Dihydroceramide accumulation and reactive oxygen species are distinct and nonessential events in 4-HPR-mediated leukemia cell death

Dihydroceramide accumulation and reactiveoxygen species are distinct and nonessentialevents in 4-HPR-mediated leukemia cell death

Aintzane Apraiz, Jolanta Idkowiak-Baldys, Naiara Nieto-Rementería,María Dolores Boyano, Yusuf A Hannun, and Aintzane Asumendi

Abstract: 4-(Hydroxyphenyl)retinamide (4-HPR) is a synthetic retinoid with a strong apoptotic effect towards different can-cer cell lines in vitro, and it is currently tested in clinical trials. Increases of reactive oxygen species (ROS) and modulationof endogenous sphingolipid levels are well-described events observed upon 4-HPR treatment, but there is still a lack ofunderstanding of their relationship and their contribution to cell death. LC–MS analysis of sphingolipids revealed that in hu-man leukemia CCRF-CEM and Jurkat cells, 4-HPR induced dihydroceramide but not ceramide accumulation even at suble-thal concentrations. Myriocin prevented the 4-HPR-induced dihydroceramide accumulation, but it did not prevent the loss ofviability and increase of intracellular ROS production. On the other hand, ascorbic acid, Trolox, and vitamin E reversed 4-HPR effects on cell death but not dihydroceramide accumulation. NDGA, described as a lipoxygenase inhibitor, exerted asignificantly higher antioxidant activity than vitamin E and abrogated 4-HPR-mediated ROS. It did not however rescue cel-lular viability. Taken together, this study demonstrates that early changes observed upon 4-HPR treatment, i.e., sphingolipidmodulation and ROS production, are mechanistically independent events. Furthermore, the results indicate that 4-HPR-driven cell death may occur even in the absence of dihydroceramide or ROS accumulation. These observations should betaken into account for an improved design of drug combinations.

Key words: 4-HPR, leukemia, dihydroceramide, ROS, cell death.

Résumé : Le 4-(hydroxyphényl)rétinamide (4-HPR) est un rétinoïde synthétique qui exerce des effets pro-apoptotiques im-portants sur différentes lignées cellulaires cancéreuses in vitro, et il fait actuellement l’objet d’études cliniques. L’augmenta-tion d’espèces réactives d’oxygène (ERO) et la modulation des niveaux de sphingolipides endogènes sont des événementsbien décrits qui sont observés à la suite d’un traitement au 4-HPR, mais on comprend encore mal leur relation et leur contri-bution à la mort cellulaire. Une analyse des sphingolipides par LC–MS a révélé que chez les cellules de leucémie CCRF-CEM et Jurkat, le 4-HPR induisait l’accumulation de dihydrocéramides mais pas de céramides, même à des concentrationssublétales. La myriocine prévenait l’accumulation de dihydrocéramides induite par le 4-HPR, mais elle ne prévenait pas laperte de viabilité et l’augmentation de la production d’ERO. D’un autre côté, l’acide ascorbique, le Trolox et la vitamine Erenversaient les effets du 4-HPR sur le plan de la mort cellulaire, mais pas l’accumulation de dihydrocéramides. Le NDGA,décrit comme inhibiteur de lipoxygénase, exerçait une activité antioxydante plus élevée que la vitamine E et abrogeait laproduction d’ERO induite par le 4-HPR. Il ne pouvait cependant pas restaurer la viabilité cellulaire. Dans son ensemble,cette étude démontre que les changements précoces observés à la suite d’un traitement au 4-HPR, c.-à-d. la modulation dessphingolipides et la production d’ERO, sont des événements indépendants d’un point de vue mécanistique. De plus, les ré-sultats indiquent que la mort cellulaire induite par le 4-HPR peut survenir même en absence d’accumulation de dihydrocéra-mides ou d’ERO. Ces observations devraient être prises en considération afin d’améliorer la conception de protocoles decombinaison médicamenteuse.

Received 8 November 2011. Revision received 31 December 2011. Accepted 3 January 2012. Published at www.nrcresearchpress.com/bcbon 19 March 2012.

Abbreviations: 4-HPR, N-(4-hydroxyphenyl)retinamide; AA, L-ascorbic acid; ATRA, all-trans retinoic acid; baic, baicalein, 5,6,7-trihydroxyflavone; C12-dhCCPS, D-erythro-2-N-[12_-(1_- D-erythro-2-N-[12_-(1_-pyridinium)dodecanoyl]-4,5-dihydrosphingosinebromide, D-erythro-C12-dihydroceramide; Cer, ceramide; DES, dihydroceramide desaturase; dhCer, dihydroceramide;dhSph, dihydrosphingosine; dhSL, dihydrosphingolipids; ER, endoplasmic reticulum; LC–MS, liquid chromatography – massspectrometry; myr, myriocin; NDGA, nordihydroguaiaretic acid; RAR/RXR, retinoic acid receptor; ROS, reactive oxygen species; SD,standard deviation; SL, sphingolipid; Sph, sphingosine; SPT, serine palmitoyltransferase; T, Trolox, ((±)-6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid); T-ALL, T-cell acute lymphoblastic leukemia; TLC, thin layer chromatography; Tm, Trolox-methylether; and vitE, vitamin E.

A. Apraiz, N. Nieto-Rementería, M.D. Boyano, and A. Asumendi. Department of Cell Biology and Histology, School of Medicine andDentistry, University of the Basque Country, Sarriena s/n, 48940 Leioa (Bizkaia), Spain.J. Idkowiak-Baldys and Y.A. Hannun. Department of Biochemistry and Molecular Biology, Medical University of South Carolina,Charleston, SC 29425, USA.

Corresponding author: Aintzane Asumendi (e-mail: [email protected]).

209

Biochem. Cell Biol. 90: 209–223 (2012) doi:10.1139/O2012-001 Published by NRC Research Press

Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Mots‐clés : 4-HPR, leucémie, dihydrocéramide, ERO, mort cellulaire.

[Traduit par la Rédaction]

IntroductionA large number of studies based on 4-(hydroxyphenyl)reti-

namide (4-HPR, fenretinide) describe this agent as a promis-ing chemotherapeutic drug, but its mechanism of actionremains not fully defined. 4-HPR is a synthetic amide deriv-ative of the all-trans retinoic acid (ATRA) that is being testedagainst several cancers in vitro (Maurer et al. 1999; Sun et al.1999; Asumendi et al. 2002; Tiwari et al. 2006) and in vivo,as well as in clinical trials (Reynolds and Lemons 2001; Ga-raventa et al. 2003; Veronesi et al. 2006; Sabichi et al. 2008).The interest in 4-HPR over other retinoids is based on 3 ma-jor advantages: (i) unlike vitamin A and many natural deriva-tives such as ATRA, 4-HPR exerts its cell growth inhibitoryeffect mainly by inducing cell death rather than differentia-tion (Ponzoni et al. 1995; Kitareewan et al. 1999), (ii) it canbe applied against retinoic acid resistant cancers (Delia et al.1993; Kitareewan et al. 1999; Reynolds and Lemons 2001),and (iii) it shows a favorable systemic toxicity profile com-pared with other retinoids that allows and supports continuityon in vivo studies as well as further application in clinicaltrials (Rotmensz et al. 1991).Despite the positive characteristics of 4-HPR, there is still

controversy about its mechanism of action (Hail et al. 2006;Kadara et al. 2007; Tiwari et al. 2008). Previous studies fromour group and others described a strong proapoptotic effect of 4-HPR on several T-cell acute lymphoblastic leukemia (T-ALL)cell lines (Asumendi et al. 2002; O’Donnell et al. 2002; Fa-derl et al. 2003). In agreement with data obtained in severalcell models (Suzuki et al. 1999; Kim et al. 2006), oxidativestress (i.e., increase of reactive oxygen species (ROS) pro-duction) seemed to be a candidate mediator of 4-HPR-driven cell death in our leukemia model (Asumendi et al.2002). Moreover, we found that upon 4-HPR treatment, theRAR/RXR independent mitochondrial apoptosis pathwaywas activated where enhanced ROS production and modula-tion of sphingolipids (SLs) levels were the earliest eventsdetected (Morales et al. 2007). While 4-HPR-mediated ROSincrease is an established event (Suzuki et al. 1999; Asu-mendi et al. 2002; Kim et al. 2006; Kadara et al. 2007), therelationship between 4-HPR and SL metabolism has not beenfully elucidated (Maurer et al. 1999; Lovat et al. 2004; Reh-man et al. 2004; Morales et al. 2007; Darwiche et al. 2007).SLs are generally described as sphinganine- or sphingo-

sine-based lipids that form a large family of molecules withboth structural and signaling functions (Hannun and Obeid2008). Among them, ceramide (Cer) (sphingosine-based SL)is currently considered the core of the SL metabolic network,and much work has been focused on its signaling role in celldeath (Obeid et al. 1993; Taha et al. 2006; Carpinteiro et al.2008). As summarized in Fig. 1, Cer can be generated by 3different pathways, i.e., de novo synthesis, sphingomyelin hy-drolysis, and the recently described salvage pathway (Hannunand Obeid 2008; Kitatani et al. 2008). Numerous studieshave implicated Cer in 4-HPR-mediated cytotoxicity (Maurer

