Levels of SCS7/FA2H-Mediated Fatty Acid 2-Hydroxylation Determine the Sensitivity of Cells to Antitumor PM02734

Research Article

Ana B. Herrero,1 Alma M. Astudillo,2 Marı´a A. Balboa,2 Carmen Cuevas,3 Jesu´s Balsinde,2 and Sergio Moreno1

1Instituto de Biologıá Molecular y Celular del Cańcer, Consejo Superior de Investigaciones Cientı´ficas/Universidad de Salamanca, Campus Miguel de Unamuno, Salamanca, Spain; 2Instituto de Biologıá y Gene´ticaMolecular, Consejo Superior de Investigaciones Cientı´ficas, and Centro de Investigacioń Biome´dicaen Red de Diabetes y Enfermedades Metabo´licas Asociadas, Valladolid, Spain; and 3PharmaMar S.A., Research and Development, Madrid, Spain

Abstract into mice. Based on these observations, and in view of its PM02734 is a novel synthetic antitumor drug that is currently acceptable nonclinical toxicity profile, PM02734 has been selected in phase I clinical trials. To gain some insight into its mode of for clinical development (2). action, we used the yeast Saccharomyces cerevisiae as a model Although PM02734 has entered phase I clinical trials with a positive therapeutic index in advanced pretreated solid tumors, system. Treatment of S. cerevisiae with PM02734 rapidly very little is known about its mechanism of action. To gain insight induced necrosis-like cell death, as also found for mammalian into the in vivo mechanism of the action of PM02734, we used the cells treated with its close analogue kahalalide F. We have yeast Saccharomyces cerevisiae as a model organism. We found that screened the complete set of 4,848 viable S. cerevisiae haploid the compound induces rapid necrosis-like cell death in yeast. We deletion mutants to identify genes involved in sensitivity or also provide evidence that PM02734 affects vesicle trafficking and resistance to PM02734. Forty-five percent of the 40 most mitochondrial functions because mutants affected in these sensitive strains identified had a role in intracellular vesicle processes were highly sensitive to the drug. Furthermore, we have trafficking, indicating that the drug severely affects this identified the scs7 mutant as the most resistant strain to the process. A mutant strain lacking the sphingolipid fatty acyl cytotoxic action of PM02734. Moreover, increased levels of Scs7 2-hydroxylase Scs7 was found to be the most resistant to turned the cells hypersensitive to the compound. These results PM02734, whereas overexpression of Scs7 rendered the cells have been validated in human cells. Scs7 and its human hypersensitive to PM02734. To validate these findings in homologue, FA2H, are hydroxylases that introduce a hydroxyl human cells, we did small interfering RNA experiments and group at the position 2of fatty acids. 2-Hydroxy fatty acids are also overexpressed the Scs7 human homologue FA2H in found almost exclusively as N-acyl chains within the ceramide human cancer cell lines. As in yeast, FA2H silencing turned moiety of a variety of sphingolipids. The functional role of fatty acid the cells resistant to the drug, whereas FA2H overexpression 2-hydroxylation has not been clearly established, but it seems led to an increased sensitivity. Moreover, exogenous addition to affect membrane conformation (3). We have found that of the 2-hydroxylated fatty acid 2-hydroxy palmitic acid to 2-hydroxylated fatty acid–containing ceramides are involved in different human cell lines increased their sensitivity to the the mechanism of action of PM02734. The implications of all these cytotoxic compound. Taken together, these results suggest that findings in the treatment of tumors with PM02734 and patient the cell membrane and, in particular, 2-hydroxy fatty acid– selection are discussed. containing ceramides are important for PM02734 activity. These findings may have important implications in the development of PM02734 because tumor cells with high Materials and Methods FA2H expression are expected to be particularly sensitive to Chemicals. PM02734 was obtained from PharmaMar, prepared as a 10 this drug. [Cancer Res 2008;68(23):9779–87] mg/mL stock solution in DMSO/ethanol (1:1), and kept at À20jC. The drug concentrations used for the experiments were in the range of those obtained after clinical administration of the compound. Syringomycin E was Introduction also supplied by PharmaMar and was dissolved in DMSO at 5 mg/mL. MMS PM02734 is a synthetic cyclic depsipeptide related to natural was supplied by Fluka and SDS was obtained from Sigma. kahalalides, especially to kahalalide F (KF), an antitumor Yeast strains and media. The 4,848 S. cerevisiae haploid deletion compound isolated from the Hawaiian mollusk Elysia rufescens mutants in nonessential genes were obtained from Invitrogen. Yeast strains j (1, 2). PM02734 has shown in vitro activity against a broad were grown at 30 C on yeast extract-peptone-dextrose (YPD) with geneticin spectrum of tumor types: breast, colon, pancreas, lung, prostate, (G418, Duchefa) or without geneticin when growing the wild-type strain. Electron microscopy. Cells (2 Â 108) were fixed in 3% glutaraldehyde etc. In addition, PM02734 shows statistically significant in vivo for 1 h at room temperature and processed as described by Wonisch et al. antitumor activity in several human cancer cell lines xenografted (4). Finally, cells were resuspended in 500 AL of 100% Spurr resin and transferred to capsules where resin polymerized for 20 h at 60jC. Ultrathin sections were stained with lead citrate and uranyl acetate and examined in a transmission electron microscope. Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). Identification of yeast strains hypersensitive or resistant to Requests for reprints: Sergio Moreno, Instituto de Biologı´a Molecular y Celular del PM02734. Deletion strains were grown in YPD with G418 at 200 Ag/mL. Ca´ncer, Consejo Superior de Investigaciones Cientificas/Universidad de Salamanca, Cells were resuspended and inoculated in YPD with 250 Ag/mL PM02734. Campus Miguel de Unamuno, 37007 Salamanca, Spain. Phone: 34-923294810; Fax: 34- Resistant strains were identified as those with a higher absorbance after 923294795; E-mail: [email protected]. I2008 American Association for Cancer Research. 24 h of growth. Sensitive strains were unable to grow after 48 h. Strains doi:10.1158/0008-5472.CAN-08-1981 hypersensitive or resistant to PM02734 were isolated to new plates to www.aacrjournals.org 9779 Cancer Res 2008; 68: (23). December 1, 2008

