance (ICR) reduction noise threshold of 1. All data were normalized using root means square (25) and intensities normalized to each other per figure as indicated in the figure legends. Images of differentially expressed lipids were generated using FlexImaging 4.0 software.
Statistical analysis
For statistical analysis of lysotracker studies, linear regression analysis was used to evaluate the association between treatment group and MFI. Main effects in the model were treatment groups and replicate (to adjust for experiment-level effects). MFI was log-transformed to adhere to regression model assumptions. P values were based on statistical significance of coefficients comparing each treatment group to the control condition using a t test. Regression diagnostics were used to ensure model assumptions were appropriate.
To adhere to assumptions of statistical models, tumor volumes and weight were transformed using a square-root transformation. Tumor volume over time was modeled using a linear regression model estimated using GEE (generalized estimating equations) with an exchangeable correlation structure to account for repeated measures over time per mouse. Main effects of treatment and time and an interaction between treatment and time were included as covariates in the model. Difference in tumor growth was evaluated using a Wald test for the interaction term in the model. Model estimates ± SEs were estimated and back-transformed to the original scale. Tumor weight was compared using a two-sample t test (on the square-root transformed tumor weight values). P values less than 0.05 were considered statistically significant.
Results
ABC294640 reduces the viability of prostate cancer cells through a nonapoptotic mechanism that involves increased lysososmal acidification
ABC294640 (Fig. 1A) is a first-in-class inhibitor of sphingosine kinase 2 (SPHK2) that has shown efficacy in preclinical models of cancer as well as inflammatory diseases (9, 11–14, 26–30). Although ABC294640 has entered the clinic, mechanistic understanding on how this inhibitor impacts castrate-resistant prostate cancer cells is limited. Because ABC294640 can influence immune function, we primarily focused our investigation on the murine castration-resistant TRAMP-C2 cell line, which can be grown in immune-competent hosts (18). As shown in Fig. 1B, treatment with ABC294640 results in a dose- and time-dependent inhibition of viability. Previous studies with ABC294640 indicated that the drug, depending on the cell line, reduces viability either via classical caspase-mediated apoptosis or autophagic cell death (6, 11, 12, 27). To evaluate apoptosis, TRAMP-C2 cells were treated with increasing concentrations of ABC294640 (up to the IC50) or 8 nmol/L docetaxel (IC50), a known inducer of apoptosis. DNA fragmentation was evaluated by flow cytometry. Docetaxel but not ABC294640 resulted in the appearance of a sub-G1 population (Fig. 1C).
The impact of ABC294640 and prostate cancer viability. A, the structure of ABC294640. B, viability of TRAMP-C2 cells. Data shown are the average ± standard deviation from two independent experiments performed in triplicate. C, assessment of DNA fragmentation in TRAMP-C2 cells 72 hours after drug exposure. Docetaxel was included as a positive control. Data represent the mean ± SD from three independent experiments. ***, P < 0.001. D, lysosomal acidification in TRAMP-C2 cells 72 hours after drug exposure. Rapamycin (4 nmol/L) was included as a positive control. Data shown are the average ± SD from three independent experiments performed in duplicate. E, Western blot analysis of TRAMP-C2 cells treated with the indicated concentration of ABC294640 for 72 hours. GAPDH serves as a loading control. Similar results were obtained in separate experiments. F, viability of 22Rv1 and C4-2 cells. Data shown are the average ± SD from two independent experiments performed in triplicate. G, Western blot analysis of 22Rv1 and C4-2 cells treated with ABC294640 for 72 hours. GAPDH serves as a loading control.
Autophagic cell death has also been reported in response to ABC294640. This response is associated with the formation of autophagosomes and is characterized by an increase in lysosomal acidification, induction of Beclin expression, and cleavage of LC-3. ABC294640 even at concentrations as low as 10 μmol/L significantly increased lysosomal acidification (Fig. 1D) but no induction of Beclin was detected by Western blot analysis (Fig. 1E). LC3-I levels increased with ABC294640 treatment but no cleavage to LC3-II was detected (Fig. 1E). These results suggested that the decrease in viability of TRAMP-C2 cells following treatment with ABC294640 at or below IC50 concentrations is neither mediated by classical apoptosis nor by autophagic cell death.
