FEBS Letters 581 (2007) 2641–2646

Evidence for a role of glycosphingolipids in CXCR4-dependentcell migration

Cristina Limatolaa,b, Valeria Massac,1, Clotilde Lauroa, Myriam Catalanoa, Anna Giovanettid,Silvia Nuccitellic, Angelo Spinedic,*

a Istituto Pasteur-Fondazione Cenci Bolognetti, Dipartimento di Fisiologia Umana e Farmacologia, Centro di Eccellenza BEMM,Universita di Roma ‘Sapienza’, Roma, Italy

b Neuromed I.R.C.C.S., Pozzilli, Italyc Dipartimento di Biologia, Universita di Roma ‘Tor Vergata’, Via della Ricerca Scientifica 1, I-00133 Roma, Italy

d Istituto di Radioprotezione, ENEA CR Casaccia, Roma, Italy

Received 16 April 2007; accepted 2 May 2007

Available online 11 May 2007

Edited by Sandro Sonnino

Abstract Chemotaxis induction is a major effect evoked bystimulation of the chemokine receptor CXCR4 with its sole li-gand CXCL12. We now report that treatment of CHP-100 hu-man neuroepithelioma cells with the glucosylceramide synthase(GCS) inhibitor DL-threo-1-phenyl-2-hexadecanoylamino-3-pyrrolidino-1-propanol inhibits CXCR4-dependent chemotaxis.We provide evidence that the phenomenon is not due to unspecificeffects of the inhibitor employed and that inhibition of GCS nei-ther affects total or plasmamembrane CXCR4 expression, norCXCL12-induced Ca2+ mobilization. The effects of the GCSinhibitor on impairment of CXCL12-induced cell migration tem-porally correlated with a pronounced downregulation of neutralglycosphingolipids, particularly glucosylceramide, and with a de-layed and more moderate downregulation of gangliosides; more-over, exogenously administered glycosphingolipids allowedresumption of CXCR4-dependent chemotaxis. Altogether our re-sults provide evidence, for the first time, for a role glyco-sphingolipids in sustaining CXCL12-induced cell migration.� 2007 Federation of European Biochemical Societies. Publishedby Elsevier B.V. All rights reserved.

Keywords: CXCR4; Chemotaxis; Glycosphingolipids;Neuroepithelioma

1. Introduction

Activation of the chemokine receptor CXCR4 by its unique

physiological ligand CXCL12 (stromal cell derived factor-1,

SDF-1) evokes pleiotropic effects on a broad range of target

tissues and fulfils a variety of functions both during develop-

Abbreviations: PPPP, DL-threo-1-phenyl-2-hexadecanoylamino-3-pyrrolidino-1-propanol; GCS, glucosylceramide synthase; GSL,glycosphingolipid; GlcCer, glucosylceramide; GalCer, galactosylcera-mide; LacCer, lactosylceramide; Gb3, globotriaosylceramide; C8-GlcCer, DD-Glucosyl-b1-1 0-N-octanoyl-DD-erythro-sphingosine; C8-Lac-Cer, DD-Lactosyl-b1-10-N-octanoyl-DD-erythro-sphingosine; HPTLC,high-performance thin layer chromatography; PBS, phosphate buf-fered saline; FACS, fluorescence-activated cell sorting

*Corresponding author. Fax: +39 062023500.E-mail address: spinedi@uniroma2.it (A. Spinedi).

1Present address: Dipartimento di Neurogenetica Molecolare, IstitutoNeurologico Nazionale ‘Carlo Besta’, Milano, Italy.

0014-5793/$32.00 � 2007 Federation of European Biochemical Societies. Pu

doi:10.1016/j.febslet.2007.05.003

ment and adulthood [1]. A prominent effect brought about

by CXCR4 activation is chemotaxis induction: by this mecha-

nism, activation of the CXCL12/CXCR4 axis is involved, for

instance, in proper homing of haemato-/lympho-poietic cells

[2], in the cytoarchitectural maturation of certain districts of

the central nervous system [2,3], as well as in mobilization of

cellular elements during the inflammatory response [4]. Re-

cently, an important role for the CXCL12/CXCR4 axis has

also been proposed in connection with tumor progression: in

fact, there is evidence that expression of CXCR4 by tumor

cells may direct their metastatization to CXCL12 producing

sites [5–7].

