Overexpression of Sphingosine-1-Phosphate Lyase ... - Eukaryotic Cell

EUKARYOTIC CELL, June 2004, p. 795–805 1535-9778/04/$08.00⫹0 DOI: 10.1128/EC.3.3.795–805.2004 Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Vol. 3, No. 3

Overexpression of Sphingosine-1-Phosphate Lyase or Inhibition of Sphingosine Kinase in Dictyostelium discoideum Results in a Selective Increase in Sensitivity to Platinum-Based Chemotherapy Drugs Junxia Min, Andrew L. Stegner, Hannah Alexander, and Stephen Alexander* Division of Biological Sciences, University of Missouri, Columbia, Missouri 65211-7400 Received 26 January 2004/Accepted 4 March 2004

The efficacy of the chemotherapy drug cisplatin is often limited due to resistance of the tumors to the drug, and increasing the potency of cisplatin without increasing its concentration could prove beneficial. A previously characterized Dictyostelium discoideum mutant with increased resistance to cisplatin was defective in the gene encoding sphingosine-1-phosphate (S-1-P) lyase, which catalyzes the breakdown of S-1-P, an important regulatory molecule in cell function and development and in the regulation of cell fate. We hypothesized that the increased resistance to cisplatin was due to an elevation of S-1-P and predicted that lowering levels of S-1-P should increase sensitivity to the drug. We generated three strains that stably overexpress different levels of the S-1-P lyase. The overexpressor strains have reduced growth rate and, confirming the hypothesis, showed an expression-dependent increase in sensitivity to cisplatin. Consistently, treating the cells with D-erythro-N,N,dimethylsphingosine, a known inhibitor of sphingosine kinase, increased the sensitivity of mutant and parent cells to cisplatin, while addition of exogenous S-1-P or 8-Br-cyclic AMP made the cells more resistant to cisplatin. The increased sensitivity of the overexpressors to cisplatin was also observed with the cisplatin analog carboplatin. In contrast, the response to doxorubicin, 5-flurouracil, or etoposide was unaffected, indicating that the involvement of the sphingolipid metabolic pathway in modulating the response to cisplatin is not part of a global genotoxic stress response. The augmented sensitivity to cisplatin appears to be the result of an intracellular signaling function of S-1-P, because D. discoideum does not appear to have endothelial differentiation growth (EDG/S1P) receptors. Overall, the results show that modulation of the sphingolipid pathway at multiple points can result in increased sensitivity to cisplatin and has the potential for increasing the clinical usefulness of this important drug. lecular genetic manipulations allowed us to isolate by insertional mutagenesis a number of mutants— with disruptions in single genes—that have increased resistance to cisplatin. The disrupted genes were identified by sequencing the DNA flanking the insertions. None of the identified genes had been associated previously with cisplatin resistance, and as such they represented potential new targets for improving therapy (20). One of the genes identified in the above study encodes the enzyme sphingosine-1-phosphate (S-1-P) lyase (sglA), which catalyzes the last step in the sphingomyelin degradation pathway, the conversion of S-1-P to phosphoethanolamine and hexadecanal (Fig. 1) (45). In addition to its resistance to cisplatin, the S-1-P lyase null mutant (sglA⌬) exhibited dramatic phenotypic changes in its growth and development. These included the inability of aggregating cells to form anterior Factin-filled pseudopods, the inability to form multicellular slugs that can phototax, and a block in late development resulting in a decrease in spore production (19). Overall, the multiple phenotypes indicated that S-1-P is a central regulatory molecule in D. discoideum development, similar to the much-studied regulatory molecule cyclic AMP (cAMP) (39). Subsequent studies have reported that the S-1-P lyase plays important roles in cell function and development in other systems as well, including yeast (7), Caenorhabditis elegans (27), Drosophila melanogaster (8), and mouse cells (16). Evidence now exists showing that S-1-P functions both as an intracellular second

Chemotherapy is frequently used to treat cancer. However, the efficacy of treatment is often limited because some tumors are intrinsically resistant to anti-tumor drugs, while in others resistance is selected for during the course of therapy. One important example is the chemotherapy drug cisplatin [cisdiamminedichloroplatinum (II)], which is widely used in the treatment of small-cell and non-small-cell lung cancer, nonHodgkin’s lymphoma, testicular cancer, ovarian cancer, head and neck cancer, esophageal cancer, and bladder cancer, among others (33). Various mechanisms of resistance have been proposed, including reduced drug concentration in the cell, drug inactivation, increased DNA repair, or failure to turn on cell death pathways (40). However, despite numerous studies, our understanding of resistance to this widely used drug remains poor, and the signaling pathways that activate the variety of proposed mechanisms of resistance are, for the most part, unknown. To some degree this is the result of the difficulty of using animal cell lines for identifying specific genetic loci associated with drug resistance. We have used the cellular slime mold Dictyostelium discoideum to identify specific genes that are involved in a cell’s response to cisplatin. The amenability of this organism to mo* Corresponding author. Mailing address: Division of Biological Sciences, 303 Tucker Hall, University of Missouri, Columbia, MO 652117400. Phone: (573) 882-6670. Fax: (573) 882-0123. E-mail: alexanderst @missouri.edu. 795

796

MIN ET AL.

EUKARYOT. CELL

that manipulations of these sphingolipid biosynthesis and biodegradation pathways could augment treatment with cisplatin. Additionally, the sglAOE mutants have an impaired growth phenotype, which results in slower proliferation in liquid cultures and smaller plaque size when grown on bacteria, confirming the pleiotropic role of S-1-P in the regulation of cellular processes. MATERIALS AND METHODS

FIG. 1. Schematic presentation of the last steps in the sphingomyelin degradation pathway. The S-1-P lyase and sphingosine kinase enzymes shown in shaded boxes are the primary focus of this study. The compounds that were used in this study (DMS and cAMP) and their effects on sphingosine kinase are shown. sglAOE is the S-1-P lyase overexpressor strain, and sglA⌬ is the S-1-P lyase null strain.

