AU5197986A

AU5197986A - Transmissible, mammalian genes associated with tumor metastasis

Info

Publication number: AU5197986A
Application number: AU51979/86A
Authority: AU
Inventors: Shelly C. Bernstein; Robert A. Weinberg
Original assignee: Whitehead Institute for Biomedical Research
Current assignee: Whitehead Institute for Biomedical Research
Priority date: 1984-11-30
Filing date: 1985-11-26
Publication date: 1986-06-18
Also published as: JPS62501399A; FI863095A0; GR852864B; DK359486A; ZA859070B; PT81584A; NO863066L; PT81584B; HUT41837A; DK359486D0; FI863095L; NO863066D0; IL77157A0; ES549438A0; WO1986003226A1; ES8705043A1; EP0203970A1

Description

TRANSMISSIBLE, MAMMALIAN GENES ASSOCIATED WITH TUMOR METASTASIS

Description

Technical Field This invention is in the field of molecular biology and more specifically relates to the detec¬ tion and isolation of discrete, transmissible, ma- alian genes associated with tumor metastasis.

Background Art Although the molecular basis for malignant transformation leading to cancer is not yet fully understood, much information about the process has been developed recently using molecular biology techniques. For example, while it has long been thought that transformation involved the alteration of critical genes, referred to as oncogenes, such discrete oncogenes have only recently been isolated and shown to cause transformation.

One recent approach to isolation of an oncogene involved the tr-ansfer of tumor cell DNA from the BJ bladder carcinoma cell line into non-transformed NIH 3 3 mouse fibroblasts. In this work, it was found that the phenotype of cellular transformation could be passed from cell to cell in this manner. Tumor DNA was able to induce foci of transformed cells in the recipient NIH monolayer cultures while DNA from normal, untransformed donor cells failed to produce foci. See Shih, C, Shilo, B. , Goldfarb, M.P.,

Dannenberg, A. and Weinberg, R.A. Proc. Natl. Acad. Sci. USA 76.5714-5718 (1979); Cooper, G.M. , Oken- guist, S. and Silverman, L. Nature 284; 418-421 (1980); Shih, C. , Padhy, L.C., Murray, M.J. and Weinberg, R.A. Nature 290; 261-264 (1981); Kron- tiris, T.G. and Cooper, G.M. Proc. Nat-1. Acad. Sci. USA 78; 118Ϊ-1184 (1981); and Perucho, M. et al. Cell 27; 467-476 (1981) . These results demonstrated oncogenic factors present in the EJ tumor cell line DNA which were apparently absent from the DNA of normal cells.

Studies which examined the sensitivity or resistance of oncogenic DNA from the EJ bladder carcinoma line to treatment of various site-specific endonucleases indicated that certain specific donor DNA sequences were involved in such cellular trans¬ formation. See Lane, M.A. , Sainten, A. and Cooper, G.M. Proc. Natl. Acad. Sci. USA 78; 5185-5189 (1981) ; and Shilo, B. and Weinberg, R.A. Nature 289; 607-609 (1981) . A discrete, definable oncogene was later directly demonstrated by molecular isolation of a transforming gene from the EJ human bladder carcinoma cell line by a method involving inter- species transfection. See Shih, C. , Weinberg, R.A.

(1982) Cell 29; 161-169.

Isolated human sequences from the c-K-ras • oncogene which are present in certain human lung tumors have been described. See Nakano, H. , Yama- moto, F. , Neville, C. , Evans, D. , Mizuno, T. , andPerucho, M. , "Isolation of Transforming Sequences of Two Human Lung Carcinomas: Strucutural and Functional Analysis of the Activated c-K-ras Onco¬ genes", Proc. Acad. Sci. USA; 71-75, January, 1984; Santos, E. , Martin-Zanca, M. , Reddy, P., Pierotti, M.A. , Delia Porta, G. , Barbacid, M. , "Malignant Activation of a K-ras Oncogene and Lung Carcinoma but Not in Normal Tissue of the Same Patient", Science 223; 661-4, February 17, 1984.

Oncogenes, such as those of the ras group, are able to induce full tumorigenic conversion of immortalized cells, such as NIH 3T3 mouse fibro- blasts. The effects of these ras oncogenes on primary embryo fibroblasts are more circumscribed; in these cells, the ras oncogene requires the collaboration of a second oncogene, such as myc to induce tumorigenicity. See Land, H. , Parada, L.F. and Weinberg, R.A. Nature (London) 304; 596-602 (1983); Ruley, H.E. Nature (London) 304; 602-606

(1983) . In either case, ras-transformed NIH 3T3 cells or ras-plus-myc transformed embryo fibro¬ blasts, the cells form localized non-metastasizing tumors in immunocompetent hosts. See Land e_t al. , supra. This suggests that the ability of such tumor cells to metastasize requires further alterations. Up until the present time, however, very little was known about what genetic alterations were involved in metastasis.

