EP0910654A1 - Enzyme combinations for destroying proliferative cells - Google Patents

Abstract

Enzyme combinations useful for destroying cells, particularly proliferative cells, are disclosed. Vectors enabling the intracellular expression and transfer of said enzyme combinations, as well as the therapeutical use thereof, particularly in anti-cancer gene therapy, are also disclosed.

Description

 Combinations of enzymes for the destruction of proliferative cells

The present invention relates to the field of gene and cell therapy. It relates in particular to combinations of enzymes which can be used for the destruction of cells, in particular proliferative cells. It also relates to vectors allowing the transfer and the intracellular expression of these combinations of enzymes, as well as their therapeutic use, in particular in anti-cancer gene therapy.

Gene therapy, which consists of introducing genetic information into an organism or cell, has experienced extraordinary development in recent years. The identification of genes involved in pathologies, the development of gene delivery vectors, the development of control or tissue-specific expression systems in particular have contributed to the development of these new therapeutic approaches. Thus, during the last 5 years, numerous clinical trials of gene or cell therapy have been undertaken in Europe as in the United States, in fields such as monogenic diseases (hemophilia, cystic fibrosis), cancer, cardiovascular pathologies or still nervous system disorders.

In the field of pathologies linked to cellular hyperproliferation (cancer, restenosis, etc.), different approaches have been developed. Some are based on the use of tumor suppressor genes (p53, Rb), others on the use of antisense directed against oncogenes (myc, Ras), still others on immunotherapy (administration of antigens tumor cells or specific immune cells, etc.). Another approach consists in introducing into the cells concerned a toxic or suicide gene capable of inducing the destruction of said cells. Such genes are for example genes capable of sensitize cells to a pharmaceutical agent. They are generally genes coding for non-mammalian and non-toxic enzymes which, when expressed in mammalian cells, transform a prodrug, initially slightly or not toxic, into a highly toxic agent. Such a prodrug activation mechanism is advantageous for several reasons: it makes it possible to optimize the therapeutic index by adjusting the concentration of prodrug or the expression of the enzyme, to interrupt the toxicity by no longer administering the prodrug and assess the mortality rate. In addition, the use of these suicide genes offers the advantage of not being specific for a particular type of tumor, but of general application. Thus, the strategies based on the use of tumor suppressor genes or anti-oncogenic antisense are applicable only to tumors having a deficiency in said suppressor gene or an overexpression of said oncogene. Likewise, approaches based on immunotherapy must be developed patient by patient, to take into account immune restrictions and skills. On the contrary, a strategy based on the use of a suicide gene is applicable to any type of tumor, and, more generally, to practically any type of cell.

Many suicide genes are described in the literature, for example the genes coding for cytosine deaminase, purine nucleoside phophorylase or thymidine kinase, such as, for example, thymidine kinases from the varicella virus or from the herpes simplex virus type 1. .

Escherichia coli cytosine deaminase is capable of catalyzing the deamination of cytosine to uracil. Cells that express the E. coli gene are therefore able to convert 5-fluorocytosine to 5-fluorouracil, which is a toxic metabolite (Mullen et al. 1992 Proc. Natl. Acad. USA 89 ρ33). The purine nucleoside phophorylase from Escherichia coli allows the conversion of non-toxic deoxyadenosine analogs to very toxic adenine analogs. As the eukaryotic enzyme does not exhibit this activity, if mammalian cells express the bacterial gene, deoxyadenine analogs such as 6-methyl-purine-2'-deoxyribonucleoside is transformed into a toxic product for these cells (Sorscher et al. . 1994 Gene Therapy 1 p233).

The thymidine kinase of the chickenpox virus allows the monophosphorylation of 6-methoxypurine arabinoside. If mammalian cells express this viral gene, this monophosphate is produced and then metabolized by cellular enzymes into a toxic compound (Huber et al.

1991 Proc. Natl. Acad. USA 88 p8039).

Among these genes, the gene coding for thymidine kinase (TK) is particularly interesting from a therapeutic point of view because, unlike other suicide genes, it generates an enzyme capable of specifically eliminating the cells being divided, since the prodrug is transformed into a non-diffusible product which inhibits DNA synthesis. The viral thymidine kinase, and in particular the varicella virus or herpes simplex virus type 1 thymidine kinases have a different substrate specificity from the cellular enzyme, and it has been shown that they are the target of guanosine analogs such as acyclovir or ganciclovir (Moolten 1986 Cancer Res. 46 p5276). Thus, ganciclovir is phosphorylated to ganciclovir monophosphate only when mammalian cells produce the enzyme HSV1-TK, then cellular kinases allow the metabolism of ganciclovir monophosphate to diphosphate then triphosphate which causes the stop of DNA synthesis and leads to cell death (Moolten 1986 Cancer Res. 46 p5276; Mullen 1994 Pharmac. Ther. 63 pl99). The same mechanism occurs with other thymidine kinases and other guanosine analogs.

Furthermore, a propagated toxicity effect ("by-stander" effect) was observed when using TK. This effect is manifested by the destruction not only of cells having incorporated the TK gene, but also of neighboring cells. The mechanism of this process can be explained in three ways: i) the formation of apoptotic vesicles which contain phosphorylated ganciclovir or thymidine kinase, originating from dead cells, then the phagocytosis of these vesicles by neighboring cells; ii) the passage of prodrug metabolized by thymidine kinase, by a process of metabolic cooperation from cells containing the suicide gene to cells not containing it and / or iii) an immune response linked to tumor regression (Marini et al . 1995 Gene Therapy 2 p655).

For those skilled in the art, the use of the suicide gene coding for the herpes virus thymidine kinase is very widely documented. In particular, the first in vivo studies in rats with a glioma show tumor regressions when the HSV1-TK gene is expressed and doses of 150 mg / kg of ganciclovir are injected [K. Culver et al. 1992 Science 256 pl550]. However, these doses are highly toxic in mice [T. Osaki et al. 1994 Cancer Research 54 p5258] and therefore totally proscribed in gene therapy in humans.

A number of therapeutic trials are also underway in humans, in which the TK gene is delivered to cells by means of various vectors such as in particular retroviral or adenoviral vectors. In human gene therapy clinical trials, these are doses much lower doses which should be administered, of the order of 5 mg / kg and for a short duration of treatment (14 days) (E. Oldfield et al. 1995 Human Gene Therapy 6 p55). For higher doses or longer treatments over time, undesirable side effects of toxicity are indeed observed.

To remedy these drawbacks, it has been proposed to synthesize more specific or more active thymidine kinase derivatives to phosphorylate guanosine analogs. Thus have been described derivatives obtained by site-directed mutagenesis. However, no precise biochemical characterization on pure enzymes has been carried out, no cell test using these mutants has been published and no functional improvement has been reported (WO95 / 30007; Black et al., 1993 Biochemistry 32 pll618 ). In addition, the inducible expression of an HSV1-TK gene, deleted from its first 45 codons, was carried out in eukaryotic cells, but the prodrug doses used remain comparable to those described in all the tests in the literature (B. Salomon et al. 1995 Mol. Cell. Biol. 15 p5322). Consequently, none of the variants described so far shows improved activity with respect to thymidine or with respect to ganciclovir.

The present invention provides an improved method of gene therapy by suicide gene. The present invention describes in particular a method making it possible to improve the efficiency of phosphorylation of guanosine analogues by thymidine kinase and thus, to improve the therapeutic potential of this treatment. The present invention provides in particular a method for triphosphorylating nucleoside analogs such as ganciclovir or acyclovir so that the triphosphorylation of these analogs is very significantly increased at doses of ganciclovir (resp. Acyclovir) i) significantly lower; ii) or likely to cause a more pronounced "bystander"effect; iii) or not leading to a cellular toxicity which could occur when wild thymidine kinase is overexpressed.

This method can be applied to cancer, to cardiovascular diseases, or to any application requiring the death of certain cells such as cells infected with a virus; this virus can be a virus of the type

HIV (human immunodeficiency virus), CMV (cytomegalovirus) VCR (respiratory syncytial virus).

The present invention is based in particular on the use of a combination of enzymes, making it possible to improve in vivo the phosphorylation reaction of nucleoside analogs.

A first subject of the invention therefore resides in a composition comprising:

- an enzyme capable of phosphorylating a nucleoside analog, to generate a monophosphate analog,

- an enzyme capable of phosphorylating said monophosphate analog, to generate a diphosphate analog, and,

- an enzyme capable of phosphorylating said diphosphate analog, to generate a toxic triphosphate analog.

More particularly, the enzyme capable of phosphorylating a nucleoside analog, to generate a monophosphate analog is a thymidine kinase, the enzyme capable of phosphorylating said monophosphate analog, to generate a diphosphate analog is a guanylate kinase and the enzyme capable of phosphorylating said diphosphate analog to generate a triphosphate analog is a nucleoside diphosphate kinase. Furthermore, the compositions according to the invention can comprise, not directly the enzyme, but a nucleic acid coding for the enzyme. In this regard, the subject of the invention is also a composition which can be used for the delivery and the in vivo production of a combination of enzymes, comprising:

. a first nucleic acid coding for an enzyme capable of phosphorylating a nucleoside analog, to generate a monophosphate analog,

a second nucleic acid coding for an enzyme capable of phosphorylating said monophosphate analogue, to generate a diphosphate analogue, and,

- A third nucleic acid encoding an enzyme capable of phosphorylating said diphosphate analog, to generate a toxic triphosphate analog.

Advantageously, the first nucleic acid codes for a thymidine kinase, the second nucleic acid codes for a guanylate kinase and the third nucleic acid codes for a nucleoside diphosphate kinase.

The nucleoside analog is generally a guanosine analog, such as, for example, ganciclovir, acyclovir or penciclovir. Other nucleoside analogs are, for example, compounds of the trifluorothymidine, 1- [2- deoxy, 2-fluoro, beta-D-arabino furanosyl] -5-iodouracil, ara-A, 1-beta-D- arabinofuranosyl thymidine type. (araT), 5-ethyl-2'deoxyuridine, iodouridine, AZT, AIU, dideoxycytidine, AraC and bromovinyldesoxyuridine (BVDU). The preferred analogs are ganciclovir, acyclovir, penciclovir and BVDU, preferably ganciclovir and acyclovir. The triphosphated form is toxic in the sense that it causes, directly or indirectly, cell death.

