CA1319631C - Method for the production of proteins by means of inducible expression systems in genetically modified eukaryotic host-cells multiplicated in-vivo - Google Patents

Abstract

ABSTRACT

In the production of proteins of biological interest by means of a stress inducible gene expression unit/eukaryotic host cell system, the transformed cell lines are multiplicated by tumour growing in immunodefficient warm-blooded animals, after which the multiplicated cells are cultured in vitro and subjected to stress, whereby expression occurs in high yield.
In vivo multiplication rates of 105-106 the innoculated quan-tity/2 weeks are reported without any loss of the latent inducible expression capacity.

Description

1319~

The present invention concerns the biochemical production of proteins of interest by the techniques of genetic engineering and, more especially, by inducibly expressing genes of interest in appropriate genetically modified eukaryotic host cells.
The use of and the advantages of inducively expressing genes under the control, for instance, of heat shock expre6sion elements for the production of ~ene-products o interest in a variety of cell types, have previously been described (see International Applications W0 87/00861 (P.
Bromley et al) and EP 118,393 ~P. Bromley et al): DREAN0 et al. Gene (1987) 49, 1-8).
Being applied to a eukaryotic expression system, this technique has the intrinsic advantages of having an efficient, general and highly inducible character. These factor~ are of considerable interest for the commercial production of proteins of biological and pharmaceutical interest, particularly where these proteins are complex, modified, unstable or are potentially toxic to the producer cell. However, their economical performance will rely to a large degree on the use of host/vector models able to en6ure a high level of production of fully competent proteins, a~oclated with low cost cell growth and protein purl~ication. For instance, the use o~ micro organisms as host-cell ~ystems has demonstrated that the ideal quality which ~hould be expected from such a system cannot be found ln a ~ingle micro organism species, and more appropriate expre~ion systems and vectors are needed. Moreover, most proteins of therapeutic interest are secreted proteins, which frequently require po~t-translational modifications for -activity and immunological specificity, such modifications possibly lacking when translation occurs in bacteria.
Hence, the development of genetically engineered eukaryotic cell lines harbouring recombinant genes constitutes the . r `- 1319631 most promising prospect, desplte the relatively primitive state of eukaryotic cell culture technology as compared to that of bacter~al systems.

However, although recent developments in tissue culture, employing chemically defined media, constitute an important breakthrough in the adaptation of tissue culture methods to industrial production, there is presently a need of lmproved culturing means for multiplication and mass culture of cells transformed with genetically engineered gene expression systems.

Some of the major problems associated with mass cultures are:
to obtaln a large number of cells (especially sterility, materials, qualified personnel, space); to reduce the cost of production, essentially due to manpower but also of culture media containing fetal calf serum (FCS can represent up to 80% of the total production cost); and to purify the expected proteins of lndugtrlal interest and partlcularly ln the case of secretable protelns to lsolate them from the whole suspenslon medla used for cell culture.

Thu5, ln an effort to develop the lnducible expression of genes of lnterest harboured by suitable host cells on an industrial scale, the present inventors have provlded a method for the blochemlcal productlon of proteins of lnterest wlth an inducible recomblnant gene expresslon unlt/host cell system, thls method comprlslng the followlng steps: ~1) Choslng a sultable cell llne of eukaryotlc orlgln, ~2) Genetically modlfying these cells by transformlng with recomblnant DNA contalnlng the followlng elements: (a) a gene codlng for a proteln of lnterest, thls gene belng under the control of a stress lnduclble transcrlptlon promoter of eukaryotic orlgln, ~b) an oncogene or an oncogenlc transformlng gene of cellular or vlral origin, ~3) Inoculating the transformed cells into selected laboratory warm-blooded anlmals capable of growing tumors ln response to lnoculatlon; ~4) Allowing the tumors to grow ln the anlmal until tumors of 10lcells or more are obtained, (5) Removing said tumors and effecting dissociation thereof, introducing the dissociated cells into in-vitro culture medium and sub~ecting them to stress to induce expression of the gene of interest; (6) Isolating the expressed protein.

Briefly stated, appropriate eukaryotic host cell lines capable of generating tumours when innoculated to immunodeficient animals are transfected with nucleic acid constructions containing at least one gene of interest under the control of a stress inducible promoter. With "normal~' cells, a transforming gene, e.g. an oncogene, is also added by co-transfection to impart to the cells unrestricted multiplication capacity. Then the transformed cells which are capable of expressing said gene of interest upon induction are innoculated to im-- 2a -~VI

1319~31 munodeficient animals whereby they are transiently multipli-cated in the form of tumours developed by the animals. Sur-prisingly, and despite the presence in the cells of transcrip-tion units not expressed under normal growth conditions, the rates of multiplication were very high (about 105-106 times the innoculated quantity in 1-2 weeks) without cell degeneracy or loss of the capacity of expressing the gene of interest under stress. This method is also particularly useful for obtaining rapidly and at relatively little expense identifica-tion and testing quantities of engineered protein variants of natural proteins.
It should be noted at this stage that the multiplication B f certain cell lines ~y tumoral transplantation in warm-blooded animals is not novel per se. ~or instaPncé,~ ~ -A-2,083,826 (HAYASHIBARA) discloses the multiplication of in-sulin prod~cing, human oncogenic cells by transplantation into warm-blooded animals such as nude mice or immunosuppressed mammals (rats, hamsters, etc.) and ~irds. However, the types of cell lines involved ~n this prior art only include easily pre~ervable stable cell lines such as insuloma, lung tissue carclnoma, lymphoblast or hybridoma cells. No corresponding multlpllcation of cells incorporating ~enes dominated by heat-shock elements has been previollsly reported, to the knowledge of the present inventors. Furthermore, the rapld in-vivo multipllcation~ of cells of the type involved here, i.e.
containing a recombinan~ stress-inducible expre3sion system, ln very high yield, without the production of undesirable inhibitory products and full preservation of heat-shock in-ducibility, constitutes a very important technical step vis-~-vis the previous techniques.
In short, the experiment~ whlch support the claimed me-thod include the following embodiment~:
Genes coding for products of intere~t (hormones, enzymes and other proteins of interest for diagnoses and pharmaceuti-cal applications) were placed into suitahle vector~ under the tranucriptional and translational control of heat-shock promo-ter elements of eukariotic origin (human, Drosophila, etc.).
Optionally, other precursor or post-translational sequences and variants thereof were also introduced lnto the nucleic acids of concern, all this being carried out by usual recombinant DNA
technlques. Most of this has been disclosed in wo 87/00861 ( Bromley et al.

