1983;43:3808-3813. Published online August 1, 1983.Cancer Res Stephan Ladisch, Baiba Gillard, Clifford Wong, et al. Cell GangliosidesShedding and Immunoregulatory Activity of YAC-1 Lymphoma
Updated Version http://cancerres.aacrjournals.org/content/43/8/3808
Access the most recent version of this article at:
Shedding and Immunoregulatory Activity of YAC-1 Lymphoma CellGangliosides1
Stephan Ladisch,2 3 Baiba Gillard, Clifford Wong, and Lisa Ulsh
Division oÃHematology-Oncology and Gwynne Hazen Cherry Memorial Laboratories, Department of Pediatrics, UCLA School of Medicine, Los Angeles, California 90024
ABSTRACT
YAC-1 lymphoma cells, both when cultured in vitro and when
passaged in ascites form in vivo, synthesize gangliosides (meansof 22.1 and 14.7 nmol lipid-bound sialic acid isolated per 108
cells, respectively) with potent inhibitory effects on mitogen- andantigen-induced lymphoproliferation: 10 to 30 nmol highly purifiedYAC-1 gangliosides/ml caused >90% inhibition of proliferative
responses of murine lymphocytes to concanavalin A, lysozyme(a soluble specific antigen), and allogeneic cells (mixed-lympho
cyte response). Measurable quantities of these gangliosideswere shed by the tumor cells in vitro and also were recoveredfrom the ascites fluid in vivo. Furthermore, the gangliosidesisolated from ascites fluid (mean of 15.3 nmol/ml) had inhibitoryactivity of a magnitude similar to that of the gangliosides isolatedfrom the tumor cells. Therefore, significant inhibition of normallymphoproliferative responses by tumor-derived gangliosides oc
curred at ganglioside concentrations which are actually presentin the fluid surrounding the tumor cells in vivo. These resultssupport the hypothesis that shedding of gangliosides may serveto protect tumor cells from host immune destruction.
INTRODUCTION
Certain gangliosides, which are sialic acid-containing glyco-
sphingolipids, are known to have potent immunoregulatory properties, including inhibitory effects on antibody production in vivo(1), on the generation of antibody-synthesizing cells in vitro (22),
and on lymphocyte proliferative responses to mitogens andantigens in vitro (17, 33). Both bovine and human brain gangliosides, and gangliosides shed by antigen-stimulated lymphocytes
(4) have been shown to exhibit such immunoregulatory properties.
The hypothesis that shedding of gangliosides may contributeto the escape of tumor cells from immune destruction in vivo (5)is supported by the above findings and by observations ofelevated levels of circulating gangliosides in both animals andhumans with cancer (8, 16, 23, 28). It has further been suggested, although not proven, that tumor-derived gangliosides
may be a cause of generalized immunosuppression frequentlyobserved in tumor-bearing hosts (17).
To directly test the hypothesis that purified gangliosides, isolated from homogeneous tumor cell populations, may have suchimmunoregulatory activity, we have studied a murine tumor
1This work was supported by USPHS Grant CA 27701 from the National Cancer
Institute, Department of Health and Human Services.2 Recipient of Research Career Development Award 1 K04 CA 00821 from the
National Cancer Institute, and Scholar of the Leukemia Society of America.3To whom requests for reprints should be addressed, at Division of Hematology/
Oncology, Department of Pediatrics, UCLA School of Medicine, Los Angeles, Calif90024.
Received September 20, 1982; accepted May 5. 1983.
system, the YAC-1 lymphoma. We have found: (a) that YAC-1
cells cultured in vitro synthesize gangliosides which inhibit mitogen- and antigen-induced lymphoproliferative responses and
which are actively shed by these cells; and (b) that inhibitory,chromatographically similar gangliosides are also found in YAC-
1 cells propagated in vivo and in the ascites fluid surroundingthese cells. Together, the results provide direct evidence supporting an immunoregulatory role of gangliosides synthesized bytumor cells in vivo.
MATERIALS AND METHODS
Cell Culture. YAC-1 lymphoma cells were cultured in complete me
dium consisting of Roswell Park Memorial Institute Medium 1640 supplemented with glutamine (2 mw) and 10% heat-inactivated FCS.4 Gan
gliosides were isolated from cells and conditioned medium harvestedduring the logarithmic phase of cell growth to minimize artifacts due tocell death. One to 2 x 105 cells/ml culture medium were incubated in a
humidified 5% CO2:95% air atmosphere. After 72 hr, during which timethe cell concentration had routinely increased 10-fold and cell viability
was >99%, as assessed by trypan blue dye exclusion, the cultures wereharvested. The washed cell pellets, isolated by centrifugation at 300 xg for 10 min, and the cell-free conditioned medium, obtained by centrifugation at 1000 x g for 15 min, were stored at -80° under N2 until the
extractions were performed.YAC-1 lymphoma cells proliferating in vivo were obtained by serially
passaging these cells in the peritoneal cavity of syngeneic A/Sn mice(The Jackson Laboratory, Bar Harbor, Maine). Ten to 14 days after theinjection of 5 x 106 cells, the ascites fluid (1 to 2 ml/mouse, containing
2 to 3 x 10e tumor cells; >95% viable) was harvested. After hypotonielysis of contaminating erythrocytes at 4°, cells not morphologically
identifiable as YAC-1 lymphoma cells were rare, and less than 0.5% of
the cells were macrophages, as identified by staining with nonspecificesterase (10).