When mammalian cells, modified to express thymidine kinase (HSV1-TK for example), are put in the presence of a nucleoside analogue (ganciclovir for example), they become capable of carrying out the phosphorylation of ganciclovir to lead to ganciclovir monophosphate. Subsequently, cellular kinases allow the metabolism of this ganciclovir monophosphate successively into diphosphate then triphosphate. The ganciclovir triphosphate thus generated, then produces toxic effects by incorporating into the DNA and partially inhibits the cellular DNA polymerase alpha which causes the arrest of DNA synthesis and therefore leads to the death of the cell. . In this mechanism, the step of monophosphorylation of the nucleoside analog is considered as the limiting step. It is for this reason that various approaches have been described in the prior art in an attempt to improve the intrinsic properties of thymidine kinase (creation of TK mutants, search for more efficient expression and administration systems, etc).

The present invention now shows that it is possible to improve the effectiveness of treatment by administering, in combination with a thymidine kinase, other enzymes involved in the phosphorylation of nucleoside analogs. The present invention thus relates to different combinations of enzymes making it possible to optimize the intracellular reaction of triphosphorylation of nucleoside analogs. Another aspect of the present invention relates to vectors for the introduction and intracellular expression of these combinations of enzymes. It may in particular be several vectors each allowing the production of an enzyme, or one or more vectors each allowing the production of several enzymes or all enzymes. The present invention also relates to a process for triphosphorylation of nucleoside analogs in the presence of a combination of enzymes, possibly produced in situ by expression of corresponding genes, as well as a process for the destruction of proliferative cells.

The phosphorylation of nucleoside monophosphates into nucleoside diphosphates and then triphosphates has been documented in vitro. These phosphorylations are carried out in the presence of i) human erythrocyte lysate in the case of ganciclovir (Cheng et al. 1983 J. Biol. Chem. 258 pl2460) or ii) enzymatic preparations of guanylate kinase and nucleoside diphosphate kinase human erythrocytes, in the case of ganciclovir and acyclovir (Miller et al. 1980 J. Biol. Chem. 255 p720; Smée et al. 1985 Biochem. Pharmac. 34 pl049). Although phosphorylation of nucleoside monophosphates into diphosphates and then triphosphates has been demonstrated in mammalian cells, these conversions do not seem to be complete with ganciclovir monophosphate or acyclovir monophosphate (Agbaria et al. 1994 Mol. Pharmacol. 45 p777; Caruso et al. 1995 Virology 206 p495; Salomon et al. 1995 Mol. Cell. Biol. 15 p5322).

The present application now describes compositions making it possible to improve the therapeutic efficacy of a thymidine kinase in vivo.

The first enzyme used in the compositions and methods according to the invention, capable of phosphorylating a nucleoside analog to generate a monophosphate analog, is advantageously a non-thymidine kinase mammal. It is preferably a thymidine kinase of viral origin, and in particular herpetic. Among the herpetic thymidine kinases that may be mentioned in particular are herpes simplex virus type 1 thymidine kinase (HSV1-TK), herpes simplex virus type 2 thymidine kinase (HSV2-TK), thymidine kinase chickenpox virus (VZV-TK), thymidine kinase from Eppstein Barr virus (EBV-TK), or thymidine kinase from herpes viruses of bovine origin (Mittal et al., J. Virol 70 (1989 ) 2901), equine Robertson et al., NAR 16 (1988) 11303), feline (Nunberg et al., J. Virol. 63 (1989) 3240) or simian (Otsuka et al., Virology 135 (1984) 316) .

The sequence of the gene coding for the thymidine kinase enzyme of the herpes simplex virus type 1 has been described in the literature (see in particular McKnight 1980 Nucl. Acids Res. 8 p5931). There are natural variants leading to proteins with comparable enzymatic activity on thymidine, or ganciclovir (M. Michael et al. 1995 Biochem. Biophys. Res. Commun 209 p966). The sequence of the gene coding for the herpes simplex virus type 2 thymidine kinase enzyme has also been described (Swain et al., J. Virol. 46 (1983) 1045).

The thymidine kinase used in the present invention can be a native thymidine kinase (the natural form of the enzyme or one of its natural variants), or a derivative form, that is to say resulting from structural modification (s) of the native form. As indicated above, various mutants or derivatives have been described in the literature. Although their intrinsic properties do not seem to be significantly modified, these molecules can be used in the context of the present invention. They are for example mutants having a modification close to the DRH region of the interaction site of a nucleoside (W 095/30007). The HRD region corresponds to the aspartic residues, arginine and histidine at positions 163, 164 and 165 of the TK. These three positions are very conserved between herpetic TK. Different mutants have been described in positions 160-162 and 168-170 (WO95 / 30007). Other artificial variants have a modification at the level of the ATP binding site (FR96 01603). Furthermore, other TK derivatives can be prepared according to conventional techniques of molecular biology, and used in the combinations of the invention. These mutants can be prepared, for example, by mutagenesis on a nucleic acid coding for a native thymidine kinase, preferably herpetic, or one of its variants. Numerous methods making it possible to carry out site-directed or random mutagenesis are known to those skilled in the art and mention may be made of site-directed mutagenesis by PCR or by oligonucleotide, random mutagenesis in vitro by chemical agents such as for example hydroy lamine or in vivo in imitating E. coli strains (Miller "A short course in bacterial genetics", Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1992). The sequences thus mutated are then expressed in a cellular or acellular system and the expression product is tested for the presence of an activity of thymidine kinase type, under the conditions described in particular in the examples. Any enzyme resulting from this process, having the ability to phosphorylate a nucleoside analog, to generate a monophosphate analog, can be used in the present invention.

Preferably, in the context of the present invention, a TK originating from the thymidine kinase of the herpes simplex virus type 1 (HSV1-TK) or a corresponding coding nucleic acid is used. It is more particularly HSV1-TK or one of its variants, such as natural variants or artificial variants. Among the artificial variants, there may be mentioned more particularly the variants P155A / F161V and F161I (Biochemistry 32 (1993) p.11618), the A168S variant (Prot. Engine. 7 (1994) p.83) or the variants having a modification at the ATP binding site such as the M60I variant. Even more preferably, a nucleic acid coding for the thymidine kinase of the herpes simplex virus type 1 (HSV1-TK) is used.

The second enzyme used in the compositions and methods according to the invention, capable of phosphorylating a nucleoside analog monophosphate to generate a diphosphate analog, is advantageously a guanylate kinase. Guanylate kinase (GMPK) has been purified from different organisms (man, rat, beef, yeast). The gene coding for GMPK has also been cloned into various cell types, and in particular into the yeast Saccharomyces cerevisae. GUK1 gene (M. Konrad 1992 J. Biol. Chem. 267 p25652). From this gene, the 20 kDa GMPK enzyme was also purified. The GMPK gene, designated gmk, has also been isolated and overexpressed in Escherichia coli (D. Gentry et al. 1993 J. Biol. Chem. 268 pi 4316). This enzyme is different from the enzyme of S. cerevisae in terms of cooperativity and oligomerization, while the sequences have strong regions of identity 46.2% on 182 residues. The sequence Al 1042, identified as coding for a factor having hematopoietic cell growth potential activity (EP0274560), has 51.9% identity over 180 residues with the GUK1 gene of S. cerevisae. and appears to be the human counterpart to GMPK, although no biochemical data have been published.

The third enzyme used in the compositions and methods according to the invention, capable of phosphorylating a nucleoside diphosphate analog to generate a triphosphate analog, is advantageously a nucleoside diphosphate kinase. Nucleoside diphosphate kinase (NDPK) is a broad substrate specific enzyme and has been purified from a wide variety of sources (M. Inouye et al. 1991 Gene 105 p31). For various organisms (Myxococcus xanthus, Drosophila melanogaster, Dictyostelium discoideum, rat, beef, man, E. coli and S. cerevisae) the gene coding for NDPK has been cloned and the corresponding enzymes are very homologous (K. Watanabe et al. 1993 Gene 29 pl41). However, only the enzymes of higher eukaryotes have a "leucine zipper" sequence. The human genes described coding for an NDPK activity are in particular nm23-Hl and nm23-H2. It is suggested that the nm23-H2 gene codes for a bifunctional protein with two independent functions which are NDPK activity and transcription factor (E. Postel et al. 1994 J. Biol. Chem. 269 p8627). The YNK gene of S-cerevisae is not an essential gene for yeast and codes for NDPK which is probably a tetrameric protein whose molecular weight of the monomers is 19 kDa (A. Jong et al. 1991 Arch. Biochem. Biophys. 291 p241).

The nucleic sequences coding for GMPK or NDPK used in the context of the invention can be of human, animal, viral, synthetic or semi-synthetic origin.

In general, the nucleic acid sequences of the invention can be prepared according to any technique known to those skilled in the art. As an illustration of these techniques, we can notably mention:

- chemical synthesis, using the sequences described in the literature and for example a nucleic acid synthesizer,

- screening of banks using specific probes, in particular as described in the literature, or - mixed techniques including chemical modification (elongation, deletion, substitution, etc.) of sequences screened from libraries.

Advantageously, the nucleic acid sequences used in the context of the invention are cDNA or gDNA sequences. The cDNA sequences are sequences lacking introns, obtained from RNA. GDNA sequences are regions of chromosome. In eukaryotes, they include one or more introns. The gDNA sequences used in the context of the invention may comprise all or part of the introns present in the natural gene, or one or more introns artificially introduced into a cDNA to increase, for example, the efficiency of expression in mammalian cells. . The nucleic acids can code for native enzymes or for variants or derivatives having the same type of activity. These analog nucleic acids can be obtained by conventional techniques of molecular biology, well known to those skilled in the art. It can be mutagenesis, directed or not, hybridization from libraries, deletion or insertion, the construction of hybrid molecules, etc. Generally, the modifications relate to less than 20% of the bases of the nucleic acid. The functionality of analog nucleic acids is determined as described in the examples by assaying the enzymatic activity of the expression product.