et al. 1999; Rehman et al. 2004; Darwiche et al. 2005; Hail etal. 2006; Morale s et al. 2007; Jiang et al. 2011) but applica-tion of more advanced liquid chromatography – mass spec-trometry (LC–MS) techniques has recently revealed that 4-HPR induces an increase in dihydroceramide (dhCer) ratherthan Cer, at least in some tumor models (Kraveka et al. 2007;Wang et al. 2008; Valsecchi et al. 2010).The aim of this work was to establish the mechanistic re-

lationship among the 4-HPR-induced oxidative stress,changes in SLs, and cell death. SL levels were analyzed byLC–MS technology in the leukemia models previouslystudied by our group. The analysis revealed accumulation ofdhCer (but not Cer) following 4-HPR treatment. Most impor-tantly, this study defines oxidative stress and dhCer accumu-lation as 2 distinct events occurring early after exposure to 4-HPR. On top of that, these data also indicate that, at least inT-ALL cell lines, cell death upon treatment may occur evenin the absence of accumulation of dhCer or ROS, indicatingthat, unlike what was previously proposed, none of theseearly events are necessarily essential mediators of 4-HPR-mediated cytotoxicity.

Materials and methods

ReagentsRPMI 1640 (No. 11835-034), red phenol-free RPMI 1640

Fig. 1. General view of sphingolipids synthesis and recycling path-ways. Thicker arrows point out the pathway analyzed in the presentstudy. SPT, serine palmitoyltransferase; dhSph, dihydrosphingosine(sphinganine); DES, dihydroceramide desaturase; dhCer, dihydrocer-amide; Cer, ceramide; complex dhSL, complex dihydrosphingolipids(including dihydrosphingomyelin, dihydroglucoceramide and deriva-tives); complex SL, complex sphingolipids (sphingomyelin, gluco-ceramide and derivatives); Sph, sphingosine; and 4-HPR, N-(4-hydroxyphenyl)retinamide.

210 Biochem. Cell Biol. Vol. 90, 2012

Published by NRC Research Press

Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

(No. 11835-063), and heat inactivated fetal bovine serum(No. 10082-174) were from GIBCO/BRL (Invitrogen, Carls-bad, Calif.). 4-HPR (No. H7779), H2O2 (No. H1009), myrio-cin (No. M1177), and antioxidants (except baicalein) werepurchased from Sigma Chemical Co. (St. Louis, Mo.). 4-HPR (10 mmol/L) was dissolved in DMSO, aliquoted, andstored at –80 °C. Myriocin (serine palmitoyltransferase inhib-itor) was prepared at 1 mmol/L (in DMSO) and aliquots werestored at –20 °C. Ascorbic acid, vitamin E, Trolox, and Tro-lox-methylether were prepared fresh prior to each experimentas follows: vitamin E (No. T1539; diluted in ethanol (1:10)prior to application); Trolox ((±)-6-hydroxy-2,5,7,8-tetrame-thylchromane-2-carboxylic acid; No. 238813; dissolved inculture media to 1 mmol/L); Trolox-methylether (No. 93510;dissolved in ethanol to 100 mmol/L due to low water solubil-ity); and L-ascorbic acid (No. 255564; dissolved in water).Nordihydroguaiaretic acid (NDGA) (No. N5023) stock solu-tion was prepared at 80 mmol/L in DMSO (stored at –20 °C). Baicalein (No. 196322; stock solution in DMSO) andAnnexin V-FITC apoptosis Detection Kit (No. PF032) werefrom Calbiochem (San Diego, Calif.). CM-H2DCFDA (No.C6827), BODIPY® 581/591 C11 (D3861), and MitoSOX(No. M36008) were obtained from Molecular Probes (Invi-trogen). The cell viability XTT assay was purchased fromRoche Molecular Biochemicals, Ind.(No. 11465015001). D-er-ythro-2-N-[12′-(1″-pyridinium)dodecanoyl]-4,5-dihydrosphingo-sine bromide (D-erythro-C12-dihydroceramide; C12-dhCCPS)was synthesized in the Lipidomics Core Facility at theMedical University of South Carolina (Szulc et al. 2006).Cells were expose to a final DMSO concentration ≤0.1%.

Cell lines and culture conditionsHuman CCRF-CEM and Jurkat acute lymphoblastic leuke-

mia cells were grown in RPMI 1640 with 2 mmol/L L-gluta-mine and supplemented with 10% heat inactivated fetalbovine serum (complete culture medium). Cells were main-tained at 37 °C in a humidified incubator containing 5%CO2. When required, cells were exposed to 2 h incubationwith antioxidants or serine palmitoyltransferase inhibitorprior to adding the drug.

Metabolic activity assayA standard XTT assay was used to determine cell viability.

For 4-HPR-only treatments, cells were plated in 96-wellplates at 750 000 cells/mL and 100 µL/well. After 4 h, treat-ments were added at 50 µL/well obtaining a final density of500 000 cells/mL and final volume of 150 µL/well. Four rep-licates were used per experimental condition. XTT reagentmixture was added 4 h before the end of the selected treat-ment period and absorbance at 490 nm was determined foreach well (in accordance with the manufacturer's instruction).A slightly modified protocol was used for the analysis of theeffect of myriocin (final concentration 100 nmol/L) or antiox-idant on 4-HPR treatment. Briefly, cells were seeded on60 mm culture dishes and myriocin or antioxidants wereadded after 4 h. 4-HPR treatment was added 2 h later andcells were plated in quadruplicates in 96-well plates(150 µL/well).

Clonogenic capacity assayPrior to seeding in the semisolid medium, cells were ex-

posed to selected 4-HPR concentrations for 16 h (5 × 106cells/sample, ∼500 000 cells/mL). The 16 h time point wasselected because of the lack of difference among cells chal-lenged to 16 h and 48 h treatment when compared at the48 h time point (data not shown). The 6-well plates werecovered by 2 mL/well of prewarmed 1.2% methylcellulosemedium and a layer of 0.6% methylcellulose medium con-taining 50 000 cells was placed on top of the 2 mL 1.2%methylcellulose layer. The 6-well plates were maintained at37 °C in a humidified incubator containing 5% CO2 andchecked up to colony growth (7–9 days). The clonogenic ca-pacity was determined by counting colonies >50 cells in 3random areas/well by an inverted Nikon Eclipse TS100 mi-croscope.

Apoptotic cell deathApoptotic cell death was determined by the annexin V-

propidium iodide (PI) staining assay in accordance withthe manufacturer’s RAPID protocol. We analyzed 10 000cells/sample using a Coulter EPICS ELITE ESP flow cy-tometer. Cells were classified as (i) healthy cells (PI (–)and FITC (–) cells), (ii) early apoptotic (PI (–) and FITC (+)cells), and (iii) late apoptotic/necrotic cells (PI (+) and FITC(+) cells). Results were analyzed by WinMDI 2.8 and Sum-mit v. 4.3 softwares.

Measurement of general oxidative stressIntracellular oxidative stress was determined by CM-

H2DCFDA. Based on available literature, CM-H2DCFDAand its former versions are able to detect hydrogen peroxide,hydroxyl radicals, peroxyl radicals, and peroxynitrite anions.Briefly, cells were seeded in red phenol-free RPMI-1640-based complete culture media (5 × 106 cells per treatment).As done for the viability assay, myriocin (final concentrationof 100 nmol/L) or antioxidants were added after 4 h of incu-bation. Following 2 h myriocin/antioxidant treatment, 4-HPRwas added and the cells were incubated for selected timeperiods. After treatment, approximately 2 × 106 cells were re-suspended in 0.8 mL PBS containing 10 µmol/L CM-H2DCFDA and incubated for 25 min at 37 °C in the dark.Samples were transferred to tubes or 96-well plates prior tofluorescence measurement. Fluorescence intensity was meas-ured either by a plate reader (Fluoroskan Ascent, Labsys-tems) or by a flow cytometer (Coulter EPICS ELITE ESP).Probe-free cells were used as a negative control and freshprepared H2O2 (200 µmol/L, 15 min) was used as a positivecontrol.

Mitochondrial superoxide productionMitochondrial superoxide production was estimated by

MitoSOX oxidation. Cells were seeded and treated as speci-fied for the analysis of the intracellular oxidative stress. Aftertreatment, approximately 2 × 106 cells were resuspended in0.8 mL Hank's buffered salt solution containing 5 µmol/LMitoSOX and incubated for 15 min at 37 °C in the dark.Samples were washed with PBS and transferred to tubes or96-well plates prior to fluorescence measurement. Fluores-cence intensity was measured either by a plate reader or by aflow cytometer.

