MXPA00009048A - Suicide gene therapy system for the treatment of brain tumours - Google Patents

Abstract

In a suicide gene therapy system, a nucleic acid molecule, in particular a DNA molecule, encoding a glycosidase is applied with a non-toxic cyanogenic glycoside substrate, which upon degradation by the glycosidase yields hydrogen cyanide, for the treatment of brain tumours in animals or humans. The DNA preferably encodes linamarase, the substrate is preferably linamarin. It is carried by a viral vector or a plasmid. The two interacting components are administered separately.

Description

SYSTEM OF THERAPY WITH SUICIDAL GENES FOR THE TREATMENT OF CEREBRAL TUMORS DESCRIPTION OF THE INVENTION The invention relates to the gene therapy sector for brain tumors. Glioblastoma is one of the most frequent brain tumors (it is affected by 1 in 50,000 people per year) and deadly (it is fatal in about 100%). In order to find a cure for this disease, a number of strategies have been developed in the past. One of the most promising strategies was to sensitize the tumor cells to specific prodrugs. Among these strategies, one of the most satisfactory systems to date is the mixture of thymidine kinase from herpes simplex (HsTK) and ganciclovir (Culver et al., 1992, Ram et al., 1993, Izquierdo et al., 1995). HsTK phosphorylates nucleoside-like compounds (such as ganciclovir or acyclovir) that normal eukaryotic cells are unable to phosphorylate. The incorporation of these analogs into DNA in replication blocks the process, killing the cell thereby. Tumors REF. 123152 malignant brain (glioblastomas) have been cured in an animal model, specifically rat, provided that the volume of the tumor does not exceed a size of 100-150 mm3 (Izquierdo et al., 1997). When the same treatment is applied to humans, complete remission of the tumor is not achieved, although in some cases there is a measurable reduction in size that is probably responsible for a considerable increase in the expected survival time of these patients (Culver et al. 1994; Izquierdo et al., 1996). In a similar manner, tumor cells can be sensitized to cytosine-arabinoside (ara-C) or 2'-2'-difluoro-deoxycytidine (gemcitabine) by delivery to the neoplastic cell of the deoxycytidine kinase gene which enables the prodrugs they are phosphorylated and therefore blocks replication. The cytosine-deasase gene, as another example, is capable of making these cells sensitive to the 5-fluoro-cytosine prodrug. ~~ To date, only four pro-drug activator systems have been used in transgenic systems for brain tumors; the herpes simplex thymidine kinase, mentioned above, for the activation of ganciclovir and the cytosine deaminase of E. col i for 5- fluoro-cytosine, the human cytochrome P450 2B1 for cyclophosphamide, and - the purine-nucleoside-phosphorylase of E. col i for the activation of 6-methyl-purine-2 '-deoxyribonucleoside. It has already been suggested by Deonarain, 1994, to explore the enzyme linamarase, which converts amygdalin into cyanide, for the treatment of cancers of hepatomas, melanoas, pancreatic and breast. The object of the present invention was to create an alternative system of murine murine therapy for the treatment of brain tumors, in particular of malignant tumors such as glioblase. To solve the problem underlying the object of the invention, an alternative suicide gene system was developed for the preparation of a medicament that is intended for the treatment of brain tumors. The present invention is directed to the use of a nucleic acid molecule encoding a glycosidase in combination with a non-toxic cyanogenic glycoside, which after degradation by glycosidase provides hydrogen cyanide, for the preparation of a medicament that is intended for the treatment of brain tumors .