A particular composition within the meaning of the invention comprises a first nucleic acid coding for a thymidine kinase and a second nucleic acid coding for a nucleoside diphosphate kinase. In this embodiment, the nucleoside diphosphate kinase is preferably of non-human eukaryotic origin. By non-human enzyme is meant an enzyme not naturally present in human cells. It can be a viral, animal enzyme, or from a lower eukaryotic organism (such as a yeast). It can also be an unnatural derivative of a human enzyme, having one or more structural modifications. More preferably, it the NDPK used in the present invention is chosen from the yeast or beef NDPK. These compositions can also comprise a nucleic acid coding for a guanylate kinase, such as for example a yeast guanylate kinase.

Another particular composition within the meaning of the invention comprises a first nucleic acid coding for a thymidine kinase and a second nucleic acid coding for a non-human guanylate kinase. The non-human GMPK can be chosen from rat, beouf, yeast, bacteria GMPKs or derivatives thereof. Preferably, GMPK comes from a lower eukaryote, in particular from yeast.

According to a particularly advantageous embodiment, the nucleoside diphosphate kinase used in the context of the present invention is of eukaryotic or animal origin. Even more preferably, it is a yeast or beef nucleoside diphosphate kinase. The Applicant has indeed demonstrated that, surprisingly, the yeast nucleoside diphosphate kinase, and in particular that of Saccharomyces cerevisae or that of beef, makes it possible to phosphorylate analogs of nucleoside diphosphates, such as ganciclovir diphosphate or acyclovir diphosphate , into nucleoside triphosphates. In addition, the results present in the examples clearly show that these enzymes have on these substrates an activity much superior to the human enzyme. Thus, in the presence of 0.675 μg of human enzyme, the percentage of GCV triphosphate obtained is 1.5%, while in the presence of 0.75 μg of yeast enzyme, it is 82.9%. Similarly, in the presence of 6.75 μg of human enzyme, the percentage of GCV triphosphate obtained is 24%, while in the presence of 1.5 μg of yeast enzyme, it is 91.1% and in the presence of 5 μg of beef enzyme, it is 92%. The same results are obtained with another nucleoside analog, acyclovir. Thus, in the presence of 6.75 μg of human enzyme, the percentage of ACV triphosphate obtained is less than 0.4%, while in the presence of 1.5 μg of yeast enzyme, it is 8%, in presence of 15 μg of yeast enzyme, it is 81% and in the presence of 5 μg of beef enzyme, it is 1.3%. These results clearly demonstrate the advantage of using, in the combinations according to the invention, a yeast or beef nucleoside diphosphate kinase. These results also show that the first stage of phosphorylation of the analog into monophosphate is not necessarily the limiting stage of the process and that the use of a combination of enzymes according to the invention makes it possible to increase the therapeutic potential. of treatment.

Furthermore, the applicant has also shown that the yeast guanylate kinase, and in particular of Saccharomyces cerevisae. also made it possible to phosphorylate analogs of nucleoside monophosphates, such as ganciclovir monophosphate or acyclovir monophosphate, into nucleoside diphosphates with good activity. Thus, in the presence of 2.5 μg of yeast enzyme, the percentage of GCV diphosphate obtained can exceed 92% and in the presence of 74 μg of yeast enzyme, the percentage of ACV diphosphate obtained is 54%. In addition, the present results show that the yeast guanylate kinase has a GCVMP phosphorylation rate 2 times greater than the human enzyme. Likewise, its affinity for GCVMP is greater by a factor at least equal to 2 than the affinity that the human enzyme has for this substrate. In total, the value of Vmax / Km of the yeast enzyme for the GCVMP is 4.4 times greater than the value of the enzyme human. For ACVMP, the value of Vmax / Km of the yeast enzyme is 7 to 9 times greater than the value that the human enzyme has for this substrate.

The Applicant has also demonstrated that coupling these two non-human enzymes with a thymidine kinase, for example the herpes virus type I thymidine kinase made it possible to phosphorylate nucleoside analogs such as ganciclovir or acyclovir in triphosphate derivatives, with very high efficiency.

According to a preferred embodiment, the compositions of the invention comprise sequences coding for yeast guanylate kinase (EC-2.7.4.8) and / or the nucleoside diphosphate kinase (EC-2.7.4.6). More preferably, these are the enzymes of the yeast S. cerevisae. These sequences are used simultaneously with a sequence (HSV1-TK) coding for the thymidine kinase of the herpes simplex virus type 1 (EC-2.7.1.21) to allow triphosphorylating the nucleoside analogs such as ganciclovir or the acyclovir.

As indicated above, the compositions according to the invention may comprise a combination of enzymes or nucleic acids allowing the production in vivo of the enzymes. They are advantageously nucleic acids. This mode of implementation is preferred since it allows in vivo production of higher levels of enzymes and thus a greater therapeutic effect.

According to a first embodiment, in the compositions of the invention, the nucleic acids are carried by the same expression vector. This embodiment is particularly advantageous since it suffices to introduce a single vector into a mammalian cell to obtain the desired therapeutic effect. In this embodiment, the different nucleic acids can constitute three distinct expression cassettes within the same expression vector. Thus, the different nucleic acids can each be placed under the control of a transcriptional promoter, a transcriptional terminator and distinct translation signals. It is also possible to insert several nucleic acids in the form of a polycistron, the expression of which is directed by a single promoter and a single transcriptional terminator. This can be achieved in particular by the use of IRES ("Internai Ribosome Entry Site") sequences positioned between the nucleic sequences. In this regard, the expression vectors of the invention may comprise a bicistronic unit directing the expression of two nucleic acids, and optionally a separate nucleic acid coding for the third enzyme. The vectors of the invention can also comprise a tricistronic unit directing the expression of the three nucleic acids. These different embodiments are illustrated in the examples.

Preferred expression vectors within the meaning of the invention are in particular:

- A vector comprising:

. a first nucleic acid coding for a thymidine kinase, and,

. a second nucleic acid encoding a non-human guanylate kinase. It is preferably a yeast guanylate kinase.

- A vector comprising:

. a first nucleic acid coding for a thymidine kinase, and,

. a second nucleic acid encoding a nucleoside diphosphate kinase. It is preferably a nucleoside diphosphate kinase non-human eukaryote. More preferably, it is a NDPK of bovine or yeast origin.

Advantageously, this vector also comprises a nucleic acid coding for a guanylate kinase.

The thymidine kinase used in the vectors of the invention is advantageously a thymidine kinase of viral origin, in particular herpes. It is preferably a thymidine kinase originating from the TK of the HSV-1 or HSV-2 virus.

As indicated above, in the vectors according to the invention, the different nucleic acids can be placed under the control of distinct promoters, or constitute a polycistronic unit under the control of a single promoter. In this regard, as indicated above, the enzymes can also be produced in coupled form, the different nucleic acids being coupled to produce a protein carrying the different enzymatic activities. In particular, a particular embodiment of the vectors according to the invention is characterized in that the nucleic acid coding for thymidine kinase of viral origin and the nucleic acid coding for non-human guanylate kinase are coupled and code for a protein carrying both TK and GUK activities. According to another variant, in the vectors of the invention, the nucleic acid coding for the thymidine kinase of viral origin and the nucleic acid coding for the nucleoside diphosphate kinase are coupled and code for a protein carrying both the activities TK and NDPK. By way of illustration, the coupling between the enzymes is carried out by means of a peptide linker, for example of structure (G ₄ S), According to another embodiment, in the compositions of the invention, the nucleic acids are carried by several expression vectors.

As indicated below, the expression vectors can be of plasmid or viral origin. As they are vectors of viral origin, they are advantageously retroviruses or adenoviruses.

Different promoters can be used in the context of the invention. These are sequences allowing the expression of a nucleic acid in a mammalian cell. The promoter is advantageously chosen from functional promoters in human cells. More preferably, it is a promoter allowing the expression of a nucleic acid sequence in a hyperproliferative cell (cancerous, restenosis, etc.). In this regard, different promoters can be used. It may for example be the own promoter of the gene considered (TK, GMPK, NDPK). They can also be regions of different origin (responsible for the expression of other proteins, or even synthetic). It can thus be any promoter or derived sequence stimulating or repressing the transcription of a gene in a specific way or not, inducible or not, strong or weak. Mention may in particular be made of the promoter sequences of eukaryotic or viral genes. For example, they may be promoter sequences originating from the genome of the target cell. Among the eukaryotic promoters, ubiquitous promoters can be used in particular (promoter of the HPRT, PGK genes, a-actin, tubulin, DHFR, etc.), promoters of intermediate filaments (promoter of the GFAP genes, desmin, vimentin, neurofilaments, keratin, etc), promoters of therapeutic genes (for example the promoter of the MDR, CFTR genes, Factor VIII, ApoAI, etc.), tissue-specific promoters (promoter of the pyruvate kinase gene, villin, intestinal fatty acid binding protein, alpha -actin from smooth muscle, etc.), promoters of specific cells of the dividing cell type such as cancer cells or of promoters responding to a stimulus (steroid hormone receptor, retinoic acid receptor, glucocorticoid receptor, etc.) or so-called inducible. Likewise, they may be promoter sequences originating from the genome of a virus, such as for example the promoters of the El A and MLP genes of adenovirus, the early promoter of CMV, or also the promoter of the RSV LTR, etc. In addition, these promoter regions can be modified by adding activation or regulatory sequences, or allowing tissue-specific or majority expression.

Another object of the invention relates to a product comprising a combination of enzymes capable of tri phosphorylating a guanosine analog, and said guanosine analog, for simultaneous, separate or spread over time administration.

The invention also relates to a composition for the in vivo production of a toxic nucleoside triphosphate analog comprising, packaged separately or together:

- a nucleoside analogue

- a nucleic acid encoding a thymidine kinase

- a nucleic acid coding for a guanylate kinase, and, - a nucleic acid coding for a nucleoside diphosphate kinase.

The invention also relates to a composition comprising a combination of enzymes involved in the phosphorylation of nucleosides, possibly generated in situ by expression of nucleic sequences, at least one of these enzymes being of non-human eukaryotic origin. The enzyme combination includes in particular a TK and an NDPK; a TK and a GMPK or a TK, a GMPK and an NDPK.

The present invention also relates to a method for the destruction of proliferative cells comprising the administration to said cells of a combination of enzymes comprising a TK and an NDPK. Preferably, the combination further comprises a guanylate kinase. The invention also relates to a method for the destruction of proliferative cells comprising administering to said cells a combination of enzymes comprising a TK and a GMPK.