Ganglioside Isolation and Quantitation. Total lipid extracts of thecell pellets (108 to 1010cells), conditioned medium, and ascites fluid were
obtained by extracting the lyophilized starting material with 20 volumes(or 10 ml/1O8 cells) CM 1:1 (all solvent ratios are by volume) for 18 hr at4°with stirring and reextracting the residue with the same volume of
CM 2:1 for 3 hr. The extracts were pooled and dried by rotary evaporationfollowed by oil pump vacuum to remove remaining traces of solvent andwater.
Gangliosides were purified from the total lipid extracts of cells andascites fluid by: (a) the method of Irwin and Irwin (6), modified by usingboth 20 column volumes CM 2:1 and then 4 column volumes of CMW65:25:4 to completely elute neutral lipids and phospholipids prior to therecovery of gangliosides in CMW 50:50:15; or (b) by DEAE-Sephadex
chromatography adapted for extraction of small samples (32). The latterprocedure demonstrated that the less polar gangliosides (e.g., GM4and
' The abbreviations used are: PCS, fetal calf serum; CM, chloroform:methanol;
CMW, chloroform:methanol:water; LBSA, lipid-bound sialic acid; TLC, thin-layerchromatography; Con A, concanavalin A; MLR, mixed lymphocyte response; HEL,hen eggwhite lysozyme. Gangliosides are identified according to the nomenclatureof Svennerholm (31).
QMS),which are not recovered in the gangliosidefraction (CMW 50:50:15)by the method of Irwin and Irwin (26), were not present in the YAC-1cells we studied, thus allowing either procedure to be used, with qualitatively and quantitatively similar recovery of YAC-1 cell gangliosides.Partial purification of gangliosides from conditioned medium, which ismore difficult because of the much lower concentrations of gangliosidesin this material, was accomplished by partition of the dried total lipidextract in diisopropyl ether:butanol:water (6:4:5) (13), which we previously found to efficiently separate less polar lipids from gangliosides inserum (34). All gangliosidepreparations were further purified by removalof CM 1:1 insoluble material by centrifugation and by Sephadex G-50gel filtration to remove low-molecular-weightcontaminants(32).Sampleswere stored under N2in CM 1:1 at -20°.
Gangliosideswere quantitated as nmol LBSA by the modified colorimetrie resorcinol assay (15). TLC (2) of gangliosides was performedusing 10 cm x 10 cm precoated Silica Gel 60 high-performance TLCplates (E. Merck, Darmstadt, Germany)which had been preactivated byheating to 90°for 45 min. The plates were developed in CM:0.25%aqueous CaCI2•2 H2O60:40:9. Gangliosideswere visualizedas purplebands with resorcinol-HCI reagent (30). Gangliosides used for the im-munologicalstudies were found to be free of contaminatingphospholipidsby molybdenum blue stain (3) and neutral glycosphingolipids by orcinolstain (29). They were also negative for protein by Coomassie blue stainof ganglioside preparations subjected to sodium dodecyl sulfate gelelectrophoresis.
Autoradiographyof YAC-1 Cell Gangliosides. YAC-1 cells werecultured in vitro (106cells/ml complete medium)for 24 hr with 0.5 ¿¡CiD-[1-14C]glucosamineHCI(specific activity, 54.2 mCi/mmol)and 0.5 /¿CiD-[1-14C]galactose(specificactivity, 56.5 mCi/mmol; New EnglandNuclear,
Boston, Mass.) per ml culture medium (7, 25). The cells were thenharvested, and the gangliosideswere isolated and purified as describedabove. YAC-1 cells passagedin vivo were radiolabeledin vivo by the i.p.injectionof 1.0 ^Ci of each sugar into a mouse bearingYAC-1 lymphomafor 10 days. Six hr later, the cells were harvested, and gangliosideswereisolated. Purified gangliosides were chromatographed as describedabove, and the plates were exposed to Kodak (Rochester, N. Y.) XRP-5 X-ray film at -80°.After autoradiography, TLC plates were visualized
with resorcinol spray reagent to allow comparison of autoradiographicbands with resorcinol-positivebands.
Lymphocyte Proliferation Assays. YAC-1 lymphomagangliosideswere tested for modulatory effects on the proliferative responses ofmurine lymphocytes to the mitogen, Con A, to allogenelccells (bidirectional MLR), and to a soluble specific antigen, HEL.
