JOURNAL TRANSCRIPT
Oncogene (1997) 15, 3047 ± 3057 1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00
Rho proteins induce metastatic properties in vivo Luis del Peso1,2, RubeÂn HernaÂndez-Alcoceba1, Nieves Embade, Amancio Carnero3, Pilar Esteve, Carmen Paje and Juan Carlos Lacal Instituto de Investigaciones BiomeÂdicas, CSIC, Madrid, Spain
Rho proteins have been implicated in the regulation of multiple signal transduction processes. Some of the members of this family, including the rho gene from Aplysia californica and the human genes (rhoA, rhoB and rac-1), are proto-oncogenes since when properly mutated they can induce cell transformation, and the generated rho-transformed cells are tumorigenic when inoculated into mice. In addition to their tumorigenic activity, there is evidence suggesting that Rho proteins may contribute to the metastatic phenotype. However, all the experiments implicating Rho proteins or Rhoregulating proteins in the induction of metastatic potential are either indirect or have been performed in vitro. In this study we investigated whether cells transformed by rho oncogenes do have metastatic potential in vivo. We present evidence that cells transformed by the Aplysia californica rho gene, when injected directly into the blood stream are able to eciently colonize lungs and secondary organs, consistent with the acquisition of the metastatic potential. Moreover, tumors derived from subcutaneous injections of these rho-transformed cells are also able to metastasize in distant organs, a strong support to the hypothesis that Rho proteins play a role in the metastatic phenotype. Finally, cells transformed by the human oncogenes dbl, vav and ost, three well-known guanine exchange factors for members of the Rho family, or cells transformed by the activated human rac-1 or rhoA genes do also have metastatic potential when injected into the blood stream. These results demonstrate that signaling pathways regulated by Rho proteins play an important role in the acquisition of the metastatic phenotype in vivo. Keywords: Rho proteins; metastasis; dbl; ost; vav
Introduction Small GTP-binding proteins are a large superfamily of proteins implicated in the regulation of diverse biological functions (Lacal and McCormick, 1993). GTP-binding proteins bind guanine nucleotides and have GTPase activity. In the basal state, GTPases are bound to GDP as an inactive complex. After the appropriate stimulus, speci®c exchange factors lead to Correspondence: JC Lacal Current addresses: 2 Department of Pathology, University of Michigan, 1301 Cathering Road, Ann Arbor, MI 48109-0602, USA; 3 Insitute of Child Health, Cancer Biology Unit, 30 Guildford Street, London WC1N 1EH, UK 1 These authors have equally contributed to this study Received 2 June 1997; revised 12 August 1997; accepted 13 August 1997
the release of the GDP molecule and its exchange for GTP. Once in the GTP-bound form they are able to activate downstream molecules or eectors. This molecular switch is disconnected when the intrinsic GTPase activity breaks down the bound GTP into GDP, and the molecule returns to its inactive form. Despite the variety of processes controlled by these proteins they are grouped in three families based on their sequence homology. Ras proteins are implicated in signal transduction processes leading to control of cell growth and dierentiation (Grand and Owen, 1991; Lacal and McCormick, 1993), Rab proteins regulate vesicle trac and secretion (Grand and Owen, 1991; Lacal et al., 1993) and Rho proteins have been implicated in a number of cell functions such as cytoskeleton regulation (Paterson et al., 1990; Ridley and Hall, 1992; Ridley, 1995; Nobes and Hall, 1995), signal transduction (Coso et al., 1995; Hill et al., 1995; Minden et al., 1995; Vojtek and Cooper, 1995; Perona et al., 1997), control of cell proliferation (Perona et al., 1993a,b; Khosravi-far et al., 1995; Lebowitz et al., 1995; Olson et al., 1995; Prendergast et al., 1995; Leeuwen et al., 1995; Qiu et al., 1995a) and apoptosis (JimeÂnez et al., 1995; Esteve et al., 1995). In agreement with a role of Rho proteins in cell proliferation, we have previously reported transforming activity of the rho gene from Aplysia californica (Perona et al., 1993a). Overexpression in NIH3T3 cells of the Aplysia rho gene induces transformation and these transformed cells are tumorigenic when injected in nude mice. These results have been con®rmed recently by expressing activated versions of rhoA or rhoB in NIH3T3 cells (Khosravi-far et al., 1995; Lebowitz et al., 1995; Prendergast et al., 1995). In addition to Aplysia rho and the human rhoA and rhoB, another human member of the Rho family, rac-1 has been shown to be oncogenic (Qiu et al., 1995a; Leeuwen et al., 1995). Moreover oncogenes such as ost, vav, and dbl, isolated from dierent human tumors are exchange factors for Rho proteins, suggesting that these oncogenes induce cellular transformation by activating Rho proteins (Horii et al., 1994; KhosraviFar et al., 1994; Quilliam et al., 1995). Finally, recent reports indicate that transformation mediated by Ras proteins depend upon the presence of Rho proteins (Khosravi-Far et al., 1995; Qiu et al., 1995a,b). Metastasis is the most common complication in human tumors leading to patient death. Oncogenic transformation is the ®rst step for the generation of tumors, but it is not always associated with the metastatic phenotype suggesting that dierent proteins may be implicated in the induction of tumorigenic and metastatic behavior of the tumor cells (Miele et al., 1996). It has been suggested that Rho proteins, in addition to their ability to induce cell transformation may have a role in metastasis. Thus, Rho proteins are
Induction of metastasis by rho genes L del Peso et al
V14-2-5
V14-2-3
TWT-1B
Several oncogenes have been described to induce the tumorigenic phenotype, however only malignant tumors are able to invade neighbour tissues and give rise to metastasis in distant organs. Previously our group has demonstrated that overexpression of the rho gene from Aplysia californica in NIH3T3 ®broblasts induces cell transformation and the tumorigenic phenotype (Perona et al., 1993a). In the present work we have studied the behavior of rho-induced tumors in vivo. We investigated the metastatic properties of some of our previously characterised, transforming cell lines, such as WT15, WT16, V5 and V3 (Perona et al., 1993a). In addition, and in order to avoid problems derived from extensive cell culture of these established cell lines, we have freshly generated new rhotransformed cell lines. V14-2-3 and V14-2-5 are two cell clones of NIH3T3 cells expressing the activated version of the rho gene from Aplysia californica (rho Val14). WT-2-1 clone was generated using the wild type gene. Also, in order to avoid clonality eects that may increase the metastatic properties of the selected cell lines, mass cultures of transfected cells expressing either the wild type rho gene (LP3-WT) or the activated
One of the most frequently used and accepted assays for analysis of metastasis potential is the determination of lung colonization by intravenously injected cells (Chen et al., 1993; Rong et al., 1994). In order to study the metastatic potential of rho genes we injected the selected rho transformed cell lines and mass cultures in the lateral tail vein of nude mice as a way to determine their ability to induce metastasis (experimental metastasis assay). Figure 2 shows the percentage of surviving mice versus days after injection. Injection of control cells, parental NIH3T3 and NIH3T3 cells transfected with the empty vectors and selected with
TV14-2
Generation of cell lines expressing the Aplysia rho gene
Induction of metastatic properties by Rho expression
LP3-V14
Results
version rhoVal14 (LP3-V14) were generated by G418 selection. Finally, tumor-derived cell lines were also established. Nude mice were injected subcutaneously with 16106 cells of LP3-WT or LP3-V14. After 45 (LP3-WT) or 57 (LP3-V14) days, tumors of about 1 cm diameter were excised, and derived cells grown in vitro and selected for G-418 resistant cells. The generated cell lines were designated as TWT1B (derived from LP3-WT) TV14-1 and TV14-2 (derived from LP3-V14). Expression of the Aplysia Rho protein was assessed by Western-blot (Figure 1). The anti-Rho antibody used is a commercial polyclonal serum raised against a 14 amino acids region from the human RhoA protein that is almost identical in the Aplysia californica Rho protein (Santa Cruz). This antibody was found to eciently recognize both the recombinant, E. coliexpressed and puri®ed RhoA and Aplysia Rho proteins (data not shown). In order to support that the high levels of protein expression were speci®cally derived from the exogenous Aplysia rho gene, the levels of mRNA expression were also investigated by Northern blot using the Aplysia gene as a probe. The observed levels of the Aplysia-speci®c mRNA expression were consistent with the levels of protein expression (data not shown). Only those cell lines that expressed the Aplysia Rho protein over basal endogenous levels were selected for further characterization.
