ORIGINAL ARTICLE

Joel de Leon Æ Audry Fernandez Æ Circe Mesa

Marilyn Clavel Æ Luis E. Fernandez

Role of tumour-associated N-glycolylated variant of GM3 gangliosidein cancer progression: effect over CD4 expression on T cells

Received: 28 March 2005 / Accepted: 07 June 2005 / Published online: 6 October 2005� Springer-Verlag 2005

Abstract Gangliosides have diverse biological functionsincluding modulation of immune system response. Thesemolecules are differentially expressed on malignant cellscompared with the corresponding normal ones and areinvolved in cancer progression affecting, in differentways, the host’s anti-tumour specific immune responses.Although in humans the N-glycolylated variant of GM3ganglioside is almost exclusively expressed in tumourtissues, the significance of this glycolipid for malignantcell biology remains obscure, while for NAcGM3 strongimmune suppressive effects have been reported. Thepresent work demonstrates, for the first time, thecapacity of NGcGM3 ganglioside to down-modulateCD4 expression in murine and human T lymphocytes,especially in non-activated T cells. Thirty and tenfoldreductions in CD4 expression were induced by purifiedNGcGM3 ganglioside in murine and human T lym-phocytes, respectively. The CD4 complete recovery inthese cells occurred after 48 h of ganglioside removal,due to neo-synthesis. Restored T cells kept similar sen-sitivity to ganglioside-induced CD4 down-modulationafter a new challenge. In addition, a clear associationbetween NGcGM3 insertion in lymphocyte plasmamembranes and the CD4 down-modulation effect wasdocumented. Notably, a possible role of this gangliosidein tumour progression, taking advantage of the X63myeloma model, was also outlined. The relevance ofthese findings, characterizing NGcGM3 as a possibletumour immunesurveillance inhibitor and supportingthe reason for its neo-expression in certain human can-cers, is contributing to this unique heterophilic gangli-oside validation as target for cancer immunotherapy.

Keywords Tumour immunology Æ Immunedysfunction Æ NGcGM3 ganglioside ÆCD4 down-modulation

Abbreviations CMP-NeuAc: Monophosphoril citidinN-acetyl sialic acid Æ D-PDMP: D-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol Æ NAcGM3:N-acetylated GM3 Æ NGcGM3: N-glycolylatedGM3 Æ NGNA: N-glycolylneuraminic acid Æ TFI:Total fluorescent intensity

Introduction

Accumulating evidences suggest the capability of certaingangliosides to promote suppressive effect over immunesystem functions [5, 10, 16, 18, 19, 30]. This is particu-larly true for tumour-derived gangliosides [21, 25, 28,31]. These glycolipids can be shed into the tumourmicroenvironment contributing to cancer progression bysuppressing anti-tumour immune responses [reviewed inRef. 2]. The N-glycolylneuraminic acid (NGNA) bio-synthesis was abrogated on humans evolution due to apartial deletion in the CMP-NeuAc hydroxylase enzymegene [8, 17]. However, expression of N-glycolylatedganglioside, together with the N-acetylated variant ofcertain gangliosides, has been adopted by humans can-cers, i.e. retinoblastoma, breast and germ cell tumours[15, 23, 26]. Detection of NGNA, basically in humantumour cells, seems to depend on absorption and met-abolic incorporation from exogenous dietary sources [1].

Among the immune suppressive influences of the N-acetylated variant of GM3 ganglioside (NAcGM3) overT-cell functions [18, 30, 35], the down-modulation ofCD4 expression has received particular attention [12,32]. Considering minor structural differences betweenNGcGM3 and NAcGM3 gangliosides, we decided toexplore the capacity of NGcGM3 ganglioside to reduceCD4 expression, in murine and human T lymphocytes.

J. de Leon (&) Æ A. Fernandez Æ C. Mesa Æ M. ClavelL. E. FernandezVaccine Department, Centre of Molecular Immunology,216 esq 15, Atabey, Playa, 16040 C. Habana, CubaE-mail: yoel@ict.cim.sld.cuTel.: +53-7-2716810Fax: +53-7-2720644

Cancer Immunol Immunother (2006) 55: 443–450DOI 10.1007/s00262-005-0041-6

Also, the possible role of this ganglioside in tumourgrowth was studied, taking advantage of the X63 mye-loma tumour model in which NGcGM3 is the maincontained ganglioside [27]. Results of this work addressrelevance of NGcGM3 as a target for current passiveand active specific cancer immunotherapy.

