J. Cell Sci. 21, 579-594 ('976)
579
Printed in Great Britain
EFFECT OF GLUTARALDEHYDE FIXATION ON
LECTIN-MEDIATED AGGLUTINATION OF
MOUSE LEUKAEMIA CELLS
W. J. VAN BLITTERS WIJK,# E. F. WALBORG J R . , |
CONSTANCE A. FELTKAMP, H. A. M. HILKMANN AND
P. EMMELOT
Division of Cell Biology, Antoni van Leeuwenhoek-Huis,
The Netherlands Cancer Institute, Amsterdam,
The Netherlands
SUMMARY
The effect of glutaraldehyde fixation on lectin-mediated agglutination of murine leukaemia
(GRSL) cells was investigated using 2 assay methods which differed in the shear forces to
which the agglutinated cells were subjected. First, lectin and cells were allowed to interact
under conditions in which shear forces were minimized and the degree of agglutination was
evaluated microscopically by the appearance and size of the cell aggregates. This assay
demonstrated that concanavalin A (con A)-, wheat germ agglutinin (WGA)- or Ricinus
communis agglutinin I (RCA^-mediated cytoagglutination was unaffected (WGA and RCAj)
or somewhat enhanced (con A) by prior fixation of the cells with glutaraldehyde. Secondly,
an electronic particle counter was used to measure the disappearance of single cells and
concomitant appearance of cell aggregates as a function of the lectin concentration. In this
assay, in which the aggregated cells are subjected to significant shear forces during dilution
and cell counting, agglutination of GRSL cells by each of the 3 lectins was drastically
inhibited by prior fixation of the cells with glutaraldehyde. This assay also demonstrated
enhanced nonlectin-induced cell aggregation after fixation. In both cytoagglutination assays
about the same lectin concentration was required for threshold agglutination of unfixed cells.
Comparatively, the results of the 2 cytoagglutination assays indicate that a fraction of the
lectin-mediated bonds between unfixed cells is shear resistant and that fixation of the cells
either weakens these bonds or inhibits their formation.
Morphologically, cells prefixed with glutaraldehyde were spherical at all lectin concentrations,
with a continuous dense distribution of cell surface-bound con A, labelled directly with
haemocyanin or indirectly using the peroxidase-diaminobenzidine reaction. Unfixed cells
showed angular and toadstool-shaped deformations, especially at the highest lectin concentrations, the agglutinating surfaces being flattened against each other over extended areas.
The distribution of con A label was continuous and dense between the apposed surfaces and
discontinuous on free surfaces. In the presence of con A the free surfaces of prefixed cells
exhibited more microvilli than the surfaces of non-prefixed cells. These results favour the
view that fixation prevents the formation of shear-resistant, lectin-mediated bonds between
cells, not by restricting the lateral mobility of lectin receptors, but by impairing the apposition
of rigid cell surfaces.
• Author to whom off-print requests should be sent.
f Recipient of an Eleanor Roosevelt International Cancer Fellowship. Present address:
Biochemistry Department, The University of Texas System Cancer Center, M.D. Anderson
Hospital and Tumor Institute, Houston, Texas 77025, U.S.A.
580
W. J. Van Blitterswijk and others
INTRODUCTION
Neoplastically transformed and trypsinized normal cells generally exhibit a higher
lectin-mediated agglutinability than do normal cells (Burger, 1973; Lis & Sharon,
1973; Nicolson, 1971; Sachs, 1974). This has been ascribed to increased lectininduced clustering of lectin receptor sites (Nicolson, 1971; Rosenblith et al. 1973;
Sachs, 1974) in the former cells, may be facilitated by an increased density of the
sites (Collard & Temmink, 1975). Recently the generality of increased clustering
in transformed cells has been opposed and the essential role of topological redistribution of lectin receptors in lectin-induced cytoagglutination has thereby been
questioned (De Petris, Raff & Malucci, 1973; Roos & Temmink, 1975; Temmink,
Collard, Spits & Roos, 1975). In addition Inbar, Shinitzky & Sachs (1973 b) found
that malignant transformation is not always accompanied by an increased rotational
diffusion of the lectin receptors.
It has been proposed that topological redistribution (clustering) of lectin receptors
is a prerequisite for lectin-induced cytoagglutination (Inbar et al. 1973 a; Nicholson,
1972). To test this hypothesis, cells can be fixed with glutaraldehyde, a treatment
which cross-links the membrane (glyco)proteins and thereby restricts their lateral
mobility. Several studies of this type have been performed, but not all results are
consistent with this hypothesis. In some cases inhibition of lectin-induced cytoagglutination was observed by fixation (Inbar et al. 1973 a; Rutishauser & Sachs,
1975); however, concanavalin A (con A)-induced agglutination of Ehrlich ascites
carcinoma cells was unaffected (Inoue, 1974) and that of NovikofF ascites hepatoma
cells was enhanced (Davis & Walborg, 1975).