LP3-WT
implicated in cell motility which may be important for migration and metastatis (Takaishi et al., 1993, 1994). met, the gene that codes for the receptor of the hepatocyte growth factor (HGF), also known as scatter factor (SF), delivers a signal into the cell that confers metastatic properties in vivo (Rong et al., 1994). At least part of the intracellular signaling of HGFreceptor is mediated by Rho proteins (Takaishi et al., 1994; Nishiyama et al., 1994). Tiam-1, a gene related to the metastatic phenotype, codes for a protein with sequence homology with Rho exchange factors (Habets et al., 1994) and with ex vivo exchange activity for several members of the Rho family including Rac-1, Rho A and Cdc42 (Michiels et al., 1995). Furthermore, rac-transformed cells have invasion properties in vitro (Michiels et al., 1995) and Rho proteins are involved in adhesion of leukocytes activated with chemoattractants through integrins (Laudanna et al., 1996). Finally CD44, a membrane antigen associated with the metastatic phenotype, interacts in the cytoplasm with a protein complex composed by Erzin, Radixin and Moesin (ERM). The translocation of this complex from cytoplasm to the membrane has been reported to be controlled by RhoA (Takaishi et al., 1995). All these results suggest a role of Rho proteins in the control of metastatic behavior. However there is no in vivo evidence available to con®rm this hypothesis. In this study we have taken the advantage of the high transforming ability of the rho gene from Aplysia californica to study the metastatic potential of Rho genes in vivo. The results shown demonstrate that activation of Rho proteins lead to the metastatic phenotype of tumor cells in vivo. We have extended our ®ndings to the human rho genes and demonstrate also the metastatic properties of NIH3T3 cells transformed by the human oncogenes rac-1, dbl, vav and ost.
Control
3048
Rho Aply — Rho A —
Figure 1 Characterization of NIH3T3 cell lines expressing Rho proteins from Aplysia californica. Western-blot of the indicated cell lines using a polyclonal antibody against the Rho protein (see text for details). LP3-WT or LP3-V14 correspond to independent mass cultures isolated from cells transfected respectively with either the wild type rho gene or the rho Val-14 mutant. TV14-2 and TWT-1B are tumor-derived cell lines induced by the injection into nude mice of LP3-V14 or LP3-WT respectively. V14-2-3 and V14-2-5 are two independent clones of NIH3T3 cells transfected with the rho Val-14 mutant. Control correspond to NIH3T3 cells. Localization of the Aplysia Rho and the endogenous RhoA proteins are indicated
Induction of metastasis by rho genes L del Peso et al
G418 (NIH3T3-Neo) had no activity, since at 70 days after inoculation all the mice injected were alive and healthy and had no detectable metastasis. However, the cells expressing the ras or met oncogenes were highly lethal when injected in nude mice as has been previously described (Rong et al., 1994). Three weeks after injection all the v-met injected mice were dead and those injected with ras-transformed cells died after 4 weeks.
Figure 2 Survival curves after lateral tail vein injection of rhotransformed cell lines. Equivalent number of cells (56105) of the indicated cell lines resuspended in 100 ml of serum-free DMEM were injected in the lateral tail vein of male Nu/Nu mice. Legend indicates the injected cell lines which correspond to those showed in Figure 1. Mice from several experiments are represented, a total of nine mice were injected with control cells (NIH3T3 or NIH3T3-Neo, similar results), ®ve mice with the met-transformed cell line, six mice with the ras-transformed cell line, ®ve mice with the V14-2-3 cell line (Aplysia rho Val14-transformed cells), six mice with the V14-2-5 cell line (Aplysia rho Val14-transformed cells), three mice with the TWT1B cell line (LP3-WT tumorderived cell line expressing the Aplysia rho WT gene), and three mice injected with the TV14-2 cell line (LP3-V14 tumor-derived cell line expressing the Aplysia rho Val14 gene)
The cell lines expressing rho can ®t into two groups, one including the V14-2-3 and V14-2-5 cell lines that are as aggressive as those expressing ras oncogenes. A second group include the cell lines TWT-1B and TV142 (described in Figure 1), which showed a less aggressive behavior since mice injected with these cell lines survived for longer times. Further characterisation of the mechanisms involved in the acquisition of the metastatic property of the Rho-expressing cells, will be required to fully understand the basis of the dierence in their aggressiveness. The results from several experiments are summarized in Table 1. Except for experiment #1 mice were monitored daily after injection until they died. Necropsy was practiced for analysis of experimental metastasis. As expected, the cause of death was multiple lung tumors in all cases (Table 1). In experiment #1, one of the mice injected with each cell line was sacri®ced two weeks after injections and the rest four weeks after injection regardless of their status. Where indicated, rho-induced metastatic lung nodules were isolated, cut into slices and cultured in petri dishes. The metastasis-derived cells were then selected for G418 resistance and designated as MCTWT-20, MC-TV14-13, and MC-2-5-20. These newly generated cell lines derived from metastases were analysed for the rho protein expression (Figure 3a) and mRNA expression (Figure 3b). Rho expression was increased in MC-TWT-20, MC-TV14-13 and MC2-5-2 compared with the original cells, LP3-WT, LP3V14 and V14-2-5 respectively. Mice were ®xed in 10% formaldehyde in PBS and microscopic analysis of histological preparations from several tissues was carried out as described under Materials and methods. Pathological analysis revealed that in addition to lung, other organs were also colonized including adipose tissue of the lung hilius and mediastinal lymph nodes. Representative lungs from mice injected with either control or rhotransformed cells are depicted in Figure 4. The
Table 1 Metastasis induced by lateral tail vein injection of rho transformed cells Micea
Metas.b
none rho Val14 rho Val14 v-met none rho WT rho Val14 rho Val14
2 3 2 2 2 3 3 3
7(0/2) +(1/3) +(1/2) +(2/2) 7(0/2) +(3/3) +(2/3) +(3/3)
rho Val14 H-ras Val12 v-met none rho Val14 H-ras Val12 none rho Val14
3 3 3 3 3 3 2 2
+(3/3) +(3/3) +(3/3) 7(0/3) +(3/3) +(2/3) 7(0/2) +(2/2)
Exp.