Splenocytes of DBA/2 and A/Sn mice (TheJackson Laboratory)wereisolated, freed of contaminatingerythrocytes by hypotonie lysis, washed,and resuspended in complete mediumconsisting of Dulbecco's modifiedEagle's medium with 4.5 g glucose/liter (Grand Island Biological Co.,Grand Island, N. Y.) and 5% heat-inactivated PCS, supplemented with25 mw 4-(2-hydroxyethyl)-1-piperazineethanesulfonicacid, 0.07 mw 2-mercaptoethanol, penicillin (50 units/ml), and streptomycin (50 Mg/ml).Lymph node cells were isolated from inguinal,periaortic, and peritonealnodes of female C57BL/6 mice at 9 days after immunization with 100tig HEL. These cells were suspended in complete medium consisting ofClick's medium with 0.5% syngeneic mouse serum (19), supplemented
as above. Cell viability was consistently >90%.Gangliosideswere dissolved in complete medium, sonicated under N2
for 2 min at room temperature in a Branson bath sonicator (ScientificProducts, Irvine, Calif.), and added to the cell suspensions immediatelybefore (HEL and MLR) or 18 hr before (Con A) addition of the stimulant.Final culture volumes, in flat-bottomed 96-well Microtiter plates (FlowLaboratories, Inglewood, Calif.), were 100 n\ for the Con A and 200 ¡Afor the HEL and MLR assays. Cell and stimulant concentrations (per ml)were 1 x 106 DBA/2 splenocytes (Con A, 3 ng), 2 x 106 lymph nodecells (HEL, 100 ^g), and 1 x 106each of DBA/2 and A/Sn splenocytesin the MLR. The plates were incubated at 37°in a humidifiedatmosphereof either 5% CO2:95%air (72 hr for Con A-induced proliferationand 120hr for the MLR) or 2% CO2:98%air (for 120 hr for HEL-inducedprolifer-
Shedding and Immunoregulation by Tumor Cell Gangliosides
ation). Cultures were pulsed with 0.5 to 1 nC\ [3H]thymidine (specific
activity, 6.7 mCi/mmol)and harvestedon glass fiber paper (A. H. Thomas,Philadelphia,Pa.). [3H]Thymidineuptake was quantitated by ¿¡-scintillation counting. Meannet [3H]thymidineuptake was calculated by subtrac
tion of the cpm incorporated by triplicate unstimulated cultures fromthose of triplicate stimulated cultures. All conditions, including cell andstimulant concentrations and culture duration, were selected to result inoptimal cellular [3H]thymidineuptake.
Statistical Analysis. Student's t test was used to determinethe
significanceof differences between mean ganglioside concentrations.
RESULTS
Characterization of YAC-1 Cell Gangliosides. Four major
gangliosides, with TLC mobility between GMi and GDia standardbrain gangliosides, were seen on TLC of the gangliosides isolatedfrom YAC-1 cells (Fig. 1). At least 8 other minor bands were also
visible, suggesting a high degree of structural diversity of the cellgangliosides. The patterns of gangliosides isolated from cellscultured in vitro and from cells passaged in vivo were very similar(Fig. 1), indicating that the ganglioside profile of YAC-1 cells is
independent of their propagation in vitro versus in vivo.Each separate preparation of YAC-1 cell gangliosides was
quantitated and qualitatively characterized by TLC, becausevariants of YAC-1 cells have been found to have strikingly
different ganglioside patterns (36). The ganglioside content ofcells cultured in vitro (n = 7) or in vivo (n = 6) [22.1 ±3.7 (S.E.)versus 14.7 ±4.7 nmol LBSA per 10" cells, respectively] was
not significantly different (p > 0.2). Moreover, the TLC mobilitiesof the major bands isolated from the individual preparations werevery reproducible, suggesting that no YAC-1 variants had developed either in vitro or in vivo during the 24-month period of these
studies.To exclude the possibility th?t the cell ganglioside patterns
beled precursor sugars into newly synthesized gangliosides bothin vitro and in vivo. As seen in Fig. 2, the autoradiographicpattern in each experiment was identical to the gangliosidepattern visualized by resorcinol staining, confirming that the
~ GD1o
" GD1b
- G-M- Origin
BFig. 1. YAC-1 lymphoma cell gangliosides. Total ganglioside fractions were
isolated as described in "Materials and Methods" from: YAC-1 lymphoma cells
cultured in vitro (Lane A) or propagated in vivo (Lane B). and bovine brain(standards, Lane C); 7 nmol LBSA were spotted per lane. All bands are resorcinolpositive.
BFig. 2. Autoradiography of YAC-1 lymphoma cell gangliosides.TLC patterns of
radiolabeledYAC-1 lymphoma cell gangliosideswere assessed both by resorcinol-HCI staining (Lanes A to C) and by autoradiography of the same plate (Lanes Dand £).Gangliosidesof YAC-1 cells cultured in vitro (4.4 nmol LBSA; 8000 cpm)visualized by resorcinol-HCIstaining (Lane B) and by autoradiography (Lane D),and of YAC-1 cells propagated in vivo (5 nmol LBSA; 8000 cpm) visualized byresorcinol-HCIstaining (LaneC) and by autoradiography (LaneE) are shown. LaneA. G.»and bovine brain standard gangliosides (5 nmol LBSA). Bands above GM1in Lane C are resorcinol negative; all other bands in Lanes A to C are resorcinolpositive.
gangliosides isolated from the cells had actually been synthesized by the cells.
