[CANCER RESEARCH 48, 5124-5128, September 15, 1988]
Correlation between Cell Deformability and Metastatic Potential in B16-F1
Melanoma Cell Variants1
Tomasz Ochalek,2 Frank J. Nordt,3 Kjell Tullberg,4 and Max M. Burger5
Biocenler of the University of Basel, Department of Biochemistry, CH-4056 Basel [T. O., K. T., M. M. B.J; and Sandoz, Ltd., Department of Preclinical Research, CH4002 Basel; [F. J. N.J Switzerland
ABSTRACT
Four B16 melanoma cell variants were investigated to determine if
there exists a correlation between their deformability and their metastatic
potential. Cell deformability was measured as the percentage of cells
traversing lO-^m diameter Nuclepore filter membranes at constant pres
sure as a function of time. A method was devised to circumvent common
problems encountered in cell filtration experiments, i.e., cell aggregation
and adhesion to the filter and failure to recover the input. Fla cells with
the lowest spontaneous metastatic rate required 44 s for 50% of the cell
input to traverse the filter, whereas No. 4 cells, featuring the highest
metastatic rate, needed 12s despite the fact that the cells had identical
dimensions. Other variants tested showed intermediate filterability which
also correlated with their metastatic potential. Cells, when pretreated
with cytochalasin B at a final concentration of 21 p\i exhibited increased
filterability (75% and 42% greater than control for Fla and No. 4 cells,
respectively). Somewhat smaller increases were observed after colchicine
treatment. The findings imply major involvement of the cytoskeleton in
the filterability and thus deformability of these B16 variants. Such
physicochemical factors may play an important role in the metastasis of
this and possibly other tumor types.
INTRODUCTION
The mechanical properties of tumor cells can influence their
ability to survive stresses occurring in the microcirculation, thus
modifying their metastatic potential (1-3). Despite other more
specific antitumor defense mechanisms, tumor cells surviving
in the microvasculature must be resistant to the shear stresses
arising in the vascular bed (4), the factional forces arising
between their peripheries and vessel walls (5) as well as have
the ability to traverse capillaries which generally are rigid and
smaller in diameter than tumor cells (1,2, 5). Depending on
the type of tumor, the deformability of tumor cells or lack of it
may play a critical role in their ability to form neoplastic foci.
In the work presented here, we have directly addressed the
question of whether a correlation exists between the deforma
bility of variants of B16 melanoma cells and their metastatic
potential. The influence and the effect of the presence of cytoskeletal elements on deformability of these B16 variants were
also examined.
Although a precise definition of cellular deformability does
not exist, numerous methods are available for measuring prop
erties of cells which are related to their deformability (1, 2, 6).
Micropipet aspiration and filtration methods have been most
widely used in this regard. The former involve, in their simplest
application, measuring the negative pressure required for a cell
Received 12/17/86; revised 4/22/88; accepted 6/3/88.
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
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' Financial Support: Swiss National Foundation Grant 3.269-0.82
2 Present address: Jagiellonian University, Institute of Molecular Biology, A1
Mickiewicza 3, 31-120 Krakow, Poland.
3To whom requests for reprints should be addressed, at Oregon Health
Sciences University, Department of Neurology, Portland, OR 97201.
4 Present address: Sandoz, Ltd., Analytical Department, CH-4002 Basel, Swit
zerland.
5 Present address: Friedrich Miescher Institute, P. O. Box 2543, CH-4002
Basel, Switzerland.
or a portion thereof to enter a pipet the opening of which is
smaller than the diameter of the cell. Alternatively, filtration
methods generally rely on the time it takes for a given volume
of a cell suspension to pass through micropore membranes at
constant pressure or measuring the pressure rise on passing a
suspension through the filter at constant flow. A limitation of
the filtration method is that information regarding material
properties of the cell, such as fluid mechanical parameters,
cannot be deduced and it is possible to obtain only average
information on a cell population, i.e., the presence or the effects
of subpopulations are easily missed. The advantages of the
micropore filtration method lies in the fact that it is simple,
economical, and well suited to answering the questions ad
dressed by this investigation. The methodology for measuring
B16 tumor cell filterability was specifically devised and adapted
with regard to differences in their adhesive and aggregative
properties from other tumor cell types investigated thus far,
e.g., ascites and leukemia cells (7). With this system we were
able to show that the deformability characteristics of variants
of the B16 melanoma system are related to their metastatic
potential.