Cell line
Gene
1
NIH3T3/Neo LP3-V14 V14-2-5 12b LP8-1 TWT1B TV14-2 V14-2-3 V14-2-5 LP8-3 12b Zip/Neo V14-2-5 LP8-3 Zip/Neo V14-2-3
2
3 4
Organc Lung Lung Lung Lung Lung Lung, Lymph N., Vascular, Adipose Lung, Lymph N. Lung, Adipose, Lymph N. Lung, Vascular Lung Lung Lung
L.T.d 0/2 0/3 1/2 0/2 0/2 1/3 1/3 0/3 0/3 0/3 1/3 0/3 0/3 0/3 0/2 1/2
All the NIH3T3 cells used as control (NIH3T3/Neo, LP8-1 and Zip/Neo) were transfected with the corresponding plasmids carrying the Neo-resistance gene, and selected for G418 resistance. a Mice, represents the total number of injected mice in each experiment for the indicated cell line. b Metas., indicated the presence (+) or absence (7) of metastases. Within brackets, the proportion of animals with metastases over the total number of animals injected as indicated. c Organ, describes the organs aected by Macroscopic metastases. In exp. #2 microscopic analysis of several tissues was carried out, macroscopic mestatases were found only in lung. Adipose and Lymph N. indicate adipose tissue of the lung hilios and mediastinal lymph nodes respectively. d L.T.: local tumor, indicates the presence of tumor in the tail base over the anus. Data indicate the mice with a local tumor over the total number of injected mice
3049
Induction of metastasis by rho genes L del Peso et al
3050
MC-2-5-20
MC-TV14-13
MC-TWT-20
V14-2-5
V14-2-3
TWT-1B
TV14-2
LP3V14
LP3WT
Control
a
— 30 Rho Aply — Rho A —
MC-2-5-20
V14-2-5
V14-2-3
TV14-2
LP3V14
Control
b
MC-TV14-13
—22
Rho Aply 28S —
28S —
18S —
Figure 3 Rho expression in metastasis-derived cell lines. (a) Expression of the Aplysia Rho protein by Western-blot in cell lines derived from metastasis and their parental cell lines. The nomenclature is the same as in Figure 1, Control is NIH3T3 cells transfected with the empty vector and selected for G418 resistance, MC-TWT-20 is a cell line derived from a metastasis induced by TWT-1B injection, MC-TV14-13 is a cell line derived from a TV14-2-induced metastasis and MC-2-5-20 is a cell line derived from a metastasis induced by V14-2-5 injection. The antibody used is a polyclonal antibody against a peptide from the RhoA protein as in Figure 1a (see text for details). The localization of the exogenous Aplysia Rho protein and the endogenous RhoA protein are shown (left) as well as the position of the molecular markers (right). (b) Expression of the Aplysia Rho by Northern-blot analysis. Nomenclature as indicated in a. Up: levels of expression of Aplysia rho mRNA. Below: control of total RNA loading
intravenous injection of the rho-transformed cells led to the development of multiple and large metastatic tumors in the lung. Some of the tumors are indicated by arrows in Figure 4a. Macroscopical analysis revealed the replacement of the normal lung structure by invasion of the tumors (Figure 4a). As shown in Figure 4b, microscopic analysis con®rmed the fibrosarcomatous pattern of the invasive tumoral tissue (black arrow) which is typical of tumors derived from
rho-transformed ®broblasts, as previously reported (Perona et al., 1993a). The metastases in organs other than lung were usually microscopic suggesting that the colonies in these organs were derived from cells that ®rst colonized lung and then left this organ to invade others. Probably large lung tumors killed the animals before these secondary metastasis grew further. Supporting this hypothesis we have found tumoral cells in lung blood vessels (Figure 4c) as well as in heart cavities (not shown) in some of the animals analysed. Moreover, cells from these metastatic tumors were successfully selected using G418, further demonstrating its speci®c origin from the injected cells. The white arrow in Figure 4c indicates the normal microscopic structure of the lung, which is being in®ltrated by the tumoral tissue. Finally, a representative lymph node in®ltrated by the same kind of metastatic cells is shown in Figure 4d. The normal lymphatic tissue, pointed out by a white arrow, is almost completely substituted by the sarcomatous cells, pointed out by a black arrow. Rho expression induces metastatic phenotype in the spontaneous metastasis assay The data presented above suggested that Rho expressing cells have metastatic properties. In addition to the metastasis assay described there are others that resemble more accurately all the steps that tumor cells follow in vivo to give rise to metastases. In these alternative assays cells are not delivered directly into the blood stream, but instead they are injected subcutaneously allowing them to grow into a primary tumor. In order to investigate whether Rho expression induces a complete metastatic potential, we injected the cell lines subcutaneously in the foot pad of the mice. When a primary tumor of about 1 cm of diameter grew up, the leg was cut over the knee to eliminate the primary tumors. Then mice were allowed to develop metastases for several weeks. We describe this assay as `spontaneous metastasis assay'. In this assay tumor cells have the chance to leave the primary tumor and give rise to metastases in other organs. The results from several experiments for the analysis of spontaneous metastasis assays are summarized in Table 2. In these experiments, some of the Rho expressing cell lines were highly aggressive with similar time courses for the generation of the primary tumors as those of the rasand met-derived cell lines. In keeping with this highly aggressive behavior, the cell lines V14-2-3 and V14-2-5 killed 80 ± 100% of the injected mice after 10 weeks, while only 50% of the mice injected with the ras- or met-transformed cells died at the end of the experiment, 18 weeks after the injections (Figure 5). No eect was observed with the control cells, NIH3T3 cells, 18 weeks after injections. As in the experimental metastasis experiments, necropsy was carried out to determine the cause of death. It is interesting to note that spleen was the target organ in several of the cases of the mice injected with the rho-transformed cells (Table 2 and Figure 6). Most of the spleen metastases detected showed a multinodular growth pattern with spindle cells intermixing into the spleen sinusoids. Other cases as those of V5, V14-2-5, V14-2-3 and TWT-1B, presented several lung metastases and sometimes lymph node
Induction of metastasis by rho genes L del Peso et al
3051
a
b
c
d
Figure 4 Analysis of metastases induced by rho-transformed cell lines. (a) Photograph of the lungs of injected animals with either control NIH3T3-Neo cells (right) or with the rho-transformed cells (left). Metastatic lung nodules (arrows) can be readily observed in the lungs from the animals injected with the Rho-expressing cells. Microscopic photographs of tumoral cells in®ltrated in the lung (b), blood vessels (c) and lymph node (d) are also shown. No pathological eects were observed in the microscopic analysis of samples from the animals injected with control cells. Black arrows indicate tumoral tissue. White arrows represent normal tissue. See text for details
