(CANCER RESEARCH 52. 4948-4953. September 15. 1992]
Cell Surface Accessibility of Individual Gangliosides in Malignant Melanoma Cells
to Antibodies Is Influenced by the Total Ganglioside Composition of the Cells1
Kenneth O. Lloyd,2 Claudia M. Gordon, Immac J. Thampoe, and Cheryl I)iBeni-detto
Memorial Sloan-Ketlering Cancer Center and Cornell University Graduate School of Medical Sciences, New York, New York 10021
ABSTRACT
The reactivity of a panel of antiganglioside monoclonal antibodies
with a number of melanoma cell lines having different ganglioside com
position profiles was studied. One cell line synthesized only GM3, one
produced both CM3 and GD2, 2 had GM3 and GD3 as their major
gangliosides, and 2 others synthesized approximately equal amounts of
GM3, GM2, GD3, and GD2 gangliosides. Antibody reactivity with
viable cells was analyzed by: (a) flow cytometry on suspension cells; and
(i) mixed hemagglutination assays or immune adherence assays on
monolayer cells in culture. GM3 was efficiently detected only in the cell
line having GM3 as its sole ganglioside. In the other cell lines, GM3
was difficult to detect even in cells in which it made up a high proportion
(up to 50%) of the total ganglioside content. GM2 was easily detectable
only in JB-RH melanoma cells (which contain only GM3 and GM2).
GD3 was the most reactive ganglioside in 2 cell lines and GD2 in 2 other
lines. In general, the most complex ganglioside present in a cell was the
one most accessible to antibody. The differential exposure at the cell
surface of specific gangliosides may have implications for antibodydirected tumor detection and therapy and for cell-protein or cell-cell
interactions that involve glycolipids.
INTRODUCTION
Melanoma cells, as well as other cell types of neuroectodermal origin, are comparatively rich in gangliosides (1). In human
melanoma tissue and cell lines, the predominant gangliosides
are GM33 and GD3, although in some instances GM2 and
GD2 can also be expressed in substantial amounts (2). mAb,4
both mouse and human, have been used extensively in the anal
ysis of ganglioside expression in melanoma (1, 3). Moreover,
mAbs to certain gangliosides have found use in the diagnosis
and therapy of tumors. Specifically, anti-GD3 mAbs have been
studied for the therapy of malignant melanoma (4), and antiGD2 has been applied to both the therapy and diagnosis of
neuroblastoma (5).
Although the reactivity of antiganglioside mAbs with specific
cells and tissue is, in general, related to ganglioside composition
of the cells, there are indications that other factors also play a
role in this process. For example, although GM3 ganglioside is
widely distributed in a variety of cell types, antibodies to GM3
react selectively with certain melanoma cell lines (6-8). This
phenomenon has been partially explained by the high density of
GM3 found on melanoma cells (7). Another important param
eter, which has not been as well studied, may be the influence of
other components of the cell surface, i.e., glycoproteins and
other glycolipids, on antibody reactivity.
Received 1/28/92; accepted 7/3/92.
The costs of publication of this article were defrayed in part by the payment of
page charges. This article must therefore be hereby marked advertisement in accord
ance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 Supported by grants from the National Cancer Institute (CA 21445 and CA
08478).
2 To whom requests for reprints should be addressed, at Memorial SloanKettering, 1275 York Avenue. New York, NY 10021.
3 Ganglioside nomenclature is according to Svennerholm (40).
4 The abbreviations used are: mAb, monoclonal antibody; PBS, phosphatebuffered saline, 0.15 M NaCI-0.05 M sodium phosphate, pH 7.2; TLC, thin layer
chromatography; Gb, globoside series glycolipid; IA. immune adherence: MHA,
mixed hemagglutination.
In this study, we have compared the total ganglioside com
position of a panel of melanoma cell lines with the ability of a
number of antiganglioside mAbs (anti-GM3, -GM2, -GD3, and
-GD2) to react with these cell lines. The results show that the
reactivity of a given mAb with a cell line is dependent not only
on the presence or absence of its cognate ganglioside antigen
but also on the expression of other gangliosides by the target
cell.
MATERIALS
AND METHODS
Cell Lines. Human melanoma cell lines SK-MEL-28, - 31, -37, and
MeWo have been described (9). B78 is a subclone of the mouse mela
noma cell line B16 (10) and was kindly provided by Dr. A. Albino,
Sloan-Kettering Institute. JB-RH is also a mouse melanoma cell line,
and it was derived by Dr. Jane Berkelhammer, University of Missouri,
Columbia, MO, and subsequently subcloned for high GM2 expression
(11). The cell lines were cultured in minimal essential medium supple
mented with fetal bovine serum (10%), L-glutamine (2 HIM),nonessential amino acids (1%), and penicillin-streptomycin (100 Mg/ml) as de
scribed previously (9).
Antibodies. mAb M2590 |anti-GM3; IgM (6)], DH2 [anti-GM3;
IgG3 (12)], and GMR6 [anti-GM3; IgM (13)] were kindly provided by
Drs. T. Itoh and M. Taniguchi, by Dr. S. Hakomori, and by Dr. T. Tai,
respectively. mAb 3F8 (anti-GD2; IgG3) was a gift from Dr. N-K.
Cheung (5). mAb R24 (anti-GD3; IgG3) and 5-3 (anti-GM2; IgM)
have been described previously (11, 14).
