Analytical Biochemistry 384 (2009) 166–169

Contents lists available at ScienceDirect

Analytical Biochemistry

journal homepage: www.elsevier .com/ locate /yabio

A secondary assay for ceramide kinase inhibitors based on cell growth inhibitionby short-chain ceramides

Christine Graf, Philipp Rovina, Frédéric Bornancin *

Novartis Institutes for BioMedical Research, A-1235 Vienna, Austria

a r t i c l e i n f o

Article history:Received 20 July 2008Available online 14 September 2008

Keywords:CeramideCeramide-1-phosphateCell deathVesicle sheddingNVP-231NVP-995

0003-2697/$ - see front matter � 2008 Elsevier Inc. Adoi:10.1016/j.ab.2008.09.008

* Corresponding author. Address: Novartis InstituForum 1, 4056 Basle, Switzerland. Fax: +41 61 32 405

E-mail address: frederic.bornancin@novartis.com (1 Abbreviations used: CerK, ceramide kinase; C1P,

ceramide; DMSO, dimethyl sulfoxide; cDNA, complememodified Eagle’s medium; FCS, fetal calf serum; MEMmedium; FACS, fluorescence-activated cell sorter.

a b s t r a c t

We recently reported that ectopic expression of ceramide kinase (CerK) in various cell lines increasestheir sensitivity to cell death induced by the exogenous addition of short-chain (e.g., C2) ceramides(Cer). Here we show that this higher sensitivity results from CerK catalytic activity and production ofC2-ceramide 1-phosphate (C2-C1P). If CerK activity is inhibited by the potent inhibitor NVP-231, C2-C1P is not produced and viability returns to control levels. The EC50 of NVP-231 in this assay is in thelow nanomolar range, consistent with the IC50 determined in activity assays in vitro using purified CerK.NVP-995, a structurally related but inactive compound, does not protect against C2-Cer-induced celldeath. This assay is robust and easy to implement and scale up, thereby providing a valuable secondaryscreen assay for CerK inhibitors.

� 2008 Elsevier Inc. All rights reserved.

Targeting ceramide kinase (CerK)1 with small molecule inhibi-tors has recently attracted interest because of involvement of CerKand/or its product, ceramide-1-phosphate (C1P), in inflammatoryprocesses and cell growth [1,2]. Some inhibitor compounds have al-ready been reported, and among these the diamino-benzothiazolederivative NVP-231 competes with ceramide (Cer) binding to CerKwith low nanomolar potency [3]. Identification of these compoundshas relied on activity assays either in vitro or in cellular systems [3–6]. There is to date no secondary throughput assay described asbeing capable of monitoring CerK activity and C1P formation usinga biological readout.

We created various cell lines that express human CerK. None ofthese lines had any growth advantage over control cells, and in factan increase in doubling time was observed [7]. Moreover, whenCerK-expressing cells were treated with short-chain ceramides,growth rates decreased more dramatically than in control cells[7]. Using this paradigm and NVP-231 as a tool compound, weinvestigated this further and have developed the first secondaryCerK assay for screening and characterizing cell-permeable CerKinhibitors.

ll rights reserved.

tes for BioMedical Research,62.F. Bornancin).

ceramide-1-phosphate; Cer,ntary DNA; DMEM, Dulbecco’s-E, Eagle’s medium essential

Materials and methods

Materials

C2-Cer was purchased from Cayman and prepared to 50 mM inabsolute ethanol. [32P]Orthophosphate (10 mCi/ml) was obtainedfrom Amersham Biosciences. Blasticidin was obtained from Invivo-Gen. Fumonisin B1, purchased from Sigma, was prepared to 10 mMin dimethyl sulfoxide (DMSO). NVP-231 and NVP-995 weresynthesized at Novartis Institutes for BioMedical Research. Bothcompounds were dissolved to 10 mM in DMSO.

Preparation and culture of stably transfected cells

CerK complementary DNA (cDNA) (GenBank accession no.AB079066) was obtained and subcloned into the pMono-Blasti vec-tor (InvivoGen) as described previously [7]. COS-1 cells were cul-tured in Dulbecco’s modified Eagle’s medium (DMEM)/10% fetalcalf serum (FCS) at 37 �C/5% CO2 in a humidified atmosphere.COS-1 cells stably expressing CerK (COS–CerK) were generated asdescribed and maintained in the presence of 3 lg/ml blasticidin[7]. Stable RBL-2H3 cells were generated using the same procedureand grown in 1� Eagle’s medium essential medium (MEM-E)/20%RPMI 1640/10% FCS at 37 �C/5% CO2.

