0 1992 Wiley-Liss, Inc.

Cytometry 13:9-14 (1992)

Invasive EDithelial Cells Show More Fast Plasma Mernbrani Movements Than Related or Parental Non-Invasive Cells’ Nicolas A.F. van Larebeke, Marc E. Bracke, and Marc M. Mareel Laboratory of Experimental Cancerology, Department of Radiotherapy and Nuclear Medicine, University of Gent, B 9000 Gent, Belgium Received for publication August 3, 1990; accepted July 30, 1991

Fast plasma membrane movements (FPMM) are involved in ruffling, blebbing, fast shape change, and fast translocation. A simple method for the quantification of FPMM was used to study the relation between FPMM and invasive capacity in five pairs of invasive and noninvasive variants from four different epithelial cell types. The human mammary cell line MCF-7/6, the ras-transformed dog kidney cell line rus-MDCK, the rustransformed mouse mammary gland cell lines NM9-rus-12 and NM-f-rus-TD, and spontaneously transformed late passage mouse lens explant MLE cells, all of which were invasive in vitro, showed more FPMM in our measurements and

Cell motility is involved in invasion as demonstrated indirectly by histology both at the light microscopic (28) and electron microscopic level (6). This relation has also been demonstrated directly by time-lapse microcinematography in vitro in organ culture and in vivo in the rabbit ear chamber and in the mesentery of rats and mice (28). As to a role for increased cell motility in the expression of the malignant phenotype, Varani et al. (35) and Hart (13) found no correlation between speed of migration in vitro and metastatic capability in vivo. In several recent publications, however, evidence is presented in favour of a positive correlation between malignancy, as defined by invasive capability sometimes leading to metastasis (16), and migration or translocation in vitro (3,7,17,21,23,30, 32,34,38). Selection of B16 melanoma cells showing faster transmigration through nucleopore filters led to cell lines with increased metastatic potential (31). Grimstad (10) selected directly for the most rapidly migrating cells from a fibrosarcoma culture by successive cycles of migration through a micropore membrane; selected cells showed increased invasiveness as well in vitro as in vivo. Translocation is not the only aspect of

displayed more ruffling activity on timelapse video films than the related or parental MCF-?/AZ, MDCK-3, NM9, and NM-f cell lines and early passage MLE cells, none of which were invasive. Interestingly, induction of invasive capacity in MCF-7/AZ cells by retinoic acid was accompanied by an increase in FPMM, but speed of translocation was not increased. Together these observations support the hypothesis that a certain level of FPMM is a prerequisite for invasive capacity.

Key terms: Cell-motility, cell-membrane, cell-surface, ruffling, cell-shape, imageanalysis, invasion, malignant, retinoicacid, translocation

cell motility that has been related to invasion and it is still not clear what aspect, if any, is critical for the invasive phenotype. Guirguis et al. (11)reported that autocrine motility factor (14) stimulated the formation of “pseudopodal protrusions’’ from cancer cells. Translocation, ruffling, and pseudopodal activity were found to be increased in Dunning rat prostatic adenocarcinoma cells compared to normal rat prostate cells (20) and in highly metastatic rat prostatic adenocarcinoma cells compared to rat prostatic adenocarcinoma cells with low metastatic ability (21,23,24).Taniguchi et al. (29)demonstrated that transfer of the v-fos oncogene to src-transformed rat 3Yl cells rendered these cells highly metastatic; these cells showed a n increased invasiveness into MatrigelR-coated filters, associated with a n increase in motility measured in terms of the

‘Supported by the Flemish Advisory Commission on Cancer Prevention, by the Fondation Philippe et Therese Lefevre, and by Actie Kom Op Tegen Kanker 1989.

10

VAN LAREBEKE ET AL.

intensity of “momentary alterations of cell shape,” the latter reflecting ruffling, pseudopodal activity, and fast cell translocation. We used the method described in the accompanying paper (33)to study fast plasma membrane movements (FPMM) in relation to invasive capacity in vitro using five pairs of cell variants, with each pair consisting of an invasive and a non-invasive related or parental cell variant. Retinoic acid has been shown to induce the invasive phenotype in MCF-71AZ human mammary carcinoma cells in vitro (5). This offered us the possibility to examine whether or not induction of invasion was accompanied by a n increase in FPMM or a n increase in speed of translocation of individual cells.

MATERIALS AND METHODS Cell Lines and Culture Media Five pairs of invasive (I+)and non-invasive (l-) variants were chosen from the following four cell lines: the MDCK dog kidney cell line derived from tubular epithelium (15);the MCF-7 human breast cancer cell line obtained from a pleural effusion (26); the NMuMG mouse mammary cell line isolated from the normal mammary gland of a n outbred mouse (22); and mouse lens explant culture (MLE) cells brought into primary culture in our laboratory by L. Messiaen (18). MDCK-3 cells were I- in organ culture (25). In contrast, rus-MDCK cells, obtained from MDCK-3 cells after transformation by Harvey murine sarcoma virus (8), were I ’ (4). The culture medium for MDCK cells was Dulbecco’s Modification of Eagle’s Medium (DMEM, Gibco BRL, Gent, Belgium), supplemented with: 10% (v/v) Fetal Bovine Serum (FBS, Gibco), 0.05 %(w/v) L-glutamine (Gibco), and 250 IUiml penicillin (hereafter called complete DMEM). Among the members of the MCF-7 cell family we chose the I- variant MCF-7/AZ and the I+ variant MCF-716 (5,9). MCF-7 cells were cultured as described by Bracke et al. (5). NM9 cells (I-) were subcloned from a n NMuMG(ATCC) culture. NM9-rus-12 cells (1’) were obtained from NM9 cells after co-transfection with plasmids containing respectively a neomycin resistance gene and the mutated c-Ha-rus oncogene (unpublished results in collaboration with F. Van Roy, Laboratory of Molecular Biology, State University, Ghent, Belgium). N M f cells (I-) were cloned like NM9 cells and selected for their fibroblastoid morphology. NM-f-ras-TD cells (I+) were transfected like NM9-ras-12 cells and selected after tumor formation in a nude mouse (36). NMuMG cell variants were cultured in complete DMEM with 10 pg bovine insulin/ml (Gibco). Till passage 16 in vitro (early passage cells) MLE cells did not invade in organ culture (1- ). After spontaneous transformation, they were, a t passage 18 and afterwards (late passage cells), invasive in organ culture (1’) (18).MLE cells were cultured a s described by Messiaen et al. (18).