Apraiz et al. 211


Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Intracellular lipid peroxidationThis assay was based on the oxidation-mediated fluores-

cence shift of an externally added undecanoic acid (BODIPY581/591 C11). Oxidation of the polyunsaturated butadienylportion of the BODIPY 581/591 dye truncates the conjugatedp-electron system, resulting in a shift of the fluorescenceemission peak from ∼590 nm to ∼510 nm.Briefly, cells were seeded and treated as described for in-

tracellular oxidative stress and mitochondrial superoxide esti-mation. Incubation for 3 h at 37 °C in the dark was requiredfor proper internalization of the probe (10 µmol/L BODIPY581/591 C11), and treatments were added before or after theprobe depending on the selected treatment exposure time.BODIPY and treatment-containing culture medium was re-placed by 1 mL PBS/sample prior to measuring fluorescence.Flow cytometric analysis of lipid peroxidation was performedin the General Research Services SGIker of the UPV/EHU(http://www.ikerkuntza.ehu.es/p273-sgikerhm/en/). Ten thou-sand cells per sample were analyzed.

In situ dihydroceramide desaturase assayC12-dhCCPS was used as a substrate for the dihydrocera-

mide desaturase (DES) enzyme. The synthetic analogue wasdissolved in 100% ethanol at a 100 mmol/L stock concentra-tion and stored at –20 °C, protected from light. Stock con-centration was diluted and added to cells in complete culturemedium. The final C12-dhCCPS concentration in mediumwas 500 nmol/L. When required, C12-dhCCPS was added to-gether with 4-HPR and incubated for the selected time peri-ods. Levels of C12-dhCCPS and its product (C12-CCPS) wereanalyzed by LC–MS as described by Szulc et al. (2006) andBielawski et al. (2006).

LC–MS analysis of endogenous sphingolipid speciesCells were seeded and treated in 60 mm culture dishes as

described above. After treatment, cells were washed twicewith PBS to avoid external SL contamination from mediaand further analysis was performed in the Lipidomics CoreFacility at the Medical University of South Carolina (www.musc.edu/BCMB/lipidomics) as described previously (Bie-lawski et al. 2006; Szulc et al. 2006). Total values were nor-malized to inorganic phosphates by an adapted Bligh andDyer lipid extraction (Van Veldhoven and Bell 1988).

StatisticsResults were represented as the mean ± standard deviation

(SD). The number of replicates was specified for each experi-ment. Statistical analysis was performed by SPSS 15.0 soft-ware. Depending on the compared set of data, a Student's ttest (for pairwise mean comparison) or ANOVA plus a Bon-ferroni/Tamhane post hoc (based on variance homogeneitytest) analysis were applied. Statistical significance was set at95% (p < 0.05) or 99% (p < 0.01) as indicated in each fig-ure.

Results

4-HPR cytotoxicity in T-ALL cell linesPrevious studies demonstrated the strong apoptotic effect

of 4-HPR in leukemia cells as well as the dose and time-

dependence of the cytotoxicity (Asumendi et al. 2002; Mo-rales et al. 2005). In this study we selected a 48 h treatmentexposure to determine acute toxicity profiles upon a widerange of 4-HPR concentrations while long-term toxicitywas determined by the clonogenic assay. As shown inFig. 2A, 4-HPR concentrations ≥1 µmol/L induced acuteloss of viability in both leukemia cell lines; after 48 h at3 µmol/L, the number of CCRF-CEM and Jurkat viablecells decreased to 15.3% ± 6.3% and 21.1% ± 10.2%, re-spectively, and cell death was almost 100% with 10 µmol/Las shown by annexin V-propidium iodide staining (data notshown). Despite the observed differences in drug-drivenacute toxicity (Fig. 2A), 4-HPR decreased the clonogeniccapacity in a similar fashion in both cell lines (p > 0.05;Fig. 2B), even at concentrations as low as 0.5 µmol/L.Next, we determined the effect of 4-HPR on the cell cycleto evaluate cell cycle arrest as a possible mechanism under-lying the observed decrease of the clonogenic capacity upontreatment. PI labeling mediated analysis of the cell cyclefollowing 24–48 h exposure to 4-HPR (0.5 µmol/L) revealed

Fig. 2. Effect of N-(4-hydroxyphenyl)retinamide (4-HPR) on cellviability (A) and clonogenic capacity (B). (A) Cell viability was es-timated by means of metabolic activity. Cells were treated with theindicated concentration of 4-HPR for 48 h and viability was deter-mined by an XTT assay. Values are shown as the percent of un-treated control cells ± SD of at least 3 independent experimentsperformed in quadruplicate (n ≥ 12). (B) For clonogenic capacity,cells were treated in normal culture conditions for 16 h with the in-dicated 4-HPR concentrations. After 16 h, treatment was removedand cells were seeded on semisolid methylcellulose medium asmentioned in the Materials and methods. Colonies were counted at7–9 days. As internal assay controls, cells for the clonogenic assaywere also subjected to 16 h viability estimation (XTT) and apoptosis(annexin V-propidium iodide) determination (data not shown). Datarepresent the average of colony-number percent relative to untreatedcells ± SD of 3 independent experiments. **, p < 0.01.



Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

no significant differences among treated and untreatedCCRF-CEM and Jurkat cells (data not shown). Thus, 4-HPR exerted not just acute but also long-term antitumor ac-tivity in selected T-ALL cell lines.

DES activity and endogenous dhCer levels upon 4-HPRtreatmentThe effect of 4-HPR on DES was described by Kraveka

and colleagues (2007) in neuroblastoma cells. As shown inFig. 3A, 4-HPR inhibited DES activity in CCRF-CEM leuke-mia cells in a dose- (p < 0.01) and time- (p < 0.05) depend-ent manner, leading to a concomitant increase of theendogenous cellular dhCer content (Fig. 3B). Interestinglyenough, DES inhibition occurred within the first hour oftreatment at 4-HPR concentrations as low as 0.5 µmol/L andwas sustained as shown by the persistent effect on the in situconversion of C12-PyrdhCer to C12-PyrCer during 4-HPRexposure (i.e., 14.8% ± 0.6% after 1 h vs. 15.3% ± 1.9%after 6 h exposure to 4-HPR; Fig. 3A). LC–MS analysis ofendogenous SLs revealed a similar pattern of dhCer accumu-lation in both CCRF-CEM and Jurkat cells (Fig. 3C) at bothsublethal and cytotoxic 4-HPR concentrations. Of note, wealso observed an increase in the level of endogenous dihy-drosphingosine (dhSph, also known as sphinganine; Supple-mentary Fig. 2A)1 in CCRF-CEM cells at 4-HPRconcentrations ≥1 µmol/L. Similar dhSph accumulation re-quired higher 4-HPR concentrations (6 h, 10 µmol/L; datanot shown) in Jurkat cells. Under the same conditions, aslight accumulation of sphingosine (Sph) was also observedin CCRF-CEM cells (Supplementary Fig. 3A)1 but not in Ju-rkat cells (data not shown). Moreover, no increase of endoge-nous Cer levels was observed in any of the tested cell linesand conditions (Fig. 3D), consistent with recent studies (Kra-veka et al. 2007). These data demonstrated that in our leuke-mia model, 4-HPR driven rapid inhibition of cellular DES isthe likeliest source for the observed accumulation of endoge-nous dhCer. Of note, the increase in dhCer levels occurredeven at 4-HPR concentrations below those required for acutetoxicity.

Fig. 3. Effect of N-(4-hydroxyphenyl)retinamide (4-HPR) on dihy-droceramide desaturase (DES) activity and endogenous dihydrocera-mide/ceramide levels. CCRF-CEM cells were incubated for 1 h withthe synthetic dihydroceramide analogue (C12-PyrdhCer or C12-dhCPPS) prior to adding the indicated 4-HPR concentrations. Cellswere collected after 1 h or 6 h of 4-HPR exposure. Synthetic (A)and endogenous (B) sphingolipids were measured by LC–MS. DESactivity was estimated as the percent of desaturated product (C12-PyrCer) related to total C12-Pyr (C12-PyrdhCer plus C12-PyrCer) incells. The dose-dependent effect of 4-HPR on endogenous dihydro-ceramide (dhCer) and ceramide (Cer) levels are represented in (C)and (D). Cells were treated with low (0.25–0.5 µmol/L) and moder-ate (1–3 µmol/L) 4-HPR concentrations and endogenous total dihy-droceramide (C) and ceramide (D) levels were measured by liquidchromatography – mass spectrometry after 6 h exposure. Data re-present the mean ± SD of 2 independent experiments performed induplicate.

1Supplementary data are available with the article through the journal Web site (www.nrcresearchpress.com/doi/suppl/10.1139/o2012-001/).

