- Email: [email protected]
Invasion of human embryonic fibroblast cell aggregates by rat tumor cells of different metastatic capacities
Cell Biology
International
Reports,
Vol. IO, No. 5, May 1986
375
Invasion of Human Embryonic Fibroblast Cell Aggregates by Rat Tumor Cells of Different Metastatic Capacities. Institute
Hans-Otto Werling, Eberhard Spiess, and Neidhard Paweletz of Cell and Tumor Biology, German Cancer Research Center Im Neuenheimer Feld 280, D-6900 Heidelberg, F.R.G. Abstract
Aggregates of normal human Wi38 cells are used as a three dimensional substrate to test in vitro the behavior of rat tumor cells which exhibit different invasive and metastatic capabilities in vivo. The invasive but non metastatic tumor cells colonize the Wi38 cell aggregates, invade and destroy them within three days. The non invasive but highly metastatic tumor cells settle in a limited number on the aggregates but show no further activites. Co-cultivation of these tumor cells with cell suspension of single Wi38 cells under aggregation conditions does not hinder the Wi38 cells in forming aggregates. The results show that the invasive process in the metastatic cascade needs more specific reaction partners and host environment than local tumor growth. The conditions of the first process cannot be mimicked by a simple model. Introduction Invasiveness and metastasis are the main threats of tumor cells toward their hosts. It is therefore a need to investigate tumor cells in respect to these properties. In vitro model systems are means to characterize tumor cells for their invasive abilities. (Gershman 1982, Jones 1982, Kramer and Nicolson 1982, Mareel 1983, Poste 1982, Vlodavsky et al. 1983). Matzku et al. (1983) have established a rat tumor cell model with clones of different metastatic and invasive behavior. One, BSp73 AS 17-4 forms large invasively expanding local tumors, but hardly metastasizes. The other, BSp73ASML 14-1, forms only small local tumors, but metastasizes via the lymphatic pathway into the lung. On this way it forms extended tumors in the lymph nodes. In a series of experiments we have tested the behavior of these two tumor cell lines (Aulenbacher et al. 1984, Paweletz et al. 1984, Paku et al. 1986a, Paweletz et al. 1986). Cell aggregates are a convenient three dimensional system to test tumor cells for their activities against normal cells. Expeiences from earlier investigations have shown that in vitro even 0309-1651/86/050375-08/$03.00/O
@ 1986 Academic
Press Inc. (London)
Ltd.
376
Cell Biology
International
Reports,
Vol. 10, No. 5, May 1986
the xenogenic character of such models does not disturb certain tumor cells in their invasive activities (Benke and Paweletz 1984, Benke et al. 1984, Gerstberger and Paweletz 1982, Kramer and Nicolson 1979, Mareel 1982, Niedbala et al. 1985). We used normal human Wi38 lung fibrolast cells, which readily form aggregates (Gerstberger and Paweletz 1981), as a substrate to investigate the behavior of the described tumor cells.
Materials
and Methods
Cell culture: Human fibrolasts: Wi38 human lung fibrolast cells (ATCC-CCL75) were cultivated in Basal Medium Eagle (BME) supplemented with Lglutamine and 10 % fetal calf serum. The cell5aggregates were produced by rotation mediated culture of 5x10 cells per 5 ml culture medium in siliconized Erlgnmeyer flasks (25 ml) on a gyratory shaker at 70 rpm and 37 C. Rat cells: Two variants of the rat tumor BSp73 (Matzku et al. 1983) were used: AS (BSp73 AS clone 17-4), and ASML (BSp73 ASML clone 14-1). Cell cultivation for propagation and cell aggregate formation was performed as described previously (Werling et al.
1985). Electron Microscopy: Specimens for scanning electron scribed earlier (Werling et al.
microscopy
were processed
as de-
1985).