Flow Cytometry. For flow-cytometric experiments, confluent cell
cultures were disrupted with 10 min EDTA-PBS, trypsin (0.1%)-EDTA
(0.02%), or Pronase (1 Mg/ml in PBS; Calbiochem, San Diego, CA) to
obtain single cell suspensions. Cells (1.0 x 106/ml) were incubated with
saturating concentrations of antibody as follows: M2590 (20 Mg/ml),
DH2 (culture supernatant), GMR6 (culture supernatant), 5-3 (ascites
fluid, 1:100), R24 (20 Mg/ml), and 3F8 (20 Mg/ml). After incubation at
0°Cfor 45 min, the cells were washed 3 times with PBS-1% bovine
serum albumin. They were then incubated for 30 min at 4°Cwith
anti-mouse IgM-fluorescein conjugate (Sigma Chemical Co., St. Louis,
MO) for mAbs M2590, GMR6, and 5-3 and anti-mouse Ig (H- and
L-chain-specific)-fluorescein conjugate (Sigma) for mAbs DH2, R24,
and 3F8. After washing as described above, the cells were suspended in
PBS-1 % bovine serum albumin at 1.0 x 106cells/ml and analyzed in an
Epics Profile II fluorocytometer.
Mixed Hemagglutination and Immune Adherence Assays. MHA
and IA assays for cell surface reactivities on adherent cells were carried
out as described previously (9, 15). The MHA method with protein
A-conjugated red blood cells was used for the IgG3 mAbs (DH2, R24,
and 3F8), whereas the IA method was used for the IgM mAbs (M2590
and 5-3).
Glycolipids. Gangliosides were isolated from cells as described pre
viously (16). Briefly, cell pellets were extracted with chloroform-methanol (2:1, 1:1, and 1:2, sequentially) and the glycolipid fraction was
isolated by Florisil chromatography of the acetylated derivatives (17).
Neutral glycolipids were separated from gangliosides by DEAESephadex ion exchange chromatography (18). The ganglioside fraction
was finally isolated by chromatography on a Sep-Pak C|8 (WatersMillipore, Milford, MA) column (19).
The sialic acid content of the ganglioside fractions was determined by
hydrolysis in 0.1 N HC1 at 80°Cfor l h and separation of the released
4948
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1992 American Association for Cancer Research.
ACCESSIBILITY
OF (¡ANGLIOSIDES IN TUMOR CELLS
sugars was done with high performance aniónexchange chromatography using a pulsed amperiometric detector (Dionex Corp., Sunnyvale,
CA) according to a modification (20) of the method of Hardy et al. (21).
Fatty acid composition was determined after methanolysis in l N
HC1 in methanol at 80°Cfor 6 h, and analysis of the released fatty acid
esters by gas chromatography-mass spectometry using a Delsi-Nermag
(Houston, TX) spectrometer.
Thin Layer Chromatography. Thin layer chromatography of the iso
lated gangliosides was carried out in chloroform-methanol-0.2% CaCl2
(60:40:9) on high performance silica gel G TLC plates (E. Merck), and
the gangliosides were visualized with resorcinol-HCI. Plates were then
scanned with a Shimadzu TLC scanner, and the percentage of individ
ual gangliosides determined by comparison with known amounts of
gangliosides. TLC immunostaining was performed using aluminumbacked silica gel G plates (E. Merck) as described (16), except that the
reactivity was visualized by using rabbit anti-mouse IgG-horseradish
peroxidase as the second reagent.
Radiolabeling of Cell Surface Components Containing Sialic Acid.
Melanoma cells were specifically labeled in cell surface sialic acid res
idues using a slight modification of the method of Gahmberg and
Andersson (22). A monolayer culture of melanoma (grown in a T-75
flask) was mixed with PBS and after cooling the flask to 4°C,3.0 ml of
2 mM sodium periodate in PBS were added. After incubating at 4°Cfor
30 min, 3 ml of 100 IHMglycerol-PBS were added. After 5 min at room
temperature, the cell layer was rinsed 4 times with PBS. Sodium [3H]borohydride (0.4 mCi, 7.7 Ci/mmol; New England Nuclear-Dupont,
Boston, MA) in PBS (3 ml) was then added and allowed to react for 30
min at room temperature. Unlabeled sodium borohydride (5 mg/ml, 1.0
ml) was then added. After 5 min, the cell layer was rinsed gently with 4
portions of PBS. The cells were then detached from the flask by scrap
ing, and acidic glycolipids were extracted from the cell pellet as de
scribed above and analyzed by TLC and fluorography (16).
RESULTS
Ganglioside Profiles of Melanoma Cell Panel. The total ganglioside profile of 6 melanoma cell lines was determined by
TLC and visualization with resorcinol-HCI (Fig. \A). The TLC
plates were also scanned and the percentage of individual gan
gliosides determined (Table 1). The identity of the individual
gangliosides was confirmed by antibody immunostaining5
(Fig. \B). The cell lines were chosen for their range of ganglioside composition. Thus, mouse B78 cell line has GM3 as its
only detectable ganglioside; mouse JB-RH has both GM3 and
GM2 in approximately equal amounts. SK-MEL-28 and
MeWo have GM3 and GD3 as their major ganglioside compo
nents [this pattern is typical of the majority of human mela
noma cell lines and tumor tissue samples (1-3)]. Even though
GD2 could not be detected in these 2 cell lines on TLC by
resorcinol-HCI reagent, small amounts of GD2 could be de
tected by immunostaining with anti-GD2 mAb (Fig. 1Ä);the
level of GD2 detected was estimated to represent less than 1%
of the total cell ganglioside. SK-MEL-31 and -37 are human
melanoma cell lines showing more complex ganglioside pat
terns with GM3, GM2, GD3, and GD2 all being expressed in
substantial amounts.