According to this method, the cells are contacted with a nucleoside analog, preferably a guanosine analog, which is converted in cells expressing the combination of enzymes into a toxic compound.

According to the invention, the enzymes can be administered to the cells by administration of nucleic acids coding for said enzymes.

The invention also resides in the use of the nucleoside diphosphate kinase or of a nucleic acid coding for it, in combination with a thymidine kinase or a nucleic acid coding for a thymidine kinase, for the preparation of a pharmaceutical composition. intended for the destruction of proliferative cells.

The invention also relates to a method of triphosphorylation of a nucleoside analog comprising bringing said analog into contact with a combination of enzymes, at least one of which being of non-human eukaryotic origin. The present invention now provides new therapeutic agents for interfering with many cellular dysfunctions. For this purpose, the nucleic acids or cassettes according to the invention can be injected as such at the site to be treated, or incubated directly with the cells to be destroyed or treated. It has in fact been described that naked nucleic acids can penetrate cells without any particular vector. Nevertheless, it is preferred in the context of the present invention to use an administration vector, making it possible to improve (i) the efficiency of cell penetration, (ii) targeting (iii) extra- and intracellular stability. In a particularly preferred embodiment of the present invention, the nucleic sequences are incorporated into a transfer vector. The vector used can be of chemical, plasmid or viral origin.

By chemical vector is meant within the meaning of the invention, any non-viral agent capable of promoting the transfer and expression of nucleic sequences in eukaryotic cells. These chemical or biochemical vectors, synthetic or natural, represent an interesting alternative to natural viruses in particular for reasons of convenience, safety and also by the absence of theoretical limit as regards the size of the DNA to be transfected. These synthetic vectors have two main functions, to compact the nucleic acid to be transfected and to promote its cellular fixation as well as its passage through the plasma membrane and, where appropriate, the two nuclear membranes. To compensate for the polyanionic nature of nucleic acids, the non-viral vectors all have polycationic charges. As a representative of this type of non-viral transfection techniques, currently developed for the introduction of genetic information, one can thus mention those involving DNA and DEAE-dextran complexes. (Pagano et al., J. Virol. 1 (1967) 891), DNA and nuclear proteins (Kaneda et al., Science 243 (1989) 375), DNA and lipids (Felgner et al., PNAS 84 (1987) 7413), the use of liposomes (Fraley et al., J. Biol. Chem. 255 (1980) 10431), etc.

The use of viruses as vectors for gene transfer has emerged as a promising alternative to these physical transfection techniques. In this regard, different viruses have been tested for their ability to infect certain cell populations. In particular, retroviruses (RSV, HMS, MMS, etc.), the HSV virus, adeno-associated viruses, and adenoviruses.

The nucleic acid or vector used in the present invention can be formulated for topical, oral, parenteral, intranasal, intravenous, intramuscular, subcutaneous, intraocular, transdermal, etc. administration. Preferably, the nucleic acid sequence or the vector is used in an injectable form. It can therefore be mixed with any pharmaceutically acceptable vehicle for an injectable formulation, in particular for a direct injection at the site to be treated. They may in particular be sterile, isotonic solutions, or dry compositions, in particular lyophilized, which, by addition as appropriate of sterilized water or physiological saline, allow the constitution of injectable solutes. Direct injection of the nucleic acid sequence into the patient's tumor is advantageous because it allows the therapeutic effect to be concentrated in the affected tissues. The doses of nucleic acid sequences used can be adapted as a function of various parameters, and in particular as a function of the vector, the mode of administration used, the pathology concerned or even the duration of the treatment sought.

The invention also relates to any pharmaceutical composition comprising a combination of enzymes as defined above. It also relates to any pharmaceutical composition comprising at least one vector as defined above.

It also relates to the use of an NDPK of yeast origin or of a GMPK of yeast origin for the in vivo phosphorylation of nucleoside analogs.

Due to their antiproliferative properties, the pharmaceutical compositions according to the invention are very particularly suitable for the treatment of hyperproliferative disorders, such as in particular cancers and restenosis. The present invention thus provides a particularly effective method for the destruction of cells, in particular hyperproliferative cells. It is thus applicable to the destruction of tumor cells or smooth muscle cells of the vascular wall (restenosis). It is particularly suitable for the treatment of cancers. By way of example, mention may be made of colon adenocarcinomas, thyroid cancer, lung carcinoma, myeloid leukemia, colorectal cancer, breast cancer, lung cancer, gastric cancer, cancer esophagus, B lymphomas, ovarian cancer, bladder cancer, glioblastoma, hepatocarcinoma, bone, skin, pancreatic cancer or kidney and prostate cancer, cancer esophagus, laryngeal cancers, head and neck cancers, HPV positive anogenital cancers, EBV positive nasopharyngeal cancers, etc.

It can be used in vitro or ex vivo. Ex vivo, it essentially consists in incubating the cells in the presence of a nucleic sequence (or of a vector, or cassette or directly of the derivative). In vivo, it consists in administering to the organism an active quantity of a vector (or of a cassette) according to the invention, preferably directly at the site to be treated (tumor in particular), before, simultaneously and / or after the injection of the prodrug under consideration, ie ganciclovir or a nucleoside analog. In this regard, the subject of the invention is also a method of destroying hyperproliferative cells comprising bringing said cells or a part of them into contact with a combination of enzymes or nucleic sequences as defined above. before, in the presence of a nucleoside analog.

The present invention will be more fully described with the aid of the examples and figures which follow, which should be considered as illustrative and not limiting.

LEGEND OF THE FIGURES:

Figure 1: Schematic representation of the pcDNA3-TK vector

Figure 2: Schematic representation of the vector pXL2854

Figure 3: Schematic representation of the vector pXL2967

Figure 4: Schematic representation of the vector pXL3081

Figure 5: Demonstration of the expression of the GMPK and NDPK proteins

Figure 6: Schematic representation of the vector pXL3098

Table 1: Kinetic constants of GMPK from yeast and human erythrocytes on GCVMP and ACVMP. Published values: [DFSmée et al. (1985) Biochem. Pharmacol. 34: 1049-1056] ^a ; [RE Boehme (1984) /. Biol. Chem. 259: 12346-

12349] ^b ; [WH Miller and RL Miller (1980) /. Biol. Chem. 255: 7204-7207] c Table 2: TK-GMPK-NDPK coupling:% of products formed

MATERIALS AND METHODS

Abbreviations

ACV: acyclovir

GCV: ganciclovir

GMPK: guanylate kinase

HSV1-TK: herpes simplex virus type 1 thymidine kinase.

NDPK: nucleoside diphosphate kinase

General molecular biology techniques

Methods conventionally used in molecular biology such as preparative extractions of plasmid DNA, centrifugation of plasmid DNA in cesium chloride gradient, electrophoresis on agarose or acrylamide gels, purification of DNA fragments by electroelution, the extraction of proteins with phenol-chloroform, the precipitation of DNA in a saline medium with ethanol or isopropanol, the transformation in Escherichia coli are well known to those skilled in the art and are abundantly described in the literature (Sambrook et al. "Molecular Cloning, a Laboratory Manual", Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989; Ausubel et al. "Current Protocols in Molecular Biology", John Wiley & Sons, New York, 1987 ). The pUC type plasmids and phages of the M13 series are of commercial origin (Bethesda Research Laboratories), the plasmids pBSK or pBKS come from Stratagen.

The enzymatic amplification of DNA fragments by the so-called PCR technique [Polymerase-catalyzed £ hain Reaction] can be carried out using a "DNA thermal cycler" (Perkin Elmer Cetus) according to the manufacturer's recommendations.

The electroporation of plasmid DNA in E. çoH cells can be carried out using an electroporator (Bio-Rad) according to the supplier's recommendations.

Verification of the nucleotide sequences can be carried out by the method developed by Sanger et al. [Proc. Natl. Acad. Sci. USA, 74 (1977) 5463-5477] using the kit distributed by Amersham or the one distributed by Applied Biosystems.

EXAMPLE 1 Construction of expression vectors for combinations of enzymes

This example describes various methods for the construction of expression and transfer vectors of the nucleic sequences of the invention in vitro or in vivo.

1.1 - Construction of plasmid vectors

For the construction of plasmid vectors, different types of expression vectors can be used. 2 types of vectors are more particularly preferred: - The vector pSV2, described in DNA Cloning, A practical approach Vol.2, DM Glover (Ed) IRL Press, Oxford, Washington DC, 1985. This vector is a eukaryotic expression vector. The nucleic acids encoding the combinations of enzymes of the invention can be inserted into this vector at the Hpal-EcoRV sites. They are thus placed under the control of the promoter to enhance the SV40 virus.

- The vector pCDNA3 (Invitrogen). It is also a eukaryotic expression vector. The nucleic sequences coding for the enzymes or the combinations of enzymes of the invention are placed, in this vector, under the control of the CMV early promoter.

1.2 - Construction of viral vectors

According to a particular embodiment, the invention resides in the construction and the use of viral vectors allowing the transfer and the expression in vivo of nucleic acids as defined above.

With regard more particularly to adenoviruses, various serotypes, the structure and properties of which vary somewhat, have been characterized. Among these serotypes, it is preferred to use, within the framework of the present invention, human adenoviruses of type 2 or 5 (Ad 2 or Ad 5) or adenoviruses of animal origin (see application WO94 / 26914). Among the adenoviruses of animal origin which can be used in the context of the present invention, mention may be made of adenoviruses of canine, bovine, murine origin (example: Mavl, Beard et al., Virology 75 (1990) 81), ovine, porcine , avian or even simian (example: after-sales service). Preferably, the adenovirus of animal origin is a canine adenovirus, more preferably a CAV2 adenovirus [Manhattan strain or A26 / 61 (ATCC VR-800) for example]. Preferably, in the context of the invention, adenoviruses of human or canine or mixed origin are used. Preferably, the defective adenoviruses of the invention comprise ITRs, a sequence allowing the encapsidation and a nucleic acid according to the invention. Even more preferably, in the genome of the adenoviruses of the invention, the El region at least is non-functional. The viral gene considered can be made non-functional by any technique known to a person skilled in the art, and in particular by total suppression, substitution, partial deletion, or addition of one or more bases in the gene or genes considered. Such modifications can be obtained in vitro (on isolated DNA) or in situ, for example, by means of genetic engineering techniques, or by treatment with mutagenic agents. Other regions can also be modified, and in particular the region E3 (WO95 / 02697), E2 (WO94 / 28938), E4 (WO94 / 28152, WO94 / 12649, WO95 / 02697) and L5 (WO95 / 02697). According to a preferred embodiment, the adenovirus according to the invention comprises a deletion in the regions E1 and E4. According to another preferred embodiment, it comprises a deletion in region E1 at the level of which the region E4 and the nucleic sequence of the invention are inserted (Cf FR94 13355). In the viruses of the invention, the deletion in the E1 region preferably extends from nucleotides 455 to 3329 on the sequence of the adenovirus Ad5.