Materials and methods

Human and murine lymphocytes

Balb/c mice, between 6 and 12 weeks of age, were pur-chased from the Centre for Laboratory Animal Pro-duction (CENPALAB, Havana, Cuba) and maintainedunder standard conditions in the animal house of theCIM (Havana, Cuba). For in vitro experiments, LN cellswere isolated from Balb/c mice and were smashed andmade into a single cell suspension in PBS. Humanperipheral blood obtained from healthy volunteer do-nors was centrifuged over a Ficoll-Hypaque densitygradient to obtain peripheral blood leukocytes (PBL) asdescribed [4].

Reagents

Ficoll-Paque Plus (density 1.077 g/mL) to isolate humanPBL from was purchased from Amersham PharmaciaBiotech AB, UK. The inhibitor of GlcCer-synthase D-PDMP (Matreya, Pleasant Gap, PA, USA) was addedto cell cultures during 72 h at 10 lM from a 2.5 mMstock solution. Cycloheximide (Sigma Chemical, Dorset,UK) was used as a protein neo-synthesis inhibitor at a10 lg/mL final concentration. Concanavalin A (Con A)mitogen was purchased from Calbiochem (Calbiochem,La Jolla, CA, USA) and was added to cell cultures at 2-lg/mL final concentration.

Gangliosides isolation and quantification analysis

The NGcGM3 and NAcGM3 gangliosides were ob-tained from horse and dog erythrocytes, respectively,according to the method of Svennerholm and Fredmanwith minor modifications [34]. Gangliosides were alsoextracted from D-PDMP-treated or untreated X63myeloma cells. Briefly, cells were pelleted, lyophilizedand then extracted twice in chloroform/methanol/water(4:8:3, v/v/v) with stirring. Insoluble material was re-moved by centrifugation and the lipid-enriched super-natant was dried and subjected to a Folch partition [11].The supernatant recovered by centrifugation was driedand dissolved in 0.2 M NaOH/methanol at 37�C for 2 h.After neutralization with 0.2 M HCl/methanol, the lipidextract was partitioned three times in hexane/methanol(1:3, v/v). Polar phase was dried and dissolved in water.Salts were eliminated by a 3-day dialysis. Silica gel 60

and DEAD-Sephadex A25 chromatographic methodswere included for final purification, yielding purityhigher than 95%. Gangliosides were as quantified by aresorcinol colorimetric assay detecting lipid-bound sialicacid [33]. High-performance thin layer chromatography(HPTLC), using silica gel HPTLC plates (Merck,Darmstadt, Germany), was performed in chloroform/methanol/0.25% aqueous KCl (5:4:1, v/v/v). Ganglio-sides were visualized with resorcinol.

Tumour growth assays

P3X63 Ag 8.653 myeloma cell line (ATCC NCRL 1580)was grown in RPMI 1640 with Glutamax I and 25 mMHEPES (Gibco, UK) supplemented with 10% FCS(HyClone, USA), 100 U/mL penicillin, 100 lg/mLstreptomycin and 50 lM 2-mercaptoethanol, andmaintained at 37�C with 5% CO2. Myeloma cells werecultured or not with added D-PDMP gangliosideinhibitor medium before being injected subcutaneouslyin Balb/c mice (106 cells/animal). The diameter of tu-mour nodules induced by inoculated tumour cells wasmeasured with callipers every 3 days.

FACS analysis

Murine and human lymphocytes were analysed forexpression of surface molecules by flow cytometry as-says, including CD3, CD4, CD8 and CD25. The pres-ence of NGcGM3 at the plasma membrane oflymphocytes and cell lines was also tested. The followingMAbs were used for staining cells at a 1:200 dilution:FITC-conjugated rat anti-mouse CD3 (145-2C11),FITC or PE-conjugated rat anti-mouse CD4 (RM4-5),PE-conjugated rat anti-mouse CD8 (53–6.7), biotiny-lated rat anti-mouse CD25 (7D4), FITC-conjugatedmouse anti-human CD3 (HIT3a), PE-conjugated mouseanti-human CD4 (RPA-T4), PE-conjugated mouse anti-human CD8 (RPA-T8) (BD PharMingen, San Diego,CA, USA). Streptavidin–RPE-conjugated was pur-chased from Dako, Denmark and used at a 1:300 dilu-tion. Cells were stained with the biotinylated-14F7 MAbat 5 lg/mL to detect NGcGM3 ganglioside on cellplasma membrane. This antibody was produced andprovided by the Centre of Molecular Immunology,Cuba [6].

In all the cases, cells were suspended in 200 lL ofPBS-NaN3 0.01%-FCS 1% and stained with eitherFITC or PE-conjugated MAbs for 15 min at 4�C. At theend of the incubation time, cells were washed with PBS-FCS 1%. The Annexin V Kit (BD PharMingen) wasused to assess the capacity of D-PDMP to increase theapoptosis rate on X63 cells culture. Annexin V+ andPI� cells were considered in early apoptotic state.