Then, suitable host cells were transfected with recombinant DNA's lncluding a hybrid gene of interest under control of a hsp 70 sequence and an oncogenic transforming gene. When tumoral host-cells were used, no additional oncogene to promote multiplication was necessay. Contrastingly, when host cells such as NIH-3T3 were used, co-transfection was brought about using an oncogenic transformant sequence of cellular or viral origin . Then, after further ln vltro expansion and testing for expression of the gene under heat shock, the transformed cell lines were in~ected into lmmunodeficlent anlmals, for instance nude mlce or lmmunodepressed rats and the tumors were allowed to grow to about 109-101 cells.

The tumors were removed, minced and the cells were dissoclated, for lnstance trysinized, and placed into culture media whereby they could subsequently be stressed. The gene of lnterest drlven by a heat-shock control element, is induced and the cells expressed the protein of interest. The product was then collected by usual means, for example afflnlty chromatography.
Detalls on the various embodlments of this inventlon will be better understood with reference to the accompanying flgures, a 11st of whlch ls summarlzed below.

BRIEF DESCRIPTION OF THE ACCOMPANYING FIGURES (FIGS 1-8 and 11~13 are photographs) Flg. 1 shows a nude mouse carrylng a sub-cutaneous (s.c.) tumour produced by cell llne ~Cl.6) ln~ectlon.

~..'.

131963~

Fig. 2 shows a Sencar mouse (Swiss-derived) with an intramuscular (i.m.) tumour induced by transplantation of C1.6 tumour cells from nude mice.

Fig. 3 shows a cyclosporine-immunosuppressed rat with a - 4a -~!

131~31 s.c. tumour induced by transplantation of Cl.6 cells from a nude mouse.
Fig. 4 shows the histological appearance of a s.c.
tumour induced by C1.6 cells injection in nude mice.
Fig. 5 ~hows a high magnification of the s.c. tumour shown in fig. 4. Note the high incidence of mitotic index.
Fig. 6 shows a s.c. tumour induced by Cl.6 cells injections in nude mice. Note the vascular spaces.
Fig. 7 shows a pulmonary tumour (fibrosarcoma) induced by intratracheal instillation of HBN2 cells in a nude mouse.
Fig. 8 shows a high magnification of the intrapulmonary tumour shown in figure 7.
Fig. 9 represents schematically the construction of plasmid pl7HBN. Fig. 9a represents plasmid pl7hgH dhfr W0 87/00861 (P. Bromley et al) containing a hsp70-hGH hybrid gene and the mouse dihydrofolate reducta~e gene under the control o~ the SV40 early promotor sequence, and terminated by the SV40 terminator sequence.
Pig. 9b represents plasmid pV69 (Meneguzzi et al. 1984, EMB0 J. 3; 365-371) containing the neomycin resistance gene under the control of the herpes simplex thymidine kinase promotor and terminator, and the 5.4 kb bovine papilloma virus ~ubgenomic fragment.
Fig. 9c represents plasmid pl7hGH NE0 W0 87/00861 (P.
Bromley et al) containing the hsp70-hGH hybrid gene and the neomycin resi~tance gene under the control of the SV40 early promotor sequence and terminated by the SV40 terminator ~equence.
Fig. 9d represents plasmid phGH BPV which results from the ligation o~ a HindIII-BamHI fragment derived from pl7hGH
dhfr and a 7.5 Kb fragment resulting from the digestion of pV69 with HindIII and partial digestion with BamHI.

~., - 131963~

Fig. 9e represents plasmid pl7K NEO which results from the ligation of a 1.8 Kb HindIII-BamHI fragment derived from pV69 (the HindIII extremity has been filled-in with DNA
polymerase I (large fragment); and a 3.4 Kb fragment resulting from the digestion of pl7hGH NEO with PvuII and BamHI.
Fig. 9f represents plasmid pl7HBN which results from the ligation of a HindIII-BamHI fragment derived from pl7K NEO, - 5a -13i9631 and a 6.8 Kb fragment ~esulting from digestion of phGH BPV
with HindIII and partial digestion with BamHI.
Fig 10 is a key to the symbols used in Fig 9 to define the various DNA sequences involved. The approximate length ~in base-pairs) of such sequences in the plasmids schematized in Fig 9 are given in Table I, below.
Fi.g 11 shows Cl.6 tumour cells after dissociation and inoculation in a Petri dish incubated in serum supplemented medium.
Fig 12 shows Cl.6 tumour cells after dissociation and inoculation in a 1 liter biofermentor with incubation in serum free medium.
Fig 13 shows Cl.6 ~umour cells after dissociation and inoculation in a 2 liter biofermentor with incubation in serum-free medium in the presence of 3g/1 of cytodex 1 micro-carrlers.

1319~31 Identification of the sequences in Figs ga-9f and restriction sites.