The defective recombinant adenoviruses according to the invention can be prepared by any technique known to those skilled in the art (Levrero et al., Gene 101 (1991) 195, EP 185 573; Graham, EMBO J. 3 (1984) 2917). In particular, they can be prepared by homologous recombination between an adenovirus and a plasmid carrying inter alia a nucleic sequence or a combination of nucleic sequences of the invention. Homologous recombination occurs after co-transfection of said adenovirus and plasmid in an appropriate cell line. The cell line used must preferably (i) be transformable by said elements, and (ii), contain the sequences capable of complementing the part of the genome of defective adenovirus, preferably in integrated form to avoid the risk of recombination. As an example of a line, mention may be made of the human embryonic kidney line 293 (Graham et al., J. Gen. Virol. 36 (1977) 59) which contains in particular, integrated into its genome, the left part of the genome an Ad5 adenovirus (12%) or lines capable of complementing the E1 and E4 functions as described in particular in applications No. WO 94/26914 and WO95 / 02697 or in Yeh et al., J. Virol. 70 (19%) 559.

Then, the adenoviruses which have multiplied are recovered and purified according to conventional techniques of molecular biology, as illustrated in the examples.

Concerning adeno-associated viruses (AAV), these are DNA viruses of relatively small size, which integrate into the genome of the cells they infect, in a stable and site-specific manner. They are capable of infecting a broad spectrum of cells, without inducing an effect on cell growth, morphology or differentiation. Furthermore, they do not seem to be involved in pathologies in humans. The AAV genome has been cloned, sequenced and characterized. It comprises approximately 4,700 bases, and contains at each end an inverted repeat region (ITR) of approximately 145 bases, serving as an origin of replication for the virus. The rest of the genome is divided into 2 essential regions carrying the packaging functions: the left part of the genome, which contains the rep gene involved in viral replication and the expression of viral genes; the right part of the genome, which contains the cap gene coding for the capsid proteins of the virus.

The use of vectors derived from AAVs for gene transfer in vitro and in vivo has been described in the literature (see in particular WO 91/18088; WO 93/09239;

US 4,797,368, US5, 139,941, EP 488,528). These applications describe various constructs derived from AAVs, in which the rep and / or cap genes are deleted and replaced by a gene of interest, and their use for transferring in vitro (on cells in culture) or in vivo (directly in an organism) said gene of interest. The defective recombinant AAVs according to the invention can be prepared by co-transfection, in a cell line infected with a human helper virus (for example an adenovirus), of a plasmid containing a nucleic sequence or a combination of nucleic sequences of the invention bordered by two inverted repeat regions (ITR) of AAV, and of a plasmid carrying the encapsulation genes (rep and cap genes) of AAV. A usable cell line is, for example, line 293. Other production systems are described, for example, in applications WO95 / 14771; WO95 / 13365; WO95 / 13392 or WO95 / 06743. The recombinant AAVs produced are then purified by conventional techniques.

Concerning herpes viruses and retroviruses, the construction of recombinant vectors has been widely described in the literature: see in particular Breakfield et al., New Biologist 3 (1991) 203; EP 453242, EP178220, Bernstein et al. Broom. Eng. 7 (1985) 235; McCormick, BioTechnology 3 (1985) 689, etc. In particular, retroviruses are integrative viruses, selectively infecting dividing cells. They therefore constitute vectors of interest for cancer applications. The genome of retroviruses essentially comprises two LTRs, an encapsidation sequence and three coding regions (gag, pol and env). In recombinant vectors derived from rctroviruses, the gag, pol and env genes are generally deleted, in whole or in part, and replaced by a heterologous nucleic acid sequence of interest. These vectors can be produced from different types of retroviruses such as in particular MoMuLV ("murine Moloney leukemia virus"; also designated MoMLV), MSV ("murine Moloney sarcoma virus"), HaSV ("Harvey sarcoma virus") ; SNV ("spleen necrosis virus"); RSV ("Rous sarcoma virus") or the Friend virus. To construct recombinant retroviruses according to the invention comprising a nucleic nucleic sequence or a combination of nucleic sequences according to the invention, a plasmid comprising in particular the LTRs, the packaging sequence and said nucleic sequence is constructed, then used to transfect a line so-called packaging cell, capable of providing trans retroviral functions deficient in the plasmid. Generally, the packaging lines are therefore capable of expressing the gag, pol and env genes. Such packaging lines have been described in the prior art, and in particular the line PA317 (US4,861,719); the PsiCRIP line (WO90 / 02806) and the GP + envAm-12 line (WO89 / 07150). Furthermore, the recombinant retroviruses may include modifications at the level of the LTRs to suppress transcriptional activity, as well as extended packaging sequences, comprising a part of the gag gene (Bender et al., J. Virol. 61 (1987) 1639). The recombinant retroviruses produced are then purified by conventional techniques.

For the implementation of the present invention, it is very particularly advantageous to use a defective recombinant adenovirus or retrovirus. These vectors indeed have properties which are particularly advantageous for the transfer of suicide genes into tumor cells.

1.3 - Chemical vectors

The nucleic acids or the plasmid expression vectors described in this example (1.1) and in example 2 can be administered as such in vivo or ex vivo. It has indeed been shown that naked nucleic acids can transfect cells. However, to improve the transfer efficiency, it is preferred to use, within the framework of the invention, a transfer vector. It can be a viral vector (example 1.2.) Or a synthetic transfection agent. Among the synthetic vectors developed, it is preferred to use, within the framework of the invention, cationic polymers of polylysine type, (LKLK) n, (LKKL) n, (PCT / FR / 00098) polyethylene imine (WO96 / 02655) and DEAE dextran or cationic or lipofectant lipids. They have the property of condensing DNA and promoting its association with the cell membrane. Among these, mention may be made of lipopolyamines (lipofectamine, transfectam, etc.), different cationic or neutral lipids (DOTMA, DOGS, DOPE, etc.) as well as peptides of nuclear origin. In addition, the concept of targeted transfection has been developed, mediated by a receptor, which takes advantage of the principle of condensing DNA thanks to the cationic polymer while directing the binding of the complex to the membrane thanks to a chemical coupling between the cationic polymer and the ligand of a membrane receptor, present on the surface of the cell type that is to be grafted. The targeting of the transferrin, insulin or hepatocyte asialoglycoprotein receptor has thus been described. The preparation of a composition according to the invention using such a chemical vector is carried out according to any technique known to those skilled in the art, generally by simple contacting of the various components.

EXAMPLE 2 Cloning of HSV1-TK and / or guanylate kinase and / or nucleoside diphosphokinase in a eukaryotic expression vector

This example describes a particular embodiment of the invention, using a plasmid expression vector system to produce in situ the combinations of enzymes of the invention.

The expression of prokaryotic or eukaryotic genes in mammalian cells is known to those skilled in the art. To optimize this expression, the vectors of the invention described below comprise the following signals i) a promoter / enhancer such as the CMV promoter which is well expressed in human cells; ii) a Kozak sequence, the consensus of which is (G / A) NNAUG (G / A); iii) the gene to be expressed; followed iv) by a polyadenylation sequence (V. Chisholm 1995 DNA cloning Vol.4, ed. D. Glover and B. Hames pi). Such constructions are possible using commercial vectors such as the pZeoSV, pcDNA3 ... vectors and have been produced with the genes coding for HSVl-TK, GMPK and NDK of S. cerevisae.

2.1 - HSVl-TK expression vector

The HSV1-TK gene coding for thymidine kinase of the herpes simplex virus type 1, derived from the plasmid pHSV-106 (Gibco-BRL) was cloned into the eukaryotic expression vector pcDNA3 (Invitrogen). This 6936 bp pcDNA3-TK plasmid was constructed by introducing the 1.5 kb EcoRI-NotI insert originating from pBTKl between the EcoRI and NotI sites of pcDNA3, see figure. The plasmid pBTKl was obtained in the following manner: After making the ends blunt, the insert BglII-NcoI of 1.5 kb originating from pHSV-106 and containing the gene HSVl-TK, the sequence of which is published by McKnight 1980 Nucl. Acids Res. 8 p5931, was cloned at the Smal site of pBSK.

The insert of the plasmid pcDNA3-TK contains i) 60 bp upstream of the HSVl-TK gene comprising the Kozak sequence (CGTATGG), ii) the gene sequence (1.13 kb) which is identical to that published by McKnight 1980 Nucl . Acids Res. 8 p5931, iii) the 3 'sequence to the gene (0.3 kb) which is also described by McKnight.

2.2 - Guanylate kinase expression vector

The GUK1 gene of 561 bp encoding the guanylate kinase of S. cerevisae derived from pGUK-1 (see example 3), was cloned into the eukaryotic expression vector pcDNA3 (Invitrogen) after introducing a Kozak consensus. This plasmid pXL2854 was obtained in the following manner. The Xbal-BamHI insert of pGUK-1 containing the GUK1 gene was cloned into the plasmid pSL301 (Invitrogen) between the Xbal-BamHI sites so that the GUKl gene can then be excised by the enzymes HindIII and BamHI and be donated between the HindIII and BamHI sites of pcDNA3 to generate a plasmid pcDNA3-GUK1. Between the HindIII and PflMI sites of this plasmid is cloned a HindHI-PflMI fragment of 150 bp containing a Kozak consensus and the 5 'region of GUK1 to form the plasmid pXL2854 of 6001 bp see FIG. 2. The HindHI-PflMI fragment of 150 bp was isolated from a 280 bp fragment amplified by PCR using the plasmid pGUK-1 and sense oligonucleotides 6915 5 '(GAG AAG CTT GCC ATG GCC CGT CCT ATC GTA A) 3' (SEQ ID n . 1) and antisense 6916 5 '(GAG GAT CCG TTT GAC GGA AGC GAC AGT A) 3' (SEQ ID n.2), hybridization taking place at 45 ° C (the Kozak consensus being emphasized on the oligonucleotide 6915 ). The nucleic sequence amplified by PCR has been sequenced and has two differences compared to the published sequence (M. Konrad 1992 J. Biol. Chem. 267 p25652) corresponding to the changes S2A (Serine in position 2 replaced by an alanine) and V34A ( Valine in position 34 replaced by an alanine).