messenger and as an extracellular ligand for the family of G protein-coupled endothelial cell growth (EDG) receptors, and a substantial body of evidence has implicated S-1-P in mammalian cell movement and tumor cell metastasis through Gprotein-coupled mechanisms (29, 41). Earlier work in animal systems has led to the rheostat model, which suggests that it is the relative level of ceramide to S-1-P (24, 42), or sphingosine to S-1-P (36), that determines a cell’s decision to proliferate or die. Thus, cells with increased levels of ceramide or sphingosine follow a pathway of cell death, while increasing the level of S-1-P counteracts this and results in cell differentiation, proliferation, and inhibition of cell death. Accordingly, we hypothesized that the increased resistance of the sglA⌬ strain to cisplatin was due to an increased level of intracellular S-1-P. If so, altering the S-1-P levels, either genetically or pharmacologically, would result in corresponding changes in cisplatin sensitivity. This approach would potentially identify new molecular targets for increasing the sensitivity of tumors to cisplatin. In this paper we describe the generation and the phenotypic analysis of mutants that stably overexpress the SglA protein to different levels. The overexpressor mutants exhibit the predicted increase in sensitivity to cisplatin and the related drug carboplatin but not to other drugs tested. Treating the parent cells with the sphingosine kinase inhibitor D-erythro-N,N,-dimethylsphingosine (DMS) mimicked the sglA overexpression (sglAOE) phenotype and resulted in increased sensitivity to cisplatin as well. The increased sensitivity was reversed by the addition of 8-Br-cAMP or by the addition of S-1-P to the mutant strains. These studies support our idea that sphingolipids are involved in the cell’s specific response to cisplatin and

Strains and culture conditions. Ax3-ORF⫹, the parent strain for the overexpressing S-1-P lyase gene, and the transformation vector pDXA3C were gifts from D. Manstein, National Institute for Medical Health, London, United Kingdom (18, 26). The 7.7-kb transformation construct carrying the myc-tagged sglA gene (pDXA3C/sglA) is pJM5. The S-1-P lyase overexpressors [sglA-myc]neor-1, [sglA-myc]neor-2, and [sglA-myc]neor-3 (strains SA601 to SA603) are referred to in the text as sglAOE-1, -2, and -3. Strain SA555 contains a homologous deletion of the sglA gene, is blasticidin resistant, and is referred to in the text as sglA⌬ (previously named cis2B [19, 20]). Strains were stored either frozen in liquid nitrogen in a mixture of 5% dimethyl sulfoxide (DMSO) in horse serum or as desiccated spores on silica gel at 4°C. Fresh cultures were started monthly from stocks. Cells were grown in HL5 medium (43). Clonal isolation of strains and some growth rate experiments were done by growing cells on SM agar in association with Klebsiella aerogenes as a food source (43). All cell growth was done at 22°C. Construction of sglA overexpression vector. A full-length cDNA for the S-1-P lyase gene (sglA; accession no. AY283052) was derived by reverse transcriptionPCR (Gibco Superscript Preamplification System; Gibco-BRL, Gaithersburg, Md.) by using total RNA from D. discoideum strain Ax4. First-strand synthesis was done using the gene-specific reverse primer 5⬘ CTCCAATGCATCGTAA GTTGATTGAGAAGG 3⬘, and the ensuing PCRs were carried out using the forward primer 5⬘ CTCGAGCTCATGGATAAAGCAAATGAT 3⬘, the reverse primer above, and the high-fidelity polymerase Easy-A Hi-Fi PCR cloning enzyme (Stratagene, La Jolla, Calif.) (underlined regions are gene-specific sequences). NsiI or SacI sites were introduced for the purpose of cloning. The amplified DNA was ligated directly into the corresponding sites of pDXA3C to generate expression vector pJM5. The construct includes the 5⬘ ATG of the sglA gene and is missing the 3⬘ stop codon. Expression is driven by the actin 15 promoter, and the protein product is myc tagged at the C terminus. Manstein et al. (26) reported high levels of expression with this vector. The entire sglA fragment was sequenced by using sglA gene-specific primers. Prior to transformation the plasmid was purified by using a QIAGEN (Valencia, Calif.) Maxi plasmid preparation kit. Generation of overexpressor strains. Logarithmically growing D. discoideum Ax3-ORF⫹ cells (5 ⫻ 106 cells/ml) were mixed with 15 ␮g (10 ␮l) of vector DNA and were immediately electroporated as described previously (17). Cells were brought up to 40 ml in DD broth (26), plated in four 100-mm petri dishes, and incubated at 22°C for 24 h. Transformants were selected by the addition of 20 ␮g of G418/ml at 22°C. Cells were fed every 4 to 5 days with DD broth-G418, and small colonies of G418-resistant cells began to appear after 10 to 12 days. Control transformations were Ax3-ORF⫹ cells transformed with pDXA3C vector. Cells from 20 individual sglA transformant colonies and from control transformant colonies were transferred to 24-well plates for expansion. These were incubated for 4 days and then were inoculated both onto 24-well plates containing SM agar and bacteria in order to examine the phenotype of the cells during growth and development on the agar and also onto a replica 24-well plate containing DD broth-G418. Ten single clones, which appeared to be growing more slowly than the control transformants, were chosen for Western analysis for the presence of the c-myc-tagged SglA protein. To this end, cells from the corresponding 24-well plates were plated clonally onto 100-mm-diameter SM agar plates with bacteria and single clones were isolated. Western blots. Putative mutant cells were plated on SM agar in association with K. aerogenes and were allowed to grow to confluence (until the plates were clear of bacteria). The cells were scraped off the plate, washed twice in SS buffer (0.6 g of NaCl/liter, 0.75 g of KCl/liter, 0.4 g of CaCl2/liter), pelleted, and kept frozen at ⫺80°C. Cell pellets were lysed in 1 ml of lysis buffer (50 mM Tris-HCl [pH 8.0], 5 mM EDTA, 0.5% Triton X-100) including protease inhibitor [1:100 dilution of 100⫻ protease inhibitor cocktail, which contained 20 mM 4-(2aminoethyl)benzenesulfonyl fluoride, 100 ␮g of pepstatin A/ml, 10 ␮g of leupeptin/ml] on ice. Protein concentration was determined by bicinchoninic acid protein assay (Pierce, Rockford, Ill.). Fifty micrograms of protein was separated

VOL. 3, 2004

S-1-P LYASE REGULATES CISPLATIN SENSITIVITY

797

FIG. 2. Expression of the SglA protein in D. discoideum. Cells of each of the sglA-myc transformants were harvested, washed, and lysed in denaturing buffer, and the protein concentration was determined. Fifty micrograms of total protein per lane was separated by sodium dodecyl sulfate–10% polyacrylamide gel electrophoresis. (A) Total protein, stained with colloidal Coomassie blue. Note that the samples have equal amounts of protein. Vector1 and Vector2 are vector control transformants. Ax3-ORF⫹ is the untransformed parental cells. MW, molecular size standards. (B) Proteins were transferred to nitrocellulose membranes and were probed with anti-c-myc antibodies. (C) The exposed film shown in panel B was scanned and quantitated. The image shows relative units of expression, using sglAOE-3 as a reference (1).