Apraiz et al. 213


Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Role of endogenous dhCer in 4-HPR-induced oxidativestress and cytotoxicityPreincubation with the serine palmitoyltransferase (SPT)

inhibitor myriocin was performed to evaluate the role ofdhCer accumulation in 4-HPR-mediated toxicity and ROSproduction. For this set of experiments, 4-HPR concentrationand treatment exposure were established based on the highlethality observed with 3 µmol/L 4-HPR following 48 h incu-bation (Fig. 2A) and the robust dhCer accumulation obtainedafter 6 h exposure to 4-HPR (Figs. 3B–3C). As shown in

Fig. 4A, myriocin (100 nmol/L) effectively blocked the denovo SL synthesis pathway and completely abolished the 4-HPR- (3 µmol/L) induced dhCer accumulation in bothCCRF-CEM and Jurkat cells (4-HPR vs. myriocin + 4-HPR)as well as previously described dhSph and Sph accumula-tions (Supplementary Fig. 2B and SupplementaryFigs. 3B.1–B.2, respectively)1. As expected, inhibition of thede novo synthesis pathway decreased endogenous Cer levelsin both cell lines; no difference in Cer content was observedin the 4-HPR-treated and untreated cells (Fig. 4A).

Fig. 4. Role of the de novo sphingolipid synthesis in N-(4-hydroxyphenyl)retinamide (4-HPR)-mediated reactive oxygen species (ROS) in-crease and cell death. Cells were treated with 100 nmol/L myriocin (myr) for 2 h, followed by 4-HPR treatment. Cellular levels of dihydro-ceramide (dhCer) and ceramide (Cer) were measured by LC–MS after 6 h of 4-HPR (3 µmol/L) treatment (A), general ROS production wasdetermined by CM-H2DCFDA after 30 min 4-HPR (3 µmol/L) exposure (B), and cellular viability was estimated by an XTT assay (after 16 hand 24 h exposure (C). Values are presented as the mean ± SD of 2 independent experiments (A) or the percent related to correspondingcontrol cells ± SD of at least 3 independent experiments performed in quadruplicate (n ≥ 12; B–C). Fifteen minutes of H2O2 (200 µmol/L)exposure was used as the internal positive control for ROS production (B). **, p < 0.01; and p > 0.05, no significant differences.

Fig. 5. Effect of antioxidants on N-(4-hydroxyphenyl)retinamide (4-HPR)-mediated cytotoxicity and dihydroceramide accumulation. CCRF-CEM and Jurkat cells were incubated for 2 h with 100 µmol/L ascorbic acid (AA) or vitamin E (vitE) prior to 3 µmol/L 4-HPR treatment.Effect of radical scavengers on ROS upon 4-HPR (30 min) exposure was measured by CM-H2DCFDA oxidation (A). Values are presented asthe percent of corresponding control cells ± SD of 3 independent experiments performed in quadruplicate (n = 12). **, p < 0.01, relative tocells exposed to 4-HPR (3 µmol/L). Cell viability following 4-HPR ± antioxidants was determined by an XTT assay (B) and apoptotic celldeath (C) estimated by annexin V-FITC/propidium iodide (PI) labeling. Data are representative of 2 independent experiments. Endogenoustotal dihydroceramide levels (D) were measured by LC–MS as described in the Materials and methods. Data represent the mean ± SD of 2independent experiments.



Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Apraiz et al. 215


Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Next, the role of 4-HPR-mediated dhCer increase in theobserved oxidative stress induction (SupplementaryFigs. 1A–1B)1 was evaluated by employing myriocin fol-lowed by 30 min 4-HPR treatment. As shown in Fig. 4B,30 min treatment with 3 µmol/L 4-HPR was enough to in-duce a 4.91 ± 0.44-fold (CCRF-CEM) and 2.35 ± 0.27-fold(Jurkat) increase in ROS production relative to untreated con-trol cells, while 10 µmol/L 4-HPR induced an even higherROS accumulation (data not shown). No difference was ob-served between the untreated control and myriocin-treatedcells (p > 0.05), and myriocin pretreatment did not prevent4-HPR-induced oxidative stress in any of the 4-HPR concen-trations or cell lines (4-HPR vs. myriocin + 4-HPR; p >0.05). Fifteen minutes of H2O2 exposure was chosen as aninternal positive control for ROS detection. Thus, inhibitionof the accumulation of dhCer and free sphingoid bases didnot affect the production of ROS, negating a role for dhCerand free sphingoid bases in the ROS response to 4-HPR.Next, we analyzed the effect of myriocin on 4-HPR-induced

acute loss of viability upon 16–24 h exposure to assess anyearly or transient effect of myriocin on 4-HPR- treatment.Myriocin treatment did not prevent 4-HPR-mediated cyto-toxicity in either cell line (CCRF-CEM and Jurkat) and atany time point (16 and 24 h) or 4-HPR concentration(1 µmol/L and 3 µmol/L) (p > 0.05; Fig. 4C). Therefore,the results suggest that the observed 4-HPR-induced dhCer,dhSph, and Sph accumulation do not mediate acute celltoxicity.

Effect of radical scavengers on 4-HPR-driven dhCeraccumulation and antitumoral activityAs shown in Fig. 5A, ascorbic acid and (+)-a-tocopherol

(vitamin E) significantly decreased the 4-HPR- (3 µmol/L)triggered increase in ROS production in both cell lines (p <0.01) although complete inhibition was achieved only with as-corbic acid. Despite partial ROS buffering capacity, vitamin Eprovided a sustained capacity to abolish 4-HPR-mediated celldeath as evaluated by both metabolic activity (Fig. 5B) andannexin V-propidium iodide staining (Fig. 5C) measure-ments. Protection by ascorbic acid, on the other hand, lastedno longer than 24 h in either cell line (Fig. 5B).Next, the effects of antioxidants on endogenous dhCer and

Cer levels were determined by LC–MS. Trolox (synthetic de-rivative of vitamin E) and Trolox-methylether (modified Tro-lox molecule without antioxidant activity) were testedtogether with the previously mentioned ascorbic acid and vi-tamin E. Cells were subjected to 2 h preincubation with anti-oxidants prior to adding 4-HPR (3 µmol/L) as performed forthe analysis of oxidative stress and cytotoxicity. Regardless oftheir capacity to buffer 4-HPR-induced ROS increase or celldeath, antioxidants were able to neither block endogenousdhCer accumulation (Fig. 5D) nor substantially modify Cerlevels (data not shown) upon 4-HPR treatment. Of note, Troloxbut not Trolox-methylether rescued cells from 4-HPR-induceddeath (data not shown). Thus, we concluded that cellularROS increase was not a direct modulator of endogenousdhCer levels. On the other hand, clear discrepanciesemerged among the selected compounds between theirROS-buffering capacity and their activity against 4-HPR-mediated cytotoxicity.

Role of ROS in the antitumoral activity of 4-HPRAs mentioned above, prevention from 4-HPR-mediated

ROS accumulation was significant but partial with vitamin E(Fig. 5A) whereas this antioxidant showed sustained capacityto block cell death induced by 4-HPR (Figs. 5B–5C). Ofnote, opposite results were obtained with ascorbic acid(Figs. 5A–5C). CM-H2DCFDA-mediated analysis of ROSprovides information regarding several types of reactive spe-cies including hydrogen peroxide and peroxyl radicals. Never-theless, a more detailed study of the contribution of specificreactive species required the use of alternative approaches.One of the distinctive characteristics of vitamin E (a-

tocopherol in this study) is that due to its hydrophobicnature, vitamin E is especially potent against oxidative dam-age to lipidic environments as it plays a role as a peroxylradical scavenger that terminates chain reactions (Traberand Atkinson 2007). Of note, lipoxygenase (LOX) -medi-ated reactions are an important source of peroxyl radicals(Kühn and Borchert 2002). LOXs, and more specifically12-LOX, have been implicated in 4-HPR-driven cell deathin neuroblastoma cells (Lovat et al. 2002). Therefore, weanalyzed the capacity of 4-HPR to induce lipid peroxidationas well as the capacity of vitamin E and ascorbic acid tomodulate the mentioned reactive species. 4-HPR treatmentgenerated a clear dose- and time-dependent increase in lipidperoxidation of both cell lines (Supplementary Fig. 1E)1.Ascorbic acid and vitamin E were both able to prevent 4-HPR (6 h) -mediated lipid peroxidation (Fig. 6A). Nonethe-less, only vitamin E retained the capacity to buffer lipidperoxidation after 24 h of 4-HPR exposure (Fig. 6B). Thesedata suggested a role for lipid peroxides in the cytotoxicitydriven by 4-HPR.To further explore the possible implication of LOXs in cell

death, we selected the following 2 inhibitors of LOXs: baica-lein (12-LOX inhibitor) and NDGA (dose-dependent inhibitorof 5-, 12-, and 15-LOX). Based on literature, at least 5-LOX is expressed in the selected leukemia model (Jurkatcells) (Cook-Moreau et al. 2007).NDGA (but not baicalein) was able to effectively block

general oxidative stress (i.e., CM-H2DCFDA oxidation)driven by 4-HPR (3 µmol/L) in both cell lines (Fig. 7A). Inparticular, NDGA (but not baicalein) was able to protect cellsfrom 24 h of 4-HPR-induced lipid peroxidation (Fig. 7B) asobserved with vitamin E (Fig. 6B). Nevertheless, NDGA (aswell as baicalein; data not shown) did not prevent eitherCCRF-CEM or Jurkat cells from dying upon 4-HPR expo-sure (Figs. 7C–7D). Thus, these data suggest a role forLOXs (but not 12-LOX) in a 4-HPR-mediated oxidative envi-ronment but not in mediating the cytotoxic responses.In a previous study, we identified mitochondria as an im-

portant source of ROS upon 4-HPR treatment in our leuke-mia model (Asumendi et al. 2002). According to availabledata, the selected probe for general oxidative stress detection(CM-H2DCFDA) is not specific for mitochondrial superoxidedetection. Consequently, the contribution of superoxidewould have been ignored and possibly underestimated in theprevious assays. At this point, we wondered whether discrep-ancies among vitamin-E- and NDGA-mediated antiapoptoticactivity could be related to their capacity to buffer mitochon-drial superoxide species induced by 4-HPR. To address thisquestion, we performed a side by side analysis of the effect



Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

of both compounds on 4-HPR-driven mitochondrial superox-ide production. 4-HPR exposure for 2 h was selected basedon the superoxide production peak observed in CCRF-CEMand Jurkat cells at this time point (Supplementary Figs. 1C–1D).1 As observed upon 30 min 4-HPR exposure (Figs. 5Aand 7A), the capacity of NDGA to protect cells from generaloxidative stress (i.e., CM-H2DCFDA oxidation) was sustainedand higher than that exerted by vitamin E (Fig. 8A) even after6 h 4-HPR treatment (data not shown). Moreover, the data re-vealed almost complete prevention of 4-HPR-triggered mito-chondrial superoxide production with NDGA but a partialor null effect with vitamin E (Fig. 8B). Therefore, we con-clude that NDGA exerts a more potent antioxidant activitythan vitamin E against mitochondrial superoxide accumula-tion after 4-HPR treatment.

DiscussionThe goal of the current study was to establish the mecha-

nistic relationship between the 2 earliest events detected after

4-HPR treatment, the increase in the production of ROS andthe changes in the endogenous profiles of SLs. We alsowanted to determine the implication of these 2 early eventsin the strong apoptotic effect of 4-HPR in selected T-ALLcell lines (CCRF-CEM and Jurkat).4-HPR has been described to regulate multiple molecular

pathways (Lovat et al. 2004; Hail et al. 2006; Tiwari et al.2006; Kadara et al. 2007) and to induce several cellular proc-esses, including well characterized apoptosis (Delia et al.1993; Wu et al. 2001; Lovat et al. 2004; Hail et al. 2006),differentiation (Chen et al. 2003), and autophagy (Tiwari etal. 2008). The complexity and abundance of cellular net-works affected by 4-HPR makes the identification of keyevents related to each of the altered cellular processes diffi-cult.Focused on leukemia, a study by Delia and colleagues

(1995) was the first one suggesting a role for oxidative stressin cell death upon 4-HPR treatment. This work also describedthe capacity of Bcl-2 to delay or to decrease the apoptotic

Fig. 6. Effect of ascorbic acid (AA) and vitamin E (vitE) on lipid peroxidation following N-(4-hydroxyphenyl)retinamide (4-HPR) treatment.CCRF-CEM and Jurkat cells were incubated for 2 h with 100 µmol/L of AA or (vitE) and co-incubated with 4-HPR (3 µmol/L) for another6 h (A) or 24 h (B). Lipid peroxidation was determined by the exogenously added fluorescent undecanoic acid (BODIPY 581/591 C11) asdescribed in the Materials and methods. Data on the 24 h 4-HPR exposure are representative of 2 independent experiments.

Apraiz et al. 217


Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

effect of the retinoid that was also observed in later studiesby our group (Morales et al. 2007). Cer-mediated cytotoxic-ity in leukemia cells was first proposed by O’Donnell andcolleagues (2002). Later studies have further evaluated themechanisms following 4-HPR-mediated increase in ROS orCer. Mechanistically, several studies describe ROS-mediatedcleavage of PKC s and downstream inactivation of Mcl-1(Kang et al. 2008; Ruvolo et al. 2010), ROS-mediated de-phosphorylation of Akt (Cao et al. 2009), or endoplasmicreticulum stress and proteasome activation (Wang et al.2009) as important events leading to leukemia cell death.Nevertheless, none of these studied the possibility for anti-tumor activity in the absence of alterations on SLs or ROSproduction.In previous studies, our group demonstrated that in T-ALL

cell lines, 4-HPR induces intrinsic mitochondrial apoptoticcell death (Asumendi et al. 2002) and that acute cell deathwas characterized by an early increase in ROS and Cer levels(Morales et al. 2007). In the present work, additional dataprove that 4-HPR exerts therapeutically meaningful antitumoractivity (i.e., reduction of clonogenic capacity) at concentra-tions as low as 0.5 µmol/L that is not necessarily linked toacute toxicity as observed in data obtained from Jurkat cells.Therefore, the data support the potential application of fenre-tinide as a chemotherapeutic agent.SLs (especially Cer and gangliosides) have been largely

linked to 4-HPR-driven cytotoxicity (Maurer et al. 1999; Wuet al. 2001; Lovat et al. 2004; Rehman et al. 2004; Darwicheet al. 2005; Hail et al. 2006; Morales et al. 2007). Previousstudies based on thin layer chromatography pointed out anincrease in endogenous Cer as a major response in 4-HPR-mediated cell death (Maurer et al. 1999; Chen et al. 2003;Rehman et al. 2004; Hail et al. 2006; Morales et al. 2007;Tiwari et al. 2008). 4-HPR has been suggested to regulateendogenous Cer levels by activating serine palmitoyltransfer-ase-ceramide synthase (first enzymes of the de novo synthe-sis pathway) (Wang et al. 2001) or acidic-sphingomyelinase(Lovat et al. 2004). Recent studies that applied LC–MS tech-nology provided much more accurate detection of differentSL species and revealed a robust accumulation of dhCer (butnot Cer) in several cell lines following 4-HPR treatment(Zheng et al. 2006; Kraveka et al. 2007; Valsecchi et al.2010). Concomitant accumulation of dhSph and dhCer hasalso been recently described in 4-HPR treatment (Wang etal. 2008). Results from the in situ activity assay and LC–MSanalysis in our leukemia model demonstrated the rapid inhib-itory effect of 4-HPR on cellular DES activity and the result-ing time- and dose-dependent dhCer and, to a smaller extent,dhSph and Sph accumulation, in agreement with the data thatwas previously mentioned. Similarly, a recent study by Rah-

maniyan and colleagues (2011) identified DES as a direct invitro target of fenretinide.Despite the large number of studies linking the Cer in-

crease to 4-HPR-driven cell death, the current data regarding4-HPR on SL modulation raises the question about the realimplication of SL on observed biological effects. The factthat myriocin inhibited the accumulation of dhCer and freesphingoid bases but did not have any effect on 4-HPR-induced ROS production indicates that in our leukemiamodel SL modulation is not upstream of the observed ROSincrease. Moreover, myriocin did not inhibit 4-HPR-induceddeath. Hence, myriocin-based data prove that the describedaccumulation of SLs is not necessarily involved in the ob-served acute cytotoxicity. Nonetheless, Kraveka and col-leagues (2007) linked the inhibition of DES to cell cyclearrest. The data presented herein indicate that 4-HPR re-duces clonogenic capacity in conditions of significant dhCeraccumulation but absence of acute cytotoxicity. Therefore, arole for dhCer not in acute cytotoxicity but in proliferationcapacity could be feasible although our data suggest the in-volvement of a mechanism distinct from cell cycle arrest.Similarly to (dh)Cer accumulation, induction of oxidative

stress has often been described as an essential mediator in 4-HPR treatment (Suzuki et al. 1999; Asumendi et al. 2002;Kim et al. 2006; Kadara et al. 2007; Jiang et al. 2011) whilethe possible cause-effect relationship among 4-HPR modu-lated ROS and SLs profiles has often been discussed (Maureret al. 1999; Wu et al. 2001; Lovat et al. 2004; Hail et al.2006; Darwiche et al. 2007; Morales et al. 2007). Rehmanand colleagues (2004) described that the effect of 4-HPR on(dh)Cer could be partially reverted by N-acetyl-L-cysteinewhile Batra and colleagues (2004) did not observe differen-ces between 4-HPR-only and 4-HPR+N-acetyl-L-cysteinetreated cells. On the other hand, we have recently shown thatoxidative stress inhibits DES activity by an indirect mecha-nism (Idkowiak-Baldys et al. 2010). As mentioned above, inthis study, the effect of myriocin on 4-HPR treatment indi-cates that ROS production is not downstream of dhCer accu-mulation in cells. Moreover, the results of antioxidants on4-HPR-mediated SL modulation reveal that endogenousdhCer levels are not regulated (at least directly) by cellularROS levels; ROS production and DES inhibition may there-fore be considered distinct events upon 4-HPR exposure.Upon detailed analysis of studies focused on 4-HPR-triggered

oxidative stress, it becomes apparent that the sources andfunction of ROS in 4-HPR-exerted antitumoral activity arenot well defined. 12-LOX (Lovat et al. 2002), NADPH oxi-dase (Kadara et al. 2008), and mitochondria (Suzuki et al.1999; Asumendi et al. 2002; Kim et al. 2006; Darwiche etal. 2007; Cuperus et al. 2010) have been proposed as origins