Among the many factors (Gordon & Marquardt, 1974; Marquardt & Gordon,
1975; Schnebli & Bachi, 1975) that govern the agglutination reaction, shear forces
have recently been found to be very important (Marquardt & Gordon, 1975; Schnebli
& Bachi, 1975). Marquardt & Gordon (1975) showed that glutaraldehydefixationof
erythrocytes slightly increased agglutinability by Con A and soybean agglutinin
when tested on unperturbed Microtiter plates, but that disruption of agglutination
was obtained by shear forces using an electronic particle counter.
The present study extends these observations to an ascites leukaemia (GRSL)
originating and maintained in GR/A mice. These cells were especially interesting
in this respect since we found highly dynamic properties (shedding, antigenic
modulation and capping) of a mammary tumour virus-induced cell surface antigen
(Calafat et al. 1976; Hilgers et al. 1975; Van Blitterswijk et al. 1975). This antigen
exhibits con A-receptor activity (Van Blitterswijk et al. unpublished). We tested the
effect of fixation on the cytoagglutination induced by con A, wheat germ agglutinin
(WGA) and Ricinus communis agglutinin I (RCAZ). Two assay methods, differing in
the shear forces to which the agglutinated cells were subjected, were utilized. The
marked differences in the results obtained using these 2 assays are reported and
discussed in relation to electron-microscopic investigations on the effect of fixation
on cell-surface morphology and the topology of surface-bound con A.
Effect of fixation on cytoagglutination
581
MATERIALS AND METHODS
Cells
GRSL2 cells from a spontaneous lymphoid leukaemia in the GR/A mouse strain were
maintained (passage 94-106) by weekly intraperitoneal transplantation in 2- to 3-month-old
GR/A mice. Cells were harvested 7 or 8 days after transplantation and washed (250 g, 10 min)
3 times with at least 25 vol. of Ca1+- and Mg^-free phosphate-buffered saline pH 7-5
(CMF-PBS) (Cronin, Biddle & Sanders, 1970).
Lectins
Concanavalin A (con A), a twice-crystallized product dissolved in saturated NaCl (Lot
no. 119), was obtained from Miles Yeda Ltd, Rehovot, Israel. The preparation of wheat
germ agglutinin (WGA) and Ricinus communis agglutinin I (RCA,) has been described
previously (Neri et al. 1976). RCAt was dissolved in CMF-PBS and insoluble material
removed by centrifugation. The concentrations of the con A and RCAX were determined
using extinction coefficients ( I % / I cm at 280 nm) of 14 (Podder, Surolia & Backhawat, 1974)
and 13, respectively. The latter extinction coefficient was determined by the supplier. The
concentration of WGA was determined gravimetrically. The haemagglutination activities of
the lectins, determined with guinea-pig erythrocytes by the method of Smith, Neri & Walborg
(i973)i were 60000 haemagglutination units (HAU)/mg of con A, 26000 HAU/mg of WGA
and 68000 HAU/mg of RCA^
Glutaraldehyde fixation of cells
Purity of the glutaraldehyde solution (Fluka A.G., Buchs, Switzerland) (Fahimi &
Drochmans, 1965) was ascertained by measuring the absorbance at 230 and 280 nm, and
only solutions exhibiting an absorbance ratio (^230 nm/^280 nm) of < 1-3 were utilized. The
cells were fixed at 4 CC with glutaraldehyde by adding to a suspension of io7 cells/ml an
equal volume of 0-5% (w/v) glutaraldehyde (23% solution in H,O diluted with CMF-PBS).
Following fixation for 15 min, the cells were centrifuged and suspended in 02 M glycine in
CMF-PBS at a concentration of io7 cells/ml and allowed to stand for 10 min at 4 °C. Finally
the cells were washed 3 times with CMF-PBS. Control cells passed through all steps, except
for the glutaraldehyde fixation. No differences in results were obtained between these control
cells and untreated control cells (kept on ice).
Cytoagglutination
Particle counter assay. At the bottom of 12-ml plastic tubes z$-fi\ aliquots of a cell suspension
(2 x io7 cells/ml) in CMF-PBS were mixed with 25-/4I aJiquots of the lectin serial dilutions
in the same buffer. The lectin concentrations indicated in Figs. 1-3 (pp. 583-585) are the
end concentrations in this 50-/4I reaction mixture. After incubation at 22 CC with gentle
shaking for 30 min the tubes were put on ice and the cytoagglutination was measured
immediately by means of an electronic particle counter (Coulter Counter model ZF, Coulter
Electronics Ltd, Harpenden, England). To this end the contents of the tubes were diluted
400-fold with Isoton (Coulter Electronics Ltd) which was pipetted down the side of the tube
in two 10-ml portions. After adding the first portion of Isoton, the tube was emptied into
a 30-ml Accuvette plastic Coulter Counter vial, rinsed with the second portion and emptied
again into this vial. Then the vial was inverted twice and ioo-/il aliquots were counted using
a ioo-/im orifice tube at 3 threshold (T) settings: 12-5 (7\), 50 (Tj), and 100 (Ta), corresponding to particles greater than 240 fim3, > 960 fim* and > 1920/tm3, respectively.
Aperture current (i) and attenuation switch (A) settings were 32 and i-o respectively. Calibration was done by means of 1245-^m polystyrene divinyl benzene latex beads (Coulter
Electronics Ltd). Threshold T1 was chosen to exclude the few erythrocytes present in the
cell suspension (< 5%), but to include all GRSL cells (mean volume of single cells 436 /tm3)
Tt and Tz were chosen arbitrarily to count clumps of > 2 and > 4 cells, respectively.