aectation. In other case (V14-2-3) a metastasis was detected in the calota growing into the subcutaneous tissue and in®ltrating the bone, although no invasion of the adjacent brain was observed. In this assay neither the ras- nor met-transformed cells were able to induce metastasis under the experimental conditions used. Since these cell lines were used only as controls, no further investigation was performed to unveil their properties. Deaths after injection of ras- or mettransformed cells were due to local recidiva of the primary tumors. In addition to those mice injected with ras and met, some of the mice injected with rho expressing cells presented local recidiva after removal of the primary tumor. This local recidiva may be a consequence of an incomplete elimination of the primary tumor. However, it is also possible that they are caused by aggressive tumors invading the surrounding tissues and/or local lymph nodes. In keeping with the last possibility, these local tumors have been shown to be in®ltrating at least in some of the cases studied (data not shown). Cells transformed by human oncogenes involved in Rhosignaling pathways have metastatic properties The above results of the in vivo metastatic properties of rho-transformed cells were obtained using the rho gene from Aplysia californica. The Aplysia gene has been found to have similar properties to the human rho genes in their transforming ability and also in their
ability to induce apoptosis (JimeÂnez et al., 1995; Esteve et al., 1997 submitted). This is most likely due to the high degree of homology among dierent Rho proteins from Aplysia and humans (Chardin, 1993; Perona et al., 1993b). However, it could be that some small sequence dierences may account for distinct biological activities and that human Rho proteins may not have metastatic properties. Thus, we decided to verify that the metastatic phenotype could also be observed in cells transformed with some of the human oncogenes that are involved in Rho-dependent signaling. The human rac-1 gene and three genes whose products are related to the regulation of Rho proteins such as dbl, vav and ost, were tested for their ability to induce metastasis. The dbl, vav and ost oncogenes belong to the family of exchange factors for Rho proteins which contain a Dbl-homology (DH)-region (reviewed by Cerione and Zheng, 1996). DH-containing proteins activate Rho-like GTP-binding proteins by promoting GDP/GTP exchange. They are activated as oncogenes by deletion of their N-terminal regions. These proteins have been reported to act on a wide range of Rho proteins under ex vivo conditions. Thus, Dbl has been reported to function as an exchange factor for RhoA and CDC42 (Hart et al., 1991), and Ost can be an exchange factor for RhoA and CDC42 (Horii et al., 1994). However, they may have a more limited speci®city when tested in whole cells as Vav, which shows selectivity for Rac (Crespo et al., 1996).
Induction of metastasis by rho genes L del Peso et al
3052
Cells transformed by the human rac-1, dbl, vav, and ost oncogenes were generated as described in Materials and methods and elsewhere (Esteve et al., submitted). Mice were injected into the lateral tail vein with 16106 cells and the appearance of any symptoms of disease scored for several weeks. Data were collected for a period of 16 weeks and are summarized in Table 3. Figure 7 shows some examples of the tumors generated by the rac1, dbl, vav, and ost-transformed cell lines both at macroscopic and microscopic levels. The tumoral tissue (black arrows) is shown in the area of in®ltration of normal parenchyma (white arrows). All the mice (n=4) injected with the dbl-transformed cells died within 10 ± 11 weeks of the injections, and all developed macroscopic lung metastasis. One of the animals was further analysed pathologically and the study indicated malignant tumors of ®brohistiocytical origin. Two animals were injected with the ractransformed cells, and both died spontaneously after 11 ± 15 weeks of injections. Macroscopic metastases were readily observed in the lung with both large and small size tumors, similar to those observed with the other genes analysed (Figure 7a). Tumor cells had enlarged and pleomorphic nuclei with a high degree of histological aggresiveness (Figure 7b right). The pathological analysis showed metastatic cells of mesenchymal origin in lung and heart tissues, with a low degree of histological malignancy. However, cytological atypias were observed. Abundant collagen ®bers with swirly appearance were observed in the lung metastases. The subendorcardic in®ltration showed abundant dysplasia and chondroic metaplasia (Figure
7b, middle). Abundant cellular atypias and some necrotic areas were also observed with the generation of tumoral embolization in the mediastinal vein (data
Figure 5 Survival curves after subcutaneous injection of rhotransformed cell lines. Equivalent number of cells (16106 cells) of the indicated cell lines resuspended in 100 ml of serum-free DMEM were injected subcutaneously in the foot pad of male Nu/Nu mice. When a primary tumor of about 1 cm of diameter grew up, the leg was cut over the knee to eliminate the primary tumors. Then mice were left to develop metastasis for the indicated time in days. Legend indicates the injected cell lines which correspond to those showed in Figures 1 and 2. Mice from several experiments are represented, a total of three mice were injected with control cells (NIH3T3-Neo), four mice with the mettransformed cell line, four mice with the ras-transformed cell line (LP8-3), ®ve mice with the Aplysia rho Val14-transformed cells (V14-2-3), seven mice with the Aplysia rho Val14-transformed cells (V14-2-5), and four mice with the tumor-derived TWT-1B cell line, expressing the Aplysia rho WT gene (LP3-WT)
Table 2 Metastasis induced by subcutaneous injection of rho-transformed cells Exp
Cell line
1
Zip/Neo WT7 V5 H-ras Zip/Neo WT15 WT16 V5 V3 V2 H-ras NIH3T3 Zip/Neo TWT-1B WT-2-1 TV14-1 TV14-2 V14-2-3 V14-2-5 H-ras LP8-3A Zip/Neo TWT-1B V14-2-3 V14-2-5 LP8-3 12b
2
3
4
Gene none rho WT rho Val14 H-ras Val12 none rho WT rho WT rho Val14 rho Val14 rho Val14 H-ras Val12 none none rho WT rho WT rho Val14 rho Val14 rho Val14 rho Val14 H-ras Val12 H-ras Val12 none rhoWT rho Val14 rho Val14 H-ras Val12 v-met
Micea 2 2 2 2 2 4 3 3 3 4 3 3 3 3 3 3 3 3 3 3 3 3 4 5 7 4 4
Tumorb
Macroc
0/2 2/2 2/2 2/2 0/2 4/4 3/3 3/3 3/3 0/4 3/3 0/3 0/3 3/3 0/3 3/3 3/3 3/3 3/3 3/3 3/3 0/3 4/4 5/5 7/7 4/4 4/4
7 7 + (1/2) 7 7 7 7 + (1/1) 7 7 7 7 7 7 7 7 7 + (1/3) + (1/3) 7 7 7 (0/3) + (1/4) + (3/4) + (6/7) 7 (0/4) 7 (0/4)
(2) (4.5) (2) (9) (15) (4) (14) (3) (4.5) (6) (5.5) (3.8) (3.1) (1.9) (3) (3.8) (3.8) (2.8) (2.4) (2.3)
Organd
Lung & Lymph N.