Results Wi38 Cell Aggregation: As found earlier (Gerstberger and Paweletz 1981) Wi38 cells have a high ability to form aggregates. In these former experiments the rotation rate was 90 rpm; under this condition the aggregate diameters increased rapidly in the initial phase reaching the maximal aggregate diameters at about 18 hours. In the new experiments, the rate of rotation was slower, 70 rpm, and hence, shearing forces had a lower impact on the cells. This leads to a steady increase of aggregate diameters (Fig. 1); after 72 to 96 hours also larger aggregates fused, forming finally a macroaggregate of considerable size (Fig. 2). In this aggregate the Wi38 cells were arranged parallel to each over wide areas in layers of different orientation. Such an aggregate is an appropriate substrate for the colonization by other cells. Confrontation of Wi38 Aggregates with AS Rat Tumor Cells: Single cell suspensions of AS cells were co-cultivated with three days old Wi38 aggregates maintaining the aggregate forming conditions. The AS cells developed both, stable homotypic con-
Cell Biology
International
Reports,
Vol. 10, No. 5, May 1986
377
378
Cell Biology
International
Reports,
Vol. 70, No. 5, May 1986
tacts between each other and heterotypic contacts to the Wi38 cell aggregate already after 30 min (Fig. 3). They settled particularly in furrows and surface cavities of the Wi38 aggregate, probably due to the lower exposure to shearing forces in these regions. From these areas of attachment the AS cells grew out over the surface layer but also undermining it. These activities resulted in a destruction of the macroaggregate which decayed into numerous small aggregates and subsequently in a complete cover of these small aggregates by the AS cells. After 72 hours only remnants of Wi38 cells were visible on the surface (Figs. 4, 5). Thus, the AS tumor cells exhibited an invasive and destructive activity, which within a short time, led to the disappearance of the host cells in this in vitro system. Confrontation of Wi38 Cell Aggregates with ASML Tumor Cells: The ASML cells were confronted with Wi38 cell aggregates under the same conditions as the AS cells. Most of these tumor cells were then only involved in homotypic contacts which led to the typical ASML cell clusters (Werling et al. 1985). Heterotypic contacts to the Wi38 cells were also found in furrows and cavities (Fig. 6). In contrast to the AS cells, however, the ASML cells remained in these places of initial settlement inactive in and lateral and vertical outgrowth respect to proliferation, (Fig. 7, 8). A destruction of the preformed Wi38 cell aggregates was not observed. To prove a possible influence of ASML cells on the aggregation process of the Wi38 cells we co-cultivated single cell suspensions of both cell types under aggregate forming conditions. Also in this case we found no irritations neither on the Wi38 cells nor on the process of their aggregation. Figures 1, 2: Scanning electron micrographs of Wi38 cell aggretheir average diameter is about gates. 1) 24 hours aggregates; 150 urn. 2) 96 hours macroaggregate; wide areas are covered by cells orientated in parallel. Figures 3 to 5: Scanning electron micrographs showing the confrontation of Wi38 cell aggregates with AS rat tumor cells. 3) Colonization of the aggregate by the tumor cells after 30 min fo co-cultivation. 4) Surface of an aggregate after 72 hrs of co-cultivation.The Wi38 cells have disappeared from the surface. 5) Detail from Fig. 4 showing the remnants of Wi38 cells between the tumor cells. Figures 6 to 8: Scanning electron micrographs showing the confrontation of Wi38 cell aggregates and ASML cell clusters after 2 hours of cultivation. 7) After 72 hours of co-cultivation the widest parts of the aggregate are still free from tumor cells and the Wi38 cells appear undisturbed. 8) Detail from Fig. 7) showing the co-existence of normal human and rat tumor cells.