Disclosure of the Invention This invention relates to the discovery of a discrete, transmissible, mammalian gene associated with tumor metastasis.^'

One embodiment of the invention comprises a method for isolating such a discrete, transmissible gene associated with tumor metastasis from the DNA of a mammalian species. In this method, donor DNA from metastatic mammalian tumor cells ("donor DNA") is fragmented into a multiplicity of fragments, at least one of which contains a discrete, transmissi- ble gene of interest because it is associated with metastasis of the tumor. The multiplicity of fragments and a selectable marker are then trans¬ mitted into recipient cells capable of phenotypi- cally expressing the presence of the selectable marker as well as the discrete, transmissible gene of interest. The recipient cells are then cultured under conditions which allow phenotypic expression of the selectable marker. Because of such pheno¬ typic expression, recipient cells can be selected which have acquired the gene of interest, in addi¬ tion to the marker; they may also contain-additional DNA sequences from the donor fragments transmitted into the recipient cells, as well as their own endogenous DNA. Recipient cells selected because of phenotypic expression of the selectable marker are introduced into a mammalian host under conditions such that the introduced cells would not be meta¬ static in the absence of the acquired gene of interest. These cells are allowed to form a primary tumor at the site of introduction in the mammalian host, and the host is then examined for metastases. DNA is then recovered from cells of any metastases formed in the host or from cultures of cells removed from the metastases. The aforementioned steps are then repeated substituting DNA from the metastases in place of original donor DNA until DNA recovered from a subsequently arising metastasis is found to contain only that portion of the original donor DNA that constitutes essentially a discrete, transmissi¬ ble, mammalian gene associated with metastasis and a donor marker indicating that this fragment origi-

» nated from the original donor DNA. DNA recovered from such metastasis is then broken into fragments, at least one of which is a fragment containing essentially only the gene of interest and donor marker, after which the gene of interest is re¬ covered. Such recovery may be, for example, by procedures of gene cloning.

In one embodiment of the invention which has actually been experimentally performed, a gene associated with metastasis of the human cell line ME-180, a human cervical carcinoma metastatic to omenturn, has been isolated. The original DNA from this cell line was serially cotransfected into NIH 3T3 mouse fibroblast cells with pSV2neo DNA, a selectable marker. Subcutaneous injection of such cells into immunocompetent mice resulted in an initial metastasis to the lung of one mouse and a subsequent metastasis to the abdomen of a second mouse. The ME180 gene does not appear to be closely related to the royc oncogene or to the ras oncogenes in that it did not exhibit reactivity in nucleic acid hybridization tests using DNA probes that are specific for the yc or ras oncogenes.

In another embodiment of the invention, which has also been experimentally performed according to the procedures described above, a gene associated with metastasis of the human cell line SK-N-MC, a human neuroblastoma metastatic to the supraorbital area, has been isolated. Subcutaneous injection of such cells into immunocompetent mice resulted in abdominal metastasis in one mouse. Isolating a mammalian gene associated with metastasis has many desirable consequences. The isolated gene can be compared with closely related sequences in normal DNA and such comparisons should lead to an understanding of what alterations occur to lead to metastasis.

Additionally, protein coded for by the gene can be produced in significant quantity so that it can be studied to understand the metabolic alterations that occur in the cell during the expression of the metastatic trait. This may also lead to insights into methods by which such a gene, gene products or cellular products affected by the gene product could be antagonized or inhibited. It is, of course, expected to lead to sensitive tests for the presence of this gene or products of this gene involving probes for the gene, RNA transcribed from the gene, or antibodies reactive with products of the gene. Brief Description of the Figure

The" figure is a photograph of a blot illus¬ trating the results of tests for the presence of human ras and myc oncogene sequences in DNA from metastases derived from secondary transfectants of EJ-6-2-Bam-6a cells.

Best Mode For Carrying Out The Invention

Although the experimental work presented below involved detection of a gene associated with meta- stasis for the ME-180 cell line and a gene associ¬ ated with the SK-N-MC cell line, any discrete, transmissible mammalian gene associated with meta¬ stasis could be detected using the techniques of this invention. A "discrete" gene is one having a contiguous sequence of base pairs located in one bloc__'of sequences of definable length.' This block may contain regions coding for protein as well as intervening sequence regions which do not encode protein. A "transmissible" gene is one which can be transmitted from cell to cell using gene transfer (transfection) techniques.

Donor DNA containing such a discrete, trans¬ missible gene can be isolated from other host cell constituents by art-recognized techniques. For example, cells grown in culture can be lysed and the viscous lysate can then be extracted with phenol and with chloroform-isoamyl alcohol. DNA can then be precipitated by ethanol precipitation.

The initial donor DNA containing the gene of interest can be fragmented by mechanical or enzy¬ maticmethods. For example, it can be passed through a narrow gauged needle so that it will be fragmented due to shear. On the other hand, DNA molecules might also be fragmented using restriction endonucleases. The important limitation on frag- mentation is that at least one fragment results each time the fragmentation is done which contains in intact form all of the gene of interest.

If the fragment containing the gene of interest does not have an associated marker sequence, one must be added. It is possible to establish experi¬ mentally a linkage between a gene and a marker. For example, a donor cell may be tagged with copies of cloned DNA sequence, such as a 0X174 bacteriophage DNA fragment. Upon co-transfection, the donor cell DNA fragments and the cloned marker or "tag" DNA fragments become linked in the recipient cell in a randomly alternating co-polymer. See Lowy, I., Pellicer, A., Jackson, J.F., Simas, I.M. G. K., Silverstein, S. and Axel, R. (1980) Cell 22; 817-823.