The total ganglioside content of the different cell lines was
determined by assaying the amount of sialic acid present in the
extracted ganglioside fractions. The lipid-bound sialic acid con
tents of the cell lines were: 0.42 tig/107 cells in B78 cells, 5.07
Mg/107 cells in JB-RH; 4.20 ¿ig/107cells in SK-MEL-28; 2.72
Mg/107 cells in MeWo; 5.00 Mg/107 cells in SK-MEL-37; and
1.08 Mg/107 cells in SK-MEL-31.
5 K. O. Lloyd, C. M. Gordon, 1.J. Thampoe, and C. DiBenedetto. unpublished
observations.
Resorcinol
B
Anti
- GD2
GM3GM2GD3GD2-
1
23456
1 2
3
Fig. 1. A, ganglioside patterns of 6 melanoma cells determined by TLC and
visualization with resorcinol-HCI reagent. Lane I, B78; Lane 2, JB-RH; Lane 3,
SK-MEL-28; Lane 4, MeWo; ¡Mne5, SK-MEL-37; Lane 6, SK-MEL-31. Ap
proximately 6 ng of NeuAc were applied per Lane except for Lane 2, which
contained 2 un NeuAc. B, TLC immunostaining of gangliosides from melanoma
cells with anti-GD2 antibody (3F8): Lane I. B78; Lane 2. JB-RH; Lane 3, SKMEL-28; Lane 4, MeWo; Lane 5, SK-MEL-31; Lane S, standard mixture of
GM3, GM2, GD3, GD2, GDla. and GTlb gangliosides. Approximately 6 Mg
NeuAc were applied per Lane.
Table 1 Ganglioside composition of panel of melanoma cell lines"
Ganglioside (%)
lineB78JB/RHSK-MEL-28MeWoSK-MEL-37SK-MEL-31GM380e4951461924GM2047WV2015GD
Cell
" Determined by TLC and detection of gangliosides by resorcinol-HCI spray as
described in "Materials and Methods."
'' Unidentified gangliosides and nonspccifically stained components.
' Percentage of total ganglioside in each cell type (average of 2 experiments).
''Small amounts of GD2 could be detected by TLC immunostaining with mAb
3F8.
The fatty acid composition of the gangliosides (GM3 and
GD3) isolated from SK-MEL-28 cells was also determined.
The major fatty acids in GM3 were C16:0 (11.9%), C18:l
(14.1%), C 18:0 (8.3%), C22:l (62.5%), C24:l (1.3%), and
C24:0 (1.9%), and in GD3 were C 16:0 (17.6%), C 18:1 (7.9%),
C18:0 (15.9%), C22:l (33.7%), C24:l (15.7%), and C24:0
(9.7%). The data on the fatty acid composition of GD3 differ
from those presented by Nudelman et al. (23), who found a
predominance of C24 fatty acids. The reason for this discrep
ancy is unclear, but it could be due to differences in culture
conditions or isolation procedures.
Reactivity of Antiganglioside Antibodies with Melanoma
Cells as Analyzed on Suspension Cells by Flow Cytometry. The
reactivity of the panel of antiganglioside mAbs with cell sus
pensions of the 6 melanoma cell lines described in the previous
section was analyzed by flow cytometry. Typical patterns are
shown in Fig. 2, and quantitative data are given in Table 2.
Several interesting aspects of the reactivity are evident. For
most of the cell lines, only one mAb reacted with the majority
of the cells in the culture with strong uniform staining. For
SK-MEL-28 and MeWo, this was mAb R24, reacting with
GD3; for SK-MEL-31, mAb 3F8 (identifying GD2) reacted
maximally; and for JB-RH only mAb 5-3 (identifying GM2)
reacted strongly. Only SK-ME-37 was an exception, showing
quite strong immunoreactivity with both mAb R24 (anti-GD3)
and 3F8 (anti-GD2). Mouse melanoma cell line B78 reacted
only with anti-GM3 mAbs (mAbs M2590, GMR6, and DH2).
The degree of reactivity depended on the antibody used: with
mAb DH2, 58.2% of the cells were positive and the intensity of
fluorescence was 7.3; with mAb M2590, 55% of the cells were
positive with a mean fluorescence intensity of 5.4; and mAb
4949
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1992 American Association for Cancer Research.
ACCESSIBILITY
OF GANGL1OSIDES
SK-MEL-28
B 78
"
•¿
/'
V
-*•
•¿-.._
v.._.
\
FLUORESCENCE
Fig. 2. Ho«c>torneine analysis of cell linci with antiganglioside antibodies
using indirect immunofluorescence method. Cell lines: A. D, and G, B78; B. E.
and H. JB-RH: C. F, and I. SK-MEL-28. Antibodies: A. B, and C, GMR6
(anti-GM3); D. E, and F. 5-3 (anti-GM2); G, H. and I. R24 (anti-GD3). The
fainter line in each panel is a control representing staining with fluoresceinatedanti-lgG alone.
GMR6 was more strongly reactive, with 90% of the cells being
positive with an intensity of 10.4. It should be noted that mAb
GMR6 shows some cross-reactivity with GDla and GMlb, but
since these gangliosides are not present in melanoma cells, this
is not a complication of the analysis.