Lung & Lumph N.
Spleen Lung
Lung Spleen, Lung, Head Spleen, Ribs, Lymph N.
Rd.e 0/2 0/2 1/2 1/2 0/2 0/4 0/3 0/3 1/3 0/4 1/3 0/3 0/3 2/3 0/3 1/3 0/2 1/3 1/3 1/3 0/3 0/3 0/4 0/5 1/7 2/4 1/4
The NIH3T3 cells used as controls (Zip/Neo) were transfected with the empty pZip/Neo plasmid and selected for G418 resistance. In experiment #3, an additional control was used, the parental NIH3T3 cells without G418 selection. a Mice, are the total number of injected mice. b Tumor, number of mice that developed a tumor in the foot pad over the total number of injected mice. Within brackets, the average weeks for the development of tumors with 1 cm diameter is indicated. c Macro, presence (+) or absence (7) of macroscopic metastasis. Within brackets, number of mice with metastases over the total mice analysed. d Organ, are the organs aected by macroscopic metastasis. Not all the organs listed were found aected in all the animals with metastasis. Lymph N. indicates mediastinal lymph nodes. e RD., indicates a local tumoral recidiva in the injected leg after the tumor was removed. Data indicate the number of mice with tumor recidiva over the total injected mice
Induction of metastasis by rho genes L del Peso et al
latency similar to that referred for rac-transformed cells. Macroscopic metastases were readily observed in the animals injected, and these metastasis were not further analysed microscopically. By contrast, none of the three animals injected with control NIH3T3 cells showed any symptoms of disease at 16 weeks of injection. The animals were sacri®ced and no detectable metastases were observed macroscopically. In an independent experiment, NIH3T3 cells were transfected with the rhoA gene (mutated at 61 codon) under the control of the EF-Tu promoter, and selected for G418 resistance. Two mass cultures, RhoA.1 and RhoA.2 were selected along with the cells transfected with the empty vector. RhoA.1 expressed 5.5-fold the basal levels of the endogenous RhoA protein, while clone RhoA.2 expressed 2.4-fold, compared to the NIH3T3 cells (data not shown). Cells were injected into the tail vein as indicated under Materials and methods and all the animals were sacri®ced 10 weeks after injection regardless of their status. Metastases were found in two out of three of the animals injected with the RhoA.1 clone and one out of three of the animals injected with the RhoA.2 clone, but none in the animals injected with the parental cell line. Also, the metastases were smaller in those scored positive, relative to those observed for the other oncogenes investigated. All these results indicate that all the human genes tested that ®t into the Rho-signaling pathways (dbl, vav, ost, rac-1 and rhoA) induced the acquisition of the metastatic phenotype under in vivo conditions.
not shown). Finally, both vav- and ost-transformed cells were able to induce also metastatic growth in the lungs with the development of large tumors with a a
b
Discussion Cell transformation is not always followed by the ability of the tumor cells to metastasize (Miele et al., 1996). In fact, not all the known oncogenes are able to induce the metastatic phenotype in addition to the tumorigenic phenotype. It has been reported also that some oncogenes are able to induce metastatic properties in some speci®c cell lines but not in others (Davies et al., 1993; Tatsuka et al., 1996).
Figure 6 Spleen metastases induced by rho-transformed cells. (a) Macroscopic photograph of the spleens from mice injected with control cells. NIH3T3-Neo (left) and rho-transformed cells (right). (b) Microscopic photograph of a metastasized spleen from mice injected with the Rho-transformed cells. Black arrows indicate tumoral tissue. White arrows represent normal tissue. See text for details
Table 3 Metastasis induced by lateral tail vein injection of cells transformed by genes related to Rho signaling Cell line
Gene
Micea
Metas.b
Organc
Latency (w)
A Zip/Neo Rac Dbl-pool Ost Vav
none rac1 dbl ost vav
3 2 4 2 2
7 + + + +
(0/3) (2/2) (4/4) (2/2) (1/2)
± Lung Lung Lung Lung
16 (sacrf) 11 ± 15 (spt.) 10 ± 11 (spt.) 14 ± 16 (spt/sacrif) 14 ± 16 (spt/sacrif)
B NIH/pCEFL Rho A.1 Rho A.2
none rho A rho A
3 3 3
7 (0/3) + (2/3) + (1/3)
Lung Lung
10 (sacrf.) 10 (sacrf.) 10 (sacrf.)
The NIH3T3 cells used as controls (Zip/Neo and NIH/pCEFL) were transfected with the empty pZIP/Neo of pCEFL plasmids and selected for G418 resistance. a Mice, represents the total number of injected mice for the indicated cell line. b Metas. indicates the presence (+) or absence (7) of metastasis. Within brackets, the proportion of animals with metastasis over total animal injected is indicated. c Organ, describes the organs aected by Macroscopic metastasis. Latency refers to the time when metastasis was analysed. Scrf. means that the animal was sacri®ced at the indicated time with no symptoms of disease. Spt. means that the animal died either spontaneously or was sacri®ced with evident signs of sickness. Cells transfected with the empty vector of the rhoA expressing vector were selected and inoculated into nude mice as indicated. The animals were sacri®ced 10 weeks after injection. No evidence of sickness was observed in these animals when the experiment was ®nished
3053
Induction of metastasis by rho genes L del Peso et al
3054
a
b.1
b.2
b.3
Figure 7 Metastasis induced by rac-1-, vav-, ost- and dbl-transformed cells. (a) Macroscopic photographs of lungs from mice injected with control cells (NIH3T3-Neo) or cells transformed by the indicated genes (RAC, VAV, DBL,OST). (b) Microscopic photographs from metastatic in®ltration of rac-transformed cells in the lung (left) and heart (middle). (Right) in®ltration of dbltransformed cells in the lung. Black arrows indicate tumoral tissue. White arrows represent normal tissue
Tumor cells must cover several steps to give rise to metastasis in distant organs (Ponta et al., 1994). Previous studies suggested that members of the rho family may induce the metastatic phenotype in in vitro assays (Michiels et al., 1995). We show here that rhotransformed cells are capable of inducing metastasis in two in vivo metastasis assays. The `experimental metastasis assay' is less restrictive than the `spontaneous metastasis assay', since the ®rst steps that tumor cells must follow in vivo to metastasize are overcome by delivering the cells directly into the blood stream. However microscopic studies have revealed that even in the less restrictive assay, rho-transformed cells are able to complete all the steps and give rise to secondary metastasis in organs other than lung. These results were also con®rmed in the spontaneous metastasis assay, where the rho gene analysed completed all the necessary steps for metastasis. Thus, the data presented in this study strongly support a role for Rho proteins in the induction of metastasis. We have used primarily the rho gene from Aplysia californica which is highly homologous to mammalian rho genes, with 89 ± 92% identity and 96 ± 97% homology in the ®rst 180 aminoacids and 33 ± 50% identity and 50 ± 75% homology in the last 12 residues where all small GTPases diverge, including the human RhoA, RhoB and RhoC proteins (Chardin, 1993; Perona et al., 1993b). In fact, these three human Rho proteins (A, B and C) are among themselves 89 ± 92% identical and 94 ± 99% homologous in the ®rst 180