Preferably, the nucleic acid molecule is a DNA molecule. Glycosidases (O-glycosyl-hydrolases) are a widely spread group of enzymes that hydrolyse the glycosidic link between two or more carbohydrates or between a carbohydrate and a non-carbohydrate moiety. The catabolism of cyanogenic glycosides is initiated by dissociation of the rest of carbohydrate by a β-glycosidase, which provides the corresponding alpha-hydroxynitrile. The hydroxynitrile decomposes to produce hydrogen cyanide and an aldehyde ketone. In a preferred embodiment of the invention, glucosidase, in particular a β-glucosidase, and the substrate is a cyanogenic glucoside. In an even more preferred embodiment of the invention, the enzyme β-glucosidase is linamarase and the cyanogenic glucoside is linamarin. Linamarase hydrolyzes the innocuous substrate linamarin (2-OH-iso-butyronitrile-β-glucopyranoside) to form the cyanohydrin of acetone and glucose (Mkpong et al., 1990; Hughes et al., 1992) (Figure 1). Acetone cyanohydrin is unstable at pH values above 6 and decomposes spontaneously in acetone and HCN (Selmar et al., 1987 a, b). The cyanogenic glucoside, linamarin, is not hydrolyzed by most tissues of mammals and therefore in the absence of linamarase, once it has been absorbed, it is excreted in an unaltered state in the urine (Frakes and Sharma, 1986, Hernández and collaborators, 1995). The β-glucosidase, for example, linamarase, can be derived from different plants, for example cassava, oleaginous flax, hevea, climbing bean, white clover, etc., (Kakes, 1985; Fields and Gerhardt, 1994; Frehner and Conn. , 1987; Hughes and Dunn, 1982). As an alternative to linamarase and linamarin, other glycosidases and their corresponding substrates can be used. Depending on the rest of the sugar and the glycosidic bond, the enzyme can be selected, among others, between monosaccharidases and di saccharidases (for example, galactosidases, lactases, fucosidases). Enzymes can be derived from plants or microorganisms. To obtain DNA that encodes an enzyme whose sequence has not yet been identified, methods based on classical molecular biology protocols can be used (Sambrook et al., 1989). An example of a method of this type was used to determine the nucleotide and amino acid sequences of linamarase from white clover (Oxtoby et al., 1991).

Leaving aside the enzymes that occur in nature, or the nucleic acid molecules, in particular the DNAs that encode them, respectively, modified enzyme molecules, or the DNAs that encode them, respectively, can be used. Such molecules can be truncated or modified, so long as they hydrolyze the cyanogenic carbohydrate substrate, so that the toxic HCN agent is released. It can be easily checked if a modified molecule is suitable for application in the present invention, in preliminary enzymatic analyzes in which the substrate is incubated with the enzyme and the release of the enzyme is detected.

HCN. Along with linamarin, the substrate can be selected from other natural cyanogenic glycosides, for example, amygdalin, sambunigrin, prunasin, holocaline, zierin, lotaus tralina, taxifilina, vicianina, durrina (Gmelin et al., 1973; Steveps et al. Nartey, 1968). Leaving aside cyanogenic glycoside substrates with natural ones, synthetic cyanogenic glycosides can be designated to be used as substrates for the enzyme.

Examples of synthetic substrates may be monosaccharides based on those having the general formula where Ri and R? , which may be different or equal to each other, designate branched or unbranched C? -C4 alkyl; R3, R4, R5, R & Ri, Rs, R9, Rio and Rn, which may be the same or different from each other, designate hydrogen, hydroxy, C1-C4 alkyl or C1-C4 alkoxy. Similarly, an expert can design corresponding disaccharide derivatives, for example derivatives of cellobiose, maltose, lactose, sucrose, genotibose, trehalose or turanose. In order for a substrate to be appropriate for the present invention, it must be non-toxic and subject to decomposition by a glycosidase to produce hydrogen cyanide. A person skilled in the art can test the suitability of a combination of enzyme and substrate in routine enzymatic assays, in which the substrate is incubated with the enzyme under the appropriate conditions, and the production of hydrogen cyanide is determined, as described , among other bibliographical citations, by Selmar et al., 1987; Itoh-Nashida et al., 1987; Mao and Anderson, 1967; Stevens et al., 1968. The invention makes use, for the first time, of a molecule of a nucleic acid, whose action on a product causes the release of hydrogen cyanide, for the treatment of brain tumors, in particular malignant brain tumors. In the following, for reasons of simplicity, the expression "linamarase gene" is used for the nucleic acid molecule encoding a glycosidase. The invention can be applied to the treatment of tumors such as glioblas toma, medulloepithelioma, medulloblas toma, neuroblas toma, germinoma, embryonal carcinoma, malignant astrocytoma, acetabulo toma, ependymoma, oligodendroglioma, plexal carcinoma, neuroepi telioma, pinooblas toma, ependimoblas toma, primitive neuroectodermal tumor, malignant meningioma, chondrosarcoma, meningeal sarcomatous, malignant melanoma and malignant schawnoma.