Shedding of Gangliosides by YAC-1 Cells. In a series ofexperiments, gangliosides in conditioned medium, generated bythe 72-hr culture of YAC-1 cells in vitro, and in control unconditioned medium (i.e., PCS gangliosides) were isolated and quan-titated as described in "Materials and Methods." A net ganglio-
side concentration of 0.36 ±0.14 nmol LBSA per ml medium(n = 3), due to shedding of gangliosides by tumor cells, wasfound. Estimating the average cell concentration during the 72hr to have been 1 x 10sper ml, these results indicate a calculatedapproximate shedding rate of 12 nmol LBSA per 108cells/24 hr,
or about 50% of the cell ganglioside content per 24 hr. Thepattern of gangliosides isolated from the conditioned mediumshowed striking similarity to the pattern of the major cell-synthesized gangliosides (Fig.3). The additional bands of higher mobilitynot seen in the cell ganglioside pattern (Fig. 3) were derived fromthe FCS which contains GM3as the major ganglioside component.Together, the quantitative and qualitative findings demonstratesubstantial shedding of cell gangliosides by YAC-1 cells culturedin vitro.
In the ascites fluid surrounding YAC-1 cells propagated in vivo,the quantity of ganglioside present (15.3 ±1.7 nmol/ml; n = 4)was much higher than in the in vitro system. To the extent thatthese gangliosides were shed by the tumor cells, this higherconcentration may be a reflection of an approximately 100-foldhigher cell density (i.e., 2 to 3 x 108/ml)in vivo. As in the in vitro
system, the TLC pattern of gangliosides isolated from the ascitesfluid included the major bands seen in the TLC pattern ofgangliosides isolated from the cells passaged ¡nvivo (Fig. 3).The most prominent ganglioside in the ascites fluid, with TLCmobility slightly greater than that of GMiganglioside, was, however, only faintly visible on the TLC of the cell gangliosides (Fig.
3). Possible explanations for this difference in the relative proportions of specific gangliosides found in the cells and in thesurrounding in vivo fluid remain to be investigated. They wouldinclude: (a) alteration of cell shedding of gangliosides when cellsare passaged in vivo instead of in vitro; (b) presence of gangliosides derived from serum or other host tissue, rather than shedby the tumor cells; and (c) biochemical metabolism of the gangliosides once shed into the ascitic fluid. Whatever the basis forthe substantial levels of gangliosides in the ascites fluid, thesefindings imply that there is a significant ganglioside concentrationin the local environment of the tumor cells in vivo.
Modulationof Lymphocyte Proliferative Responses by TumorCell Gangliosides. Modulationof the proliferativeresponseof normal murine splenocytes to the mitogen, Con A, by gangliosides isolated from YAC-1 lymphoma cells is shown in Chart 1.A dose-related inhibition of this lymphocyte proliferative responsewas observed. Concentrations of 25 to 30 nmol YAC-1 cellgangliosides/ml, or 2.5 to 3 nmol/2 x 105 splenocytes/culture,caused 90% inhibition of net [3H]tnymidine uptake by stimulated
splenocytes. This marked inhibitory activity was characteristic ofgangliosides isolated both from YAC-1 cells cultured in vitro andthose passaged in vivo (Chart 1). Interestingly, gangliosidesisolated from the ascites fluid surrounding YAC-1 cells prolifer-
GM3
GM1
GD1a
GDIÃ’GT1
Origin
~ GM3
—Origin
Fig.3. Shedding of YAC-1 lymphoma cell gangliosides. Gangliosidesof YAC-1lymphoma cells cultured in vitro (Lane B) or propagated in vivo (Lane D) arecompared with gangliosidesisolated from the respective surrounding fluid obtainedin the same experiment: conditioned medium (Lane C) and ascites fluid (Lane E);Lanes A and F, Gunand bovine brain standard gangliosides. Each lane contained10 nmol LBSA. Both thin-layer chromatograms (Lanes A to C and D to F) werestainedwith resorcinol-HCI.Bands above GUI in Lane C are resorcinol negative; allother bands are resorcinol positive.
-0.5
-1.0-
9
3-1.5
25 50 T5 0 25Lipid-bound sialtc acid (nmol/ml)
50
50co
!ö£C
90 S5
95
>99
Chart 1. Inhibition of Con A-induced murine lymphoproliferative responses bygangliosides isolated from YAC-1 lymphoma cells and from ascites fluid. Points,mean and bars, range (1 to 3 separate experiments, each performed in triplicate)of the change in net [3H]thymidineuptake of Con A-stimulated splenocyte culturescausedby incubation in the presenceof the indicatedconcentration of gangliosides.A. gangliosides isolated from cells cultured in vitro (•);B, gangliosides isolatedfrom cells propagated in vivo (•)and from the surrounding ascites fluid (V).