Cell growth and cell death assays

For viability measurement, we used resazurin, a fluorescent re-dox indicator (Alamar Blue, Serotec). Cells were incubated with

Table 1NVP-231 protects COS–CerK against C2-Cer-induced cell death

Annexin V Annexin V and 7-AAD

Vehicle 1.7 ± 0.3 6.2 ± 1.0NVP-231 (50 nM) 1.1 ± 0.2 (ns) 1.1 ± 0.5***

NVP-995 (50 nM) 2.0 ± 0.4 (ns) 6.8 ± 0.9 (ns)

Note. COS and COS–CerK cells were seeded to a 24-well plate at 1.2 � 105 cells perwell, and 24 h later the serum content of the medium was reduced to 1% and cellswere incubated with vehicle (0.1% ethanol + 0.01% DMSO) or 10 lM C2-Cer for 9 hin the presence or absence of 50 nM NVP-231 or 50 nM NVP-995. After a wash inphosphate-buffered saline (PBS), cells were detached, washed again twice with coldPBS, resuspended in 1� binding buffer, and analyzed by FACS as described inMaterials and methods. Data represent the ratio of the percentage of positive cellsfor the markers indicated, of C2-Cer-treated cells versus vehicle-treated cells,obtained in triplicates ± standard deviations. Statistical analysis was performedwith a t test (***P < 0.001; ns, not significant).

Secondary assay for ceramide kinase inhibitors / C. Graf et al. / Anal. Biochem. 384 (2009) 166–169 167

10% (v/v) Alamar Blue at 37 �C for approximately 2 h. Fluorescencewas recorded (excitation 530 nm, emission 600 nm, cutoff 590 nm)in a SpectraMax Gemini XS fluorescence microplate reader (Molec-ular Devices). Apoptotic and necrotic cell death was analyzed byfluorescence-activated cell sorter (FACS) using the Annexin V PEApoptosis Detection Kit I (cat. no. 559767, Beckton Dickinson)according to the manufacturer’s instructions.

Viability screen assay development

COS-1 cells (0.33 � 104) and COS–CerK cells (0.40 � 104), previ-ously maintained in 3 lg/ml blasticidin-containing medium, wereseeded into wells of a 96-well plate (black with clear bottom, cat.no. 3603, Costar) and grown in the presence of 10% FCS. The 20%higher seeding density of COS–CerK cells takes into account theslower doubling rates of these cells compared with control cells[7]. The next day, the cell monolayers were rinsed with 1% FCS-containing medium and then incubated in quadruplicates in 1%FCS-containing medium with the indicated amounts of C2-Cerand compound (vehicle, fumonisin, NVP-231, NVP-995). After7 h, an equal volume of medium containing identical concentra-tions of C2-Cer and compounds was added and incubation wasprolonged up to 24 h. Afterward, an Alamar Blue solution wasadded to 10% final, incubation was allowed to proceed for 2 h,and the plate was read with a fluorescent plate reader (see above).

Results

C2-Cer is a potent inducer of cell death in CerK-expressing cells

In cell lines stably transfected with CerK, treatment with N-acetyl sphingosine (C2-Cer) is a potent inducer of cell death whencompared with control-transfected cells [7]. Typically, a 24-h treat-ment with 5 lM C2-Cer reduces growth by nearly 30% in controlcells, whereas reduction reaches 60% in COS–CerK cells (Fig. 1).That cell death is occurring was measured by FACS analysis (Table1) but is also clear from inspection of the cell culture dish, that is,the layer of COS–CerK cells displaying significant amounts ofrounded and floating cells at the end of the incubation period. In-creased cell death does not result from a deacylation/reacylationcycle that can produce long-chain ceramides—and thus long-chain

0

20

40

60

80

100

Viab

ility

(% v

s un

trea

ted)

COS COS-CerK

**

— + — + — +— — — — ++

C2-CerFumonisin

120

ns

Fig. 1. C2-Cer is a potent inducer of cell death in CerK-expressing cells. COS andCOS–CerK cells were seeded, grown, and treated as described in the ‘‘ViabilityScreen Assay Development” section of Materials and methods. Viability wasmeasured using Alamar Blue. Vehicle (�) (0.1% ethanol + 1% DMSO); Fumonisin B1(100 lM); C2-Cer (5 lM). Means of quadruplicates ± standard deviations areshown. Statistical analysis was performed with a t test (**P < 0.01; ns, notsignificant).

