A plasmid containing the K-fgJ proto-oncogene linked to the dihydrofolate reductase gene has been constructed, and used in transfection experiments to investigate the effects of K-fgf expression on the tumorigenic and metastatic properties of NIH-3T3 fibroblasts. Analysis of cells transfected with K-fgf revealed that expression of the K-fgJ proto-oncogene can, in a single step, induce both tumorigenic and metastatic characteristics, as determined in soft agar cloning experiments, and in tumorigenicity and experimental lung metastasis assays with BALB/c n u / n u mice. Selection for resistance to increasing concentrations of methotrexate lead to the isolation of a series of cell lines containing amplifications of both the dihydrofolate reductase gene and the linked K-fgdf gene, which synthesized elevated levels of growth factor message and protein. The most highly resistant and gene amplified cell lines exhibited lower than expected levels of K-fgJ mRNA, and also appeared to have down-regulated cell surface growth factor receptors. Further support for the concept that altered K-fgf expression can induce fully malignant and metastatic cells was obtained in experimental metastasis assays, where K-fgf transfected and gene amplified cell lines were highly aggressive.
Introduction
Metastasis is a complex process that results in tumor growth at sites distant from the primary neoplasm. The spread of cancer is responsible for the majority of cancer deaths because conventional therapy is often unsuccessful in the treatment of widely disseminated disease. It now appears that many of the modified properties of metastatic populations may be controlled through alterations in expression of a relatively small number of key genes [1,2], normally important in the regulation of diverse cellular functions including prolif- eration, differentiation, cell-cell communication and motility [3,4]. These genes are frequently referred to as cellular oncogenes and tumor supressor genes [1,2,4,5]. In addition, there is evidence that growth factors can play a critical role in cellular transformation through a mechanism of autocrine stimulation, and many tumor cell populations appear to release potently mitogenic growth factors, which contribute to the malignant state
Correspondence: J.A. Wright, Manitoba Institute of Cell Biology, University of Manitoba, 100 Olivia Street, Winnipeg, Manitoba R3E 0V9, Canada.
[6-8]. Indeed, oncogenes have been described which code for polypeptide growth factors or their receptors [2,9]. Recently, transfection experiments carried out with sequences encoding the fibroblast growth factors bFGF and kFGF have shown that they can exhibit potent transforming activity [6,10-12], and the k-fgf gene has been found amplified in a variety of human tumors [13-16].
Amplification of localized discrete chromosomal segments is essentially an aberrant process which fre- quently occurs during the advanced steps in the emer- gence of highly malignant tumors [2,14-17]. The con- cept that amplification of cellular genes might confer a growth advantage for proliferation of tumor cells and, therefore is related to disease prognosis, is a popular theme in cancer research (e.g., Ref. 18). However, more work needs to be done at the molecular level, to further test the relationship between increasing proto- oncogene amplification and metastatic potential. Al- though transfection studies with k-fgf have shown that transformed cells expressing K-fgf also produce subcu- taneous tumors in mice [11,19], the metastatic potential of K-fgf transfected and gene amplified lines has not been documented. In the present investigation we have evaluated the ability of K-fgf to convert NIH-3T3 cells
104
to the metastatic phenotype, following transfection and stable integration. The proto-oncogene, K-fgf was am- plified in NIH-3T3 cells, on its own promoter, using linkage to the dihydrofolate reductase (dhfr) gene, which also allowed us to investigate K-fgf expression and amplification characteristics in relationship to tu- mor progression in an experimental metastasis model.
Materials and Methods
Source of plasmids. The Kaposi-fibroblast growth factor (k-fgf) cDNA (pG3(B)-SacI) used as a probe as well as the genomic k-fgf gene pG6.6, was kindly provided by Dr. C. Basilico, N.Y. University School of Medicine, and has been described by Delli-Bovi and Basilico [20]. The dhfr-containing amplification plas- mid containing the dhfr mini-gene, MG4, in the pGEM3 vector was a generous gift from Dr. R. John- ston at the University of Calgary, Alberta. The probes required for experiments were obtained from plasmid preparations cut with the appropriate restriction en- donucleases. The pG3(B)-Sac I plasmid containing the cDNA sequence for the k-fgf gene was digested with Eco R1 and Sacl to yield the 600 bp band correspond- ing to k-fgf cDNA. The MG4 amplification plasmid was cut with Hind III and Pst I to yield the 4 kb murine genomic dhfr gene (R. Johnson, personal com- munication), which also was used as a probe.