Mean fluorescent intensity (MFI) and percent ofpositive stained cells were determined in a FACScaninstrument (Becton Dickinson, USA). The WinMDI 2.8

444

program was used to analyse a total of 104 cells acquiredon every FACS assay. Total Fluorescent Intensity (TFI)values were calculated as: MFI (PE or FITC emissionchannel)·fraction of positive stained cells.

Modulation of CD4 expression on mouseand human T lymphocytes

The LN cells isolated from Balb/c mice were culturedduring 72 h in a Con A added RPMI 1640 completemedium supplemented with 10% FCS. Pre-activatedand not activated LN cells were re-suspended at 106

cells/mL in PBS. Purified NAcGM3 or NGcGM3 gan-gliosides were added at different concentrations to cellsuspensions. In parallel, gangliosides isolated from D-PDMP-treated or untreated X63 myeloma cells werediluted in 250 lL of PBS and added at 1:12, 1:24 and1:48 dilution (v/v) to lymphocyte suspensions. Dilution1:12 is equivalent to around 50 lg/mL concentration ofX63 extracted gangliosides. Incubation periods werealways 1 h, at 37�C in CO2 5%. Control cells incubatedin the absence of gangliosides were included in everyexperiment. Human peripheral lymphocytes were re-suspended in PBS at 106 cells/mL and incubated withdifferent concentrations of purified NAcGM3 orNGcGM3 gangliosides. After 1 h, cells were washed andincubated in 24-well plates (Nunc, Denmark) at 106

cells/mL in the absence or in the presence of cyclohexi-mide added medium during 24 or 48 h. After the end ofincubation periods, cells were stained to evaluate CD4molecule modulation by flow cytometry.

NGcGM3 presence at the plasma membrane wasdetected in both murine and human T lymphocytesincubated for 1 h at 37�C in the presence of differentconcentrations of purified NGcGM3 ganglioside. Cellswere stained with the NGcGM3-specific 14F7 MAb asdescribed above.

Statistical analyses

Differences in mean tumour volumes between groups ofanimals, inoculated with D-PDMP-treated or untreatedX63 cells, were analysed by Student’s t-test. The Mann–Whitney U method was used as a non-parametric testfor pair-wise comparisons. To determine statistical cor-relation, the Pearson test was applied. In all cases, valueswere considered statistically significant when P £ 0.05.

Results

Ganglioside synthesis inhibitor D-PDMP reduces lipid-bound sialic acid content of treated X63 myeloma cells

In X63 myeloma cells cultured for three days in RPMImedium containing D-PDMP, a twofold reduction in

total gangliosides content was observed (Fig. 1a).NGcGM3 is the major ganglioside in X63 parental cellline, the P3/NS1 myeloma (NGcGM3/NAcGM3, 85:15ratio) [27]. Levels of this ganglioside in X63 cells showeda threefold decrease (expressed as MFI) after D-PDMPtreatment as indicated by FACS analyses with theNGcGM3-specific 14F7 MAb (Fig. 1b).

X63 murine myeloma tumour growth was inhibited afterthe cells’ treatment with D-PDMP

NGcGM3 contribution to tumour growth was assessedin murine X63 myeloma model. Balb/c mice were SC-inoculated with 106 normal or D-PDMP-treated X63cells and tumour growth followed thereafter. Ganglio-side content reduction after the cells’ treatment with theglucosylceramide synthase inhibitor significantly af-fected tumour growth in vivo (Fig. 2). Eleven days aftercell inoculation, mean tumour volume was 4.5 timeshigher in mice inoculated with untreated cells than inthose injected with D-PDMP-treated cells (P=0.004).At D-PDMP-tested concentration (10 lM), neither anincrease in apoptotic cell frequency (detected by Ann-exin V and PI staining) (Table 1) nor an impact inmyeloma cells in vitro proliferation rate (data not

Fig. 1 Ganglioside content on X63 myeloma cells treated with theganglioside synthesis inhibitor D-PDMP. a Assessment of gangli-oside content in X63 myeloma cells after D-PDMP treatment bythe resorcinol colorimetric assay for lipid-bound sialic acid. Mann–Whitney test was used for statistical analysis; b Detection ofNGcGM3 expression on X63 myeloma cells by FACS analysis withthe biotinylated 14F7 MAb. Non-stained cells were included ascontrol (filled histogram)

445

shown) were detected. These results suggest that theobserved D-PDMP-treated X63 myeloma cells in vivogrowth decrease is not caused by differences in inocu-lated cells’ proliferative capacity.

