Autotaxin (ATX), an exo-nucleotide pyrophosphatase and phosphodiesterase, was originally isolated as a potent stimulator of tumor cell motility. In order to study whether ATX expression affects motility-dependent processes such as invasion and metastasis, we stably transfected full-length ATX cDNA into two non-expressing cell lines, parental and ras-transformed NIH3T3 (clone7) cells. The effect of ATX secretion on in vitro cell motility was variable. The ras-transformed, ATX-secreting subclones had enhanced motility to ATX as chemoattractant, but there was little difference in the motility responses of NIH3T3 cells transfected with atx, an inactive mutant gene, or empty vector. In MatrigelTM invasion assays, all subclones, which secreted enzymatically active ATX, demonstrated greater spontaneous and ATX-stimulated invasion than appropriate controls. This difference in invasiveness was not caused by differences in gelatinase production, which was constant within each group of transfectants. In vivo studies with athymic nude mice demonstrated that injection of atx-transfected NIH3T3 cells resulted in a weak tumorigenic capacity with few experimental metastases. Combination of ATX expression with ras transformation produced cells with greatly amplified tumorigenesis and metastatic potential compared to ras-transformed controls. Thus, ATX appears to augment cellular characteristics necessary for tumor aggressiveness.
Metastasis is a complex process during which tumor cells move out of the primary site into compatible secondary loci, degrading tissue barriers along their route (Woodhouse et al., 1997). Cellular motility is thought to be crucial for the early stages of invasion when tumor cells move from the primary tumor into local tissue stroma. Similarly, cell motility is a likely component of intravasation and extravasation, which are necessary for initiation of secondary tumors at sites capable of sustaining their growth.
Tumor cell motility, much like cellular proliferation, appears to be aberrantly regulated, leading us to hypothesize that tumor cells produce autocrine factors important to this process (Liotta et al., 1986). We identified and cloned a potent cytokine, autotaxin (ATX), that stimulates in vitro migration of the human melanoma producer cell line, A2058, at sub-nanomolar levels (Murata et al., 1994). DNA sequence analysis determined that ATX was homologous to a family of exo/ecto nucleotide-pyrophosphatase phosphodiesterases (NPPs) that includes the B cell activation marker, PC-1, and the neural differentiation antigen, B10 (Stefan et al., 1999). Like other members of the NPP family, ATX has multiple enzymatic activities (Clair et al., 1997). The phosphodiesterase (PDE) activity appears to be essential for motility stimulation, since a single point mutation in the enzymatic active site at threonine-210 results in a mutant molecule (T210A-ATX) which lacks both functions (Lee et al., 1996a).
Recombinant ATX (rATX) shows potent motility-stimulating activity in several tumor cell lines derived from human breast (MacDonald et al., 1996) and prostate (Mulvaney et al., 1998) carcinomas, as well as neuroblastoma (Kawagoe et al., 1997). The motility-stimulating activity of ATX suggests indirectly that it may be important in tumor invasion and metastasis. In the present study, we sought more direct evidence of such a role for ATX. To accomplish this, we produced stable transfectants from both non-transformed and v-ras-transformed NIH3T3 (clone7) cells with plasmids containng wild-type atx or T210A-ATX mutant cDNA. Clone7 cells were chosen for the transfections because of their low tumorigenic potential (Ju et al., 1991). Our results strongly suggest that ATX enhances invasion and metastatic dissemination of tumors.
Expression of ATX in transfectants
Neither the parental clone7 nor the ras-transformed NIH3T3 cells expressed detectable ATX when their respective total RNAs were tested by RT–PCR amplification (data not shown). ATX cDNA was stably transfected into both, resulting in clones that expressed and secreted ATX. Partially purified conditioned media from NIH3T3 cells which were transfected with atx (3T3-ATX1) or with inactive mutant atx (3T3-T210A), as well as atx-transfected ras-transformed cells (3T3-RAS-ATX1, 2, and 3), all contained a protein recognized by ATX anti-peptide antibody (Figure 1a,b, respectively). In contrast, the empty-vector transfectants, including both the parental NIH3T3 cells (3T3-M) and ras-transformants (3T3-RAS-M1, 2, and 3), did not secrete detectable ATX. The lower panel in Figure 1b depicts a Ras-specific immunoblot of lysates from these same cells and indicates that each transfectant expresses approximately equivalent concentrations of both c-Ras (lower band) and v-Ras (upper band) proteins. Transfectants of the parental NIH3T3 cells were found to synthesize only c-Ras (data not shown).