Isolation and ligation of K-fgf and the dhfr plasmid. The dhfr amplification plasmid consists of the MG4 dhfr minigene [21] in a pGEM3 vector (Promega, Madison, WI), in the multicloning region at the Hind Ill and Pst I site. The genomic k-fgf gene in the pG6.6 plasmid [20] was also cloned into a pGEM3 vector in the multicloning region at the Sal I site. To construct a k-fgf amplification plasmid the k-fgf gene was isolated and purified from the pG6.6 plasmid and inserted into the dhfr plasmid at the Sal I site of the multicloning region. More specifically, 10 p,g of the pG6.6 plasmid was cut with 8 units/p,g Sal I, for 3 h and purified using a Geneclean kit (Bio 101, La Jolla, CA). The resulting insert was resuspended in TE buffer (10 mM Tris-HC1 pH 8.0 plus 1 mM EDTA) at a concentration of approx. 0.5 # g / u l . The vector, dhfr plasmid, was also cut with 8 uni ts /ug Sal I in the presence of TA buffer (TA buffer contains 33 mM Tris-acetate (pH 7.9), 66 mM potassium acetate, 10 mM magnesium acetate, 0.5 mM 2-mercaptoethanol and 50 p.g per ml BSA) for 2 h. The vector was then dephosphorylated using HK T M phosphatase (Bio /Can Scientific, Missis- sauga, Ont.) by first adding 5 mM CaC12 to the cut vector and then 1 unit / /xg HK T M phosphatase. The vector was incubated for 1 h at 30°C and the HK T M
phosphatase was deactivated by incubation of the vec- tor at 65°C for 30 min. The vector and cut k-fgf gene were then mixed in a ratio of 2 : 1, vector to insert, in a
final concentration of 50 /zg/ml total DNA in a total of 20/~1. As a control, vector alone was also diluted to 50 /z l /ml in a total of 20 #1 TE buffer. Ligase, 1 × ligase buffer (1 × ligase buffer is 0.05 M Tris-HC1 (pH 7.6), 10 mM MgC12, 1 mM ATP, 1 mM dithiothreitol and 5% polyethylene glycol-8000) and 10 mM dATP, final concentration, were added to the vector and vec- tor, insert mixture and the samples were incubated at 12°C overnight. The next day the mixture was used to transform HB101 E. coli [22]. The ligated plasmids were diluted to 100 /zl with TE and mixed with the competent E. coli culture and set on ice for 30 min. The E. coli cells were then heat shocked at 42°C for 2 rain. and incubated at 37°C in 1 ml LB broth [22]. The mixture was centrifuged, resuspended in 100 ~1 LB broth, and spread on LB plates containing 50 p.g/ml ampicillin. The plates were incubated overnight at 37°C, after which colonies were picked and used to inoculate 5 ml LB medium containing ampicillin, so that a mini plasmid preparation could be carried out to identify recombinant plasmids [22].
Transfection into mouse cell lines. Co-transfection of the pKFD (K-fgf, dhfr) plasmid and the hygromycin resistance drug marker plasmid pY3 [23] was carried out by the lipofection method [24,25]. After approx. 10-14 days in the presence of 0.2 m g /m l hygromycin, drug resistant colonies were isolated and analyzed by Southern and Northern blot experiments to confirm integration and expression of transfected DNA.
Nucleic acid preparation and Southern and Northern blot analysis. Genomic DNA was isolated from cells by phenol-chloroform extraction [26]. Southern blot analy- sis was carried out as we have previously described [27] and either 2 or 20 p.g of DNA was digested to comple- tion with 3-4 units/ /zg DNA of the desired restriction endonuclease. Total cellular RNA was extracted from logarithmically growing cells using a rapid method for mRNA isolation [28] and Northern blot analysis was performed as we have described previously [27].