Shedding and Immunoregulation by Tumor Cell Gangliosides
TabteiYAC-1lymphoma cell gangliosides inhibit mitogen- and antigen-induced lymphoproliferativeresponses
GangliosideSequential
purificationsteps6Control
(no addedganglioside)Modified
method of Irwin andIrwin0Sephadex
G-50 gelfiltration8Preparative
TLC'Concentration
(nmol LBSA/ml)3030
201030Con
A ConA74
111.3(88)"7.9
(89) 2.6 (76)40 (45)70 (5)[3H]Thymidine
uptake8Stimulant
Con AMLR20
532.1
(90) 2.530
421.7(92)
0.4(95)
(43)(21)(>99)HEL430.01
(>99)0.74 (98)0.44(99)0.02
(>99)a Mean net cpm x 10~3of triplicate cultures; S.E. <10%." Sequentialpurification procedures do not abrogate inhibitory activity.'' Removalof phospholipids,neutral lipids. and neutral glycosphingolipids.d Numbers in parentheses, percentageof inhibitionversus control.e Removalof nucleotide sugars, low molecularweight peptides, and salts; removed fraction was not inhibitory (not shown).' Compared to modified Irwin and Irwin method alone, the additional gel filtration and preparativeTLC steps resulted in a 45-fold reduction in protein, to 0.002 mol%.
ating in vivo had inhibitory effects on normal mitogen-induced
lymphoproliferation which were at least as potent as those ofthe cell-derived gangliosides: 25 nmol/ml caused 90% inhibition
of lymphoproliferation (Chart 1). At their actual average concentration in the ascites fluid (15 nmol/ml), these gangliosidescaused 78% inhibition of normal lymphoproliferation, providingstrong evidence supporting a potential in vivo inhibitory effect ofthese gangliosides.
Inhibition of [3H]thymidine uptake by gangliosides was docu
mented to reflect actual inhibition of cell proliferation, in that lackof an increase in the cell count in cultures exposed to Con A andgangliosides (30 nmol/ml) was observed, in comparison to adoubling of the cell count in cultures exposed to Con A alone. Incontrast, the viability of unstimulated splenocytes (assessed bytrypan blue dye exclusion) was not affected by culture in thepresence of these gangliosides, thereby excluding nonspecificcytotoxicity as the cause of the inhibition. In separate experiments in which the concentration of Con A was varied from 0.75to 50 fig/ml (15 times the optimal concentration), the degree ofinhibition caused by 30 nmol YAC-1 cell gangliosides/ml was
constant (85 to 90%). The lack of decreasing inhibition withincreasing Con A concentration provides evidence that, as observed previously in the case of mixed brain gangliosides (17,33), the tumor cell gangliosides did not inhibit the lymphoproliferative responses to Con A merely by direct binding to, andinactivation of, the mitogen.
The effects of YAC-1 cell gangliosides on lymphoproliferative
responses to 2 other stimulants, allogeneic cells (MLR) and asoluble specific antigen (HEL), were also studied to determinewhether inhibitory effects of these tumor cell gangliosides wererestricted to proliferative responses induced by Con A. In parallelexperiments, proliferative responses to all 3 stimulants weremarkedly inhibited by the YAC-1 gangliosides (Table 1). There
fore, inhibition of proliferation was not dependent on the natureof the stimulant and, thus, reflects a general inhibitory effect onmitogen- and antigen-induced lymphoproliferation. Moreover, the
results in Table 1 suggest an even more marked inhibitory effectof these tumor cell gangliosides on the proliferative responsesto specific antigens (i.e., MLR and HEL) than to the nonspecificmitogen, Con A, an observation also made in testing these samegangliosides in human lymphoproliferation assays (14).
Recently, it has been shown that protein contamination may
account for some of the biological effects attributed to gangliosides (11). To exclude the possibility that contaminating proteinmight be responsible for the inhibition of lymphoproliferation, wetested the effect of sequential purification procedures on proteincontent and on inhibitory activity (Table 1) of gangliosides isolated from YAC-1 cells cultured in vitro. Gangliosides recovered
in CMW 50:50:15 and redissolved in CM 1:1 contained approximately 4 /¿gprotein/100 nmol LBSA. The further purificationprocedures described in "Materials and Methods," coupled with
an additional final step, preparative TLC, resulted in a 45-fold
reduction in protein content (to approximately 0.1 ng/iQQ nmolLBSA, or 0.002 mol% protein) in the purified gangliosides.5 The
quantitation method [autoradiographic visualization (21) of theprotein which had been 125l-labeled(20) and resolved by sodium
dodecyl sulfate gel electrophoresis (21)] also demonstrated thatthe minute quantity of protein still associated with the gangliosides after preparative TLC was qualitatively identical (apparentmolecular weight of the major band M, = 68,000) to that presentat the earlier stages of the purification.5 This purification se
quence resulted in no reduction of inhibitory effects of the YAC-
1 cell gangliosides on lymphoproliferative responses to Con A,MLR, or HEL (Table 1). Therefore, the results provide very strongevidence supporting the conclusion that the inhibitory activity isdue to ganglioside.