The medicament of the present invention represents a combined preparation of two ingredients, namely the linamarase gene and the linamarin substrate, which are administered separately. The two components interact directly and form a functional unit. In the present invention, the linamarase gene can be used in combination with any vector system that is suitable for delivering genes to brain tumor cells. In a preferred embodiment the nucleic acid molecule encoding linamarase is a DNA or RNA molecule carried by a viral vector. In a particularly preferred embodiment, the vector is a retroviral vector. Retroviruses, which have RNA as their nucleic acid and use the reverse transcriptase enzyme to copy their genome to be integrated into the DNA of the chromosomes of host cells, only effectively infect the proliferating cells (even though they could enter inactive cells) since they require cell division for the integration and expression of genes. This property makes retroviruses the most convenient vectors for the treatment of rapidly dividing cells in malignancies that arise within a post-mydotic tissue, in an anatomical compartment (the brain) that is partially separated from the rest of the body by the barrier hemato-encephalic. When compared to other types of viral vectors, the retroviral vectors are more selective for the transgenic delivery to brain tumor cells, it is inherited by all cells of the progeny of the infected cells. The vector must meet the general requirement of being defective and therefore incapable of providing infective offspring. It has two long terminal repeats (LTRs, of Long Terminal Repea ts) that act as potent promoters and a selection gene, for example, the puromycin resistance gene (puromycin-N-acetyl transaminase), together with the linamarase gene Therapeutically active. A large variety of retroviral vectors have been designed for gene therapy applications and are available to a person skilled in the art. Examples of suitable vectors that can be used in the present invention and methods for constructing such vectors have been reported by Vile and Russel 1995, and by Braas et al., 1996. Most retroviral vectors that have been designed for Gene therapy are based on the Murine Leukemia Virus (Mo-MLV of Moloney Murine Leukemia Virus). A study on retroviral vectors, which need to be fulfilled in order to make them suitable for brain tumor gene therapy, is presented in the articles by Kramm et al., 1995, the disclosure of which, including references cited as relevant, is expressly incorporated herein by reference. The retroviral vector can be applied as it is, that is as a constituent of the pharmaceutical composition, together with an agent such as protamine or polybrene to facilitate a viral infection, or in a suitable physiological buffer, usually a buffer of saline solution. The virus titre should not be less than 10b colony forming units (u.f.c.) per me. The preferred route of administration is intratumoral injection of the virus and infusion to the substrate. In order to counteract the limitations for the use of retroviral vectors alone, that is to say the direct application of the virus by intratumoral injection, which are due to the relatively low titers and the instability of the infectious particles, the retroviral vectors are applied preferably by intratumoral grafting of packaging cells (producing cells) that have been infected with the vector and allowed to grow. The packaging cells are injected as a suspension in a saline buffer and can be of human or murine origin, preferably human, for example cells of the human fibrosarcoma cell line FLYA13 (Cosset and collaborators, 1995). An example of a mouse cell line is? CRIP (Danos and Mulligan, 1988). Human cells are preferred, since they are not rejected as easily as murine producing cells. The cells should provide titers not lower than 106 colony forming units per ml, and the number of cells injected is approximately 109 or greater. In another embodiment, the viral vector is an adenoviral vector. Due to the natural tropism of adenoviruses, their accessibility to easy manipulation and the high titers of virus that can be obtained, adenoviral vectors have found an increased interest for cancer gene therapy as an alternative to retroviral vectors, whose use is limited due to its requirement to present active cell division and therefore the virus titles that can be obtained. A large variety of adenoviral vectors are available for use in the present invention, being given by Kovesdi et al., 1997, a study on the recent developments in vectors that takes into account the structural, immunological and target orientation problems that are raised by the applications to gene therapy. Strategies for the therapy of cancer using adenoviral vectors were also assessed by Descamps et al., 1996; specifically, it has been shown that adenoviral vectors are useful for applying the suicide system of HSV-tk and ganciclovir in the treatment of glioma (Chen et al., 1994; Bonnekoh et al., 1995). The transfer of genes, mediated by adenovirus, to the brain has been further discussed by Kramm et al., 1995, and Yung, 1994; and it has been methodologically determined by Peltékian et al., 1997. The vectors described in the cited references can be used to insert the linamarase gene according to the present invention. Since adenoviruses with larger numbers of titers can be produced, they can be applied as they are, ie there is no requirement to implant producer cells.