2.3 - Expression vector of the nucleoside diphosphokinase

The YNK gene coding for the nucleoside diphosphokinase of S. cerevisae. derived from the plasmid pAD1-YNK (K. Watanabe et al. 1993 Gene 29 pl41), was cloned into the eukaryotic expression vector pcDNA3 (Invitrogen) after having introduced a Kozak consensus. This plasmid pXL2967 was obtained in the following manner. PCR amplification was carried out with the plasmid pAD1-YNK for template and sense oligonucleotides 7017 5 '(AAG GAT CCA CCA TGG CTA GTC AAA CAG AAA) 3' (SEQ ID n. 3.) and antisense 7038 5 '( AAG AAT TCA GAT CTT CAT TCA TAA ATC CA) 3 '(SEQ ID no. 4) at the hybridization temperature of 40 ° C (the Kozak consensus being underlined on the oligonucleotide 7017). The amplified fragment of 477 bp is digested with BamHI and EçoRI then cloned between the BamHI and EcoRI sites of pcDNA3 to generate the plasmid pXL2967 of 5861 bp, see FIG. 3. The sequence of the fragment amplified by PCR is the same as that published for the YNK gene except at position 4 which corresponds to a change S2A S2A (Serine in position 2 replaced by an alanine) for the protein NDPK (K. Watanabe et al. 1993 Gene 29 pl41).

2.4 - Vector for the co-expression of 2 genes

Gene coexpression can be achieved in several ways known to those skilled in the art. A preferred embodiment consists in introducing, between the sequences to be expressed, internal ribosome entry sites, IRES sequences (Mountford et al., TIG 11 (1995) 179). An IRES sequence and the YNK gene are introduced 3 ′ of the GUK1 gene cloned into pcDNA3 to generate a vector allowing the co-expression of GUK1 and YNK, in the form of a bicistronic unit (vector pGUKl-YNK). More specifically, the sequence of TIRES (internai ribosome entry site) of the EMCV (encephalomyocartis virus) originating from the plasmid pCITE (Novagen) was recloned by PCR between EcoRI and Ncol sites and introduced into the EcoRI and EcoRV sites of the plasmid pBluescript ( Stratagene) to generate the plasmid pXL3065. The 477 bp NcoI-EcoRV fragment containing the YNK gene derived from the plasmid pXL2967 is cloned between the NcoI and EcoRV sites of pXL3065 to form the plasmid pXL3079. The 1 kb BamHI-EcoRV fragment containing the IRES and the YNK gene of the plasmid pXL3079 is then cloned between the BamHI and EcoRV sites of the plasmid pXL2854 to generate the plasmid pXL3081. This plasmid contains the CMV and T7 promoters upstream of the GUK1 gene coding for guanylate kinase of S. cerevisae itself followed by TIRES and the YNK gene coding for the nucleoside diphosphokinase of S. cerevisae see FIG. 4. The expression of the proteins GMPK and NDPK was tested using the plasmids pXL2854, pXL2967 and pXL3081 a transcription / translation system for reticulocytes originating from Promega, see FIG. 5. The results obtained show that the GMPK and NDPK proteins are coexpressed with the plasmid pXL3081.

The same approach is used to generate a vector coexpressing TK and YNK (vector pTK-YNK). or TK and GUKl (vector pTK-GUKl).

2.5 - Vector for the co-expression of 3 genes

The sequence coding for TK is inserted into the vector pGUKl-YNK of example 2.4 above to generate a vector capable of expressing the 3 enzymatic activities (vector pTK-GUKl-YNK).

2.6 - Vector for the expression of a HSVl-TK / S.cerevisiae GMPK fusion

The construction of fusion protein is well known to those skilled in the art and is accomplished by the creation of a peptide linker between the C-terminal part of one protein and the N-terminal part of the other protein (1989 Nature 339 p394). Such a protein allows the cellular co-localization of enzymes and can also favor the "tunneling" of substrate (Ljungcrantz et al. 1989 Biochemistry 28 p8786). A fusion protein was produced by linking using the linker - (Gly) 4-Ser-

(Gly) 4-Ser- (Gly) 4 (SEQ ID No. 9) the C-terminal sequence of HSV1-TK (Asn376) and the N-terminal sequence of the GMPK of S. cerevisiae (Ser2). The plasmid pXL3098 allowing the production of this fusion protein was constructed as follows. The 3 'sequence of the HSVl-TK gene (positions 1108 to 1128) was cloned by hybridization of the 5 'sense oligonucleotides (CCG GGA GAT GGG GGA GGC TAA CGG AGG TGG CGG TTC TGG TGG CGG AGG CTC CG) 3' (SEQ ID n ° 5) and antisense 5 '(GAT CCG GAG CCT CCG CCA CCA GAA CCG CCA CCT CCG TTA GCC TCC CCC ATC TC) 3 '(SEQ ID No. 6), so that the Asn codon of the HSVl-TK gene (position 1128 bp) is followed by the codons coding for amino acids ((Gly) 4Ser) 2Gly. The 58 bp fragment is thus cloned between the Xmal and BamHI sites of pNEB193 (Biolabs) to generate the plasmid pTKL +. The 5 'sequence of the GUKl gene of S. cerevisiae is amplified by PCR using the pXL2854 template and 5' sense oligonucleotides (GAG AAT TCC GGA GGC GGT GGC TCC CGT CCT ATC GTA) 3 '(SEQ ID no. 7) and 5 'antisense (GAG GAT CCG TTT GAC GGA AGC GAC AGT A) 3' (SEQ ID No. 8), so that the codon Ser at position 2 of the GMPK is preceded by the codons (Gly) 3Ser . This 0.27 kb fragment is cloned between the BamHI and EcoRI sites of pUC19 (Biolabs) to form a plasmid which is then cut with PflMI and BamHI in order to introduce therein the 0.44 kb PflMI-BamHI fragment of pXL2854 containing the 3 'sequence of GUKl. and generate the plasmid pGUKL-. The 0.58 kb BspEI-Xbal fragment of pGUKL- is inserted between the BspEI and Xbal sites of pTKL + to create pTKLGUK. The 0.63 kb Xmal-BamHI insert of pTKLGUK is then cloned between the Xmal and BamHI sites of the expression vector pET11a, to form the plasmid pXL3098, see FIG. 6. This plasmid is then introduced into the strain BL21DE3met- and leads to the production of the protein TK- ((Gly) 4Ser) 3-GMPK.

The fusion protein is then purified to homogeneity and the kinetic parameters of GCV and TACV phosphorylation of this enzyme are compared to the kinetic parameters obtained with the HSV1-TK and GMPK proteins of S. cerevisiae. 2.7 - Transfer and expression in vivo

The vectors described in Examples 2.1 to 2.6 are used for the transfer and expression in vivo of combinations of enzymes according to the invention. For this purpose, various compositions comprising said vectors are prepared:

a composition comprising the vector pcDNA3-TK, the vector pXL2854 and the

Lipofectamine,

a composition comprising the vector pcDNA3-TK, the vector pXL2967 and the Lipofectamine,

- A composition comprising the vector pcDNA3-TK, the vector pXL2854, the vector pXL2967 and Lipofectamine

a composition comprising the vector pTK-GUKl and the Lipofectamine, optionally in combination with the vector pXL2967,

a composition comprising the vector pTK-YNK and the Lipofectamine, optionally in combination with the vector pXL2854, and,

- A composition comprising the vector pTK-GUKl-YNK and Lipofectamine

Lipofectamine can be replaced by another chemical vector as described in Example 1.3.

These different compositions are used in vivo or ex vivo for the intracellular transfer and expression of combinations of enzymes according to the invention. They can also be used on cell cultures, and for example on cultures of NIH3T3 fibroblast cells or cells of human colon carcinoma, HCT116. After transfer of the vectors, the nucleoside analog is administered and cell destruction is demonstrated.

EXAMPLE 3 Purification of guanylate kinase

The acellular extracts of the E. coli strains overexpressing the GMPK of S. cerevisae can be prepared in various ways, among which there may be mentioned, lysozyme lysis in the presence of EDTA, the use of Menton type grinding devices -Golin, French Press, X-Press, or the action of ultrasound. More particularly, the acellular extracts of the L \ coli strains overexpressing the GMPK of S. cerevisae were prepared in the following manner:

The E. coli BL21 (DE3) pGUK-1 strain is cultivated as described by M. Konrad in J. Biol. Chem. 267 p25652 in 1992. After centrifugation (5000 xg; 20 min), the cells obtained from 1 l of culture are resuspended in 20 ml of Tris / HCl buffer 20 mM pH 7.5, containing ImM EDTA and sonicated for 4 min at 4 ° C. . After centrifugation (50,000 x g; 1 h), the supernatant is injected onto a MONO Q HR 10/10 column (Pharmacia) equilibrated in 20 mM Tris / HCl buffer pH 7.5. The proteins are eluted with a linear gradient from 0 to 500 mM NaCl in the 20 mM Tris / HCl buffer pH 7.5. The fractions containing the GMPK activity are combined and concentrated, then chromatographed on a column of Superdex 75 HR16 / 10 (Pharmacia) eluted with Tris / HCl buffer 50 mM pH 7.5, 150 mM NaCl. The fractions containing the GuK activity are grouped together. After this stage, the preparation has a single visible band on SDS-PAGE, after revelation with Coomassie Blue, and this band migrates with an apparent molecular weight of approximately 21,000.

The activity of GMPK is conventionally assayed using an assay protocol described in the literature, Agarwal et al. 1978 Meth. Enzymol. vol.LI p 483. EXAMPLE 4 Determination of the kinetic constants of guanylate kinase from S. cerevisae.