DISCUSSION
The YAC-1 lymphoma was selected as the tumor system for
the present studies of tumor cell gangliosides and their effect onnormal lymphocyte proliferative responses, because it had beenshown that YAC-1 cells propagated i.p. in vivo release, into the
surrounding ascites fluid, membrane vesicles enriched in lipidcontent over that of the membrane itself (24). Therefore, itseemed possible to us that immunologically active polar glycosphingolipids (gangliosides) might also be shed by these cells insubstantial quantities. We investigated both the gangliosides ofYAC-1 cells cultured in vitro and of YAC-1 cells proliferating in
vivo: (a) in order to allow comparison of the characteristics ofthe cell gangliosides isolated under the 2 conditions; and (b)because the in vitro system would allow isolation of gangliosides
synthesized strictly by the tumor cells rather than by normal orreactive host cells, as might be the case in vivo; while (c) the invivo system, by yielding data concerning ganglioside concentrations in the actual local environment of the tumor as well, mightprovide important information regarding the potential in vivosignificance of our immunological findings.
YAC-1 cells, whether cultured in vitro or propagated in vivo,
contained similar quantities of gangliosides, with similar qualitative patterns. In both cases, identity between ganglioside patterns visualized by autoradiography and by resorcinol stainingdocumented that the cell-derived gangliosides had been synthesized de novo by the YAC-1 cells. This finding is of importance,
because it suggests that the immunoregulatory activity of thegangliosides we have isolated represents a true tumor cell-
associated effect, rather than effects of gangliosides in theculture medium (in vitro) or derived from the host (in vivo).
Gangliosides which we had thus shown to have been synthesized by the tumor cells in vitro were tested in lymphoproliferationassays. The marked inhibitory effect of these highly purifiedgangliosides, which was independent of the nature of the stimulating antigen and not due to trace contaminating protein,therefore provides an unequivocal demonstration of the immu-nosuppressive activity of strictly tumor cell-derived gangliosides.
Using highly purified gangliosides isolated from cells cultured invitro avoided the problem, recognized by others (12), associatedwith ascribing immunoinhibitory activity of gangliosides isolatedfrom tumor-containing tissue (i.e., AKR lymphomatous thymus)
to gangliosides of the tumor cells as opposed to gangliosides ofother cells present in the tissue. Furthermore, the demonstrationof quantitatively similar dose-related inhibition of lymphoproliferation by gangliosides isolated from YAC-1 cells which had been
propagated in vivo shows that tumor cells proliferating in vivoalso synthesize such immunoregulatory gangliosides.
Another major objective of our studies was to assess sheddingof gangliosides in the YAC-1 system. Shedding of gangliosides
by tumors has been suggested by findings of increased circulating ganglioside levels in tumor-bearing hosts (8, 16, 23, 28).More recently, the shedding of a specific tumor-associated gan
glioside was documented by the binding, to sera of patients withcolon carcinoma, of a monoclonal antibody found to be directedto a ganglioside specific for colon carcinoma (9,18). In the tumorsystem we studied, the quantities of gangliosides shed by thetumor cells in vitro were such as to be detectable by the muchless-sensitive standard TLC staining techniques, suggesting that
the degree of shedding of gangliosides by tumor cells may begreater than previously recognized.
The demonstration of ganglioside synthesis and shedding byYAC-1 cells in vitro provides the background for addressing the
question of the potential in vivo relevance of immunoregulatorytumor cell gangliosides, i.e., do gangliosides associated withproliferating tumor cells in vivo have immunoregulatory properties, and are these gangliosides present in biologically significantquantitites in vivo? If such were the case, ganglioside shedding,which has been proposed as a possible mechanism of tumor cellescape from immune surveillance (5), could indeed, as has beensuggested (17), contribute to the widely reported inhibitory effects of malignant ascites fluid and plasma of tumor-bearinghosts on immune responses such as mitogen-induced lympho
proliferation. Our results showing significant inhibitory activity ofthe purified gangliosides isolated from the ascites fluid surround
ing the YAC-1 cells proliferating in vivo, with a very similar dose-
response curve to that of the cells themselves, are in accordwith this hypothesis.