In all experimental cell lines, PDE activity in conditioned media was assessed utilizing the p-nitrophenyl-TMP colorimetric (Clair et al., 1997) assay (data not shown). These assays confirmed the immunoblot data and established that enzymatically active ATX was expressed by the atx-transfected clones but not by the parental, ras-transformed, or empty vector-transfected cells. The conditioned media from 3T3-T210A cells, which secreted an ATX that was recognizable by the anti-peptide antibody used in immunoblots, demonstrated no PDE activity above background levels, confirming that these cells secreted the inactive mutant form of ATX.
In vitro migratory and invasive properties
The effects of ATX expression on in vitro cellular migration and invasion were assessed using Boyden chamber chemotaxis and MatrigelTM invasion assays, respectively.
In the experiments depicted in Figure 2a, the chemotactic response of 3T3-M cells was compared to both 3T3-ATX1 and 3T3-T210A cells. All three cell lines responded to 0.4–1.5 nM recombinant ATX (rATX) as chemoattractant. Repeated assays demonstrated no significant difference in the motility response of any of the native ATX transfectants (3T3-ATX1, 2, and 3) compared to the inactive mutant ATX transfectant (3T3-T210A). These subclones appeared to be slightly more motile than 3T3-M cells, but they were not significantly different from NIH3T3 parental cells.
We also tested the response of A2058 cells to partially purified conditioned media from each of the transfected NIH3T3 cell lines (data not shown). As expected, only supernatants from the 3T3-ATX transfectants stimulated motility. Media from parental NIH3T3 cells, 3T3-M transfectants, or 3T3-T210A transfectants were identical in stimulating no significant motility response. In order to determine whether the T210A-ATX had negative dominant properties, we mixed the supernatants from 3T3-T210A cells with those from 3T3-ATX cells. The T210A-ATX had no inhibitory effect whatever and appears to be an inactive protein rather than negative dominant.
Though there were no significant differences in the motility responses of the NIH3T3 subclones to ATX stimulation, in vitro invasive properties were significantly different. As shown in Figure 2c, all three of the 3T3-ATX subcloned cell lines were significantly more invasive than either 3T3-M or 3T3-T210A (P<0.001). A portion of this enhanced response was due to increased spontaneous invasion with no ATX placed in the lower well (Figure 2c, stippled area). In addition (Figure 2c, unshaded area), the 3T3-ATX subclones were also significantly more invasive in response to rATX stimulation (P<0.01).
In the experiments depicted in Figure 2b,d, the results of analogous assays are shown for the 3T3-RAS-M and the 3T3-RAS-ATX subclones. The migratory responses of each ATX-secreting subclone to rATX stimulation were significantly greater than the responses of the empty vector-transfected lines (Figure 2b). In vitro invasion assays (Figure 2d) demonstrated that the 3T3-RAS-ATX cells also showed greater spontaneous and ATX-stimulated invasion than the 3T3-RAS-M cells (P<0.001 and P<0.01, respectively).
The augmented invasiveness, for both the 3T3-ATX and 3T3-RAS-ATX subclones, had two components. The non-specific component of spontaneous invasion was independent of attractant (Figure 2c,d, stippled area). All of the atx-transfected subclones demonstrated this elevated background invasiveness. However, the specific component was only augmented by native ATX in the lower chamber. In contrast, the NIH3T3 parental and transfected cells all demonstrated invasion in response to either 50 nM fibronectin or 20 ng/ml platelet-derived growth factor. The atx-transfected cells had no greater specific response to these ligands than the appropriate control cells (data not shown).
The differences seen in the invasive properties of atx-transfectants cannot be accounted for by differences in growth rate since there is no significant difference in growth for any of the clones over the time course of these experiments (data not shown). Likewise, the difference in invasiveness was not due to differences in gelatinase secretion. When equal quantities of protein from cell supernatants were compared utilizing zymograms (Novex), the gelatinase activity was similar within each experimental group of singly or doubly transfected subclones (data not shown). Similarly, supernatants taken from either the upper or lower well of the invasion chambers showed no differences between cell lines in the relative amounts of activated or precursor gelatinase secreted over the course of the assay.