Microtest II plate assay. Lectin-induced cytoagglutination was determined at 22 °C using
38
CEL
21
582
W. J. Van Blittersvnjk and others
a modified version of the method of Wray & Walborg (1971). The degree of agglutination
was determined at a fixed time as a function of the lectin concentration. This assay minimizes
the shear forces to which the agglutinated cells are subjected. The cytoagglutination reaction
was performed in wells of No. 3040 Microtest II Tissue Culture Plates (Falcon Plastics,
Oxnard, California). These plates contain 96 flat-bottomed wells with a fluid capacity of
0-4 ml per well. To 25 fil of CMF-PBS or a saccharide solution in CMF-PBS (in the case
of cytoagglutination inhibition assays) were added 25 ftl of a serial dilution of lectin in
CMF-PBS. An aliquot (25 /tl) of a suspension of GRSL cells containing 8 x io8 cells/ml was
then added and the plate agitated by gently tapping against the palm of the hand. The
reaction was allowed to proceed undisturbed for 15 min at which time the plates were again
agitated in the same manner. The plates were allowed to stand an additional 15 min after
which the degree of agglutination was evaluated by microscopic examination using a substage microscope (10 x objective, 8 x oculars). Wells in which less than 25% of the cells
were present in aggregates of 3-10 cells were scored as o. Wells in which at least 25% of the
cells were in aggregates of 3-10, 10-25, o r 25~5° cells were scored as + 1 , +2 or +3, respectively. Maximal agglutination ( + 4) indicated that at least 25% of the cells were in aggregates
of > 50 cells with few single cells or small (3-10 cells) aggregates visible. All assays were
performed in triplicate and scored by the same observer. The lectin concentrations indicated
in Fig. 3 represent the final concentrations in the 75-//.I reaction mixture.
Electron microscopy
Prefixation of cells was performed in 0-25% and in 2 5 % glutaraldehyde (Ladd Res. Ind.
Inc., Burlington, Vermont, U.S.A.; diluted from 70%) in PBS (Dulbecco; Oxoid, London)
for 15 min at 4 °C. The cells were washed 3 times with large volumes of PBS.
Prefixed and non-prefixed cells (8 x 10* per tube) were incubated for 15 min at room temperature with 100 fig con A, final volume 1 ml. After 3 washes with PBS, bound con A was
labelled with peroxidase (PO; from horse-radish) according to Bernhard & Avrameas (1971)
or with haemocyanin (HC; from Helixpomatia) according to Smith & Revel (1972). Incubations
with HC (500/ig/oi ml PBS) were performed for 15 min at room temperature. The cells
were washed 3 times with PBS, and fixed in a mixture of glutaraldehyde and OsO4 (Hirsch &
Fedorko, 1968). Incubation with PO (50 fig/i-o ml PBS, for 15 min at room temperature)
was followed by 3 washes with PBS, fixation in 2-5% glutaraldehyde, and final incubation
with 3,3'-diaminobenzidine tetrahydrochloride (DAB; 500 /ig/10 ml PBS) for 8 min at room
temperature. After washing, the contrast of the reaction product was increased by fixation
of the cells in 1 % OsO< for 30 min at 4 °C.
After each incubation or washing the cells were centrifuged at 250 g for 1-2 min at 4 °C.
The final pellets were dehydrated and embedded in a mixture of Epon 812 and Araldite. Thin
sections of HC-incubated cells were doubly stained with uranyl acetate and lead hydroxide.
Sections of PO-reacted cells were observed without further staining. Observations were
performed with a Philips EM 300 at 60 kV using a $o-/im objective aperture.
RESULTS
Cytoagglutination assays
The effect of glutaraldehyde fixation on lectin-mediated agglutination of GRSL cells
was investigated by 2 assay methods, i.e. the plate assay and the particle counter assay.
Fixation with 0-25% glutaraldehyde for 15 min proved to be suitable. Increasing
the fixation time to 60 min did not significantly change the results obtained by both
assays, whereas 2- 5 % glutaraldehyde resulted in higher nonlectin-induced aggregation
(see below).
In both assays the lectin-induced cytoagglutination reactions were inhibited by
the various lectin-specific saccharides: methyl-a-D-glucopyranoside for con A, Nacetyl-D-glucosamine for WGA and D-galactose for RCAj (not illustrated).
Effect of fixation on cytoagglutination
583
Particle counter assay
Cytoagglutination was measured in the electronic particle counter above 3 thresholds, 7\, T2 and T3 which scored aggregates of > 1, > 2 and > 4 cells, respectively
(see Materials and methods). Increase of agglutination of unfixed cells with increasing
lectin concentrations was monitored by the decrease of single cells (number of
1300
r
20 0-5
1
5
10
RCA | conc,//g/ml
50
100
500
Fig. 1. Agglutination of GRSL cells as a function of RCA! concentration. [Data
were from experiment A (Table 1). D—D (2VT,), single cells; • — • (T^-T,),
clumps of 2 < cells ^ 4; O—O T3, clumps of cells > 4. The amounts counted
in the absence of lectin were taken as 100. Distribution between Tu Tt and T3 can
be deduced from Table 1.
aggregates measured at Tx minus that measured at T2, abbreviated as 7\-T 2 for
convenience) and the increase of T3 agglutination. Typical data for RCAX are
presented in Fig. 1. The amounts of Tj-T2, T2-T3, and T3 counted in the absence
of lectin were taken as 100. In contrast to the linear response of T3 agglutination
induced by RCA^ the corresponding response for WGA and con A consistently
showed non-linearity (Fig. 3, top).