C1P—from C2-Cer [8]. Indeed, treatment with fumonisin B1, aninhibitor of ceramide synthase [9], failed to prevent the growthinhibitory effect of C2-Cer (Fig. 1, right). Therefore, the C2-Cer ef-fect on cell death is short-chain dependent. We found this assayto be robust, amenable to miniaturization, and thus useful for sec-ondary screening of CerK inhibitors, as shown below.

Phosphorylation of C2-Cer by CerK is responsible for the observedgrowth inhibitory effects—Reversal by NVP-231

Measuring residual viability as a function of increasing C2-Cerconcentrations showed that C2-Cer is at least threefold more po-tent on COS–CerK cells as compared with control COS cells [7].Remarkably, this increased sensitivity to C2-Cer in COS–CerK cellswas abrogated in the presence of 50 nM of the recently identifiedCerK inhibitor NVP-231 (Fig. 2A). In these cells, NVP-231 decreasedthe growth inhibitory effect of all C2-Cer concentrations tested,whereas it had no effect on control COS cells, consistent with thevery low CerK activity present in the latter [7]. The protective effectof 50 nM NVP-231 against C2-Cer-induced cell death in COS–CerKcells could also be exemplified by monitoring ongoing apoptosisand necrosis (Table 1). Similar findings were observed in CerK-expressing RBL-2H3 cells in the presence of 50 nM NVP-231(Fig. 2B).

C2-Cer permeates cells and is readily phosphorylated by CerK.We found that NVP-231 completely suppresses C2-Cer phosphory-lation (Fig. 2C). This strongly suggests that the growth inhibitoryeffect of C2-Cer in COS–CerK cells is dependent on production ofC2-C1P.

NVP-231 is equipotent in activity assays and in the C2-Cer cell-basedassay

The potency of NVP-231 in the C2-Cer cell growth assay wasinvestigated on treatment with 10 lM C2-Cer. This concentrationwas chosen based on the results shown in Fig. 2. NVP-231 wasremarkably potent in this assay, reaching nearly complete efficacyat 10 nM (Fig. 3). This finding is consistent with the low nanomolarIC50 previously characterized in vitro [3]. NVP-995, a weakly activederivative of NVP-231 with the benzamide aromatic ring substi-tuted by two methoxy groups (Fig. 3), had no effect up to 50 nM,also consistent with the micromolar potency measured in activityassays [3]. Accordingly, this analog had no effect on apoptosis/ne-crotic readouts (Table 1). The failure of NVP-995 to protect againstC2-Cer-mediated cell death, whereas NVP-231 does afford protec-tion, demonstrates the dependency of the growth inhibitory effectof C2-Cer on its phosphorylation by CerK. Altogether, these datashow the suitability of a C2-Cer-based survival assay for secondaryscreening and characterization of CerK inhibitors.

C2-Cer (µM)

COS-CerK

COS

— C2-c

erC2

-cer

+ N

VP-2

31

0

100

200

300

400

0.0 2.5 5.0 7.5 10.0 12.5 15.0

vehicleNVP-231

0

100

200

300

400

Viab

ility

(Ala

mar

Blu

e F.

U.)

0

250

500

750

1000

0

250

500

750

1000

0.0 2.5 5.0 7.5 10.0 12.5 15.0

RBL

RBL-CerK

A B

C

***

***

**

**

*** ** ** *****

***

Fig. 2. NVP-231 prevents C2-C1P formation and reverses the C2-Cer-induced growth inhibitory effect in COS–CerK and RBL–CerK cells. (A) COS and COS–CerK cells wereseeded, grown, and treated with graded concentrations of C2-Cer in the presence of vehicle (0.1% ethanol + 0.01% DMSO) or NVP-231 (50 nM), as described in the ‘‘ViabilityScreen Assay Development” section of Materials and methods. Viability was measured using Alamar Blue. Means of quadruplicates ± standard deviations are shown. Statisticalanalysis between NVP-231-treated samples and vehicle-treated samples at each C2-Cer concentration was performed with a t test (*P < 0.05, **P < 0.01, ***P < 0.001). F.U.,fluorescence units. (B) As in panel A with RBL-2H3 (RBL)- and CerK-expressing RBL-2H3 cells (RBL–CerK). (C) COS–CerK loaded ± C2-Cer (10 lM) ± NVP-231 (100 nM) in thepresence of [32P]orthophosphate (300 lCi/ml) for 2 h at 37 �C before lipid extraction and thin layer chromatography analysis, performed as published previously [3].