Total gangliosides fraction extracted from X63 cells andpurified NGcGM3 reduce CD4 expression in mouse andhuman T lymphocytes

Considering that subcutaneous growth of X63 myelomadepends on CD4 T cells (data not shown) and alsoprevious described effects of certain gangliosides overCD4 expression [21, 36], we directly explore the myelo-ma total ganglioside fraction capacity to reduce CD4expression on T cells. Gangliosides either isolated fromuntreated or D-PDMP-treated cells were incubated withmouse T lymphocytes. As shown in Fig. 3a, the lipid-bound sialic-acid-containing fraction induced a drasticreduction in CD4 expression on T cells. The intensity ofthis effect was almost half diminished when gangliosidesobtained from D-PDMP pre-treated myeloma cells wereused (Fig. 3b).

The next step was determining the particular contri-bution of NGcGM3 ganglioside to the above-describedeffect. For this purpose, murine and human T lympho-cytes were incubated with NGcGM3 and CD4 levelsmeasured thereafter. As shown in Fig. 4a and b, inter-action of this purified ganglioside with mouse and hu-man lymphocytes provoked CD4 down-modulation inboth cellular populations. Notably, significant reduc-tions in TFI values on stained CD4+ T cells (30 and 10times in mouse and human lymphocytes, respectively)were observed. Furthermore, almost identical resultswere obtained comparing NGcGM3 and NAcGM3 ef-fects over CD4 (Fig. 4c–e) and CD8 expression(Fig. 4d–f). As shown, modifications on CD8 levels werenot induced by GM3 variants. Expression of CD3 wasalso unmodified at any assayed ganglioside concentra-tions (data not shown).

Purified NGcGM3 ganglioside differentially affects CD4levels in naıve or CD4+ CD25+ - activated T cells

Purified NGcGM3’s capacity to down-modulate CD4levels on naıve or activated murine T lymphocytes wastested. The LN naıve cells from Balb/c mice were ob-tained and activated CD4+ CD25+ T cells were gener-ated after in vitro culture of these cells with Con Amitogen. Interestingly, the addition of NGcGM3 pref-erentially reduces CD4 molecule expression in naive Tcells (a 66% drop in expression), compared with almostno effect over CD4+ CD25+ mitogen-activated T lym-phocytes (Fig. 5). Modification in CD25 levels on acti-vated T lymphocytes was not induced by theganglioside.

Fig. 3 Down-modulation of CD4 expression in mouse T lympho-cytes, incubated with gangliosides isolated from X63 myelomacells. a Extracted gangliosides were suspended in 250 lL of PBSand diluted 12, 24 or 48 times on lymphocyte suspension. Filledhistogram corresponds to CD4 expression in control cells. bDifferential CD4 molecule down-modulation induced by ganglio-sides obtained from D-PDMP-treated (solid symbols) or untreatedX63 cells (open symbols); TFI data from two independent assayswere considered for statistical analysis (P<0.01 at both dilutionstested, Mann–Whitney test)

0 5 10 15 200

300

600

900

1200

1500

Days after cell inoculation

Tum

our

volu

me

(mm

3 )

****

*

*

Fig. 2 Ganglioside content relevance for X63 myeloma growth.Balb/c mice were inoculated with 106 D-PDMP-treated (triangles)or untreated (circles) X63 myeloma cells. Tumour diameters weremeasured with callipers every 3 days after tumour palpation (*P<0.05, ** P<0.01, Student’s t-test)

Table 1 Effect of D-PDMP treatment in X63 cells apoptosis invitro

AnnexinV+ PI�a AnnexinV+ PI+b

X63c 7.6±0.8 9.9±1.0X63+D-PDMPc 10 lM 6.7±0.3 11.2±1.2X63+D-PDMP 50 lM 13.9±1.6 24.2±1.7

a Percent of early apoptotic cells in three independent assaysb

Percent of necrotic and late apoptotic cells in three independentassaysc X63 myeloma cells were cultured during 72 h with normalor D-PDMP-supplemented culture medium. Ganglioside synthesisinhibitor was tested at 10 lM in all in vivo or in vitro assays, exceptfor apoptosis detection where a 50 lM concentration was alsoanalysed. Cells were washed and stained with FITC-conjugatedAnnexin V and PI

446

Human lymphocyte CD4 expression recovering afterremoval of the NGcGM3 is due to new molecule syn-thesis

CD4 molecules recovery kinetic in human peripherallymphocytes, incubated for 1 h with purified NGcGM3ganglioside (50 lg/mL), was evaluated. Cytofluoromet-ric assays, developed 24 h after the treatment, revealed athreefold increase in CD4 molecules level on ganglio-side-treated cells (Fig. 6). After 48 h, more than 85% ofCD4 expression recovery was observed.