Cell Biology
international
Reports,
Vol. 10, No. 5, May 1986
379
Discussion In the simple test system described here, the AS tumor cells exhibited a behavior which is in accordance with their behavior in vivo: expansive and destructive action towards the normal host cells. Such a behavior was also observed in the other in vitro test system (Aulenbacher et al. 1984, Paku et al. 1986a, Paweletz et al. 1986) but, not to such an extent. In similar studies Wi38 cells were likewise attacked by HeLa tumor cells (Benke and Paweletz 1984, Benke et al. 1984). In contrast to the AS cells the ASML cells were inactive. This was also observed in other test systems where cell monolayers (Benke et al. 1984), or parts of blood or lymph vessel or diaphragma material was used for confrontation (Aulenbacher et al. 1984, Paku et al. 1986a). However, lung material (excised lung cubes) (Paweletz et al. 1986) provided a substrate against which these ASML cells showed penetrative and infiltrative activities. These studies also revealed a certain selectivity of binding of the ASML cells: they were found attached to a distinct type of cells on the surface of the lung cubes and from these attachment points gained access to the underlying basal lamina. Selectivity for organs and cells was also shown by murine B16 melanoma cells (McGuire et al. 1984, Nicolson et al. 1985). Detailed studies of the metastasizing process in vivo (Paku et al. 1986b) have shown these tumor cells penetrating the basal lamina of blood capillaries in the lung and lymph vessels from both directions. These in vivo and in vitro experiments with lung material give no clear information which parts in these processes are played by the tumor cells and which by the host, or in other words, the dependence of the tumor cells from the host and vice versa. An in vitro phagokinetic assay has clearly proven the ability othe ASML cells to recognize a distinct type of lung cells, and to attach firmly to these cells (Werling et al. 1986). The observed penetration of lung tissue in vitro could be ostensible and in fact be the consequence of the extensive growth of the lung cells. In vivo, the penetration of blood and lymph vessels and the translocation in the tissue might be facilitated by means of the host, e. g. tissue and blood vessel pressure but this is unlikely to be the only cause. Evidence has been presented that tumor cells produce proteases which lyse components of the basement membrane and hence could enable its perforation (Liotta et al 1980, Nicolson 1984, Thorgeiersson et al. 1985, Turpeenniemi-Hujanen et al. 1985). Such proteases have also been found in ASML cells, even more pronounced than in the AS cells (Koppel et al. 1984). Effects of such proteases have not yet been recognized in our in vitro experiments. This may have several reasons: suppression of enzyme production, inactivation of the enzyme, inappropriate reaction substrates, and/or insufficient detection
380
Cell Biology
international
Reports,
Vol. 10, No. 5, May 1986
systems. But altogether we have indirect evidence to propose that the ASML cells also have an active role in the penetration process: at least they select the appropriate site at which they can weaken or perforate the basal lamina by secreted proteases; the subsequent translocation may then be provided by the host. All proposed parameters are obviously absent in the model with the Wi38 aggregates. Thus, our experiments demonstrate a limited use of the model. Aggressivity towards normal cells and invasion into such tissue, as it is carried out by AS cells can be proven easily. The penetration of the basal lamina and the translocation through tissue as the ASML cells practice so successfully for metastasis in vivo cannot be shown. This sequence of events obviously needs very specific conditions and cellular reaction partners which we are not yet able to mimic in a model. References Aulenbacher, P., Werling, H.-O., Paweletz, N., Spiess, E. (1984) Invasive activities of metastasizing and non metastaszing tumor cell variants in vitro. Anticancer Res. 4, 75-82. Benke, R., Paweletz, N. (1984) Scanning electron microscopic observations on cells grown in vitro. VII. HeLa cells can 33, 52-54. penetrate Wi38 fibroblasts. Eur. J. Cell Biol. Paweletz, N. (1984) Studies on conBenke, R., Wering, H.-O., frontations of tumor cells with human diploid fibroblasts. Anticancer Res. 4, 241-246. Gershman, H. (1982) Three dimensional models for the study of invasion and metastasis. In: L.A. Liotta and I.R. Hart (eds.) Tumor Invasion and Metastasis, pp. 231-250. Martinus Nijhoff Publ., The Hague microsGerstberger, R., Paweletz, N. (1981) Scanning electron copic observations on cells grown in vitro. VI. Aggregate formation in confrontation cultures of human diploid and tumor cells. Eur. J. Cell Biol. 26, 136-143. Jones, P.A. (1982) In vitro assay of invasion using endothelial and smooth muscle cells. In: L.A. Liotta and I.R. Hart (eds.) Tumor Invasion and Metastasis, pp. 252-265. Martinus Nijhoff Publ., The Hague. R. (1984) Kdppel, P., Baici, A., Keist, R., Matzku, S. and Keller, Cathepsin-B-like proteinase as a marker for metastatic tumor cell variants. Exptl. Cell Biol. 52, 293-299. G.L. (1979) Interactions of tumor cells Kramer, R.H., Nicolson, with vascular endothelial cell monolayers: a model for metastatic invasion. Proc. Natl. Acad. Sci. (USA) 76, 5704-5708. Liotta, L., Tryggvason, K,. Garbisa, S., Hart I., Foltz, C., Shafie, S. (1980) Metastatic potential correlates with enzymatic degradation of basement membrane collagen. Nature, 284, 67-68. Mareel, M.M. (1982) The use of embryo organ cultures to study
Cell Biology
International
invasion
in vitro.