Anti-GM3 mAbs DH2. M2590, and GMR6 did not react
significantly with any of the other cell lines tested. mAb M2590
reacted with 35% of SK-MEL-37 and GMR6 reacted with 76%
of MeWo cells, but the mean intensities were only 2.4 and 2.9,
respectively. The low reactivity of anti-GM3 with all the cell
lines, except B78, was quite suprising, considering that the
GM3 ganglioside content was quite high, ranging from about
50% of total ganglioside content for SK-MEL-28 and JB-RH to
19% for SK-MEL-37. A similar phenomenon was noted with
some of the other antibodies. Thus, even though SK-MEL-31
had substantial amounts of GD3 (20% of total gangliosides),
anti-GD3 mAb (R24) reacted only weakly with this cell line—
71% of the cells were positive but the average intensity was only
4.1. Likewise, SK-MEL-37 had 19% of its ganglioside content
as GM2, but anti-GM2 (mAb 5-3) reacted only weakly (19.7%
positive cells with a mean intensity of 6.8). As expected, antiGD2 reacted very strongly with SK-MEL-31 and SK-MEL-37,
which contain large amounts of GD2. More suprisingly, a high
proportion of SK-MEL-28 and MeWo cells, which have ex
tremely low GD2 levels, reacted with anti-GD2, although the
intensity of fluorescence was rather low. From these data, it
appears that the reactivity of an antigangliosidc mAb with a
melanoma cell is strongly influenced by the nature of the other
gangliosides expressed by the cell. There was a tendency for a
mAb to react most strongly with the most complex ganglioside
in the cell when more than one ganglioside was present.
IN TUMOR CELLS
These data were derived from cell suspensions prepared using
either an EDTA-containing or trypsin-EDTA buffer. Essen
tially the same results were obtained using either procedure.
The effect of more vigorous protease treatment was examined
by treating the cell suspensions with Pronase (using a maximum
of 1 Mgenzyme/ml enzyme, inasmuch as higher concentrations
resulted in cell lysis). Under these conditions, the reactivity of
the 2 cell lines tested (B78 and SK-MEL-28) with anti-GM3
was slightly enhanced, but reactivity with anti-GD3 was un
changed (data not shown).
Reactivity of Antiganglioside Antibodies as Analyzed on Mel
anoma Motiola) IT Cells by Mixed Hemagglutination and Im
mune Adherence Assays. In order to determine whether the
accessibility of cell surface gangliosides to antibodies was the
same on monolayer cells as with cell suspensions, the reactivity
with mAbs was studied on melanoma cells growing in the wells
of microtiter plates using two red blood cell rosetting assays:
either the MHA or the 1A. The former was used for the IgG3
antibodies using protein A-coupled erythrocytes and the latter
assay was used for IgM antibodies. In general, the results (Table
3) were very similar to those obtained with flow cytometry.
Thus, anti-GM3 mAbs were strongly reactive only with B78
cells; JB-RH cells reacted only with anti-GM2 and not with
anti-GM3 antibodies, and SK-MEL-31 was reactive only with
anti-GD2 mAb despite having substantial amounts of the other
3 gangliosides.
Accessibility of Gangliosides to Cell Surface Labeling by Periodate-[3H]Borohydride. Since GM3 was poorly recognized
by antibodies in a number of cell lines, the accessibility of this
ganglioside to chemical reagents was analyzed. Specifically, the
radiolabeling of gangliosides by treatment with sodium perio
date and sodium [3H]borohydride was analyzed in 3 cell lines
(B78, JB-RH, and SK-MEL-28). This procedure has been dem
onstrated to specifically label cell surface components contain
ing sialic acid, at least in erythrocytes (22). The results showed
that GM3 was efficiently labeled in all 3 cell lines (Fig. 3). This
was despite the fact that GM3 was poorly accessible to anti
body in JB-RH and SK-MEL-28 cells. GM2 was efficiently
labeled in JB-RH cells and GD3 was radiolabeled, although
not strongly, in SK-MEL-28 cells. These data confirm that
these gangliosides are cell surface components in the cell lines
examined.
DISCUSSION
The phenomenon
of the "crypticity" of some glycolipids in
the plasma membrane of mammalian cells has been known
for some time. The classic example is the inaccessibility of
globoside, the major neutral glycolipid of adult human eryth
rocytes, to antigloboside antibodies (24). In this instance, treat
ment of the cells with trypsin or sialidase enhances globoside
Table 2 Reactivity of antiganglioside antibodies with melanoma cells as assayed by fluorescence staining and flow cytometry
Anti-GM3DH2Cell
lineB78
Pos"58.3*
Pos0.558.7
2.4
JB/RH
6.4
2.92.5
2.2
SK-MEL-28
24.7
0.88
2.1
1.3
MeWo
11.9
76.0
3.7
5.8
1.8
SK-MEL37
1.2
3.8
5.5
35.0
2.4
SK-MEL-31%
0.4Int.6.3* I.IM2590%Pos55.1
0.6Int.5.4 1.1GMR6%Pos90.010.4
1.1Int.10.4
% Pos. percentage of cell showing positive fluorescence in comparison to staining
h Average of 2 separate experiments.
Pos2.6
7.7
12.5
19.8
2.7
6.0
4.4
1.9
4.4
4.0
97.7
36.7
64.2
2.53.2
2.5
4.1
97.6
64.8
61.7
2.9
90.0
21.1
91.1
1.1
19.7
23.8
6.8
2.0Anti-GM25-3%
70.5Anti-GD3R24Int.8.4
4.1%Pos5.0 74.1Int.3.4 3.6Anti-GD23F8%
97.8Int.3.6 35.5
with lluoresceinated anti-Ig alone; Int. mean fluorescence intensity.