In the case of RCAj, the number of aggregates of 2-4 cells (T2-T3) remained
about constant at the various lectin concentrations (Fig. 1). This intermediary state
of agglutination was not significantly dependent on the type of lectin, but was primarily
determined by the level of spontaneous cell aggregation present in the absence of
lectin, which could vary from one experiment to another (as expressed by the
percentage ratios T2/T1 and T3jTly Table 1). A relatively high spontaneous (nonlectin38-2
W. J. Van Blitterszoijk and others
Table i. Enhanced nonlectin-induced aggregation of
GRSL cells after glutaialdehyde fixation
% TJT,
/o
TJT,
A
^
Experiment
Fixed
A
B
C
D
E
28
Control
77
Fixed
^
Control
3'4
o-2
2-4
55
26
23
32
18
12
38
47
o-6
o-6
14
i-5
'•5
7'5
11
In all experiments aliquots of 2-3 x 10' cells were counted in quadruplicate using the
Coulter Counter at thresholds Tx (Ss 1 cell), T, (> 2 cells) and T 3 ( > 4 cells). Standard
deviation was 5 % which includes errors and variations introduced by attachment of cells to
plastic, heterogeneity of cell suspension, pipetting and in the Coulter Counter.
10
50 100
Con A conc.,/(g/ml
500 1000
Fig. 2. Agglutination of fixed (—) and unfixed (
) GRSL cells as a function of the
concentration of con A. O, % TJTlt ratio of particles of > 2 cells to particles > 1 cell.
• » % T,ITlt ratio of particles greater than 4 cells to particles > 1 cell. Fixation
of cells resulted in enhanced nonlectin-induced aggregation without further lectininduced agglutination. Data were from experiment B (Table 1).
induced) cell aggregation resulted in a significant decrease of T2-Ts for all 3 lectins
as a function of their concentration, in contrast to Fig. 1 which depicts an experiment
with cells showing low spontaneous aggregation (experiment A).
Table 1 and Fig. 2 (ordinate values) demonstrate that glutaraldehyde fixation
enhanced nonlectin-induced aggregation of cells, except for experiment E. Furthermore, fixation resulted in complete inhibition of lectin-induced cytoagglutination
as measured at the 3 thresholds in the particle counter for lectin concentrations up
to 2 mg/ml. This is demonstrated in Fig. 2 for con A by the percentage ratios
and Tg/Tj, similar curves being obtained for WGA and RCAj (not illustrated).
Effect of fixation on cytoagglutination
585
Of the various curves (Fig. 1), that depicting Tz agglutination was the steepest
for all 3 lectins. This, together with the observed increase of nonlectin-induced cell
aggregation by fixation (Table 1) led us to use the T3 agglutination as the most
sensitive indicator for comparing the agglutinability of fixed and unfixed cells, as
illustrated in Fig. 3 (top).
i300 r
i300 r
0-5 1
50 100
500
Fig. 3. Agglutination of fixed (A—A) and unfixed (O—O) GRSL cells as a
function of lectin concentration (/ig/ml) as assayed by the particle counter (A, C, E)
and on Microtest II Tissue Culture Plates (B, D, F). A, B, con A; c, D, WGA; E, F,
RCAT. Data were from experiments B (A, C) and C (E) (Table 1). Amounts counted
in the particle counter in the absence of lectin were taken as 100.
Plate assay
Parallel with the measurements in the electronic particle counter and using cells
from the same batches, cytoagglutination was also measured on Microtest II plates
using the criteria outlined in Materials and methods.
586
W. J. Van Blittersmjk and others
The results, illustrated in Fig. 3 (bottom), were strikingly different from those
obtained with the particle counter with respect to the effect of fixation. As measured
in the plate assay, fixation had no effect on cytoagglutination by RCAX or WGA,
whereas con A-induced cytoagglutination was somewhat enhanced. At this point
28r
80
16
60
n
40
& 12
20i
1
-i
1
0
:
i
' • ' • - - ' •
3
4
5
'
6
7
-
i
•
'
•
8
i
i
9
:
1
.
_
.