on sodium dodecyl sulfate–10% polyacrylamide gel electrophoresis and was blotted onto nitrocellulose (OPTITRAN; Schleicher & Schuell, Keene, N.H.) for 4 h at 63 V. The blots were developed with monoclonal anti-c-myc antibodies, clone 9E10 (Sigma-Aldrich, St. Louis, Mo.), and horseradish peroxidase-conjugated immunopure goat anti-mouse secondary antibody (Pierce). The reaction was developed by using BM chemiluminescence blotting substrate (Roche, Indianapolis, Ind.). The films were scanned, and the relative density of the bands was determined by using Metamorph version 4.6. Growth rate measurements. Growth on a solid surface was measured by plating cells at low density (about 40 cells per plate) on 100-mm-diameter SM agar plates in association with bacteria. As cells consume the bacteria they form plaques. The plates were scanned daily on a flatbed scanner starting at day 3, when the plaques were first visible. The diameters of 10 random plaques per strain were measured, and areas were calculated. For measuring growth rate in liquid cultures, cells were inoculated in HL5 medium at 5 ⫻ 104 cells/ml, and duplicate samples were counted daily in a hemocytometer. Nuclei counting. Logarithmically growing cells (3 ⫻ 104 cells) were washed in LPS buffer (43), deposited on coverslips, and allowed to settle and adhere. The cells were fixed in 3.7% formaldehyde for 3.5 min, washed thoroughly with LPS, permeabilized with 0.5% NP-40 in LPS, and stained with 20 ␮g of 4⬘,6-diamidino-2-phenylindole (DAPI)/ml. Random fields of cells were examined on a Zeiss IM microscope with a 100⫻ Neofluor lens, and the number of nuclei in each cell was counted. The results are the averages of 300 cells for each strain. Drug sensitivity. All drugs used in this study were purchased from SigmaAldrich (St. Louis, Mo.). Stock solutions were constructed as follows: cisplatin, 1 mg/ml (3.3 mM) in Pt buffer (3 mM NaCl, 1 mM NaPO4 [pH 6.5]); carboplatin, 1 mg/ml (2.69 mM) in Pt buffer; doxorubicin, 2 mg/ml (3.45 mM) in Pt buffer; etoposide, 30 mg/ml (50.97 mM) in DMSO; 5-flurouracil (5-FU), 25 mg/ml (192.2 mM) in DMSO. For each drug we ran a preliminary experiment on Ax4 cells to determine the optimal drug concentration and pH which resulted in a level of cell death where increased sensitivity or resistance could be detected (data not shown). In all cases the drugs had a stronger effect when the experiments were done in buffer rather than in growth medium. However, because sudden starvation of the cells causes stress by itself, all the experiments were done in HL5 growth medium. Duplicate 10-ml cultures of parent or mutant strains were then treated with drugs at the concentrations and times indicated in the figures while being shaken at 200 rpm at 22°C and were assayed for viability by using a rapid plaque viability assay in 24-well plates (1). Survival was calculated as the percentage of the untreated culture. All viability determinations were done in duplicate, and the results were averaged. This assay allows the quantitation of viability over 4 orders of magnitude and is therefore extremely sensitive. Error was calculated by using the statistical tool in Microsoft Excel, and standard error bars are presented in all the graphs. P values were calculated by using Student’s t test in Microsoft Excel. Cotreatments of cells with cisplatin and sphingolipid metabolism-related

compound. 8-Br-cAMP was dissolved to 40 mg/ml (93 mM) in HL5. DMS was dissolved to 2.5 mg/ml (7.63 mM) in DMSO, and S-1-P was dissolved to 1 mg/ml (2.64 mM) in DMSO. All experiments were done in duplicate in 2-ml cultures in 20-ml glass scintillation vials. Each experiment consisted of an untreated culture, a culture treated with cisplatin alone, a culture treated with the tested chemical alone, and a culture treated with both. For the combination treatments, the chemicals being tested were added to growing cultures 1 h prior to the addition of cisplatin. The rest of the experiment was carried out as described in the above section on drug sensitivity. Survival was calculated as a percentage of the untreated culture. Fold change in resistance or sensitivity was calculated as the ratio of the survival after combination treatment to the expected survival if each of the compounds acted independently, e.g., survival after DMS and cisplatin together divided by (survival after DMS alone times survival after cisplatin alone). Standard error and P values were calculated as described above.

RESULTS Stable overexpression of S-1-P lyase. The S-1-P lyase null mutant was originally obtained in our random insertional mutagenesis selection for cisplatin resistance. In addition to its increased resistance to the drug, the mutant displayed aberrant development. The aim of this study was to construct mutants that overexpressed the S-1-P lyase to test the hypothesis that altering the levels of S-1-P in the cells would result in increased sensitivity to cisplatin. In addition, we predicted that sglA overexpression would provide additional clues to the role of this enzyme in cell function. Three mutant strains that overexpress a myc-tagged fusion protein of the S-1-P lyase were isolated. These strains produced the expected 58-kDa fusion protein to different levels and were named sglAOE-1, sglAOE-2, and sglAOE-3. Figure 2B shows the levels of c-myc expression in the three strains as detected by reactivity with anti-myc antibody. The level of expression in strain sglAOE-1 and sglAOE-2 was approximately 15 and 7 times that of strain sglAOE-3 (Fig. 2C). Figure 2A shows the Coomassie blue-stained proteins in the samples used for the Western analysis and confirms that the level of protein in each sample was equal. Moreover, the increasing amounts of the recombinant SglA-myc fusion protein can be observed in

798

MIN ET AL.

EUKARYOT. CELL

FIG. 3. Growth rate of the sglAOE strains. The three sglAOE strains, the vector control transformation strains, and the parent strain were plated on SM agar plates in association with K. aerogenes. Plates were scanned daily, and the diameter of the plaques was measured. (A) Photographs of the plates at day 4. (B) Plaque size. The results are presented as the area of the plaque over 3 days, and each point is the average of 10 randomly selected plaques. Inset is a visual representation of the results at day 3, when the plaques were still very small. (C) The three sglAOE strains, the vector control transformation strain, and the parent strain were inoculated at a density of 105 cells/ml in 20 ml of HL5 medium and were grown with shaking at 200 rpm at 22°C. Cultures were counted daily. When the cultures reached approximately 5 ⫻ 106 cells/ml they were diluted down to 5 ⫻ 104 cells/ml so that the culture could be followed through three consecutive passages. The last passage was allowed to go through stationary phase. (D) Cells of the three sglAOE strains, the vector control transformation strain, and the parent strain were harvested from HL5 medium at a density of 3 ⫻ 105 (mid-log phase) and washed, and 100 ␮l was allowed to settle on coverslips, where they were fixed with 3.7% formaldehyde. The cells were stained with 20 ␮g of DAPI/ml. The number of nuclei per cell was counted in 300 cells from each strain on a Zeiss IM microscope with a Chroma Technology Corp. lucifer yellow filter. Results are presented as the percentages of the total number of cells.