Reports,
In:
Invasion and Metastasis, The Hague. Mareel, M.M. (1983) Invasion
Vol. IO, No. 5, May 1986
381
L.A. Liotta and I.R. Hart (eds.) Tumor pp. 207-230. Martinus Nijhoff Publ.
in vitro. Methods of analysis. Cancer metastasis Rev. 2, 201-218. Mareel, M.M., Bruyneel, E.A., Dragnonetti, C.H., DeBruyne, G.K., Van Cauwenberge, R.M.-L., Smeets, L.A., Van Rooy, H. (1984) Effect of temperature on invasion of MO -mouse fibrosarcoma cells in organ culture. Clin. Expl. Met%stasis 2, 107-125. Matzku, S. Komitowski, D., Miltenberger, M., Zoller, M. (1983) Characterization of BSp73, a spontaneous rat tumor and its in vivo selected variants showing different metastasizing capacities. Invasion Metastasis 3, 109-123. McGuire, E.J., Mascali, I.J., Grady, S.R., Nicolson, G.L. (1984) Involvement of cell-cell adhesion molecules in liver colonization by metastatic murine lymphoma/lymphosarcoma variants. Clin. Expl. Metastasis 2, 213-222. Nicolson. G.L. (1984) Cell surface molecules and tumor metastasis Exptl. Ce 11 Res: 150, 3-22. K., Basson, C., Welch, D.R. (1985) PreNicolson, G.L , Dulski, ferentia 1 organ attachment and invasion in vitro by B16 melanoma cells selected for differing metastatic colonization and invasive properties. Invasion Metastasis 5, 144158. Niedbala, M.J . , Crickard, K., Bernacki, R.J. (19851 Interactions of human ovarian tumor cells with-human mesothe 1 ial cells grown on extracellular matrix. Expl. Cell Res. 160, 499-513. Paku, S., Werling,.H.-O., Aulenbacher, P., Paweletz, N Spiess E. (1986a): Invasive activities of metastaszing ani non metaII. Stud es on contasizing tumor cell variants in vitro. frontations with aorta, vein, ductus thoracicus , diaphragm, and lung. Anticancer Res. In Press. Paku, S., Paweletz, N., Spiess, E., Aulenbacher, P., Werling, H.-O. (1986b) Ultrastructural analysis of invasion in the rat lung by tumor cells metastasizing lymphatically. Submitted for publication. Paweletz, N., Werling, H.-O., Aulenbacher, P., Spiess, E. (1984) Morpholigical and behavioral characteristics of two rat tumor cell lines with different metastatic capacities. Scanning Electron Microscopy 1984/11, 783-792. Paweletz, N., Paku, S., Werling, H.-O., Spiess, E. (1986) Experimental approaches to problems of invasion and metastasis. Anticancer Res. In press. Poste, G. (1982) Methods and models for studying tumor invasion. In: L.A. Liotta and I.R. Hart (eds.) Tumor Invasion and Metastasis, pp. 147-171. Marinus Nijhoff Publ., The Hague. Thorgeiersson, U.P., Turpeenniemi-Hujanen, T., Liotta, L.A. (1985) Cancer cells, components of basement membranes, and proteolytic enzymes. Int. Rev. Expl. Pathol. 27, 203-234. Turpeenniemi-Hujanen, T. Thorgeiersson, U.P., Hart, I.R., Grant S.S. and Liotta, L.A. (1985) Expression of collagenase IV
Cell Biology
International
Reports,
Vol. 70, No. 5, May 1986
(basement membrane collagenase) activity in murine tumor cell hybrids that differ in metastatic potential. J. Natl. Cancer Inst. 75, 99-103. I., Schirrmacher, Vl ,odavsky, V., Ariav, Y., Fuks, Z. (1983) Lymphoma cell interaction with cultured vascular endothelial cells and with the subendothelial basal lamina: attachment invasion and morphological appearance. Invasion Metastasis 3, 81-97. H.-O., Spiess, E., Aulenbacher, P. and Paweletz, N. Wer ling, (1985) Adhesion and spreading characterization of a rat tumor cell system exhibiting different metastazising behavior.
Invasion Metastasis 5, 270-294.
N. (1986) In vitro H.-O., Spiess, E., Paweletz, Wer ling, of BSp73 rat tumor cells with different metastatic ties. Submitted for publication.