•••
10 11 12 13
103. Ai.T.
4-
5
6
7
8
9 10 11 12 13
103. A.I.T.
Fig. 4. Distribution of aggregate sizes in GRSL cytoagglutination as measured by
the Coulter Counter. D , control without lectin; El, with 250 fig/ml WGA. / \ J I ,
no. of particles present at the various windows formed by combinations of A = 1,
i — 32, T = 125, 50, 100 and A = 2, i = 128, T = 10, 12, 14, 16 etc. A.i.T. and
threshold particle volume are linearly proportional. A.i.T. = 1000 corresponds to
a threshold particle volume of 600 /tm8. The mean volume of a single GRSL cell
is 436 /tm3. The inset graph with larger ordinate scale represents and illustrates the
increase in T, (compare Figs. 1 and 3, top) caused by the lectin.
it should be noted that since the experimental variation of this assay is approximately
33% (Wray & Walborg, 1971) only conditions which resulted in at least 2-fold
alterations in the concentrations of lectins required for threshold or half-maximal
agglutination were considered significant. The maximum agglutination of cells by
con A was always less than that caused by the 2 other lectins.
Effect of fixation on cytoagglutination
587
It is noteworthy that about the same lectin concentration yielded threshold
agglutination in both cytoagglutination assays using unfixed cells (Fig. 3, top and
bottom).
A further difference was found in the mean size of the aggregates present under
the conditions of the two assay methods. Fig. 4 shows the distribution patterns of
particle sizes measured in the particle counter after cytoagglutination mediated by
a high concentration (250 /ig/ml) of WGA as compared with control cells. Here the
mean size (3-5 cells/cluster) was much smaller than that of the lectin-mediated
aggregates seen in the plate assay (estimated as ^ 40 cells/cluster at +4 degree of
agglutination).
Light microscopy revealed that a fraction of the unfixed cells exhibited deformations
in shape when incubated with lectin. These alterations resembled that described
by Loor (1974), i.e. the cells became angular and toadstool-shaped. These alterations,
which probably reflect capping (see also electron-microscope results), were most
evident at high concentrations of con A and RCAj and were confined primarily to
non-agglutinated cells. Fixed cells remained spherical at all concentrations of each
lectin.
Electron microscopy
On the surface of cells which were prefixed with glutaraldehyde the distribution
of con A, as visualized by haemocyanin (HC) or by the peroxidase-diaminobenzidine
(PO-DAB) reaction product, was very regular and dense (Figs. 5-7). HC molecules
were attached with very small interspace to microvilli and to smooth parts of the
cell surface, while the PO-DAB reaction product formed a continuous dense layer
over the cell surface. Tangential sections of this layer showed that it was composed
of small electron-dense spheres (Fig. 7).
No differences were observed between the reactions of cells prefixed with 2-5 or
with 0-25% glutaraldehyde. The prefixed cells were spherical with many rather
short microvilli, generally radiating perpendicularly from the cell surface. No signs
of agglutination were observed. All cells appeared completely free from each other.
Cells which were not prefixed showed a strong con A-mediated agglutination
{Fig. 8). The agglutinated cells had changed in shape; the agglutinating surfaces
were flattened against each other over extended areas, giving the cells a more angular
outline. Only the free surfaces remained spherical. The number of microvilli was
much lower than on prefixed cells. Between agglutinating surfaces microvilli were
often absent, while at the free surfaces they were sometimes pressed against the cell
body.
The distribution of con A label on non-prefixed cells was variable (Figs. 8-10).
Between agglutinated cells a continuous distribution of PO-DAB reaction product
was found, while HC molecules could be absent at these sites. Probably, the size of
the HC molecules inhibited their penetration between agglutinated cells. On free
surfaces the distribution of both HC and the PO-DAB reaction product was irregular; some parts of the cell surface showed a rather high concentration of label,
while others contained much less or were completely free of label.
588
W. J. Van Blittersvrijk and others
Fig. 5. Fixed GRSL cells. The reaction product of PO-DAB which visualizes
con A forms a continuous layer over the surface of the spherical free cells, x 8250.
Fig. 6. Fixed GRSL cells, con A labelled by HC. The HC molecules are rather
regularly, with small interspaces, distributed over microvilli and smooth parts of
the cell surface, x 30000.
Fig. 7. The reaction product of PO-DAB on fixed cells is composed of small
electron-dense spheres, x 30000.
Ejfect of fixation on cytoagglutination
Fig. 8. Unfixed GRSL cells, con A visualized by PO-DAB. The cells are agglutinated
and deformed, the agglutinating surfaces are flattened against each other over
extended areas. The PO-DAB reaction product is regular and dense between
apposed surfaces of agglutinated cells, and irregular on free surfaces, x 11250.
Fig. 9. Unfixed GRSL cells, con A labelled by HC. The HC molecules are localized
between apposed surfaces of agglutinated cells and, in irregular patches, on the
free surface of the cells, x 30coo.
Fig. 10. Unfixed GRSL cell, con A visualized by PO-DAB. The reaction product
on this free cell is concentrated in a cap over one pole of the cell, which shows a
villous membrane. The opposite pole has a smooth, spherical membrane and is
virtually free from label, x 10200.
59°
W. J. Van Blitterswijk and others
A small number of con A-treated non-prefixed cells, especially the non-agglutinated
ones, had an appearance (Fig. 10) which was suggestive of capping. These cells
exhibited an irregular, villous plasma membrane at one pole, and were smooth
and spherical at the opposite pole. The smooth pole was virtually free from con A
label, while the label was more dense at the villous pole.