the samples from all three overexpressing strains. The variations in SglA-myc expression are stable and are presumably due to differences in copy number. Overexpression of S-1-P lyase results in altered growth regulation. The S-1-P lyase-overexpressing strains have slower growth rates than the parent strain. This is reflected both in the size of the plaque for bacterially grown cells and in a slower growth rate in liquid. Figure 3A and B depicts the change in plaque size over 3 days for each strain and shows that each overexpressing strain has a different growth rate. The decrease in plaque size reflects the level of overexpression of the SglA protein. The inset in Fig. 3B magnifies the differences at day 3, when the plaques were very small. Figure 3C shows the growth rate of the strains when growing axenically in HL5 medium. The sglAOE strains also grow more slowly in liquid medium, although the difference becomes less pronounced at densities over 5 ⫻ 106 cell/ml. This growth

behavior was observed in three consecutive passages. Again, the degree of growth inhibition is parallel to the level of SglA overexpression, although the differences are not as pronounced as when the cells are growing on agar. Importantly, the cell density at which the cells enter stationary phase is altered in the overexpressing strains, with the higher levels of overexpression causing a lower stationary phase density. SglA overexpression also results in a delayed decrease in cell number in late stationary phase. The change in growth phenotype is accompanied by an increasing number of multinucleate cells, as is shown in Fig. 3D. S-1-P lyase overexpression increases sensitivity to the chemotherapeutic agents cisplatin and carboplatin but not to other drugs. The original sglA⌬ mutant had increased resistance to cisplatin, and we hypothesized that the lack of S-1-P lyase resulted in the buildup of S-1-P, a lipid associated with growth promotion and inhibition of cell death. Based on this hypoth-

VOL. 3, 2004

FIG. 4. sglAOE strains are sensitive to cisplatin. Cultures of 106 cells/ml in HL5 medium were treated with increasing concentrations of cisplatin (0, 75, 150, and 300 ␮M) for 4 h, serially diluted, and plated for viability on SM agar in 24-well plates with K. aerogenes as the food source. Viability is expressed as the percentage of surviving cells relative to an untreated culture. sglA⌬ is the SglA null strain. Because the sglA⌬ and sglAOE strains had different parents, we established that the Ax3-ORF⫹ and Ax4 strains had identical sensitivities to cisplatin (data not shown).

esis, it is predicted that cells overexpressing S-1-P lyase should be more sensitive to cisplatin. Cells were treated with three concentrations of cisplatin for 4 h and were scored for viability. The results shown in Fig. 4 confirm our hypothesis and show that the SglA-overexpressing cells are indeed more sensitive to cisplatin than the parent. The increase in sensitivity parallels the level of expression of the myc-tagged SglA protein in each strain, with slgAOE-1 being 10-fold more sensitive, followed by sglAOE-2 (3.3-fold) and sglAOE-3 (1.6-fold). Confirming the results of previous studies (20), sglA⌬ cells assayed in the same experiment are more resistant to cisplatin than Ax4 cells. Cross-resistance to drugs is an important aspect of chemo-


799

therapy and can considerably limit the options for treatment. Thus, it was determined whether the sglAOE and sglA⌬ mutants have increased sensitivity or resistance to other drugs in addition to cisplatin. To this end we tested the response of the two mutants and the Ax4 parent in parallel to a 24-h treatment with cisplatin, carboplatin (a cisplatin analog widely used in chemotherapy), and three other non-platinum drugs (Fig. 5). The data show that the same altered sensitivity to cisplatin was observed in the case of carboplatin. Carboplatin is less toxic to the Ax4 cells (65% survival at 300 ␮M), but sglAOE-1 is more sensitive to carboplatin and sglA⌬ is more resistant to the drug, similar to the results obtained with cisplatin. In contrast, neither mutant showed altered sensitivity or resistance to doxorubicin, 5-FU, or etoposide. Therefore, altering SglA expression did not affect the cells’ responses to these drugs. Inhibition of the sphingosine kinase synergistically increases sensitivity to cisplatin. Deleting or overexpressing the sglA gene resulted in increased resistance or increased sensitivity to cisplatin, respectively. These results suggested that inhibiting sphingosine kinase should mimic the sglAOE phenotype and would result in increased sensitivity to cisplatin. To this end we tested the effect of the sphingosine kinase inhibitor DMS on cells of the sglAOE-1, sglA⌬, and parental strains, and the results are shown in Fig. 6. (i) Parental wild-type cells. High levels of DMS are toxic to D. discoideum cells. Treatment with concentrations of 20 and 50 ␮M resulted in complete cell death (data not shown), underscoring the importance of sphingosine kinase for normal cell growth. To test the effect of DMS on cisplatin sensitivity, lower doses were chosen, which resulted in a lower level of cell death. Figure 6A and B shows that DMS kills the parental wild-type cells in a dose- and time-dependent manner. Concentrations of 2.5, 5, and 10 ␮M DMS result in 97, 89, and 28% survival after 5 h and 113, 74, and 11% after 24 h. Treatment of parallel cultures with 150 ␮M cisplatin alone reduced cell viability to 74% percent after 5 h and 36% after 24 h. Cotreatment of cells with both DMS and cisplatin resulted in a synergistic increase in drug sensitivity. For example, at 5 ␮M DMS

FIG. 5. Overexpression of sglA renders the cells more sensitive to cisplatin and carboplatin but not to other drugs. Ten milliliters of cultures of 106 cells/ml in HL5 medium were treated with the indicated drugs for 5 h, serially diluted, and plated for viability measurements as described for Fig. 4. Survival was calculated as a percentage of the untreated culture. (A) 300 ␮M cisplatin; (B) 300 ␮M carboplatin; (C) 450 ␮M doxorubicin; (D) 400 ␮M 5-FU; (E) 300 ␮M etoposide. Strain names are the same as those shown in Fig. 4. P values (by Student’s t test) for cisplatin and carboplatin were ⱕ0.05 and ⬍0.001, respectively.

800

MIN ET AL.

EUKARYOT. CELL

FIG. 6. Pretreatment with DMS increased the cell response to cisplatin. Growing cultures of sglAOE-1, sglA⌬, and the parent strain at 106 in HL5 medium were treated with 0, 2.5, 5, and 10 ␮M DMS for 1 h prior to adding 150 ␮M cisplatin. The cultures were assayed for viability at 5 and 24 h as described in the legend to Fig. 4. (A and B) parent strain; (C and D) sglAOE-1; (E and F) sglA⌬. Closed circles, 150 ␮M cisplatin with increasing concentrations of DMS. Open circles, increasing concentrations of DMS alone. (G) Photograph of the plaques of the viable cells at day 3 after plating. Note that increasing concentrations of DMS result in smaller plaques.