residues and only 33% identical and 42 ± 75% homologous in the last C-terminal 12 residues. Finally, the major putative eector domain of Rho proteins is identical in the Rho protein from Aplysia and that of Rho A, B and C proteins. Despite this homology, Aplysia rho genes are as tumorigenic as the mammalian rhoA or rac-1 genes (Perona et al., 1993a; Khosravi-far et al., 1995; Lebowitz et al., 1995; Prendergast et al., 1995; Qiu et al., 1995). This makes the Aplysia gene appropriate for the in vivo assays. Metastatic activity was observed in cell lines transfected with either the wild type gene and the mutated version. However, the ratio of metastasis observed in the mice injected with the wild type version was four out of 22 (18%) mice injected, while the mutated version gave rise to metastasis in 28 out of 51 (55%) mice injected. This is in agreement with our previous ®ndings that although both wild type and activated mutants of Rho proteins are tumorigenic, the mutated version is more potent (Perona et al., 1993a,b). The signi®cant and expected dierence found for the metastatic activity between wild type and the mutant, suggest the involvement of some of the known eectors for Rho proteins in this eect. We have observed in vivo metastatic properties also with the human rhoA and rac-1-derived cells. Finally, cell lines transformed by human oncogenes known as guanine exchange factors for the Rho proteins, such as dbl (Hart et al., 1991), ost (Horii et al., 1994), and vav (Crespo et al., 1996), were also tumorigenic and
Induction of metastasis by rho genes L del Peso et al
showed metastatic activity under similar conditions. These results are a very strong support for a role of Rho-dependent signaling in the regulation of metastasis in vivo. Rho proteins have been shown to aect a variety of cellular functions such as cell morphology, formation of stress ®bers and focal adhesions, cell motility, membrane ruing, cytokinesis, cell aggregation, smooth-muscle contraction, cytotoxicity and neurite contraction (reviewed in Takai et al., 1995). Changes in cytoskeleton and motility may be important to complete the multi-step process required for the establishment of metastasis (Ponta et al., 1994). In addition, several intracellular target enzymes for Rho proteins have been described, such as phosphatidylinositol 4-phosphate 5-kinase (PIP5K) (Chong et al., 1994), phosphotidylinositol 3-kinase (PI3K) (Zhang et al., 1993), and phospholipase D(PLD) (Bowman et al., 1993; Malcolm et al., 1994; Esteve et al., 1995; Kuribara et al., 1995). PIP5K has been found increased in malignant tumors and several studies suggest a role for PLD in the induction of metastatic properties (Imamura et al., 1993; Pai et al., 1994; Williger et al., 1995). These results suggest that the ability of Rho-transformed cells to induce metastasis is mediated by several intracellular Rho-activated enzymes that may include PLD and/or PIP5K. In addition, cytoskeletal alterations may also play a critical role in the metastatic process. RhoA has been proposed to aect the membrane localization of the complex Ezrin/Radixin/Moesin (ERM), which directly interacts with the CD44 glycoprotein (Takaishi et al., 1995). CD44 is involved in cell adhesion and has been suggested to play a critical role in metastasis in both murine and human tumors (Gunthert et al., 1991; Matsamura Tarin 1992; Wielenga et al., 1993). Furthermore, Rho proteins have been related to spreading of human monocytes (Aepfelbacher et al., 1994) and leukocyte adhesion through integrins (Laudanna et al., 1996), further supporting their putative role in migration and metastasis. However, it is possible that other proteins related to adhesion and cell-to-cell contact, as well as some of the proteases associated to the metastatic phenotype such as collagenases and stromolisine, may be also activated by Rho proteins. This possibility should be explored and deserves further research eorts. The involvement of Rho proteins in the regulation of cell growth and transformation has become clear from dierent approaches. NIH3T3 cells overexpressing the Aplysia californica rho gene are tumorigenic when injected into nude mice (Perona et al., 1993a). RhoB is expressed transiently after stimulation by growth factors and transformation with oncogenes such as v-src and v-fms (Jahner and Hunter, 1991), and Ras-induced transformation requires Rho proteins as downstream elements (Khosravi-Far et al., 1995; Qiu et al., 1995a,b). Rho proteins seem to be involved in the degradation of p27Kip1 (Hirai et al., 1997), a key protein involved in the regulation of the G1 to S transition. Finally, other intracellular enzymes that have been reported to be activated by Rho proteins, including several kinases (Chong et al., 1994; Zhang et al., 1993) and transcription factors (Hill et al., 1995; Perona et al., 1997), could also play an important role in Rho-induced metastasis.
Besides the strong evidence linking signaling cascades regulated by Ras and Rho proteins, Rhoinduced metastasis seems to be an uncoupled signal from Ras-dependent transformation, at least in the system investigated. These results are also in agreement with the recently reported eects on SRE, where Rasand Rho-induced activation of SRE follow independent pathways (Hill et al., 1995). Also, Ras- and Rhoinduced activation of NF-kB follow independent pathways (Perona et al., 1997, and unpublished results). Further work would be needed to determine the eectors of Rho proteins which may be responsible for the induction of the in vivo metastatic phenotype described in this study. Materials and methods Cell culture and reagents Normal NIH3T3 mouse ®broblasts as well as all NIH3T3 derived cell lines were grown in Dulbecco's Modi®ed Eagle's medium (DMEM) supplemented with 10% new born calf serum (Gibco) under standard conditions of temperature (378C) humidity (95%) and carbon dioxide (5%). Some of the rho Aplysia expressing cells were described previously (Perona et al., 1993) and others were freshly generated using an LTR-driven vector (pZIP). Transfections were carried out by the calcium phosphate method (Chen and Okayama, 1987) with minor modi®cations. The ras- and mettransformed cell lines have been described previously (Peso et al., 1996). The cell lines transformed by rac-1, dbl, ost and vav used in this study are mass cultures of NIH3T3 cells transfected with the corresponding activated genes and selected with 1 mg/ml of G418. These cell lines have been fully described elsewhere (Esteve et al., submitted). Western-blot analysis of protein expression Cells were grown under standard conditions until they reached con¯uence. Then cells were washed twice with icecold TD buer (137 mM NaCl, 5 mM KCl, 1 mM Na2HPO4, 20 mM Tris, pH 7.4) and lysed in 300 ml of ice-cold lysis buer (50 mM Tris, pH 7.4, 0.25% NP-40, 0.25% SDS, 150 mM NaCl, 15 mM b-Glycerophosphate, 10 mM Na PPi, 50 mM Na F, 10 mg/ml aprotinin, 1 m M PMSF). Nuclei and detergent-insoluble material was removed by centrifugation at 13 000 r.p.m. in a microfuge for 15 min. The resulting supernatants were assayed for estimation of total cell protein (Bio-Rad) and equal amounts of cell lysates (30 mg) were boiled at 958C for 5 min in SDS ± PAGE sample buer. For Western blot analysis, proteins were electrophoresed onto 10 or 12% SDS ± PAGE gels poured in 20620 cm glass plates. Resolved proteins were transferred to nitro-cellulose and blots were blocked for 1 h in 4% non-fat dried milk in TTBS (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.05% Tween 20). Blots were incubated 1 h with a polyclonal sera raised against 119 ± 132 aminoacids in RhoA protein (Santa Cruz, SC-179) and developed by ECL (Amersham). Northern-blot analysis of mRNA expression Con¯uent dishes were washed twice with ice-cold TD buer (137 mM NaCl, 5 mM KCl, 1 mM Na2HPO4, 20 mM Tris, pH 7.4) and lysed in 1 ml of UltraspecTM (UltraspecTM-II RNA isolation system, Biotecx) and mRNA extracted according with manufacturer instructions. Ten mg of total RNA were denatured by treatment with formaldehyde and formamide and electrophoresed in 1% agarose-formaldehyde gels (2.2% formaldehyde).