Anchorage independent cell growth
The ability to survive and sustain proliferation without anchorage to a substratum is a hallmark characteristic of the transformed state. The effect of ATX expression on transformation was assessed using a soft agar colony-formation assay. The results are shown in Table 1. Clone7, 3T3-M, 3T3-T210A, and 3T3-ATX1 failed to produce colonies in soft agar. In contrast, all of the ras-transformed clones produced anchorage independent colonies. Among these transformed cells, the 3T3-RAS-ATX clones appeared to form colonies slightly more efficiently than vector-transfected lines (P<0.05 by ANOVA analysis with Tukey's post test). These data indicate that ATX expression is not sufficient to induce transformation, but it may slightly enhance the ability of already transformed cells to thrive.
In vivo tumorigenicity, growth, and metastasis
In order to determine whether the ATX-induced invasive properties would be sustained in an in vivo cancer and metastasis model, the transfected cells were injected into athymic nude mice. Subcutaneous injections were utilized to assess tumorigenicity and in vivo tumor growth rate. The results are shown in Table 2. No localized tumor formation was detected when parental NIH3T3 (clone7), 3T3-M, or 3T3-T210A cells were injected, whether the inocula contained 4×105 or 106 cells. Tumor appearance was followed in these mice for as long as 4 months post injection. However, the 3T3-ATX1 clone produced solid tumors in 8/10 mice and 3T3-ATX3 clone produced a tumor in a single mouse (1/10). In each case, a slow-growing tumor formed at the injection site, becoming detectable 1 month or later post injection. This mass invariably became immobile, suggesting invasion into surrounding skin and muscle layers, and grew steadily until the mouse was sacrificed between 45 days and 84 days post injection.
All subcutaneous injections of the ras-transfected cell lines utilized 2.5×104 cells for each inoculum. The results of a typical experiment can be seen in Figure 3a. Mice injected with 3T3-ATX-RAS subclones developed large tumors in 12 days. These tumors were hyperemic almost as soon as they became detectable and tended to develop areas of central necrosis relatively early. Mice injected with 3T3-RAS-M cells also developed solid tumors, but these tumors tended to become hyperemic later in their course and developed central necrosis more slowly, if at all. The 3T3-RAS-ATX clones all produced tumors more rapidly than any of the 3T3-RAS-M lines (P<0.001).
In order to verify that the autotaxin expression was sustained throughout the course of these experiments, mRNA was extracted from tumors removed from four mice, representing four of the six clones. Northern blot analysis, utilizing ATX-specific oligonucleotides as probes, determined that 3T3-RAS-ATX1 and 3T3-RAS-ATX2 tumors each had strong positive ATX mRNA bands, but, as expected, ATX mRNA was not detected in either 3T3-RAS-M tumor (Figure 3b).
None of the transfected cells produced visually detectable lung or liver metastases after subcutaneous injection. The ras-transfected cell lines developed life-threatening solid tumors so quickly that there may not have been time for metastases to form. The capacity of each clone to form experimental metastases was, therefore, determined by tail vein injections of 105 cells/mouse. As seen in Table 2, neither parental NIH3T3 cells (clone7) nor 3T3-T210A produced any detectable lung colonies. Two out of five mice injected with 3T3-ATX1 cells developed 4–8 colonies in their lungs after 40 days. Experiments with ras-transformed cells were terminated between 12 and 15 days. Figure 3c shows a typical experimental result. The parental ras-transformed cells (3T3-RAS) produced results similar to the three 3T3-RAS-M clones. Among these four cell lines, 16/40 mice developed lung colonies (range: 1 to 16 colonies/mouse). In contrast, all of the 3T3-RAS-ATX mice developed lung colonies (range: 10 to >200 colonies/mouse).
These data indicate that ATX may confer a range of phenotypic changes when transfected into non-transformed cells, including the ability to form slow-growing tumors and to establish sparse metastatic colonies. Though the expression of ATX, by itself, may not induce a transformed phenotype, it appears to contribute greatly to the aggressive and metastatic character of cells that are already transformed.
In this study, full-length, wild-type ATX cDNA was transfected into non-tumorigenic NIH3T3 (clone7) cells and into ras-transformed cells from the same parental line. In vitro studies with the resultant cloned cells indicated that ATX expression and secretion had little effect on the motility responses of the transfected NIH3T3 subclones, though it increased the ATX-specific chemotactic response of atx-transfected ras-transformed cells. However, transfection of full-length native ATX cDNA resulted in greatly augmented invasion through MatrigelTM barriers. Subcutaneous injection of parental NIH3T3 cells transfected with atx alone gave rise to slow-growing, immobile sarcomatous tumors in 30% of athymic nude mice. Though injection of the ras-transformed cells resulted in the formation of solid tumors in all mice, the combination of ras transformation and atx expression induced more rapidly growing tumors. Intravenous injection of parental NIH3T3 cells transfected with atx alone resulted in sparse lung colonies in a few mice 40 days post injection. Approximately half of the mice receiving empty vector-transfected ras-transformed cells developed lung colonies (≤16/mouse) within 2 weeks; however, all mice that received atx-transfected ras-transformed cells developed experimental metastases and several had greater than 200 visible colonies.