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2008 American Association for Cancer Research. Cancer Research determine their drug tolerance phenotype more accurately using different 1 h at room temperature and developed using enhanced chemilumines- concentrations of the drug. cence Western blotting detection reagents (Amersham). For caspase-3 and Overexpression of SCS7 in S. cerevisiae. Full-length SCS7 gene was poly(ADP-ribose) polymerase (PARP) detection, 30 Ag of protein from placed under control of the GAL1 promoter using a PCR-based method (5). HCT116 or U937 cells were used. Anti-caspase-3 antibody (Cell Signaling A PCR fragment containing 49 nucleotides 142bp upstream of the SCS7 Technology) and anti-PARP antibody (Sigma) were used at 1:1,000 and start codon together with the KanMX6 module and 50 nucleotides ending 1:2,000, respectively. just downstream the start codon was amplified using the pair of Cell viability assays. Determination of cell viability was carried out oligonucleotides scs7-F4 (5¶-AACGGGGTTTTGCGCGGATTTACT- following the recommendations of the CellTiter 96 Aqueous Non- GACGTGTGACATCTTTTCGCAGTTAAGAATTCGAGCTCGTTTAAAC-3¶) Radioactive Cell Proliferation Assay (Promega Biotech Iberica). Cells seeded and scs7-R2(5 ¶-GTACCGTCTTTTTTGAAAACAGTTCCAAAGTCTTGGAAG- in 24-well plates at the indicated concentration in the overexpression or TATTAGTCGACATTTTGAGATCCGGGTTTT-3¶). Plasmid pFA6a-KanMX6- siRNA experiments were treated with different concentrations of cytotoxic PGAL was used as template. S. cerevisiae BY4741 wild-type strain was compounds. To determine viability, 100 AL of a combined 3-(4,5- transformed with the PCR fragment and selected in medium containing dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tet- G418. Correct integration in the yeast chromosome was verified by PCR. razolium salt/phenazine methosulfate solution were added to each well Cell cultures. HCT116 cells were grown in McCoy’s 5A medium containing 500 AL of culture medium. After 3 h of incubation, the formation (Invitrogen). U937 and HeLa cells were grown in RPMI 1640. Media were of the formazan product was assayed by measuring absorbance at 490 nm supplemented with 10% calf bovine serum, 100 units/mL penicillin, and 100 in a 24-well plate reader (Tecan Ultra Evolution). Ag/mL streptomycin. Cells were maintained under humid conditions of 5% RNA extraction and Northern blots. Total RNA from cells was obtained