4950
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1992 American Association for Cancer Research.
ACCESSIBILITY OF GANGLIOSIDES IN TUMOR CELLS
Table 3 Reactivity of antiganglioside monoclonal antibodies with melanoma cell lines assayed by red blood cell rosetting methods (MHA and IA)
Anti-GM3
DH2*
M2590"
Cell line
Anti-GM2, 5-3°
Anti-GD3, R24*
Anti-GD2, 3F8*
(2+f0032
B78JB/RHSK-MEL-28MeWoSK-MEL-37SK-MEL-3164'
<+)8
(+)32
<+)2
(±)0128(2+)0004
(+)16
(+)0128(3+)04
(+)>
(3+)1024(3+)
1024
(+)001024(3+)128(3+)>1024(3+)2 (±)00512(3+)32
" Assayed on monolayer cells with IA method.
* Assayed on monolayer cells with protein A-red blood cell MHA method.
' Last dilution showing positive cells.
d Estimate of strength of reaction: 3+, very strong; 2+, strong; +, weak; ±,very weak.
I04"-B(3H)4
GM3GM2GD3-
^^
1
2
3
Fig. 3. Fluorogram of TLC plate showing cell surface radiolabeling of gangliosides with sodium periodate-['H]borohydride
method. Lane l, B78; Lane 2,
JB-RH; Lane 3, SK MEL-28.
expression, showing that membrane proteins, glycoproteins, or
gangliosides may be shielding the globoside molecules (24).
Similarly, Stein et al. (25) showed that GM1 expression in
murine lymphocytes is enhanced after treatment of the cells
with trypsin, and Urdal and Hakomori (26) found that the
expression of Cer in murine lymphoma cells Gg3 is greatly
influenced by existing glycoconjugates. Cell surface glycolipids
also vary in their degree of exposure to enzymes such as
neuraminidase (27) and galactose oxidase (28). In the studies
presented here on melanoma cells (summarized in Fig. 4), the
effect is more specific in that we show that the accessibility of a
given ganglioside to antibody is influenced by the coexpression
of other gangliosides in the cell. The phenomenon is unlikely to
be due to the influence of other cell surface glycoproteins, in
asmuch as protease treatments had little effect on the results.
The possibility that the carbohydrate structures detected by the
various antibodies are present on glycoproteins, as well as on
glycolipids, is also unlikely since Pronase treatment enhances,
rather than diminishes, antibody reactivity. Moreover, our stud
ies have failed to demonstrate reactivity of antiganglioside an
tibodies with cell surface glycoproteins on melanoma cells using
a variety of approaches.6
The observed phenomenon was most marked with GM3.
Thus, although all the melanoma cell lines examined contained
GM3, only in B78 cells (in which GM3 is the sole ganglioside
present) was GM3 readily available to antibody. In JB-RH and
SK-MEL-28 cells, in which GM3 is one of 2 major gangliosides
present (representing about 50% of the total ganglioside com6'K. O. Lloyd, unpublished observations.
position), this ganglioside was essentially unreactive with antiGM3 antibodies. A similar finding was the relatively poor re
activity of anti-GM2 antibody with SK-MEL-37 cells and of
anti-GM2 and anti-GD3 with SK-MEL-31, even though these
cell lines contained substantial amounts of the target antigens.
In all cell lines examined, it appeared that the more complex
ganglioside present was the most readily available for reactivity
with antibody. A striking example of this was the ready detec
tion of GD2 in SK-MEL-28 and MeWo cells by anti-GD2
antibody (albeit with low fluorescence intensity), even though
this ganglioside was undetectable in the ganglioside fraction
extracted from these cells by chemical means (resorcinol-HCI)
and could only be detected by immunostaining of TLC plates
with anti-GD2 antibody (Fig. 15).
These results show that the organization of gangliosides in
the plasma membrane influences the exposure of a specific
ganglioside at the cell surface. Since the effect is found on cell
suspensions, as well as with monolayer cells, it appears unlikely
that some gross change in membrane architecture or interfer
ence by membrane proteins is responsible for this phenomenon.
Also the possibility that the undetected gangliosides are not
found in the plasma membrane, but are located intracellularly,
was discounted by showing, in 3 cell lines, that GM3 was avail
able at the cell surface for radiolabeling by small chemical re
agents that are thought not to penetrate the cell (22).
Although this phenomenon has not been studied systemati
cally, previously published studies indicate that it has been ob
served in other systems also. For instance, Kannagi et al.
(29) studied the reactivity of an antigangliotriosylceramide
(GgjOse) antibody with sublines of a mouse lymphoma cell line
and showed that its reactivity was reduced in sublines also ex
pressing gangliotetraosylceramide and GM1. Other examples,
although not commented on by the authors, can be noted in the
study of Kawashima et al. (30) on the characteristics of antiGD2 antibodies reacting with melanoma and neuroblastoma
cell lines. They showed quite strong reactivity of an anti-GD2
antibody (A 1-425) with MeWo and G361 cells even though
GD2 was very low or undetectable on TLC plates by resorcinolHCI reagent.