The CD4 molecules’ recovery in plasma membranesafter ganglioside removal was also measured in thepresence of cycloheximide, a chemical inhibitor of pro-tein synthesis. As shown in Fig. 6, recuperation of CD4molecules level was abolished in cultures containingcycloheximide. These results suggest that T cells require

new CD4 molecules synthesis to restore their normallevels.

After a first round of CD4 levels restoration, lym-phocytes did not become refractory to future contactswith NGcGM3. In fact, repeated incubation of thesecells with the ganglioside promoted recurrent cycles ofCD4 down-modulation and recovery (data not shown).

Fig. 4 Effect of purified NAcGM3 and NGcGM3 gangliosidesover CD4 and CD8 expression in murine and human Tlymphocytes. Mice LN cells (a) or human PBL (b) were incubatedwith depicted concentrations of NGcGM3 to evaluate effect overCD4 expression by FACS. Data from three independent assays, foreach cell population, were considered for statistical analysis. Asignificant reduction in CD4 levels (P<0.05, Mann–Whitney test)was detected at any NGcGM3 concentration tested. Comparativeeffect of NAcGM3 and NGcGM3 gangliosides on mouse (c, d) andhuman (e, f) T cells is represented in histograms: filled (control);continuous line (NAcGM3 at 50 lg/mL); detached line(NGcGM3 at 50 lg/mL)

NGcGM3 concentration (µg/mL)0 12.5 25 50

0

10

20

30

40

50

60

70

%C

D4+

T c

ells

0 12.5 25 500

10

20

30

40

50

%C

D4+

T c

ells

a b

Fig. 5 Differential effect of NGcGM3 over CD4 expression onnaıve (a) or activated (b) murine T lymphocytes. The LN cells wereactivated with Con A for 72 h prior to incubation with purifiedNGcGM3 ganglioside. Percent of CD4+ cells on activatedpopulation corresponds to the fraction of CD4 T lymphocytesgated as CD25+ cells. Results shown correspond to a representa-tive FACS assay. At the higher ganglioside concentration tested,CD4 reductions on naıve T cells were 66, 72 and 59% vs. 13, 18 and15% reduction on activated T cells (three independent experimentsdata)

TF

Ie/T

FIc

cycloheximide + +- - -

0

0,2

0,4

0,6

0,8

1.0

1h 24h 24h 48h 48h

Fig. 6 CD4 recovery dependence on new protein synthesis inganglioside-treated human peripheral lymphocytes. Cells wereincubated in the presence of NGcGM3 ganglioside (50 lg/mL;filled bars) or in their absence as control (empty bars). Gangliosidesremoved from cells were then cultured in medium alone or incycloheximide added medium during 48 or 96 h. TFIe/TFIc totalfluorescence intensity experimental/control. Data are representativeof the other two independent assays with CD4 recoveries higherthan 85% and lower than 35%, in the absence or in the presence ofcycloheximide, respectively

447

CD4 down-modulation correlates with NGcGM3 gan-glioside insertion into lymphocyte plasma membranes

In order to understand the physical reason why purifiedNGcGM3 down-modulated CD4 expression in T cells,the insertion capacity of this ganglioside into mouse andhuman cells plasma membranes was studied. The LNcells obtained from Balb/c mice and human peripherallymphocytes were incubated with different gangliosideconcentrations and NGcGM3 ganglioside detected inmouse and human T lymphocyte plasma membranes(Fig. 7). Notably, in human T cells, insertion of exoge-nous added ganglioside was higher than in murine Tlymphocytes. In fact, at 25 lg/mL of NGcGM3 gangli-oside, MFI values detected in 14F7 MAb stained cellsincreased more than 50 times in murine but almost 200times in human T lymphocytes relative to correspondingcontrol cells.

It is worth noting that a significant inverse correlationbetween CD4 molecules down-modulation and the in-creased ganglioside association to murine and human Tlymphocytes membranes was observed, in a concentra-tion-dependent manner (Fig. 8).

Discussion

Tumour-derived NAcGM3 ganglioside impairs immunesystem functions [28, 31], exerting an extensively de-scribed effect over CD4 expression [12, 32]. However,the role of NGcGM3 as an inhibitor of tumour im-munesurveillance has been less extensively studied [35,36]. This work focused on the description and charac-terization of CD4 down-modulation induced byNGcGM3 ganglioside on mouse and human T cells.