Fig. 7. Effect of lypoxygenase (LOX) inhibitors on N-(4-hydroxyphenyl)retinamide (4-HPR)-driven oxidative stress and acute cytotoxicity. T-cell acute lymphoblastic leukemia (T-ALL) cells were incubated for 2 h with the indicated LOX inhibitors (baic, baicalein, 12-LOX inhibitor;nordihydroguaiaretic acid (NDGA), dose dependent 5-LOX > 12-LOX, 15-LOX inhibitor) or vitamin E (vitE; 100 µmol/L) prior to adding 4-HPR (3 µmol/L). Cellular oxidative stress status (A) was estimated by CM-H2DCFDA oxidation after 30 min 4-HPR treatment while lipidperoxidation (B) was analyzed by BODIPY 581/591 C11 in CCRF-CEM cells after 24 h exposure to 4-HPR. Cytotoxicity of selected treat-ments was determined by an XTT assay (C) and annexin V-FITC/propidium iodide (PI) labeling (D; CCRF-CEM cells) following 24 h 4-HPRexposure. Reactive oxygen species (ROS) and viability data were related to their corresponding control values. (A and C) Data represent themean ± SD of 3 independent experiments performed in quadruplicate (n = 12). **, p < 0.01 (relative to 4-HPR-treated cells); and p > 0.05,no statistical difference among the indicated groups. (B and D) Representative data of 2 independent experiments.



Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

for ROS following 4-HPR treatment. In this study, we per-formed an in-depth analysis of different types of ROS thatbrought to light clear discrepancies among the capacity to

buffer 4-HPR-mediated oxidative stress and the capacity toprotect from 4-HPR-driven cell death. Ascorbic acid,despite being able to completely block an early oxidative

Apraiz et al. 219


Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

environment, offered only a transient protection against celldeath following 4-HPR treatment. This could be linked toits inability to prevent later oxidative events. Vitamin E, onthe other hand, exerted a sustained protection from lipidperoxidation and limited capacity against mitochondrialsuperoxide but effectively protected from cytotoxicity medi-ated by 4-HPR. On the other hand, the LOX inhibitorNDGA (Salari et al. 1984) demonstrated a clear and sus-tained antioxidant capacity against all the tested reactivespecies but had no effect on the cytotoxic response. Bycontrast, baicalein was neither able to block ROS nor toprevent 4-HPR-driven leukemia cell death. All together, thepresent data indicate that LOXs (but not 12-LOX) could bethe source of the detected increase in ROS. Nevertheless,direct radical scavenger activity by NDGA cannot be ex-

cluded as it has been described as a potent in vitro scav-enger of a broad range of reactive radical species(Floriano-Sánchez et al. 2006). However, despite its signifi-cant capacity to buffer ROS, NDGA was clearly incapableof rescuing leukemia cells from 4-HPR-induced cell death.In this regard, Cuperus and colleagues (2010) also observedthat Trolox could just partially prevent 4-HPR-driven loss ofviability at conditions previously described to abrogate ROSproduction (Cuperus et al. 2008). Moreover, ATP produc-tion was also decreased in cells with no significant increaseof mitochondrial or general ROS production (e.g., FISKneuroblastoma cells) (Cuperus et al. 2010). These observa-tions point out the fact that 4-HPR-driven acute cytotoxicitymay occur even in the absence of ROS production, againstthe generalized idea.

Fig. 8. Side by side comparison of the antioxidant capacity of vitamin E (vitE) and nordihydroguaiaretic acid (NDGA) against general andmitochondrial increase of reactive oxygen species (ROS). CCRF-CEM cells were incubated for 2 h with the indicated concentrations of vitEor NDGA and N-(4-hydroxyphenyl)retinamide (4-HPR) was added that was followed by an additional 2 h of incubation in presence of anti-oxidants. General oxidative status (A) was estimated by CM-H2DCFDA while mitochondrial superoxide production (B) was determined byMitoSOX as described in the Materials and methods. Data are representative of 2 (A) or 3 (B) independent experiments.



Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

In conclusion, this study provides a clear differentiationbetween 2 major early events after 4-HPR treatment, the in-crease in ROS and the modulation of endogenous SL levels.It also offers a new view regarding the contribution of ROSand SLs to the antitumor effect of 4-HPR. We have shownthat 4-HPR induces a rapid DES inhibition resulting in astrong dhCer accumulation that is independent of the increasein cellular ROS. Moreover, we have shown that 4-HPR-mediated acute cytotoxicity may occur even in the absenceof dhCer or ROS accumulation. Data on dhSph and Sphalso keep them apart from 4-HPR-induced ROS or acutecell death in selected leukemia cell models. Characterizationof tumor cell death driving and nondriving events following4-HPR treatment is necessary to plan successful therapiesbased on drug combination or to understand the most likelysources of treatment failure.

AcknowledgementsThis work was conducted at the Lipidomics Core Facility

at MUSC that is supported by the National Institutes ofHealth, grant No. C06 RR018823 from the Extramural Re-search Facilities Program of the National Center for ResearchResources. The work was also partly supported by NIH grantCA087584. Also, technical and human support provided bySGIker (UPV/EHU, MICINN, GV/EJ, ERDF, and ESF) isgratefully acknowledged. Financial support was FPU grantNo. AP-2004-6497 from the Spanish Ministry of Scienceand Innovation, NIH grant No. CA97132, grant No. UFI11/44 from the University of the Basque Country, and grantNo. IT423-07 from the Basque Government.

ReferencesAsumendi, A., Morales, M.C., Alvarez, A., Aréchaga, J., and Pérez-

Yarza, G. 2002. Implication of mitochondria-derivated ROS andcardiolipin peroxidation in N-(4-hydroxyphenyl) retinamide-induced apoptosis. Br. J. Cancer, 86(12): 1951–1956. doi:10.1038/sj.bjc.6600356. PMID:12085192.

Batra, S., Reynolds, C.P., and Maurer, B.J. 2004. Fenretinidecytotoxicity for Ewing's sarcoma and primitive neuroectodermaltumor cell lines is decreased by hypoxia and synergisticallyenhanced by ceramide modulators. Cancer Res. 64(15): 5415–5424. doi:10.1158/0008-5472.CAN-04-0377. PMID:15289350.

Bielawski, J., Szulc, Z.M., Hannun, Y.A., and Bielawska, A. 2006.Simultaneous quantitative analysis of bioactive sphingolipids byhigh-performance liquid chromatography-tandem mass spectro-metry. Methods, 39(2): 82–91. doi:10.1016/j.ymeth.2006.05.004.PMID:16828308.

Cao, J., Xu, D., Wang, D., Wu, R., Zhang, L., Zhu, H., et al. 2009.ROS-driven Akt dephosphorylation at Ser-473 is involved in 4-HPR-mediated apoptosis in NB4 cells. Free Radic. Biol. Med. 47(5):536–547. doi:10.1016/j.freeradbiomed.2009.05.024. PMID:19482076.

Carpinteiro, A., Dumitru, C., Schenck, M., and Gulbins, E. 2008.Ceramide-induced cell death in malignant cells. Cancer Lett. 264(1):1–10. doi:10.1016/j.canlet.2008.02.020. PMID:18353542.

Chen, S., Samuel, W., Fariss, R.N., Duncan, T., Kutty, R.K., andWiggert, B. 2003. Differentiation of human retinal pigmentepithelial cells into neuronal phenotype by N-(4-hydroxyphenyl)retinamide. J. Neurochem. 84(5): 972–981. doi:10.1046/j.1471-4159.2003.01608.x. PMID:12603822.

Cook-Moreau, J.M., El-Makhour Hojeij, Y., Barrière, G., Rabino-vitch-Chable, H.C., Faucher, K.S., Sturtz, F.G., and Rigaud, M.A.

2007. Expression of 5-lipoxygenase (5-LOX) in T lymphocytes.Immunology, 122(2): 157–166. doi:10.1111/j.1365-2567.2007.02621.x. PMID:17484769.

Cuperus, R., Tytgat, G.A., Leen, R., Brites, P., Bras, J., Caron, H.N.,and Van Kuilenburg, A.B. 2008. Pleiotropic effects of fenretinidein neuroblastoma cell lines and multicellular tumor spheroids. Int.J. Oncol. 32(5): 1011–1019. PMID:18425327.