Although the presence of other gangliosides in the cell mem
brane seems to affect the reactivity of a specific ganglioside with
its cognate antibody, the mechanism underlying this phenom
enon is still uncertain. An obvious explanation would be that
the longer gangliosides physically block accessibility to the
shorter gangliosides, particularly when the glycolipids are clus
tered together in the membrane. In fact, Kannagi et al. (29)
proposed such a mechanism to explain the lower reactivity with
antibody of gangliotriosyl ceramide in some sublines of a mouse
lymphoma cell line as discussed above. It seems unlikely, how
ever, that a glycolipid longer than the inaccessible glycolipid by
one sugar unit (e.g., GM2 versus GM3, or GD3 versus GM3)
4951
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1992 American Association for Cancer Research.
ACCESSIBILITY OF GANGLIOSIDES IN TUMOR CELLS
Fig. 4. Summary of ganglioside content and
immunoreactivity (flow cytometry) of mela
noma cells. A, ganglioside composition (per
cent of total ganglioside in cell line), l, GM3;
2, GM2; 3, GD3; 4, GD2. fi. flow-cytometric
analysis with antibodies (percent of cells posi
tive compared with negative control). 1. DH2
(anti-GM3); 2, 5-3 (anti-GM2); 3, R24 (antiGD3); 4, 3F8 (anti-GD2). *. intensity of reac
tivity with antibodies is low (2-3 intensity
units).
B78
JB/RH
SK-MEL-28
MeWo
SK-MEL-37
B78
JB/RH
SK-MEL-28
MeWo
SK-MEL-37
SK-MEL-31
CD
O)
m
£
o
o; '
O.
B
SK-MEL-31
II
J
could be physically blocking the smaller by virtue of its size
alone. Indeed, molecular modeling studies have indicated that
the oligosaccharide moiety of GD3 is not appreciably longer
than that of GM3, inasmuch as the conformation of GD3 oli
gosaccharide was calculated to be hook-shaped with the termi
nal Neu Ac being folded back onto the G M 3 sequence.7 A pos
sible explanation of this phenomenon lies in an understanding
of the nature of glycolipid-glycolipid interactions in the cell
membrane and of interactions of the oligosaccharide chain with
the lipid surface of the membrane. These are both very poorly
understood processes. Pertinent to this topic are the recent
studies of Strömberg et al. (31) who, using molecular modeling
techniques, showed that the orientation of the saccharide chain
of globo-series glycolipids plays an important role in determin
ing their binding to Escherichia coli. Specifically, they sug
gested that the sugar chain of globotriosyl-ceramide (GbOse3) is
oriented parallel to the membrane surface, whereas the sugar
moieties of longer glycolipids (GbOse4 and GbOse5) have more
vertical orientations. In the context of the present study, it
could be suggested that the presence of longer glycolipids in a
cell causes the shorter glycolipids to assume a more horizontal,
and therefore less accessible, orientation at the cell surface.
Biophysical studies, such as those that have been carried out on
model membrane systems, may help in understanding this phe
nomenon (32-34). The tendency of certain glycolipids to coa
lesce in clusters or patches, as demonstrated in these studies,
could influence their reactivity with antibodies.
A number of studies have suggested that subtle differences in
ceramide structure, such as fatty acyl chain length, can influ
ence the antigenicity of glycolipids (29). In the present study, we
compared the fatty acid composition of GM3 and GD3 isolated
from SK-MEL-28 melanoma cells. The overall fatty acid com
position of the 2 gangliosides was found to be quite similar.
Although the GD3 sample had a slightly greater proportion of
C24 fatty acids, this difference does not seem to be large enough
to result in greater exposure of the majority of the GD3 mole
cules to antibody.
Although the function of glycolipids is still poorly under
stood, recent work has shown that they can play important roles
in cell-cell interactions by virtue of glycolipid-glycolipid (35)
and glycolipid-protein (selectin) recognition (36). They also
serve as important targets for the reactions of cells with anti
bodies. Specifically, since tumors often express altered glycolipid patterns, antibodies to glycolipids provide important
reagents for the analysis, diagnosis, and therapy of tumors.
Obviously, the degree of exposure of individual glycolipids at
the cell surface will be an important parameter in both cell-cell
and cell-antibody interactions. Specifically, the presence of a
certain glycolipid in the cell membrane does not necessarily
mean that the glycolipid is able to participate in these interac
tions. A specific example of this could be the selectively shown
by anti-GM3 monoclonal antibodies for a few melanoma cell
lines even though GM3 is widely distributed in many cell types
(6, 8, 12). Ñores et al. (7) have explained this phenomenon on
the need to exceed a certain antigen density before an antibody
can react effectively with a target cell. While this could be one
important parameter, the data presented here indicate that the
total ganglioside composition of the cells is another contribut
ing factor. The question of accessibility to antibody could also
be important in selecting antibodies for use in tumor diagnosis
and therapy. Thus, when a cell type has more than one major
glycolipid, then the use of an antibody to the more complex
glycolipid in the target cell could be most effective. To some
extent this has already happened in that anti-GD3 is the anti
body of choice for studying malignant melanoma cells (4, 14),
even though GM3 is found in approximately equal amounts in
the melanoma cells (14, 37). Also anti-GD2 is widely used in
studies on neuroblastoma (5, 38), even though GM2 is an
equally prominent ganglioside in this cell type (39). Neverthe
less, attention to the question of antigen accessibility could help
in the choice of antibodies or combination of antibodies for use
in tumor diagnosis and therapy in the future.
ACKNOWLEDGMENTS
We thank Marvin Olsen (Sloan-Kettering Institute) for assistance
with the mass spectrometric analysis of fatty acids and Ken Class for the
flow cytometric data.
REFERENCES
'S. L. Laban. H. M. Dunniway, K. O. Lloyd, and T. A. W. Koerner, Conformational analysis of ganglioside oligosaccharides using a molecular mechanics
force field, manuscript in preparation.