Plasmid Se~uences :kb pl7hGH dhfr D:0.6; B:<0.1; 1:0.7;
C:0.7(terminator):0.4(promotor) ~:0.7; ~:2.3 pV69 L:O.5(promotor):0:3(terminator) J:l.0; M:5.4:A:2.2 pl7h~H NE0 ~:0.6; 8:<0.1; 1:0.7;
C:0.4(promotor):0.7(terminator);
J:1.5; A:2.8 phGH 8PV C:0.7~terminator); 1:0.7;
M:5.4; A:2.2 pl7K NE0 D;0.6; B:~0.1;
L:0.5~promotor):0.3~terminator);
J:l.0; A:2.8 pl7H8N D:0.6; B:<().l; 1:0.7 C:0.7~terminator);
L:0.5~promotor):0.3~terminator);
J:1.0; M:5.4; A:2.8 BI : ~amHI
BII : Bgl II
HIII : Hind III
PII : Pv~II
. , 1319~31 For harbouring the gene recombinant DNA constructions, suitable cell lines had to be selected. Cell lines (NIH3T3) producing high-levels of human growth hormone (hG~) under heat shock control, have been described by Dreano et al., (Gene, (1987) 49, 1-8). In brief, one transformed 3T3 mouse cell line clone 6 (Cl.6) can secrete 2 to 5 ~g hGH per lo6 cells over a 20 hour period following a 2 hours heat shock at 42C. ~his concentration is 1200 time the hGH concentration measured in the medium of non heat-treated Cl.6 cells. In addition, these cells can be utilized for repeated induction cyles and they were therefore highly suitable for use in the present invention.
We report below the results of transplantation prodedures o~ three genetically engineered mouse cells lines:
clone 6 (C1.6), HBN2 and Clone 18 (Cl.18). These three cell line~ expressed the human hsp70-hGH hybrid gene, and contain the following additional genes; in Cl.6, the human Harvey c-ra6 oncogene (Tabin et al. 1982, Nature, 300; 143-149); in HBN2 the bovine papilloma virus (BPV) subgenomic fragment, and a neomycln re~istance gene (used ~or G418 resistance ~olectlon), and in Cl.18 the neomycin resistance gene (Cl.18 ls used as a control). Obviously, other genes of interest can also be placed under the control o~ the hsp70 promotor sequence~ in similar DNA construction6 and expre~sed simllarly. 8uch genes o~ interest include for instance viral proteins, hormones, enzymes, blood proteins and others.
Regarding heat-shock control sequences, it is obvious that other heat-~hock control elements can be used in place o~ the seguences mentioned in the specific experiments illustrating the present invention. Genetically engineered variant~ obtained by nucleotide deletion, mutation and insertlon, as disclosed in WO 87/00861 (P. Bromley et al), are al~o suitable.

Transplantation of the two first cell lines into nude mice, induced tumour in injected animals. After proteolytic dissociation, or treatment with collagen the tumour cells were able to produce, after heat induction, the æame quantity 5 of hGH as the parental cells. Using this methodology we were able to obtain approximatively lolO tumour cells per tumour.
Finally we showed that these cells retrieved from tumour~ in animals, can be easily cultivated on Cytodex (a trademark) microcarrieres in 1 liter or 2 liter bioreactors and, in addition, in chemically defined media serum free medium (SFM), or serum and protein free medium (SPFM) (see CHESSEBEUF and PADIEU., in published PCT/841 901,363.6 and FR-A-83/04843). Repeated heat regulated production of hGH
was once again obtained.
1. Clone 6 The BALB NIH-3T3 cell~ (obtained from the ATCC
organization) were co-trans~ected with plasmid pl7hGHdhfr, ~ee ~ig. 9a, carrying the human growth hormone (hGH) driven by the human heat shock protein 70 (hsp70) promotor (plasmid pl7hGHdh~r i~ di~closed in WO 87/00861 (P. Bromley et al)), and by a plasmid carrying the human c-ras oncogene. ~wo week~ later, foci were i~olated, cultured and analyzed for hGH produ¢tion. Clone 6 was ~ound to secrete 3 ~g o~ hGH per 106 cell~ in a 15 hour period ~ollowing a single heat treatment o~ 2 hours at 42C.

1.1 Transplantation into nude mice Spontaneously transformed 3T3 cells injected into nude mice are able to induce tumour formation (Rubin and Arnstein, 1982, Cancer Re~., 300; 143-149, Rubin 1984, J. Natl. Cancer Inst., 72, 375-381).

_ g _ ', ~,.

l.l.1 Subcutaneous trans~lantation;

106 cells in 0.1 ml of culture medium without serum were injected subcutaneously in the dorsal area of nude mice. The first tumour was clearly visible 3 weeks after inoculation (fig. l) when the tumour reached a diameter of 20 mm (weighing about 3g), it was dissected. A part of this 1st generation tumour (designated Gl/n) was minced in culture medium and one ' "

