In VitroCell.Dev.Biol.28A:364-368, May 1992 9 1992 TissueCultureAssociation 0883-8364/92 $01.50+0.00
EFFECT OF SCATTER FACTOR AND HEPATOCYTE GROWTH FACTOR ON MOTILITY AND MORPHOLOGY OF MDCK CELLS YUAN LI, ANSAMMA JOSEPH, MADHU M. BHARGAVAl, ELIOT M. ROSEN, TOSHIKAZU NAKAMURA, A~O ITZHAK GOLDBERG
Department of Radiation Oncology, Long Island Jewish Medical Center, New Hyde Park, New York 11042 (Y. L., A. J., M. M. B., I. G.); Department of Therapeutic Radiology, Yale University School of Medicine, New Haven, Connecticut 06510 (E. M. R.); and Faculty of Science, Kyushu University, Fukuoka, Japan (T. N.) (Received 13 November 1991; accepted 17 December 1991)
SUMMARY
We investigated the effects of human placental scatter factor (hSF), mouse scatter factor (mSF) and recombinant human hepatocyte growth factor (HGF) on motility and morphology of individual Madin-Darby canine kidney cells using a computerized cell tracking system. All three factors increased the velocity of individual cells and the ratio of moving to stationary cells. Similarly, all three factors caused changes in morphologic features of cells, leading to increased area, flatness, and polarity. Increases in area and flatness but not polarity were slightly greater with HGF than with hSF or mSF. These results suggest that SFs and HGF have similar effects on motility and morphology of isolated epithelial cells.
Key words: scatter factor; hepatocyte growth factor; cell motility; cell morphology. Preparation of SFs and HGF. Mouse SF (mSF) was purified from serum-free conditioned medium from ras-transformed mouse N1H/2 3T3 cells (clone D4) (Rosen et al., 1990b). Human placental SF was purified as described (Rosen et al., 1990c; Bhargava et al., 1991). Recombinant human HGF was purified from culture supernatants from C127 mouse fibroblasts transfected with a complete human HGF cDNA (Nakamura et al., 1989). Scatter assay. Scatter activity was quantitated by the MDCK serial dilution scatter assays as described previously (Stoker et al., 1987; Rosen et al., 1990b). In this assay, small colonies of MDCK cells are incubated for 20 h with serial dilutions of test sample in 96-well plates, stained, and examined for scatter effect (i.e. spreading of normally tight cohesive colonies, dispension of colonies into individual cells, and conversion to fibroblastic appearance). Scatter activity at the limiting dilution is defined as 0.5 MDCK scatter units per milliliter, allowing calculation of the activities in the original undiluted sample. Preparation of cellsfor tracking. Single cell suspensions were prepared from subconfluent cultures of MDCK cells and 2000 to 5000 cells in DMEM + 10% FBS were seeded into 25-cm 2 culture flasks (Nunclon 152094, capacity 40 ml). Cells were incubated for 4 h at 37 ~ C to allow attachment. Flasks were then filled completely with fresh medium with or without factors, capped tightly, and incubated for another 2 h. Flasks were then placed in a temperature-controlled enclosure in the cell-tracking system. Cell tracking. A multipurpose automated microscope system, Dynamic Microscope Image Processing Scanner (DMIPS), developed by the Cancer Imaging Section, B.C. Cancer Research Center, Vancouver, British Columbia (Thurston et al., 1986, 1988) was used to investigate cell motility and morphology. This system consists of an optical scanner, a high precision computer-controlled microscope stage, a fast image processing unit, a host computer, an incubator with temperature control (+0.1" C), automatic data acquisition, and an image recognition software package. Before placing the flask on the stage, the incubator housing stage was warmed to 37 ~ C for 30 min. The X-Y controller was used to move the stage manually to select individual cells to be tracked. Selected cells were positioned near the microscope's cross hair. For better recognition results, only round and flattened cells, which had an approximate area of 100 gm 2
INTRODUCTION Scatter factors (SFs) are hetcrodimcric glycoproteins, produced by fibroblasts and smooth muscle cells which disperse (scatter) cohesive colonies of epithelium and stimulate cell motility (Stoker et al., 1987; Gherardi et al., ]989; Rosen et al., 1989a, 1990b,e). SFtreated cells show membrane ruffling, and protrude and retract conspicuous cytoplasmic processes (Warn and Dowrick, 1989; Dowrick and Warn, 1991). SF-induced motility may be mediated in part by new protein synthesis, alterations in protein phosphorylation, and eytoskeletal reorganization (Rosen et al., 1990a). To better understand the action of SF it is important to determine its effects on individual cells. We utihzed a computerized cell tracking system (Thurston et al., 1986, ]988) to quantitate the motility and morphologic characteristics of large numbers of isolated Madin-Darby canine kidney (MDCK) epithelial cells. In view of considerable sequence homology between SFs and hepatocyte growth factors (HGFs) (Gherardi and Stoker, 1990; Rosen et al., 1990b; Weidner et al., 1990) we also studied the effect of HGF on the motility and morphology of MDCK cells. MATERIALSAND METHODS
Cell culture and reagents. MDCK epithelial cells were obtained from American Type Culture Collection, Rockville, MD. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), 100 U. m1-1 penicillin, 100 gg. ml-l streptomycin, and 0.1 mM nonessential amino acids as described previously (Rosen et al., 1990b). Cells were incubated at 37 ~ C in a humidified atmosphere of 5% CO2:95% air.
l To whom correspondence should be addressed.