The kinetic constants of the yeast GMP kinase purified as described in Example 3 are determined under the following enzymatic assay conditions:

The yeast GMP kinase is incubated for 10 min at 30 ° C in 100 μl of 50 mM Tris / HCl buffer pH 7.8 containing 4 mM ATP, 10 mM Mg C12, 100 mM KC1, 1 mg / ml BSA (bovine albumin serum) and 5-100 μM [8-3H] GCVMP (40 nCi / nmol) or 200-3200 μM [8-3HJACVMP (40 nCi / nmol). The reaction is stopped by heating the reaction mixture for 3 min at 80 ° C, 50 μl of 10 mM potassium phosphate buffer pH 3.5 are added, and after centrifugation for 2 min at 10,000 xg, 100 μl of supernatant are analyzed by high pressure liquid chromatography (HPLC) in the following system:

Stationary phase: Partisphere SAX (WHATMAN) - particle diameter: 5μm - Dimensions: 4.6 x 125 mm.

Mobile phase:

Buffer A: KH2PO40.01 M pH3.5 (adjusted using concentrated H3PO4)

Buffer B: KH2PO40.75 M ρH3.5 (adjusted using concentrated H3PO4)

Flow rate: lml / min

Gradient:

Detection:

UV: 265 nm

Radiochemical: detection of tritium

scintillating flow (Berthold Optisafe 1) 1ml / min

For the calculation of the kinetic constants, the quantity of yeast GMP kinase introduced into the enzymatic reaction is adjusted so as to transform at most 10% of the substrate introduced at the start. The Michaelis curves are adjusted to the experimental points using the Grafit software (Sigma). The results are presented in Table 1. They show that the yeast GMP kinase is capable of phosphorylating GCVMP and TACVMP. In addition, it has a GCVMP phosphorylation rate 2 times greater than the human enzyme. Likewise, its affinity for GCVMP is greater by a factor at least equal to 2 than the affinity that the human enzyme has for this substrate. In total, the value of Vmax / Km of yeast Tenzyme for GCVMP is 4.4 times greater than the value of human Tenzyme. For substrates entering into competition, the constant Vmax / Km determines the specificity of Tenzyme with respect to these substrates. She is known as "constant specificity" [A. Fersht, Enzyme Structure and Mechanism, 1985, WH Fréeman an Co., London].

In the same way, the yeast GMP kinase has a capacity to phosphorylate TACVMP much superior to the human enzyme. The value of the Vmax / Km of yeast Tenzyme for TACVMP is 7 to 9 times higher than the value that the human Tenzyme has for this substrate.

EXAMPLE 5 - Coupling of TK enzymatic activity. GMPK and NDK:

In a preferred coupling mode, the incubation is carried out in 100 μl of Tris / HCl buffer 50 mM pH 7.8, containing 1 mg / ml of BSA (bovine serum albumin), 5 mM ATP, 4 mM MgCl 2, 12 mM KCl, 2 mM DTT, 600 μM EDTA , lOOμM [8-3H] -GCV (40nCi / nmol) or lOOμM [2-3H] -ACV 40nCi / nmol and different amounts of TK, GuK and NDPK (Cf Table 2). The NDPKs of 3 organisms were used (enzymes marketed by SIGMA), as indicated in Table 2.

The coupling of the HSVl-TK and GMPK enzymes of S. cerevisae allows 90% phosphorylation of ganciclovir to ganciclovir diphosphate. And the nucleoside diphosphokinase from baker's yeast, i.e. from S. cerevisae. coupled with the enzymes HSVl-TK and GMPK, allows phosphorylation to ganciclovir triphosphate with better activity than does the human nucleoside diphosphokinase of erythrocytes. Comparable results are obtained with acyclovir. These results clearly demonstrate (a) that the combination of enzymes according to the invention brings a significant improvement in phosphorylation, indicating that the only modification of the TK would not be sufficient to improve the properties of the system, (b) that the system of the invention makes it possible to increase the efficiency of treatment with a TK suicide gene (c) that certain Non-human enzymes have better activity for nucleoside analogs, making their use particularly advantageous.

EXAMPLE 6 Purification of the HSV1-TK / S.cerevisiaeGMPK fusion protein

The acellular extracts of the E. çoH strains overexpressing the fusion protein can be prepared in various ways, among which may be mentioned, lysozyme lysis in the presence of EDTA, the use of Menton-Golin type grinding devices , French Press, X-Press, or the action of ultrasound. More particularly, the acellular extracts of the E. coli strains overexpressing the fusion protein were prepared as follows:

The E. coli BL21 DE3 pXL3098 strain is cultured in LB medium. After centrifugation (5000xg; 20 min), the cells obtained from 11 of culture are resuspended in 20 ml of buffer A: 50 mM Tris / HCl pH 7.8, containing 5 mM DTT, 4 mM MgCl ₂ , 10% glycerol (v / v), Benzamidine 2mM, E64 50μl / l (lOOμg / 1 solution N- [N- (L-3-trûπs-carboxyoxiran-2-carbonyl) -L-leucyl] -4-aminobutylgua nidine, Péfabloc 0,2mM, STI (Soybeam trypsin inhibitor), Leupeptin 2mg / ml, and sonicated for 4 min at 4 ° C. After centrifugation (50000xg; lh) the supernatant is injected onto a MONO Q HR 10/10 column (Pharmacia) balanced in buffer A. The proteins are eluted with a linear gradient from 0 to 400 mM NaCl in the Tris / HCl buffer 20 mM pH 7.5. The fractions containing the TK and GMPK activity are combined and concentrated then chromatographed on a column of Superdex 200 HILoad 26/60 (Pharmacia) eluted with buffer A containing 150 mM NaCl. The fractions containing the TK and GMPK activity are combined. At this stage the preparation presents u only band visible on SDS-PAGE, after revelation with Coomassie Blue, and this band migrates with an apparent molecular weight of approximately 61,000. EXAMPLE 7 Study of phosphorylation by the HSV1-TK / GMPK fusion

This example describes a study of the kinetic parameters of phosphorylation of GCV and of TACV of the fusion, compared with the kinetic parameters obtained with the proteins HSV1-TK and GMPK of S. cerevisiae.

7.1. TK activity assay

The activity of TK is assayed as follows: an enzymatic extract containing approximately 0.1 unit of TK is incubated for 15 min at 37 ° C. in 100 μl of Tris / HCl buffer 50 mm pH 7.8 containing lmg / ml of BSA ( bovine serum albumin), 5mM ATP, 4mM MgC12, 12mM Kcl, 2mM DTT, 600μM EDTA and lOOμM GCV + [8-3H] -GCV 40nCi / nmol. The reaction is stopped by the addition of 10 μl of 50 mM Tris / HCl buffer pH 7.8 containing ImM of non-radioactive thymidine. The phosphorylated species are fixed on a DEAE sephadex column (400 μl of gel) then, after washing the column, these species are eluted with 2 ml of 1M HCl. The radioactivity in the sample is then counted by liquid scintillation.

7.2. Determination of GMPK activity

The activity of GMPK is conventionally assayed using an assay protocol described in the literature, K.C. Agarwal et al. (Methods In Enzymology (1978) Vol. LI 483-490).

7.3. Determination of the activity of the fusion protein and of the co-incubation of the TK and GMPK enzymes

The capacity of the TK-GMPK fusion protein to transform GCV or TACV into GCVDP or ACVDP relative to a synthetic mixture of TK and GMPK is determined as follows: Incubation takes place in 200 μl of Tris / HCl buffer 50 mM pH 7.8 containing 1 mg / ml of BSA (bovine serum albumin), 5 mM ATP, 4 mM MgC12, 12 m MKC1, 2 mM DTT, 600 μM EDTA, 1 to 100 μM of GCV + [8-3H ] GCV (40nCi / nmol) or 1 to 100μM ACV + [2-3H] ACV 40nCi / nmol and different amounts of TK, GMPK and equivalent amounts of fusion protein. The reaction is stopped by heating for 3 min at 80 ° C., after centrifugation 100 μl of incubate are analyzed in the following system:

Stationary phase: SAX Partisphere (WHATMAN) - particle diameter: 5μm.

Dimensions: 4.6 x 125 mm.

Mobile phase:

Buffer A: KH2PO40.01 M pH3.5 (adjusted using concentrated H3PO4) Buffer B: KH2PO40.75 M pH3.5 (adjusted using concentrated H3PO4)

Flow rate: 1ml / min

Gradient:

Detection:

UV: 265 nm

Radiochemical: detection of tritium flow of scintillant (Optisafe 1 from Berthold) 1ml / min

The coupling and combination of the HSVl-TK and GMPK enzymes of S. cerevisiae allow the phosphorylation of ganciclovir to ganciclovir diphosphate (see table 3). Comparable results are obtained with acyclovir (see Table 3).

These results clearly show that the TK-GMPK fusion protein retains the properties of the two original enzymes. In addition, depending on the operating conditions, the TK-GMPK fusion protein brings a significant improvement in phosphorylation a) up to a factor of 1.8, for GCV b) up to a factor of 1.2 for TACV versus the coincubation of the wild-type TK and GMPK enzymes.

The fusion protein allows in vivo co-localization of enzymatic activities in the cell, the separate enzymes being distributed, in the nucleus for THSVl-TK and in the cytosol for GMPK. This construction therefore makes it possible to increase the effectiveness of treatments by suicide gene.

All of these results clearly demonstrate the therapeutic interest of the present invention, making it possible both to reduce the doses of nucleoside and of enzymes and to obtain a significant pharmacological benefit.

LIST OF SEQUENCES

(1) GENERAL INFORMATION:

(α) DEPOSITOR:

(A) NAME: RHONE POULENC RORER S.A.

