Path. Res. Pract. 188,657-662 (1992)
Metastatic Capacity and Differentiation in Murine Melanoma Cell Lines A Morphometric Study A. Maiorana, V. Cavallari1, M. C. Maioranal, R. A. Fano, S. Scimone1 , R. Fante and S. Garbisa2 Istituto di Anatomia Patologica, Universita di Modena, Italy; lDipartimento di Patologia Umana, Universita di Messina, Italy; 21stituto di Istologia, Universita di Padova, Italy
SUMMARY A morphometric analysis was carried out on electron micrographs of cells of the Fl (low metastatif capacity) and Fl 0 (high metastatic capacity) variant sublines of the murine B16 melanoma, both in in-vitro cultures and in lung-metastatic nodules developed after the intravenous injection of neoplastic cells in syngeneic C57 black male mice. A group of 28 morphometric parameters was derived to describe quantitatively each neoplastic cell profile. No qualitative difference was observed between the two cell lines. The quantitative expression of subcellular organelles was dissimilar in the two sublines, being consistently characterized, both in in-vitro cultured cells and in lung-metastatic colonies, by a significant decrease in the mean values of parameters related to melanosomes in the high metastatic capacity cell line (Bl6-FlO). Moreover, in in-vitro cultured cells, indices describing heterochromatin masses and cytoplasmic membranous compartments displayed statistically significant differences between the two sublines. In this experimental system, an inverse relationship between metastatic capacity and differentiation is detected, since cells with a more aggressive metastatic behavior exhibit a decreased degree of differentiation.
Introduction A major advance in research into metastasis has come from the isolation of tumor cell subpopulations that display markedly different metastatic behaviour, such as the FI and FlO variant sublines of the murine Bl6 melanoma 7,8. After injection into the venous system of syngeneic mice I2,n,n, these lines give rise, to low (FI) and high (FlO) numbers of metastatic colonies in the lungs. In spite of an accurate characterization of their biochemical properties l2 , antigenic expression 18 and invasive activitylO, experimental studies correlating the morphological features of tumor cell subpopulations with their biological © 1992 by Gustav Fischer Verlag, Stuttgart
properties are lacking. A cytological similarity between the parent Bl6 melanoma and its variant FI subline has been reported 28 . Studies that relate the degree of differentiation to metastatic behaviour have yielded contradictory results, showing that a higher differentiative pattern can be accompanied by either a decrease 6 or an enhancement2 of the metastatic capacity in experimental tumors. This work aimed at establishing whether a different quantitative expression of subcellular structures is present in tumor subpopulations that exhibit different metastatic capacities. The morphometric analysis was carried out on electron micrographs of Bl6-melanoma Fl and FlO 0344-0338/92/0188-0657$3.5 0/0
658 . A. Maiorana et al.
variant sublines, both in in-vitro cultured cells and in lung-metastatic nodules developed after the i. v. injection of neoplastic cells in mice. Material and Methods Cell cultures ilnd experimental metastasis (ormation. The isolation and properties of Bl 6-F1 (low metastatic capacity) and B16-F 10 (high metastatic capacity) variant sublines of the murine B16 melanoma have been described elsewhere 7 • 8 . Both cell lines were grown in plastic flasks in Dulbccco's modified Eagle's minimal essential medium, supplemented with 10 % fetal calf serum without antibiotics at 37 DC in a humidified atmosphere of 5 % CO 2 in air. B16-Fl and B16-FI0 melanoma cells were harvested by 3-4 minute incubation with 2 mM EDT A solution in Ca++ Mg++-free phosphate buffered saline (PBS) and subsequently washed in PBS. Fl and Fl 0 cell suspensions were divided into 2 aliquots: a) for in vitro studies on cell cultures, an aliquot of Fl and FlO cells was centrifuged. The resulting pellets were fixed and routinely processed for ultrastructural studies. Cells were observed with the electron microscope. b) for in vivo studies on experimental lung metastases, the remaining cell suspensions were checked for viability by trypan blue exclusion. Only suspensions containing >95 % viable cells were used. Cell concentration (determined by hemocytometer) was adjusted to 1 x 10" viable cells/ml Hanks' balanced salt solution. Five-hundred microliters of 5 x 10 5 viable cell suspension were inoculated via the lateral tail vein in unanesthetized C57 black male