3055
Induction of metastasis by rho genes L del Peso et al
3056
RNAs were transferred by capillary action to 0.45 mm nylon membranes (Nytran NY 13N, Schleicher & Shuell) and ®xed to the membrane by u.v. irradiation. RNA loading was controlled by staining on membranes with methylene blue or by ethidium bromide staining of RNAs. Membranes were pre-incubated 15 min in 16PSE (330 mM sodium phosphate pH 7.2, 7% SDS, 1 mM EDTA) prewarmed to 658C. Probes were labeled with g-32P-dCTP (3000 Ci mmol71, Amersham) by random priming (Ready to goTM DNA labeling kit, Pharmacia) using 40 ng of the 570 bp BamHI insert from pZIP-neo-rho V14 (Perona et al., 1993) containing the full-length Aplysia rho Val 14 cDNA. Filters were hybridized for 12 ± 24 h at 658C in 16PSE without formamide, washed at 658C in 0.16PSE, and exposed to X-ray ®lm with an intensifying screen. In vivo metastasis assays The in vivo metastasis assays were carried out as previously described (Chen et al., 1993; Rong et al., 1994). Before injection cells were trypsinized for a short time and Trypsin inactivated by serum. Then cells were pelleted by centrifugation and resuspended in serum-free DMEM at 106 cells in 50 ml (`spontaneous metastasis assay') or at 5 ± 106105 cells in 100 ml (`experimental metastasis assay'). Just after resuspension in serum-free DMEM cells were injected subcutaneously in the foot (`spontaneous metastasis assay') or directly in blood stream (`experimental metastasis assay') by injection into the lateral tail vein. In the spontaneous metastasis assay, injected cells were left to yield a tumor of 1 cm in diameter, then the tumor was
excised by cutting the leg over the knee. After tumor removal mice were followed until they died. All experiments were done in nu/nu male mice and animals treated according to the current laws about animal experimentation. Preparation of samples for pathological analysis Mice were ®rst analysed for macroscopic metastasis and scored as positive or negative. Where indicated, mice were further analysed for the con®rmation and characterization of metastatis. Tissue specimens were ®xed in 10% formalin in phosphate buered solution (PBS) for at least 24 h at room temperature, dehydrated and embedded in paran. Paran sections were cut, mounted on poly-lysine coated glass slides, and deparanized. Hematoxylin and eosin stained sections of dierent organs were evaluated for the presence of metastatic cells. Acknowledgements We greatly appreciate S RamoÂn y Cajal and J Regadera for the pathological analysis of the tumors, and R Perona and C SuaÂrez for the generation of the expression plasmids and some of the ras-and rho-transformed cells used in this study. This work was supported by Grants 93/0293 and 96/ 2136 from Fondo de InvestigacioÂn Sanitaria (FIS) from the Spanish Department of Health, Grant PB94-0009 from DGICYT and Grant AE 00387/95 from the ConsejerõÂ a de EducacioÂn of Comunidad de Madrid. LP is a fellow from DGICYT, RH is a fellow from FundacioÂn RamoÂn Areces and NE is a fellow from Gobierno del Pais Vasco.
References Aepfelbacher M, Vauti F, Weber PC and Glomset JA. (1994). Proc. Natl. Acad. Sci., 91, 4263 ± 4267. Bowman EP, Uhlinger DJ and Lambert JD. (1993). J. Biol. Chem., 268, 21509 ± 21512. Cerione RA and Zheng Y. (1996). Curr. Op. Cell Biol., 8, 216 ± 222. Coso OA, Chiariello M, Yu JC, Teramoto H, Crespo P, Xu N, Miki T and Gutkind S. (1995). Cell, 81, 1137 ± 1146. Crespo P, Bustelo XR, Aaronson DS, Coso OA, LopezBarahona M, Barbacid M and Gutkind S. (1996). Oncogene, 13, 455 ± 460. Chardin P. (1993). In The ras Superfamily of GTPases. Lacal JC and McCormick F. (eds.). CRC Press, Boca Raton, Florida, USA, pp. 203 ± 231. Chen C and Okayama H. (1987). Mol. Cell. Biol., 7, 2745 ± 2752. Chen L, Ashe S, Singhal MC, Galloway DA, HellstroÈm I and HellstroÈm KE. (1993). Proc. Natl. Acad. Sci. USA, 90, 6523 ± 6527. Chong LD, Traynor-Kaplan A, Bokoch GM and Schwartz MA. (1994). Cell, 79, 507 ± 513. Davies BR, Davies MPA, Gibbs FEM, Barraclough R and Rudland PS. (1993). Oncogene, 8, 999 ± 1008. Esteve P, Peso L and Lacal JC. (1995). Oncogene, 11, 2657 ± 2665. Esteve P, Embade N, Perona R, SuaÂrez C, JimeÂnez B, LeoÂn J, del Peso L, Arends M, Bustelo X, Miki T and Lacal JC. (submitted). Grand RJA and Owen D. (1991). Biochem. J., 279, 609 ± 631. Gunthert U, Hofmann M, Rudy W, Raber S, ZoÈ ller M, Haussman I, Matzku S, Wenzel A, Ponta H and Herrlich P. (1991). Cell, 65, 13 ± 24. Habets GM, Scholtes EHM, Zuydgeest D, Kammen RA, Stam JC, Berns A and Collard JG. (1994). Cell, 77, 537 ± 549. Hart MJ, Eva A, Evans T, Aaronson SA and Cerione R. (1991). Nature, 354, 311 ± 314.