Treatment with ABC294640 reduces expression of c-Myc and the androgen receptor
Recently, ABC294640 was shown to decrease expression of c-Myc through increased proteosomal degradation (16). Upon treatment with ABC294640, a dose-dependent decrease in c-Myc expression was observed in TRAMP-C2 cells (Fig. 1E). To determine whether a decrease in c-Myc expression occurs upon treatment with ABC294640 in other castration-resistant prostate cancer cell lines, we extended our study to include 22Rv1 and LNCaP-derived C4-2 cells. The IC50 was comparable among the three prostate cancer cell lines: 22Rv1 (29 μmol/L), C4-2 (32 μmol/L), TRAMP-C2 (28 μmol/L). A time-dependent decrease in viability of 22RV1 and C4-2 cells is shown in Fig. 1F and similar to TRAMP-C2 cells, a decrease in c-Myc expression was observed (Fig. 1G). Microarray analysis has recently shown that c-Myc levels strongly correlate with androgen receptor (AR) expression in castration-resistant tumors (31). Because AR expression in TRAMP-C2 cells is relatively low, we investigated the impact of ABC294640 on expression of this protein in the two human castration-resistant prostate cancer cell lines. In both 22Rv1 and C4-2 cells, AR expression was reduced upon treatment with ABC294640 (Fig. 1G).
ABC294640 results in accumulation of dihydroceramides through inhibition of dihydroceramide desaturase activity
The dual sphingosine kinase inhibitor SKI-II increases dihydroceramide levels and inhibits the activity of dihydroceramide desaturase (DEGS), the enzyme responsible for converting dihydroceramide into ceramide (32). To investigate whether ABC294640 also inhibits DEGS, we initially performed sphingolipid analysis. Treatment with ABC294640 resulted in a dose-dependent decrease in sphingosine-1-phosphate (S1P; Fig. 2A) but did not significantly impact on overall levels of ceramide (Fig. 2B). However, at 30 μmol/L ABC294640, the distribution of ceramide species was altered. The major ceramide species in TRAMP-C2 cells were C24:1-Cer (>40%), C24:0-Cer (20%–30%), C16-Cer (10%–15%), C22:0-Cer (6%–7%), and C22:1-Cer (3%–4%). All other ceramide species (C14, C18, C20, C26) were 2% of the total or less and were collectively grouped as “minor” ceramides. As shown in Fig. 2C, at 30 μmol/L ABC294640 significantly increased in C16- and C22-ceramides, which occurred at the expense of C24-ceramides. In contrast to ceramide, total dihydroceramides increased by more than 3-fold with concentrations as low as 10 μmol/L ABC294640 (Fig. 2D). This increase was observed across all species of dihydroceramide and a significant shift in species distribution was only apparent at 30 μmol/L ABC294640 (Table 1). Accumulation of dihydroceramides suggested that DEGS is inhibited in cells treated with ABC294640 and we therefore measured the activity of the enzyme. Treatment with ABC294640 decreased DEGS activity but not expression (Fig. 2E and F). To test whether DEGS was also responsible for the decrease observed in c-Myc expression, we treated cells with C8-CPPC at concentrations known to reduce enzyme activity to 5% (17), but no appreciable effect on levels of c-Myc expression was observed (Fig. 2G). Finally, to address whether SPHK2 expression is required for the increase in dihydroceramides following treatment with ABC294640, we obtained MEFs from wild-type and SPHK2-deficient mice. Although MEFs were significantly more resistant to ABC294640 (IC50 > 60 μmol/L at 120 hours), drug treatment produced a similar increase in dihyroceramide irrespective of SPHK2 expression (Fig. 2H).
Sphingolipid and dihydroceramide desaturase activity analysis following ABC294640 treatment. TRAMP-C2 cells were analyzed for intracellular levels of S1P (A), total ceramides (B), distribution of ceramide species (C), total dihydroceramides (D), dihydroceramide desaturase activity (E), DEGS expression following treatment with ABC294640 for 72 hours (F), and expression of c-Myc following treatment with C8-CCPC (G). Similar results were observed in multiple experiments. H, MEF isolated from wild-type (wt) or SPHK2-deficient mice (SK2 KO) were treated with 70 μmol/L ABC294640 for 5 days and dihydroceramide was quantitated in triplicate cultures. *, P < 0.05; **, P < 0.005; ***, P < 0.0005.