CXCR4 is a seven-transmembrane-domain receptor coupled

to heterotrimeric proteins of the Gi family and its engagement

by CXCL12 evokes activation of a multiplicity of signalling

pathways: among these, activation of phosphoinositide 3-ki-

nase and/or extracellular signal-regulated kinases have been

shown to provide key events for chemotaxis induction in a

variety of cell systems [3,5,8–10]. Notably, evidence has been

provided that signalling through CXCR4 may be sensitive to

the receptor lipid environment. In particular, it has been

shown that cholesterol may be crucial for maintaining a recep-

tor conformation with high affinity for its ligand [11]; more-

over, it has been reported that CXCR4 signalling, at least in

the context of chemotaxis induction, may require receptor

recruitment into lipid rafts [12].

Glycosphingolipids (GSLs) are major components of cell

membranes and of lipid rafts as well; nevertheless, at present,

the issue of possible functional relationships between the

CXCR4/CXCL12 axis and GSLs, in the context of cell migra-

tion, has been fairly disregarded. This issue has been herein ad-

dressed having as an experimental model the human

neuroepithelioma cell line CHP-100 [13]: CXCR4, in fact, is

constitutively expressed by CHP-100 cells [14] and, in this sys-

tem, it has been reported to mediate CXCL12-induced chemo-

taxis [8].

2. Materials and methods

2.1. MaterialsMaterial for cell culture, Sephadex G-50 and all natural lipids were

from Sigma. DL-threo-1-phenyl-2-hexadecanoylamino-3-pyrrolidino-1-propanol (PPPP) was from Calbiochem–Novachem. DD-Glucosyl-

blished by Elsevier B.V. All rights reserved.

2642 C. Limatola et al. / FEBS Letters 581 (2007) 2641–2646

b1-1 0-N-octanoyl-DD-erythro-sphingosine (C8-GlcCer) and DD-Lactosyl-b1-1 0-N-octanoyl-DD-erythro-sphingosine (C8-LacCer) were from Avan-tiPolar. Silica gel 60 high-performance thin layer chromatography(HPTLC) plates were from Merck. CXCL12 was from Peprotech.The anti-CXCR4 antibody (C20) was from Santa Cruz Biotechnology.The R-phycoerythrin-conjugated mouse anti-human CXCR4 mono-clonal antibody was from Pharmingen. The Apotest FITC Kit wasfrom Nexins Research B.V. The chemiluminescence ECL detectionsystem was from Amersham. Fura-2/AM was from Molecular Probes.

2.2. Cell cultureCells were grown at 37 �C in RPMI 1640 medium, supplemented

with 10% heat-inactivated fetal calf serum, 2 mM glutamine, 100i.u./ml penicillin and 100 lg/ml streptomycin, in a humidified atmo-sphere with 5% CO2. PPPP, C8-GlcCer and C8-LacCer were adminis-tered to cells from dimethylsulfoxide stock solutions (1 mM, 40 mMand 40 mM, respectively); CXCL12 and ganglioside GM2 were admin-istered from aqueous stock solutions (12.5 lM and 50 mM, respec-tively).

2.3. Lipid analysisControl and PPPP-treated cells were washed twice with phosphate

buffered saline (PBS), pH 7.4, trypsinized and pelleted at 600 · g for5 min at 4 �C. Lipids were extracted by the method of Folch et al.[15]. For ganglioside separation, the Folch’s upper phase was dried un-der vacuum, redissolved in distilled water and desalted by gel filtration.Gangliosides recovered in the void volume were dried under vacuum,resuspended in chloroform/methanol (2:1, v/v), spotted onto HPTLCplates, developed in chloroform/methanol/0.22% CaCl2 (60:40:9, v/v/v) and visualized by staining with the resorcinol reagent [16]. For anal-ysis of neutral GSLs and ceramide, the Folch’s lower phase was dried,lipids resuspended in 0.1 M methanolic KOH, subjected to mild alka-line hydrolysis for 1 h at 37 �C and re-extracted by the Folch’s method.Ceramide was resolved by developing samples in chloroform/aceticacid (9:1, v/v) and visualized, after charring, with 3% cupric acetatein 8% phosphoric acid [17]. For separation of neutral GSLs, sampleswere developed in chloroform/methanol/water (65:25:4, v/v/v). In someanalyses, to achieve a better resolution between glucosyl- and galacto-syl-ceramide, neutral GSLs were resolved on HPTLC plates dipped in2.5% (w/v) methanolic boric acid and activated by heating at 110 �Cfor 30 min [18]. Neutral GSLs were visualized by the a-naphthol re-agent [16]. For phospholipid analysis, aliquots of the Folch’s lowerphase were developed in chloroform/methanol/acetic acid/water(60:27:8:2, v/v/v/v). Cholesterol was separated by developing aliquotsof the Folch’s lower phase in petroleum ether/diethyl ether (96:4,v/v) for full plate length and, after drying, for half-length in the samedimension, in petroleum ether/diethyl ether/formic acid (80:20:2, v/v/v). Both phospholipids and cholesterol were visualized, after charring,with 3% cupric acetate in 8% phosphoric acid [17]. Lipid quantitationwas performed by densitometric analysis, after HPTLC separation andspot visualization as described above, using a computerized image pro-cessing system (Gel-Pro Analyzer) and, when necessary, referring toknown amounts of authentic lipid standards developed along withsamples.