Received:
21.1.86
Accepted:
3.3.66
motility abili-
International
Reports,
Vol. IO, No. 5, May 1986
375
Invasion of Human Embryonic Fibroblast Cell Aggregates by Rat Tumor Cells of Different Metastatic Capacities. Institute
Hans-Otto Werling, Eberhard Spiess, and Neidhard Paweletz of Cell and Tumor Biology, German Cancer Research Center Im Neuenheimer Feld 280, D-6900 Heidelberg, F.R.G. Abstract
Aggregates of normal human Wi38 cells are used as a three dimensional substrate to test in vitro the behavior of rat tumor cells which exhibit different invasive and metastatic capabilities in vivo. The invasive but non metastatic tumor cells colonize the Wi38 cell aggregates, invade and destroy them within three days. The non invasive but highly metastatic tumor cells settle in a limited number on the aggregates but show no further activites. Co-cultivation of these tumor cells with cell suspension of single Wi38 cells under aggregation conditions does not hinder the Wi38 cells in forming aggregates. The results show that the invasive process in the metastatic cascade needs more specific reaction partners and host environment than local tumor growth. The conditions of the first process cannot be mimicked by a simple model. Introduction Invasiveness and metastasis are the main threats of tumor cells toward their hosts. It is therefore a need to investigate tumor cells in respect to these properties. In vitro model systems are means to characterize tumor cells for their invasive abilities. (Gershman 1982, Jones 1982, Kramer and Nicolson 1982, Mareel 1983, Poste 1982, Vlodavsky et al. 1983). Matzku et al. (1983) have established a rat tumor cell model with clones of different metastatic and invasive behavior. One, BSp73 AS 17-4 forms large invasively expanding local tumors, but hardly metastasizes. The other, BSp73ASML 14-1, forms only small local tumors, but metastasizes via the lymphatic pathway into the lung. On this way it forms extended tumors in the lymph nodes. In a series of experiments we have tested the behavior of these two tumor cell lines (Aulenbacher et al. 1984, Paweletz et al. 1984, Paku et al. 1986a, Paweletz et al. 1986). Cell aggregates are a convenient three dimensional system to test tumor cells for their activities against normal cells. Expeiences from earlier investigations have shown that in vitro even 0309-1651/86/050375-08/$03.00/O
@ 1986 Academic
Press Inc. (London)
Ltd.
376
Cell Biology
International
Reports,
Vol. 10, No. 5, May 1986
the xenogenic character of such models does not disturb certain tumor cells in their invasive activities (Benke and Paweletz 1984, Benke et al. 1984, Gerstberger and Paweletz 1982, Kramer and Nicolson 1979, Mareel 1982, Niedbala et al. 1985). We used normal human Wi38 lung fibrolast cells, which readily form aggregates (Gerstberger and Paweletz 1981), as a substrate to investigate the behavior of the described tumor cells.
Materials
and Methods
Cell culture: Human fibrolasts: Wi38 human lung fibrolast cells (ATCC-CCL75) were cultivated in Basal Medium Eagle (BME) supplemented with Lglutamine and 10 % fetal calf serum. The cell5aggregates were produced by rotation mediated culture of 5x10 cells per 5 ml culture medium in siliconized Erlgnmeyer flasks (25 ml) on a gyratory shaker at 70 rpm and 37 C. Rat cells: Two variants of the rat tumor BSp73 (Matzku et al. 1983) were used: AS (BSp73 AS clone 17-4), and ASML (BSp73 ASML clone 14-1). Cell cultivation for propagation and cell aggregate formation was performed as described previously (Werling et al.
1985). Electron Microscopy: Specimens for scanning electron scribed earlier (Werling et al.
microscopy
were processed
as de-
1985).