Impact of ABC29460 on individual dihydroceramide species
Elevated dihydroceramide levels following treatment with ABC294640 are also detected in vivo
To evaluate the impact of ABC294640 on growth of TRAMP-C2 tumors in vivo, one million cells were injected subcutaneously and treatment with ABC294640 was initiated 3 days later. ABC294640 has been investigated in acute (7 day) and chronic (28 day) toxicity studies in mice and rats at concentrations up to 250 mg/kg (9). In mice, a dose of 50 mg/kg has frequently been used in other studies (6, 12, 26, 27, 33), but higher concentrations (75 or 100 mg/kg) have also been administered (28, 34). Qin and coworkers have shown that tumor growth is significantly reduced when the drug is administered by either oral gavage or intraperitoneal injection (27), suggesting that the drug has similar availability by either route. On the basis of these previous studies, we elected to treat mice with 50 mg/kg given via intraperitoneal injection 5 days/week for four cycles. The average body weight for vehicle-treated mice at the conclusion of the experiment was 24.42 ± 1.71 g, and ABC294640-treated mice weighed 23.69 ± 1.37 g. There was no significant difference in body weight between the groups at any other time points throughout the experiment, suggesting that treatment with ABC294640 did not adversely affect the animals. Measurable tumors began to appear on day 20 (treatment day 17) in both the vehicle and ABC294640-treated groups. Tumors in the ABC294640-treated group progressed more slowly compared with vehicle treated tumors (P = 0.03) and were significantly smaller in size and weight (P = 0.015) at the termination of the experiment than tumors in the control group (Fig. 3).
Impact of ABC294640 on tumor growth in vivo. TRAMP-C2 cells were implanted subcutaneously into syngeneic hosts, which were treated and monitored as described in Materials and Methods. Tumor measurements over time (A) and tumor size at experiment termination (B). Data shown are combined from two separate experiments representing 9 to 11 total animals per treatment group.
On the last day of the experiment, mice were treated with vehicle or ABC294640, sacrificed one hour later, and tumors and plasma were collected for further analyses. Tumors were analyzed by MALDI imaging mass spectrometry, which can be used to detect ABC294640 as well as sphingolipids. As shown in Fig. 4, detectable levels of ABC294640 were present within tumor tissues. Similar to treatment with ABC294640 in vitro, dihydroceramides were increased with ABC294640 treatment but no consistent changes in C16-ceramide were observed. Analysis of plasma indicated a trend toward increased total ceramide (Fig. 5A) that was primarily due to significant increases in C16- (Fig. 5B) and C24:1-ceramides (Fig. 5C). Plasma dhC16-ceramide remained unchanged (Fig. 5B).
MALDI-IMS analysis of ABC294640-treated tumors. Tumor tissue was processed as described in Materials and Methods and frozen sections analyzed for ABC294640 and ceramides. Phosphatidylcholine (PC) serves as an internal control.
Analysis of plasma ceramides. Plasma was collected at the termination of the experiment and analyzed for ceramides by LC/MS. Total ceramides (A), C16-ceramides (B), and very long-chain ceramides (C). Data shown are the average ± SD from three animals per group. *, P < 0.05.
Discussion
In this study, we evaluated the effects of the small-molecule inhibitor ABC294640 on castration-resistant prostate cancer cells. The IC50 of ABC294640 (at 72 hours) in prostate cancer cells was between 28 μmol/L and 32 μmol/L, which is consistent with an IC50 of 28 μmol/L recently reported for castration-resistant PC3 cells (35). The IC50 in nonmalignant MEF was not reached upon exposure to 70 μmol/L ABC294640 for 72 hours, suggesting that normal cells are more resistant to the effects of this inhibitor. In contrast with previous studies, we were unable to confirm that ABC294640 reduced viability through induction of apoptosis or autophagic cell death (Fig. 1; refs. 6, 9, 11, 12, 16, 27). In agreement with our results, Gestaut and colleagues showed that caspase-9 was not activated following treatment with ABC294640 although Ki-67–staining decreased (35). Similarly, decreased viability in response to ABC294640 in acute lymphoblastic leukemia did not result in morphologic features associated with apoptosis nor did the pan-caspase inhibitor ZVAD prevent the decrease in viability (36). The same study reported that LC3 cleavage occurred at 50 and 70 μmol/L ABC294640, but the autophagy inhibitor 3MA failed to inhibit cell death. Taken together, these findings suggest that ABC294640 has the capacity to reduce proliferation without inducing apoptosis or autophagic cell death.