2.4. Western Blot analysisTo monitor CXCR4 expression, control or PPPP-treated cells were

washed with PBS and directly lysed with hot SDS sample buffer. Equalamounts of protein [19] from each sample were processed and analyzedas previously reported [8].

2.5. Chemotaxis assayControl and PPPP-treated cells were trypsinized, washed twice with

RPMI 1640 and plated (500000/ well) on 12 mm transwells (8 lm poresize filters) in RPMI 1640, enriched with 0.5 lM PPPP, in the case ofPPPP-pretreated cells, and containing or not the glucosylceramide syn-thase (GCS) inhibitor at the above-reported concentration, in the caseof control cells, as indicated in the text. C8-GlcCer (40 lM) or C8-Lac-Cer (40 lM) or ganglioside GM2 (50 lM) were also added to the med-ium, when indicated. The lower chambers contained RPMI 1640medium enriched with 60 nM CXCL12 or vehicle (water). After 5 hof incubation at 37 �C, cells were treated with ice-cold 10% trichloro-acetic acid for 5 min. Cells adhering to the upper side of the filter werescraped off, while cells on the lower side were stained with a solution

containing 50% isopropanol, 1% formic acid and 0.5% (w/v) brilliantblue R 250. Stained cells were counted in more than 20 fields with a40· objective.

2.6. Analyses by fluorescence-activated cell sorting (FACS)For evaluating membrane CXCR4 expression, control and PPPP-

treated cells were washed twice with RPMI 1640, trypsinized and cen-trifuged at 600 · g for 5 min at 4 �C. The pellets were washed twicewith PBS, resuspended in 1 ml of fluorescence-activated cell sorting(FACS) buffer (5% fetal bovine serum and 0.1% sodium azide inPBS) and subdivided in two aliquots of 0.5 ml each. The two aliquotswere centrifuged at 600 · g for 5 min at 4 �C; thereafter, one samplewas resuspended in 15 ll of FACS buffer containing the phycoery-thrin-conjugated anti-CXCR4 antibody diluted 1:11, and the otherin 15 ll of FACS buffer alone (for autofluorescence estimation). Afterincubation at 4 �C for 30 min, cells were pelleted, washed twice andresuspended in FACS buffer. Samples were analyzed on a FACSCali-bur (Becton-Dickinson, USA), with a CellQuest software. Cell viabilitywas estimated by dual staining with annexin V-FITC and propidiumiodide, using a commercially available Kit (Apotest FITC), accordingto the manufacturer’s instructions.

2.7. Ca2+ measurementsControl and PPPP-treated cells were incubated for 45 min at 37 �C

in RPMI 1640 containing the membrane permeant dye Fura-2/AM(4 lM). Fluorescence determinations were made using a conventionalsystem driven by Axon Imaging Workbench software (Axon Instr.).All optical parameters and digital camera settings were maintainedthroughout, to avoid non-homogeneous data. The changes in freecytosolic Ca2+ concentration ([Ca2+]i) were expressed as DR (i.e. theratio between the images acquired with two excitation wavelengths,340 nm and 380 nm, monitoring emission at 510 nm).

3. Results

3.1. Long-term treatment of CHP-100 cells with PPPP induces

general GSL depletion and inhibits CXCL12-induced cell

migration but not Ca2+ mobilization

As a first approach to our study, CHP-100 cells were cul-

tured for 5 d in the presence of the specific GCS inhibitor

PPPP (0.5 lM) [20], in order to achieve general GSL depletion;

the chemotactic response of GSL-depleted cells to CXCR4 was

then monitored and compared to that of control cells.