According to a further embodiment of the invention, adenoviral and retroviral vector chimeras can be used which combine the advantageous characteristics of both the adenoviral vectors and the retroviral vectors, and have recently been shown to provide a novel strategy for achieving transfection in vi stable and highly efficient vo (Bilbao et al., 1997). In another embodiment of the invention, the viral vector is a vector of the herpes simplex virus (HSV from Herpes Simpl ex Virus). Methods for constructing HSV vectors, which can be used to introduce the linamarase gene and deliver the gene to tumor cells, have been described by Latchaman, 1994; the use of HSV vectors for brain tumors has been described by Kramm et al., 1995; and by Yung, 1994. With respect to the other components of the pharmaceutical composition containing a viral vector carrying the linamarase gene, these are essentially the same as for the retroviral vector compositions. The pharmaceutical composition may contain a polycationic agent that enhances virus binding to host cells, such as polybrene (hexadimethrine bromide) or protamine, which may be present at a concentration of about 4-8 μg / ml. When such agents are applied, the risk of possible side effects, for example, bleeding in the case of protamine, has to be weighed against the benefit of an increase in the efficiency of gene transfer that is achieved by the addition of that agent . Leaving aside viral vector systems, non-viral gene delivery systems can be used in the present invention to deliver a plasmid carrying the linamarase gene to brain tumor cells. In this case, the linamarase gene is carried by plasmid DNA. By using non-viral systems, some of the disadvantages of viral vectors can be avoided, for example design restrictions and security issues. A non-viral system, which can be used in the present invention, is based on an endocytosis mediated by a receptor. Such a system is described in PCT patent document WO 93/07283 which is based on plasmid DNA complexes and polycations which are conjugated to ligands that bind to cell surfaces and is equipped with an endosomolytic function, such as an inactivated adenovirus. or an active peptide in membranes. For the present invention, the cell surface binding ligand, which serves as a targeting function, can be selected from ligands that bind to brain cells, for example, glioma-specific antibodies. Another group of non-viral gene transfer systems, which can be used in the present invention, is based on lipophilic gene delivery vehicles. Examples of such delivery vehicles are liposomes that have been suggested for various gene therapy applications (Felgner et al., 1994, San et al., 1993). The DNA plasmid carrying the linamarase gene can be encapsulated in liposomes according to known methods (Hug and Sleigth, 1991). Other examples are cationic lipids such as lipofectin (Felgner et al., 1987). To increase the incorporation of non-viral delivery vehicles and to enhance the expression of the linamarase gene, various routes of administration can be applied in the present invention that take into account the specific requirements posed by the need to open the blood-tumor barrier. in the brain, which have been suggested for the application of drugs to brain tumors as an accompanying measure. Such routes have been studied by Kramm et al., 1995; an example is the application of bradykinin or of bradykinin-like compounds. Another example of improvement. the incorporation of the gene by brain tumor cells is the conjugation of targeting molecules with the vehicle to deliver, for example, liposomes coupling with an antibody specific for gliomas, as has been suggested by Yoshida and Mizzumo, 1994. The targeting molecules can also be monoclonal antibodies to proteins that are more abundant in gliomas, such as the fibrillar acid protein (Abaza et al., 1997). An alternative route to apply the linamarase plasmid is the transfer of naked DNA. In the main, this method of transfer would be an ideal system for various economic and technical reasons, and it is also advantageous due to the low antigenicity of the naked DNA. The disadvantage of the low efficiency of this gene transfer pathway has lately been overcome by a method that combines the injection of the plasmid followed by in vivo electroporation into the tumor tissue. By means of electroporation, the cell membranes are permeabilized, thereby significantly increasing the transfer of drugs and / or genes (Nishi et al., 1996).