VOL. 3, 2004


801

FIG. 7. Pretreatment with S-1-P increases resistance of cells to cisplatin. Growing cultures at 106 cells/ml in HL5 medium were treated with 0, 10, 20, and 50 ␮M S-1-P 1 h prior to the addition of 150 ␮M cisplatin. Cells were sampled at 5 and 24 h and were assayed for viability as described in the legend to Fig. 4. (A and B) Parental strain at 5 and 24 h, respectively. (C and D) sglAOE-1 at 5 and 24 h, respectively. Closed circles, 150 ␮M cisplatin with increasing concentrations of S-1-P. Open circles, increasing concentration of S-1-P alone.

survival is reduced 2.8-fold after 5 h and 8.9-fold after 24 h over what would be predicted if the two drugs were working independently. Clearly, inhibiting sphingosine kinase increases sensitivity to cisplatin in the parental cells, as was predicted. (ii) SglA overexpressor cells. The effect of DMS on the sglAOE-1 cells is shown in Fig. 6C and D. Again, the sglAOE-1 cells were killed by DMS in a dose- and time-dependent manner. As shown above, these cells are initially more sensitive to cisplatin. In contrast to the results with the parental cells, there is no synergistic effect of cotreatment, and the killing with both DMS and cisplatin was not greater than the expected combined killing of the two drugs alone. This is presumably because the level of S-1-P in these cells is already very low due to the increase in the S-1-P lyase enzyme, such that inhibition of the kinase has no additional effect. (iii) SglA⌬ cells. SglA⌬ mutant cells are more resistant to cisplatin than the parent cells, and 5 h of treatment with 150 ␮M cisplatin resulted in virtually no cell killing. Similar to the results described for the parent cells, cotreatment with DMS and cisplatin resulted in a synergistic reduction in viability of 6.5-fold at 5 h and 4.2-fold at 24 h when treated with 5 ␮M DMS. Cells treated with DMS produced small plaques, and the size of the plaques decreased with increasing DMS concentrations (Fig. 6G). This is in agreement with the smaller plaque size that was seen with the sglAOE strains and is consistent with the idea that reducing the level of S-1-P results in a decrease in growth rate. Exogenous S-1-P increases resistance to cisplatin. Based on the results with the sglAOE and null mutants, it was predicted that adding S-1-P to cells would increase resistance to cispla-

tin—essentially mimicking the sglA⌬ mutant. It was previously shown that the developmental timing phenotype of the sglA⌬ mutant could be mimicked by adding exogenous S-1-P to wildtype cells (19). The results showed that S-1-P does indeed make parental and sglAOE-1 cells more resistant to cisplatin, as was predicted. Figure 7A and B depicts the results with the parent strain. At 5 h with 150 ␮M cisplatin there was little killing, and therefore adding S-1-P had no obvious effect. However, at 24 h there was approximately 50% killing with cisplatin alone, and S-1-P at even the lowest dose of 2 ␮M reversed this to almost the untreated level, i.e., made the cells more resistant. In the case of the initially more-cisplatin-sensitive sglAOE-1 cells (Fig. 7C and D), S-1-P also increased resistance. The sensitivity to cisplatin was never fully reversed in the sglAOE-1 mutant. This is presumably because of the high levels of S-1-P lyase. 8-Br-cAMP increases resistance to cisplatin. cAMP has been reported to be an activator of sphingosine kinase (25), and this suggested that increasing levels of cAMP could increase resistance to cisplatin in a fashion similar to that of adding S-1-P to the cells. Therefore, the membrane-permeable analog 8-Br-cAMP was used to increase intracellular concentrations of cAMP. 8-Br-cAMP has been shown to be effective at mimicking the effects of cAMP in D. discoideum development, including the late stages of spore development, where 20 mM levels were used for maximum effect (15). The results depicted in Fig. 8 show that 8-Br-cAMP does indeed make both parental and sglAOE-1 cells more resistant to cisplatin. We initially found that high levels of 8-Br-cAMP (10 to 20 mM) were toxic to mitotically dividing cells and therefore chose to test lower concentrations. The effect of 8-Br-cAMP on the

802

MIN ET AL.

EUKARYOT. CELL

FIG. 8. Pretreatment with 8-Br-cAMP increases resistance of cell sensitivity to cisplatin. Growing cultures of 106 cell/ml in HL5 medium were treated with 0, 2, and 5 mM 8-Br-cAMP for 1 h prior to the addition of 150 ␮M cisplatin. Viability and strain names are as described in the legend to Fig. 4. Closed circles, 150 ␮M cisplatin with increasing concentrations of 8-Br-cAMP. Open circles, increasing concentrations of 8-Br-cAMP alone.

parent strain (Fig. 8A) was mild, and the effect on the sglAOE strain (Fig. 8B) was more pronounced, where survival increased from 30 to 70%. DISCUSSION Genetic studies using insertional mutagenesis revealed new and unexpected genes and pathways that are involved in modulating a cell’s sensitivity to the chemotherapeutic drug cisplatin (20). These genes included those encoding sphingosine-1-P lyase, cAMP phosphodiesterase (regA), golvesin (a Golgi protein), a P2Y purine receptor homolog, and a CAAX prenyl protease homolog. Similar mutational strategies with yeast have also identified cAMP phosphodiesterase (3) as well as a novel copper transporter, CTR1 (9, 21). The S-1-P lyase null mutant was chosen for additional study, because the defective enzyme in this strain represents the last step in a pathway that has been previously shown to play an important role in cell signaling, cell proliferation, and cell death via controlling the levels of ceramide, sphingosine, and S-1-P (36, 42). The D. discoideum genome encodes homologs to the genes encoding the enzymes of this pathway found in mammals (28), including the sphingomylinase, ceramidase, sphingosine kinase, S-1-P phosphatase, and S-1-P lyase enzymes. The resistance of the S-1-P null mutant suggested that this pathway could be manipulated either pharmacologically or genetically to increase drug sensitivity and that this information could ultimately be used to increase the efficacy of treat-