DISCUSSION
Lectin-mediated agglutination of GRSL cells was measured by 2 methods;
electronically in a particle counter and microscopically in the wells of tissue culture
plates. In contrast to the former assay method, the plate assay was performed under
conditions in which external shear forces were minimized. The mean size of the
lectin-induced aggregates of unfixed cells, detected by the particle counter was
3-5 cells at the highest lectin concentrations (Fig. 4). These aggregates were much
smaller (approximately a factor of 10) than those seen at high lectin concentration in
the plate assay. Apparently large aggregates are disrupted into much smaller ones
by the shear forces applied during dilution and counting, suggesting that apart from
shear-resistant bonds, relatively weaker shear-sensitive bonds may operate between
cells in lectin-mediated agglutination. It should be noted that this distinction of
bonds does not necessarily infer the involvement of distinct cell surface lectin
receptor molecules but may reflect distinct binding capacities between the apposed
surfaces of agglutinated cells (see below).
After glutaraldehyde fixation of cells a complete inhibition of lectin-mediated
cytoagglutination was found using the particle counter assay, but according to the
plate assay cytoagglutination was unaffected (WGA, RCAj) or even enhanced
(con A). These seemingly contradictory results can be explained by assuming that
especially the lectin-mediated bonds between unfixed cells which are shear-resistant
are weakened or their formation is inhibited by prior fixation.
Fixation by glutaraldehyde is expected to cross-link (glyco)proteins on the cell
surface, thus restricting (1) the motility (plasticity) of the cell surface, and (2) lateral
mobility of lectin receptors (Inbar et al. 1973b; Sachs, 1974). Inbar et al. (1973a)
suggest that this second factor, especially the ability of lectin sites to form clusters,
is required for cell agglutination. In our opinion it is very difficult to determine
which of the above two factors is the most important in this respect. It could depend
on the type of cell, the specific lectin and the density of the lectin receptors on the
cell surface.
Electron microscopically we found a dense continuous distribution of con A
label on the surface of fixed non-agglutinated cells. A similar dense distribution was
observed at the contacts between non-fixed agglutinated cells. A conspicuous trait
in the latter, absent from fixed cells, was their mutual deformation, the agglutinating
surfaces being flattened against each other over extended areas and often completely
devoid of microvilli. Based on these observations we favour the view that the plasticity
required for mutual apposition of cell surfaces is the main factor that allows the
formation of lectin-mediated shear-resistant bonds, and this is consequently prevented
Effect of fixation on cytoagglutination
591
by a 'frozen' cell surface as obtained by glutaraldehyde fixation. However, if lateral
mobility of lectin receptors is a relevant factor in the agglutination of GRSL cells which cannot be excluded - we would prefer the notion of a short-range redistribution
or alignment of complementary receptors between apposed faces of agglutinating
cells rather than that of a longer-range clustering of the receptors. Recently Marquardt
& Gordon (1975) reached similar conclusions for lectin-induced agglutination of
erythrocytes, namely that after fixation stable aggregate formation is prevented by
the rigidity of the cell surface, and that agglutination is not dependent on clustering
of lectin receptors. Our electron-microscopical results as well as the conclusions
drawn therefrom are also in agreement with those of De Petris et al. (1973) on
fibroblasts.
Using unfixed cells and a given lectin the particle counter assay and the plate
assay indicated that about the same lectin concentration was required for threshold
cytoagglutination. Schnebli & Bachi (1975) reported that mechanical shear markedly
increased the concentration of con A necessary to achieve threshold agglutination
of erythrocytes. This is conceivable because in contrast to our assay methods they
used only one (macroscopic) criterion to score cytoagglutination. Their results can
thus be explained by assuming that the majority of the con A-mediated bonds are
shear-sensitive, probably in part due to the already relatively rigid membrane of
the native erythrocyte as compared with the lymphoid GRSL cell.
In contrast to the linear agglutination response curves induced by increasing
concentrations of RCA1( those induced by con A and WGA consistently showed
non-linearities as measured in the particle counter. A maximum in the curve and
decreasing agglutinability at highest concentrations of these 2 lectins was often
observed. The latter phenomenon has also been found by Kaneko, Hayatsu &
Ukita (1975) and is comparable to the 'prozone effect' generally encountered in
immunological reactions (immunoprecipitation, antibody-induced cytotoxicity and
cap formation (Hilgers et al. 1975; Taylor, Duffus, Raff & De Petris, 1971) which
results from an excess of ligand over ligand receptor sites.
The biphasic shape of the agglutination curves at intermediate concentrations of
con A and WGA could result from the existence of 2 kinds of receptor sites with
different affinities for the lectins, both contributing to the shear-resistant bonds in
cytoagglutination. A similar biphasic curve has been found for con A-mediated
cytoagglutination of mouse sarcoma cells (Hwang, Murphree & Sartorelli, 1974).
A phenomenon distinct from lectin-induced cytoagglutination is the enhancement
of 'spontaneous', nonlectin-induced cell aggregation by prior glutaraldehyde fixation.