Alternatively and preferably, the fragment containing the gene of interest will contain a naturally associated donor-specific marker. For example, human DNA contains over 300,000 copies of the Alu sequence interspersed throughout the entire genome which can serve as a specific marker indicating the presence of human DNA segments. Thus, almost every gene is linked closely, i.e., less than 10 kilobases, to a copy of this repeated sequence which can serve as marker. See Houck, CM., Rinehart, F.D., and Sch id, C.W. (1979) J^ Mol. Biol. 132; 289-306. Since this Alu sequence is not present in mouse DNA, it is species-specific to human DNA. In fact, a molecular clone of the human Alu sequences probe in Southern blots has been used to detect the presence in mouse cells of introduced human onco¬ genes of bladder and colon carcinoma origin and of promyelocytic leukemia origin. See Murray, M.J., Shilo, B.Z., Shih, C, Cowing, D., Hsu, H.W. and Weinberg, R.A. (1981) Cell 25; 355-361. Each of these, when resolved from the mouse sequence back¬ ground, was determined to be affiliated with its own characteristic array of human Alu segments. This species-specific marker, described above, is used in order to insure that a gene finally recovered after isolation procedures did in fact originate with the original mammalian DNA known to be endowed with multiple copies of this specific marker.

Independent of this, fragmented donor DNA can be transmitted into cells with a selectable marker which can be phenotypically expressed by the cells. In the work described herein, the selectable marker employed was the plasmid pSV2neo. Subjecting cultures to G418 selection kills all cells except those carrying this plasmid because of cytotoxic effects of the G418 drug, which drug is inactivated in cells that have acquired the pSV2neo gene. See Southern, E.M. and Berg, P. (1982) J. Mol. Appl. Genet. 1; 327-341.

Of course, other selectable markers could be employed such as the Ecogpt marker which confers resistance to the cytostatic effects of the drug mycophenolic acid. The preferred route for transmitting donor DNA fragments and the selectable marker into cells is transfection, such as calcium phosphate facilitated transfection. DNAs of any sequence or biological origin can be introduced into mammalian cells by transfection. The transfected DNA need not have any sequence homology with the genome of the recipient cell. Co-transfection is the co-mingling of two or more DNAs prior to their being introduced into mammalian cells, such as by the calcium phosphate facilitated transfection procedure.

EJ-6-2-Bam-6a cells were employed as recipients for the genes associated with metastasis in the experiments described herein. Such cells are NIH 3T3 mouse fibroblast cells transformed with the EJ-Ha-ras oncogene, whose presence allows these cells to fpr tumors which are localized and non- metastatic. Furthermore, these cells contain only the 6.6 kb fragment of the EJ-Ha-ras oncogene, devoid of any Alu sequences. NIH 3T3 cells, and their derivatives, are unusual because they take up and allow efficient stable expression of transfected DNAs.

Another suitable cell line for transfection is known as Rat-1 cell line. Although most cell lines are relatively refractory to transfected DNAs, it is believed that other cell lines will be found which allow expression of transfected genes.

After co-transfected cells have phenotypically expressed the presence of the selectable marker, cells showing such expression are selected. For example, a colony expressing the phenotype can be physically picked by standard techniques and the cells of the chosen colony can then be used to seed a large scale culture. Those cells which have expressed the selectable marker can be shown to' have concomitantly acquired substantial amounts of the donor cell genomic DNA whose presence may not be manifested in phenotypic alterations of these cells that are apparent in monolayer culture.

Such selected recipient cells can then be introduced into a mammalian host under conditions such that these recipient cells would be tumorigenic but not be metastatic unless they have acquired a donor gene associated with tumor metastasis. Subcutaneous injection is one suitable technique, but the experimental work described herein also documents that intramuscular injection, intraperi- toneal inoculation as well as injection into the foot pad of mice were suitable also. Intravenous injection was not suitable since otherwise non- metastasizing tumors created metastases when the tumor cells were intravenously injected.

DNA is then recovered from metastatic cells of metastases formed in the host and employed in place of the original mammalian donor DNA for a second round of transfection into the tumorigenic non-meta- static EJ-6-2-Bam-6a cells. In many cases it is preferred to culture cells from the metastasis in order to obtain larger amounts of DNA than might be obtained directly from cells from the metastasis. The preceding steps are repeated, as required. By repeating all of these steps, serial transmission of the gene of interest, together with any donor marker sequence, are accomplished. When transfection techniques of human DNA into mouse fibroblasts are done, it has been found that two transfections are usually sufficient in order to create a recipient cell which has the donor gene of interest as well as its associated marker, but lacks virtually all other extraneous donor DNA segments, including extraneous donor segments carrying donor specific marker sequences. Although there may be very small amounts of extraneous donor DNA present in addition to the donor gene of interest, such small fractions should not impede expression by the gene of interest, nor the detection of this specific gene of interest by nucleic acid hybridization procedures. When a cell has been selected which contains essentially only the gene of interest and its associated donor marker, the marker may be employed in order to identify and recover the gene. For example, a marker gene may be a human repetitive DNA sequence of the Alu type such as that which could be detected by use of the BLUR probe; alternatively, the gene of interest could be experimentally linked to a marker gene prior to the first of the two transfections and this marker gene could then be identified by use of a cloned probe that is reactive with this marker gene. Recovery of the gene can be done by creating- a genomic library. The gene of interest or a portion of it, may be contained within a phage also carrying a donor specific marker sequence such as an Alu sequence or an experi¬ mentally added marker. This phage may be identified employing standard hybridization procedures and the phage can be amplified, yielding the human DNA segment in cloned form.