mice (Charles River, Calco, CO, Italy), weighing 16-18 g. B16-Fl cells'were injected in 9 mice and an identical number of animals was treated with B16-F10 cells. Mice were sacrificed 21 days later by cervical dislocation, the chest was opened and cold fixative was instilled endotrachcally. Subsequently, lungs were removed and immersed in cold fixative. The number of melanoma lung metastases, easily recognized by the deep black color, was counted under a dissecting microscope. Tumor nodules were then removed, measured, trimmed in small fragments (less than 0.5 mm 3 ) and stored in cold fixative. Electron microscopy. The fixative used was cold Karnovsky fluid (pH 7.2, 200 mOsm/I). All materials (cell pellets and lung tumor fragments) were left in cold fixative for 24 hours, then routinely processed for electron microscopy (post-fixation in 1 % OS04, dehydration in graded ethanols, embedding in epoxy resin). Thin sections were collected on uncoated 200-mesh copper grids and viewed under the Siemens 102 electron microscope operating at 80 KV. Nucleated neoplastic cell profiles were randomly selected and photographed at 4,000 x and 10,000 x magnification. Melanophages, lymphocytes and neoplastic cell profiles with artefacts or regressive alterations were not photographed. Approximately 6-8 photographs were required to record a single cell profile. Negative 6 x 9 films were printed on 18 x 24 cm photographic paper (final magnification: 12,000 x and 30,000 x). A test specimen for calibration (Balzers Union, Fiirstentum Liechtenstein) with 2,160 crossing lines/mm was included in each photographic session to check the uniformity of the magnification. Morphometry. A computer-assisted semi-automatic method was used. The hardware consisted of a Numonics 2210-1217 graphics tablet (Numonics Corp., PA, USA) equipped with a 4-button cursor and interfaced with a Lemon 386 32-bit personal computer (Belton Electronics, Montelupone, MC, Italy). The software was developed in the laboratory of Morphometry, Dept.
of Human Pathology, University of Messina. Details of the method have been described elsewhere4,5. Control studies to test the sampling program and to assess the consistency and reproducibility of the measurements were carried out, as reported in reference 4 . The program provided measurements of perimeters, areas and linear distances utilizing the matrix of actual coordinates of single points of the profiles under investigation. Electron micrographs were placed on the plane of the digitizer and the following structures were traced with the cursor: cell membrane, nuclear membrane, nucleolus (in the lower magnification photographs), R.E.R., S.E.R., Golgi complex, mitochondria, heterochromatin masses and melanosomes (in the higher magnification photographs). An example is shown in Fig. 1. Measuring options were sequentially selected by the operator. Values of area, perimeter and, in some cases, center of gravity related to the subcellular stucture under examination were shown on the screen and
Table 1. List of morphometric parameters studied 1) 2) 3) 4) 5) 6) 7)
Nuclear Nuclear Nuclear Nuclear Nuclear Nuclear Nuclear
Vv (x 100) Sv Sv/V v
c.r.
relative eccentricity axes ratio mean area (~2)
8) Heterochromatin masses Vv (x 100) 9) Heterochromatin masses mean area (~2) 10) Nucleolar Vv (x 100) 11) Nucleolar relative eccentricity 12) Nucleolar mean area (~2) 13) 14) 15) 16) 17) 18)
Rough endoplasmic reticulum (R.E.R.) Vv (x 100) Rough endoplasmic reticulum (R.E.R.) Sv Smooth endoplasmic reticulum (S.E.R.) Vv (x 100) Smooth endoplasmic reticulum (S.E.R.) Sv Golgi complex Vv (x 100) Golgi complex Sv
19) 20) 21) 22) 23)
Mitochondria Mitochondria Mitochondria Mitochondria Mitochondria
Vv (x 100) Sv Sv/Vv
24) 25) 26) 27)
Melanosomes Melanosomes Melanosomes Melanosomes
Vv (x 100) Sv NA mean area (~2)
c.r.
mean area
(~2)
28) Cell axes ratio Vv = volume density, Sv = surface density, Sv/V v = surfaceto-volume ratio; NA = numerical density; c.r. = contour index. It is a measure of the regularity of the profile, being 3.54 in perfectly regular circles and higher than 3.54 in irregular profiles. It was derived by dividing the perimeter by the square root of the area. - Relative eccentricity = it is a dimension-indepedent indicator of the position of an inscribed profile, when the inscribing one approximates the circular shape. . . . r - absolute eccentricity re Iatlve eccentnClty = -----------'--r
where: r = radius of the inscribing profile, absolute eccentricity = linear distance bet\veen the center of gravity of the inscribing profile and that of the inscribed one.
