Ultrastructural Studies of Surface Features of Human Normal and Tumor Cells in Tissue Culture by Scanning and Transmission Electron Microscopy 1,2 Matthew A. Gonda,3 Stuart A. Aaronson/ Nelson
Ellmore,~
SUMMARY-Human tumors of a variety of histopathologic types have been established in tissue culture. The sUrface features of these cell lines were investigated by scanning electron microscopy (SEM) with the use of new techniques for specimen preparation. Tumor cells demonstrated striking degrees of surface activity with numerous microvilli, filopodia, blebs, and ruffles. Intercellular contacts were also prominent in cultures of most solid tumors observed by SEM. At low cell density, normal human fibroblasts exhibited some surface features such as microvilli and blebs, but at higher cell density they lacked extensive surface modifications. By transmission electron microscopy (TEM), the cytoskeleton of normal fibroblasts was :shown to be well organized, with parallel orientation of microfilaments, filaments, and microtubules. These structures were also in tumor cells, but they lacked the degree oJ organization of fibroblasts. Desmosomes were readily demonstrated in normal fibroblasts and carcinoma cells in culture but not in sarcomas, melanomas, or tumors of neural origin. These studies have provided the first correlative SEM and TEM analyses of solid human tumor cells of diverse pathologic types in vitro.-J Natl Cancer Inst 56: 245-263, 1976.
Scanning electron microscopy (SEM) has become a valuable tool in studies of cell organization (1-7), cell-tocell interactions (1~ 8~ 9), and the quantitation of biologic entities (10). To date, there are few reports in which transformed and normal cells have been analyzed by SElYI (6~ 11 ~ 12). In particular, studies of tumor cells of human origin have been possible only with the .relatively recent development of tissue-culture techniques which isolate solid tumor cell lines from a variety of histopathologic types (13-16). In the present study, we report observations by SEM and transmission electron microscopy (TEM) of the ultrastructural surface features of human normal and tumor cells grown in vitro. MATERIALS AND METHODS
Cells.-Humall fibroblast strain 5011' was obtained as a punch biopsy from a normal adult, and strain M413 was derived from human embryo lung. The fibroblasts had normal karyotypes, exhibited a finite life-span, and were studied within the first 15 cell generations in culture. Designation of the pathologic type of each human tumor cell line was based on histopathologic diagnosis of the original tumor and, in most cases, also on the histopathology of tumors formed after the cell line was injected into immunosuppressed mice (14). Each tumor cell possessed other properties characteristic of malignant cells, including abnormal chromosome pattern [(14); Nelson-Rees W, Stulberg C: Personal communication] and grew as established lines in culture (>250 generations). The tumor cells were studied after their establishment as permanent lines. Malignant cell lines included epidermoid carcinoma A388, lung carcinoma A549, malignant melanomas A375 and A875, rhabdomyosarcoma A204, and glioblastoma A172. We derived lines Al186 and Al207 from a rhabdomyosarcoma and a
Victor H. Zeve,4 and Kunio Nagashima 3,
·6
glioblastoma, respectively, using techniques similar to those described in (14). Culture methods.-Cells were grown in Dulbecco's modification of Eagle's medium supplemented with 10% calf serum at 37° C in a humidified 10% CO 2 atmosphere. For electron microscopy (EM), cells were grown in 60-mm petri dishes (Falcon Plastics, Los Angeles, Calif.) containing carbon-coated 5-mm square glass cover slips. Cells were seeded at densities of approximately 4 X 10 4 cells/dish. Cultures were selected for fixation in late exponential growth; the medium was routinely changed 48 hours before fixation. vVe observed at least eight independent cultures of each cell line and found consistent results over a 12-month observation. EM.