CO2 at 37jC. following the recommendations of the RNeasy Mini kit (Qiagen). Five Overexpression experiments. HCT116 cells were transiently transfected micrograms of each sample were separated on a agarose formaldehyde gel with pcDNA3 or pcDNA3-FA2H (6) using Lipofectamine 2000 (Invitrogen). and Northern blotting was carried out following the protocol of the Briefly, plasmid/Lipofectamine 2000 complexes (0.5 Ag/1.5 AL per well) were ExpressHyb Hybridization Solution (BD Biosciences). The FA2H cDNA prepared in 100 AL Opti-MEM (Life Technologies, Invitrogen) and probe was an 862-bp fragment obtained by PCR using the plasmid pcDNA3- distributed in 24-well plates. HCT116 cells were trypsinized and resus- FA2H as template (6) and the oligonucleotides FA2H-5 (5¶-GCAGGGCTC- pended in normal growth medium to 2.5 Â 105/mL. Five hundred CATGGAGAACGAGCCTGTAGC-3¶)andFA2H-3(5¶-TGGGAGTTGT- microliters were then added to each well. For the generation of stable cell CACTGCGTCTTCAGGTGGGG-3¶). Probes were labeled with [a-32P]dCTP lines, HCT116 or HeLa cells were transfected as above and then subjected to using the Rediprime II Random Prime Labeling System kit (Amersham). G418 selection (400 or 500 Ag/mL for HCT116 or HeLa). Single colonies were isolated and analyzed for FA2H expression by Western blot. Results Incubation of cells with fatty acids. Cells were incubated with several concentrations of fatty acids (either 2-hydroxylated or not) for 24 h and then PM02734 inhibits cell growth in budding yeast. To analyze washed, placed in serum-free medium, and incubated with different whether PM02734 affected the growth of S. cerevisiae cells, we concentrations of PM02734 for an additional 24 h. Cell viability was constructed growth curves using different concentrations of the assessed by the CellTiter 96 Aqueous Non-Radioactive Cell Proliferation drug (Fig. 1A, left). The minimum concentration of PM02734 Assay, as described below. observed to completely inhibit cell growth over at least 24 hours To determine the incorporation of 2-hydroxy fatty acids into lipid was 250 Ag/mL (157 Amol/L). To determine whether the effect of A 14 species, the cells were labeled with 0.1 Ci/mL 2-hydroxy [1- C]palmitic PM02734 in yeast cells was cytostatic or cytotoxic, a survival test acid (specific activity, 50–60 mCi/mmol; American Radiochemicals, Inc.) for after incubation with several concentrations of PM02734 was done 24 h. As a positive control, the cells were incubated with 0.1 ACi/mL (Fig. 1A, right). Viability declined with increasing concentrations of [1-14C]oleic acid (Amersham). Then, lipids were extracted according to the method of Bligh and Dyer (7) and separated by TLC using three different PM02734, indicating that the compound was cytotoxic to yeast mobile phases: (a) diethyl ether/petroleum ether/acetic acid (50:50:1; to cells. separate 2-hydroxy fatty acids from nonhydroxylated fatty acids; ref. 8), (b) PM02734 rapidly induces necrosis in yeast cells. Following n-hexane/diethyl ether/acetic acid (70:30:1; to separate polar lipids from lethal injury, two modes of cell death can be distinguished in nonpolar lipids; ref. 9), and (c) chloroform/formic acid (95:5; to separate mammalian cells: apoptosis and oncosis, or primary necrosis (11). ceramides from glycerophospholipids; ref. 10). Radioactivity in the plates During the development of oncosis, membrane integrity is lost and was visualized by autoradiography. The position of the different radioactive cells become permeable, first to small molecules such as propidium spots was compared with that of commercial standards. iodide (PI) and then to larger molecules such as high-molecular Small interfering RNA knockdown experiments. HCT116 cells were weight dextrans (12). To test whether PM02734 permeabilized yeast transfected with small interfering RNA (siRNA) oligonucleotides using to PI, we treated cells with concentrations of PM02734 ranging Lipofectamine 2000 and following the recommendations of the Silencer A siRNA Starter kit (Ambion). Briefly, siRNA/Lipofectamine 2000 complexes from 50 to 350 g/mL and then did PI treatment. Flow cytometry were dispensed in 24-well plates and then mixed with 5 Â 104 cells per well. analysis revealed that PM02734 permeabilized cells to PI because a Three different siRNAs (Ambion) against FA2H were tested for the population of PI-positive cells appeared (Fig. 1B). PM02734- reduction of mRNA levels 24, 48, or 72 h after transfection. The most dependent PI permeation was found to be very quick (PI-positive effective was 5¶-CCACGGUUCAAAGUGGUGGTT-3¶ (sense) and 5¶-CCACCA- cells appeared after 1 hour of PM02734 treatment) and concentra- CUUUGAACCGUGGTT-3¶ (antisense) at 30 nmol/L 48 h after transfection. tion dependent. The percentage of PI-positive cells did not increase Protein extracts and Western blots. Cells transfected with pcDNA3 or after incubation times longer than 5 hours. PI-positive cells pcDNA3-FA2H or stable clones were washed twice with PBS and lysed with corresponded to dead cells, as shown by methylene blue staining A 100 L of lysis buffer [10 mmol/L Tris-HCl (pH 7.4), 1% Triton X-100, 0.1% (Fig. 1C, darkcells ). In summary, PM02734-induced cell death SDS, 150 mmol/L NaCl, 25 Ag/mL phenylmethylsulfonyl fluoride]. Protein results in the uptake of methylene blue and PI, suggesting that the concentrations were measured using the bicinchoninic acid protein assay kit (Pierce). For FA2H detection, 30 Ag of total protein extract were run on a cells become metabolically inactive and lose their integrity, which 10% SDS-PAGE gel, transferred to nitrocellulose, and incubated with anti- is different from mammalian apoptosis and yeast apoptosis-like human FA2H antibodies (1:1,000; ref. 6) for 2 h at room temperature. cell death, in which only cell viability is lost (13). Different studies Horseradish peroxidase–linked donkey anti-rabbit IgG antibodies were used carried out with the PM02734 analogue, KF, have indicated that the as secondary antibodies at 1:2,000 dilution. Inmunoblots were incubated for compound induces cell death with no markers of apoptosis but

Cancer Res 2008; 68: (23). December 1, 2008 9780 www.aacrjournals.org

Figure 1. A to C, effects of PM02734 in S. cerevisiae. A, growth curves of S. cerevisiae in the presence of the indicated concentrations of PM02734 (left) and percentage of viable cells after 5 h of treatment with PM02734 (right). B, fluorescence-activated cell sorting analysis showing the permeability of yeast cells to PI. Cells (108) were resuspended in 5 mL YPD and treated with PM02734 at the indicated concentrations, washed with water, and resuspended in 50 mmol/L sodium citrate containing 2 Ag/mL PI. Abscissas show FL2-H values (orange-red fluorescence, log scale). PI-negative peaks (PIÀ) correspond to background fluorescence. PI-positive peaks (PI+) represent the number of cells that became permeable to PI due to the effect of PM02734. C, cells treated with 250 Ag/mL PM02734 for 2 h were stained with PI and methylene blue. Dead cells are stained blue (dark cells in the black and white picture) and living cells are unstained. D, Western blots analysis of caspase-3 and PARP in extracts from HCT116 or U937 cells untreated or treated for the indicated doses and times with PM02734 and BEL. Actin was used as a loading control.