Cuperus, R., Leen, R., Tytgat, G.A., Caron, H.N., and vanKuilenburg, A.B. 2010. Fenretinide induces mitochondrial ROSand inhibits the mitochondrial respiratory chain in neuroblastoma.Cell. Mol. Life Sci. 67(5): 807–816. doi:10.1007/s00018-009-0212-2. PMID:19941060.

Darwiche, N., Abou-Lteif, G., Najdi, T., Kozhaya, L., Abou Tayyoun,A., Bazarbachi, A., and Dbaibo, G.S. 2005. Human T-celllymphopropic virus type I-transformed T-cells have a partialdefect in ceramide synthesis in response to N-(4-hydroxyphenyl)-retinamide. Biochem. J. 392(1): 231–239. doi:10.1042/BJ20050578. PMID:16086670.

Darwiche, N., Abou-Lteif, G., and Bazarbachi, A. 2007. Reactiveoxygen species mediate N-(4-hydroxyphenyl)retinamide-inducedcell death in malignant T cells and are inhibited by the HTLV-1oncoprotein Tax. Leukemia, 21(2): 261–269. doi:10.1038/sj.leu.2404472. PMID:17122865.

Delia, D., Aiello, A., Lombardi, L., Pelicci, P.G., Grignani, F.,Grignani, F., et al. 1993. N-(4-hydroxyphenyl)retinamide inducesapoptosis of malignant hemopoietic cell lines including thoseunresponsive to retinoic acid. Cancer Res. 53(24): 6036–6041.PMID:8261419.

Delia, D., Aiello, A., Formelli, F., Fontanella, E., Costa, A.,Miyashita, T., et al. 1995. Regulation of apoptosis induced bythe retinoid N-(4-hydroxyphenyl) retinamide and effect ofderegulated bcl-2. Blood, 85(2): 359–367. PMID:7811993.

Faderl, S., Lotan, R., Kantarjian, H.M., Harris, D., Van, Q., andEstrov, Z. 2003. N-(4-Hydroxylphenyl)retinamide (fenretinide, 4-HPR), a retinoid compound with antileukemic and proapoptoticactivity in acute lymphoblastic leukemia (ALL). Leuk. Res. 27(3):259–266. doi:10.1016/S0145-2126(02)00162-5. PMID:12537979.

Floriano-Sánchez, E., Villanueva, C., Medina-Campos, O.N., Rocha,D., Sánchez-González, D.J., Cárdenas-Rodríguez, N., and Ped-raza-Chaverrí, J. 2006. Nordihydroguaiaretic acid is a potent invitro scavenger of peroxynitrite, singlet oxygen, hydroxyl radical,superoxide anion and hypochlorous acid and prevents in vivoozone-induced tyrosine nitration in lungs. Free Radic. Res. 40(5):523–533. doi:10.1080/10715760500419365. PMID:16551579.

Garaventa, A., Luksch, R., Lo Piccolo, M.S., Cavadini, E., Montaldo,P.G., Pizzitola, M.R., et al. 2003. Phase I trial and pharmacoki-netics of fenretinide in children with neuroblastoma. Clin. CancerRes. 9(6): 2032–2039. PMID:12796365.

Hail, N., Jr, Kim, H.J., and Lotan, R. 2006. Mechanisms offenretinide-induced apoptosis. Apoptosis, 11(10): 1677–1694.doi:10.1007/s10495-006-9289-3. PMID:16850162.

Hannun, Y.A., and Obeid, L.M. 2008. Principles of bioactive lipidsignalling: lessons from sphingolipids. Nat. Rev. Mol. Cell Biol.9(2): 139–150. doi:10.1038/nrm2329. PMID:18216770.

Idkowiak-Baldys, J., Apraiz, A., Li, L., Rahmaniyan, M., Clarke, C.J.,Kraveka, J.M., et al. 2010. Dihydroceramide desaturase activity ismodulated by oxidative stress. Biochem. J. 427(2): 265–274.doi:10.1042/BJ20091589. PMID:20105137.

Jiang, L., Pan, X., Chen, Y., Wang, K., Du, Y., and Zhang, J. 2011.Preferential involvement of both ROS and ceramide in fenretinide-induced apoptosis of HL60 rather than NB4 and U937 cells.Biochem. Biophys. Res. Commun. 405(2): 314–318. doi:10.1016/j.bbrc.2011.01.036. PMID:21237137.

Kadara, H., Lacroix, L., Lotan, D., and Lotan, R. 2007. Induction of

Apraiz et al. 221


Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

endoplasmic reticulum stress by the pro-apoptotic retinoid N-(4-hydroxyphenyl)retinamide via a reactive oxygen species-dependentmechanism in human head and neck cancer cells. Cancer Biol.Ther. 6(5): 705–711. doi:10.4161/cbt.6.5.3963. PMID:17404501.

Kadara, H., Tahara, E., Kim, H.J., Lotan, D., Myers, J., and Lotan, R.2008. Involvement of Rac in fenretinide-induced apoptosis. CancerRes. 68(11): 4416–4423. PMID:18519704.

Kang, M.H., Wan, Z., Kang, Y.H., Sposto, R., and Reynolds, C.P.2008. Mechanism of synergy of N-(4-hydroxyphenyl)retinamideand ABT-737 in acute lymphoblastic leukemia cell lines: Mcl-1inactivation. J. Natl. Cancer Inst. 100(8): 580–595. doi:10.1093/jnci/djn076. PMID:18398104.

Kim, H.J., Chakravarti, N., Oridate, N., Choe, C., Claret, F.X., andLotan, R. 2006. N-(4-hydroxyphenyl)retinamide-induced apopto-sis triggered by reactive oxygen species is mediated by activationof MAPKs in head and neck squamous carcinoma cells.Oncogene, 25(19): 2785–2794. doi:10.1038/sj.onc.1209303.PMID:16407847.

Kitareewan, S., Spinella, M.J., Allopenna, J., Reczek, P.R., andDmitrovsky, E. 1999. 4HPR triggers apoptosis but not differentia-tion in retinoid sensitive and resistant human embryonal carcinomacells through an RARgamma independent pathway. Oncogene,18(42): 5747–5755. doi:10.1038/sj.onc.1202981. PMID:10523855.

Kitatani, K., Idkowiak-Baldys, J., and Hannun, Y.A. 2008. Thesphingolipid salvage pathway in ceramide metabolism andsignalling. Cell. Signal. 20(6): 1010–1018. doi:10.1016/j.cellsig.2007.12.006. PMID:18191382.

Kraveka, J.M., Li, L., Szulc, Z.M., Bielawski, J., Ogretmen, B.,Hannun, Y.A., et al. 2007. Involvement of dihydroceramidedesaturase in cell cycle progression in human neuroblastoma cells.J. Biol. Chem. 282(23): 16718–16728. doi:10.1074/jbc.M700647200. PMID:17283068.

Kühn, H., and Borchert, A. 2002. Regulation of enzymatic lipidperoxidation: the interplay of peroxidizing and peroxide reducingenzymes. Free Radic. Biol. Med. 33(2): 154–172. [Review.]doi:10.1016/S0891-5849(02)00855-9. PMID:12106812.

Lovat, P.E., Oliverio, S., Ranalli, M., Corazzari, M., Rodolfo, C.,Bernassola, F., et al. 2002. GADD153 and 12-lipoxygenasemediate fenretinide-induced apoptosis of neuroblastoma. CancerRes. 62(18): 5158–5167. PMID:12234979.

Lovat, P.E., Di Sano, F., Corazzari, M., Fazi, B., Donnorso, R.P.,Pearson, A.D., et al. 2004. Gangliosides link the acidicsphingomyelinase-mediated induction of ceramide to 12-lipox-ygenase-dependent apoptosis of neuroblastoma in response tofenretinide. J. Natl. Cancer Inst. 96(17): 1288–1299. doi:10.1093/jnci/djh254. PMID:15339967.

Maurer, B.J., Metelitsa, L.S., Seeger, R.C., Cabot, M.C., andReynolds, C.P. 1999. Increase of ceramide and induction of mixedapoptosis/necrosis by N-(4-hydroxyphenyl)retinamide in neuro-blastoma cell lines. J. Natl. Cancer Inst. 91(13): 1138–1146.doi:10.1093/jnci/91.13.1138. PMID:10393722.

Morales, M.C., Pérez-Yarza, G., Nieto-Rementería, N., Boyano, M.D.,Jangi, M., Atencia, R., and Asumendi, A. 2005. Intracellularglutathione levels determine cell sensitivity to apoptosis inducedby the antineoplasic agent N-(4-hydroxyphenyl)retinamide.Anticancer Res. 25(3B): 1945–1951. PMID:16158929.