MATERIALS AND METHODS
Investigations were carried out on three variants of F l B16 melanoma
cells which were selected through culturing on Nuclepore filters, as
described earlier (8). These variants were designated No. 1, No. 2, and
No. 4 and a fourth nonselected variant of Fl was designated FIA.
intraThe metastatic capacity of all variants was assayed in vivo by
i.m. injection and removal of primary tumor and a s.c. injection assay
with the primary tumor left intact (8, 9).
Cell Culture. Cells were cultured as described earlier (9). Briefly, the
cells were grown in a humidified incubator in 100- x 20-mm Petri
dishes in Eagle's minimum essential medium (GIBCO) supplemented
with 10% heat-inactivated calf serum, 1% glutamine, vitamins, and
antibiotics in 5% CO?. An initial cell concentration which allowed cells
to approach subconfluence 2 days after seeding was used. Care was
taken to control cell density strictly since it was found to influence the
filterability. Cells were harvested by rinsing the monolayers with an
EDTA solution (50 ¿tM)
and incubating them at 37°Cfor 5 min. After
resuspension, the cells were transferred to new culture dishes at the
desired concentrations.
Preparation of Cells for the Filtration Assay. The medium was aspi
rated and the cells were washed with 10 ml of Ca2+-Mg2+ free PBS6
(CMF-PBS) and then incubated with 5 ml of CMF-PBS for 5-10 min.
Cells harvested in this manner were then suspended and allowed to rest
at room temperature for 30 min in 20 ml serum-free minimum essential
medium which previously had been supplemented with 5 HIM 4-(2hydroxyethyl)-l-piperazineethanesulfonic
acid and filtered through
0.45-Mm Millipore membranes. The viability of these cells was checked
by the trypan blue dye exclusion test.
Drug Treatment of Cells before Filtration. Cells at a concentration of
2 x 10s cells/ml, in a total volume of 10 ml were incubated for 45 min
with 100 JIMcolchicine (Sigma) end concentration; 21 ¿tM
cytochalasin
B (CB) end concentration or 100 p\i colchicine + 21 pM CB end
' The abbreviations used are: PBS, phosphate buffered saline; CB, cytochalasin
B; CMF-PBS, calcium magnesium free phosphate buffered saline; 7°»,
time in
seconds at which 50% of the total cell input passed the niter.
5124
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1988 American Association for Cancer Research.
TUMOR CELL DEFORMABILI!"* AND METASTASIS
concentrations, at 22°Cor at 4°Cwithout drugs. All cell suspensions,
including controls, contained 20 ¿dof dimethyl sulfoxide, which was
used as the solvent. After the incubation period, cell viability was
checked by dye exclusion and the cell concentration was appropriately
adjusted. Drug or cold-treated cells were filtered immediately after
treatment to avoid cell regeneration. A 10-fold dilution of the incubated
cell suspension which was normally performed in making concentration
adjustments still maintained high enough drug concentrations to dis
rupt the cytoskeleton (10).
Mean cell size distributions of all investigated cell variants in pro
and postfiltrates and after drug or cold treatments were determined
using a Coulter Counter Model ZBI with a C-1000 Channelizer inter
faced with an Apple He microcomputer (11). Size measurements of
nuclei were performed in the same manner following cell disruption
with Zapoglobin (Coulter), a commercial cell lysing agent.
Selection of Filters. One lot (51 F2D9) of 13-mm diameter Nuclepore
filters with 10-^im diameter pores was used. It was found that although
the pore diameter was uniform there was considerable variability in
pore density. Therefore, a device for preselection of filters was con
structed according to a principle previously described (7). It consisted
of an air pump, a water manometer, a pressure regulator, and an outlet
equipped with a valve connected to the filter holder shown in Fig. \A.
Initially, the system was maintained under a constant air pressure of
30 cm HiO. With the filter in place, opening the valve resulted in a
rapid decrease in pressure which was proportional to the total pore
area. Only filters which gave a uniform decrease in pressure of 8-10
cm HiO were used in all experiments.