clear features of oncosis (14–16). One of the signature events of cell caspase-3 active fragments were readily detected (Fig. 1D, left). death by apoptosis is the activation of a cascade of proteolytic These data suggest that PM02734-induced HCT116 cell death is not enzymes, the so-called caspases, which are synthesized as of the apoptotic type. To confirm this, cleavage of the caspase-3 zymogens (procaspases) and undergo proteolytic maturation. To substrate PARP was studied as well. The digestion fragment of 85 examine the features of PM02734-induced HCT116 cell death, kDa only appeared in the positive controls but not in PM02734- analysis of procaspase-3 and its active fragment was studied by treated cells. immunoblot. Figure 1D shows that treating the cells with PM02734 PM02734 causes invaginations of the plasma membrane. did not induce cleavage of caspase-3. As a control of these To gain further insight into the alterations caused by PM02734 in experiments, we used cells exposed to 25 Amol/L bromoenol S. cerevisiae cells, we did electron microscopy studies (Fig. 2). lactone, a compound that induces apoptosis in a variety of cells by Untreated cells displayed a large nucleus, one or more large interfering with membrane phospholipid metabolism (17, 18), and vacuoles, and an intact membrane system (Fig. 2A). Surviving cells www.aacrjournals.org 9781 Cancer Res 2008; 68: (23). December 1, 2008

Figure 2. Electron micrographs of S. cerevisiae cells. A, untreated cells. B to E, cells treated with 250 Ag/mL PM02734 for 2 h. Arrows, plasma membrane invaginations. Scale bars are shown. exposed to PM02734 exhibited the same characteristics as by genes that participate in mitochondrial functions, including untreated cells (Fig. 2B, left). In contrast, cells killed by the action ATP5, ATP7, ATP4,orATP15, which are involved in ATP of PM02734 showed extensive cell disorganization and most biosynthesis, and COQ9, CAT5, and COQ4, with a role in the intracellular structures were destroyed, which is characteristic of biosynthesis of ubiquinone (coenzyme Q). Another functional necrotic cells. The most remarkable effect of PM02734 in yeast cells group included the following genes playing a role in cell wall was seen at the level of the plasma membrane, where we observed biosynthesis: FLC2, CWH41, CCW12, ROT2, HOC1, CHS1, ALG6, several membrane invaginations (Fig. 2B–E, arrows). Multiple YUR1, SKT5, and OST3. Finally, a group of genes participating in membrane vesicles were also observed inside PM02734-treated sphingolipid biosynthesis was found: CSG2 (this mutant was cells (Fig. 2C–E). already found in the top 40), FEN2, SUR2, SUR1,orBST1. Screening for PM02734-hypersensitive deletion mutants. Identification of PM02734-resistant mutants. To identify We did a genome-wide PM02734 sensitivity screening by replica strains resistant to PM02734, the whole collection of deletion plating the 4,848 deletion mutants into 96-well plates containing mutants was grown in the presence of 250 Ag/mL PM02734 for 24 250 Ag/mL PM02734. By doing so, we identified 199 mutants that hours (detailed information about this screening is supplied in were hypersensitive to the compound (Supplementary Table S1; Supplementary Data). Supplementary Table S3 shows the top 20 detailed information about the screening is supplied in Supple- most resistant strains to PM02734. In contrast to the sensitive mentary Data). Supplementary Table S2shows the top 40 most strains, these 20 mutants were not grouped in functional categories sensitive strains to PM02734. When we analyzed the whole list of by the Gene Ontology Term Finder tool. To determine their degree 199 hypersensitive strains (Supplementary Table S1), we found a of resistance more accurately, mutants in genes that presented a total of 39 genes involved in vesicle-mediated transport, of which clear human ortholog (tcb3, scs7, ste24, and ygr054w) were grown in 25 were specifically involved in Golgi trafficking (underlined). We rich medium, in the presence or absence of 250 Ag/mL PM02734, also found some other functional groups. One of them was formed and growth curves were obtained (Fig. 3). We found that scs7 was

Cancer Res 2008; 68: (23). December 1, 2008 9782 www.aacrjournals.org

constructed a S. cerevisiae strain in which the SCS7 gene was placed under the control of a regulatable GAL1 promoter. This promoter is repressed by glucose and induced in the presence of galactose (23). Scs7 overexpression resulted in increased sensitivity to PM02734, as revealed by spot assays (Fig. 4, YPGal panel), whereas repression of Scs7 resulted in resistance to the drug, like the scs7 mutant (Fig. 4, YPGlc panel). These cells were found to be resistant to PM02734 and impermeable to PI after treatment with the compound (data not shown). The drugs syringomycin E, MMS, and SDS were used as controls. Syringomycin E is an antifungal cyclic lipodepsinonapeptide produced by Pseudomonas syringae (24). Interestingly, the scs7 mutant has also been found to be resistant to this compound (3). As expected, PGALSCS7 cells grown on glucose behaved like the scs7 mutant in terms of resistance to syringomycin E (Fig. 4, YPGlc panel) and Scs7 overexpression slightly increased the sensitivity of the cells to syringomycin E (Fig. 4, YPGal panel). As controls, we used MMS, a DNA alkylating Figure 3. Growth curves of S. cerevisiae strains in the presence or absence of agent, and SDS, a detergent that causes the cell membrane to break PM02734. down. As shown in Fig. 4, cells lacking or overexpresing Scs7 behaved as the wild-type in terms of sensitivity to both MMS and SDS. These results indicated that levels of Scs7-mediated lipid the strain most resistant to PM02734 because it exhibited the hydroxylation specifically affect the sensitivity of the cells to lowest degree of growth inhibition. tcb3 was also found to be highly syringomycin E and PM02734. resistant, whereas ste24 and ygr054w were found to be only slightly Overexpression of FA2H increases the sensitivity of mam- resistant to the compound. malian cells to PM02734. To validate the results found in yeast in SCS7 encodes a sphingolipid fatty acid 2-hydroxylase, located in human cells, we decided to look at the effect of FA2H over- the endoplasmic reticulum, which introduces an a-OH into the expression and FA2H knockdown on the sensitivity of mammalian characteristic very long chain fatty acids of yeast sphingolipids cells to PM02734. It has previously been shown that COS7 cells (19, 20). Scs7 is a well-conserved protein and FA2H has been transfected with pcDNA3-FA2H contain higher levels of 2- identified as its human ortholog (6). tcb3 belongs to the yeast hydroxyceramides and 2-hydroxy fatty acids compared with family of tricalbins. Tricalbins are thought to be the yeast orthologs control cells, which indeed shows that the FA2H gene encodes a of mammalian synaptotagmins (21), a family with at least 14 fatty acid 2-hydroxylase (6). Transient overexpression of FA2H in mammalian variants involved in membrane-trafficking events (22). HCT116 cells leads to an increased sensitivity to PM02734 but has The high level of resistance of the scs7 mutant to PM02734, no effect on sensitivity to the alkylating agent MMS nor to together with its putative role in defining membrane conformation syringomycin E (see Supplementary Fig. S1A). Similar results were (3) and its high conservation in mammalian cells (FA2H), obtained in two additional mammalian cell lines: COS1 and prompted us to focus on Scs7/FA2H as a putative target of embryonic A293T (data not shown). PM02734. To examine in more detail the effect of FA2H overexpression on Sensitivity of S. cerevisiae cells to PM02734 depends on Scs7 the sensitivity of human cells to PM02734, we generated stable levels. Because SCS7 deletion led to resistance to PM02734 in S. clones overexpressing FA2H not only in HCT116 cells but also in cerevisiae cells, we hypothesized that overproduction of Scs7 could HeLa cells (Fig. 5). Clones overexpressing FA2H were identified by lead to hypersensitivity. To test this hypothesis, we first Western blot (Fig. 5A) and examined for sensitivity to PM02734. As