-Cells selected for EM were fixed at 37° C for 30 minutes in 2% glutaraldehyde in 0.05 X Dulbecco's phosphate-buffered saline and 0.5 M sodium cacodylate, pH 7.2. After glutaraldehyde fixation, the cultures were rinsed several times with 0.1 M cacodylate buffer, pH 7.2, and then post fixed for 5 minutes with 1'% osmium in 0.2 M cacodylate buffer. The osmium was quickly replaced by cacodylate buffer, and the cover slips were placed in a critical point-drying (CPD) specimen holder submerged in cacodylate buffer. We carefully avoided air drying. The remaining fixed cells in petri dishes were refrigerated at 10° C for later dehydration and in situ embedding for TElY!. Cells for SElYI were rapidly dehydrated in graded ethanols with three final changes of 100% ethanol. Using a Bomar SPC-EX-900 CPD device (Bomar Co., Tacoma, Wash.), we preserved specimens against air-drying artifacts by CPD with Freon, according to the method of Cohen et al. (17). After CPD, the specimens were attached to I3-mm aluminum stubs with silver conducting paint and lightly carbon coated in a vacuum evaporator. The stubs were then removed, placed in a diode sputtering device, Conductavac I (Seevac Corp., :Monroeville, Pa.) , and coated at 175-,u pressure, 75 rnA at a distance of 4.5 em from the cathode source for 6 minutes with 75-100 A goldpalladium. We have found that vacuum evaporation of conductive metals onto the surface of biologic specimens caused peeling and cracking, due to the high temperatures involved in this process [(1); unpublished results]. By sputtering conductive coatings onto the surface of biologic specimens, we minimized heat damage; here temperatures at the specimen surface rarely reached 42° C, the temperature used for CPD. Furthermore, sputReceived May I, 1975; accepted September 5, 1975. Supported by Public Health Service contracts i\'01-C02.H23 from the Office of the Director of the National Cancer Institute (NCI), and NCI-E-73-3212 from the Virus-Cancer Program within the NCI. 3I'rederick Cancer Research Center (Litton Bionetics, Inc.), P.O. Box B, Frederick, Md. 21701. 4 NCI, National Institutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Bethesda, Md. 20014. 5 Hazelton Laboratories, 9200 Leesburg Pike, Vienna, Va. 22180. 6 We acknowledge the excellent assistance of Mrs. Luann DeFife in editing and typing the manuscript. 1
2
JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 56, NO.2, FEBRUARY 1976
245
246
GONDA, AARONSON, ELLMORE, ZEVE, AND NAGASHIMA
tering was more uniform in coating the cell surface, since this technique allowed penetration of small crevices and even the undersides of exposed surfaces; this was not attainable with vacuum evaporation, even with rotary tilt stages. The sputtering technique also provided more consistent thicknesses of conductive coatings, since parameters such as pressure, specimen distance from the metal source, time, and amperage could be set and monitored. DeNee and Walker (18), using sputtering, also noted superior coatings of lung tissue. Specimens were photographed in an Etec Autoscan oPerated at 17 kV and equipped with a 90° goniometer tilt stage. :Micrographs were taken at a 35° tilt, unless otherwise specified. Wi th the tumor cells visualized, variations in density among micrographs were not associated with differences in cell morphology. The SEM micrographs were selected from areas of each culture that most closely represented the culture as a whole. For TEM, cells for in situ embedding in petri dishes were poststained in a sucrose-buffered 0.25% uranyl acetate (VA) solution for 2 hours, dehydrated in graded ethanols, and infiltrated with 100% Epon 812 at room temperature for 2 days, to volatilize the remaining ethanol. The infiltrated specimens were then polymerized at 50° C for 24 hours and at 70° C for 48 hours. Thin sections were cut perpendicular to the growth plane of cells on an LKB Vltrotome III equipped with a diamond knife, mounted on carbon-coated Formvar 300-mesh copp~r grids, and stained with a saturated solution of UA'~nd Reynolds lead citrate (19). Observations and mi
Ellmore,~
SUMMARY-Human tumors of a variety of histopathologic types have been established in tissue culture. The sUrface features of these cell lines were investigated by scanning electron microscopy (SEM) with the use of new techniques for specimen preparation. Tumor cells demonstrated striking degrees of surface activity with numerous microvilli, filopodia, blebs, and ruffles. Intercellular contacts were also prominent in cultures of most solid tumors observed by SEM. At low cell density, normal human fibroblasts exhibited some surface features such as microvilli and blebs, but at higher cell density they lacked extensive surface modifications. By transmission electron microscopy (TEM), the cytoskeleton of normal fibroblasts was :shown to be well organized, with parallel orientation of microfilaments, filaments, and microtubules. These structures were also in tumor cells, but they lacked the degree oJ organization of fibroblasts. Desmosomes were readily demonstrated in normal fibroblasts and carcinoma cells in culture but not in sarcomas, melanomas, or tumors of neural origin. These studies have provided the first correlative SEM and TEM analyses of solid human tumor cells of diverse pathologic types in vitro.-J Natl Cancer Inst 56: 245-263, 1976.