364
365
SCATTER FACTOR INDUCED CELL MOTILITY ~.
0.8-
peak on the monitor. The position of the cell and its morphologic and photometrical characteristics were recorded in the computer memory. During the system's automatic revisit program, 18 features were used for cell detection and discrimination (Poon et al., 1987). If a cell was not detected in the search area by a window of 96 • 96 #m 2 centered at the previous recorded cell position, a larger area (200 • 200#m 2) was searched. If the cell was still not detected, it was considered as "lost" and not revisited. Sixty cells per experimental condition were tracked every 10 min for 8 h. Two independent experiments were performed, and the final plots and statistical evaluation were based on data for 120 cells from each treatment group. At the end of the experiment, the quantitative "path" of the cell movement (walk pattern) was determined. The dynamic morphology of moving cells was extracted and quantitated at different stages of cell movement (see below). Classification of cells into motile and stationary groups. Cells were classified as "moving" or "stationary" if their average velocity was greater or less, respectively, than the global average velocity of all cells tracked. Determination of morphologic features. Morphologic changes were quantitated by measuring the area, flatness, and circularity of each cell at 10-min intervals during the 8-h tracking period. These values were then averaged for the 120 cells in each treatment group. The area of a cell in the field (96 X 96 t t m 2 window) was determined at the threshold of 8 grey level units above background intensity. Cell flatness is a measure of the roughness of surface. To determine the flatness at each point of the perimeter the sobel edge detection values were compared to the input threshold, to provide a sum of pixels having a high gradient (F = A/V% where F is cell flatness, A is area in pixels, and V is volume). A higher value (measured in percentage) represents a more flattened cell. Circularity is defined as perim-
Z
:~ tn 7
0.7-
o
0.6.
-~
0.5.
n,'-
0.4. 0
o _J
0.3.
w
0.2
,,,
EFFECT OF SCATTER FACTOR AND HEPATOCYTE GROWTH FACTOR ON MOTILITY AND MORPHOLOGY OF MDCK CELLS YUAN LI, ANSAMMA JOSEPH, MADHU M. BHARGAVAl, ELIOT M. ROSEN, TOSHIKAZU NAKAMURA, A~O ITZHAK GOLDBERG
Department of Radiation Oncology, Long Island Jewish Medical Center, New Hyde Park, New York 11042 (Y. L., A. J., M. M. B., I. G.); Department of Therapeutic Radiology, Yale University School of Medicine, New Haven, Connecticut 06510 (E. M. R.); and Faculty of Science, Kyushu University, Fukuoka, Japan (T. N.) (Received 13 November 1991; accepted 17 December 1991)
SUMMARY
We investigated the effects of human placental scatter factor (hSF), mouse scatter factor (mSF) and recombinant human hepatocyte growth factor (HGF) on motility and morphology of individual Madin-Darby canine kidney cells using a computerized cell tracking system. All three factors increased the velocity of individual cells and the ratio of moving to stationary cells. Similarly, all three factors caused changes in morphologic features of cells, leading to increased area, flatness, and polarity. Increases in area and flatness but not polarity were slightly greater with HGF than with hSF or mSF. These results suggest that SFs and HGF have similar effects on motility and morphology of isolated epithelial cells.