Potential mechanisms by which proliferation may be inhibited in response to treatment with ABC294640 include diminished expression of the AR or c-Myc, both of which are important therapeutic targets in prostate cancer. The AR remains active even after castration resistance develops (37) and our results show that treatment with ABC294640 decreased AR expression in both 22Rv1 and C4-2 prostate cancer cells (Fig. 1F). The dual sphingosine kinase inhibitor SKI-II also diminishes AR expression independent of apoptotic or autophagy signaling (38). Because exogenous S1P was unable to restore AR expression, a role for intracellular S1P was implicated (38). Nuclear localization of both SPHK2 and the AR suggests the possibility that the SPHK2-derived pool of S1P could be important for its expression. However, at 10 μmol/L the SPHK2-specific inhibitor (R)-FTY720 methyl ether (ROME) failed to decrease AR expression and decreases at 50 μmol/L were thought to occur as a consequence of an off-target effect (38). Future studies using RNAi approaches will be needed to mechanistically investigate the relationship between sphingosine kinases and AR expression. Although the AR may be a target affected by ABC294640, we also observed a decrease in the expression of c-Myc, which functions as an important growth effector downstream of the AR (39) and is amplified in up to 72% of castrate-resistant prostate cancers (40). Recently, SPHK2 was shown to regulate expression of the oncogene c-Myc in acute lymphoblastic leukemia (36). ABC294640 reduced expression of c-Myc in all castration-resistant prostate cancer cell lines tested (Fig. 2).
One of the most striking results was that ABC294640 increased levels of dihydroceramides by approximately 3- to 5-fold (Fig. 2D, Table 1). A significant and persistent increase in dhC16-ceramide was observed during the initial characterization of ABC294640 but other dihydroceramide species were not investigated (9). Our results show that all species of dihydroceramides were significantly elevated upon treatment with ABC294640 and that a 3-fold increase in dihydroceramide was achieved with as little as 10 μmol/L ABC294640 (Table 1). The increase in dihydroceramide inversely correlated with a decrease in dihydroceramide synthase (DEGS) activity (Fig. 2E). ABC294640 treatment did not reduce the expression of DEGS (Fig. 2F), supporting the idea that the drug results in enzymatic inhibition. The dual sphingosine kinase inhibitor SKI-II also diminishes DEGS activity, an effect not observed with the sphingosine kinase 1-selective inhibitor PF-543 (32). Initial testing of ABC294640 indicated that the small-molecule inhibitor was selective for SPHK2 over SPHK1 and did not impair the activity of various other kinases, including protein kinase B/C, cyclin-dependent kinases, and MAPKs (9). However, our results show that the capacity of ABC294640 to increase dihydroceramides was retained in SPHK2-deficient cells (Fig. 2H), suggesting that ABC294640 can impact DEGS independent of SPHK2 expression. This “off-target” effect would not have been discovered during initial characterization of ABC294640, since screening focused on kinases and not enzymes involved in sphingolipid metabolism. Cingolani and coworkers suggested that some of the effects attributed to decreases in S1P following inhibition of SPHK2 by SKI-II could stem from increases in dihydroceramides (32). Our results with ABC294640 support their conclusion. For example, the increase in lysosomal acidification or growth inhibition following treatment with ABC294640 could be due to inhibition of SPHK2, DEGS, or both. In contrast, since inhibition of DEGS by C8-CPPC did not reduce expression of c-Myc (Fig. 2G), we hypothesize that the decrease in c-Myc is a consequence of reduced SPHK2 activity. Further studies will be needed to dissect which consequences following treatment with ABC294640 are due to inhibition of SPHK2 and which are due to DEGS inhibition. Whether ABC294640 directly impacts DEGS or acts on upstream enzymes, as has been suggested for the dual sphingosine kinase inhibitor SKI-II (32), is currently under investigation.