As shown in Fig. 1A, control CHP-100 cells displayed gan-

glioside species comigrating with GM3, GM2, GM1 and

GD1a standards (band multiplicity and slight migration shifts

likely reflect heterogeneous length of the ceramide mojeties);

notably, in cells cultured for 5 d in the presence of PPPP, the

levels of all the above-mentioned ganglioside species fell below

the detection limit of the staining method. Fig. 1B shows that

exposure of CHP-100 cells to 0.5 lM PPPP for 5 d also in-

duced a dramatic downregulation of the levels of major neutral

GSL species, namely glucosylceramide (GlcCer), lactosylcera-

mide (LacCer) and globotriaosylceramide (Gb3); moreover,

sustained GCS inhibition did not redirect GSL metabolism to-

ward galactosylceramide (GalCer) synthesis, as conversely ob-

served in other cell systems [21] (Fig. 1B).

Consistently with previous findings [20,22], we found that

ceramide levels were not significantly increased in PPPP-trea-

ted cells (0.95 ± 0.30 lg/mg protein, n = 3) as compared to

controls (0.59 ± 0.14 lg/mg protein, n = 3); this was apparently

due to ceramide utilization for synthesis of sphingomyelin,

whose levels rose from 8.4 ± 0.7 lg/mg protein (n = 3), as ob-

served in control cells, to 20.1 ± 3.6 lg/mg protein (n = 3) in

PPPP-treated cells (P < 0.01 by Student’s t test). Most notably,

Fig. 1. Long-term treatment with PPPP downregulates GSLs withoutaffecting total or plasmamembrane CXCR4 expression. (A) Aliquotsof the Folch’s upper phases, corresponding to equal amounts ofprotein, from control CHP-100 cells (lane 2) or from cells cultured for5 d in the presence of 0.5 lM PPPP (lane 3), were analyzed forgangliosides. Lane 1 shows the migration of authentic gangliosidestandards. Ganglioside nomenclature is that of Svennerholm [26]. (B)Aliquots of the Folch’s lower phases corresponding to equal amountsof protein, from control CHP-100 cells (lane 1) or from cells culturedfor 5 d in the presence of 0.5 lM PPPP (lane 2), were analyzed forneutral GSLs. Migration of GalCer and GlcCer standards are shownin lanes 3 and 4, respectively; lane 5 shows the migration of LacCerand Gb3 standards. (C) Equal amounts of protein from lysates ofcontrol CHP-100 cells or cells grown for 5 d in the presence of 0.5 lMPPPP were analyzed by Western Blot for CXCR4 expression. (D)Surface CXCR4 expression, as monitored by FACS analysis, incontrol CHP-100 cells (upper panel) or cells grown for 5 d in thepresence of 0.5 lM PPPP (lower panel), using a phycoerythrin-conjugated anti-CXCR4 antibody (a-CXCR4-PE), as reported in thetext. Cell autofluorescence is also shown (Baseline).

Fig. 2. GSL depletion inhibits CXCL12-induced cell migration but notCa2+ mobilization. CHP-100 cells, either left untreated (None) orpretreated for 5 d with 0.5 lM PPPP (PPPP) were subjected tochemotaxis assay either in the absence (Vehicle) or presence of a 60 nMCXCL12 gradient (CXCL12). PPPP (0.5 lM) was always present inthe medium of PPPP-pretreated cells, whereas migration of untreatedCHP-100 cells was performed either in the absence (�) or presence (+)of 0.5 lM PPPP. Data are normalized setting at 100 the number ofcontrol CHP-100 cells migrating in the absence of both CXCL12 andPPPP in the assay medium (basal) and are expressed as means and S.D.from four different experiments. Statistical significance is indicated byasterisks (P < 0.01, as compared to basal, after Student’s t test). (B)Representative Ca2+ transients induced by addition of 60 nM CXCL12(arrows) in CHP-100 cells either left untreated (Control) or pretreatedfor 5 d with 0.5 lM PPPP (PPPP) (C) Histograms of the maximalamplitude of Ca2+ transients induced by 60 nM CXCL12 in CHP-100cells either left untreated (Control) or pretreated for 5 d with 0.5 lMPPPP (PPPP). Data are means and S.D. from seven different deter-minations.