Metastatic Capacity and Cell Differentiation . 659
IIAQIlrtCATlOIt
OrDAtJlIIi IIOUtJlIE
RD . !>Z4IH 13
1
2
3
~
_ 6.3159111E-lI2 IIOlIIDAIIY 2 . 22J 13'3
21. 5363
of
5
r.
7
8 'I
~ 11 12 13 14 15 16 17 18 1'1 211 21 22
recorded. Measurements derived from each micrograph were independently stored in memory as a matrix of raw numerical variables. Subcellular components of each cell profile were described by the set of variables obtained from the sequential accumulation of partial measurements performed on the group of corresponding low- and high-magnification micrographs. Each cell profile was recorded under a separate record number and each block was identified by the file-name. In pelleted material directly derived from cells cultured in flasks, 35 cell profiles of the B16-Fl1ine and an identical number of cell profiles of the B16-FI0 line were studied. In lung metastases developed in mice: after i.v. injection of the two cell lines, a total of 12 blocks (6 from each cell line ) corresponding to 12 metastatic nodules (each from a separate mouse) and 200 cell
Fig. 1. a) Section of the cytoplasm of a neoplastic melanocyte (electron micrograph: 35,000 X ). b) profiles of mitochondria and endoplasmic reticulum from the same section are shown on a high-resolution screen, after having been traced with the cursor on the graphic tablet. Values of area and boundary length of the subcellular structure under investigation are shown on the right side (top).
profiles (110 of the B16-Fl· cell line and 90 of the B16-Fl 0 line) were studied. From the raw data, a group of 28 morphometric parameters was derived (Table 1). Stereo logic parameters, such as volume density (V v), surface density (Sv) and numerical density (N A), of subcellular structures were calculated according to WeibeW. Four hierarchically related reference spaces were used: cell area (for calculation of nuclear Vy and nuclear Sy), nuclear area (to derive nucleolar indices), sampled nuclear area (for heterochromatin masses Vy) and sampled cytoplasmic area (to obtain parameters describing cytoplasmic organelles). "Sampled" areas were nuclear and/or cytoplasmic regions present on each highmagnification photograph, after exclusion of previously marked zones that overlapped in adjacent fields. The surface-to-volume
660 . A. Maiorana et al. ratio 16 was obtained by dividing the Sv of a subcellular profile by its Vv. The calculation of other parameters, such as contour index]7 and relative eccentricity, is explained in Table 1. Mean values and standard deviations of all parameters were calculated both in: a) in-vitro cultured cells, and b) in-vivo experimental lung metastases. In this last case, individuallyrecorded cell profiles from metastatic nodules of the F1 cell line (110 elements) and lung metastases of the FlO line (90 elements) were pooled and compared. Statistical analysis was carried out utilizing the Student's t test for unpaired data. All morphometric parameters (Table 1) were entered in the statistical calculations. Probability values above 0.05 were considered non significant.