As for viral vectors, the preferred route of administration for the aforementioned non-viral vectors is intratumoral injection. The administration of the prodrug, for example, linamarin, is preferably the local route, although it can also be peritoneal or intravenous. The preference for the local administration route is based on the estimation and experimental confirmation in other systems (tk and ganciclovir, or tk and 5-2-bromovinyl-2'-deoxyuridine) that the amount of prodrug needed to eradicate 1 mm3 of tumor is a thousand times less if the prodrug is administered locally than it is intraperitoneally, which substantially reduces the cost of treatment and reduces the possible harmful effects on other tissues. Local administration can be achieved by infusion with a cannula from the tumor cavity to the outside, connected for several hours per day with a continuous infusion pump, which will supply approximately 0.2 ml / hour. The treatment lasts approximately one week. The total dose is approximately 50 mg of linamarin per day. This is equivalent to 5 mg of cyanide per day, which are far from toxic doses of 50-60 mg. The prodrug is preferably administered in a saline physiological solution, for example PBS (phosphate buffered saline = Ph ospha t e Buffered Sal ine). In general, 50 mg of the prodrug constitutes the minimum dose per day that is administered for the treatment of a tumor. This dose, which is well below the toxic concentration in terms of the released cyanide, can be increased in order to eradicate larger tumors. Depending on the specific requirements of the treatment, ie the individual patient, the size and location of the tumor, the number of tumors, etc., - the mode of application of the substrate can be continuous, which is the preferred mode of application, or consecutive in several doses per day, or it can be, in some cases, a single dose. Accordingly, the concentration of the substrate may vary. In the case of continuous administration, the minimum concentration of the substrate is 10 mg / ml. If the substrate is administered in a single dose or in several doses, the concentration can be correspondingly increased up to about 30 mg / ml. The substrate is administered after a sufficient period of time for the linamarase gene to be expressed in the tumor cells. If, according to the preferred embodiment of the invention, the linamarase gene is administered by means of a retroviral vector, this period of time is determined by the rate of infection of the tumor cells by the virus. Depending on the accessibility of the tumor cells, this period of time is usually from several days to approximately one week. In a preferred embodiment, the invention comprises a kit (ki t), wherein a first container contains the preparation with the nucleic acid molecule encoding glycosidase, preferably linamarase, and a second container contains the preparation with the substrate, preferably linamarin. Both preparations are found in pharmaceutically acceptable formulations, for example physiological saline solutions. In a preferred embodiment the container contains retroviral vectors carrying the linamarase gene, with an ole title>. 106 colony forming units. In a preferred embodiment, the first container contains producer cells, infected with the retroviral vectors, which are provided in a deep frozen form. The number of cells contained in the preparation can be the number of cells ready for application, ie 109 cells or more, alternatively, a cell preparation is provided, which must be grown in the appropriate medium, the which can be provided in another container, to obtain the number of cells required for the specific application. In another preferred embodiment, the first container contains a preparation of adenoviral vectors. In another preferred embodiment, the first container contains a plasmid DNA preparation, optionally together with a non-viral gene delivery system, for