Since all cell lines utilized in the present study responded to ATX as chemoattractant, all presumably express the necessary cell surface receptors. Augmentation of ATX expression resulted in significantly increased spontaneous and ATX-specific invasion without affecting specific invasion toward other motility-stimulating ligands. This ligand specificity suggests that ATX secretion may modulate the number or function of its own receptor. This could be a partial explanation for the increased aggressiveness of ATX-secreting cells. Another clue to the ATX mechanism of action may lie in the observation that transfection of cDNA from the inactive ATX mutant (T210A-ATX) had no effect on invasion or metastasis. These data imply that the ATX-induced phenotypic changes require chemotactically and enzymatically active protein. In addition, at least for the atx-transfected and ras-transformed cells, the augmented motility response to ATX stimulation suggests that autocrine stimulation of motility plays a direct role in the observed effect on cell aggressiveness. Unlike the invasion assays, the motility assays revealed no increase in baseline motility. ATX is predominantly chemotactic, with chemokinetic properties as well. Increased secretion of ATX could effectively decrease the concentration gradient, particularly in a short-term assay during which there is insufficient time for complete equilibration. Unexpectedly, the atx-transfected NIH3T3 cells were more invasive without being more chemotactic when stimulated by ATX. There was some indication that the 3T3-ATX cell lines secreted ATX more slowly than the 3T3-RAS-ATX cells. The 3T3-ATX cell supernatants required partial purification and nearly 100-fold concentration in order to detect ATX on immunoblots though unconcentrated and unpurified supernatants from 3T3-RAS-ATX cells were sufficient for detection. Since the chemotaxis assay is significantly shorter than the invasion assay, the lack of motility response by 3T3-ATX subclones could just mean that these cells do not secrete sufficient concentrations of ATX to induce motility during the shorter incubation period.
ATX is a member of the NPP family of ecto-enzymes, which were originally identified as cell surface differentiation antigens (Kindler-Rohrborn et al., 1985; Takahashi et al., 1970) or as a secreted motility factor (Stracke et al., 1992). As a group, these proteins share nearly identical PDE active sites (Stracke et al., 1997) as well as both PDE and nucleotide pyrophosphatase activities. They appear to have multiple physiological roles. In developing mice, increased ATX expression is seen at sites of active myelination (Fuss et al., 1997), as well as sites of precartilage condensation and tooth formation (Bachner et al., 1998). PC-1 expression has been correlated to regulation of inorganic pyrophos-phate production in cartilage and bone (Huang et al., 1998) and to nucleotide salvage by activated T cells (Deterre et al., 1996). Based on the largely circumstantial evidence now available, it appears that the NPPs might be important in differentiation and development and might have a complex relationship to carcinogenesis. Overexpression of B-10 has been noted on the cell surface of ethylnitrosourea-induced rat brain tumors (Kindler-Rohrborn et al., 1994). ATX expression is elevated in human neuroblastoma and stimulates cell motility in a neuroblastoma cell line (Kawagoe et al., 1997). A chemotactic response to ATX has been described for human melanoma (Liotta et al., 1986), breast carcinoma (MacDonald et al., 1996), and prostate carcinoma (Mulvaney et al., 1998) cell lines. However, to date, there have been few in vivo studies to elucidate whether the motility and differentiation effects associated with these proteins could be physiologically relevant to tumor cells. Recently Deissler et al. (1999) reported experiments in which B-10 cDNA was transfected into NIH3T3 and two rat glioma cell lines. Transfectants from the NIH3T3 cell line and one of the glioma lines displayed an altered cellular morphology and increased in vitro invasion into type I collagen gels. We have now shown that ATX expression augments invasiveness even in non-transformed cells, and enhances both tumorigenesis and metastatic potential in transformed cells. The increased invasive response in vitro is ATX-specific, suggesting that the increased tumorigenicity and metastasis observed in vivo are also responses to ATX. This argues strongly for the importance of ATX-directed autocrine signaling in cancer, presenting a novel target for therapeutic intervention.