Figure 4. Sensitivity of S. cerevisiae strains to different drugs. The SCS7 gene was placed under control of the regulatable GAL1 promoter. In the presence of galactose (YPGal), the promoter is on and PGALSCS7 cells express high levels of Scs7. When cells are grown in glucose, the promoter is off and the PGALSCS7 strain behaves like the scs7 null mutant. WT, PGALSCS7,orscs7D cells were grown in rich medium with either glucose or galactose. Five microliters of 4 Â 106 cells/ mL and 1:10 serial dilutions were spotted onto plates containing either YPGal or YPGlc. www.aacrjournals.org 9783 Cancer Res 2008; 68: (23). December 1, 2008

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2008 American Association for Cancer Research. Cancer Research shown in Fig. 5B, these clones were clearly more sensitive to the of action of the antitumor drug PM02734. Here, we report that drug in both cell lines. Microscopic examination of these cells after PM02734 induces a rapid necrosis-like cell death in S. cerevisiae 1 hour of treatment showed a swollen phenotype also reported for and HCT116 cells, as also found for other human cell lines treated cells treated with the PM02734 analogue KF (15, 16, 25). In with the closest PM02734 analogue, KF. Yeast cells were found to be conclusion, these experiments indicate that overexpression of sensitive to PM02734 at concentrations much higher than human FA2H highly increases the sensitivity of mammalian cells to cells. This difference might be due to the presence of the yeast cell PM02734. wall because some of the mutant strains hypersensitive to PM02734 Exogenous addition of 2-hydroxy fatty acids increases the were deleted in genes involved in cell wall organization and sensitivity of mammalian cells to PM02734. In view of the biogenesis. results obtained in the overexpression experiments, we decided to The alterations caused by the PM02734 analogue, KF, in human investigate whether increasing the cellular levels of 2-hydroxylated cells (cytoplasmic swelling, vesiculation of cytoplasmic organelles, lipids, by incubating different cell lines (HCT116, U937, and HeLa) vacuolization, and mitochondrial and endoplasmic reticulum with exogenous 2-hydroxylated fatty acids, would result in an damage) prompted Suarez et al. (15) to suggest that the compound enhanced sensitivity to PM02734. Exogenous addition of fatty acids induces changes in the osmotic balance of the cell. Accordingly, the (2-hydroxylated or not) up to 50 Amol/L did not compromise cell mechanism of action of KF has been postulated to be similar to viability even in the absence of serum (data not shown). HCT116 those of other cytotoxic peptides, which induce cell death through cells that had been incubated in the presence of 2-hydroxy palmitic the formation of new ion channels in the membrane and/or by acid showed a marked increase in their sensitivity to PM02734 changing the activity of existing channels (14, 26). In particular, compared with cells incubated with palmitic acid or without any Garcia-Rocha et al. (25) suggested that KF may act similarly to fatty acid (Fig. 5D, left). This effect was specific for PM02734 monensin, a Na+/H+ ionophore produced by Streptomyces cinna- because the addition of 2-hydroxy palmitic acid did not change the monensis. This compound slows down intracellular transport, sensitivity to MMS. Similar results were obtained with HeLa and principally within the Golgi complex (27, 28). We found that 39 of U937 (data not shown). the top 199 strains most sensitive to PM02734 were deleted in To study the metabolic fate of 2-hydroxy palmitic acid in the genes involved in intracellular vesicle trafficking, 25 of them in cells, 14C-labeled fatty acid was used and the different lipid vesicle targeting to, from, or within the Golgi apparatus. These products into which the fatty acid could potentially incorporated results strongly indicate that PM02734 also affects intracellular were separated by TLC. Interestingly, 2-hydroxy palmitate incor- transport in yeast and that it might cause damage to the Golgi porated almost exclusively into the ceramide fraction in the three compartment, as also found for monensin. However, other cell lines used. Results with HCT116 cells are shown (Fig. 5D, right). organelles, such as mitochondria, may also be subject to the effect Essentially no 2-hydroxy palmitic acid remained in the cell as free of PM02734 treatment because mutants in genes exerting a role in fatty acid or bound to phospholipid. This distribution strongly mitochondrial functions were also found to be hypersensitive to contrasts with that of a nonhydroxylated fatty acid, such as PM02734. Mitochondrial dysfunctions have been associated with palmitic or oleic acid, where, as expected, most of the label cancer, and drugs targeting mitochondria have been suggested to appeared in the phospholipid and triacylglycerol fractions. These be important to inhibit cancer cell proliferation (29). If PM02734 results indicate that cell membranes with an increased pool of does affect mitochondria in human cells, this might explain, at 2-hydroxy fatty acid in ceramides are more sensitive to PM02734. least in part, its higher effectiveness in cancer cells because the Knockdown of FA2H increases the resistance of HCT116 cells rapid and continuous growth of tumor cells is highly energy to PM02734. To address whether FA2H depletion had an effect on dependent. the sensitivity of the cells to PM02734, we did RNA interference Interestingly, many of the yeast genes identified that participate experiments. Three different siRNAs were tested for a reduction in in intracellular transport are highly conserved in humans (see FA2H mRNA levels at 24, 48, or 72 hours after transfection. Northern Supplementary Data). Some of them, such as AP3D1 or SYNJ2, are blot analysis revealed a significant reduction in FA2H mRNA after expressed at low levels in certain tumors (30, 31). It will be very 48 hours in FA2H siRNA treated with one of the tested oligos versus interesting to analyze whether lowering the levels of the the nontargeting control siRNA (data not shown). HCT116 cells corresponding genes in human cells leads to increased sensitivity transfected with this siRNA were assayed for sensitivity to PM02734, to PM02734, as found in yeast. If this turns out to be the case, the MMS, or syringomycin E (Fig. 6A and B) and analyzed by Northern prediction will be that tumors expressing low levels of these blot (Fig. 6C). Survival assays showed that FA2H knockdown proteins will show a good response to PM02734 treatment. significantly increased the resistance to PM02734 (Fig. 6A). This Although PM02734 could interact with the lipid bilayer, it is effect was only observed for PM02734 because treatment with MMS unlikely to form pores by itself because the molecule is too small or syringomycin E did not affect HCT116 sensitivity to these drugs. to directly span the lipid bilayer. A minimum of 20 amino acids is Figure 6B shows that cells transfected with control siRNA cells were required for this action (26) and PM02734 is formed by only 14 completely swollen when treated with a high dose (7 Amol/L) of amino acids. One possibility is that PM02734 forms multimers. It PM02734 for 1 hour. In contrast, the majority of the cells transfected is also possible that the drug interferes with some component of with FA2H siRNA had the same appearance as the untreated cells, the membrane, resulting in the formation, or modification, of ion which visually confirmed that depletion of FA2H results in increased channels. It is known that the composition of the phospholipid resistance to PM02734. bilayer strongly influences the formation of channels by other cytotoxic peptides (26). In the present work, we have shown that sensitivity to PM02734 both in yeast and humans largely depends Discussion on the levels of Scs7/FA2H-mediated fatty acid hydroxylation. A In the present study, we took advantage of the cell biology and role for FA2H in membrane conformation has been suggested genetics of the budding yeast S. cerevisiae to clarify the mechanism based on the observation that 2-hydroxylation participates in