example, liposomes, in which the plasmid DNA has been incorporated. In the case of a gene delivery system, based on a receptor-mediated endocytosis, the components that facilitate the entry of DNA into the cells and their expression, for example a polycation conjugated with a cellular ligand, can be present in a additional container. The second container contains the substrate, preferably linamarin. In a preferred embodiment, linamarin is present in solid form in the total dose to be applied, and the buffer is provided separately. In this embodiment, the substrate solution can be prepared, prior to administration, at an appropriate concentration according to the route of administration. It has been shown in the experiments of the present invention that the neoplastic retroviral cells expressing the linamarase gene have, after their introduction into tumor cells, the potential to cause a wide surrounding effect between neighboring cells. The data obtained suggest a substantial annihilating effect that would counteract the poor estimates (of 1-10%) made for an infection with retroviruses of solid tumors. Experiments performed in the present invention demonstrate that the transduction of mouse or rat cells by retroviral vectors expressing the cassava linamarase gene, cas 5 (Hughes et al., 1992) dramatically increases the sensitivity of these cells to cytotoxic effects of the linamarina in vi tro and in vi vo. Neighboring cells can also be affected by the released cyanide, thereby amplifying the killing effect. In the experiments of the present invention, a therapeutic retrovirus was engineered by introducing the linamarase gene (cas 5) into the pBabepuro vector based on Moloney murine leukemia (Morgenstern and Land, 1990). The resulting recombinant retroviral plasmid carries the ca s 5 gene expressed under the control of the LTR promoter (long terminal repeat, of the terminal on repea t) at 5 'and the selection gene pa c (puromycin-N-acetyl-transferase) ) under the SV40 promoter. Amphotropic retroviral packaging cells? CRIP were grown in increasing amounts of linamarin to determine the sensitivity of these cells to the compound. A value of 2 mg / ml of linamarin does not substantially decrease the survival values of the cells. No cyanide was detected in any of the types of cell lines tested (from C6,? CRIP and 3T3) or even after treatment with very high concentrations of linamarin, demonstrating the stability of the substrate and the absence of linamarase activity in these mammalian cells. The line of ecotropic murine packaging cells? CRE was transfected with the retroviral plasmid vector pL1 inSp and the preceding supernatants from? CRE turfs were harvested to infect individual clones producing? CRIP. Some clones showed a marked toxicity against the presence of linamarin, with estimated concentrations that resulted in a 50% cytotoxicity (IC 50) of 7-60 μg / ml (Figure 3a). The toxicity is due to the production of cyanide as measured by the Lambert method (Lambert et al., 1975). In a similar manner, C6 cells of rat glioblastoma were infected with the? CRElin. Clones from C61in generally gave IC50 values that were 50 times higher than individual clones from? CRIP. The cells of the packaging cell line FlyA13 (human fibrosarcoma) has an IC50 intermediate value of 195 μg / ml at a pH of 7.4 which decreases to 77 μg / ml at a pH of 6.6 (Figure 3c). Other cell lines whose CI5o values were estimated showed values closer to those of the C6 than those of the? CRIP (Table 1). The value of IC50 extraordinarily low, estimated for the? CRIP, could be related to a very high permeability for linamarin of these particular cells. After incubation with linamarin, the cell extracts from the C61in produced even greater amounts of cyanide than the? CRIPlin suggesting a very efficient expression of the gene in these cells. Isolation of the mRNA and Northern blot hybridization confirmed the colorimetric results.