Materials and methods
The A2058 cell line (Todaro et al., 1980) as well as ras-transformed and parental NIH3T3 clone7 cells were maintained in DMEM supplemented by 2 mM glutamine, 1× penicillin/streptomycin and 10% (v/v) heat-inactivated fetal bovine serum (complete DMEM). The ras-transformed and parental NIH3T3 clone7 cells lines were kindly provided by Dr Douglas Lowy (NIH, Bethesda, MD, USA).
For in vivo studies, 6–7 week-old female (Balb/c) athymic nude mice were obtained from Charles Rivers Labs (Fredrick, MD, USA). Food and water were administered ad libitum.
Plasmid construction and DNA transfection
The pcDNA3 (Invitrogen, San Diego, CA, USA) eukaryotic expression vector was utilized for the establishment of stable ATX-expressing transfected clones, while the zeomycin-sensitive pcDNA3.1/Zeo vector was used for ras-transformed NIH3T3 cells. For pcDNA3 constructs, cDNA of full-length human atx or its T210A mutant form was prepared by HindIII–EcoRV digestion and ligated into identically cleaved vector. For pcDNA3.1/Zeo constructs, ATX cDNA and vectors were digested with HindIII and XbaI restriction enzymes. LipofectamineTM was utilized for all transfections, following the manufacturer's recommended protocol (Life Technologies, Inc.). Concurrent transfections with the appropriate empty vector served as controls. Selection was begun 3 days after transfection with medium containing 1 mg/ml geneticin (GIBCO, BRL) for pcDNA3 constructs and 0.31 mg/ml zeocin (Invitrogen) for pcDNA3/Zeo constructs. Individual clones were isolated and characterized by Western blot analysis.
Soft agar colony formation assay
For each cell line, 5×103 cells were seeded in triplicate into 24-well plates containing 0.3% low melting agarose layered over 0.7% low melting agarose. Colonies were counted after a 14-day incubation at 37°C.
Tumorigenicity and spontaneous metastases
Unanesthetized mice were randomized into experimental groups and inoculated subcutaneously into the left flank. Inocula contained 4×105 (or 1×106) cells in 0.1 ml DMEM for the NIH3T3-derived clones or 2.5×104 cells for ras-transformed clones. Three-dimensional mean volumes of visible solid tumors were measured every other day with a caliper. At indicated times, the animals were sacrificed and tumors were resected, measured and preserved in 10% formalin in PBS. Liver, kidneys, lungs, and spleen were examined for each mouse, no visible metastatic nodules were found in any control or experimental mouse after subcutaneous injection. Cell viability was determined prior to injections by Trypan blue exclusion and found to be greater than 95%. The order of injection was randomized to eliminate any inadvertent selection bias.
Experimental pulmonary metastases
Unanesthetized athymic nude mice each received 1×105 cells via tail vein injection. Mice were monitored daily and sacrificed 15–40 days post injection. Lungs were removed, washed with PBS, and fixed in neutralized 10% formalin in PBS. Metastases were confirmed by pathological examination, and the numbers of superficial lung colonies were counted. Cell viability was determined as above.
In vitro motility and invasion assays
Experiments were performed in triplicate. Chemotaxis was assayed in 48 well microchemotaxis chambers with gelatin-coated 8 μm polyvinylpyrrolidine-free polycarbonate filters (Neuroprobe, Cabin John, MD, USA) as described previously (Aznavoorian et al., 1996). Invasion was assayed in Biocoat MatrigelTM invasion chambers (Mulvaney et al., 1998). The lower chambers contained either partially purified rATX (Lee et al., 1996a) diluted in 0.1% bovine serum albumin-DMEM or this same medium alone. Chemotaxis assays were incubated for 4 h, invasion chambers for 20–24 h at 37°C. The membranes were fixed, stained with Diff-Quik, and counted by direct visualization of nuclei in four high power (400×) fields.
Western blot analysis
Conditioned media from each clone was collected and concentrated 100-fold in Amicon ultrafiltration chambers. The media were partially purified by lectin affinity chromatography with concanavalin A-agarose (Vector Labs), samples were separated by gel electrophoresis, and immunoblots were prepared as previously described (Lee et al., 1996a).