Hill CS, Wynne J and Treisman R. (1995). Cell, 81, 1159 ± 1170. Hirai A, Nakamura S, Noguchi Y, Yasuda T, Kitagawa M, Tatsuno L, Oeda T, Tahara K, Terano T, Narumiya S, Kohn LD and Saito Y. (1997). J. Biol. Chem., 272, 13 ± 16. Horii Y, Beeler JF, Sakaguchi K, Tachibana M and Miki T. (1994). EMBO J., 13, 4776 ± 4786. Imamura F, Horai T, Mukai M, Shinkai K, Sawada M and Akedo H. (1993). Biochem. Biophys. Res. Commun., 193, 497 ± 503. JimeÂnez B, Arends M, Esteve P, Perona R, Sanchez-Prieto R, RamoÂn y Cajal S., Wyllie A and Lacal JC. (1995). Oncogene, 10, 811 ± 816. Khosravi-Far R, Chrzanowska-Wodnicka M, Solski PA, Eva A, Burridge K and Der CJ. (1994). Mol. Cell. Biol., 14, 6848 ± 6857. Khosravi-Far R, Solski PA, Clark GJ, Kinch M and Der CJ. (1995). Mol. Cell. Biol., 15, 6443 ± 6453. Kozma R, Ahmed S, Bet A and Lim L. (1995). Mol. Cell. Biol., 15, 1942 ± 1952. Kumagai N, Morii N, Fujisawa K, Nemoto Y and Narumiya S. (1993). J. Biol. Chem., 268, 24535 ± 24538. Kuribara K, Tago T, Yokozeki T, Sasaki T, Takai Y, Morii N, Narumiya S, Katada T and Kanaho Y. (1995). J. Biol. Chem., 270, 25667 ± 25671. Lacal JC and McCormick F. (1993). CRC Press, Florida, USA. Laudanna C, Campbell JJ and Butcher EC. (1996). Science, 271, 981 ± 983. Lebowitz PF, Davide JP and Prendergast GC. (1995). Mol. Cell. Biol., 15, 6613 ± 6622. Leeuwen FN, Kammen RA, Habets GGM and Collard JG. (1995). Oncogene, 11, 2215 ± 2221. Malcolm KC, Ross AH, Qiu RG, Symons M and Exton JH. (1994). J. Biol. Chem., 269, 25951 ± 25954. Matsamura Y and Tarin D. (1992). Lancet, 340, 1053 ± 1057.
Induction of metastasis by rho genes L del Peso et al
Michiels F, Habets GM, Stam JC, Kammer RA and Collard JG. (1995). Nature, 375, 338 ± 340. Miele ME, Robertson G, Lee JH, Coleman A, McGary CT, Fisher PB, Lugo TG and Welch DR. (1996). Mol. Carcinog., 15, 284 ± 299. Minden A, Lin A, Claret FX, Abo A and Karin M. (1995). Cell, 81, 1147 ± 1157. Nishiyama T, Sasaki T, Takaishi K, Kato M, Yaku H, Araki K, Matsumura Y and Takai Y. (1994). Mol. Cell. Biol., 14, 2447 ± 2456. Nobes CD and Hall A. (1995). Cell, 81, 53 ± 62. Olson MF, Ashworth A and Hall A. (1995). Nature, 269, 1270 ± 1272. Pai JK, Frank EA, Blood C and Chu M. (1994). Anti-Cancer Drug Desing., 9, 363 ± 372. Paterson HF, Self AJ, Garrett MD, Just I, Aktories K and Hall A. (1990). J. Cell Biol., 111, 1001 ± 1007. Perona R, Esteve P, JimeÂnez B, Ballestero R and Lacal JC. (1993a). Oncogene, 8, 1285 ± 1292. Perona R, Ballestero RP and Lacal JC. (1993b). The rho gene family. In The ras Superfamily of GTPases. Lacal JC and McCormick F (eds.). CRC Press, Boca Raton, Florida, USA. Perona R, Montaner S, Saniger L, SaÂnchez-PeÂrez I, Bravo R and Lacal JC. (1997). Genes and Development, 15, 463 ± 475. Peso L, Esteve P, HernaÂndez R and Lacal JC. (1996). Biochem. J., 322, 519 ± 528. Prendergast GC, Khosravi-Far R, Solski PA, Kurzawa H, Lebowitz PF and Der CJ. (1995). Oncogene, 10, 2289 ± 2296. Ponta H, Hofman M and Herrlich P. (1994). Eur. J. Cancer, 30, 1995 ± 2001.
Qiu RG, Chen J, Kirn D, McCormick F and Symons M. (1995a). Nature, 374, 457 ± 459. Qiu RG, Chen J, McCormick F and Symons M. (1995b). Proc. Natl. Acad. Sci. USA, 92, 11781 ± 11785. Quilliam LA, Khosravi-Far R, Hu SY and Der CJ. (1995). Bioessays, 17, 395 ± 403. Ridley AJ. (1995). Curr. Opin. Gen. Dev., 5, 24 ± 30. Ridley AJ, Comoglio PM and Hall A. (1995). Mol. Cell Biol., 15, 1110 ± 1122. Ridley AJ and Hall A. (1992). Cell, 70, 389 ± 399. Rong S, Segal S, Anver M, Resau J and Vande Woude G. (1994). Proc. Natl. Acad. Sci. USA, 91, 4731 ± 4735. Takai Y, Sasaki T, Tanaka K and Nakanishi H. (1995). TIBS, 20, 227 ± 231. Takaishi K, Kikuchi A, Kuroda S, Kotani K, Sasaki T and Takai Y. (1993). Mol. Cell. Biol., 13, 72 ± 79. Takaishi K, Sasaki T, Kameyama T, Tsukita S, Tsukita S and Takai Y. (1995). Oncogene, 11, 39 ± 48. Takaishi K, Sasaki T, Kato M, Yamochi W, Kuroda S, Nakamura T, Takeichi M and Takai Y. (1994). Oncogene, 9, 273 ± 279. Tatsuka M, Ota T, Yamagishi N, Yashikara Y, Wada M, Matsuda N, Mitsui H, Seiki M and Odashima S. (1996). Mol. Carcinog., 15, 300 ± 308. Vojtek AB and Cooper JA. (1995). Cell, 82, 527 ± 529. Williger BT, Reich R, Neeman M, Bercovici T and Liscovitch M. (1995). J. Biol. Chem., 270, 29656 ± 29659. Wielenga UJM, Heider KH, Oerhaus GJA, Adolf GR, Vanderberger FM, Ponta H, Herrlich GR and Pals ST. (1993). Cancer Res., 53, 4754 ± 4756. Zhang J, King WG, Dillons S, Hall A, Feig L and Rittenhouse SE. (1993). J. Biol. Chem., 268, 22251 ± 22254.
3057