Once isolated, the mammalian gene associated with the Alu sequence, such as the mammalian gene expressing the metastatic phenotype, can be further cloned into other suitable recombinant DNA vectors. Such cloning employs fundamental gene splicing techniques, such as those described by Cohen and Boyer in U.S. Patent No. 4,227,224. Suitable recombinant DNA vectors include bacterial plasmids, phages, animal viruses and yeast vectors. In such techniques, hosts would be employed which allow the recombinant DNA vector to multiply. Using cloning and expression techniques, significant amounts of the proteins coded for by the gene isolated by the method described herein can be produced. Such proteins can be studied and their effects on modifying other cellular constituents and regulating their level and activity can be eluci¬ dated. Moreover, such proteins can be used to produce antibodies by standard antibody production techniques. Thus, for producing polyclonal anti¬ bodies, such proteins would be employed to immunize a host, such as a rabbit or a rat, and antibodies to the protein would be collected from serum obtained from the host.

Alternatively, monoclonal antibodies could be produced employing cells which produce antibodies to the protein produced by the isolated gene in typical fusion techniques for forming hybridoma cells. Basically, these techniques involve the fusing of the antibody producing cell with a cell having immortality, such as a myeloma cell, to provide a fused cell hybrid which has immortality and is capable of producing the desired antibody, in this case an antibody to the protein coded for by the isolated gene. The hybrid cells are then cultured under conditions conducive to the production of antibody which is subsequently collected from the cell culture medium. Such techniques for producing monoclonal antibodies have been well described in the literature. See, for example, U.S. Patent Nos. 4,172,124 and 4,196,265 issued to Hilary Koprowski et al. , the teachings of which are hereby incorpor¬ ated by reference. A significant use for the antibodies produced_, to such protein is in assays to detect the presence of protein coded for by the isolated gene associated with metastasis. Such assays include immunoassays, such as those radioimmunoassays employing labelled antibodies or viral antigens.

Probes could also be employed for detecting the gene associated with metastasis or mRNA transcribed by said gene. Such probes might comprise, for example, a labelled polynucleotide complementary to at least a portion of a mammalian gene associated with tumor metastasis.

Metastasis may be prevented or dimininshed by employing antibodies capable of neutralizing protein expressed by a mammalian gene associated with metastasis of the tumor. Alternatively, metastasis may be prevented or diminished by antagonizing the gene associated with metastasis, or by antagonizing any cellular product or intermediate affected by protein expressed by said gene.

This invention will now be further and more specifically described in the following examples.

EXAMPLES

EXAMPLE 1 ABILITY OF NIH 3T3 CELLS AND EJ TRANSFECTANTS TO FORM TUMORS AND METASTASES IN NUDE AND NFS/NCr MICE

The ability of NIH 3T3 cells and an NIH 3T3 subline transformed by the Ha-ras oncogene isolated from the EJ human carcinoma cell line was investi¬ gated. Isolation of the Ha-ras oncogene from the EJ human bladder carcinoma cell line has been described by Shih and Weinberg. See, Shih, C, Weinberg, R.A. (1982) Cell 29; 161-169.

The EJ transfectant, termed EJ-6-2-Bam-6-a, was an NIH 3T3-derived tertiary transfectant carrying the EJ-Ha-ras bladder carcinoma oncogene. It was obtained by passing DNA by transfection in two serial cycles through NIH 3T3 cells. DNA of the secondary transfectant was cleaved with the restric¬ tion endonuclease Bam HI prior to the third cycle of transfection, freeing the Ha-ras oncogene from linkage to the human Alu repeat sequences.

The cells were inoculated into either immuno- competent histocompatible mice, which were 6-8 weeks old NFS/NCr mice or immunoincompetent mice, which were NIH nude (nu/nu) mice previously irradiated with 500 rad. Each mouse was inoculated by inject- ing subcutaneously into the flanks of the mice 10 cells in 0.2 ml phosphate buffered saline. The results were as follows:

Mouse Metastatic

Cells Injected Strain Tumorigenicity Ability

NIH 3T3 nu/nu 0/6 -

EJ-6-2-Bam-6a nu/nu 6/6 6/6

NIH 3T3 NFS/NCr 0/8 -

EJ-6-2-Bam-6a NFS/NCr 62/62 1/62.