1. Furukawa, K., and Lloyd, K. O. Gangliosides in melanoma. In: S. Ferrane
(ed.). Human Melanoma. Immunology, Diagnostics, and Therapy, pp. 1530. Heidelberg: Springer, 1989.
2. Tsuchida. T., Ravindrananth, M. H., Saxton. R. E., and Irie, R. F. Gangli
osides of human melanoma: altered expression in vivo and in vitro. Cancer
Res.. 47: 1278-1281, 1987.
3. Lloyd, K. O.. and Old, L. J. Human monoclonal antibodies to glycolipids and
other carbohydrate antigens: dissection of the humoral immune response in
cancer patients. Cancer Res., 49: 3445-3451, 1989.
4. Houghton, A. N., Mintzer, D., Cordon Cardo, C., Welt, S., Fliegel, B.,
Vadhan, S., Carswell, E.. Melamed, M. R., Oettgen, H. F., and Old, L. J.
Mouse monoclonal lgG3 antibody detecting GD3 ganglioside: a phase I trial
in patients with malignant melanoma. Proc. Nati. Acad. Sci. USA, H2: 12421246, 1985.
5. Cheung. N. K., Neely, J. E., Landmeier, B., Nelson, D., and Miraldi, F.
Targeting of ganglioside GD2 monoclonal antibody to neuroblastoma. J.
NucÃ-.Med., 28: 1577-1583, 1987.
6. Wakabayashi, S., Saito. T., Shinohara, N., Okamoto, S., Tomioka, H., and
Taniguchi, M. Syngeneic monoclonal antibodies against melanoma antigens
with species specificity and interspecies cross-reactivity. J. Invest. Dermatol.,
83: 128-133. 1984.
4952
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1992 American Association for Cancer Research.
ACCESSIBILITY
OF GANGLIOSIDES
7. Ñores, G. A., Dohi. T., Taniguchi, M., and Hakomori, S. Density-dependent
recognition of cell surface GM3 by a certain anti-melanoma antibody, and
GM3 lactone as a possible immunogen: requirements for tumor-associated
antigen and immunogen. J. Immunol., 139: 3171-3176, 1987.
8. Yamaguchi, H., Furukawa, K., Fortunato. S. R., Livingston, P. O.. Lloyd, K.
O., Oettgen, H. F., and Old, L. J. Cell-surface antigens of melanoma recog
nized by human monoclonal antibodies. Proc. Nati. Acad. Sci. USA, 84:
2416-2420, 1987.
9. Carey, T. E., Lloyd, K. O., Takahashi. T., Travassos, L. R., and Old, L. J. AU
cell-surface antigen of human malignant melanoma: solubilization and par
tial characterization. Proc. Nati. Acad. Sci. USA, 76: 2898-2902, 1979.
10. Graf, L., Kaplan, P.. and Silagi, S. Efficient DNA-mediated transfer of se
lectable genes and unselected sequences into differentiated and undifferentiated mouse melanoma clones. Somat. Cell Mol. Genet., 10: 139-151, 1984.
11. Natoli, E. J., Jr., Livingston. P. O., Pukel. C. S.. Lloyd, K. O.. Wiegandt. H..
Szalay, J., Oettgen. H. F.. and Old, L. J. A murine monoclonal antibody
detecting /V-acetyl- and /V-glycolyl-GM2: characterization of cell surface re
activity. Cancer Res.. 46: 4116-4120. 1986.
12. Dohi. T., Ñores, G., and Hakomori. S. An IgG3 monoclonal antibody estab
lished after immunization with GM3 lactone: immunochemical specificity
and inhibition of melanoma cell growth in vitro and in vivo. Cancer Res., 48:
5680-5685, 1988.
13. Tai, T., Kawashima, !.. Ozawa, H., and Kotani, M. Generation of murine
monoclonal antibodies to ganglio-series gangliosides. Glycoconjugate J.. 8:
176, 1991.
14. Pukel, C. S.. Lloyd, K. O., Travassos, L. R., Dippold, W. G., Oettgen, H. F.,
and Old, L. J. GD3. a prominent ganglioside of human melanoma. Detection
and characterization by a mouse monoclonal antibody. J. Exp. Med.. 755:
1133-1147, 1982.
15. Real, F. X., Oettgen, H. F., and Old, L. J. Detection and analysis of anti
bodies against cancer cell surface antigens in the sera of cancer patients. ///.
N. R. Rose, H. Friedman, and J. L. Fahey (eds.). Manual of Clinical Immu
nology, pp. 798-802. Washington, DC: American Society of Microbiology,
1986.
16. Furukawa, K., Clausen, H., Hakomori, S., Sakamoto, J., Look, K., Lundblad.
A., Mattes. M. J.. and Lloyd, K. O. Analysis of the specificity of five murine
anti-blood group A monoclonal antibodies, including one that identifies type
3 and type 4 A determinants. Biochemistry'. 24: 7820-7826, 1985.
17. Saito, T.. and Hakomori, S-l. Quantitative isolation of total glycosphingolipids from animal cells. J. Lipid Res., 12: 257-259, 1971.
18. Ledeen, R. W., and Yu, R. K Gangliosides: structure, isolation, and analysis.
Methods Enzymol.. 83: 139-191, 1982.
19. k muÃu, S. K.. and Suzuki, A. Simple micro-method for the isolation of
gangliosides by reverse-phase chromatography. J. Chromatogr., 224: 249256. 1981.