Purified NGcGM3 induced a significant and con-centration-dependent CD4 down-modulation on mouseand human T lymphocytes. Complete recovery of CD4expression in human T cells was observed 48 h afterNGcGM3 ganglioside removal and depends on newCD4 molecules synthesis. T cells remained sensible toCD4 down-modulation induced by NGcGM3, evenafter total CD4 recovery, suggesting that a long-lastingeffect over T cells function could be induced in vivo byNGcGM3-containing tumours. These results high-lighted the possibility that tumour-associated NGcGM3can act in favour of cancer progression, reducing CD4expression and affecting T-cells functionality.

A mechanism to explain GM1 and GM3 effect overCD4 expression has been described [12, 13, 29]. Struc-tural similarities between N-acetyl and N-glycolyl vari-ants of GM3, and our results showing an almostcomplete identity in the effect of both gangliosides overCD4 and CD8 expression, suggest that a similar mech-anism could explain CD4 down-modulation induced bythese gangliosides. The relationship of this mechanismwith the known capacity of shed tumour gangliosides tobe inserted into lymphoid cells plasma membrane [3, 9,

22] was demonstrated by the finding of a statisticalcorrelation between NGcGM3-induced CD4 down-modulation and the presence of different amounts of thisganglioside at T-cell surfaces. Especially in humanlymphocytes, lipid rafts composition disturbanceprompted after external NGcGM3 ganglioside insertioninto the cell surface is possible, considering that NAc-GM3 is the major ganglioside in human T cells; mean-while the N-glycolylated variant of GM3 is ratherabsent.

The CD4 expression dependence of NGcGM3 con-tact in T cells with distinct phenotypes was investigated.Surprisingly, more than 60% reduction in CD4 level wasdetected on naıve T cells, a significantly higher effectthan that observed on mitogen-activated lymphocytes.Several proteins inserted into lipid rafts modify theirdistribution and functionality after cell activation [14,20, 24]. Re-distribution of molecules at the cell surface,after Con A-induced activation, could modify CD4sensibility to the ganglioside-induced down-modulation.

Fig. 7 Externally added NGcGM3 ganglioside detection at plasmamembranes of T lymphocytes. The NGcGM3 ganglioside expres-sion in murine (a) and human (b) T lymphocytes was detected after1 h incubation with depicted ganglioside concentrations. TheNGcGM3 expression was evaluated in gated CD3+ lymphocytesby FACS assays. Cell staining was conducted with biotinylated-14F7 MAb

448

It could be interesting to find if NGcGM3 can reduceCD4 expression in antigen-specific activated T lympho-cytes.

In human normal cells, CMP-NeuAc hydroxylase iscompletely inactive due to N-terminal deletion of a 92-base-pair-long exon in the genomic DNA coding theenzyme sequence [17]. However, N-glycolylated gan-gliosides have been detected in human cells, preferen-tially tumour cells [15, 23, 26], depending on the dietaryingestion of NGNA-containing food [1]. The relevanceof this ganglioside for tumour progression was apparentsince SC inoculation in mice of ganglioside synthesisinhibitor D-PDMP-treated X63 myeloma cells exhibiteda significant reduction in tumour growth when comparedwith untreated cells. While in these tumour cellsNGcGM3 is the major ganglioside [27], a possibleexplanation of this result could be related to the quan-titative reduction in this immunosuppressive gangliosidecontent, at the myeloma cell membrane level. On theother hand, the influence of CD4+ T cells over subcu-taneously growing X63 tumours (data not shown) sug-gests a possible relationship with CD4 T-cell functions.

In fact, we found that the lipid-bound sialic acid-con-taining fraction isolated from X63 cells provoked aprofound down-modulation of CD4 molecule on murineT cells. However, in vivo effects of NGcGM3 gangliosideover tumour growth might not be restricted to CD4 Tcells. Experiments focused on finding whether NGcGM3can affect APC functions or alter cytokine pattern se-creted by T cells are currently in progress in our lab.

A remaining question is why NGcGM3 is adopted bycertain advanced tumours [7] in spite of its relativelyhigh immunogenicity? A hypothetic explanation couldbe related to NGcGM3 immunosuppressive capacity,contributing to cancer progression. All these resultssupported our current application of anti-cancer thera-pies targeting NGcGM3 ganglioside [6, 7].

Acknowledgements We thank Armando Lopez for excellent tech-nical assistance and Dr Blanca Tormo for skilful manuscript edi-tion. This work was supported by the Centre of MolecularImmunology (Havana, Cuba).