- 9a -lo 131963~

portion was resuspended to i.nocul~te a rnonolayer culture for subsequent heat shock, and the other portion was reinjected su~cutaneously into other nude mice (to prod~ce G2/n), finally the remaininq part of the tumour was fixed with formaldehyde for histological examination. These tumour cells had a great capacity to anchor on plastic dishes. Indeedr 30 min. after seeding in complete medium most of the cells were attached, and presented a fibro.hlastic aspect. After one night, cells covered the dish and only few dead cel.ls were observed (fig.
11). After heat treatment of tumour cells (from Gl/n) in cell culture, they were shown to secrete 2,5 to 5 ~g of hGH per 106 cells; this is within the .same range of production as that of the original Cl.6 cells. In addition, tumour slices (0.5 to lg) were found, after heat treatment, to secrete 20 to 100 ~g of hGH in suspension med.i.um. Identical e~periments were per-formed in cells from tulaours passaged up to 8 times in nude mice ~G8/n). Table IT (see Experi.~ental part) shows that, with Cl.6 the heat requlated production is maintained at a constant rate durinq 24 generations in nude mice, and the a generations in sencar mice; with HF3N2, another clone to be described below, a slmilar result was also observe~ with cells from the first 2 generations in nude mice. After successive subcuta-neou~ pa~sages into nude mice, the t.i.me to tumour onset was reduced to only 2 week~ for Ihe production of a 30 mm tumour, thls w~en the tumour cel.l~ came frolll another nude mouse, compared to the 4 weeks needed to obtain tumour~ derived from cell cultures. The histological appe~rance (fig. 4 to 6) of the ~ubcutaneoùs tumours remained the same (a fihrosarcoma with a high mitotic index) duri.nq passaqe in nude mice. In the tumours up to a diameter of 30 mm produced in nude mice with Cl.6, no ma~or tumour necrosis was ohserve~. Moreover, the periphery of the tumour was well irriga~ed by blood capil-laires ~fig. 6).

1.1.2 IntraPeritoneal injection Cl.6 cells, derived from tissue culture (106 cells), were al50 in~ected intraperitoneally into a nude mouse. Three weeks 1319~31 after injection, when a swelling of the abdomen was observed, the mouse was sacrificed. At necrosy, tumours with a marked ascite were noted, and tumour metastases were found in the peritoneal lymph nodes and around the kidneys, as well as in the liver. Microscopical]y this tumour was similar to the tumours obtained by subcutaneous inject;ion of Cl.6 cells.

1.2 Transplantation into sencar mice The Sencar strain of mice (~wiss derived strain, as ex-plained in the experimental section), which is not genetically lmmunodeficient, was used as a host for genetically engineered transformed cell~, as a test of whether these cells can grow in normal mice, after previous passage in nude mice. An at-tempt to produce tumours in Sencar mice directly from cell cultures was unsuccessful. However, Cl.6 cells obtained after a first generation tumour in nude mice. (Gl/n), injected subcutaneously (s.c.) or intraperitoneally (i.p.) into Sencar mice, produced respectively s.c. and i.p. tumours in Sencar mice (denominated Gl/s/~c and Gl/8/ip fig. 2). This experlment was repeated wlth a 2nd generation tumoue in nude mice (G2/n) whlch, in~ected ~.c. into a Sencar mouse, also produced an s.c. tumour in gencar mice. A direct passage of Cl.6 tumours from Sencar to Senc~r was also successflll since an i.p. tumour ln ~encar mouse ~Gl/s/ip) wa~ transplan~ed into several other Sencar mlce by s.c., i.p. and intramuscular ~i.m.) routes, to produce the G2/s tumour~. I.m. and s.c tumours have been observed and Sencar to Sencar transplantations can be continu-ed. The histological structure of the tumour~ in Sencar mice is similar to that of those induced in the nude mice.

1.3 Tran~P].antation into_rats In order to verify if interspecies xenografts of geneti-cally engineered, transformed cells could be produced, tumours grown in nude mice (G4/n) were subcutaneously transplanted ~pieces of about 75 mm3) into three albino rats. One rat received cyclosporine for immunosuppression ~see Experimental 1319~1 part), the two other rats did not received the drug. Two weeks after transplantation, the tumour had regressed in the two normal rats, while a large tumour (12 cm diameter after 3 weeks) was developed in the cyclosporine-immunosuppressed rat ~fig. 3). At necropsy this tumour was found to consist of a large necrotic area surrounded ~y a layer of about 2 cm of living tumour cells (totalling about 40 g). Its histological structure was once again that of a fi~rosarcoma, similar to the tumours induced in mice previously. The heat-induced pro-duction of hGH from the cell of this tumour was similar to that from nude mice tumors. Immunosuppressed animals are inte-rest;Lng hosts for growing genetically engineered cells and, in addition to rat, other warm-blooded animals are also usable.

2 Clone HBN_Z

HBN2 is also a clone derived from the NIH-3T3 cell line, with incorporation of a plhsmid pl7HBN, whose construction is descrihed in fig. 9a-9f, containing 3 transcription units~
the hGH gene driven ~y the human hsp70 promotor, ~2) the neomycin res1stance gene under control of the herpes ~implex viru thymidine kinase promotor (~Jsed for cell selection via G,418 re~i~tance) and (~) a su~genomic fragment of bovine paplloma virus carryin~3 sequences responsible for the mainte-nance of multicopy pLasmids in a st~le form in mouse fibro-blasts ~Meneguzzi et al. l9B4, F~Bn J., 3, 365-371).

2.1 TransPlantation into nude mice Two types of transplantation of HBN2 cells from culture ~using 106 cells in 0.1 ml of culture medium without serum) were performed. In one case, the ce]ls were injected subcuta-neously in the footpad of nude mice. One month later, a growth was noted, this first generation tumour (H~NZ/Gl/n) was taken, minced with scissors, and tumour cells were lnjected into the back of nude mice to prnduce the second generation tumours (HBN2/G2/n).
In the second case, 106 HBNZ cells Ln 0.5 ml of medium 13 ~3~ 9631 without serum were introduced intra-tracheally into the lungs of a nude,mouse. When the animal was necropsied two months later, a large pulmorlary tumour (10 mm diameter) invading a pulmonary lobe was found. The tumour (HBN2/Gl/s/it) was asses-sed microscopically to be a sarcoma (fig. 7 and 8). Again, ,cells isolated from tumour derived from clone HBN2 were stable and provided the desired protein in high yield under heat induction in culture medium. (see Table III) 3 Clone 18 (control) NIH-3T3 cells were co-transEected with plasmid pl7hGH
dhfr together with a plasmid carrying the neomycin resistance gene under control of t;he early SV40 promotor (Southern and Berg 1982, J. Mol. Appl. Gen. 1; 327-341). Two weeks later G41~-resistant (200 ~g per ml of suspension medium) clones were isolated and expanded. Clone 18 was found to secrete approximately 1,5 ~g of hGH per 106 cells after heat treat-ment. Two nude mice were injected ~.c. in the foot paw as di~closed above. Three months later no tumour had grown at the in~ection site wh,Lch indicated that the transplantation was unsuccessful.