Compound (nM)

0

100

200

300

400

0 10 20 30 40 50

NVP-231NVP-995

Viab

ility

(Ala

mar

Blu

e F.

U.)

NNH

SNH

OO

NNH

SNH

OO

OMe

OMe

NVP-231IC50= 12 nM

NVP-995IC50 = 2 µM

*******

**

Fig. 3. NVP-231 is equipotent in CerK activity or C2-Cer-induced cell death assays. Left: Structure of NVP-231 and NVP-995 with their corresponding IC50 values determinedin activity assays using purified CerK and [32P]ATP [3]. Right: COS–CerK cells were seeded, grown, and treated with 10 lM C2-Cer in the presence of graded concentrations ofeither NVP-231 or NVP-995, as described in the ‘‘Viability Screen Assay Development” section of Materials and methods. Viability was measured using Alamar Blue. Means ofquadruplicates ± standard deviations are shown. Statistical analysis between NVP-231- and NVP-995-treated samples was performed with a t test (**P < 0.01, ***P < 0.001).F.U., fluorescence units.

168 Secondary assay for ceramide kinase inhibitors / C. Graf et al. / Anal. Biochem. 384 (2009) 166–169

Discussion

The finding that phosphorylation of C2-Cer by CerK enhancesrather than protects against cell death was unexpected. Treatmentof cells with exogenous C1P has long been recognized to stimulatecell proliferation [10,11]. Furthermore, there has been significantprogress to characterize the underlying signaling pathways [12–14]. We showed in this work that production of long-chain Cerafter deacylation and reacylation of C2-Cer cannot be responsiblefor the inhibitory effect of C2-Cer on COS–CerK cell growth becausefumonisin B1, a ceramide synthase inhibitor, did not have any ef-fect. Therefore, in this experimental setting, it is unlikely thatendogenous (long-chain) Cer concentrations became high enough

to drive COS–CerK cell death. Under different conditions, such aswhen glucosylceramide synthase is inhibited, these cells becomesensitive to inhibition of CerK by NVP-231 and the combined inhi-bition leads to effective growth inhibitory Cer levels [3]. Hence, theexplanation for the higher toxicity of C2-Cer in COS–CerK com-pared with COS cells may have to do with the biophysical proper-ties of this short-chain Cer analog and its accumulation inmembrane compartments. C2-Cer readily permeates cell mem-branes and reaches Cer metabolizing enzymes such as CerK[7,15]. Phosphorylation of C2-Cer by CerK will prevent its redistri-bution and lead to accumulation of C2-C1P in cytoplasmic mem-brane leaflets. This may lead to excessive vesicular budding andrelease into the medium, precluding normal cell growth. Thus,

Secondary assay for ceramide kinase inhibitors / C. Graf et al. / Anal. Biochem. 384 (2009) 166–169 169

the previously reported alterations in morphology and growth rateof COS–CerK [7] may in fact result from dysregulated membraneshape and vesicle budding rate in COS–CerK cells. A similar mech-anism was also proposed in a recent study by Van Overloop andVan Veldhoven [16].

Based on work with COS cells, this study showed that treatmentof CerK-expressing cells with C2-Cer and use of a viability readoutprovides a robust secondary screen assay for CerK inhibitors. Suchan assay was also developed in RBL-2H3 cells with similar results(Fig. 2B). This assay is not meant to substitute for activity assaysthat remain of primary importance for screening purposes. How-ever, it is easy to implement and reproduce, and it can be readilyminiaturized, thereby providing a valuable option for advancedcompound profiling. Importantly, the EC50 found in this assay forthe reference compounds we developed—the potent CerK inhibitorNVP-231 and its inactive analog NVP-995—is similar to the IC50

measured in vitro on purified CerK [3]. Therefore, we anticipatethat this assay may help in the identification and characterizationof CerK inhibitors that are sought after increasingly.

References

[1] N.F. Lamour, C.E. Chalfant, Ceramide-1-phosphate: The ‘‘missing” link ineicosanoid biosynthesis and inflammation, Mol. Interv. 5 (2005) 358–367.