References

1. Bardor M, Nguyen DH, Diaz S, Varki A (2005) Mechanism ofuptake and incorporation of the non-human sialic acid N-glycolylneuraminic acid into human cells. J Biol Chem280:4228–4237

2. Birkle S, Zeng G, Gao L, Yu RK, Aubry J (2003) Role oftumour-associated gangliosides in cancer progression. Bio-chimistry 85:455–463

3. Black PH (1980) Shedding from normal and cancer-cell sur-faces. N Engl J Med 303:1415–1416

4. Boyum A (1968) Isolation of mononuclear cells and granulo-cytes from human blood. Scand J Clin Lab Invest Suppl 97:77–89

5. Caldwell S, Heitger A, Shen W, Liu Y, Taylor B, Ladisch S(2003) Mechanisms of ganglioside inhibition of APC function.J Immunol 171:1676–1683

6. Carr A, Mullet A, Mazorra Z, Vazquez AM, Alfonso M, MesaC, Rengifo E, Perez R, Fernandez LE (2000) A mouse IgG1monoclonal antibody specific for N-glycolyl GM3 gangliosiderecognized breast and melanoma tumours. Hybridoma 19:241–247

7. Carr A, Rodriguez E, Arango Mdel C, Camacho R, Osorio M,Gabri M, Carrillo G, Valdes Z, Bebelagua Y, Perez R, Fer-nandez LE (2003) Immunotherapy of advanced breast cancerwith a heterophilic ganglioside (NeuGcGM3) cancer vaccine. JClin Oncol 21:1015–1021

8. Chou HH, Takematsu H, Diaz S, Iber J, Nickerson E, WrightKL, Muchmore EA, Nelson DL, Warren ST, Varki A (1998) Amutation in human CMP-sialic acid hydroxylase occurred afterthe Homo-Pan divergence. Proc Natl Acad Sci USA 95:11751–11756

9. Dolo V, Li R, Dillinger M, Flati S, Manel J, Taylor BJ, PavanA, Ladisch S (2000) Enrichment and localization of gangliosideGD3 and caveolin-1 in shed tumour cell membrane vesicles.Biochim Biophys Acta 1486:265–274

10. Dumontet C, Rebbaa A, Bienvenu J, Portoukalian J (1994)Inhibition of immune cell proliferation and cytokine produc-tion by lipoprotein-bound gangliosides. Cancer Immunol Im-munother 38:311–316

11. Folch PJ, Arsove S, Meath JA (1951) Isolation of brain stradin:a new type of large molecular tissue component. J Biol Chem191:819–831

12. Garofalo T, Sorice M, Misasi R, Cinque B, Giammatteo M,Pontieri GM, Cifone MG, Pavan A (1998) A novel mechanism

NG

cGM

3 expression (TF

I)CD

4 ex

pres

sion

(TF

I)

P<0.001

r=-0.979

0

10

20

30

40

50a

b

0

60

120

180

240

300

360

CD

4 ex

pres

sion

(T

FI)

NGcGM3 concentration (µg/mL)

0 0.3

0.7

1.5 3 6

12.5 25 50

NG

cGM

3 expression (TF

I)

P<0.001

r=-0,941

0

10

20

30

40

0

200

400

600

800

1000

1200

1400

Fig. 8 Correlation between CD4 down-modulation and ganglio-side incorporation to T lymphocytes plasma membranes. CD4molecule (solid squares) and NGcGM3 ganglioside (open squares)were detected in murine (a) and human (b) T lymphocytes afterincubation with different ganglioside concentrations. The TFIvalues were calculated for CD4- and NGcGM3-positive cells. TheMFI values of CD4+ cells range from 76 (control) to 8.2 in murinecells and 96.6 (control) to 17.7 in human lymphocytes; MFI valuesof 14F7 Mab-stained cells range from 6.1 (control) to 263 in murinecells and 2.5 (control) to 1193 in human lymphocytes. Pearsoncorrelation test was used for statistical analyses. Comparableresults were obtained in two independent assays

449

of CD4 down-modulation induced by monosialogangliosideGM3. J Biol Chem 273:35153–35160

13. Garofalo T, Sorice M, Misasi R, Cinque B, Mattei V, PontieriGM, Cifone MG, Pavan A (2002) Ganglioside GM3 activatesERKs in human lymphocytic cells. J Lipid Res 43:971–978

14. Glebov OO, Nichols BJ (2004) Lipid raft proteins have a ran-dom distribution during localized activation of the T-cellreceptor. Nat Cell Biol 6:238–243

15. Higashi H, Sasabe T, Fukui Y, Maru M, Kato S (1988)Detection of gangliosides as N-glycolylneuraminic acid-specifictumour-associated Hanganutziu-Deicher antigen in humanretinoblastoma cells. Jpn J Cancer Res 79:952–956

16. Irani DN, Lin KI, Griffin DE (1996) Brain-derived gangliosidesregulate the cytokine production and proliferation of activatedT cells. J Immunol 157:4333–4340

17. Irie A, Koyama S, Kozutsumi Y, Kawasaki T, Suzuki A (1998)The molecular basis for the absence of N-glycolylneuraminicacid in humans. J Biol Chem 273:15866–15871