C. Limatola et al. / FEBS Letters 581 (2007) 2641–2646 2643

long-term treatment with PPPP had no secondary effects on

the levels of cellular cholesterol, whose downregulation has

been reported to inhibit CXCR4-dependent chemotaxis [12]:

cholesterol was in fact estimated 30.9 ± 1.6 lg/mg protein in

control cells (n = 3) and. 30.5 ± 1.6 lg/mg protein in GSL-de-

pleted cell (n = 3).

We found that treatment of CHP-100 cells with 0.5 lM

PPPP for 5 d had no effect on their viability and gross mor-

phology (not shown); moreover, general GSL depletion did

not affect total (Fig. 1C) or plasmamembrane (Fig. 1D)

CXCR4 levels. On the basis of these results, we tested the abil-

ity of control CHP-100 cells and cells treated for 5 d with

0.5 lM PPPP to migrate along a 60 nM CXCL12 gradient,

over a 5-h period. Fig. 2A shows that the extent of basal (i.e.

chemokine-independent) migration was the same in control

CHP-100 cells and the GSL-depleted counterpart. However,

while exposure to the CXCL12 gradient fairly doubled the

amount of migrating control cells, as compared to basal,

CXCL12-induced migration was fully suppressed in GSL-de-

pleted cells. To prevent GSL-depleted cells from resuming

GSL synthesis, their migration test had to be performed in

the presence of 0.5 lM PPPP. Therefore, a further control

had to be introduced, namely CHP-100 cells not previously ex-

posed to the GCS inhibitor but migrating in the presence of

0.5 lM PPPP. As shown in Fig. 2A, the presence of PPPP

did not affect CXCL12-induced migration in CHP-100 cells

not previously depleted of GSLs.

A signalling pathway associated with CXCR4 stimulation,

but only seldomly related to chemotaxis induction [5], involves

Ca2+ mobilization driven by phospholipase C-b activation.

Therefore, to address the possibility that GSL depletion might

generally compromise CXCR4 function, we studied whether

general GSL downregulation also affected [Ca2+]i variations

induced by CXCL12 (60 nM). In this respect, no significant

difference in the number of responsive elements was found be-

tween control CHP-100 cells (66 ± 13%, n = 25) and the GSL-

depleted counterpart (52 ± 1.6%, n = 27). Moreover, neither

the time course of CXCL12-induced Ca2+ transients

(Fig. 2B), nor their amplitude (Fig. 2C) were affected by

GSL depletion.

Fig. 3. Dependence of inhibition of CXCL12-induced chemotaxis andGSL downregulation on the time of cell pretreatment with PPPP. (A)Untreated CHP-100 cells or cells preincubated for the indicated timeswith 0.5 lM PPPP were subjected to chemotaxis assay, either in theabsence or presence of a 60 nM CXCL12 gradient, in media containing0.5 lM PPPP. The chemotactic index (i.e. the ratio of the number ofcells migrated in the presence of CXCL12 to the number of cellsmigrated in the absence of CXCL12) was calculated for each time ofpreincubation with PPPP. Data are means and S.D. from five differentexperiments. Statistical significance is indicated by asterisks (P < 0.01,as compared to cells not preincubated with PPPP, after Student’s ttest). (B) Relative changes of neutral GSL levels, as detected bydensitometric analysis after HPTLC separation, in cells incubated with0.5 lM PPPP for the indicated times. Data are means and S.D. fromfour different determinations. Statistical significance is indicated byasterisks (P < 0.01, as compared to cells not preincubated with PPPP,after Student’s t test). (C) Relative changes of ganglioside levels, asdetected by densitometric analysis after HPTLC separation, in cellsincubated with 0.5 lM PPPP for the indicated times. S.D.s, not shownfor sake of clarity, never exceeded 15% of the mean values, as obtainedfrom four different determinations. Statistically significant decrease ofthe levels of all ganglioside species, as observed after 24 h of cellincubation with PPPP, is indicated by the asterisk after the bracket(P < 0.01, as compared to cells not preincubated with PPPP, afterStudent’s t test).