4. Tumour cells~E~L,on Petri d _he_ in serum free media Tumour cells (Cl.6) were dis~clat:ed by succes~ive tryp-~ln î;reatments; one portiorl from each trypsination step was inoculated on Petri dishes in three different ~ypes of media:
- serum supplemented medium ~SSM) constituted of Williams'E
~olution plus 5% of fetal calf serllm and 5% new born calf serum ~see fig.ll) - ~erum free medium (SFM) constit,uted of the Williams'E medium plu~ 4 g/l of bovine serum albumltl, Ind 7.fi ~mole /1 of a mixture of 6 free fatty acids (FF'A). These six FPA were found indispensable to insure the Krebs cycle functions providing the necessary acetyl-CoA needed in the abscence of serum lipids, 9 6 ~

-serum and protein free medium (SPFM) constituted of the ~illiams'E mediu~ and containing 50 mg/l of dextran and 7.6 ~mole/l of 6 FFA. After cultivation, the cells were su~jected to stress as usual. Results are summarized in Table II (see Experimental part).

The advantages associated with the use of SFM and SPFM are :

~1) Easy selection of a basal synthetic medium for optimal growth or cell function (2) Reduction of the cost by eleminating the use of fetal calf ~erum ~3) Use of a defined culture medium which eliminate the many known and unknown components of serum (4) Supplementation with known amo-mts of effectors and inhi-bitors ~5) Simplification of the extraction process especially for ~ecreted proteins such as h~H
~6) ~limination of the extrinsic effects of the serum ~7) Easy formulati.on and use of a selective media ~) Elimination of many contaminating su~stances bound to albumin (g) Continuation of fundamental research on phenotypic expre~-~ion of diploid cell line ~lO) Innovation ln the appli.ed research of new synthetic ba~al medlum which meets the WH~ and FDA requirements when normal or recombinant eukaryotic cells are cultured for the production of therapeutic or bioloyically acti.ve molecules for human administration. These advanta~e~ are taught by CHESE~EUF and PADI~U WOB4/03710). After reinoculation on Petri dishe~, Cl.6 had exactly the same behaviour i.n SFM anA in SPFM as in SSM, i.e. very rapid anchorage, recovering a fibroblastic pattern and complete colonization of the surface of the dish. Tumour cells were similar to the parental Cl.6 cells, without any vi~lble contaminating cells. I~erefore using SFM and SPFM
culture media for cultivating the cells expanded according ta the present method were technically and economically 131963~

advantageous.

Cl 6 Cells in a biofermentor __ __.

5,1 Biofermenta or with mi _ocarriers After each trypsinization step portion of about 3.109 cells were introduced into 3 biofermentators containing re-spectively 250ml of c,ne of each of the 3 different media, together with 3 to 5 g/l of cytodex microcarriers. 12 hours after inoculation, th~ microcarriers were fully covered with tumour cells. Trypan blue coloration indicated the presence of less than 3~ of dead cell,s in any medium (fig 13). Cell3 can be maintained in biofermentors for more than 23 days and are sub~ected, during this period, to several 4 hrs treatment at 42C. The production of hGH was noted usin~ an R]A (Radioim-mùne assay) procedure. Table III (see the experimental part) shows that cells secreted hGH after each heat treatments; the productlon in SSM was more important than in SFM (see fig 12) and even more so than in SPFM.
It is lnteresting to note that in the SSM biofermentor, about 3 mg of hGH was fiecreted into the medium. It should be n~ted that these cells were derived from less than 1 complete tumour, and that a heat treatment could be applled daily.

5.2 Riofermentator wit,hout microcarKiers Dissociated tumour cells can also be maintained alive in biofermentator without microcarriers although the conditions there are less favorable. For instance, it wa~ observed that the cells were ~till, alive 5 days after innoculation.

EXPERIMENTAL,DETAILS

Method, to construct P] 7HBN ~ f iq 9 ) Plasmid pBPV hGH : pl7hGH d'hfr (Dreano et al., ibid) was digested with HindIII , and therl partially with BglII. A 1.4 kb frag~en.t containing the human hsp70 promotor, the hGH qene and the SV40 termination signal was extracted from low melting agarose (Sigma, type VII). Plasmid V6g (Meneguzzi et al., ibid) was digested with HindIII and partially with BamHI, a 7.6 kb fragment containing plasmid vector sequences and the 5.4kb bovine papi]loma vi,rus (BPV) su~genomic fragment was isolated and ligated with the above fra(3ment.
Plasmid pl7K NE0: pV6g was digested with HindIII, treated with the DNA polymerase large fragment and then digested with BamHI. A 1.8 kb resulting fragment carrying the neomycin resistance gene controlled by the herpes simplex thymidine kinase transcription signals (promotor and terminator), was purified, and ligated to a PvuII BamHI purified fragment from pl7hGH NE0 containing the human hsp70 promotor and pBR322 derived sequences.
pl7HBN : pl7K NE0 was linearised by digestion with HindIII and BamHI. This fragment including the hsp70 promotor ant the neomgcin resistance transcription unit, was ligated with a 6.8 kb fragment from pBPV hGII (obtained by total di~es-tion wlth HindIII and partial digest3on with BamHI) containing the hGH gene, SV40 termination si~nals, and the BPV subgeono-mlc fragment.
The resulting plasmid pl7H~N contain3 three complete "tran~crlptlon units" :
- the human h~p70 -hCH hy~rid gene with SV40 terminator - the neomycln resistance gene with TK transcription slgnals - "the BPV subgenomic fragment"