18. Kanda N (1999) Gangliosides GD1a and GM3 induce inter-leukin-10 production by human T cells. Biochem Biophys ResCommun 256:41–44

19. Kawaguchi T, Nakakuma H, Kagimoto T, Shirono K, Hor-ikawa K, Hidaka M, Iwamori M, Nagai Y, Takatsuki K (1989)Characteristic mode of action of gangliosides in selectivemodulation of CD4 on human T lymphocytes. Biochem Bio-phys Res Commun 158(3):1050–1059

20. Krishnan S, Nambiar MP, Warke VG, Fisher CU, Mitchell J,Delaney N, Tsokos GC (2004) Alterations in lipid raft com-position and dynamics contribute to abnormal T cell responsesin systemic lupus erythematosus. J Immunol 172:7821–7831

21. Kudo D, Rayman P, Horton C, Cathcart MK, Bukowski RM,Thornton M, Tannenbaum C, Finke JH (2003) Gangliosidesexpressed by the renal cell carcinoma cell line SK-RC-45 areinvolved in tumour-induced apoptosis of T cells. Cancer Res63:1676–1683

22. Ladisch S, Wu ZL, Feig S, Ulsh L, Schwartz E, Floutsis G,Wiley F, Lenarsky C, Seeger R (1987) Shedding of GD2 gan-glioside by human neuroblastoma. Int J Cancer 39:73–76

23. Marquina G, Waki H, Fernandez LE, Kon K, Carr A, ValienteO, Perez R, Ando S (1996) Gangliosides expressed in humanbreast cancer. Cancer Res 56:5165–5171

24. Martin M, Schneider H, Azouz A, Rudd CE (2001) CytotoxicT lymphocyte antigen 4 and CD28 modulate cell surface raftexpression in their regulation of T cell function. J Exp Med194:1675–1681

25. McKallip R, Li R, Ladisch S (1999) Tumour gangliosides in-hibit the tumour-specific immune response. J Immunol163:3718–3726

26. Miyake M, Hashimoto K, Ito M, Ogawa O, Arai E, Itomi S,Kannagi R (1990) Abnormal ocurrence and the differentiationdependent distribution of N-acetyl and N-glycolyl species ofganglioside GM2 in human germ cell tumours. Cancer 65:499–505

27. Muthing J, Steuer H, Peter-Katalinic J, Marx U, Bethke U,Neumann U, Lehmann J (1994) Expression of gangliosidesGM3 (NeuAc) and GM3 (NeuGc) in myelomas and hybrido-mas of mouse, rat, and human origin. J Biochem (Tokyo)116:64–73

28. Peguet-Navarro J, Sportouch M, Popa I, Berthier O, SchmittD, Portoukalian J (2003) Gangliosides from human melanomatumours impair dendritic cell differentiation from monocytesand induce their apoptosis. J Immunol 170:3488–3494

29. Saggioro D, Sorio C, Calderazzo F, Callegaro L, Panozzo M,Berton G, Chieco-Bianchi L (1993) Mechanism of action of themonosialoganglioside GM1 as a modulator of CD4 expression.J Biol Chem 268:1368–1375

30. Sharom FJ, Chiu AL, Ross TE (1990) Gangliosides and gly-cophorin inhibit T-lymphocyte activation. Biochem Cell Biol68:735–744

31. Shurin GV, Shurin MR, Bykovskaia S, Shogan J, Lotze MT,Barksdale EM Jr (2001) Neuroblastoma-derived gangliosidesinhibit dendritic cell generation and function. Cancer Res61:363–369

32. Sorice M, Pavan A, Misasi R, Sansolini T, Garofalo T, Lenti L,Pontieri GM, Frati L, Torrisi MR (1995) Monosialoganglio-side GM3 induces CD4 internalization in human peripheralblood T lymphocytes. Scand J Immunol 41:148–156

33. Svennerholm L (1957) Quantitative estimation of sialic acid: acolorimetric resorcinol-hydrocloridric acid method. BiochimByophys Acta 24:604–611

34. Svennerholm L, Fredmand P (1980) A procedure for thequantitative isolation of brain gangliosides. Biochim ByophysActa 617:97–109

35. Takahashi K, Ono K, Hirabayashi Y, Taniguchi M (1988)Escape mechanisms of melanoma from immune system bysoluble melanoma antigen. J Immunol 140(9):3244–3248

36. Taniguchi M, Sakatsume M, Harada Y, Nores GA, HakomoriS (1988) Analysis of melanoma antigen and its involvement intumour-escape mechanisms. Princess Takamatsu Symp 19:247–254

450