Results
Qualitative Study a) Cell culture studies: Neoplastic cells in pellets were round shaped and showed isolated cytoplasmic protrusions. Nuclei were round, with prominent nucleoli, and the cytoplasm exhibited a well-developed endoplasmic reticulum with numerous mitochondria and melanosomes in different phases of maturation. No qualitative feature was found that allowed one to differentiate in vitro cultured B16-F1 cells from the B16-F10 cells. b) Experimental lung metastases: A total of 29 metastatic colonies was detected in the lungs of 9 mice (mean: 3.22 ± 2.05) intravenously injected with melanoma cells of the B16-F1 line, whereas in 9 mice treated with the B16-F10 line 217 lung metastases (mean: 24.11 ± 15.09) were found. The difference between the two groups is statistically significant (p
Metastatic Capacity and Differentiation in Murine Melanoma Cell Lines A Morphometric Study A. Maiorana, V. Cavallari1, M. C. Maioranal, R. A. Fano, S. Scimone1 , R. Fante and S. Garbisa2 Istituto di Anatomia Patologica, Universita di Modena, Italy; lDipartimento di Patologia Umana, Universita di Messina, Italy; 21stituto di Istologia, Universita di Padova, Italy
SUMMARY A morphometric analysis was carried out on electron micrographs of cells of the Fl (low metastatif capacity) and Fl 0 (high metastatic capacity) variant sublines of the murine B16 melanoma, both in in-vitro cultures and in lung-metastatic nodules developed after the intravenous injection of neoplastic cells in syngeneic C57 black male mice. A group of 28 morphometric parameters was derived to describe quantitatively each neoplastic cell profile. No qualitative difference was observed between the two cell lines. The quantitative expression of subcellular organelles was dissimilar in the two sublines, being consistently characterized, both in in-vitro cultured cells and in lung-metastatic colonies, by a significant decrease in the mean values of parameters related to melanosomes in the high metastatic capacity cell line (Bl6-FlO). Moreover, in in-vitro cultured cells, indices describing heterochromatin masses and cytoplasmic membranous compartments displayed statistically significant differences between the two sublines. In this experimental system, an inverse relationship between metastatic capacity and differentiation is detected, since cells with a more aggressive metastatic behavior exhibit a decreased degree of differentiation.
Introduction A major advance in research into metastasis has come from the isolation of tumor cell subpopulations that display markedly different metastatic behaviour, such as the FI and FlO variant sublines of the murine Bl6 melanoma 7,8. After injection into the venous system of syngeneic mice I2,n,n, these lines give rise, to low (FI) and high (FlO) numbers of metastatic colonies in the lungs. In spite of an accurate characterization of their biochemical properties l2 , antigenic expression 18 and invasive activitylO, experimental studies correlating the morphological features of tumor cell subpopulations with their biological © 1992 by Gustav Fischer Verlag, Stuttgart
properties are lacking. A cytological similarity between the parent Bl6 melanoma and its variant FI subline has been reported 28 . Studies that relate the degree of differentiation to metastatic behaviour have yielded contradictory results, showing that a higher differentiative pattern can be accompanied by either a decrease 6 or an enhancement2 of the metastatic capacity in experimental tumors. This work aimed at establishing whether a different quantitative expression of subcellular structures is present in tumor subpopulations that exhibit different metastatic capacities. The morphometric analysis was carried out on electron micrographs of Bl6-melanoma Fl and FlO 0344-0338/92/0188-0657$3.5 0/0
658 . A. Maiorana et al.
variant sublines, both in in-vitro cultured cells and in lung-metastatic nodules developed after the i. v. injection of neoplastic cells in mice. Material and Methods Cell cultures ilnd experimental metastasis (ormation. The isolation and properties of Bl 6-F1 (low metastatic capacity) and B16-F 10 (high metastatic capacity) variant sublines of the murine B16 melanoma have been described elsewhere 7 • 8 . Both cell lines were grown in plastic flasks in Dulbccco's modified Eagle's minimal essential medium, supplemented with 10 % fetal calf serum without antibiotics at 37 DC in a humidified atmosphere of 5 % CO 2 in air. B16-Fl and B16-FI0 melanoma cells were harvested by 3-4 minute incubation with 2 mM EDT A solution in Ca++ Mg++-free phosphate buffered saline (PBS) and subsequently washed in PBS. Fl and Fl 0 cell suspensions were divided into 2 aliquots: a) for in vitro studies on cell cultures, an aliquot of Fl and FlO cells was centrifuged. The resulting pellets were fixed and routinely processed for ultrastructural studies. Cells were observed with the electron microscope. b) for in vivo studies on experimental lung metastases, the remaining