Northern blot analysis
Total cellular RNA was isolated from 0.1 gm of tumor tissue using the Micro RNA Isolation Kit (Stratagene, La Jolla, CA, USA) following the manufacturer's protocol. The RNA (30 μg) was separated by electrophoresis through a 1% agarose/formaldehyde gel and transferred to a nylon membrane. Hybridizations were carried out with previously described cDNA probes (Lee et al., 1996b) for 1 h at 68°C. The membrane was washed twice for 15 min at room temperature with 2× SSC containing 0.1% SDS and then for 30 min at 60°C with 0.1× SSC containing 0.1% SDS. The blots were analysed by PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA) after overnight exposures. The relative amounts of mRNA were normalized to internal glyceraldehyde 3-phosphate dehydrogenase (G3PDH) controls.
Aznavoorian S, Stracke ML, Parsons J, McClanahan J and Liotta LA . 1996 J Biol Chem 271: 3247–3254.
Bachner D, Ahrens M, Schroder D, Hoffmann A, Lauber J, Betat N, Steinert P, Flohe L and Gross G . 1998 Dev Dyn 213: 398–411.
Clair T, Lee HY, Liotta LA and Stracke ML . 1997 J Biol Chem 272: 996–1001.
Deissler H, Blass-Kampmann S, Bruyneel E, Mareel M and Rajewsky MF . 1999 FASEB J 13: 657–666.
Deterre P, Gelman L, Gary-Gouy H, Arrieumerlou C, Berthelier V, Tixier JM, Ktorza S, Goding J, Schmitt C and Bismuth G . 1996 J Immunol 157: 1381–1388.
Fuss B, Baba H, Phan T, Tuohy VK and Macklin WB . 1997 J Neurosci 17: 9095–9103.
Huang RP, Fan Y, Peng A, Zeng ZL, Reed JC, Adamson ED and Boynton AL . 1998 Int J Cancer 77: 880–886.
Ju WD, Velu TJ, Vass WC, Papageorge AG and Lowy DR . 1991 New Biol 3: 380–388.
Kawagoe H, Stracke ML, Nakamura H and Sano K . 1997 Cancer Res 57: 2516–2521.
Kindler-Rohrborn A, Ahrens O, Liepelt U and Rajewsky MF . 1985 Differentiation 30: 53–60.
Kindler-Rohrborn A, Blass-Kampmann S, Lennartz K, Liepelt U, Minwegen R and Rajewsky MF . 1994 Differentiation 57: 215–224.
Lee HY, Clair T, Mulvaney PT, Woodhouse EC, Aznavoorian S, Liotta LA and Stracke ML . 1996a J Biol Chem 271: 24408–24412.
Lee HY, Murata J, Clair T, Polymeropoulos MH, Torres R, Manrow RE, Liotta LA and Stracke ML . 1996b Biochem Biophys Res Commun 218: 714–719.
Liotta LA, Mandler R, Murano G, Katz DA, Gordon RK, Chiang PK and Schiffmann E . 1986 Proc Natl Acad Sci USA 83: 3302–3306.
MacDonald NJ, Freije JMP, Stracke ML, Manrow RE and Steeg PS . 1996 J Biol Chem 271: 25107–25116.
Mulvaney PT, Stracke ML, Nam SW, Woodhouse E, O'Keefe M, Clair T, Liotta LA, Khaddurah-Daouk R and Schiffmann E . 1998 Int J Cancer 78: 46–52.
Murata J, Lee HY, Clair T, Krutzsch HC, Arestad AA, Sobel ME, Liotta LA and Stracke ML . 1994 J Biol Chem 269: 30479–30484.
Stefan C, Gijsbers R, Stalmans W and Bollen M . 1999 Biochim Biophys Acta 1450: 45–52.
Stracke ML, Clair T and Liotta LA . 1997 Adv Enzyme Regul 37: 135–144.
Stracke ML, Krutzsch HC, Unsworth EJ, Arestad A, Cioce V, Schiffmann E and Liotta LA . 1992 J Biol Chem 267: 2524–2529.
Takahashi T, Carswell EA and Thorbecke GJ . 1970 J Exp Med 132: 1181–1190.
Todaro GJ, Fryling C and De Larco JE . 1980 Proc Natl Acad Sci USA 77: 5258–5262.
Woodhouse EC, Chuaqui RF, and Liotta LA . 1997 Cancer 80: 1529–1537.
We would like to thank Dr Elliott Schiffmann for his contributions to experimental design and his suggestions in the preparation of the manuscript.
About this article
Cite this article
Nam, S., Clair, T., Campo, C. et al. Autotaxin (ATX), a potent tumor motogen, augments invasive and metastatic potential of ras-transformed cells. Oncogene 19, 241–247 (2000). https://doi.org/10.1038/sj.onc.1203263