(B) STREET: 20, Avenue Raymond Aron

(C) CITY: ANTONY (E) COUNTRY: FRANCE

(F) POSTAL CODE: 92165

(G) TELEPHONE: 01.55.71.70.36 (H) FAX: (01.55.71.72.91 (ii) TITLE OF THE INVENTION: COMBINATIONS OF ENZYMES FOR THE DESTRUCTION OF PROLIFERATIVE CELLS

(iii) NUMBER OF SEQUENCES: 9 (iv) FORM DETERMINABLE BY COMPUTER:

(A) TYPE OF SUPPORT: Tape

(B) COMPUTER: IBM PC compatible

(C) OPERATING SYSTEM: PC-DOS / MS-DOS

(D) SOFTWARE: Patentin Release # 1.0, Version # 1.30 (EPO)

(2) INFORMATION FOR SEQ ID NO: 1:

(i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 31 base pairs

(B) TYPE: nucleotide

(C) NUMBER OF STRANDS: single

(D) CONFIGURATION: linear (ii) TYPE OF MOLECULE: cDNA

(xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 1: GAGAAGCTTG CCATGGCCCG TCCTATCGTA A 31

(2) INFORMATION FOR SEQ ID NO: 2:

(i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 28 base pairs (B) TYPE: nucleotide

(C) NUMBER OF STRANDS: single

(D) CONFIGURATION: linear

(ii) TYPE OF MOLECULE: cDNA

(xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 2: GAGGATCCGT TTGACGGAAG CGACAGTA 28 (2) INFORMATION FOR SEQ ID NO: 3:

(i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 30 base pairs

(B) TYPE: nucleotide

(C) NUMBER OF STRANDS: single

(D) CONFIGURATION: linear (n) TYPE OF MOLECULE: cDNA

(XI) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 3: AAGGATCCAC CATGGCTAGT CAAACAGAAA 30

(2) INFORMATION FOR SEQ ID NO: 4:

(i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 29 base pairs (B) TYPE: nucleotide

(C) NUMBER OF STRANDS: single

(D) CONFIGURATION: linear

(ii) TYPE OF MOLECULE: cDNA

(xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 4: AAGAATTCAG ATCTTCATTC ATAAATCCA 29

(2) INFORMATION FOR SEQ ID NO: 5:

(i) CHARACTERISTICS OF THE SEQUENCE:

(A) LENGTH: 53 base pairs

(B) TYPE: nucleotide (C) NUMBER OF STRANDS: single

(D) CONFIGURATION: linear

(ii) TYPE OF MOLECULE: cDNA (xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 5:

CCGGGAGATG GGGGAGGCTA ACGGAGGTGG CGGTTCTGGT GGCGGAGGCT CCG 53

(2) INFORMATION FOR SEQ ID NO: 6:

(i) CHARACTERISTICS OF THE SEQUENCE:

(A) LENGTH: 53 base pairs

(B) TYPE: nucleotide

(C) NUMBER OF STRANDS: simple (D) CONFIGURATION: linear (11) TYPE OF MOLECULE: cDNA

(xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 6: GATCCGGAGC CTCCGCCACC AGAACCGCCA CCTCCGTTAG CCTCCCCCAT CTC 53

(2) INFORMATION FOR SEQ ID NO: 7: (i) CHARACTERISTICS OF THE SEQUENCE:

(A) LENGTH: 36 base pairs

(B) TYPE: nucleotide

(C) NUMBER OF STRANDS: single

(D) CONFIGURATION: linear

(ii) TYPE OF MOLECULE: cDNA

(xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 7: GAGAATTCCG GAGGCGGTGG CTCCCGTCCT ATCGTA 36

(2) INFORMATION FOR SEQ ID NO: 8:

(i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 28 base pairs

(B) TYPE: nucleotide

(C) NUMBER OF STRANDS: single

(D) CONFIGURATION: linear (n) TYPE OF MOLECULE: cDNA

(xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 8: GAGGATCCGT TTGACGGAAG CGACAGTA 28

(2) INFORMATION FOR SEQ ID NO: 9:

(i) CHARACTERISTICS OF THE SEQUENCE:

(A) LENGTH: 14 amino acids

(B) TYPE: amino acids (D) CONFIGURATION: linear

(n) TYPE OF MOLECULE: peptide

(xi) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 9: Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser Gly Gly Gly Gly 1 5 10 VO

Table 1 - CINETIC CONSTANTS OF GUANYLATE KINASES OF YEAST TΛ o AND HUMAN ERYTHROCYTES FOR GCVMP and ACVMP

ro

π

H

VO

Bone

Table 2 - TK-GMPK-NDPK COUPLING:% BY NI OF PRODUCTS FORMED

vo

Table 3: Phosphorylation of GCV and ACV by coupling or combining the enzymes TK and GMPK

(Λ

O

* -

4>

t n H

"" pmmol of GCVDP formed 73

Claims

1. Composition for the delivery and in vivo production of a combination of enzymes, comprising:

a first nucleic acid coding for an enzyme capable of phosphorylating a nucleoside analog, to generate a monophosphate analog,

a second nucleic acid coding for an enzyme capable of phosphorylating said monophosphate analogue, to generate a diphosphate analogue, and,

- A third nucleic acid encoding an enzyme capable of phosphorylating said diphosphate analog, to generate a toxic triphosphate analog.

2. Composition according to claim 1 characterized in that the first nucleic acid codes for a thymidine kinase.

3. Composition according to claim 1 characterized in that the second nucleic acid codes for a guanylate kinase.

4. Composition according to claim 1 characterized in that the third nucleic acid codes for a nucleoside diphosphate kinase.

5. Composition according to claim 1 characterized in that the nucleic acids are carried by the same vector.

6. Composition according to claim 1 characterized in that the nucleic acids are carried by several vectors.

7. Composition according to claim 5 or 6 characterized in that the vectors are plasmid or viral vectors.

8. Composition comprising a nucleic acid coding for a thymidine kinase and a second nucleic acid coding for a nucleoside diphosphate kinase.

9. Composition according to claim 8 characterized in that the nucleoside diphosphate kinase is of non-human eukaryotic origin.

10. Composition according to Claim 9, characterized in that the nucleoside diphosphate kinase is chosen from NDPK from yeast or from beef.

11. Composition according to claim 10 characterized in that the nucleic acid coding for thymidine kinase and the nucleic acid coding for nucleoside diphosphate kinase are coupled and code for a protein carrying both TK and NDPK activities.

12. The composition according to claim 8 further comprising a nucleic acid coding for a guanylate kinase.

13. Composition according to claim 12 characterized in that the nucleic acid codes for a yeast guanylate kinase.

14. Composition comprising a nucleic acid coding for a thymidine kinase and a second nucleic acid coding for a non-human guanylate kinase.

15. Composition according to claim 14 characterized in that the guanylate kinase is of yeast origin.

16. Composition according to claim 15 characterized in that the nucleic acid coding for thymidine kinase and the nucleic acid coding for guanylate kinase are coupled and code for a protein carrying both TK and GUK activities.

17. Composition according to one of claims 1 to 16 characterized in that the thymidine kinase is of viral origin.

18. Vector including:

a first nucleic acid coding for a thymidine kinase, and,

- a second nucleic acid coding for a non-human guanylate kinase.

19. Vector including:

a first nucleic acid coding for a thymidine kinase, and,

- a second nucleic acid coding for a nucleoside diphosphate kinase.

20. Vector according to claim 19 characterized in that the nucleoside diphosphate kinase is of non-human eukaryotic origin, preferably bovine or yeast.

21. Vector according to claim 19 or 20 characterized in that it further comprises a nucleic acid coding for a guanylate kinase.

22. Vector according to one of claims 18 to 21 characterized in that the thymidine kinase is of viral origin.

23. Vector according to claims 18 to 22, characterized in that the different nucleic acids are placed under the control of distinct promoters.

24. Vector according to claims 18 to 22, characterized in that the different nucleic acids form a polycistronic unit under the control of a single promoter.

25. Vector according to claims 18, 22 and 24 comprising a first nucleic acid coding for a thymidine kinase of viral origin and a second nucleic acid coding for a non-human guanylate kinase, said nucleic acids being coupled and coding for a protein carrying both TK and GUK activity.

26. Vector according to claims 18 to 25 characterized in that it is a plasmid or viral vector.

27. Composition including

. a combination of enzymes capable of triphosphorylating a nucleoside analog, and

. said nucleoside analog,

for simultaneous, separate or staggered administration.

28. Composition for the in vivo production of a nucleoside triphosphate analog comprising, packaged separately or together: - a nucleoside analogue

- a nucleic acid encoding a thymidine kinase

a nucleic acid coding for a guanylate kinase, and,

- a nucleic acid encoding a nucleoside diphosphate kinase.

29. Composition for the in vivo production of a nucleoside triphosphate analog comprising, packaged separately or together:

- a nucleoside analogue

a nucleic acid coding for a thymidine kinase, and

- a nucleic acid encoding a nucleoside diphosphate kinase.

30. Composition according to claims 27 to 29 characterized in that the nucleoside analog is a guanosine analog.

31. Composition according to claim 30 characterized in that the guanosine analog is chosen from ganciclovir, acyclovir and penciclovir.

32. Composition comprising a combination of enzymes involved in the phosphorylation of nucleosides, at least one of these enzymes being of non-human eukaryotic origin.

33. A method for the destruction of proliferative cells comprising administering to said cells a combination of enzymes comprising a TK and an NDPK and a nucleoside analog.

34. Method according to claim 33 characterized in that the combination further comprises a guanylate kinase.

35. A method for the destruction of proliferative cells comprising administering to said cells a combination of enzymes comprising a TK and a non-human GMPK and a nucleoside analog.

36. Method according to claims 33 to 35 characterized in that the nucleoside analog is a guanosine analog.

37. Method according to claim 36 characterized in that the guanosine analog is chosen from ganciclovir, lacyclovir and penciclovir.

38. Method according to claims 33 to 37 characterized in that the enzymes are administered to the cells by administration of nucleic acids coding for the said enzymes.

39. Process for the triphosphorylation of a nucleoside analog comprising bringing said analog into the presence with a combination of enzymes, at least one of them being of yeast origin.

40. Use of the nucleoside diphosphate kinase or a nucleic acid coding for it, in combination with a thymidine kinase or a nucleic acid coding for a thymidine kinase and a nucleoside analog, for the preparation of a pharmaceutical composition for to the destruction of proliferative cells.

41. Nucleic acid encoding a protein for coupling between a thymidine kinase of viral origin and a guanylate kinase.

42. Nucleic acid according to claim 41 characterized in that the guanylate kinase is of yeast origin.

43. Nucleic acid encoding a protein for coupling between a thymidine kinase of viral origin and a nucleoside diphosphate kinase.

44. Protein encoded by a nucleic acid according to claims 41 to 43.