ment of human tumors that are resistant to cisplatin. The studies described here confirm the idea that this pathway represents new targets for modulating cisplatin action. The increased resistance of the sglA⌬ mutant to cisplatin suggested that this was due to an increase in S-1-P. Thus, we predicted that either increasing the level of the SglA protein or reducing the activity of the sphingosine kinases would make cells more sensitive to cisplatin. Indeed, overexpression of the S-1-P lyase as well as inhibition of the sphingosine kinases has validated these predictions and strongly suggests that it is the level of S-1-P that ultimately modulates drug sensitivity. The isolation of three stable sglAOE strains that produce different levels of the SglA protein allowed us to show that the level of expression parallels the level of increased cisplatin sensitivity. The sglAOE-1 strain with the highest level of SglA overexpression was up to 10-fold more sensitive to cisplatin. Consistent with this result is a recent report that overexpression of S-1-P lyase in HEK-293 cells resulted in increased stress-induced apoptosis (38). The sglAOE strains also exhibit an expression-dependent decrease of growth rate and stationary phase density and an increase in the average number of nuclei per cell. The increase in the number of nuclei per cell may be related to a decrease in the rate of cytokinesis, which would account for the slower growth rate. Whether the decrease in growth rate directly relates to an increase in drug sensitivity remains to be determined. A similar decrease in cell growth rate has been observed in HEK-293 cells overexpressing the S-1-P lyase gene (38), and yeast lacking the sphingosine kinase gene are delayed from entering S phase (12). The level of protein in the three sglAOE strains that was detected by Western blotting is not precisely linear with the increases in drug sensitivity or growth rate, but it is possible that the signaling systems affected become saturated so that additional expression has no effect. Treating the cells with the sphingosine kinase inhibitor DMS essentially mimicked the phenotype of the sglAOE strains and showed that modulating the activity of the sphingosine kinases has the predicted effect of also making cells more sensitive to cisplatin. Experiments are in progress with sphingosine kinase null mutants to further test this idea. The data on the direct addition of S-1-P further strengthens the idea that the pathway of S-1-P synthesis and degradation is intimately involved in modulating the cellular response to cisplatin. Exogenous addition of S-1-P renders the cells more resistant to cisplatin in a manner similar to that of the increase in resistance we observed in the sglA⌬ mutant. It was previously shown that exogenous addition of S-1-P to developing Dictyostelium cells can mimic the aberrant developmental phenotype of the sglA⌬ mutant (19). S-1-P most likely functions intracellularly after uptake by pinocytosis, because there are no obvious homologs of the G-protein-coupled EDG receptors (S1P receptors) in D. discoideum. This also supports the idea that the changes in cisplatin sensitivity observed in the sglA⌬ and sglAOE mutants are due to S-1-P acting as an intracellular second messenger, a function demonstrated in mammalian cells by overexpression of the sphingosine kinase gene in mouse cells lacking the receptors (32). cAMP has been reported to activate sphingosine kinase (25), and the addition of 8-Br-cAMP to cells had the predicted effect of increasing resistance to cisplatin. It is important to note that

VOL. 3, 2004


803

FIG. 9. Schematic model of the interactions that address the specificity of the resistance of cells to cisplatin with respect to sphingolipid signaling.

in a previous paper it was shown that deletion of the regA cAMP phosphodiesterase resulted in increased cisplatin resistance (20), and this is in agreement with the discovery that the yeast PDE2 null cells are cisplatin resistant. Thus, cAMP may also be signaling through protein kinase A to modulate cisplatin resistance. All the previous data show that increasing the S-1-P lyase or decreasing the sphingosine kinase results in altered sensitivity to cisplatin. Consistent with this, sphingosine kinase mRNA has been shown to be overexpressed in a variety of solid tumors (6). This suggests that cotreatment with cisplatin and compounds that lower the levels of S-1-P in cells could result in increased antitumor activity of cisplatin. Cisplatin is generally used in therapy at the maximum allowable dose, and therefore its dosage cannot be increased to treat resistant tumors. Thus, DMS or other sphingosine kinase inhibitors could be useful in combination with cisplatin to increase its efficacy in antitumor therapy—either with tumors such as lymphomas, malignant melanoma, and prostate cancer that do not respond to treatment or in cases such as ovarian tumors that initially respond and then become more resistant (33). Additional sphingosine kinase inhibitors have been reported recently (6), and one inhibitor, phenoxodiol, has been shown to promote apoptosis

in ovarian cancer cells (13). Other enzymes of the pathway of sphingosine metabolism conceivably could also be targeted for the purpose of increasing cisplatin sensitivity. Cross-resistance to drugs is an important aspect of chemotherapy, and it can limit the options for treatment. The specificity of this pathway for the sensitivity to cisplatin is intriguing. The sglAOE strains also showed increased sensitivity for the related platinum drug carboplatin but, interestingly, did not show increased sensitivity for other drugs that we tested. Carboplatin is a closely related derivative of cisplatin, which undergoes aquation and has the same platination nucleotide specificity as cisplatin (37). Therefore, it was expected that changing sensitivity to cisplatin would change sensitivity to carboplatin as well, although its toxicity level to wild-type cells is less than that of cisplatin. Indeed, the sglAOE cells are more sensitive to carboplatin and the sglA⌬ cells are more resistant. It will be interesting to see if this relationship holds for the many cisplatin analogs, which have a wide range of toxicities and structures (31). In contrast to the platinum drugs, the sglAOE and sglA⌬ cells did not show altered sensitivity to doxorubicin (DNA intercalator and topoisomerase II inhibitor [2, 5]), 5-FU (inhibits thymidylate synthase and incorporation of fluordeoxyuridine triphosphates into DNA [22]), and eto-

804

MIN ET AL.

poside (topoisomerase II inhibitor [35]). Thus, although all three drugs affect DNA metabolism and structure, the resulting cell death does not appear to be modulated by the level of S-1-P lyase. Consistent with this, it was previously shown that the sglA⌬ mutant did not show significant cross-resistance to UV light, the alkylating agent methylmethane sulfonate, or H2O2 (20). Two major issues arise from these studies: (i) how do sphingolipids affect the cellular response to cisplatin, and (ii) how can these interactions account for the specificity for cisplatin. These questions are particularly important for translating the results from a model system, like D. discoideum, to human tumor cells. Figure 9 presents a model that aims to address these questions and makes predictions for further experiments. Cisplatin exerts its cytotoxic effect by causing damage to DNA. The DNA damage is recognized by a host of enzymes— some that will target the lesions for repair and others that will target the cell to die. Damaged DNA-dependent protein kinases activate a cascade of signaling that ultimately activates the mitogen-activated protein (MAP) kinase family of proteins, including the MAP kinase ERK family as well as the stress-activated protein kinases of the JNK and p38 families. It has been suggested that the relative balance between p38-JNK and ERK determines the response to cisplatin (40). The lipids ceramide and S-1-P have also been shown to regulate these MAP kinases. Ceramide activates JNK and p38 and inhibits ERK, while S-1-P activates ERK and downregulates the stress activated enzymes JNK and p38, such that the balance between the MAP kinases reflects the lipids rheostat (4). Thus, in a cell treated with a lethal dose of cisplatin the balance between ceramide and S-1-P usually favors the activation of the p38-JNK enzymes and results in cell death. Based on this model, interfering with the balance of these lipids either by overexpression of the S-1-P lyase or by inhibiting the sphingosine kinase (both of which should result in reduced levels of S-1-P) would increase cell death. In contrast, deleting the S-1-P lyase, activating sphingosine kinase, or adding exogenous S-1-P should result in the downregulation of the JNKp38-related proteins and in the upregulation of ERK and should result in decreased sensitivity to the drug (34). The specificity of the sphingolipid pathway for regulating sensitivity to cisplatin can be attributed to several elements of this model. (i) The cytotoxic effect of cisplatin, but not doxorubicin or taxol, is mediated primarily through p38 (23). (ii) protein kinase C, phosphatidylinositol 3-kinase, and Akt-PKB are all upregulated by S-1-P (4) and have been linked to cisplatin resistance (11, 40, 46). (iii) Glutathione S-transferase is upregulated by ERK (44). Glutathione S-transferase, in turn, acts as an inhibitor of JNK in addition to interacting with glutathione to directly inactivate cisplatin (10, 14). Overall, the hypothesis emerging from these studies is that even though the sphingolipid pathway is involved in the response to a number of stresses, the sensitivity of cells to different chemotherapeutic agents may be independently controlled. It is possible that some of these behaviors will end up being cell type and/or drug specific, but together these findings are of considerable significance for the planning of therapeutic strategies when a particular drug is found to be ineffective. We are presently translating the findings of this study to mammalian cells.