Cell Filtration. A 20-ml suspension with an appropriate cell concen
tration was prepared. The cells were counted and sized using the Coulter
Counter. The preselected filters were placed in a custom-designed
plexiglass filter holder (Fig. 1). The active filtration area of the filter
was 19.6 mm2 which corresponds to 1.96 x IO4 pores (calculated on
the basis of data provided by Nuclepore). The chamber of the holder
was filled with 1 ml of a cell suspension and connected with silicon
tubing to a reservoir filled with filtration medium. The pressure was
kept constant during filtration by dropwise addition of fresh medium
to the reservoir. Fractions of filtrate were collected in siliconized tubes
at timed intervals using a fraction collector. Cell concentrations in the
filtrate were then determined immediately. A cumulated percentage
curve as a function of time for each cell variant was constructed from
the fraction counts. To determine if cells were lost when collecting
fractions, control counts were performed using a single tube which then
was compared to the cumulative data obtained from fractions. Viability
of the cells in filtrates was checked by combining fractions, centrifuging,
and resuspending them in 0.3 ml of 0.2% trypan blue in PBS solution.
Cells not subjected to filtration were treated similarly and the results
were compared.
To determine the optimal filtration conditions No. 4 and FIA cells,
representing the extremes of metastatic potential, were filtered using
driving pressures of 5, 20, and 40 cm H;( ) at temperatures of 4, 22,
and 37°Cat a cell to pore ratio of 1:20 and 1:1.
RESULTS
Metastatic Potential. The metastatic capacity of the cells was
assayed using an i.m. and a s.c. method and was reported
previously (8). Two assays were used due to the drawback that
the i.m. assay, although less time consuming than the s.c. assay,
involves surgical removal of the primary tumor with the inher
ent risk of accidental seeding of cells and other artifacts ( 12).
Furthermore, the position of the primary tumor does not cor
respond to the origin of a melanoma, namely the skin.
The results, expressed as the percentage of animals with
metastasis of FIA cells, were 28 ±2 in the i.m. assay and 17 ±
16 in the s.c. assay. For No. 1 cells the values were 56 ±11
and 50, for No. 2 cells 64 ±1 and 88 ±9, and for No. 4 cells
79 ±9 and 94 ±5 (8).
Cell and Nucleus Size. The mean cell diameter of all variants
was the same and equalled 17.4 ±0.21 pm. Size distribution
Fig. 1. Construction of the membrane niter holder and arrangement for cell
nitration. A, niter holder consisting of threaded parts 1-3 which are screwed
together for final assembly. The niter is placed between 2 and 3. Part 1 has a
sample injection port (4) and a metal tube to connect to the reservoir (5). B,
experimental setup. The cells are injected with a 1-ml tuberculin syringe (/),
which is left in position. The reservoir (2) is connected and the valve opened after
the working pressure (3) above the filter level has been adjusted. Filtered cells are
recovered at given intervals with the fraction collector (4).
Fig. 2. Relative size distributions of cells before and after filtration. Equivalent
spherical diameters of the cells as presented in the text were calculated based on
calibration of the Coulter system with polystyrene latices as standard particles
(10). A, FIA; B, No. 1; C, No. 4; D, No. 2. Dashed lines, cells before filtration;
solid lines, cells after filtration.
curves of cells in pre- and postfiltrates were essentially identical
(Fig. 2). Treatment with colchicine and CB and incubation at
4°Cdid not result in measurable changes in cell size distribution
or mean cellular diameter (data not shown). The mean diameter
of the cell nuclei was 9.8 ±0.27 pm for all the cell variants and
did not change as a result of filtration or drug treatment.
Cell Filtration. The standard deviations of the number of cells
filtered were less than 5% and that of the filtration times less
than 7% of the mean when filter membranes were preselected
as described above. Pilot studies with different cell-to-pore
ratios, driving pressures, and temperatures were designed to
establish optimal conditions of filtration. Based on the analysis
of filtration times, filtration curve slopes, and percentages of
cells filtered, the following optimal conditions were found: 20
cm FÕ2Odriving pressure; cell-to-pore ratio, 1:1; and tempera
ture, 22°C.