2. Procedur,e,u,sed to_o,b~"ain cell_l_nes 2.1 Selecti,on via focus formation Cells were co-transfected with pl7hGH dhfr ~Dreano et al.
1986) and pEJ ~Tabin et al. 1982, Nature 300, 143-149) us~ng CaCl~ procedure (Graham and van der ~b, 1973, Vlrol. 52, 456-467). After two weeks, clones that were capa~le of focus formation were isolated, and expanded. 106 cells of each clone 131~631 were seeded into 20 cm2 dishes, and aft;er 24 hrs in culture were heat treated for 2 hours at 42C, and subsequently in-cubated at 37 C for 15 hours. Clone 6 produced 3 ~g of hGH
per 106 cells under t~ese conditions of heat shock.

2.2 G418 Selection Cells were co-transfected with pl7hGH dhfr and pSV2 NE0 (Southern and Berg., 1982 J. Mol. Appl.Genet. 1, 327-341), or were transfected with p].7HHN using the CaC12 procedure. After two weeks of culture in complete suspension medium, containing 200 mg of G413 ~GIBC0), visihle resultant clones were observ-ed after 2 weeks; they were i.solated and analyzed further as above. Cl.18 and HBN2 were found to inducibly secrete respec-tively 1.4 and 0.7 ~q of hGH per 106 cells.

3 ExPeriments usinq m_ e and rats 3.1 Genetically enqln e_ed cell transplantations into nude mi.ce Nude (nu/nu) mice aged 4 to 6 weeks, from Iffa-Credo ~L'Arbresle, France) were housed in Macrolon cages with a cover filter and placed in an air-filtered cabinet. They received an autoc].aved diet and normal tAp water ad lib.
Portion~ of 106 geneti.cally modified cells were trypsi-nated, rinced twice in ~ul,becco's modified Eagle's medium without serum and resusE)erlded in about 0.1 ml of the same medium for intramuscular (i.m.), intra perltoneal ~i.p.) and subcutaneous (s.c) injectlons, a.nd in 0.5 ml for intratracheal instillation.

~ 3 ~ Transe_a,,n,t,ation~f~r,~om n,U~e,,,,,,m,i,c,e t,n,-nu,de or to _sen,car ,mice Clone 6 cells, s.c. transplanted into nude mice, deve-loped s.c. tumours. Wherl a tumour reached a size of 20 mm in diameter ~about 4 weeks after inoculation), the mouse was 3~

sacrificed and the tumour tabout 3g) was dissected, minced in culture medium without serum, and a part of the minced tumour ~l/lO) was reinjected into nude or sencar mice.
Battelle-Ceneva ~reeds the Sencar strain of mice, and albino mouse derived from the Swiss mouse by selection for its sensiti~ity to skin carcinogens. (Sencar results from "Sensi-tive to Carcinogens"). Contrary to the nude mice, this strain of mouse is not genetically immurlodepriv~.

3.3 Transplantation into rats Pieces of abou~ 75mm3 of a Cl.6 tumour, which was deve-loped in nude mice (4th qeneration), were transplanted subcutaneously into Sprague Dawley albino rats (IFFA-CRED0, France). The transplantation was performed in 3 rats; 2 normal rats, and one rat chemically immunosuppressed, by ten daily s.c. in~ection of 60 mg/kg of cyclosporine (Sandoz, Bennet et al., Cancer Res., 45, 49fi3-4g69).

3.4 Hi~toloqY

When a tumour was taken for retransplantation, a portion was placed lnto a fixative solution (lO~ bufered formalin) then processed and sections were stained by haemalin-phloxin-saffron for microscopic observations as described in M. GABE
~l968)~ Techniques Histo1,ogiques Ed. Masson et Cie Paris.

4 Detailed method ,t,o transfer tumour cells from mice to ,pioferme,,ntor~ , 4.1 DescriPtion of the three _1ture medium The composition of the three culture media used in thi~
section are :

- serum ~upplemented medium (SSM) constituted of Williams E
medium ~Gibco) plus 5% of fetal calf serum (FCS) and 5% new lg 1319631 born calf serum (both from Boehringer).
- serum free medium (SFM) consisting of the ~ame medium, with the addition of 4 g/l of bovine serum albumin (BSA) fraction V
(Sigma)(equivalent to 10% FCS) and 7.6 ~ mole/l of a mixture of 6 long chain free fatty acids (FFA) in molar proportions clo~e to those of ra~ plasma (exc~pt increased cis-linolenic acid) i.e. palmitic acid 31.0%, cis-palmitoleic acid 2.8%, stearic acid 11.6%, cis-oleic acid 13.4%, cis-linoleic acid 35.6%, cis-linolenic acid 5.6% (Sigma).
- serum and protein free medium (SPFM) constituted of the same medium containing 50 mg/l of 2.10 6 dalton dextran (Pharmacia) which was as efficient as 4 g/l BSA to solubilize 7.6 ~mole/l of 6 FFA. Each medium contains 50~1g/ml of gentamicin (gental-line, Unilabo) as antibiotic.