2644 C. Limatola et al. / FEBS Letters 581 (2007) 2641–2646

3.2. Temporal relationship between PPPP-induced inhibition of

CXCL12-induced chemotaxis and GSL downregulation

To assess whether general GSL depletion, as observed after

5 d of cell treatment with PPPP, was a stringent requirement

for inhibition of CXCR4-dependent chemotaxis, in a subse-

quent set of experiments CHP-100 cells were preincubated

with 0.5 lM PPPP for shorter times, prior to test their ability

to migrate along a 60 nM CXCL12 gradient. Remarkably, we

found that inhibition of CXCR4-dependent cell migration

started to be observed after relatively short times of pretreat-

ment with the GCS inhibitor. In this respect, Fig. 3A shows

that attenuation of CXCR4-dependent chemotaxis was

clearly observed after 4–5 h of cell pretreatment with the

GCS inhibitor. Moreover, preincubation with 0.5 lM PPPP

for 10 h already produced a pronounced inhibition of

CXCL12-induced cell migration, that turned maximal by

24 h (Fig. 3A).

Based on the results reported above, we studied the kinetics

of GSL downregulation caused by cell exposure to the GCS

inhibitor over a 24 h period. Densitometric analysis after

HPTLC separation indicated that GlcCer level was rapidly

and dramatically downregulated within the time-period exam-

ined, already decreasing to a 30% of its initial value after cell

exposure for 10 h to PPPP, and to about 15% after 24 h

(Fig. 3B). LacCer and Gb3 were also markedly downregulated,

albeit less pronouncedly than GlcCer (Fig. 3B). A somewhat

different response was observed for gangliosides, whose levels

were scarcely affected by cell treatment with 0.5 lM PPPP

within 10 h, being however downregulated to a 60–70% of

their initial values, upon prolonging the incubation to 24 h

(Fig. 3C).

Fig. 4. Inhibition of CXCL12-induced cell migration after PPPPpretreatment is reversed by exogenous GSLs. CHP-100 cells pretreatedfor 24 h with 0.5 lM PPPP (+) or left untreated (�) were subjected tochemotaxis assay either in the presence (+) or absence (�) of a 60 nMCXCL12 gradient. Migration tests were performed in the presence of40 lM C8-GlcCer or 40 lM C8-LacCer or 50 lM ganglioside GM2, orin the absence of any exogenously added GSL (None), as indicated inthe upper part of the figure. Media of all samples contained 0.5 lMPPPP, during the migration test. Data are means and S.D. from fourdifferent experiments. Statistical significance is indicated by asterisks(P < 0.02, as compared to the counterparts migrating in the absence ofCXCL12).

C. Limatola et al. / FEBS Letters 581 (2007) 2641–2646 2645

3.3. PPPP-treated cells resume CXCL12-induced chemotaxis

after delivery of exogenous GSLs

To achieve evidence for a causal relationship between GSL

depletion and inhibition of chemokine-dependent cell migra-

tion, we studied whether PPPP-treated CHP-100 cells were

able to resume chemotaxis after addition of exogenous C8-Glc-

Cer or C8-LacCer, two synthetic and short-chain neutral GSL

analogues; moreover, we monitored the effects of exogenously

administered GM2, namely the major ganglioside species in

CHP-100 cells, on the phenomenon. To this aim, cells were

pretreated for 24 h with PPPP and then exposed to a 60 nM

CXCL12 gradient for 5 h in media enriched with the individual

above-mentioned GSL species. As reported in Fig. 4, we ob-

served that all the exogenously administered GSLs were able

to restore the chemotactic response to CXCL12 in PPPP-trea-

ted cells.

4. Discussion

Results reported above provide evidence, for the first time,

that GSL downregulation, achieved by pharmacological

GCS inhibition, associates with inhibition of CXCL12-induced

chemotaxis. The following lines of evidence support the view

that the above-mentioned outcome, as observed in human

CHP-100 neuroepithelioma cells, is not due to unspecific ef-

fects of the GCS inhibitor employed but may truly reflect a

role for GSLs in sustaining chemokine-induced cell migration.

First, we have reported that, even after long-term treatment

with PPPP, cell viability was not affected; moreover, CXCR4

functional state was preserved, as shown by the evidence that

GSL-depleted cells normally mobilize Ca2+ in response to

stimulation with CXCL12. Second, we have shown that sus-

tained GCS inhibition had no secondary effects on the levels

of cholesterol, whose downregulation has been previously

demonstrated to inhibit CXCR4-dependent chemotaxis [12],

or on those of ceramide, a pleiotropic active lipid [23]. Third,

we have demonstrated that the effect of PPPP on CXCL12-in-

duced chemotaxis was dependent on the time of cell pre-expo-

sure to the GCS inhibitor and correlated with downregulation

of GSL levels; moreover, addition of exogenous GSLs to

PPPP-treated cells restored chemokine-dependent cell migra-

tion.