Gangliosides isolated from normal tissue, such as brain, alsoinhibit Con-A-induced murine lymphoproliferation (17), with the
degree of inhibition being quantitatively similar to that of theYAC-1 cell gangliosides we studied. This suggests that, inde
pendent of their cell of origin and molecular structure, gangliosides may inhibit lymphoproliferation by a common mechanism.However, it should be emphasized that, despite such similarinhibitory activities, an important difference does exist in vivobetween brain and tumor cell gangliosides, i.e., while braingangliosides are normally not found in the peripheral circulation,tumor cell gangliosides have been documented to be present inthe local environment of the tumor (present study) and in theperipheral circulation of tumor-bearing hosts (9, 18, 23). There
fore, gangliosides shed by tumor cells would be much more likelythan normal brain gangliosides to come into contact with, andsignificantly affect, the immune system in vivo. The potential forimmunosuppression by tumor-associated gangliosides in vivo is
further supported by the finding that the actual gangliosideconcentrations in the malignant ascites fluid associated with theYAC-1 lymphoma are high enough to be expected to have a
significant inhibitory effect on murine lymphoproliferative responses.
A direct relationship between total membrane sialic acid (protein- and lipid-bound) and metastatic potential of tumor cells has
recently been suggested (37), although conflicting results hadbeen obtained when the total cell ganglioside content was measured (27, 35). The documentation of immunoregulatory activityand of the shedding of substantial quantities of gangliosides byYAC-1 cells suggests an alternate explanation for how ganglio
sides might relate to metastasis, i.e., it may be that quantitativeor qualitative characteristics of ganglioside shedding, rather thancell or membrane ganglioside content alone, are important factors modulating, possibly by an immunological mechanism, themetastatic potential of tumor cells.
ACKNOWLEDGMENTS
We are very grateful to Dr. Mary Ann K. Markwell for the quantitation of proteinin lipid samples and Dr. Christine Berte for the HEL lymphoproliferation studies.We also thank Anna-Marie Delay, Eileen Schwartz, and Robert Weinberger forassistance in these studies, and Dorothy Ross for preparation of the manuscript.
REFERENCES
1. Agarwal, M. K., and Neter, E. Effect of selected lipids and surfactants onimmunogenicity of several bacterial antigens. J. Immunol., 707: 1448-14561971.
2. Ando, S.. Chang. N. C., and Yu. R. K. High performance thin-layer chromatog-raphy and densitometric determination of brain ganglioside composition ofseveral species. Anal. Biochem., 89: 437-450, 1978.
3. Dittmer, J. C., and Lester, R. L. A simple, specific spray for the detection ofphospholipids on thin-layer chromatograms. J. Lipid Res., 5. 126-127. 1964.
4. Esselman, M. J., and Miller, H. C. Modulation of B cell responses by glycolipidreleased from antigen-stimulated T cells. J. Immunol., 119: 1994-2000, 1977.
5. Hakamori. S. Structures and organization of cell surface glycolipids: dependency on cell growth and malignant transformation Biochim. Biophys. Acta,477. 55-89, 1975.
6. Irwin, C. C., and Irwin, L. N. A simple rapid method for ganglioside isolationfröresmall amounts of tissue. Anal. Biochem.. 94: 335-339, 1979.
7. Ker^j, S. F., and Stoolmiller, A. C. Biosynthesis of glycosphingolipids incultured mouse neuroblastoma cells. J. Biol. Chem., 257: 7626-7631, 1976.
Shedding and Immunoregulation by Tumor Cell Gangliosides
mammary carcinomas. Proc. Nati. Acad. Sei. U. S. A., 74: 3011-3013, 1977.
9. Koprowski, H.. Herìyn,M., Steplewski, Z., and Sears, H. F. Specific antigen inserum of patients with colon carcinoma. Science (Wash. D. C.), 272: 53-55,
1981.10. Koski, I., Poplack, D. G., and Blaese, R. M. A non-specific esterase stain for
the identification of monocytes and macrophages. In: B. R. Bloom and J. R.David (eds.), In Vitro Methods in Cell-mediated and Tumor Immunity, pp. 359-362. New York: Academic Press, Inc., 1976.
11. Krishnaraj, R., and Kemp, R. G. Gangliosides do not raise cyclic AMP levelsduring inhibition of lymphocyte mitogenesis. Biochem. Biophys. Res. Commun., 104: 1453-1460,1982.
12. Krishnaraj, R., and Saat, Y. A. Murine thymic leukemogenesis: endogenousganglioside immunosuppression. IRCS Med. Sci. Cancer, 9: 314, 1981.
13. Ladisch, S., andGillard, B. A simple solvent partition method for the purificationof nanomolar quantities of gangliosides. In: Seventh International Symposiumon Glycoconjugates, Lund, Sweden. July, 1983 (Abstract), in press, 1983.
14. Ladisch, S., Wong, C., Dish, L., and Delay, A-M. Suppression of antigen-specific lymphocyte blastogenic responses by tumor cell-derived gangliosides.Clin. Res., 29: 528A, 1981.
15. Ledeen, R. W., and Yu, R. K. Gangliosides: structures, isolation and analysis.Methods Enzymol., 83: 139-191, 1982.
16. Lengle, E. E. Increased levels of lipid-bound sialic acid in thymic lymphocytesand plasma from leukemic AKR/J mice. J. Nati. Cancer Inst.. 62: 1565-1567.