The differences in filterability of cells of the four investigated
melanoma variants manifested themselves in their filtration
time and kinetics of filtration (Fig. 3 and Table 1). In all cases
5125
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1988 American Association for Cancer Research.
TUMOR CELL DEFORMABILITY
the final percentage of filtered cells, however, was similar and
equalled about 90%. The filtration time of FlA-variant cells,
which exhibited the lowest metastatic potential, was about 1.7
times longer than that of No. 4 cells which had the highest
metastatic potential (approximately 500 versus 300 s). Such
times correspond to a complete circulation turnover, e.g., about
5 min in humans, so that they correlate well with haemodynamic kinetics. Filtration times of the No. 2 variant, the met
astatic potential of which was similar to that of No. 4, were
also not statistically different from the No. 4 filtration times.
The initial slope of the cumulated percentage of filtered cells
as a function of time of No. 4 cells was 1.7 times steeper than
the slope of FIA cells (Table 2). The initial slope of variant
No. 1 was intermediate to the two latter ones, but closer to the
slope of FIA cells.
Analysis of 7*5o,i.e., the time necessary for 50% of the cells
AND METASTASIS
of which destroy the microtubular network (Tables 1 and 2).
Following CB treatment, the filtration time of the four variants
was significantly reduced in comparison with the filtration time
of untreated cells. The initial slope of FIA cells treated with
CB becomes nearly equal to that of untreated No. 4 cells (Table
2). The TÕOvalues of CB-treated cells calculated from the
experimental data presented in Fig. 2 are not only shorter, but
the range of the means is smaller in comparison to untreated
cells (Table 3). As can be seen from Table 3, the difference
between FIA and No. 4 is reduced to 4.3 s in comparison with
32.2 s for untreated cells.
Colchicine-treated cells express a pattern of changes which
resembles that of cells which were incubated at 4°C.A decrease
in filtration time and steeper initial slopes are observed for all
four variants, but these changes are smaller when compared
with those observed after CB treatment (Tables 1 and 2). The
to pass through the filter shows even more pronounced differ
TSOvalues of colchicine treated cells are shorter than that of
ences in filtration behavior. The values are presented in Table
untreated cells, but longer and more dispersed when compared
with the TÕO
values of CB-treated cells (Table 3). Incubating
3.
Treating the cells with agents which disrupt the cytoskeletal
cells simultaneously with CB and colchicine resulted in changes
elements of the cells (colchicine and CB or incubation at 4°C) similar to those caused by CB treatment alone, although a bit
reduced filtration times and increased the slopes of the curves,
more pronounced (Tables 1 and 2).
but the percentage of cells passing through the filter still re
mained about the same (Fig. 4, Tables 1 and 2). CB, which
DISCUSSION
damages microfilaments, modifies the filterability to a greater
extent than colchicine and incubation at low temperature, both
It was the aim of this investigation to assess melanoma cell
deformability and to determine whether this parameter is as
sociated with metastatic potential. The relevance of cell filterability in vitro to the Theological or other types of behavior of
cells /'// vivo is controversial and has been widely discussed.
Certainly cell filterability cannot be regarded as a sufficient
model for a number of aspects of cell behavior in the microvasculature (13, 14). For example, tumor cell filterability cannot
be applied for selecting cell variants on the basis of their
resistance to mechanical trauma, because the resistance to
25 mechanical stress is not heritable (15). However, it was not the
intention here to consider micropore filtration as a model of
the microvasculature, but to study tumor cell deformability in
relation to metastatic potential.
Considerable attention has been paid to eliminating, as far
as
possible, interference by cell properties other than deforma
Fig. 3. Time course of nitration as cumulated percentage of total cell input.