As can be seen from the above data, subcutan¬ eous inoculation with NIH 3T3 cells into_, immuno¬ competent and immunoincompetent mice produced no tumors. Subcutaneous- inoculation of NIH 3T3 cells transformed by the Ha-ras oncogene isolated from the EJ human bladder carcinoma cell line resulted in the formation of tumors at the site of inoculation for all inoculated immunocompetent and immunoincompetent mice. All immunoincompetent mice were found to have metastases in the lungs. However, only one of the 62 immunocompetent histocompatible mice having tumors at the site of inoculation had evidence of metastatic spread. This indicates that the immuno- competence of these mice stood as an effective barrier to metastasis.

The ability of the inoculated cells to form tumors at the site of inoculation, termed tumorigen¬ icity, was determined by making a ratio of the number of mice in which tumors grew over the number of mice inoculated. Metastatic ability of the tumors was rated by making a ratio of the number of mice found to have metastases over the number of mice found to have tumors at the site of inocula- tion. Metastasis was determined by sacrificing moribund mice up to 6 weeks after inoculation and subjecting them to necropsy.

EXAMPLE 2 METASTATIC ABILITY OF NORMAL CELLULAR DNA AND DNA OF pSV2neo PLASMID TRANSFECTED INTO EJ-6-2-Bam-6-a

CELLS

Since it was desired to co-introduce DNA from human metastatic tumors together with the selectable marker pSV2neo plasmid into EJ-6-2-Bam-6a cells, the ability of normal mouse cellular DNA and DNA from pSV2neo plasmid to cause metastasis was first investigated. Some of the procedures were those of Example 1, but others were different as discussed below. NIH 3T3 cells were cotransfected with normal cellular DNA from NIH 3T3 cells and pSV2neo plasmid DNA following the procedures of Andersson e_t al. See Andersson, P., Goldfarb, M.P., and Weinberg, R.A. (1979) Cell 16; 63-75. Briefly, 75 ug of NIH 3T3 DNA and 1 ug of pSV2neo plasmid DNA were applied to 7.5 x 10⁵ EJ-transformed NIH 3T3 cells, EJ-6-2- Bam-6-a (2 x 10 cm dishes) . Cells were split in a ratio of 1:6. The following day, the cultures were subjected to G418 selection, which killed all cells except those carrying the pSV2neo DNA. See South- ern, E.M. and Berg, P. (1982) J. Mol. Appl. Genet. 1_: 327-341. Colonies resistant to G418 were counted 14 days after transfection.

The G418-resistant cells were collected after they grew to form a dense confluent monolayer

(approximately 21 days after transfection) . Tryp- sinized cells were washed with phosphate-buffered saline. All of the colonies of an individual culture dish were pooled and one million cells were 0 injected into a single animal. These pooled cells were injected NFS/NCr mice at several sites: subcutaneous (SC) ; intravenous (IV) ; intramuscular

(IM) ; foot pad (FP) ; and intraperitoneal (IP). These results were as follows: 5 Route of Metastatic

Cells Injected Inoculation Tumorigenicity Ability

EJ-6-2-Bam-6a/Neo SC 160/166 2/160

IV 4/8 4/4

IM 4/4 0/4 0 FP 4/4 0/4

IP 2/6 0/2.

As can be seen, subcutaneous injection led to metastatic spread in only one mouse out of 38 that had primary tumors. This led to the conclusion that 5 introduction of DNAs, such as NIH 3T3 DNA or pSV2neo DNA, via transfection does not itself significantly affect the metastatic ability of the ras-transformed NIH 3T3 cells. Twelve primary tumors were selected from the 38 mice inoculated and allowed to grow for up to 3 months. Some were left undisturbed, while others were resected either totally or partially. Many tumors grew so large as to penetrate the intraperi- toneal cavity, yet all remained encapsulated without evidence of intraperitoneal seeding or distant metastatic spread. This reinforced the conclusion that these ras-transformed NIH 3T3 did not have significantly high rates of metastasis formation when inoculated subcutaneously.

Mice inoculated intramuscularly in the foot pad or intraperitoneally did not form metastases.

While 4 out of 8 mice intravenously injected in the lateral tail vein formed lung metastases without the presence of a primary tfcmor, this did not bear on subsequent experiments, all of which employed injection by a subcutaneous route.

EXAMPLE 3 METASTATIC ABILITY INDUCED BY HUMAN TUMOR DNA

The ability of genetic information from a meta¬ static tumor to confer metastatic ability to EJ-6-2-Bam-6a cells was determined. These cells were well suited for such experiments because they had a low background of spontaneous metastasis

(Examples 1 and 2) when subcutaneously injected into mice and, being of NIH 3T3 origin, were able to take up efficiently and express exogenous DNA. See Smotkin, D. , Gianni, A.M., Rozenblatt, S. and Weinberg, R.A. (1975) Proc. Natl. Acad. Sci. USA 72: 4910-4913. DNAs from a variety of human metastatic tumor cell lines were applied to cultures of the Ha-ras transformant (EJ-6-2-Bam-6a) , using the cotransfec- tion procedure of Example 2 except substituting- a variety of human metastatic tumor cell line DNAs for the normal NIH 3T3 cell line DNA. Human metastatic cell lines whose DNAs were employed included: ME-180, human cervical carcinoma, metastatic to omentum; SK-N-MC, human neuroblastoma, metastatic to supraorbital soft tissue; HuTu 80, human metastatic duodenal adenocarcinoma; Capan-1, human pancreatic adenocarcinoma, metastatic to liver; Calu-1, human lung carcinoma, metastatic to pleura; SK-N-SH, human neuroblastoma, metastatic to bone marrow; Hs0895, human melanoma, metastatic to lung; Hs0891, human renal cell carcinoma, metastatic to lymph node; and

SK-MEL-5, human melanoma, metastatic to lymph node.