EUKARYOT. CELL ACKNOWLEDGMENTS This work was supported by NIH grants GM53929 and CA95872. We thank Jason Edward for assistance during his undergraduate internship (College of Arts and Science Undergraduate Research Mentor Program). We also thank the Dictyostelium cDNA sequencing project in Tsukuba, Japan (30), for clones throughout the course of this work and the Baylor/Sanger/Jena multinational Dictyostelium DNA sequencing consortium for sequence data (http://www.dictybase .org). REFERENCES 1. Alexander, H., A. N. Vomund, and S. Alexander. 2003. Viability assay for Dictyostelium for use in drug studies. BioTechniques 35:464–470. 2. Arcamone, F. 1981. Doxorubicin anticancer antibiotics, medicinal chemistry, a series of monographs, vol. 17. Academic Press, New York, N.Y. 3. Burger, H., A. Capello, P. W. Schenk, G. Stoter, J. Brouwer, and K. Nooter. 2000. A genome-wide screening in Saccharomyces cerevisiae for genes that confer resistance to the anticancer agent cisplatin. Biochem. Biophys. Res. Commun. 269:767–774. 4. Cuvillier, O., G. Pirianov, B. Kleuser, P. G. Vanek, O. A. Coso, S. Gutkind, and S. Spiegel. 1996. Suppression of ceramide-mediated programmed cell death by sphingosine-1-phosphate. Nature 381:800–803. 5. Doroshow, J. H. 1996. Anthracyclines and anthracenediones, p. 500–537. In B. A. Chabner and D. L. M. Longo (ed.), Cancer chemotherapy and biotherapy: principles and practice, 3rd ed. Lippincott Williams and Wilkins, Philadelphia, Pa. 6. French, K. J., R. S. Schrecengost, B. D. Lee, Y. Zhuang, S. N. Smith, J. L. Eberly, J. K. Yun, and C. D. Smith. 2003. Discovery and evaluation of inhibitors of human sphingosine kinase. Cancer Res. 63:5962–5969. 7. Gottlieb, D., W. Heideman, and J. D. Saba. 1999. The DPL1 gene is involved in mediating the response to nutrient deprivation in Saccharomyces cerevisiae. Mol. Cell. Biol. Res. Commun. 1:66–71. 8. Herr, D. R., H. Fyrst, V. Phan, K. Heinecke, R. Georges, G. L. Harris, and J. D. Saba. 2003. Sply regulation of sphingolipid signaling molecules is essential for Drosophila development. Development 130:2443–2453. 9. Ishida, S., J. Lee, D. J. Thiele, and I. Herskowitz. 2002. Uptake of the anticancer drug cisplatin mediated by the copper transporter Ctr1 in yeast and mammals. Proc. Natl. Acad. Sci. USA 99:14298–14302. 10. Ishikawa, T. 1992. The ATP-dependent glutathione S-conjugate export pump. Trends Biochem. Sci. 17:463–468. 11. Isonishi, S., K. Ohkawa, T. Tanaka, and S. B. Howell. 2000. Depletion of protein kinase C (PKC) by 12-O-tetradecanoylphorbol-13-acetate (TPA) enhances platinum drug sensitivity in human ovarian carcinoma cells. Br. J. Cancer 82:34–38. 12. Jenkins, G. M., and Y. A. Hannun. 2001. Role for de novo sphingoid base biosynthesis in the heat-induced transient cell cycle arrest of Saccharomyces cerevisiae. J. Biol. Chem. 276:8574–8581. 13. Kamsteeg, M., T. Rutherford, E. Sapi, B. Hanczaruk, S. Shahabi, M. Flick, D. Brown, and G. Mor. 2003. Phenoxodiol—an isoflavone analog—induces apoptosis in chemoresistant ovarian cancer cells. Oncogene 22:2611–2620. 14. Kartalou, M., and J. M. Essigmann. 2001. Mechanisms of resistance to cisplatin. Mut. Res. Fund. Mol. Mech. Mutagen. 478:23–43. 15. Kay, R. R. 1989. Evidence that elevated intracellular cyclic AMP triggers spore maturation in Dictyostelium. Development 105:753–759. 16. Kihara, A., M. Ikeda, Y. Kariya, E. Y. Lee, Y. M. Lee, and Y. Igarashi. 2003. Sphingosine-1-phosphate lyase is involved in the differentiation of F9 embryonal carcinoma cells to primitive endoderm. J. Biol. Chem. 278:14578– 14585. 17. Kuspa, A., and W. F. Loomis. 1992. Tagging developmental genes in Dictyostelium by restriction enzyme-mediated integration of plasmid DNA. Proc. Natl. Acad. Sci. USA 89:8803–8807. 18. Leiting, B., I. J. Lindner, and A. A. Noegel. 1990. The extrachromosomal replication of Dictyostelium plasmid Ddp2 requires a cis-acting element and a plasmid-encoded trans-acting factor. Mol. Cell. Biol. 10:3727–3736. 19. Li, G., C. Foote, S. Alexander, and H. Alexander. 2001. Sphingosine-1phosphate lyase has a central role in the development of Dictyostelium discoideum. Development 128:3473–3483. 20. Li, G. C., H. Alexander, N. Schneider, and S. Alexander. 2000. Molecular basis for resistance to the anticancer drug cisplatin in Dictyostelium. Microbiology 146:2219–2227. 21. Lin, X., T. Okuda, A. Holzer, and S. B. Howell. 2002. The copper transporter CTR1 regulates cisplatin uptake in Saccharomyces cerevisiae. Mol. Pharmacol. 62:1154–1159. 22. Longley, D. B., D. P. Harkin, and P. G. Johnston. 2003. 5-Fluorouracil: mechanisms of action and clinical strategies. Nat. Rev. Cancer 3:330–338. 23. Losa, J. H., C. P. Cobo, J. G. Viniegra, V. J. Sanchez-Arevalo Lobo, S. Ramon y Cajal, and R. Sanchez-Prieto. 2003. Role of the p38 MAPK pathway in cisplatin-based therapy. Oncogene 22:3998–4006. 24. Maceyka, M., S. G. Payne, S. Milstien, and S. Spiegel. 2002. Sphingosine