cell suspensions were checked for viability by trypan blue exclusion. Only suspensions containing >95 % viable cells were used. Cell concentration (determined by hemocytometer) was adjusted to 1 x 10" viable cells/ml Hanks' balanced salt solution. Five-hundred microliters of 5 x 10 5 viable cell suspension were inoculated via the lateral tail vein in unanesthetized C57 black male mice (Charles River, Calco, CO, Italy), weighing 16-18 g. B16-Fl cells'were injected in 9 mice and an identical number of animals was treated with B16-F10 cells. Mice were sacrificed 21 days later by cervical dislocation, the chest was opened and cold fixative was instilled endotrachcally. Subsequently, lungs were removed and immersed in cold fixative. The number of melanoma lung metastases, easily recognized by the deep black color, was counted under a dissecting microscope. Tumor nodules were then removed, measured, trimmed in small fragments (less than 0.5 mm 3 ) and stored in cold fixative. Electron microscopy. The fixative used was cold Karnovsky fluid (pH 7.2, 200 mOsm/I). All materials (cell pellets and lung tumor fragments) were left in cold fixative for 24 hours, then routinely processed for electron microscopy (post-fixation in 1 % OS04, dehydration in graded ethanols, embedding in epoxy resin). Thin sections were collected on uncoated 200-mesh copper grids and viewed under the Siemens 102 electron microscope operating at 80 KV. Nucleated neoplastic cell profiles were randomly selected and photographed at 4,000 x and 10,000 x magnification. Melanophages, lymphocytes and neoplastic cell profiles with artefacts or regressive alterations were not photographed. Approximately 6-8 photographs were required to record a single cell profile. Negative 6 x 9 films were printed on 18 x 24 cm photographic paper (final magnification: 12,000 x and 30,000 x). A test specimen for calibration (Balzers Union, Fiirstentum Liechtenstein) with 2,160 crossing lines/mm was included in each photographic session to check the uniformity of the magnification. Morphometry. A computer-assisted semi-automatic method was used. The hardware consisted of a Numonics 2210-1217 graphics tablet (Numonics Corp., PA, USA) equipped with a 4-button cursor and interfaced with a Lemon 386 32-bit personal computer (Belton Electronics, Montelupone, MC, Italy). The software was developed in the laboratory of Morphometry, Dept.
of Human Pathology, University of Messina. Details of the method have been described elsewhere4,5. Control studies to test the sampling program and to assess the consistency and reproducibility of the measurements were carried out, as reported in reference 4 . The program provided measurements of perimeters, areas and linear distances utilizing the matrix of actual coordinates of single points of the profiles under investigation. Electron micrographs were placed on the plane of the digitizer and the following structures were traced with the cursor: cell membrane, nuclear membrane, nucleolus (in the lower magnification photographs), R.E.R., S.E.R., Golgi complex, mitochondria, heterochromatin masses and melanosomes (in the higher magnification photographs). An example is shown in Fig. 1. Measuring options were sequentially selected by the operator. Values of area, perimeter and, in some cases, center of gravity related to the subcellular stucture under examination were shown on the screen and
Table 1. List of morphometric parameters studied 1) 2) 3) 4) 5) 6) 7)
Nuclear Nuclear Nuclear Nuclear Nuclear Nuclear Nuclear
Vv (x 100) Sv Sv/V v
c.r.
relative eccentricity axes ratio mean area (~2)
8) Heterochromatin masses Vv (x 100) 9) Heterochromatin masses mean area (~2) 10) Nucleolar Vv (x 100) 11) Nucleolar relative eccentricity 12) Nucleolar mean area (~2) 13) 14) 15) 16) 17) 18)
Rough endoplasmic reticulum (R.E.R.) Vv (x 100) Rough endoplasmic reticulum (R.E.R.) Sv Smooth endoplasmic reticulum (S.E.R.) Vv (x 100) Smooth endoplasmic reticulum (S.E.R.) Sv Golgi complex Vv (x 100) Golgi complex Sv
19) 20) 21) 22) 23)
Mitochondria Mitochondria Mitochondria Mitochondria Mitochondria
Vv (x 100) Sv Sv/Vv
24) 25) 26) 27)
Melanosomes Melanosomes Melanosomes Melanosomes
Vv (x 100) Sv NA mean area (~2)
c.r.
mean area
(~2)
28) Cell axes ratio Vv = volume density, Sv = surface density, Sv/V v = surfaceto-volume ratio; NA = numerical density; c.r. = contour index. It is a measure of the regularity of the profile, being 3.54 in perfectly regular circles and higher than 3.54 in irregular profiles. It was derived by dividing the perimeter by the square root of the area. - Relative eccentricity = it is a dimension-indepedent indicator of the position of an inscribed profile, when the inscribing one approximates the circular shape. . . . r - absolute eccentricity re Iatlve eccentnClty = -----------'--r
where: r = radius of the inscribing profile, absolute eccentricity = linear distance bet\veen the center of gravity of the inscribing profile and that of the inscribed one.