The first 11 fractions (10 in the case of FIA) were taken at 2.3-s intervals except
bility, such as their adhesive or aggregative behavior, which
the Sth which was taken after a 3.5-s interval. After this point, i.e., 26.5 s (24.2
may become manifest, depending on the experimental condi
for FIA) total filtration time, longer intervals were chosen as the number of cells
filtered per unit time decreased. /'*,,values are indicated with arrowed symbols on
tions. In contrast to previous studies (2, 6, 7) filtration time is
the abscissa and with intersections on the curves. Bars, SD of the mean. N = 5
not defined here as the time necessary for passing a given
for FI A (A) and No. 2 (•).N = 6 for No. 1 (O) and No. 4 cells (D).
volume of cell suspension through a filter, but as the time
necessary for filtering all cells capable of passing through the
Table 1 Final filtration times ofBI6 melanoma cell variants
filter under given conditions, i.e., driving pressure, pore size,
Successive fractions were collected and counted with a Coulter counter until
and cell-to-pore ratio. It was found in pilot studies that attempts
cell numbers were only twice the background count. The time in seconds at which
at correlating deformability of melanoma cells with filtration
this occurred was defined as the final filtration time.
Treatment of cells prior to filtration"
time as defined by the passage of a given volume of a cell
suspension
did not only yield unsatisfactory results, but made
CB47 +
CellFIANo.
them difficult to interpret. This appeared to result from the
33*(5)370
±
±9(7)47
±6(9)38
16(7)180(7)102(6)90(7)Colchicine390*(7)191
±
increase in cell concentration in prefiltrates (higher cell-to-pore
16(8)90(6)90(8)CB57
±
±6(9)42(7)42(7)Colchicine
±6(9)30(9)30(9) ratio) which led to cellular aggregation, a problem which could
1No.
±31(6)307
potentially be eliminated by lowering the cell concentration
2No.
23(5)292
±
thereby decreasing the cell-to-pore ratio and making cell-cell
4None504
12(6)4-C373
±
interactions less likely to occur. However, when this is done,
not all potentially filterable cells are in fact filtered since the
" All treatments resulted in significantly reduced filtration times at P < 0.001
driving fluid, i.e., the suspension medium passes ahead of the
with Student's t test. Among the untreated cells there is no significant difference
cells through unoccupied pores. The unlimited supply of sus
between No. 2 and No. 4 cells, whereas the others differ at P < 0.001 and No. 1
and No. 2at/><0.005.
pending medium used here eliminates this possibility. It also
'' Mean ±SD derived from the number of experiments shown in parentheses.
ensures that the problem of counting cells in the final stages of
' For values without SD, the endpoint was reached at the same fraction number,
filtration becomes the only limiting factor since the number of
i.e., at the same time, for all nitrations on that particular cell variant.
5126
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1988 American Association for Cancer Research.
TUMOR CELL DEFORMABILITY AND METASTASIS
Table 2 Initial slope of filtration curves ofB16 variants
In order to obtain values of cell passage rates at the beginning of filtration, the slope during the 0-4.6-s interval (corresponding to two fractions) was calculated
from all experiments. Additional filtrations were run with untreated cells for this purpose and for the /'<„
calculations presented in Table 3.
Treatment of cells prior to filtration
CB6.76 +
CellFIA
±0.19*
±0.32°
(12)
(7)
4.91 ±0.10(7)
1No.
No.
4.41 ±0.23
(19)
7.50 ±0.18
2
5.62 ±0.33
(20)
(6)
6.10 ±0.29
7.05 ±0.20
No. 4None3.69
(12)4'C3.91
(7)Colchicine4.24
" Mean ±SD derived from the number of experiments shown in parentheses.
'' Incubation of FIA cells at 4*C did not lead to a significant change; in all other
+ 0.17
(7)
5.39 ±0.17
(8)
7.55 ±0.23
(6)
7.68 ±0.19
(8)CB5.96
±0.24
(7)
7.25 ±0.22
(9)
8.53 ±0.12
(7)
8.41 ±0.29
(7)Colchicine
±0.22
(9)
8.41 ±0.15
(9)
9.78 ±0.12
(9)
9.76 ±0.25
(9)
cases significant increases in initial slopes versus untreated cells were observed at
P < 0.001. Statistical analysis of the differences among untreated cells showed P< 0.001 in all possible combinations.
Fig. 4. Time course of nitration after cells have been treated with CB. Bars,
SD of the mean. N = 7 for FIA (A), No. 2 (•),and No. 4 (D). N = 9 for No. 1
(O) cells. Other details are the same as described for Fig. 3 except that the short
interval sampling was ended at the tenth fraction for all variants.