To eliminate those cells from the transfected cultures which had not taken up and fixed donor human tumor DNA, the metastatic human tumor DNA was cotransfected with pSV2neo plasmid and G418 selec¬ tion was applied to the transfected cultures. All untransfected cells were thus killed which allowed the outgrowth of large numbers of colonies carrying the pSV2neo marker and a substantial amount of concomitantly acquired human tumor DNA.

It was estimated that each of these cotrans¬ fected colonies acquired approximately one one-thousandth of a donor tumor cell genome. See Perucho, M. , Hanahan, D. and Wigler, M. (1980) Cell 22: 309-317. Thus, several thousand such colonies needed to be tested in order to be sure that at least one carried a single copy donor gene of interest. Each transfection tested yielded 1500 to 4000 G418-resistant colonies among 12 culture dishes. The colonies of each dish were pooled and 1,000,000 cells from such a pool were inoculated subcutaneously into immunocompetent NFS/NCr mice at subcutaneous sites. Metastases were scored 2-6 weeks later by examination of the abdominal and thoracic organs.

The results were as follows:

Metastatic

DNA Transfected Tumorigenicity Ability ME-180 + pSV2neo 32/32 2/32 SK-N-MC + pSV2neo 24/24 1/24 HUTu 80 + pSV2neo 23/24 0/23 Capan-1 + pSV2neo 24/24 0/24 Calu-1 + pSV2neo 24/24 0/24 SK-N-SH + pSV2neo 24/24 0/22 Hs0895 + pSVneo 24/24 0/24 Hs0891 + pSV2neo 23/24 0/23 SK-MEL-5 + pSV2neo 24/24 0/24.

As can be seen, a large proportion of inocu- lated mice demonstrated primary tumor formation.

Initially, however, only one exhibited a metastasis. This single mouse, inoculated with primary trans¬ fectants containing human ME-180 DNA formed a metastasis to the lung. Subsequently, one addi- tional mouse inoculated with primary transfectants containing human ME180 DNA and one mouse inoculated with primary transfectants containing SK-N-MC DNA formed metastases to the peritoneal cavity. EXAMPLE 4 SECONDARY (2°) AND TERTIARY (3°) TRANSFECTANTS EMPLOYING DNA FROM THE METASTASIS IN THE PRIMARY (1°) TRANSFECTANT

Experiments were conducted to determine whether the original metastasis derived from the first inoculation of primary transfectants of human ME180 DNA in the single mouse of Example 3 was induced by human DNA sequences acquired in transfection or developed spontaneously and independently of intro¬ duced genes. Thus, tests were made to determine whether the metastatic phenotype could be further passed from this initially arising metastasis to other ras-transformed NIH 3T3 cells. ^' DNA was prepared from the original metastasis and cotransfected into Ha-ras transformed NIH 3T3 cells (EJ-6-2-Bam-6a) with pSV2neo employing the procedures of Example 3. The culture of these cells was divided into 12 sub-cultures after 1 day and then placed under G418 selection. Nine of the resulting cultures were inoculated into NFS/NCr mice employing the inoculation procedures of Example 3; six mice has large tumors. Upon autopsy, 4 of the 6 mice carried extensive metastases that were detect- able 14-21 days after inoculation.

The four cultures yielding these metastatic cells were retested by inoculation into 38 mice. Of these, 35 displayed tumors and 14 of the tumor- bearing mice carried metastases. This led to the conclusion that the metastatic behavior of cells in these 4 cultures was a reproducible phenomenon. Moreover, it appeared that the metastasis-inducing determinants, which were present in low concentra¬ tion in the original human ME-180 tumor DNA, were detectable in enhanced amounts in the DNA of the primary transfectants. Such increase in gene dosage is often seen after transfection of a variety of genes, in that sequences present in single copy and the genomes of donor cells are fixed in multiple copies in the genomes of transfected recipients. See Perucho, M. , Hanahan, D. and Wigler, M. (1980) Cell 22: 309-317.

DNAs from 2 of the 4 metastases arising from secondary transfectants were cotransfected once again with pSV2neo DNA and a third cycle of transfection and the resulting cultures tested for metastatic ability in NFS/NCr mice. One secondary transfectant yielded DNA which was used to induce a group of tertiary transfectants. These tertiary transfectants induced primary tumors in 18 of 22 mice inoculated and metastasis was observed in 8 of the 18 mice having primary tumors. In the case of the other secondary transfectant, the derived tertiary transfectants resulted in primary tumors in 11 of 12 mice inoculated and metastasis was observed in 3 of the 11 mice having these primary tumors. Histological examination of the pulmonary and intraperitoneal metastases revealed that these metastatic tumors were fibrosarcomas, as were the primary non-metasizing tumors.