Key words: scatter factor; hepatocyte growth factor; cell motility; cell morphology. Preparation of SFs and HGF. Mouse SF (mSF) was purified from serum-free conditioned medium from ras-transformed mouse N1H/2 3T3 cells (clone D4) (Rosen et al., 1990b). Human placental SF was purified as described (Rosen et al., 1990c; Bhargava et al., 1991). Recombinant human HGF was purified from culture supernatants from C127 mouse fibroblasts transfected with a complete human HGF cDNA (Nakamura et al., 1989). Scatter assay. Scatter activity was quantitated by the MDCK serial dilution scatter assays as described previously (Stoker et al., 1987; Rosen et al., 1990b). In this assay, small colonies of MDCK cells are incubated for 20 h with serial dilutions of test sample in 96-well plates, stained, and examined for scatter effect (i.e. spreading of normally tight cohesive colonies, dispension of colonies into individual cells, and conversion to fibroblastic appearance). Scatter activity at the limiting dilution is defined as 0.5 MDCK scatter units per milliliter, allowing calculation of the activities in the original undiluted sample. Preparation of cellsfor tracking. Single cell suspensions were prepared from subconfluent cultures of MDCK cells and 2000 to 5000 cells in DMEM + 10% FBS were seeded into 25-cm 2 culture flasks (Nunclon 152094, capacity 40 ml). Cells were incubated for 4 h at 37 ~ C to allow attachment. Flasks were then filled completely with fresh medium with or without factors, capped tightly, and incubated for another 2 h. Flasks were then placed in a temperature-controlled enclosure in the cell-tracking system. Cell tracking. A multipurpose automated microscope system, Dynamic Microscope Image Processing Scanner (DMIPS), developed by the Cancer Imaging Section, B.C. Cancer Research Center, Vancouver, British Columbia (Thurston et al., 1986, 1988) was used to investigate cell motility and morphology. This system consists of an optical scanner, a high precision computer-controlled microscope stage, a fast image processing unit, a host computer, an incubator with temperature control (+0.1" C), automatic data acquisition, and an image recognition software package. Before placing the flask on the stage, the incubator housing stage was warmed to 37 ~ C for 30 min. The X-Y controller was used to move the stage manually to select individual cells to be tracked. Selected cells were positioned near the microscope's cross hair. For better recognition results, only round and flattened cells, which had an approximate area of 100 gm 2
INTRODUCTION Scatter factors (SFs) are hetcrodimcric glycoproteins, produced by fibroblasts and smooth muscle cells which disperse (scatter) cohesive colonies of epithelium and stimulate cell motility (Stoker et al., 1987; Gherardi et al., ]989; Rosen et al., 1989a, 1990b,e). SFtreated cells show membrane ruffling, and protrude and retract conspicuous cytoplasmic processes (Warn and Dowrick, 1989; Dowrick and Warn, 1991). SF-induced motility may be mediated in part by new protein synthesis, alterations in protein phosphorylation, and eytoskeletal reorganization (Rosen et al., 1990a). To better understand the action of SF it is important to determine its effects on individual cells. We utihzed a computerized cell tracking system (Thurston et al., 1986, ]988) to quantitate the motility and morphologic characteristics of large numbers of isolated Madin-Darby canine kidney (MDCK) epithelial cells. In view of considerable sequence homology between SFs and hepatocyte growth factors (HGFs) (Gherardi and Stoker, 1990; Rosen et al., 1990b; Weidner et al., 1990) we also studied the effect of HGF on the motility and morphology of MDCK cells. MATERIALSAND METHODS
Cell culture and reagents. MDCK epithelial cells were obtained from American Type Culture Collection, Rockville, MD. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), 100 U. m1-1 penicillin, 100 gg. ml-l streptomycin, and 0.1 mM nonessential amino acids as described previously (Rosen et al., 1990b). Cells were incubated at 37 ~ C in a humidified atmosphere of 5% CO2:95% air.
l To whom correspondence should be addressed.
364
365
SCATTER FACTOR INDUCED CELL MOTILITY ~.
0.8-
peak on the monitor. The position of the cell and its morphologic and photometrical characteristics were recorded in the computer memory. During the system's automatic revisit program, 18 features were used for cell detection and discrimination (Poon et al., 1987). If a cell was not detected in the search area by a window of 96 • 96 #m 2 centered at the previous recorded cell position, a larger area (200 • 200#m 2) was searched. If the cell was still not detected, it was considered as "lost" and not revisited. Sixty cells per experimental condition were tracked every 10 min for 8 h. Two independent experiments were performed, and the final plots and statistical evaluation were based on data for 120 cells from each treatment group. At the end of the experiment, the quantitative "path" of the cell movement (walk pattern) was determined. The dynamic morphology of moving cells was extracted and quantitated at different stages of cell movement (see below). Classification of cells into motile and stationary groups. Cells were classified as "moving" or "stationary" if their average velocity was greater or less, respectively, than the global average velocity of all cells tracked. Determination of morphologic features. Morphologic changes were quantitated by measuring the area, flatness, and circularity of each cell at 10-min intervals during the 8-h tracking period. These values were then averaged for the 120 cells in each treatment group. The area of a cell in the field (96 X 96 t t m 2 window) was determined at the threshold of 8 grey level units above background intensity. Cell flatness is a measure of the roughness of surface. To determine the flatness at each point of the perimeter the sobel edge detection values were compared to the input threshold, to provide a sum of pixels having a high gradient (F = A/V% where F is cell flatness, A is area in pixels, and V is volume). A higher value (measured in percentage) represents a more flattened cell. Circularity is defined as perim-
Z
:~ tn 7
0.7-
o
0.6.
-~
0.5.
n,'-
0.4. 0
o _J
0.3.
w
0.2
,,,