Table 3 Tw values
Values are defined as the time in seconds for 50% of the cell input to pass the
filter.
Treatment of cells prior to filtration
CellFIA
±12.5°
±1.0
±0.6
No. 1
17.3 ±1.3
13.6 ±0.4
8.6 ±0.2
11.8 ±0.6
8.4 ±0.4
6.6 ±0.2
No. 2
No. 4None43.711.5 ±0.5
7.9 ±0.2
6.7 ±0.3
14.6*
11.0
7.6
(7.9-19.2)CB11.0
(6.7-11.0)
(11.5-43.7)Colchicine19.2
* Mean ±SD derived from / <„
values calculated from the individual experi
ments presented in Table 2. All drug treatments led to significant decreases in
T"»
vs. untreated cells at P < 0.001. Untreated cells differed from each other at P
< 0.001 except for No. 2 and No. 4 cells which did not differ significantly.
* Median values of the means in each category with the range in parentheses.
cells passing through the filter per unit time is small at that
moment.
The results show that the ability of the variants of B16
melanoma cells to pass through pores having a diameter equal
to about 57% of the cell correlates with their metastatic poten
tial (r2 = 0.97 and 0.98 for the percentage metastasis versus
final filtration time using the i.m. and s.c. assays, respectively,
P < 0.05 in both cases). Filtration curves of all the variant cell
types exhibit steep slopes in the beginning and then decrease
so that 50% of cell input passes through the filter after 10-50
s, i.e., over 10 times shorter than the total filtration time. The
changes in slope with time suggests that the cell variants tested
may be comprised of several subpopulations differing in their
filterability. It appears that the cells which undergo filtration
within TÕO
are more deformable. However, the proportion of
these cells are different in the various cell types. For example,
T50for No. 4 is 11.5 s at which point only 34% FIA cells have
passed through the filter.
With time the conditions of filtration undergo changes. Mi
croscopic examination of filters fixed with 4% formaldehyde in
PBS at a time corresponding to T50shows that about 20% of
the pores are occupied by cells. This transiently decreases the
chances of succeeding cells which reach the filter to pass
through the pores and results in some delay of filtration of
potentially equally deformable cells. At the filtration end point
only a few percent of the pores remain obstructed by cells.
The influence of varying filtration conditions (filter pore size,
temperature, pressure, cell adhesion) on the results has also
been discussed earlier (6, 7). Our observations made in prelim
inary experiments were consistent with these observations.
Therefore it needs only to be mentioned that no appreciable
differences were found in the percentage of cells filtered or in
the shapes of filtration curves at 37 and 22°C(data not shown).
On the other hand, filtration at 4°Cwas characterized by a
decrease in the percentage of filtered cells compared to filtration
carried out at 22°C.This decrease occurred not only when cells
were exposed to 4°Cimmediately before filtration, but also with
cells incubated at 4°Cfor 45 min. Therefore, the increase in
suspending medium viscosity was not compensated by slightly
greater cell deformability resulting from disruption of the microtubular system by low temperature.
The differences in cellular Theological properties of the tumoi
cell variants reflect the sum of the differences in the physicochemical properties of the cytoskeleton, the nucleus, and the
plasma membrane. Cell nuclei are regarded as being less de
formable than the surrounding cytoplasm (6, 15). The ability
of melanoma cells to migrate through pores having a diameter
five times smaller than their nucleus (8) shows that significant
deformation can take place under certain conditions. Filtration
of Ehrlich Ascites cells and LI210 cells through pores having
a diameter two times smaller than their nuclei results in cell
damage revealing itself through inhibition of DNA synthesis
and only later through changes in the integrity of the plasma
membrane (14, 15). This suggests that the nucleus is more
susceptible to mechanical stress than had previously been
thought. Considering the low viscosity of the cytoplasm and
the pore size chosen here, which roughly equals the diameter
of the nucleus, it seems unlikely that differences in nucleus
pliability would be the determinant responsible for the observed
differences in deformability.
The aim of treating cells with CB or colchicine or incubating
them at low temperature was to examine the effect that microtubules and microfilaments have on cell deformability. Treat
ment of cells with CB disrupts microfilament function (16, 17)
whereas incubation at 4°Cor treatment with colchicine results
5127
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1988 American Association for Cancer Research.