Western blot analysis. For Western blot analysis 8 ml of conditioned medium from 4 plates of cells was concentrated through a centricon 10 microconcentrator (Amicon, Danvers, MA). Upon concentration to a final volume of 100 /H/sample, 30 /~1 of each sample was prepared fo r SDS-polyacrylamide gel electrophoresis as we have described previously [29,30]. The gel was then transferred to nitrocellulose membranes by the method of Towbin et al. [31]. After transfer, the nitro- cellulose filter was blocked (as described previously, 30) in TBS-Tween (50 mM Tris-C1, pH 7.6, 150 mM NaC1 and 0.5% v / v Tween 20) plus 1% w / v bovine serum albumin (BSA) for 1 h. The filter was then incubated in the TBS-Tween containing the kFGF antibody, 1 : 200 dilution of the 682 antibody obtained from Dr. C. Basilico as described by Quarto et al. [11], for 1 h at room temperature. The filter was then
washed three times for 30 min each in TBS-Tween after which the second antibody conjugated to alkaline phosphatase was incubated with the filter for 1.5-3 h. After washing three times, the filter was placed in 100 ml of 10 mM NaHCO 3 and 0.1 mM MgCI2, the filters were developed using nitrobluetetrazolium and 5- bromo-4-chloro-3-indolyl phosphate as we have previ- ously reported [30].
Preparation of conditioned medium from kFGF pro- ducing cells. Conditioned medium was prepared from exponentially growing cells. The cells were harvested, counted and an aliquot of 5 • 105 cells was added to a 60 mm plate in culture growth medium containing 10% FCS. After 3 h the growth medium was removed and the cells were washed twice with PBS. 2 ml of defined medium consisting of a-MEM plus 10/xg/ml insulin, 5 /xg/ml transferrin and 10 ~g /ml heparin was then added to the cells, and the ceils were incubated at 37°C for 20 h. Following 20 h incubation, the medium was removed and referred to as conditioned medium.
125I-bFGF binding assay. Cells were harvested in the presence of 0.2 mM EDTA and washed twice with PBS. 5" 105 cells were resuspended in 0.5 ml of serum-free a-MEM containing 5 /zg /ml transferrin, 10 /xg/ml insulin, 0.15% gelatin and 10 /xg/ml heparin. Cells were equilibrated for 2 h at 37°C, and then washed twice with 2 M NaC1 in 20 mM sodium acetate to remove endogenous growth factor bound to the receptor. The cells were washed twice again with PBS, resuspended in 0.5 ml of the medium described above in 25 mM Hepes buffer (pH 7.4), 10 ng of bFGF and 0.1/xCi of 125I-bFGF was added for an incubation time of 2 h at 37 ° C, after which time, the cells were washed three times with PBS and resuspended in 1% Triton-X in PBS overnight at 37°C. Radioactivity was then deter- mined in a gamma scintillation counter. As a control
105
for nonspecific binding, this procedure was followed with CHO cells, which do not contain bFGF receptor sites [32], and the counts detected were subtracted from the results obtained with the various 3T3 lines.
Tumorigenicity and experimental metastasis assay. Tumor growth rates, using 9 to 10 week old female BALB/c nu /nu (Life Sciences, St. Petersberg, FL) mice were determined as we have previously described [33,34]. The latency, or time at which a subcutaneous tumor first appears, size 2 × 2 mm, was recorded and the average reported. For the experimental metastasis assay, cells were injected in a vol. of 0.2 ml into the tail veins of BALB/c nu /nu mice, which were sacrificed after 3 weeks by ether anesthesia and Bouins solution (picric acid, formaldehyde, acetic acid (15:5:1)) in- jected intratracheally [33,34]. Lungs were then removed and examined for the presence of metastatic foci.
Cloning in low percentage agar. Colony forming abil- ity was determined with a 0.5% Bacto-agar (Difco Laboratories, Detroit, MI)-a-MEM containing 10% FCS base layer and a 0.33% agar-a-MEM containing 10% FCS growth layer as we have previously described [33].