Nonlectin-induced aggregation was detected only by the particle counter assay,
indicating that this type of bond between the cells is shear-resistant. Several
mechanisms may be responsible for this enhanced aggregation, e.g. fixation may
alter the repulsive forces between cells by some unknown mechanism or, more
likely, aldehyde groups introduced on the cell surface may not be completely
inactivated by reaction with glycine, leading to reaction between cells.
Dr J. G. Collard in our Institute also observed enhanced nonlectin-mediated cell
aggregation after glutaraldehyde fixation using lymphoblast lines RAJI and EB3
592
W. J. Van Blitterstmjk and others
(unpublished). In this case the nonlectin-induced aggregates present after fixation
could be disrupted by shaking, in contrast to the con A-induced aggregates of
unfixed cells. By this procedure it was found that fixation inhibited con A-mediated
cytoagglutination. The nonlectin-induced aggregate bonds in the case of the GRSL
cells, however, are apparently much more stable than those of the aforementioned
cell lines, since they were not disrupted in the particle counter, and shaking of the
fixed GRSL cells on the plate led to massive clumping of the cells into one big
aggregate (not shown) in contrast to unfixed cells.
Thus the strength of the cell-cell bonds in nonlectin-induced cell aggregation,
which are shear-resistant in the case of GRSL cells, may vary according to the
type of cell. This may also be the case for the lectin-mediated bonds in cytoagglutination. In general, however, the distinction between shear-resistant and
shear-sensitive lectin-mediated as well as nonlectin-mediated bonds may well be
applicable to other cells and in fact provides an explanation for the conflicting results
obtained on the effect of fixation on lectin-mediated cytoagglutination (Inbar et al.
19736; Rutishauser & Sachs, 1975, versus Davis & Walborg, 1975; Inoue, 1974).
At least the present results emphasize that for the interpretation of cytoagglutination
data the type of assay method should be taken into account.
The haemocyanin used in this study was a generous gift from Dr E. F. J. van Bruggen,
University of Groningen, The Netherlands.
The technical help of Miss Cora Koning and Mr I. V. van de Pavert is gratefully acknowledged.
This collaborative research was aided by a travel grant to E. F. W. from the Paul and Mary
Haas Foundation.
REFERENCES
BERNHARD, W. & AVRAMEAS, S. (1971). Ultrastructural visualization of cellular carbohydrate
components by means of ConcanavaJin A. Expl Cell Res. 64, 232-236.
BURGER, M. M. (1973). Surface changes in transformed cells detected by lectins. Fedn Proc.
Fedn Am. Socs exp. Biol. 32, 91-101.
CALAFAT, J., HILGERS, J., VAN BLITTERSWIJK, W. J., VERBEET, M. & HACEMAN, P. C. (1976).
Antibody-induced modulation and shedding of mammary tumor virus antigens on the
surface of GR ascites leukaemia cells as compared with normal antigens. J. natn. Cancer
Inst. 56 (in Press).
COLLARD, J. G. & TEMMINK, J. H. M. (1975). Differences in density of Concanavalin A-binding
sites due to differences in surface morphology of suspended normal and transformed 3T3
fibroblasts. J. Cell Sci. 19, 21-32.
CRONIN, A. P., BIDDLE, F. & SANDERS, F. K. (1970). Wheat germ agglutinin conjugated with
ftuorescein isothiocyanate: staining of normal and transformed Chinese hamster cells.
Cytobios a, 225-231.
DAVIS, E. M. & WALBORG, E. F., JR. (1975). Lectin-induced agglutination of Novikoff
hepatoma cells: effect of glutaraldehyde fixation. Proc. Am. Ass. Cancer Res. 16, 47, Abstr.
188.
D E PETRIS, S., RAFF, M. C. & MALUCCI, L. (1973). Ligand-induced redistribution of con-
canavalin A-receptors on normal, trypsinized and transformed fibroblasts. Nature, Netu
Biol. 244, 275-280.
FAHIMI, H. D. & DROCHMANS, P. (1965). Essais de standardisation de la fixation au glutaraldehyde. I. Purification et determination de la concentration du glutaraldehyde. J. Microscopic
4, 725-736.
Effect of fixation on cytoagglutination
593
J. A. & MARQUARDT, M. D. (1974). Factors affecting hemagglutination by Concanavalin A and soybean agglutinin. Biochim. biophys. Acta 332, 136-144.
GORDON,
HILGERS, J., VAN BLITTERSWIJK, W. J., BONT, W. S., THEUNS, G. J., NUSSE, R., HAVERMAN, J.
& EMMELOT, P. (1975). Distribution and antibody-induced redistribution of a mammary
tumor virus-induced and a normal antigen on the surface of mouse leukemia cells. J. vatn.
Cancer Inst. 54, 1335-1341HIRSCH, J. G. & FEDORKO, M. E. (1968). Ultrastructure of human leukocytes after simultaneous fixation with glutaraldehyde and osmium tetroxide and ' postfixation' in uranyl
acetate. J. Cell Biol. 38, 615-627.
HWANG, K. M., MURPHREE, S. A. & SARTORELLI, A. C. (1974). A quantitative spectrophotometric method to measure plant lectin-induced cell agglutination. Cancer Res. 34, 3396—3402.