Drugs MCF-7iAZ cells were treated with all-trans-retinoic acid (RA; Sigma, St. Louis, MO, cat. No. R-2625) at 1OP6M(final concentration) dissolved in ethanol. The final concentration of ethanol in the culture medium was 0.1%. Cell Motility Terminology Terms describing different aspects of cell motility have been used according to the definitions given in the accompanying paper (33). Microscopy, Time-Lapse Video, and Quantification of Fast Plasma Membrane Movements Methods were a s described in the accompanying paper (33).Briefly, video image 2, taken 28 seconds after image 1, was subtracted from image 1, and the surface area of brighter and darker spots in the subtraction image, corresponding to areas of movement, was measured leading to the parameter motile urea in Fm2 per cell. Measurement of Translocation Distance We made time-lapse video films of microscopic fields selected a t random from the cell cultures to be analysed. Through the frame grabber of a Vidas image analysis computer two video images of the same microscopic field a t times t, and t,, separated by 30 min, were digitized; resulting pairs of images were carefully screened for artefactual movement and were rejected if there was the slightest hint that such movement had occurred. Cell contours were interactively traced using a Vidas “edit” function. The geometric center of gravity was determined for each cell at both times t, and t,, and distance of translocation was calculated as the distance in pm between the geometric centre of gravity at times tl and t,. Experiments FPMM of two cell type variants were compared using cultures in 25 cm2 tissue culture plastic flasks containing 5 ml medium. Measurements were done on matched flasks 3 dafter seeding a t the same cell density. Results are presented under the form of scatter graphs, in which each point represents the result of a measurement of the motile area in pm2 per cell on a group of cells in one randomly chosen video-microscopic field. The corresponding value was obtained by dividing the surface of the total area detected as involved in movement (sum of the areas of all detected objects in the field) by the number of cells in the corresponding field. For MLE cells, results of three experiments, using different optical conditions, were pooled, and motile area per cell of a group of cells in one field was expressed as a percentage of the mean value for the parameter motile area in Fm2 per cell in the corresponding experiment for both types of cells taken together. The number of fields mea-

... . -... .... .. --....

FAST PLASMA MEMBRANE MOVEMENTS AND INVASION

700

600

- 500 -

0 a,

2

5 400

c ._

m

QI

300

.--

a, .0

=

200

100

0

.. . -......

.. -... .

1 l1 a6 00 l 1401

3 % 120 {

5

.E 1 0 0

?

1

MCF-716

I

60

oJ MC F-7lAZ

11

-____ MDCK-3

r a s -MDCK

FIG.1. FPMM of MCF-7IAZ (I-) and MCF-716 (1') human breast carcinoma cells. Flasks seeded with 500,000 cells. A x 32 objective with phase contrast was used. Median values are indicated by a horizontal line (P=0.027).

FIG.2. FPMM of MDCK-3 (I-) and ras-MDCK (I') dog kidney cells. Flasks seeded with 50,000 cells. A x32 objective with phase contrast was used. Median values are indicated by a horizontal line (P = 0.0001).

s u e d and the total number of cells present in them are mentioned for each set of measurements. To study the effect of RA on speed of translocation and FPMM, translocation distance was measured and FPMM were quantified on the same flask, within 2 h, about 72 h after respectively adding retinoic acid or solvent alone to 25 cm2 flasks that were seeded 3 d earlier with 25,000 cells per flask.

fiber-like cells (18). Fiber-like cells showed more motility than flat cells in both early and late passage cultures. Quantification of FPMM with the 32 x p h setting (in earlier experiments with MCF-7 and MDCK lines) or 32 x bfsetting (for NM9, NM9-rus-12, NM-f, NM-f-rusTD and MLE cells) allowed objective confirmation of the observed differences. Motile area in p m 2 per cell (mean value ? standard deviation) was 254 pm2 ? 93 pm2 for MCF-7/AZ (nine fields with in total 556 cells) against 443 pm2 2 178 pm2 for MCF-7/6 (seven fields with 714 cells) (P= 0.027) (Fig. 1); 90.3 pm2 2 25.3 pm2 for MDCK-3 (ten fields with in total 75 cells) against 155.1 pm2 5 21.8 pm2 for rus-MDCK (nine fields with in total 96 cells) ( P = O . O O O l ) (Fig. 2); 4.91 pm2 ? 1.45 pm2 for NM9 (13 fields with in total 2,390 cells) against 15.79 pm2 k 4.07 pm2 for NM9-ras-12 (ten fields with 1,523 cells) (P