Results and Discussion
Construction and isolation of K-fgf amplified cell lines. Resistance to methotrexate (MTX) selects for amplifi- cation of the dhfr gene in mammalian cells [21,35], and for transfected genes linked to dhfr in cells containing appropriately constructed plasmid vectors [36]. The pKFD plasmid containing the K-fgf proto-oncogene linked to dhfr was constructed as described in Materi- als and Methods, and transfected by lipofection into normal NIH-3T3 mouse fibroblast cells, in a 5 : 1 molar ratio of pKFD to the drug marker, Py3 [23]. Hy-
Fig. 1. Southern analysis of a series of clones, isolateta Ioitowmg transfection of pKFD. 20 /xg of DNA was digested to completion with the restriction enzyme Pst I and the blot probed with the cDNA for k-fgf. Lane 1 shows the wild type NIH-3T3 line. The arrowhead denotes the
2 . 0 - Fig. 2. Southern blot analysis of positive p K F D transfected clones tor dhfr sequences. 20 Izg of D N A was digested to complet ion with the restriction enzyme Xba I. Lane 1 is the wild type NIH-3T3 cell line. The remaining lanes are of a series of positive transfected clones in the absence: ( - ), and in the presence: ( + ), of growth in MTX as indicated below in nM concentrat ions: Jane 2 ( - ) 3T3.26; ( + ) 3T3.26-40; lane 3 ( - ) 3T3.29; ( + ) 3T3.29-40; lane 4 ( - ) 3T3.3A; ( + ) 3T3.3A-20; lane 5 ( - ) 3T3.3G; ( + ) 3T3.3G-40; lane 6 ( - ) 3T3.4B; ( + ) 3T3.4B-20; lane 7 ( - ) 3T3.4C; ( + 3T3.4C-20; lane 8 ( - ) 3T3.210; ( + ) 3T3.210-40; lane 9 ( - ) 3T3.212; ( + ) 3T3.212-80; lane 10 ( - ) 3T3.1A; ( + ) 3T3. IA-20; lane
11 ( - ) 3T3.1D; ( + ) 3T3.1D-40. The four endogenous dhfr sequences are indicated by arrowheads.
gromycin resistant clones were isolated and tested for acquisition of the pKFD plasmid using Southern analy- sis. Fig. 1 shows some of the clones investigated for K-fgf sequences; an endogenous K-fgf band as well as bands representing transfected sequences were ob- served. Transfected clones were subjected to progres- sively increasing concentrations of MTX. At each step in drug selection, a fraction of the cells capable of proliferation in the presence of MTX was stored at -70°C in medium containing 10% DMSO for further analysis. Fig. 2 shows Southern blots performed with DNA prepared from some of these lines, using the
dhfr probe. Four endogenous dhfr bands were ob- served when DNA was digested to completion with Xba I. In most cases, it was apparent that the endoge- nous dhfr genes were amplified during selection with MTX. Only one clone showed transfected dhfr amplifi- cation without endogenous dhfr gene amplification during drug selection, clone 3T3.3G (Fig. 2, lane 5). This cell line also appeared to have an obvious trans- formed morphology (appeared to be more spindle shaped and refractile). As a control for future experi- ments, the 3T3.29 line (Fig. 2, lane 3) was maintained in culture and drug selection. Southern analysis per-
A 1 2
- +
B 2
- - . + 3
- - I -
23.1 -
9 . 4 -
6 . 6 -
4 .3 -
23.1 -
9 4 - 6~6 4 . 3 -
2 , 3 - 2.0-
2 , 3 - - 2 . 0 -
Fig. 3. (A) Southern blot analysis of k-fgf sequences in the wild type line NIH-3T3, lane 1 and the transfected clone 3T3.29, lane (2) grown in the absence ( - ) of MTX, and previously selected in the presence ( + ) of MTX to a concentrat ion of 40 nM. 20 /.~g of D N A was digested to completion with the restriction enzyme Xba I. (B) Southern blot analysis of dhfr sequences for the wild type line NIH-3T3, lane 1, and the control clones: dB7, lane (2) grown in the absence ( - ) of M T X and previously selected in the presence ( + ) of MTX to a concentrat ion of 40 nM MTX, and dC2, lane (3) grown in the absence ( - ) of M T X and previously selected in the presence ( + ) of MTX to a concentrat ion of 10 nM
MTX. In all cases 20/~g of D N A was digested to complet ion using the restriction enzyme Xba I.
formed to detect K-fgf sequences (Fig. 3A), showed that these cells had lost the transfected K-fgf gene during MTX selection. Furthermore, as an added con- trol for transfection experiments NIH-3T3 cells were also transfected with the dhfr containing plasmid lack- ing K-fgf sequences, and two clones, dB7 and dC2 were isolated. Southern blots indicated that dB7 and dC2 cells had incorporated and amplified the dhfr transfected plasmid following selection in the presence of MTX (Fig. 3B).