Cancer Res 2008; 68: (23). December 1, 2008 9784 www.aacrjournals.org

Figure 5. Increased sensitivity to PM02734 due to FA2H overexpression. A, level of FA2H protein in stably transfected clones of HCT116 and HeLa. B, cell viability of HCT116 and HeLa clones after 1 h of treatment with PM02734. Survival rates were calculated as percentages of untreated cells. Points, mean of the experiment done in triplicate; bars, SD. C, phase-contrast micrographs of HCT116 and HeLa clones treated with PM02734 for 1 h. D, left, effect of fatty acids on the survival rate of HCT116 cells treated with PM02734 (top) or MMS (bottom). The cells were incubated with 50 Amol/L palmitic acid (n), 50 Amol/L 2-hydroxy palmitic acid (E), or neither (control; 5) for 24 h. After this time, the cells were washed and exposed to the indicated concentrations of PM02734 or MMS, and viability was assayed as described in Materials and Methods. Right, TLC separation of 14C-labeled 2-hydroxy palmitate (left) and [14C]oleate-labeled (right) lipids from HCT116 cells. The plates were developed in the indicated solvents (top of each figure), and radioactivity was visualized by autoradiography. The position of authentic standards is indicated. FFA, free fatty acid; PL, phospholipids; DAG, diacylglycerol. www.aacrjournals.org 9785 Cancer Res 2008; 68: (23). December 1, 2008

Figure 6. Increased resistance to PM02734 by FA2H knockdown. HCT116 cells were transiently transfected with a siRNA against FA2H or with a nontargeting control siRNA. Twenty-four hours after transfection, medium was replaced by fresh medium, and 48 h after transfection, cells were assayed for sensitivity to different drugs. A, survival after treatment with PM02734, MMS, or syringomycin E was determined as described in Fig. 5B. B, phase-contrast micrographs of transfected cells untreated or treated with 7 Amol/L PM02734 for 1 h. C, effect of siRNA on FA2H mRNA levels was analyzed by Northern blot.