DNA was also prepared from the second metastasis derived from primary transfectants of human ME180 DNA and cotransfected into Ha-ras transformed NIH 3T3 cells (EJ-6-2-Bam-6a) with pSV2neo employing the procedures of Example 3. The culture of these cells was divided into 12 sub¬ cultures after one day and then placed under G148 selection. The cultures were inoculated into

NFS/NCr mice employing the inoculation procedures of Example 3. All twelve mice had evidence of tumor formation; six of the twelve mice carried extensive metasases detectable upon autopsy 14-21 days after inoculation.

Similarly, DNA was prepared from the single metasasis derived from primary transfectants of human SK-N-MC DNA and co-transfected^•into Ha-ras transformed NIH 3T3 cells (EJ-6-2-Bam-6a) with pSV2neo employing the procedures of Example 3. The culture of these cells was divided into 12 sub- cultures after one day and then placted under G148 selection. The cultures were inoculated into NFS/NCr mice employing the inoculation procedures of Example 3. All twelve mice had evidence of tumor formation; seven of the twelve mice carried exten¬ sive metastases detectable upon autopsy 14-21 days after inoculation.

EXAMPLE 5 TESTS FOR THE PRESENCE OF HUMAN ras AND myc

ONCOGENE SEQUENCES IN METASTASES

Tests were conducted for the presence of human ras and myc oncogene sequences in metastases derived from secondary transfectants of EJ-6-2-Bam-6a cells. The results are illustrated in Figure 1. DNAs were digested with the eudonucleases Xba I (lanes a. - £) and Eco RI (lanes d-1) and analyzed by Southern transfer.

The procedures for Southern transfer analysis have been previously described. See Murray, M.J., Shilo, B.-A. , Shih, C. , Cowing, D., Hsu, H.-W. and Weinberg, R.A. (1981) Cell 25; 355-61. Briefly, whole cell DNA was prepared from cells grown from metastases. Following restriction endonuclease digestion, 10 ug of the DNA of each metastasis were resolved by electrophoresis through a 1% agarose gel in 40 mM Tris (pH 7.9), 50 mM sodium acetate and 1 mM EDTA. After electrophoresis, the DNA was trans¬ ferred to nitrocellulose by the method of Southern. See Southern, E.M. (1975) J. Mol. Biol. 98; 503-518. The resulting blots hybridized with probes f<_>r human repetitive DNA. . The filters were incubated with probes specific for the following human genes Ha-ras (lanes a.-£) , Ki-ras (lanes d-f_) , N-ras (lanes -.i) , myc (lanes 2.-1) . The DNAs analyzed here are as follows: a_, ά , £, j_, EJ-6-2-Bam-6a tertiary trans¬ fectant derived from EJ bladder carcinoma DNA; lanes __.' ' ii' £' ME180-2°Metl2L, metastasis derived from secondary transfectants of EJ-6-2-Bam-6a cells transfected with ME-180 metastatic carcinoma DNA; lanes £, _f, i_, _1, ME-180 human metastatic cervical carcinoma. Size markers (in kilobases) are indi¬ cated on the ordinate.

An Ha-ras homologous segment was identified in all metastases derived from the secondary trans¬ fectants, confirming the presence of the EJ-Ha-ras gene that had been present in the initially used recipient cells. No other Ha-ras genes had been acquired beyond those that were present in the recipient cells prior to these transfections. The positive reactivity with the PBR322 probe confirmed the presence of the introduced pSV2neo gene.

EXAMPLE 6 TESTING ENDONUCLEASE FRAGMENTS OF DNA FROM METASTASES FOR METASTATIC ABILITY

DNAs from two of the four metastases derived from the secondary transfectants were digested with the restriction endonucleases EcoRI, BamHI, and Hindlll. Digested DNAs were religated with genomic NIH 3T3 DNA. The resulting DNAs were then co¬ transfected with pSV2neo DNA into EJ-6-2-Bam-6^'a cells and placed under G418 selection. Cells selected were inoculated subcutaneously into NFS/NCr mice (10 cells/mouse) . Two to six weeks later, mice were examined for metastasis. Four out of 12 mice that formed tumors when injected with trans- fectants containing Hindlll cleaved DNA formed metastases; three out of 11 mice that formed primary tumors formed metastases when injected with trans¬ fectants contaning EcoRI-cleaved DNA; none of the twelve mice that formed primary tumors when injected with transfectants containing BamHI-cleaved DNA formed metastases. Thus, both the Hindlll and EcoRI-cleaved DNA segments were biologically active. Industrial Applicability

The invention described herein is useful in detecting and isolating discrete, transmissible mammalian genes associated with tumor metastasis. Detection and isolation of such genes makes it possible to develop probes for such genes, or mRNA transcribed from such genes, as well as the produc¬ tion of antibodies against proteins encoded by such genes. This allows sensitive diagnostic techniques for metastasis and potential antagonism of such genes, gene products, or cellular products or intermediates affected by such gene products.

Equivalents

Those skilled in the art will recognize or be able to ascertain, using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed within the scope of this invention.