20. Lloyd, K. O., and Savage, A. High performance anión exchange (HPAE)
chromatography of reduced oligosaccharides from sialomucins. Glycoconju
gate J., 8: 493-498. 1992.
21. Hardy. M. R., Townsend, R. R.. and Lee, Y. C. Monosaccharide analysis of
glycoconjugates by aniónexchange chromatography with pulsed amperometric detection. Anal. Biochem.. 170: 54-62, 1988.
22. Gahmberg, C. G., and Andersson, L. C. Selective labeling of cell surface
sialoglycoprotcins by periodate-tritiatcd borohydride. J. Biol. Chem., 252:
5888-5894. 1977.
23. Nudelman, E., Hakomori. S-L. Kannagi. R., Levery, S., Yeh. M-Y.. Hell
ström, K. E., and Hellström, I. Characterization of a human melanoma-
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
IN TUMOR CELLS
associated ganglioside antigen defined by a monoclonal antibody, 4.2. J. Biol.
Chem., 257: 12752-12756. 1982.
Hakomori, S-L, Teather, C., and Andrews, H. Organizational difference of
cell surface "hematoside" in normal and virally transformed cells. Biochem.
Biophys. Res. Commun.. 33: 563-568. 1968.
Stein, K. E., Schwarting. G. A., and Marcus, D. M. Glycolipid markers of
murine lymphocyte subpopulations. J. Immunol., 120: 676-679, 1978.
Urdal, D. L., and Hakomori, S-l. Characterization of tumor-associated ganglio-n-triaosylccramide in mouse lymphoma and the dependency of its expo
sure and antigcnicity on the sialosyl residues of a second glycoconjugate. J.
Biol. Chem., 258: 6869-6874, 1983.
Barton, N. W., and Rosenberg, A. Action of Vibrio cholerae neuraminidase
(sialidase) upon the surface of intact cells and their isolated sialolipid com
ponents. J. Biol. Chem., 248: 7353-7358, 1973.
Gahmberg. C. G., and Hakomori. S. External labeling of cell surface galac
tose and galactosamine in glycolipid and glycoprotein of human erythrocytes.
J. Biol. Chem., 248: 4311-4317, 1973.
Kannagi, R.. Stroup. R., Cochran, N. A., Urdal, D. L., Young, W. W., Jr.,
and Hakomori. S. Factors affecting expression of glycolipid tumor antigens:
influence of ceramide composition and coexisting glycolipid on the antigenicity of gangliotriosyl-ceramide in murine lymphoma cells. Cancer Res., 43:
4997-5005, 1983.
Kawashima, L, Tada. N.. Fujimori, T., and Tai, T. Monoclonal antibodies to
disialogangliosides: characterization
of antibody-mediated
cytotoxicity
against human melanoma and neuroblastoma cells in vitro. J. Biochem., 108:
109-115, 1990.
Strömberg, N., Nyholm. P-G., Pascher, J., and Normark, S. Saccharide
orientation at the cell surface affects glycolipid receptor function. Proc. Nati.
Acad. Sci. USA, 88: 9340-9344, 1991.
Maggio, B., Ariga, T., and Yu. R. K. Ganglioside GD3 lactones: polar head
group mediated control of the inlermolccular organization. Biochemistry, 29:
8729-8734, 1990.
Ali, S., Brockman, H. L., and Brown, R. E. Structural determinants of
miscibility in surface films of galactosylceramide and phosphatidylcholine:
effect of unsaturation in the galactosylceramide acyl chain. Biochemistry, 30:
11198-11205, 1991.
Rock, P., Allietta, M., Young, W. W.. Jr., Thompson, T. E.. and Tillack. T.
W. Organization of glycosphingolipids in phosphatidylcholine bilayers: use
of antibody molecules and Fab fragments as morphologic markers. Biochem
istry, 29: 8484-8490. 1990.
Kojima, N., and Hakomori, S. Cell adhesion, spreading, and motility of
GM3-expressing cells based on glycolipid-glycolipid interaction. J. Biol.
Chem., 266: 17552-17558. 1991.
Brandley, B. K., Swiedler, S. J., and Robbins, P. W. Carbohydrate ligands of
the LEC cell adhesion molecules. Cell, 63: 861-863, 1990.
INni. nikalian. J., Zwingelstein. G., and Dore. J. Lipid composition of human
malignant melanoma tumors at various levels of maligant growth. Eur. J.
Biochem.. 94: 19-23, 1979.
Schulz, G., Cheresh, D. A., Varki, N. M., Yu, A.. Staffileno, L. K., and
Reisfeld, R. A. Detection of ganglioside GD2 in tumor tissues and sera of
neuroblastoma patients. Cancer Res., 44: 5914-5920, 1984.
Wu, Z-L., Schwartz, E., Seeger. R., and Ladisch, S. Expression of GD2
ganglioside by untreated primary human neuroblastomas. Cancer Res., 46:
440-443, 1986.
Svennerholm, L. Chromatographie separation of human brain gangliosides.
J. Neurochem.. 10: 613-623, 1963.
4953
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1992 American Association for Cancer Research.
Cell Surface Accessibility of Individual Gangliosides in
Malignant Melanoma Cells to Antibodies Is Influenced by the
Total Ganglioside Composition of the Cells
Kenneth O. Lloyd, Claudia M. Gordon, Immac J. Thampoe, et al.
Cancer Res 1992;52:4948-4953.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/52/18/4948
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1992 American Association for Cancer Research.