Clones of the 3T3.3G line were progressively se- lected for the ability to proliferate in the presence of 10, 20, 40, 60, 80 and 120 nM MTX. The control lines, 3T3.29, dB7 and dC2 were, in comparison, extremely difficult to step-up for growth in the presence of in- creasing concentrations of MTX. However, 3T3.29 cells were eventually selected for the ability to proliferate in growth medium containing 80 nM MTX. In addition, dB7 and dC2 cell lines capable of growing in the presence of 40 and 10 nM MTX, respectively, were also obtained. Southern analysis to detect dhfr and K-fgf sequences in these clones was performed (Fig. 4A and B). As expected amplification of both the K-fgf and dhfr genes occurred in the 3T3.3G series of clones. For example, compared to the transfected but non-drug selected 3T3.3G line, densitometric analysis indicated that the 3T3.3G-80 cell line had increased the K-fgf and dhfr gene copy numbers by about 15- and 8-fold, respectively. The control cell lines, 3T3.29-80, dC2 and dB7 exhibited increased dhfr gene copy numbers of approx. 5, 3 and > 10 fold when compared to the 3T3.3G cell line.
K-fgf expression in transfected cell lines. Northern blot analysis of K-fgf mRNA levels was carried out with NIH-3T3, 3T3.3G, 3T3.3G-10 to 120, 3T3.29-80 and dB7 cell lines. Only the 3T3.3G series of clones produced detectable levels of K-fgf mRNA. As shown in Fig. 5, the level of K-fgf message was markedly elevated above wild type and 3T3.3G cell levels in all the K-fgf amplified lines. Western blot analysis of conditioned medium from 3T3.3G, 3T3.3G-20 and 3T3.3G-80 ceils was performed to confirm the presence of secreted kFGF protein in the medium from trans- fected cells. As shown in Fig. 6, the protein was clearly present in the medium of 3T3.3G-20 cells, consistent with the high K-fgf mRNA levels in these cells (Fig. 5). Also, in keeping with message levels, the kFGF protein concentration in the medium from 3T3.3G-80 cells was markedly above the 3T3.3G level, but lower than de- tected in medium from 3T3.3G-20 cells (Fig. 6). Inter- estingly, the highest level of K-fgf mRNA observed in Northern blots was detected in the least gene amplified and MTX resistant 3T3.3G-10 and 3T3.3G-20 lines, when compared to the more highly drug resistant and K-fgf gene amplified 3T3.3G-40 to 120 cell lines. Al- though several explanations are possible, these findings
_.~ 1 2 3 4 5 6 7
23.1 -
9 . 4 - 6 . 6 - 4 . 3 -
2 , 3 - 2 .0 -
107
]3 1 2 3 4 5 6 7 8 9
23.1 -
9 A - 6 . 6 -
4 . 3 -
2.3-- 2.0-
Fig. 4. (A) Southern blot analysis of K-fgf sequences following digestion of 2 ~zg of DNA to completion with Xba 1. Lane 1 is the wild type NIH-3T3 line, and the transfected clones are: lane (2), 3T3.3G; lane (3), 3T3.3G-10; lane (4), 3T3.3G-20; lane (5), 3T3.3G-40; lane (6), 3T3.3G-60; lane (7), 3T3.3G-80; (B) Southern blot analysis of dhfr sequences following digestion of 2/zg of DNA to completion with Xba I. Lane 1 is the wild type NIH-3T3 line, and the transfected clones: lane (2), 3T3.3G; lane (3), 3T3.3G-10; lane (4), 3T3.3G-20; lane (5), 3T3.3G-40; lane (6), 3T3.3G-60; lane (7), 3T3.3G-80; lane
(8), 3T3.29; lane (9), 3T3.29-80.