VOL. 3, 2004

25. 26. 27. 28. 29. 30.

31.

32.

33. 34. 35.

kinase, sphingosine-1-phosphate, and apoptosis. Biochim. Biophys. Acta 1585:193–201. Machwate, M., S. B. Rodan, G. A. Rodan, and S. I. Harada. 1998. Sphingosine kinase mediates cyclic AMP suppression of apoptosis in rat periosteal cells. Mol. Pharmacol. 54:70–77. Manstein, D. J., H. P. Schuster, P. Morandini, and D. M. Hunt. 1995. Cloning vectors for the production of proteins in Dictyostelium discoideum. Gene 162:129–134. Mendel, J., K. Heinecke, H. Fyrst, and J. D. Saba. 2003. Sphingosine phosphate lyase expression is essential for normal development in Caenorhabditis elegans. J. Biol. Chem. 278:22341–22349. Merrill, A. H., Jr. 2002. De novo sphingolipid biosynthesis: a necessary, but dangerous, pathway. J. Biol. Chem. 277:25843–25846. Morii, T., and L. Weissbach. 2003. Sphingosine 1-phosphate and cell migration: resistance to angiogenesis inhibitors. Biochem. Biophys. Res. Commun. 310:884–888. Morio, T., H. Urushihara, T. Saito, Y. Ugawa, H. Mizuno, M. Yoshida, R. Yoshino, B. N. Mitra, M. Pi, T. Sato, K. Takemoto, H. Yasukawa, J. Williams, M. Maeda, I. Takeuchi, H. Ochiai, and Y. Tanaka. 1998. The Dictyostelium developmental cDNA project: generation and analysis of expressed sequence tags from the first-finger stage of development. DNA Res. 5:335– 340. Najajreh, Y., J. M. Perez, C. Navarro-Ranninger, and D. Gibson. 2002. Novel soluble cationic trans-diaminedichloroplatinum(II) complexes that are active against cisplatin resistant ovarian cancer cell lines. J. Med. Chem. 45:5189–5195. Olivera, A., H. M. Rosenfeldt, M. Bektas, F. Wang, I. Ishii, J. Chun, S. Milstien, and S. Spiegel. 2003. Sphingosine kinase type 1 induces G12/13mediated stress fiber formation, yet promotes growth and survival independent of G protein-coupled receptors. J. Biol. Chem. 278:46452–46460. Perry, M. C. (ed.). 1997. The chemotherapy source book, 2nd ed. Williams & Wilkins, Baltimore, Md. Pitson, S. M., P. A. B. Moretti, J. R. Zebol, H. E. Lynn, P. Xia, M. A. Vadas, and B. W. Wattenberg. 2003. Activation of sphingosine kinase 1 by ERK1/ 2-mediated phosphorylation. EMBO J. 22:5491–5500. Pommier, Y. G., F. Goldwasser, and D. Strumberg. 2001. Topoisomerase II


36.

37. 38.

39. 40. 41. 42. 43. 44. 45. 46.

805

inhibitors: epipodophyllotoxins, acridines, ellipticines, and bisdioxopeperazines, p. 538–578. In B. A. Chabner and D. L. Longo (ed.), Cancer chemotherapy and biotherapy: principles and practice, 3rd ed. Lippincott Williams and Wilkins, Philadelphia, Pa. Pyne, S. 2002. Cellular signaling by sphingosine and sphingosine 1-phosphate. Their opposing roles in apoptosis, p. 245–268. In A. Quinn and A. Kagan (ed.), Phospholipid metabolism in apoptosis, vol. 36. Kluwer Academic/Plenum, New York, N.Y. Reed, E. 2001. Cisplatin and analogs, p. 447–465. In B. A. Chabner and D. L. Longo (ed.), Cancer chemotherapy and biotherapy: principles and practice, 3rd ed. Lippincott Williams and Wilkins, Philadephia, Pa. Reiss, U., B. Oskouian, J. Zhou, V. Gupta, P. Sooriyakumaran, S. Kelly, E. Wang, A. H. Merrill, Jr., and J. D. Saba. 2003. Sphingosine-phosphate lyase enhances stress-induced ceramide generation and apoptosis. J. Biol. Chem. 279:1281–1290. Saran, S., M. E. Meima, E. Alvarez-Curto, K. E. Weening, D. E. Rozen, and P. Schaap. 2002. cAMP signaling in Dictyostelium. Complexity of cAMP synthesis, degradation and detection. J. Muscle Res. Cell Motil. 23:793–802. Siddik, Z. H. 2003. Cisplatin: mode of cytotoxic action and molecular basis of resistance. Oncogene 22:7265–7279. Spiegel, S., D. English, and S. Milstien. 2002. Sphingosine 1-phosphate signaling: providing cells with a sense of direction. Trends Cell Biol. 12:236– 242. Spiegel, S., and S. Milstien. 2002. Sphingosine 1-phosphate, a key cell signaling molecule. J. Biol. Chem. 277:25851–25854. Sussman, M. 1987. Cultivation and synchronous morphogenesis of Dictyostelium under controlled experimental conditions. Methods Cell Biol. 28:9– 29. Townsend, D. M., and K. D. Tew. 2003. The role of glutathione-S-transferase in anti-cancer drug resistance. Oncogene 22:7369–7375. van Veldhoven, P. P., and G. P. Mannaerts. 1993. Sphingosine-phosphate lyase. Adv. Lipid Res. 26:69–98. Yuan, Z. Q., R. I. Feldman, G. E. Sussman, D. Coppola, S. V. Nicosia, and J. Q. Cheng. 2003. AKT2 inhibition of cisplatin-induced JNK/p38 and Bax activation by phosphorylation of ASK1: implication of AKT2 in chemoresistance. J. Biol. Chem. 278:23432–23440.