[2] A. Gomez-Munoz, Ceramide 1-phosphate/ceramide, a switch between life anddeath, Biochim. Biophys. Acta 1758 (2006) 2049–2056.

[3] C. Graf, M. Klumpp, M. Habig, P. Rovina, A. Billich, T. Baumruker, B. Oberhauser,F. Bornancin, Targeting ceramide metabolism with a potent and specificceramide kinase inhibitor, Mol. Pharmacol. 74 (2008) 925–932.

[4] J.W. Kim, Y. Inagaki, S. Mitsutake, N. Maezawa, S. Katsumura, Y.W. Ryu, C.S.Park, M. Taniguchi, Y. Igarashi, Suppression of mast cell degranulation by anovel ceramide kinase inhibitor, the F-12509A olefin isomer K1, Biochim.Biophys. Acta 1738 (2005) 82–90.

[5] N. Munagala, S. Nguyen, W. Lam, J. Lee, A. Joly, K. McMillan, W. Zhang,Identification of small molecule ceramide kinase inhibitors using a

homogeneous chemiluminescence high throughput assay, Assay Drug Dev.Technol. 5 (2007) 65–73.

[6] A.S. Don, H. Rosen, A fluorescent plate reader assay for ceramide kinase, Anal.Biochem. 375 (2008) 265–271.

[7] C. Graf, P. Rovina, L. Tauzin, A. Schanzer, F. Bornancin, Enhanced ceramide-induced apoptosis in ceramide kinase overexpressing cells, Biochem. Biophys.Res. Commun. 354 (2007) 309–314.

[8] B. Ogretmen, B.J. Pettus, M.J. Rossi, R. Wood, J. Usta, Z. Szulc, A. Bielawska, L.M.Obeid, Y.A. Hannun, Biochemical mechanisms of the generation of endogenouslong chain ceramide in response to exogenous short chain ceramide in theA549 human lung adenocarcinoma cell line: Role for endogenous ceramide inmediating the action of exogenous ceramide, J. Biol. Chem. 277 (2002) 12960–12969.

[9] A.H. Merrill Jr.,, G. van Echten, E. Wang, K. Sandhoff, Fumonisin B1 inhibitssphingosine (sphinganine) N-acyltransferase and de novo sphingolipidbiosynthesis in cultured neurons in situ, J. Biol. Chem. 268 (1993) 27299–27306.

[10] A. Gomez-Munoz, P.A. Duffy, A. Martin, L. O’Brien, H.S. Byun, R. Bittman, D.N.Brindley, Short-chain ceramide-1-phosphates are novel stimulators of DNAsynthesis and cell division: antagonism by cell-permeable ceramides, Mol.Pharmacol. 47 (1995) 833–839.

[11] A. Berger, R. Bittman, R.R. Schmidt, S. Spiegel, Structural requirements ofsphingosylphosphocholine and sphingosine-1-phosphate for stimulation ofactivator protein-1 activity, Mol. Pharmacol. 50 (1996) 451–457.

[12] A. Gomez-Munoz, J.Y. Kong, K. Parhar, S.W. Wang, P. Gangoiti, M. González, S.Eivemark, B. Salh, V. Duronio, U.P. Steinbrecher, Ceramide-1-phosphatepromotes cell survival through activation of the phosphatidylinositol 3-kinase/protein kinase B pathway, FEBS Lett. 579 (2005) 3744–3750.

[13] P. Gangoiti, M.H. Granado, S.W. Wang, J.Y. Kong, U.P. Steinbrecher, A. Gomez-Munoz, Ceramide 1-phosphate stimulates macrophage proliferation throughactivation of the PI3-kinase/PKB, JNK, and ERK1/2 pathways, Cell. Signal. 20(2008) 726–736.

[14] P. Gangoiti, M.H. Granado, L. Arana, A. Ouro, A. Gomez-Munoz, Involvement ofnitric oxide in the promotion of cell survival by ceramide 1-phosphate, FEBSLett. 582 (2008) 2263–2269.

[15] H. Van Overloop, Y. Denizot, M. Baes, P.P. Van Veldhoven, On the presence ofC2-ceramide in mammalian tissues: Possible relationship toetherphospholipids and phosphorylation by ceramide kinase, Biol. Chem.388 (2007) 315–324.

[16] H. Van Overloop, P.P. Van Veldhoven, Ceramide-dependent release of ceramidekinase from cultured cells, Biochem. Biophys. Res. Commun. 364 (2007) 169–174.