intramolecular and intermolecular hydrogen bonding (32, 33). the mechanism of action of PM02734: (a) fatty acid 2- Interestingly, it has previously been shown that sensitivity to hydroxylation allows PM02734 hydrogen bonding directly to another cyclic peptide, the pore-forming antifungal agent sphingolipids to facilitate ion channel formation or permeability syringomycin E, also depends on the presence of fatty acid to the drug and (b) 2-hydroxylation influences the formation 2-hydroxylation because the scs7 mutant was also found to be of some kind of membrane microdomain, not necessarily lipid resistant to this drug (3). One hypothesis, proposed by Hama rafts, present in the lipid bilayer that is important for drug- et al. (3), is that 2-hydroxylation would enhance the formation of membrane association. In this regard, we have shown that sterol/sphingolipid-rich membrane domains (lipid rafts), which PM02734 induces the formation of plasma membrane invagina- may serve as sites for syringomycin E binding and channel tions in yeast (Fig. 2). These structures may represent the formation. This hypothesis is supported by the fact that mutants theoretical microdomains. in other genes involved in sphingolipid biosynthesis are also Our results highlight the importance of the study of 2- resistant to the compound. However, this is not the case for hydroxylation in membrane conformation. The role of FA2H has PM02734. We found that mutants in sphingolipid biosynthesis only been investigated in the brain, where FA2H is a myelination- were hypersensitive to the compound, indicating that disruption associated gene (34), and in the skin, where it participates in the of lipid rafts by the depletion of sphingolipids does not lead formation and function of epidermal permeability barrier (35). to resistance but to hypersensitivity to PM02734. In other words, Sphingolipids containing 2-hydroxy fatty acids are uniquely lipid rafts may constitute a barrier to the action of PM02734 and abundant in the brain, skin, intestinal tract, and certain cancers not a platform for its action. Another difference between (36). Our results may have important implications in PM02734 syringomycin E and PM02734 is that the levels of FA2H-mediated therapy because tumor cells with high levels of sphingolipid fatty sphingolipid fatty acid 2-hydroxylation did not affect the acid 2-hydroxylation are expected to be highly sensitive to this sensitivity of human cells to syringomycin E. Based on our drug. FA2H could then become a marker of PM02734 activity and results, we propose one of the two following scenarios to explain help guide patient selection.

Cancer Res 2008; 68: (23). December 1, 2008 9786 www.aacrjournals.org

Disclosure of Potential Conflicts of Interest Grant support: PharmaMar. The costs of publication of this article were defrayed in part by the payment of page No potential conflicts of interest were disclosed. charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Acknowledgments We thank Hiroko Hama for the pcDNA3-FA2H plasmid and anti-FA2H antibodies and Jose Marıá Fernańdez-Sousa, Alberto Munõz, and members of our laboratory for Received 5/26/2008; revised 8/13/2008; accepted 8/14/2008. interesting discussions during this work.