Indeed, with respect to the latter point, a finding that de-

serves clarification deals with the observation that, although

inhibition of CXCR4-dependent cell migration was already

observed when gangliosides were not yet substantially down-

regulated, resumption of chemokine-induced chemotaxis could

be induced not only by exogenous GlcCer and LacCer, but

also by ganglioside GM2. On the one hand, this result would

suggest a redundant function for different GSL species in sus-

taining chemokine-induced chemotaxis; thus, one could envis-

age the possibility that the average level of total GSLs, rather

than the level of any particular GSL species, is critical for sus-

taining CXCL12-evoked cell migration. On the other hand,

what remains to be investigated is whether exogenously admin-

istered gangliosides undergo metabolism after their uptake,

possibly leading to enrichment of the pool of (a) neutral

GSL species, critical for chemotaxis occurrence. In this respect,

it might be of particular relevance the notion that ganglioside

catabolism along the salvage pathway may lead to generation

of neutral GSLs, particularly GlcCer, that are recycled after

escaping lysosomes [24,25]. Studies are currently in progress

to clarify this point, not only in CHP-100 cells but also in

U87MG human glioblastoma cells, that we also found to re-

spond to PPPP treatment with inhibition of CXCR4-depen-

dent chemotaxis (not shown).

Acknowledgements: This work was partly supported by MIUR fundsto Prof. Fabrizio Eusebi (University of Rome ‘Sapienza’) and to Ange-lo Spinedi. The authors thank Dr. Emiliano Basso (ENEA), Dr. ElenaRomano (University of Rome ‘Tor Vergata’), Dr. Antonella Arcella,Dr. Ilenia Del Riccio (Neuromed, Pozzilli), Prof. Sergio Fucile andDr. Antonietta Sucapane (University of Rome ‘Sapienza’) for theirkind assistance in some experiments performed. Thanks are also dueto Prof. Gerry Melino (University of Rome ‘Tor Vergata’) for provid-ing the CHP-100 cell line and the phycoerythrin-conjugated anti-CXCR4 antibody.

References

[1] Klein, R.S. and Rubin, J.B. (2004) Immune and nervous systemCXCL12 and CXCR4: parallel roles in patterning and plasticity.Trends Immunol. 25, 306–314.

[2] Zou, Y.R., Kottmann, A.H., Kuroda, M., Taniuchi, I. andLittman, D.R. (1998) Function of the chemokine receptorCXCR4 in haematopoiesis and in cerebellar development. Nature393, 595–599.

[3] Luo, Y., Cai, J., Xue, H., Miura, T. and Rao, M.S. (2005)Functional SDF1a/CXCR4 signaling in the developing spinalcord. J. Neurochem. 93, 452–462.

[4] Suratt, B.T., Petty, J.M., Young, S.K., Malcolm, K.C., Lieber,J.G., Nick, J.A., Ponzalo, J.A., Henson, P.M. and Worthen, G.S.(2004) Role of the CXCR4/SDF-1 chemokine axis in circulatingneutrophil homeostasis. Blood 104, 565–571.

[5] Lee, B.C., Lee, T.H., Avraham, S. and Avraham, H.K. (2004)Involvement of the chemokine receptor CXCR4 and its ligandstromal cell-derived factor 1a in breast cancer cell migrationthrough human brain microvascular endothelial cells. Mol.Cancer Res. 2, 327–338.

[6] Liang, Z., Yoon, Y., Votaw, J., Goodman, M.M., Williams, L.and Shim, H. (2005) Silencing of CXCR4 blocks breast cancermetastasis. Cancer Res. 65, 967–971.

[7] Ehtesham, M., Winston, J.A., Kabos, P. and Thompson, R.C.(2006) CXCR4 expression mediates glioma cell invasiveness.Oncogene 25, 2801–2806.

[8] Floridi, F., Trettel, F., Di Bartolomeo, S., Ciotti, M.T. andLimatola, C. (2003) Signalling pathways involved in the chemo-tactic activity of CXCL12 in cultured rat cerebellar neurons andCHP100 neuroepithelioma cells. J. Neuroimmunol. 135, 38–46.