Scanning electron microscopy (SEM) has become a valuable tool in studies of cell organization (1-7), cell-tocell interactions (1~ 8~ 9), and the quantitation of biologic entities (10). To date, there are few reports in which transformed and normal cells have been analyzed by SElYI (6~ 11 ~ 12). In particular, studies of tumor cells of human origin have been possible only with the .relatively recent development of tissue-culture techniques which isolate solid tumor cell lines from a variety of histopathologic types (13-16). In the present study, we report observations by SEM and transmission electron microscopy (TEM) of the ultrastructural surface features of human normal and tumor cells grown in vitro. MATERIALS AND METHODS
Cells.-Humall fibroblast strain 5011' was obtained as a punch biopsy from a normal adult, and strain M413 was derived from human embryo lung. The fibroblasts had normal karyotypes, exhibited a finite life-span, and were studied within the first 15 cell generations in culture. Designation of the pathologic type of each human tumor cell line was based on histopathologic diagnosis of the original tumor and, in most cases, also on the histopathology of tumors formed after the cell line was injected into immunosuppressed mice (14). Each tumor cell possessed other properties characteristic of malignant cells, including abnormal chromosome pattern [(14); Nelson-Rees W, Stulberg C: Personal communication] and grew as established lines in culture (>250 generations). The tumor cells were studied after their establishment as permanent lines. Malignant cell lines included epidermoid carcinoma A388, lung carcinoma A549, malignant melanomas A375 and A875, rhabdomyosarcoma A204, and glioblastoma A172. We derived lines Al186 and Al207 from a rhabdomyosarcoma and a
Victor H. Zeve,4 and Kunio Nagashima 3,
·6
glioblastoma, respectively, using techniques similar to those described in (14). Culture methods.-Cells were grown in Dulbecco's modification of Eagle's medium supplemented with 10% calf serum at 37° C in a humidified 10% CO 2 atmosphere. For electron microscopy (EM), cells were grown in 60-mm petri dishes (Falcon Plastics, Los Angeles, Calif.) containing carbon-coated 5-mm square glass cover slips. Cells were seeded at densities of approximately 4 X 10 4 cells/dish. Cultures were selected for fixation in late exponential growth; the medium was routinely changed 48 hours before fixation. vVe observed at least eight independent cultures of each cell line and found consistent results over a 12-month observation. EM.-Cells selected for EM were fixed at 37° C for 30 minutes in 2% glutaraldehyde in 0.05 X Dulbecco's phosphate-buffered saline and 0.5 M sodium cacodylate, pH 7.2. After glutaraldehyde fixation, the cultures were rinsed several times with 0.1 M cacodylate buffer, pH 7.2, and then post fixed for 5 minutes with 1'% osmium in 0.2 M cacodylate buffer. The osmium was quickly replaced by cacodylate buffer, and the cover slips were placed in a critical point-drying (CPD) specimen holder submerged in cacodylate buffer. We carefully avoided air drying. The remaining fixed cells in petri dishes were refrigerated at 10° C for later dehydration and in situ embedding for TElY!. Cells for SElYI were rapidly dehydrated in graded ethanols with three final changes of 100% ethanol. Using a Bomar SPC-EX-900 CPD device (Bomar Co., Tacoma, Wash.), we preserved specimens against air-drying artifacts by CPD with Freon, according to the method of Cohen et al. (17). After CPD, the specimens were attached to I3-mm aluminum stubs with silver conducting paint and lightly carbon coated in a vacuum evaporator. The stubs were then removed, placed in a diode sputtering device, Conductavac I (Seevac Corp., :Monroeville, Pa.) , and coated at 175-,u pressure, 75 rnA at a distance of 4.5 em from the cathode source for 6 minutes with 75-100 A goldpalladium. We have found that vacuum evaporation of conductive metals onto the surface of biologic specimens caused peeling and cracking, due to the high temperatures involved in this process [(1); unpublished results]. By sputtering conductive coatings onto the surface of biologic specimens, we minimized heat damage; here temperatures at the specimen surface rarely reached 42° C, the temperature used for CPD. Furthermore, sputReceived May I, 1975; accepted September 5, 1975. Supported by Public Health Service contracts i\'01-C02.H23 from the Office of the Director of the National Cancer Institute (NCI), and NCI-E-73-3212 from the Virus-Cancer Program within the NCI. 3I'rederick Cancer Research Center (Litton Bionetics, Inc.), P.O. Box B, Frederick, Md. 21701. 4 NCI, National Institutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Bethesda, Md. 20014. 5 Hazelton Laboratories, 9200 Leesburg Pike, Vienna, Va. 22180. 6 We acknowledge the excellent assistance of Mrs. Luann DeFife in editing and typing the manuscript. 1
2
JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 56, NO.2, FEBRUARY 1976
245
246
GONDA, AARONSON, ELLMORE, ZEVE, AND NAGASHIMA
tering was more uniform in coating the cell surface, since this technique allowed penetration of small crevices and even the undersides of exposed surfaces; this was not attainable with vacuum evaporation, even with rotary tilt stages. The sputtering technique also provided more consistent thicknesses of conductive coatings, since parameters such as pressure, specimen distance from the metal source, time, and amperage could be set and monitored. DeNee and Walker (18), using sputtering, also noted superior coatings of lung tissue. Specimens were photographed in an Etec Autoscan oPerated at 17 kV and equipped with a 90° goniometer tilt stage. :Micrographs were taken at a 35° tilt, unless otherwise specified. Wi th the tumor cells visualized, variations in density among micrographs were not associated with differences in cell morphology. The SEM micrographs were selected from areas of each culture that most closely represented the culture as a whole. For TEM, cells for in situ embedding in petri dishes were poststained in a sucrose-buffered 0.25% uranyl acetate (VA) solution for 2 hours, dehydrated in graded ethanols, and infiltrated with 100% Epon 812 at room temperature for 2 days, to volatilize the remaining ethanol. The infiltrated specimens were then polymerized at 50° C for 24 hours and at 70° C for 48 hours. Thin sections were cut perpendicular to the growth plane of cells on an LKB Vltrotome III equipped with a diamond knife, mounted on carbon-coated Formvar 300-mesh copp~r grids, and stained with a saturated solution of UA'~nd Reynolds lead citrate (19). Observations and mi