Results Wi38 Cell Aggregation: As found earlier (Gerstberger and Paweletz 1981) Wi38 cells have a high ability to form aggregates. In these former experiments the rotation rate was 90 rpm; under this condition the aggregate diameters increased rapidly in the initial phase reaching the maximal aggregate diameters at about 18 hours. In the new experiments, the rate of rotation was slower, 70 rpm, and hence, shearing forces had a lower impact on the cells. This leads to a steady increase of aggregate diameters (Fig. 1); after 72 to 96 hours also larger aggregates fused, forming finally a macroaggregate of considerable size (Fig. 2). In this aggregate the Wi38 cells were arranged parallel to each over wide areas in layers of different orientation. Such an aggregate is an appropriate substrate for the colonization by other cells. Confrontation of Wi38 Aggregates with AS Rat Tumor Cells: Single cell suspensions of AS cells were co-cultivated with three days old Wi38 aggregates maintaining the aggregate forming conditions. The AS cells developed both, stable homotypic con-
Cell Biology
International
Reports,
Vol. 10, No. 5, May 1986
377
378
Cell Biology
International
Reports,
Vol. 70, No. 5, May 1986
tacts between each other and heterotypic contacts to the Wi38 cell aggregate already after 30 min (Fig. 3). They settled particularly in furrows and surface cavities of the Wi38 aggregate, probably due to the lower exposure to shearing forces in these regions. From these areas of attachment the AS cells grew out over the surface layer but also undermining it. These activities resulted in a destruction of the macroaggregate which decayed into numerous small aggregates and subsequently in a complete cover of these small aggregates by the AS cells. After 72 hours only remnants of Wi38 cells were visible on the surface (Figs. 4, 5). Thus, the AS tumor cells exhibited an invasive and destructive activity, which within a short time, led to the disappearance of the host cells in this in vitro system. Confrontation of Wi38 Cell Aggregates with ASML Tumor Cells: The ASML cells were confronted with Wi38 cell aggregates under the same conditions as the AS cells. Most of these tumor cells were then only involved in homotypic contacts which led to the typical ASML cell clusters (Werling et al. 1985). Heterotypic contacts to the Wi38 cells were also found in furrows and cavities (Fig. 6). In contrast to the AS cells, however, the ASML cells remained in these places of initial settlement inactive in and lateral and vertical outgrowth respect to proliferation, (Fig. 7, 8). A destruction of the preformed Wi38 cell aggregates was not observed. To prove a possible influence of ASML cells on the aggregation process of the Wi38 cells we co-cultivated single cell suspensions of both cell types under aggregate forming conditions. Also in this case we found no irritations neither on the Wi38 cells nor on the process of their aggregation. Figures 1, 2: Scanning electron micrographs of Wi38 cell aggretheir average diameter is about gates. 1) 24 hours aggregates; 150 urn. 2) 96 hours macroaggregate; wide areas are covered by cells orientated in parallel. Figures 3 to 5: Scanning electron micrographs showing the confrontation of Wi38 cell aggregates with AS rat tumor cells. 3) Colonization of the aggregate by the tumor cells after 30 min fo co-cultivation. 4) Surface of an aggregate after 72 hrs of co-cultivation.The Wi38 cells have disappeared from the surface. 5) Detail from Fig. 4 showing the remnants of Wi38 cells between the tumor cells. Figures 6 to 8: Scanning electron micrographs showing the confrontation of Wi38 cell aggregates and ASML cell clusters after 2 hours of cultivation. 7) After 72 hours of co-cultivation the widest parts of the aggregate are still free from tumor cells and the Wi38 cells appear undisturbed. 8) Detail from Fig. 7) showing the co-existence of normal human and rat tumor cells.