TUMOR CELL DEFORMABILITY
in microtubule dissociation (18, 19). The results strongly sug
gest that the cytoskeleton dominami) influences cellular deformability in the melanoma system. Damage to cytoskeletal
elements increases the deformability of all cell variants in
comparison to untreated cells. Disruption of microfilaments by
CB is known to reduce heterotypical adhesion, homotypical
aggregation, and agglutinability by wheat germ agglutinin in
the B16F10 melanoma system, in vitro (6). In the system
described here, CB treatment not only increases cell filterability,
but also decreases the differences in deformability of the mela
noma cells in comparison with untreated cells. The differences
in TÕO
between FIA and No. 4 cells decrease from 32.2 to 4.3 s
while the slope of CB-treated FIA cells become equal to that
of untreated No. 4 cells. In agreement with the literature,
changes caused by CB are not followed by a change in cell
volume ( 10). Incubation of cells with colchicine does not change
differences between variants to the same degree as treatment
with CB. This observation points to the fact that it is the CBsensitive microfilaments of the B16 melanoma variants, rather
than the microtubular structures, which are overridingly re
sponsible for their different deformability characteristics.
It remains to be seen why the small differences in the kinetics
of nitration remain after treatment with CB. Previous investi
gations have shown that the degree of actin organization of IS
clones of melanoma K1775 cells was inversely correlated with
their metastatic potential (20, 21). A similar relationship has
also been shown for lectin-resistant mutants of the B16 mela
noma (22). Therefore, it is possible that the amounts of the
different distribution of the microfilament degradation products
in cells with low metastatic potential results in greater rigidity
of these cells in comparison to cells which have a poorly
organized cytoskeleton. An influence of the different surface
properties or architecture on intercellular interactions and on
the interactions of the cells with the filter membrane or pore
channels cannot presently be ruled out either. Furthermore,
differences may exist in the rate at which the tumor cells deform
under the given experimental conditions, thus affecting their
filtration kinetics. This possibility is underscored by results
reported by Weiss and Clement (23) who showed RPMI No.
41 cells to have bulk dilatant properties. The expression of
these rheological properties depended on the magnitude as well
as a threshold value of the applied shear stress. Other cell types
have also been shown to have thixotropic properties (24).
In conclusion, the results of the present study suggest an
important role of the cytoskeleton, especially the microfilament
system, in the deformability of B16 tumor cells. The experi
mental design allowed us to circumvent two problems com
monly encountered in cell filtration experiments, namely cellcell adhesion and failure to detect all filterable cells. This
approach may prove to be a valuable tool for investigating the
role of tumor cell deformability in metastasis or survival in the
circulation.
AND METASTASIS
ACKNOWLEDGMENTS
We wish to thank Dr. Gradimir N. Misevic for critical reading of
the manuscript and Dr. Geraldine Shantz who did some preliminary
tests on tumor cell filtration not reported here.
REFERENCES
1. Sato, II., and Suzuki, M. Deformability and viability of tumor cells by
transcapillary passage, with reference to organ affinity of metastasis in cancer.
In: L. Weiss (ed.). Fundamental Aspects of Metastasis, pp. 311-317. Am
sterdam: North-Holland, 1976.
2. Sato, II.. Minio, .1.. Sato, T., and Suzuki, M. Deformability and filterability
of tumor cells through "Nuclepore" filter, with reference to viability and
metastatic spread. Gann. Monogr. Cancer Res., 20: 3-13, 1977.
3. Weiss, L., and Claves, D. Cancer cell damage at the vascular endothelium.
Ann. NY Acad. Sci., 416:681-692, 1983.
4. Brooks, D.E. The biorheology of tumor cells. Biorheology, 21:85-91,1984.
5. Weiss, L., and Dimitrov, D. S. A fluid mechanical analysis of the velocity,
adhesion, and destruction of cancer cells in capillaries during metastasis. Cell
Biophys., 6: 9-22, 1984.
6. Erkell, L. J., Ryd, W., and Hagmar, B. Comments on the filter test for tumor
cell deformability. Invasion Metastasis, 2: 260-267, 1982.