In vivo, ABC294640 reduced the rate of tumor growth and using MALDI-IMS, we were able to detect ABC294640 within tumor tissues indicating that the drug did indeed reach its target. C16-ceramide levels in tumors varied and no consistent increase was observed in vivo. Because increases in C16-ceramide were observed only at 30 μmol/L in vitro, our results suggest that intratumoral drug concentrations were likely below the IC50 concentration of 28 μmol/L. In contrast, levels of intra-tumoral dihydroceramide increased, suggesting that concentrations reaching TRAMP-C2 tumors in vivo are sufficient to decrease DEGS activity. Dihydroceramide has been associated with growth inhibition in vitro (17) and in vivo (41) and may have contributed to reduced growth rates and tumor size in ABC294640-treated animals.
Plasma levels of dhC16-ceramide were unchanged in ABC294640-treated mice but C16-and C24:1-ceramides increased significantly. It is unclear whether increased intracellular dhC16-ceramide served as a source of plasma C16-ceramide. Because ABC294640 impacts tumors as well as the immune system, further studies are needed to determine the source of increased C16- and C24:1-ceramides in the plasma. It might be useful to incorporate monitoring of plasma ceramide levels into clinical trials to determine whether increases in plasma C16- and C24:1-ceramides are useful biomarkers of ABC294640 activity.
Several studies have investigated combinations of sphingosine kinase inhibition and other agents. ABC294640 treatment of ovarian cancer cells, which failed to induce apoptosis as a single agent, led to activation of caspase-9 when used in combination with paclitaxel (42). Recently, ABC294640 was reported to enhance the antitumor effects of TRAIL in non–small lung cancer (43). Similar results have been observed in colon cancer cells (CVJ, unpublished observation). In glioblastoma, combination of the DNA alkylating agent temozolomide with a sphingosine kinase inhibitor, both used at subtoxic doses, led to potentiation of cell death (44). Interestingly, neither the pro-death sphingolipids ceramide nor sphingosine but their dihydro-forms appeared to be crucial for cytotoxicity. These studies suggest that combination studies with ABC294640 in prostate cancer are warranted.
In conclusion, we have shown that ABC294640 reduces the growth of castration-resistant prostate cancer cells. On the basis of in vitro analysis, ABC294640 neither induced apoptosis nor autophagic cell death but rather appeared to slow proliferation of tumor cells. This growth-inhibitory effect may stem from both inhibition of SPHK2 and DEGS, resulting in decreased intracellular S1P and increased dihydroceramides, respectively. Future studies using ABC294640 in combination with other treatment regimens will determine the clinical potential of ABC294640 for the treatment of advanced prostate cancer.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: C.D. Smith, C. Voelkel-Johnson
Development of methodology: M. Rahmaniyan, E.E. Jones, P. Lu, R.R. Drake, C.D. Smith, C. Voelkel-Johnson
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Venant, M. Rahmaniyan, E.E. Jones, P. Lu, J.M. Kraveka, C. Voelkel-Johnson
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Venant, M. Rahmaniyan, E. Garrett-Mayer, R.R. Drake, J.M. Kraveka, C.D. Smith, C. Voelkel-Johnson
Writing, review, and/or revision of the manuscript: H. Venant, M. Rahmaniyan, M.B. Lilly, E. Garrett-Mayer, J.M. Kraveka, C.D. Smith, C. Voelkel-Johnson
Study supervision: M.B. Lilly, C. Voelkel-Johnson
Grant Support
The Hollings Cancer Center's Cancer Center Support Grant P30 CA138313 in part supported this pilot project (awarded to C. Voelkel-Johnson), the Lipidomics Facility, the Proteomic Shared Resource and the Flow Cytometry & Cell Sorting Shared Resource at the Medical University of South Carolina. Additional support was from National Center for Research Resources and the Office of the Director of the NIH through Grant Number C06 RR015455 and R01 CA135087 (to R.R. Drake), CA154778 (to C. Voelkel-Johnson) and by grants from the Rally Foundation for Childhood Cancer Research (to J.M. Krakeva), Chase After A Cure Foundation (to J.M. Kraveka), Hugs for Harper Endowment (to J.M. Krakeva), and Hyundai Hope on Wheels (to J.M. Krakeva and M. Rahmaniyan).
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.
Acknowledgments
The authors thank Dr. Marion Cooley from the COBRE in Lipidomics and Pathobiology Animal Core at MUSC for supplying the mouse embryo fibroblasts. This core is supported by the NIH grant 5P30GM103339-04 (NIGMS).
- Received April 6, 2015.
- Revision received August 5, 2015.
- Accepted October 6, 2015.
- ©2015 American Association for Cancer Research.