may result from differences among the various cell lines in K-fgf transcriptional effeciencies and /or K-fgf message stabilities. Further studies are required to address this question. The 3T3.3G line synthesized relatively low but detectable levels of K-fgf mRNA, when compared to wild type non-transfected cells (compare lanes 1 ° and 2 °, Fig. 5). This resembled the measurable but relatively low levels of secreted kFGF protein found in the medium from 3T3.3G cells when compared to 3T3.3G-20 or 3T3.3G-80 medium (Fig. 6). Also, as expected, kFGF protein was not detected in medium obtained from wild type NIH-3T3 cells (Fig. 6). Northern analysis of dhfr message levels in the
108
1 2 3 4 5 6 7 8 1 ° 2 ° 3 °
1 2 3 4 5 6 7 8
Fig. 5. Northern blot analysis of (top) K-fgf message, and (bottom) glyceraldehyde-3 phosphate dehdrogenase message as a control for RNA loading. R N A was isolated from each cell line 20 h after changing the growth medium. Lane 1 indicates the wild type line NIH-3T3. The transfected cell lines are as follows: lane (2), 3T3.3G; lane (3), 3T3.3G-10;lane (4), 3T3.3G-20; lane (5), 3T3.3G-40; lane (6) 3T3.3G-60; lane (7), 3T3.3G-80; lane (8), 3T3.3G-120; Lanes 1 °, 2 o and 3 ° are an overexposure of lanes 1, 2 and 3 following film
exposure for 5 days as opposed to 24 h.
_1 2 3 4
- 3 0
_21.5 Fig. 6. A photograph of a Western blot showing kFGF protein isolated from identical amounts of conditioned medium from: lane (1), 3T3.3G-80; lane (2), 3T3.3G-20; lane (3), 3T3.3G: lane (4), NIH-3T3 wild type. The polyclonal antibody, 682, was a gift from Dr. C. Basilico (Quarto et al. 1989), and it was used at a 1:200 dilution. The locations of molecular mass protein markers (Bio-Rad) of 30
kDa and 21.5 kDa are shown on the right.
various cell lines showed that, as expected, they also expressed increased levels of dhfr mRNA (data not shown).
Tumorigenic and metastatic potential of K-fgf trans- fected cell lines. To determine the effects of amplifica- tion and overexpression of the K-fgf gene in relation to tumorigenic and metastatic potential of NIH-3T3 cells, the following lines: wild type NIH-3T3, 3T3.3G, 3T3.3G-10, 3T3.3G-20, 3T3.3G-80, 3T3.3G-120, 3T3.29-80, dB7-40 and dC2-10, were injected into BALB/c nu /nu mice, either subcutaneously or intra- venously as outlined in Materials and Methods. Since a correlation between anchorage-independent growth
TABLE I
Tumorigenicity and metastatic characteristics of the wild type N1H3T3 cell line and the transfected clones 3T3.3G, 3T3.3G-lO, 3T3.3G-20, 3T3.3G-80. 3T3.29-80, dB7-40 and dC2-10
Cell line Tumorigenicity Experimental metastas is /105 cells Freq. of cloning in low
Freq. of Latency Freq. of mice No. of lung Mean -+ standard error percentage agar /103 cells
mice (days mean with tumors tumors * of the log with _+ S.E.) (mean_+ S.E.) (no. t umor s+ 1) tumors *
* 105 cells (grown for more than 2 months in the continuous presence of drug (MTX)) were injected subcutaneously or intravenously into n u / n u mice after changing growth medium 20 h prior to injection.
** An analysis of variance (ANOVA) was performed and t-t-tests done to examine differences in experimental metastasis between cell lines. From the mean_+ standard error of the log (no. t umor s+ 1) it was determined that there is no significant difference between the clones 3T3.3G-10, -20, -80 and -120. There is a significant difference at the 5% level between the clone 3T3.3G and all the drug selected clones listed above. The average fold difference as analysed by the analysis of variance was determined to be 3.5 times.
109
and metastatic potential has been observed in some studies [37,38], we have also assessed the cloning effi- ciency of three of the K-fgf transfected lines in soft agar. The results shown in Table I demonstrate that none of the control cell lines (NIH-3T3, 3T3.29-80, dB7-40 or dC2-10) exhibited tumorigenic or metastatic properties. However, all of the K-fgf transfected lines were both tumorigenic and metastatic. For example, the 3T3.3G cell line, which expressed the least amount of K-fgf mRNA and kFGF protein (Figs. 5 and 6) exhibited a rather short tumor latency period (6.5 days) following subcutaneous injections, and formed an aver- age of 6-7 lung metastases in the experimental metas- tasis assay. The other K-fgf transfected and gene am- plified lines, which expressed high levels of K-fgf mRNA (Fig. 5) and protein (Fig. 6) showed tumor latency characteristics ranging from 5.7 to 12.3 days, and exhibited on the average 5-times the number of the lung colonies compared to the 3T3.3G line in the experimental lung metastasis assay. Statistical analysis of the metastasis data indicated that there was a signif- icant difference between all the K-fgf transfected lines and wild type NIH-3T3 cells, and a significant differ- ence at the 95% confidence level between the 3T3.3G cell line and the drug selected K-fgf transfected lines, but not a significant difference between the highly K-fgf amplified lines 3T3.3G-10, 3T3.3G-20, 3T3.3G-80 and 3T3.3G-120.