4.2 Transfer of tumour cells from mice to biofermentors The mouse was sacrificed, the skin was cut and the tumour ea~ily extracted. After finely mincing with scissors, tumour fragment were transferred to a trypsinization bottle of 50 ml, provided with a magnetic ~tirrer. To this was added an equal volume of a ~olution diluted 3 times by trypsin rinsing medium 15250 (equlvalent to trypsl.n which hydrolyses 250 times its welght of ca~eln) o})t;ained from the MLcrohiological As~ociates Company, Bethesda, Mr~, U~A. ~t was stirred for 10 minutes at 100 rpm, at 37C, then the suspension of cells was decanted into "Ham F10 medium" cooled on ice (2 volumes of medium per volume of cells) and was centrifuged at 30 g. In parallel, the trypsinization procedure was renewed in the bottle. The se-quential digestion with the trypsin was repeated until the dissociation was complete, in general 10 to 13 tlmes. After each centrifugation step, cells were sus~ended in about 4.3 ml of the culture me~ium. 1/lO ~as ~ransferred onto a 20cm2 Petri dish (Falcon), and 9/10 into a 1 liter biofermentor (Techne, Cambridge, U.K.). After some trypsination steps trypsin so-lution was no longer diluted. In another procedure, comparab1e dissociation was ohtained using a collagenase solution (Boeh-ringer, 0.55g/1).

1 31963~

When the dissociation was c~mplete, suspension medium (SSM, SFP.or SPFM, qsp 250ml) cytoAex microcarriers (Pharmacia France, F7~340, 3 to 5 g/l) were added. Media are treated with C2 by bubbling a 5% C02/air mixture for about 10 min.
Finally, cells are incuhated at 37C overnight. Media are changed and biofermentors are placed in a 42C water bath for 3 or 4 hours followed Ijy a 20 hour period of incubation at 37C. Media are removed, and replaced with fresh medium ~efore heat treatment or not as described in Table II.

Table II : quantification of hGH proAuced by Cl.6 cells after pass~ge in nude or sencar mice. Tumours were retrieved from mice, minced fine]y with scissors, and treated with trypsin. Dissociated cells were seeded in 20cm2 Petri dish ~Falcon) with 106 cells in 5 ml of ~ulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. The next day cells were heat treated for 1 to 2 hours at 42C, or not treated ~controls). After 20 hours at 37C, hGH secreted in sample medium was quantlfied by means of a hGH RIA kit from Cambridge Inc., Mass. IJ.S.A. The number in the cell identifi-cation code indicate the tumor generation order (see sect~on 1.1.1) .

131 ~ 63~

Mice Cells Heat shock control , (hGH ~/20 ml/lO6cells _ Nude Cl-6 2 to S O.Ol Gl/n 2 to 5 ---G21n 1.3 0.15 G7/n 6 0.03 G8/n 2.5 0.03 G9/n 2.2 0.03 Gll/n l.4 0.02 Gl3tn 0.7 O.Ol Gl4/n l.0 O.~l Gl6/n 1.. 7 0 Gl8/n l.9 0.04 G20/n ~.9 o G22/n l.2 0.0002 G24/n 0.7 0.0005 Sencar G2/s l.l ~.Ol G2/s lO O.OB
~5t~ ~1.3 0.0~
G7/~ 4.5 O.Ol GB/~ 7.8 0.02 Ta~le III : hGH ~ecretion in transformed NIH 3T3 cells maintained in a biofermentor. Each biofermentor was seeded with 2.2 lO9 tumour cells in a volume of 250 ml of each of the three culture media described before containing 3 to 5 g/l of cytodex microcarriers. Cells were heat treated, some days after inoculation ~ee ~'irst columm in Ta~le,III), for 3 to 4 hour~ at 42C and kept at 37C 20 hours afterward~ and, final-ly, put in fresh media. '~e C02 concent,ration in the medium wa~ maintained b,y bubbling a 5% C02 in air mixture into the medium. Twelve hours after inoculation,, the microcarri.ers were found to be covered with tumour cell~. ~ trypan blue stain indicated the presence of less than 3% c~f dead cells in any of the media utilized.

~319~3~

In experiment 2, each biofermentor was seeded with 50.106 tumour cells in a volume of lO0 ml of the ~edia shown below containing 2 g/l of citodex microcarriers. The cells were heat treated during 2 h ~t 43C, kept at 37C for 24 h more, a~d finally put into fresh media. Sllspension media were tested for hGH presence by RIA (pr-)vided hy T~lsti~:u~ Pasteur Production), or ElA.
I~e cells were m~intained in Williall)s' E medium (~xperi-ment l) or in William's E/HAM F-10 (50:50, v/v) (experiment 2) supplemented with:

- SSM; 5% of fetal calf serllm ar~c~ 5~ Gf new born calf serum, - ~FM; 4 g/1 of ~ovine serum ~ umin and 7.6 ~mole/l of six FFA (fr~e fatty ac;~s) - SPFM; 50 mg/l. of dextran and 7.6 ~mol~/l of six FFA.