7. Khato, J., Sato, II. Suzuki, M., and Sato, H. Filterability and flow charac
teristics of leukemic and non-leukemic tumor cells suspension through
polycarbonate filters in relation to hematogenous spread of cancer. Tohoku
J. Exp. Med., 128:273-284, 1979.
8. Tullberg, K. F., and Burger, M. M. Selection of B16 melanoma cells with
increased metastatic potential and low intercellular cohesion using Nuclepore
filters. Invasion Metastasis, 5:1-15, 1985.
9. Tao, T. W., and Burger, M. M. Lectin-resistant variants of mouse melanoma
cells. I. Altered metastasizing capacity and tumorigenicity. Int. J. Cancer,
29:425-430, 1982.
10. Hart, I. K . Raz, A., and Fidler, I. J. Effect of cytoskeleton-disrupting agents
on the metastatic behavior of melanoma cells. J. Nati. Cancer. Inst., 64:891899, 1980.
11. Wenger, J., Nowak, J. S . Kai, O., and Franklin, R. M. Display and analysis
of cell size distributions with a Coulter Counter interfaced to an Apple II
microcomputer. J. [umiliimi. Methods, 54: 385-392, 1982.
12. Stackpole, C. W. Distinct lung-colonizing and lung-metastasizing cell popu
lations in B16 mouse melanoma. Nature, 289: 798-800, 1981.
13. Chien, S. Principles and techniques for assessing erythrocyte deformability.
Blood Cells, 3: 71-99, 1977.
14. Gabor, H., and Weiss, L. Mechanically induced trauma suffered by cancer
cells in passing through pores in polycarbonate membranes. Invasion Metas
tasis, 5: 71-83, 1985.
15. Gabor, II.. and Weiss, L. Survival of L1210 and Ehrlich ascites cancer cells
after mechanical trauma is a random event. Invasion Metastasis, 5: 84-95,
1985.
16. Weber, K., Rathke, P. C, Osborn, M., and Franke, W. W. Distribution of
actin and tubulin in cells and in glycerinated cell models after treatment with
cytochalasin B (CB). Exp. Cell Res., 102: 285-297, 1976.
17. Wessells, N. K., Spooner, B. S., Ash, J. F., Bradley, M. O., Luduena, M. A.,
Taylor, E. L., Wrenn, J. T., and Yamada, K. M. Microfilaments in cellular
and developmental processes. Science (Wash. DC), 171: 135-143, 1971.
18. Borisy, G. G., Olmsted, J. B., and Klugman, R. A. In vitro aggregation of
cytoplasmic microtubule subunits. Proc. Nati. Acad. Sci. USA, 69: 28902894, 1972.
19. Murivi. M. M., and De Mets, M. Effect of microtubule inhibitors on invasion
and on related activities of tumor cells. Int. Rev. t viol., 90:125-168, 1984.
20. Raz, A., and Geiger, B. Altered organization of cell-substrate contacts and
membrane-associated cytoskeleton in tumor cell variants exhibiting different
metastatic capabilities. Cancer Res., 42: 5183-5190,1982.
21. Volk, T., Geiger, B., and Raz, A. Motility and adhesive properties of highl
and low-metastatic murine neoplastic cells. Cancer Res., 44: 811-824,1984.
22. Tao, T. W., Jenkins, J. M., Vosbeck, K., Matter, A., Miller, M., Jockusch,
B. M., Shen, /... and Burger, M. M. Lectin-resistant variants of mouse
melanoma cells. II. In vitro characteristics. Int. J. Cancer, 31:239-247,1983.
23. Weiss, L., and Clement, K. Studies on cell deformability. Some rheological
considerations. Exp. Cell Res., 58: 379-387, 1982.
24. Curtis, A. S. G. Cell contact and adhesion. Biol. Rev., 37: 82-129, 1962.
5128
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1988 American Association for Cancer Research.
Correlation between Cell Deformability and Metastatic Potential
in B16-F1 Melanoma Cell Variants
Tomasz Ochalek, Frank J. Nordt, Kjell Tullberg, et al.
Cancer Res 1988;48:5124-5128.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/48/18/5124
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 16, 2017. © 1988 American Association for Cancer Research.