1979.17. Lengle, E. E., Krishnaraj, R., and Kemp, R. G. Inhibition of the lectin-induced
mitogenic response of thymocytes by glycolipids. Cancer Res., 39: 817-822,
1979.18. Magnani, J. L, Brockhaus, M., Smith. D. F., Ginsburg. V., Blaszczyk, M.,
Mitchell, K. F., Steplewski, Z., and Koprowski, H. A monosialoganglioside is amonoclonal antibody-defined antigen of colon carcinoma. Science (Wash.D. C.). 272: 55-56, 1981.
19. Maizels, R. M., Clarke, J. A., Harvey, M. A., Miller. A . and Sercarz, E. E.Epitope specificity of the T cell proliferative response to lysozyme: proliferativeT cells react predominantly to different determinants from those recognizedby B cells. Eur. J. Immunol.. 70: 509-515, 1980.
20. Markwell, M. A. K. A new solid-state reagent to iodinate proteins. I. Conditionsfor the efficient labelling on antiserum. Anal. Biochem.. 725: 427-432, 1982.
21. Markwell, M. A. K., Svennerholm, L., and Paulson. J. C. Specific gangliosidesfunction as host cell receptors for Sendai virus. Proc. Nati. Acad. Sei. U. S. A.,78:5406-5410, 1981.
22. Miller, H. C., and Esselman, W. J. Modulation of the immune response byantigen reactive lymphocytes after cultivation with gangliosides. J. Immunol.,775:839-843. 1975.
23. Portoukalian, J., Zwingelstein, G , Abdul-Malak, N., and Dore, J. F. Alteration
of gangliosides in plasma and red cells of human bearing melanoma tumors.Biochem. Biophys. Res. Commun., 85: 916-920,1978.
24. Raz, A., Goldman, R., Yuli, I., and Inbar, M. Isolation of plasma membranefragments and vesicles from ascites fluid of lymphoma-bearing mice and theirpossible role in the escape mechanism of tumors from host immune rejection.Cancer Immunol. Immunother., 4: 53-59, 1978.
25. Rosenfelder, G., Ziegler, A., Wernet, P., and Braum, D. G. Ganglioside patterns:new biochemical markers for human hematopoietic cell lines. J. Nati. CancerInst., 68: 203-209, 1982.
26. Saito, M., and Rosenberg, A. Glycolipids and their developmental patterns inchick thigh and leg muscles. J. Lipid Res., 23: 3-8, 1982.
27. Skipski, V. P., Gitterman, C. O., Prendergast, J. S., Bitet-Yen, K., Lee, G.,Luell, G., and Stock, C. C. Possible relationship between glycosphingolipidsand the formation of metastasis in certain human experimental tumors. J. Nati.Cancer Inst., 65: 249-256, 1980.
28. Skipski, V. P., Katopodis, N., Prendergast, J. S., and Stock. C. C. Gangliosidesin blood serum of normal rats and Morris hepatoma 5123tc-bearing rats.Biochem. Biophys. Res. Commun., 67: 1122-1127, 1975.
29. Svennerholm, L. The quantitative estimation of cerebrosides in nervous tissue.7: 42-53. 1956.
30. Svennerholm, L. Quantitative estimation of sialic acids. II. A colorimetrieresorcinol-hydrochloric acid method. Biochim. Biophys. Acta, 24: 604-611,1957.
31. Svennerholm, L. Chromatographie separation of human brain gangliosides. J.Neurochem., 70: 613-623, 1963.
32. Ueno, K., Ando, S., and Yu, R. K. Gangliosides of human, cat, and rabbit spinalcords and cord myelin. J. Lipid Res., 79: 863-871,1978.
33. Whisler, R. L., and Yates, A. J. Regulation of lymphocyte responses by humangangliosides. J. Immunol.. 725: 2106-2111. 1980.
34. Wong, C. G., and Ladisch, S. Retention of gangliosides in serum delipidatedby diisopropyl ether/1-butanol extraction. J. Lipid Res., 24: 666-669, 1983.
35. Yogeeswaran, G. Incorporation of asialo GKKand gangliosides in cell surfaceof cultured metastatic and nonmetastatic Balb/3T3 cell lines: altered adhesionto substrate in vitro and subcutaneous tumor cell take. J. Nati. Cancer Inst.,66:303-310, 1981.
36. Yogeeswaran, G., Gronberg, A., Hasson, M.. Dalianis, T., Kiessling, R., andWelsh, R. M. Correlation of glycosphingolipids and sialic acid in YAC-1 lym-phoma variants with their sensitivity to natural killer cell mediated lysis. Int. J.Cancer, 28:517-526, 1981.
37. Yogeeswaran, G.. and Salk, P. L. Metastatic potential is positively correlatedwith cell surface sialylation of cultured murine tumor cell lines. Science (Wash.D. C.), 272: 1515-1516, 1981.