Cell Biology
international
Reports,
Vol. 10, No. 5, May 1986
379
Discussion In the simple test system described here, the AS tumor cells exhibited a behavior which is in accordance with their behavior in vivo: expansive and destructive action towards the normal host cells. Such a behavior was also observed in the other in vitro test system (Aulenbacher et al. 1984, Paku et al. 1986a, Paweletz et al. 1986) but, not to such an extent. In similar studies Wi38 cells were likewise attacked by HeLa tumor cells (Benke and Paweletz 1984, Benke et al. 1984). In contrast to the AS cells the ASML cells were inactive. This was also observed in other test systems where cell monolayers (Benke et al. 1984), or parts of blood or lymph vessel or diaphragma material was used for confrontation (Aulenbacher et al. 1984, Paku et al. 1986a). However, lung material (excised lung cubes) (Paweletz et al. 1986) provided a substrate against which these ASML cells showed penetrative and infiltrative activities. These studies also revealed a certain selectivity of binding of the ASML cells: they were found attached to a distinct type of cells on the surface of the lung cubes and from these attachment points gained access to the underlying basal lamina. Selectivity for organs and cells was also shown by murine B16 melanoma cells (McGuire et al. 1984, Nicolson et al. 1985). Detailed studies of the metastasizing process in vivo (Paku et al. 1986b) have shown these tumor cells penetrating the basal lamina of blood capillaries in the lung and lymph vessels from both directions. These in vivo and in vitro experiments with lung material give no clear information which parts in these processes are played by the tumor cells and which by the host, or in other words, the dependence of the tumor cells from the host and vice versa. An in vitro phagokinetic assay has clearly proven the ability othe ASML cells to recognize a distinct type of lung cells, and to attach firmly to these cells (Werling et al. 1986). The observed penetration of lung tissue in vitro could be ostensible and in fact be the consequence of the extensive growth of the lung cells. In vivo, the penetration of blood and lymph vessels and the translocation in the tissue might be facilitated by means of the host, e. g. tissue and blood vessel pressure but this is unlikely to be the only cause. Evidence has been presented that tumor cells produce proteases which lyse components of the basement membrane and hence could enable its perforation (Liotta et al 1980, Nicolson 1984, Thorgeiersson et al. 1985, Turpeenniemi-Hujanen et al. 1985). Such proteases have also been found in ASML cells, even more pronounced than in the AS cells (Koppel et al. 1984). Effects of such proteases have not yet been recognized in our in vitro experiments. This may have several reasons: suppression of enzyme production, inactivation of the enzyme, inappropriate reaction substrates, and/or insufficient detection
380
Cell Biology
international
Reports,
Vol. 10, No. 5, May 1986
systems. But altogether we have indirect evidence to propose that the ASML cells also have an active role in the penetration process: at least they select the appropriate site at which they can weaken or perforate the basal lamina by secreted proteases; the subsequent translocation may then be provided by the host. All proposed parameters are obviously absent in the model with the Wi38 aggregates. Thus, our experiments demonstrate a limited use of the model. Aggressivity towards normal cells and invasion into such tissue, as it is carried out by AS cells can be proven easily. The penetration of the basal lamina and the translocation through tissue as the ASML cells practice so successfully for metastasis in vivo cannot be shown. This sequence of events obviously needs very specific conditions and cellular reaction partners which we are not yet able to mimic in a model. References Aulenbacher, P., Werling, H.-O., Paweletz, N., Spiess, E. (1984) Invasive activities of metastasizing and non metastaszing tumor cell variants in vitro. Anticancer Res. 4, 75-82. Benke, R., Paweletz, N. (1984) Scanning electron microscopic observations on cells grown in vitro. VII. HeLa cells can 33, 52-54. penetrate Wi38 fibroblasts. Eur. J. Cell Biol. Paweletz, N. (1984) Studies on conBenke, R., Wering, H.-O., frontations of tumor cells with human diploid fibroblasts. Anticancer Res. 4, 241-246. Gershman, H. (1982) Three dimensional models for the study of invasion and metastasis. In: L.A. Liotta and I.R. Hart (eds.) Tumor Invasion and Metastasis, pp. 231-250. Martinus Nijhoff Publ., The Hague microsGerstberger, R., Paweletz, N. (1981) Scanning electron copic observations on cells grown in vitro. VI. Aggregate formation in confrontation cultures of human diploid and tumor cells. Eur. J. Cell Biol. 26, 136-143. Jones, P.A. (1982) In vitro assay of invasion using endothelial and smooth muscle cells. In: L.A. Liotta and I.R. Hart (eds.) Tumor Invasion and Metastasis, pp. 252-265. Martinus Nijhoff Publ., The Hague. R. (1984) Kdppel, P., Baici, A., Keist, R., Matzku, S. and Keller, Cathepsin-B-like proteinase as a marker for metastatic tumor cell variants. Exptl. Cell Biol. 52, 293-299. G.L. (1979) Interactions of tumor cells Kramer, R.H., Nicolson, with vascular endothelial cell monolayers: a model for metastatic invasion. Proc. Natl. Acad. Sci. (USA) 76, 5704-5708. Liotta, L., Tryggvason, K,. Garbisa, S., Hart I., Foltz, C., Shafie, S. (1980) Metastatic potential correlates with enzymatic degradation of basement membrane collagen. Nature, 284, 67-68. Mareel, M.M. (1982) The use of embryo organ cultures to study
Cell Biology
International
invasion
in vitro.