[9] Gomez-Mouton, C., Lacalle, R.A., Mira, E., Jimenez-Baranda,S., Barber, D.F., Carrera, A.C., Martinez-A, C. and Manes, S.(2004) Dynamic redistribution of raft domains as organizingplatform for signaling during cell chemotaxis. J. Cell Biol. 164,759–768.

[10] Peng, S.B., Peek, V., Zhai, Y., Paul, D.C., Lou, Q., Xia, X.,Eessalu, T., Kohn, W. and Tang, S. (2005) Akt activation, but notextracellular signal-regulated kinase activation, is required forSDF-1a/CXCR4-mediated migration of epitheloid carcinomacells. Mol. Cancer Res. 3, 227–236.

[11] Nguyen, D.H. and Taub, D. (2002) CXCR4 function requiresmembrane cholesterol: implications for HIV infection. J. Immu-nol. 168, 4121–4126.

[12] Gomez-Mouton, C., Abad, J.L., Mira, E., Lacalle, R.A.,Gallardo, E., Jimenez-Baranda, S., Illa, I., Bernad, A., Manes,S. and Martinez-A, C. (2001) Segregation of leading-edge anduropod components into specific lipid rafts during T cellpolarization. Proc. Natl. Acad. Sci. USA 98, 9642–9647.

[13] Thiele, C.J., McKeon, C., Triche, T.J., Ross, R.A., Reynolds,C.P. and Israel, M.A. (1987) Differential protooncogene expres-sion characterizes histopathologically indistinguishable tumors ofthe peripheral nervous system. J. Clin. Invest. 80, 804–811.

2646 C. Limatola et al. / FEBS Letters 581 (2007) 2641–2646

[14] Catani, M.V., Corasaniti, M.T., Navarra, M., Nistico, G.,Finazzi-Agro, A. and Melino, G. (2000) gp120 induces cell deathin human neuroblastoma cells through the CXCR4 and CCR5chemokine receptors. J. Neurochem. 74, 2373–2379.

[15] Folch, J., Lees, M. and Sloane-Stanley, G.H. (1957) A simplemethod for the isolation and purification of total lipids fromanimal tissues. J. Biol. Chem. 266, 497–509.

[16] Ando, S. and Saito, M. (1987) TCL and HPTLC of phospholipidsand glycolipids in health and disease in: Chromatography ofLipids in Biomedical Research and Clinical Diagnosis (Kuksis,A., Ed.), pp. 266–305, Elsevier Science Publisher, BV, Amster-dam.

[17] Macala, L.J., Yu, R.K. and Ando, S. (1983) Analysis of brainlipids by high performance thin layer chromatography anddensitometry. J. Lipid Res. 24, 1243–1250.

[18] Kean, E.L. (1966) Separation of gluco- and galactocerebrosidesby means of borate thin-layer chromatography. J. Lipid Res. 7,449–452.

[19] Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J.(1951) Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193, 265–275.

[20] Lee, L., Abe, A. and Shayman, J.A. (1999) Improved inhibitors ofglucosylceramide synthase. J. Biol. Chem. 274, 14662–14669.

[21] Grazide, S., Terrisse, A.D., Lerouge, S., Laurent, G. andJaffrezou, J.P. (2004) Cytoprotective effect of glucosylceramidesynthase inhibition against daunorubicin-induced apoptosis inhuman leukemic cell lines. J. Biol. Chem. 279, 18256–18261.

[22] Li, R., Manela, J., Kong, Y. and Ladisch, S. (2000) Cellulargangliosides promote growth factor-induced proliferation offibroblasts. J. Biol. Chem. 275, 34213–34223.

[23] Ogretmen, B. and Hannun, Y.A. (2004) Biologically activesphingolipids in cancer pathogenesis and treatment. Nat. Rev.,Cancer 4, 604–616.

[24] Trinchera, M., Ghidoni, R., Sonnino, S. and Tettamanti, G.(1990) Recycling of glucosylceramide and sphingosine for thebiosynthesis of gangliosides and sphingomyelin in rat liver.Biochem. J. 270, 815–820.

[25] Tettamanti, G., Bassi, R., Viani, P. and Riboni, L. (2003) Salvagepathways in glycosphingolipid metabolism. Biochimie 85, 423–437.

[26] Svennerholm, L. (1963) Chromatographic separation of humanbrain gangliosides. J. Neurochem. 10, 613–623.