Metastatic Capacity and Cell Differentiation . 659
IIAQIlrtCATlOIt
OrDAtJlIIi IIOUtJlIE
RD . !>Z4IH 13
1
2
3
~
_ 6.3159111E-lI2 IIOlIIDAIIY 2 . 22J 13'3
21. 5363
of
5
r.
7
8 'I
~ 11 12 13 14 15 16 17 18 1'1 211 21 22
recorded. Measurements derived from each micrograph were independently stored in memory as a matrix of raw numerical variables. Subcellular components of each cell profile were described by the set of variables obtained from the sequential accumulation of partial measurements performed on the group of corresponding low- and high-magnification micrographs. Each cell profile was recorded under a separate record number and each block was identified by the file-name. In pelleted material directly derived from cells cultured in flasks, 35 cell profiles of the B16-Fl1ine and an identical number of cell profiles of the B16-FI0 line were studied. In lung metastases developed in mice: after i.v. injection of the two cell lines, a total of 12 blocks (6 from each cell line ) corresponding to 12 metastatic nodules (each from a separate mouse) and 200 cell
Fig. 1. a) Section of the cytoplasm of a neoplastic melanocyte (electron micrograph: 35,000 X ). b) profiles of mitochondria and endoplasmic reticulum from the same section are shown on a high-resolution screen, after having been traced with the cursor on the graphic tablet. Values of area and boundary length of the subcellular structure under investigation are shown on the right side (top).
profiles (110 of the B16-Fl· cell line and 90 of the B16-Fl 0 line) were studied. From the raw data, a group of 28 morphometric parameters was derived (Table 1). Stereo logic parameters, such as volume density (V v), surface density (Sv) and numerical density (N A), of subcellular structures were calculated according to WeibeW. Four hierarchically related reference spaces were used: cell area (for calculation of nuclear Vy and nuclear Sy), nuclear area (to derive nucleolar indices), sampled nuclear area (for heterochromatin masses Vy) and sampled cytoplasmic area (to obtain parameters describing cytoplasmic organelles). "Sampled" areas were nuclear and/or cytoplasmic regions present on each highmagnification photograph, after exclusion of previously marked zones that overlapped in adjacent fields. The surface-to-volume
660 . A. Maiorana et al. ratio 16 was obtained by dividing the Sv of a subcellular profile by its Vv. The calculation of other parameters, such as contour index]7 and relative eccentricity, is explained in Table 1. Mean values and standard deviations of all parameters were calculated both in: a) in-vitro cultured cells, and b) in-vivo experimental lung metastases. In this last case, individuallyrecorded cell profiles from metastatic nodules of the F1 cell line (110 elements) and lung metastases of the FlO line (90 elements) were pooled and compared. Statistical analysis was carried out utilizing the Student's t test for unpaired data. All morphometric parameters (Table 1) were entered in the statistical calculations. Probability values above 0.05 were considered non significant.
Results
Qualitative Study a) Cell culture studies: Neoplastic cells in pellets were round shaped and showed isolated cytoplasmic protrusions. Nuclei were round, with prominent nucleoli, and the cytoplasm exhibited a well-developed endoplasmic reticulum with numerous mitochondria and melanosomes in different phases of maturation. No qualitative feature was found that allowed one to differentiate in vitro cultured B16-F1 cells from the B16-F10 cells. b) Experimental lung metastases: A total of 29 metastatic colonies was detected in the lungs of 9 mice (mean: 3.22 ± 2.05) intravenously injected with melanoma cells of the B16-F1 line, whereas in 9 mice treated with the B16-F10 line 217 lung metastases (mean: 24.11 ± 15.09) were found. The difference between the two groups is statistically significant (p