Reports,
In:
Invasion and Metastasis, The Hague. Mareel, M.M. (1983) Invasion
Vol. IO, No. 5, May 1986
381
L.A. Liotta and I.R. Hart (eds.) Tumor pp. 207-230. Martinus Nijhoff Publ.
in vitro. Methods of analysis. Cancer metastasis Rev. 2, 201-218. Mareel, M.M., Bruyneel, E.A., Dragnonetti, C.H., DeBruyne, G.K., Van Cauwenberge, R.M.-L., Smeets, L.A., Van Rooy, H. (1984) Effect of temperature on invasion of MO -mouse fibrosarcoma cells in organ culture. Clin. Expl. Met%stasis 2, 107-125. Matzku, S. Komitowski, D., Miltenberger, M., Zoller, M. (1983) Characterization of BSp73, a spontaneous rat tumor and its in vivo selected variants showing different metastasizing capacities. Invasion Metastasis 3, 109-123. McGuire, E.J., Mascali, I.J., Grady, S.R., Nicolson, G.L. (1984) Involvement of cell-cell adhesion molecules in liver colonization by metastatic murine lymphoma/lymphosarcoma variants. Clin. Expl. Metastasis 2, 213-222. Nicolson. G.L. (1984) Cell surface molecules and tumor metastasis Exptl. Ce 11 Res: 150, 3-22. K., Basson, C., Welch, D.R. (1985) PreNicolson, G.L , Dulski, ferentia 1 organ attachment and invasion in vitro by B16 melanoma cells selected for differing metastatic colonization and invasive properties. Invasion Metastasis 5, 144158. Niedbala, M.J . , Crickard, K., Bernacki, R.J. (19851 Interactions of human ovarian tumor cells with-human mesothe 1 ial cells grown on extracellular matrix. Expl. Cell Res. 160, 499-513. Paku, S., Werling,.H.-O., Aulenbacher, P., Paweletz, N Spiess E. (1986a): Invasive activities of metastaszing ani non metaII. Stud es on contasizing tumor cell variants in vitro. frontations with aorta, vein, ductus thoracicus , diaphragm, and lung. Anticancer Res. In Press. Paku, S., Paweletz, N., Spiess, E., Aulenbacher, P., Werling, H.-O. (1986b) Ultrastructural analysis of invasion in the rat lung by tumor cells metastasizing lymphatically. Submitted for publication. Paweletz, N., Werling, H.-O., Aulenbacher, P., Spiess, E. (1984) Morpholigical and behavioral characteristics of two rat tumor cell lines with different metastatic capacities. Scanning Electron Microscopy 1984/11, 783-792. Paweletz, N., Paku, S., Werling, H.-O., Spiess, E. (1986) Experimental approaches to problems of invasion and metastasis. Anticancer Res. In press. Poste, G. (1982) Methods and models for studying tumor invasion. In: L.A. Liotta and I.R. Hart (eds.) Tumor Invasion and Metastasis, pp. 147-171. Marinus Nijhoff Publ., The Hague. Thorgeiersson, U.P., Turpeenniemi-Hujanen, T., Liotta, L.A. (1985) Cancer cells, components of basement membranes, and proteolytic enzymes. Int. Rev. Expl. Pathol. 27, 203-234. Turpeenniemi-Hujanen, T. Thorgeiersson, U.P., Hart, I.R., Grant S.S. and Liotta, L.A. (1985) Expression of collagenase IV
Cell Biology
International
Reports,
Vol. 70, No. 5, May 1986
(basement membrane collagenase) activity in murine tumor cell hybrids that differ in metastatic potential. J. Natl. Cancer Inst. 75, 99-103. I., Schirrmacher, Vl ,odavsky, V., Ariav, Y., Fuks, Z. (1983) Lymphoma cell interaction with cultured vascular endothelial cells and with the subendothelial basal lamina: attachment invasion and morphological appearance. Invasion Metastasis 3, 81-97. H.-O., Spiess, E., Aulenbacher, P. and Paweletz, N. Wer ling, (1985) Adhesion and spreading characterization of a rat tumor cell system exhibiting different metastazising behavior.
Invasion Metastasis 5, 270-294.
N. (1986) In vitro H.-O., Spiess, E., Paweletz, Wer ling, of BSp73 rat tumor cells with different metastatic ties. Submitted for publication.
Received:
21.1.86
Accepted:
3.3.66
motility abili-