EXPE~IM~NT ]

Days of HS SSM SF`M SPFM
after ~nc~c~lation mg/l ~Ig/ltJfi m(~ g/lOhc m~ g/106c 1 0.35 ().0~ 0.44 0.05 0.70 0.08 Z 0.35 0.0~ ().2~ t~.()3 0.2~ 0.03 3 l.l 0.l3 0.53 ().h6 0.53 0.0~
7 2.3 0.26 l.9 h.22 0.79 0.09 g 2.9 0.33 ~.. 7 0.19 0.18 0.02 11 2.1 Q.24 1.9 0.22 0.79 0.09 1.3 4.2 0.~8 0.35 0.0 F.'XPr~.'RIM~NI' 2 1 0.25 0.5 0.35 0.7 0.25 0.5 2.95 5.~ 0.4S ().g 0.30 0.6 7 0.40 0.~ 0.6 1.2 2.25 ~.5 ].2 0.25 0.5 1.2 2.~ 3.65 7.3 1319~31 5. In another series of experiments, a DNA fragment coding for the hepatitis B virus ~urface antigen (HBsAG) was placed under the control of the human 70 kb heat shock protein promotor. The resulting plasmid construct (pl7MS neo carrying a neomycin reæistance selection gene) was used in trasfection experiments to establish a stable amnion cell line of human origin (Wish), expressing the HBsAg gene in a heat-re~ulated fashion. Post-translational modifications, such as assembly, glycosylation, secretion and production of both major and middle S proteins appeared to function normally. In addition, the production of HBsAg under various protocols of heat induction was found to be possible. After inoculation into nude mice, development of tumours was observed at the site of injection. Tumour cells, dispersed by means of collagenase or trypsin treatment from excised tumours and subsequently seeded into Petri dishes did secrete the ~ame quantities of HBsAg after heat induction as cells of the original cell line.

Re~ult~ on the production of the antigen by the tumoral cells are given in Table I~ below.

Table IV: HBsAg secreted by WB4 cells before and after passge a~ tumors into nude mice. Parental cells (around 106) were in~ected into the back area o~ nude mice; three weeks later the mice developed tumour6. The mice were sacrificed, the tumours were ~inely minced with sci~sor and tumour fragment~ were either injected into nude mice to produce the next generation of tumours, or sequentially treated with trypsin to dissociate tumour cells which were seeded in~Petri dishes. Tumour cells at a density of Io6 cells per 25 cm2 dish were heat treated (HS) or not ~Control) ~or 2 h at 43C
and po~t-incubated overnight at 37C. HBsAg was measured using an ELISA kit obtained from Abbot.

13~9~3~

Generation WB4 n~ HRSs~/106cells Control. ~IS

parental cells 0 80 Gl 0 100 G2 nd nd Complete detai].s on the construction of plasmi.d pl7MS neo are published in M. ~.EANO et al., Virus Research (1.9R7) ,8, 43-5~.

Claims (16)

1. A method for the biochemical production of proteins of interest with an inducible recombinant gene expression unit/host cell system, this method comprising the following steps:
(1) Choosing a suitable cell line of eukaryotic origin, (2) Genetically modifying these cells by transforming with recombinant DNA containing the following elements:
(a) a gene coding for a protein of interest, this gene being under the control of a stress inducible transcription promotor of eukaryotic origin, (b) an oncogene or an oncogenic transforming gene of cellular or viral origin, (3) Inoculating the transformed cells into selected laboratory warm-blooded animals capable of growing tumors in response to inoculation;
(4) Allowing the tumors to grow in the animal until tumors of 1010 cells or more are obtained, (5) Removing said tumors and effecting dissociation thereof, introducing the dissociated cels into in-vitro culture medium and subjecting them to stress to induce expression of the gene of interest;
(6) Isolating the expressed protein.

2. The method of claim 1, wherein the selected cell lines (1) are tumor cells, whereby element (b) is omitted.

3. The method of claim 1, wherein the transcription promotor (a) is a human 70 Kd heat-shock (hsp70) promoter fragment.

4. The method of claim 3, wherein the oncogene or transforming gene is selected from (i) the human c-ras oncogene and (ii) a subgenomic fragment of bovine papilloma virus.

5. the method of claim 3, wherein the gene of interest (a) is a human growth hormone (hGH) gene or a hepatitis B virus surface antigen (HBsAG).

6. The method of claim 3, wherein said recombinant DNA (2) is selected from (i) plasmid p17hGH dhfr carrying the hGH
gene driven by the hsp70 promotor; (ii) plasmid p17HBN
carrying the hGH gene under control of the hsp70 promotor, the neomycin resistance gene driven by the herpes simplex virus thymidine kinase promotor and the transformant fragment of bovine papilloma virus (clone HBN2).

7. The method of claim 3, wherein said transformation in step (2) is effected by co-transfection with plasmid p17hGH
dhfr and plasmid pEJ, the latter carrying the human c-ras oncogene (clone 6).

8. The method of Claim 2, wherein said transformation in step (2) is effected by co-transfection with plasmid p17hGH
dhfr and plasmid pSV2NEO, a plasmid carrying the neomycin resistance gene under control of early SV40 promotor.

9. The method of claim 1, wherein the selected cell lines is NIH-3T3.

10. The method of claim 3, wherein the selected animals include nude mice and chemically immunosuppressed animals.

11. The method of claim 3, including after step (4) the further intermediate step of:
(4b) removing the tumor and innoculating a portion thereof into immune animals such as SENCAR mice.

12. The method of claim 3, in which step (5) involves cultivation in culture media selected from (i) serum supplemented medium (SSM), (ii) serum free medium (SFM) supplemented with fatty acids and (iii) serum and protein free medium supplemented with fatty acids and dextran.
(SPFM)

13. The method of claim 12, including operating a fermentator in the presence of homogeneously distributed microcarriers (CYTODEX) or hollow fiber culture system.

14. The method of claim 13, wherein isolation of the protein of interest is effected by techniques selected from affinity chromatography, dialysis, ion exchange, and gel chromatography.

15. The method of claim 3, wherein said recombinant DNA (2) is a plasmid p17MS Neo carrying the HBsAg gene driven by the hsp70 promotor.

16. The method of claim 3, wherein the host cell lines are selected from NIH-3T3, and WISH (ATCC-CCL25).