References thelial cells during chemical hypoxia. Cell Death Differ and biosynthesis by peptide and polyketide synthetases. 2001;8:850–8. Microbiol Mol Biol Rev 1999;63:266–92. 1. Hamann MT, Otto CS, Scheuer PJ, Dunbar DC. 13. Madeo F, Frohlich E, Ligr M, et al. Oxygen stress: a 25. Garcia-Rocha M, Bonay P, Avila J. The antitumoral Kahalalides: bioactive peptides from a marine mollusk regulator of apoptosis in yeast. J Cell Biol 1999;145: compound kahalalide F acts on cell lysosomes. Cancer Elysia rufescens and its algal diet Bryopsis sp. J Org 757–67. Lett 1996;99:43–50. Chem 1996;61:6594–600. 14. Sewell JM, Mayer I, Langdon SP, Smyth JF, Jodrell DI, 26. Kourie JI, Shorthouse AA. Properties of cytotoxic 2. Yin J, Aviles P, Lee W, et al. Development of a liquid Guichard SM. The mechanism of action of kahalalide F: peptide-formed ion channels. Am J Physiol Cell Physiol chromatography/tandem mass spectrometry assay for variable cell permeability in human hepatoma cell lines. 2000;278:C1063–87. the quantification of PM02734, a novel antineoplastic Eur J Cancer 2005;41:1637–44. 27. Tartakoff AM. Perturbation of vesicular traffic with agent, in dog plasma. Rapid Commun Mass Spectrom 15. Suarez Y, Gonzalez L, Cuadrado A, Berciano M, the carboxylic ionophore monensin. Cell 1983;32:1026–8. 2006;20:2735–40. Lafarga M, Munoz A. Kahalalide F, a new marine-derived 28. Muren E, Oyen M, Barmark G, Ronne H. Identifica- 3. Hama H, Young DA, Radding JA, et al. Requirement of compound, induces oncosis in human prostate and tion of yeast deletion strains that are hypersensitive to sphingolipid a-hydroxylation for fungicidal action of breast cancer cells. Mol Cancer Ther 2003;2:863–72. brefeldin A or monensin, two drugs that affect syringomycin E. FEBS Lett 2000;478:26–8. 16. Janmaat ML, Rodriguez JA, Jimeno J, Kruyt FA, intracellular transport. Yeast 2001;18:163–72. 4. Wonisch W, Zellnig G, Kohlwein SD, et al. Ultrastruc- Giaccone G. Kahalalide F induces necrosis-like cell 29. Dias N, Bailly C. Drugs targeting mitochondrial tural analysis of HNE-treated Saccharomyces cerevisiae death that involves depletion of ErbB3 and inhibition of functions to control tumor cell growth. Biochem cells reveals fragmentation of the vacuole and an Akt signaling. Mol Pharmacol 2005;68:502–10. Pharmacol 2005;70:1–12. accumulation of lipids in the cytosol. Cell Biochem 17. Fuentes L, Perez R, Nieto ML, Balsinde J, Balboa MA. 30. Petrenko AA, Pavlova LS, Karseladze AI, Kisseljov FL, Funct 2001;19:59–64. Bromoenol lactone promotes cell death by a mechanism Kisseljova NP. Downregulation of genes encoding for 5. Longtine MS, McKenzie A III, Demarini DJ, et al. involving phosphatidate phosphohydrolase-1 rather subunits of adaptor complex-3 in cervical carcinomas. Additional modules for versatile and economical PCR- than calcium-independent phospholipase A2. J Biol Biochemistry (Mosc) 2006;71:1153–60. based gene deletion and modification in Saccharomyces Chem 2003;278:44683–90. 31. Rossi MR, Hawthorn L, Platt J, Burkhardt T, Cowell cerevisiae. Yeast 1998;14:953–61. 18. Perez R, Melero R, Balboa MA, Balsinde J. Role of JK, Ionov Y. Identification of inactivating mutations in 6. Alderson NL, Rembiesa BM, Walla MD, Bielawska A, group VIA calcium-independent phospholipase A2in the JAK1, SYNJ2, and CLPTM1 genes in prostate cancer Bielawski J, Hama H. The human FA2H gene encodes arachidonic acid release, phospholipid fatty acid incor- cells using inhibition of nonsense-mediated decay and a fatty acid 2-hydroxylase. J Biol Chem 2004;279: poration, and apoptosis in U937 cells responding to microarray analysis. Cancer Genet Cytogenet 2005;161: 48562–8. hydrogen peroxide. J Biol Chem 2004;279:40385–91. 97–103. 7. Bligh EG, Dyer WJ. A rapid method of total lipid 19. Mitchell AG, Martin CE. Fah1p, a Saccharomyces 32. Nyholm PG, Pascher I, Sundell S. The effect of extraction and purification. Can J Biochem Physiol 1959; cerevisiae cytochrome b5 fusion protein, and its hydrogen bonds on the conformation of glycosphingo- 37:911–7. Arabidopsis thaliana homolog that lacks the cyto- lipids. Methylated and unmethylated cerebroside stud- 8. Soberman RJ, Harper TW, Betteridge D, Lewis RA, chrome b5 domain both function in the a-hydroxylation ied by X-ray single crystal analysis and model Austen KF. Characterization and separation of the of sphingolipid-associated very long chain fatty acids. J calculations. Chem Phys Lipids 1990;52:1–10. arachidonic acid 5-lipoxygenase and linoleic acid N-6 Biol Chem 1997;272:28281–8. 33. Brown DA, London E. Functions of lipid rafts in lipoxygenase (arachidonic acid 15-lipoxygenase) of 20. Dunn TM, Haak D, Monaghan E, Beeler TJ. Synthesis biological membranes. Annu Rev Cell Dev Biol 1998;14: human polymorphonuclear leukocytes. J Biol Chem of monohydroxylated inositolphosphorylceramide (IPC- 111–36. 1985;260:4508–15. C) in Saccharomyces cerevisiae requires Scs7p, a protein 34. Maldonado EN, Alderson NL, Monje PV, Wood PM, 9. Balsinde J, Fernandez B, Solis-Herruzo JA, Diez E. with both a cytochrome b5-like domain and a Hama H. FA2H is responsible for the formation of 2- Pathways for arachidonic acid mobilization in zymosan- hydroxylase/desaturase domain. Yeast 1998;14:311–21. hydroxy galactolipids in peripheral nervous system stimulated mouse peritoneal macrophages. Biochim 21. Schulz TA, Creutz CE. The tricalbin C2domains: myelin. J Lipid Res 2007;49:153–61. Biophys Acta 1992;1136:75–82. lipid-binding properties of a novel, synaptotagmin-like 35. Uchida Y, Hama H, Alderson NL, et al. Fatty acid 2- 10. Penzo D, Petronilli V, Angelin A, et al. Arachidonic yeast protein family. Biochemistry 2004;43:3987–95. hydroxylase, encoded by FA2H, accounts for differenti- acid released by phospholipase A(2) activation triggers 22. Sudhof TC. Synaptotagmins: why so many? J Biol ation-associated increase in 2-OH ceramides during Ca(2+)-dependent apoptosis through the mitochondrial Chem 2002;277:7629–32. keratinocyte differentiation. J Biol Chem 2007;282: pathway. J Biol Chem 2004;279:25219–25. 23. Johnston M. A model fungal gene regulatory 13211–9. 11. Majno G, Joris I. Apoptosis, oncosis, and necrosis. An mechanism: the GAL genes of Saccharomyces cerevisiae. 36. Alderson NL, Walla MD, Hama H. A novel method for overview of cell death. Am J Pathol 1995;146:3–15. Microbiol Rev 1987;51:458–76. the measurement of in vitro fatty acid 2-hydroxylase 12. Nishimura Y, Lemasters JJ. Glycine blocks opening of 24. Bender CL, Alarcon-Chaidez F, Gross DC. Pseudomo- activity by gas chromatography-mass spectrometry. J a death channel in cultured hepatic sinusoidal endo- nas syringae phytotoxins: mode of action, regulation, Lipid Res 2005;46:1569–75.

www.aacrjournals.org 9787 Cancer Res 2008; 68: (23). December 1, 2008

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2008 American Association for Cancer Research. Levels of SCS7/FA2H-Mediated Fatty Acid 2-Hydroxylation Determine the Sensitivity of Cells to Antitumor PM02734

Ana B. Herrero, Alma M. Astudillo, María A. Balboa, et al.

Cancer Res 2008;68:9779-9787.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/68/23/9779

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2008/11/25/68.23.9779.DC1

Cited articles This article cites 36 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/68/23/9779.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/68/23/9779.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/68/23/9779. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.