Binding of 1:5I-bFGF. The mitogenic effects of growth factors like bFGF and kFGF occur through binding to cell surfaces and the activation of receptors. Basic FGF and kFGF compete for binding to common cell receptors, and both growth factors are able to induce receptor down-regulation [32,39]. Therefore, we examined the abilities of suspension cultures of NIH- 3T3, 3T3.3G and 3T3.3G-80 cell lines to bind labelled bFGF. As shown in Table II, 3T3.3G ceils bound approx. 50% of 125I-bFGF compared to wild type NIH- 3T3 cells and the 3T3.3G-80 cell line bound only 13% of the wild type level of 125I-bFGF. This apparent down-regulation of cell surface receptors in the K-fgf transfected cell lines may be a necessary control for the autocrine mechanism of action that has been suggested for this growth factor [19].
In summary, growth factors have an enormous influ- ence on the biological properties of cells, and can for
TABLE II
I:5I-bFGF Binding to NIH-3T3, 3T3.3G and 3T3.3G-80 cells
* The results are from two independent experiments done in tripli- cate.
example, modify cell proliferation related functions in a positive or negative fashion [40] or modulate au- tocrine regulation of the metastatic phenotype of cells in either a stimulatory or inhibitory manner [6]. These observations indicate the importance of examining the contribution of cellular oncogenes encoding polypep- tide growth factors to malignancy, including the metastatic process. This point is further emphasized by reports that altered K-fgf expression through gene amplification appear to play an important role in the development of some human cancers [13-16]. The re- sults presented in this study confirm the potent tumori- genic potential of the K-fgf cellular oncogene and conclusively demonstrate that altered expression of this gene in NIH-3T3 cells can induce highly aggressive cells capable of metastatic spread. The interesting molecular and cellular properties of the variant cell lines described in this study indicate that they will be useful for examining key regulatory features of K-fgf expression. Furthermore, the approach outlined for constructing K-fgf amplified variant lines should be applicable for establishing a variety of in vitro models for investigating the role of K-fgf gene amplification in mechanisms of growth factor regulation and tumor progression.
Acknowledgements
This work was supported by the N.C.I. of Canada (JAW and AHG) and N.S.E.R.C. (JAW). JED ac- knowledges N.S.E.R.C. and the University of Manitoba for Studentship support. JAW and AHG are Terry Fox Senior Research Scientists of the N.C.I. of Canada. We thank Drs. Claudio Bascilico of the N.Y. University Medical Center and Randal Johnston of the University of Calgary, for advice and for generously supplying the antibody and plasmids used in this study. We are also grateful for help provided by Dr. Shanti Samuel, Dr. Maureen Spearman and Mr. Arthur Chan.
References
1 Greenberg, A.H., Egan, S.E. and Wright, J.A. (1989) Invasion Metastasis 9, 360-378.
10 Delli-Bovi, P., Curatola, M., Newman, K.M., Sato, Y., Moscatelli, D., Hewick, R.M., Rifkin, D.B. and Basilico, C. (1988) Mol. Cell. Biol. 8, 2933-2941.
11 Quarto, N., Talarico, D., Sommer, A., Florkiewicz, R., Basilico, C. and Rifkin, D.B. (1989) Oncogene Res. 5, 101-110.
12 Rogelji, S., Weinberg, R.A., Fanning, P. and Klagsbrun, M. (1988) Nature 331, 173-175.
13 Adelaide, J., Mattei, M-G., Marics, I., Raybaud, F., Planche, J., De Lapeyriere, O. and Birnbaum, D. (1988) Oncogene 2, 413-416.