Morales, M.C., Pérez-Yarza, G., Rementería, N.N., Boyano, M.D.,Apraiz, A., Gómez-Muñoz, A., et al. 2007. 4-HPR-mediatedleukemia cell cytotoxicity is triggered by ceramide-inducedmitochondrial oxidative stress and is regulated downstream byBcl-2. Free Radic. Res. 41(5): 591–601. doi:10.1080/10715760701218558. PMID:17454142.

O’Donnell, P.H., Guo, W.X., Reynolds, C.P., and Maurer, B.J. 2002.N-(4-hydroxyphenyl)retinamide increases ceramide and is cyto-

toxic to acute lymphoblastic leukemia cell lines, but not to non-malignant lymphocytes. Leukemia, 16(5): 902–910. doi:10.1038/sj.leu.2402485. PMID:11986953.

Obeid, L.M., Linardic, C.M., Karolak, L.A., and Hannun, Y.A. 1993.Programmed cell death induced by ceramide. Science, 259(5102):1769–1771. doi:10.1126/science.8456305. PMID:8456305.

Ponzoni, M., Bocca, P., Chiesa, V., Decensi, A., Pistoia, V.,Raffaghello, L., et al. 1995. Differential effects of N-(4-hydroxyphenyl)retinamide and retinoic acid on neuroblastomacells: apoptosis versus differentiation. Cancer Res. 55(4): 853–861. PMID:7850799.

Rahmaniyan, M., Curley, R.W., Jr, Obeid, L.M., Hannun, Y.A., andKraveka, J.M. 2011. Identification of dihydroceramide desaturaseas a direct in vitro target for fenretinide. J. Biol. Chem. 286(28):24754–24764. doi:10.1074/jbc.M111.250779. PMID:21543327.

Rehman, F., Shanmugasundaram, P., and Schrey, M.P. 2004.Fenretinide stimulates redox-sensitive ceramide production inbreast cancer cells: potential role in drug-induced cytotoxicity. Br.J. Cancer, 91(10): 1821–1828. doi:10.1038/sj.bjc.6602212. PMID:15505623.

Reynolds, C.P., and Lemons, R.S. 2001. Retinoid therapy ofchildhood cancer. Hematol. Oncol. Clin. North Am. 15(5): 867–910. doi:10.1016/S0889-8588(05)70256-2. PMID:11765378.

Rotmensz, N., De Palo, G., Formelli, F., Costa, A., Marubini, E.,Campa, T., et al. 1991. Long-term tolerability of fenretinide (4-HPR) in breast cancer patients. Eur. J. Cancer, 27(9): 1127–1131.doi:10.1016/0277-5379(91)90309-2. PMID:1835622.

Ruvolo, V.R., Karanjeet, K.B., Schuster, T.F., Brown, R., Deng, Y.,Hinchcliffe, E., and Ruvolo, P.P. 2010. Role for PKC d infenretinide-mediated apoptosis in lymphoid leukemia cells. J.Signal Transduct. 2010: 584657. PMID:20844597.

Sabichi, A.L., Lerner, S.P., Atkinson, E.N., Grossman, H.B.,Caraway, N.P., Dinney, C.P., et al. 2008. Phase III preventiontrial of fenretinide in patients with resected non-muscle-invasivebladder cancer. Clin. Cancer Res. 14(1): 224–229. doi:10.1158/1078-0432.CCR-07-0733. PMID:18172274.

Salari, H., Braquet, P., and Borgeat, P. 1984. Comparative effects ofindomethacin, acetylenic acids, 15-HETE, nordihydroguaiareticacid and BW755C on the metabolism of arachidonic acid inhuman leukocytes and platelets. Prostaglandins Leukot.Med. 13(1):53–60. doi:10.1016/0262-1746(84)90102-1. PMID:6424136.

Sun, S.-Y., Li, W., Yue, P., Lippman, S.M., Hong, W.K., and Lotan,R. 1999. Mediation of N-(4-hydroxyphenyl)retinamide-inducedapoptosis in human cancer cells by different mechanisms. CancerRes. 59(10): 2493–2498. PMID:10344763.

Suzuki, S., Higuchi, M., Proske, R.J., Oridate, N., Hong, W.K., andLotan, R. 1999. Implications of mitochondria-derived reactiveoxygen species, cytochrome C and caspase-3 in N-(4-hydroxy-phenyl)retinamide-induced apoptosis in cervical carcinoma cells.Oncogene, 18(46): 6380–6387. doi:10.1038/sj.onc.1203024.PMID:10597238.

Szulc, Z.M., Bielawski, J., Gracz, H., Gustilo, M., Mayroo, N.,Hannun, Y.A., et al. 2006. Tailoring structure-function andtargeting properties of ceramides by site-specific cationization.Bioorg. Med. Chem. 14(21): 7083–7104. doi:10.1016/j.bmc.2006.07.016. PMID:16919460.

Taha, T.A., Mullen, T.D., and Obeid, L.M. 2006. A house divided:ceramide, sphingosine, and sphinosine-1-phosphate in pro-grammed cell death. Biochim. Biophys. Acta, 1758(12): 2007–2036.

Tiwari, M., Kumar, A., Sinha, R.A., Shrivastava, A., Balapure, A.K.,Sharma, R., et al. 2006. Mechanism of 4-HPR-induced apoptosisin glioma cells: evidences suggesting role of mitochondrial-mediated pathway and endoplasmic reticulum stress. Carcinogen-



Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

esis, 27(10): 2047–2058. doi:10.1093/carcin/bgl051. PMID:16675469.

Tiwari, M., Bajpai, V.K., Sahasrabuddhe, A.A., Kumar, A., Sinha, R.A.,Behari, S., and Godbole, M.M. 2008. Inhibition of N-(4-hydroxyphenyl)retinamide-induced autophagy at a lower doseenhances cell death in malignant glioma cells. Carcinogenesis,29(3): 600–609. doi:10.1093/carcin/bgm264. PMID:18174255.

Traber, M.G., and Atkinson, J. 2007. Vitamin E, antioxidant andnothing more. Free Radic. Biol. Med. 43(1): 4–15. [Review.]doi:10.1016/j.freeradbiomed.2007.03.024. PMID:17561088.

Valsecchi, M., Aureli, M., Mauri, L., Illuzzi, G., Chigorno, V.,Prinetti, A., and Sonnino, S. 2010. Sphingolipidomics of A2780human ovarian carcinoma cells treated with synthetic retinoids. J.Lipid Res. 51(7): 1832–1840. doi:10.1194/jlr.M004010. PMID:20194109.

Van Veldhoven, P.P., and Bell, R.M. 1988. Effect of harvestingmethods, growth conditions and growth phase on diacylglycerollevels in cultured human adherent cells. Biochim. Biophys. Acta,959(2): 185–196. PMID:3349097.

Veronesi, U., Mariani, L., Decensi, A., Formelli, F., Camerini, T.,Miceli, R., et al. 2006. Fifteen-years results of a randomized phaseIII trial of fenretinide to prevente second breast cancer. Ann.Oncol. 17(7): 1065–1071. doi:10.1093/annonc/mdl047. PMID:16675486.

Wang, H., Maurer, B.J., Reynolds, C.P., and Cabot, M.C. 2001. N-(4-hydroxylphenyl)retinamide elevates ceramide in neuroblastomacell lines by coordinate activation of serine palmitoyltransferaseand ceramide synthase. Cancer Res. 61(13): 5102–5105. PMID:11431347.

Wang, H., Maurer, B.J., Liu, Y.Y., Wang, E., Allegood, J.C., Kelly,S., et al. 2008. N-(4-Hydroxyphenyl)retinamide increases dihy-droceramide and synergizes with dimethylsphingosine to enhancecancer cell killing. Mol. Cancer Ther. 7(9): 2967–2976. doi:10.1158/1535-7163.MCT-08-0549. PMID:18790777.

Wang, K., Fang, H., Xiao, D., Zhu, X., He, M., Pan, X., et al. 2009.Converting redox signaling to apoptotic activities by stress-responsive regulators HSF1 and NRF2 in fenretinide treatedcancer cells. PLoS ONE, 4(10): e7538. doi:10.1371/journal.pone.0007538. PMID:19844581.

Wu, J.M., DiPietrantonio, A.M., and Hsieh, T.C. 2001. Mechanism offenretinide (4-HPR)-induced cell death. Apoptosis, 6(5): 377–388.doi:10.1023/A:1011342220621. PMID:11483862.

Zheng, W., Kollmeyer, J., Symolon, H., Momin, A., Munter, E.,Wang, E., et al. 2006. Ceramides and other bioactive sphingolipidbackbones in health and disease: lipidomic analysis, metabolismand roles in membrane structure, dynamics, signaling andautophagy. Biochim. Biophys. Acta, 1758(12): 1864–1884.PMID:17052686.

Apraiz et al. 223


Bio

chem

. Cel

l Bio

l. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y Si

mon

Fra

ser

Uni

vers

ity o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Documents

Dihydroceramide accumulation and reactive oxygen species are distinct and nonessential events in 4-HPR-mediated leukemia cell death