INBAR, M., HUET, C , OSEROFF, A. R., BEN-BASSAT, H. & SACHS, L. (1973a). Inhibition of
lectin agglutinability by fixation of the cell surface membrane. Biochim. biophys. Acta
3 " , 594-599INBAR, M., SHINITZKY, M. & SACHS, L. (19736). Rotational relaxation time of Concanavalin A
bound to the surface membrane of normal and malignant transformed cells. J. molec. Biol.
81, 245-253.
INOUE, M. (1974). Cell agglutination mediated by Concanavalin A and the dynamic state of
the cell surface. J. Cell Sci. 14, 197-202.
KANEKO, I., HAYATSU, H. & UKJTA, T. (1975). A quantitative assay for Concanavalin A- and
Ricinus communis agglutinin-mediated agglutinations of rat ascites hepatoma cells. Relationship between Concanavalin A binding and cell agglutination. Biochim. biophys. Acta 392,
131-140.
Lis, H. & SHARON, N. (1973). The biochemistry of plant lectins (phytohemagglutinins).
A. Rev. Biochem. 42, 541-574.
LOOR, F. (1974). Binding and redistribution of lectins on lymphocyte membrane. Eur. J.
Immun. 4, 210-220.
MARQUARDT, M. D. & GORDON, J. A. (1975). Glutaraldehyde fixation and the mechanism of
erythrocyte agglutination by Concanavalin A and soybean agglutinin. Expl Cell Res. 91,
310-316.
NERI, G., GUILIANO, M. D., CAPETILLO, S., GILLIAM, E. B., HIXSON, D. C. & WALBORG,
E. F., JR. (1976). Effect of neuraminidase and papain treatment on lectin-induced agglutination of Novikoff tumor cells and assay of lectin receptor activity of the glycopeptides
released from the cell surface by papain. Cancer Res. 36, 263-268.
NICOLSON, G. L. (1971). Difference in topology of normal and tumour cell membranes
shown by different surface distributions of ferritin-conjugated Concanavalin A. Nature,
Neiv Biol. 233, 244-246.
NICOLSON, G. L. (1972). Topography of membrane Concanavalin A sites modified by proteolysis. Nature, New Biol. 239, 193-197.
PODDER, S. K., SUROLIA, A. & BACKHAWAT, B. K. (1974). On the specificity of carbohydratelectin recognitions: the interaction of a lectin from Ricinus communis beans with simple
saccharides and Concanavalin A. Eur. J. Biochem. 44, 151-160.
Roos, E. & TEMMINK, J. H. M. (1975). Cytochemical comparison between wheat germ
agglutinin and Concanavalin A bound to mouse fibroblasts in vitro. Expl Cell Res. 94,
140-146.
ROSENBLITH, J. Z., UKENA, T. E., YIN, H. H., BERLIN, R. D. & KARNOVSKY, M. J. (1973).
A comparative evaluation of the distribution of Concanavalin A-binding sites on the
surfaces of normal, virally-transformed, and protease-treated fibroblasts. Proc. natn. Acad.
Sci. U.S.A. 70, 1625-1629.
RUTISHAUSER, U. & SACHS, L. (1975). Cell-to-cell binding induced by different lectins.
J. Cell Biol. 65, 247-257.
SACHS, L. (1974). Lectins as probes for changes in membrane dynamics in malignancy and
cell differentiation. In The Cell Surface in Development (ed. A. A. Moscona), pp. 127-139.
New York and London: John Wiley.
SCHNEBLI, H. P. & BXCHI, T. (1975). Reaction of lectins with human erythrocytes. Expl Cell
Res. 91, 175-183.
594
W. J. Van Blitterswijk and others
D. F., NERI, G. & WALBORC, E. F., JR. (1973). Isolation and partial chemical characterization of cell-surface glycopeptides from AS-30D rat hepatoma which possess binding
sites for wheat germ agglutinin and Concanavalin A. Biochemistry, N.Y. 12, 2111-2118.
SMITH, S. B. & REVEL, J. P. (1972). Mapping of Concanavalin A-binding sites on the surface
of several cell types. Devi Biol. 27, 434-441.
SMITH,
TAYLOR, R. B., DUFFUS, W. P. H., RAFF, M. C. & DE PETRIS, S. (1971). Redistribution and
pinocytosis of lymphocyte surface immunoglobulin molecules induced by anti-immunoglobulin antibody. Nature, New Biol. 233, 225-229.
TEMMINK, J. H. M., COLLARD, J. G., SPITS, H. & Roos, E. (1975). A comparative study of
four cytochemical detection methods of Concanavalin A binding sites on the cell membrane.
Expl Cell Res. 92, 307-322.
VAN BLITTERSWIJK, W. J., EMMELOT, P., HJLGERS, J., KAMLAG, D., NUSSE, R. & FELTKAMP, C. A.
(!975)- Quantitation of virus-induced (MLr) and normal cell surface antigens in isolated
plasma membranes and the extracellular ascites fluid of mouse leukemia cells. Cancer Res.
35, 2743-275IWRAY, V. P. & WALBORG, E. F., JR. (1971). Isolation of tumor cell surface binding sites for
Concanavalin A and wheat germ agglutinin. Cancer Res. 31, 2072-2079.
(Received 19 January 1976)