14 All, I.U., Merlo, G., Calahan, R. and Lidereau, R. (1989) Onco- gene 4, 89-92.
15 Theillet, C., Le Roy, X., De Lapeyriere, O., Grosgeorges, J., Adnane, J., Raynaud, S.D., Simony-Lafontaine, J., Golfarb, M., Escot, C., Birnbaum, D. and Gaudray, P. (1989) Oncogene 4, 915-922.
16 Yoshida, M.C., Wada, M., Satoh, H., Yoshida, T., Sakamoto, H., Miyagawa, K., Yokota, J., Koda, T., Kakinuma, M., Sugimura, T. and Terada, M. (1988) Proc. Natl. Acad. Sci. USA 85, 4861-4864.
17 Wright, J.A., Smith, H.S., Watt, F.M., Hancock, M.C., Hudson, D.L. and Stark, G.R. (1990) Proc. Natl. Acad. Sci. USA 87, 1791-1795.
18 Slamon, D.J., Godolphin, W., Jones, L.A., Holt, J.A., Wong, S.G., Keith, D.E., Levin, W.J., Stuart, S.G., Udove, J., Ullrich, A. and Press, M.F. (1989) Science 244, 707-712.
19 Wellstein, A., Lupu, R., Zugmaier, G., Flamm, S.L., Cheville, L., Delli-Bovi, P., Basilico, C., Lippman, M.E. and Kern, F.G. (1990) Cell Growth and Differentiation 1, 63-71.
20 Delli-Bovi, P. and Basilico, C. (1987) Proc. Natl. Acad. Sci. USA 84, 5660-5664.
21 Gasser, C.S., Simonsen, C.C., Schilling, J.W. and Schimke, R.T. (1982) Proc. Natl. Acad. Sci. USA 79, 6522-6526.
22 Maniatis, T., Fritch, E.F. and Sambrook, J. (1982) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York.
23 Blocklinger, K. and Diddelman, H. (1984) Mol. Cell. Biol. 4, 2929-2931.
24 Chang, A.C.Y. and Brennar, D.G. (1988) Focus 10, 4, p66. 25 Feigner, P.L., Gadek, T.R., Holm, M., Roman, R., Chan, H.W.,
Wenz, M., Nothrop, J.P., Ringold, G.M. and Danialsen, M. (1987) Proc. Natl. Acad. Sci. USA 84, 7413-7417.
26 Blin, N. and Stafford, D.W. (1976) Nucleic Acid Res. 3, 2303- 2308.
27 Wright, J.A., Alam, T.G., McClarty, G.A., Tagger, A.Y. and Thelander, L. (1987) Somatic Cell Mol. Genet. 13, 155-165.
28 Gough, N.M. (1988) Anal. Biochem. 173, 93-95. 29 Choy, B.K., McClarty, G.A., Chan, A.K., Thelander, L. and
Wright, J.A. (1988) Cancer Res. 48, 2029-2035. 30 McClarty, G.A., Chan, A., Engstrom, Y., Wright, J.A. and The-
lander, L. (1987) Biochemistry 26, 8004-8011. 31 Towbin, H., Stachlin, T. and Gordon, J. (1979) Proc. Natl. Acad.
Sci. USA 76, 4350-4354. 32 Mansukhani, A., Moscatelli, D., Talarico, D., Le~ska, V. and
Basilico, C.A. (1990) Proc. Natl. Acad. Sci. USA 87, 4378-4382. 33 Egan, S.E., McClarty, G.A., Jarolim, L., Wright, J.A., Spiro, I.,
Hager, G. and Greenberg, A.H. 11987) Mol. Cell Biol. 7, 830-837. 34 Egan, S.E., Wright, J.A., Jarolim, L., Yanagihara, K., Bassin,
R.H. and Greenberg, A.H. (1987) Science 238, 202-205. 35 Schimke, R.T. (1988) J. Biol. Chem. 263, 5989-5992. 36 Wurm, F.M., Gwinn, K.A. and Kingston, R.E. (1986) Proc. Natl.
Acad. Sci. USA 83, 5414-5418. 37 Cifone, M.A. and Fidler, l.J. (1980) Proc. Natl. Acad. Sci. 77,
J. Cancer 35, 667-674. 39 Moscatelli, D. and Quarto, N. (1989)J. Cell Biol. 109, 2519-2527. 40 Schwarz, L.C., Gingras, M.-C., Goldberg, G., Greenberg, A.H.
and Wright, J.A. (1988) Cancer Res. 48, 6999-7003.