Claims

1. A method of isolating a discrete, transmissi¬ ble, mammalian gene associated with tumor ^• . metastasis, comprising: a. forming a multiplicity of fragments of donor DNA from metastatic mammal¬ ian tumor cells to provide at least one fragment containing a discrete, transmissible, mammalian gene asso- ciated with metastasis; b. exposing non-metastasizing tumori¬ genic cells capable of phenotypically expressing the presence of a select¬ able marker to a selectable marker ^* and to said donor DNA fragments under conditions whereby some of said exposed cells will be recipients of said selectable marker and donor DNA fragments; c. culturing said exposed cells under conditions sufficient to allow phenotypic expression of said select¬ able marker in those cells which are recipients of said selectable marker and donor DNA fragments; d. selecting recipient cells which have phenotypically expressed the presence of said selectable marker; e. introducing said selected recipient cells into a mammalian host under conditions such that said recipient cells would not be metastatic without

5 the presence of an acquired, dis¬ crete, transmissible, mammalian gene associated with tumor metastasis; f. recovering DNA from metastatic cells of metastases formed in said host in

10 response to the introduction of said selected recipient cells; g. repeating steps (a)-(f) employing DNA recovered in step (f) in place of DNA from said metastatic cells in step

15 (a) until the DNA recovered in step

(f) contains donor DNA containing . essentially only said discrete, transmissible, mammalian gene associ¬ ated with metastasis and a donor

20 marker indicating that said fragment originated from the original donor DNA of step (a) ; and h. fragmenting DNA recovered finally in step (f) into fragments at least one

25 of which is a fragment containing essentially only said discrete, transmissible mammalian gene associ¬ ated with metastatis and a donor marker.

2. A method of Claim 1 wherein said discrete, transmissible mammalian gene associated with metastasis is a human gene.

3. A method of Claim 2 wherein said non-metasta- sizing transformed cells of step (b) are cotransfected with said selectable marker and donor DNA fragments.

4. A method of Claim 3 wherein said selectable marker comprises a DNA sequence necessary for the survival of cells under certain culturing conditions and said certain culturing condi¬ tions are employed in step (c) .

5. A method of Claim 4 wherein at least the final cotransfection of human donor DNA is made into non-human mammalian cells.

6. A method of Claim 5 wherein said donor marker comprises a DNA sequence present in human DNA.

7. A method of Claim 6 wherein said donor marker comprises an Alu sequence from human DNA.

8. A method of Claim 7 wherein the final trans¬ fection is a transfection of the discrete transmissible human gene associated with metastasis and the Alu marker into mouse fibroblast cells.

9. Isolated mammalian gene associated with tumor metastasis.

10. Isolated protein coded for by a mammalian gene associated with metastasis.

11. Antibody against the protein of Claim 10.

12. Antibody of Claim 11 wherein said antibody comprises monoclonal antibody.

13. A probe for a mammalian gene associated with metastasis.

14. A probe for mRNA transcribed from a mammalian gene associated with metastasis.

15. A probe comprising a labelled polynucleotide complementary to at least a portion of a mammalian gene associated with tumor meta- stasis.

16. A method of assessing the metastatic propensity of a tumor, comprising: a. obtaining cells from said tumor; b. testing said cells for the presence of a gene associated with metastasis of said tumor or for a product of said gene associated with metastasis of said tumor.

17. A method of Claim 16 employing a probe compris¬ ing a polynucleotide complementary to said gene or mRNA transcribed by said gene.

18. In a method of preventing metastasis of a tumor, the improvement of employing an antibody capable of neutralizing protein expressed by a mammalian gene associated with said metastasis.

19. In a method of Claim 18, the improvement wherein said antibody is a monoclonal antibody.

20. In a method of preventing metastasis of a tumor, the improvement of antagonizing a discrete, transmissible mammalian gene asso¬ ciated with said metastasis.

21. The improvement of Claim 20 wherein said antagonism is achieved by using a pharmaco¬ logical agent.

22. In a method of preventing metastasis of a tumor, the improvement of antagonizing a cellular product or intermediate affected by the gene product of a discrete, transmissible mammalian gene associated with metastasis of said tumor.

23. A method of screening for the presence of a discrete, transmissible, mammalian gene associ- ated with tumor metastasis, comprising: a. forming a multiplicity of fragments of donor DNA from metastatic mammal¬ ian tumor cells to provide at least one fragment containing a discrete,

5 transmissible, mammalian gene asso¬ ciated with metastasis; b. exposing non-metastasizing tumori¬ genic cells capable of phenotypically expressing the presence of a select-

10 able marker to said selectable marker and to said donor DNA fragments under conditions whereby some of said exposed cells will be recipients of said selectable marker and donor DNA

15 fragments; c. culturing said exposed cells under conditions sufficient to allow phenotypic expression of said select¬ able marker in those cells which are

20 recipients of said selectable marker and donor DNA fragments; d. selecting recipient cells which have phenotypically expressed the presence of said selectable marker;

25 e. introducing said selected recipient cells into a mammalian host under conditions such that said recipient cells would not be metastatic without the presence of an acquired, dis-

30 crete, transmissible, mammalian gene associated with tumor metastasis; and f. detecting the presence or absence of metastasis in said host in response to the introduction of said selected recipient cells.