In addition, in the experiments of the present invention, the toxicity of sodium cyanide on the cells was evaluated. It was found that cell lines of glioblas tumors (from C6, L9 or U373 MG) as well as an explant of glioblastoma from a human patient are 10 to 500 times more resistant to the toxic effect of exogenous sodium cyanide than the? CRIP. This finding could be related to the observation that cells from malignant tumors (C6) have a low oxidative respiration and a decrease in the number of mitochondria, making them quite resistant to metabolic poisons. such as KCN (Lichtor and Dorman, 1986; Maduh et al., 1991), a property that can be reversed in part by the conditions, low pH maintained within the tumor. The neoplastic cells that form a tumor are exposed to a substantial and consistent decrease in extracellular pH from 7.4 to 6.6 in the case of a glioblastoma, or lower for other types of tumors (Hu et al., 1988; Ger eck and Seetharaman, 1996 ) resulting in a greater sensitivity (50% or greater) to the inhibitory effects of sodium cyanate.The depletion of nutrients due to poor vascularization within solid tumors is probably the main cause of this acidity, a limited diffusion of Oxygen causes the induction of anaerobic glycolysis and lactic acid accumulation.In vi tro, crops are normally kept closer to pH 7.4 than pH 6.6, however, decreasing the pH of the culture to 6.6 reduced the IC50 of the cultures. cell lines tested The lower interstitial pH in tumors than in normal tissues would result in a greater and more specific sensitivity to the toxic effect of cyanide on e tumor cells. This is very convenient, since it will selectively increase the specificity of the therapeutic action towards cells of internal tumors instead of towards normal surrounding cells. The type of euthanasic effect expected in the present system is quite different from the surrounding effect of a prototype, dependent on the expression of canexins and cellular communication through interstitial joints (Wygoda et al., 1997; Vrionis et al., 1997). The toxicity to the neighbors of a cell containing linamarase is due to the release of cyanide, and the action of cyanide does not require cell-to-cell contact or interstitial joints. In order to determine the surrounding effect, experiments were carried out in which linamarase producing cells were mixed in a ratio of initial cells with non-producing fifty / fifty cells. The results of these experiments indicate that the surrounding effect in this system is very efficient, being proportional to the amount of cyanide released by the producer cells and the lethal dosage for cyanide in the neighboring cells. Based on the encouraging data obtained in vi tro, the in vi ve sensitivity of the system was evaluated using an intracerebral tumor model. Two types of experiments were performed: in the first of the types, C6 cells already transfected with pLlinSP and selected for resistance to puromycin, were inoculated into the brain of an animal and allowed to form a tumor. The glioblastomas were subsequently treated with linamarin. In the second type of experiments, brain tumors were induced by ordinary C6 cells, and the cells producing retroviruses of the? CRIPlin were inoculated into the growing tumor, 3 or 5 days later, the tumors were exposed to linamarin. Cells transfected with the C61in were inoculated into the white matter of the right cerebral hemisphere of Wistar rats. A few weeks later, an intracranial tumor was clearly visible by magnetic resonance imaging (magnitude of volume approximately 70 mm3). After the supply of linamarin (a total of 2.4 mg) over a period of 8 days, magnetic resonance imaging, taken three weeks after treatment, revealed a total absence of the tumor mass. One of the animals that had a similar tumor was treated with only 1 mg of linamarin. The animal finished the treatment, but eventually died as a consequence of the tumor, indicating a lower dosage limit for the prodrug. No apparent toxicity was observed by local administration of up to 9 g of linamarin in brains of normal rats under the same conditions as in the experimental animals. In the second type of experiments, C6 cells were inoculated into the right hemisphere of Wistar rats and allowed to develop to form tumors. Three of the animals that had larger tumors, with approximate magnitudes of 70 mm3 volume, were selected for further inoculation with retrovirus producing cells? CRIPlin and local treatment with linamarin. Total tumor eradication was achieved for the larger glioblastomas and for the smaller ones. Control animals with tumors that had approximately the same volume magnitudes, treated only with linamarin (6 mg) under the same conditions as described above, died shortly thereafter. The cured animals were well found for several weeks after treatment, with no apparent neurological damage and no recurrence. With the present invention, it was possible for the first time to eradicate a rat glioblastoma of the dimensions described. The toxicity for normal brain tissue was estimated by direct inoculation of cells producing the? CRIPlin, followed by treatment with linamarin. The images obtained by magnetic resonance of the treated animals serve as proof that the treatment itself does not produce any appreciable damage to the animals' brains. The regimen of transduction of tumor cells by retroviral vectors in vi is usually in the range of 1-10% of the population of tumor cells. An efficient surrounding effect (from "Good Sa aritano" or euthanasia) has been established as a hypothesis to explain the total remission of tumors in animal models by the Hs tk and gan ci cl ovi r system. The transfer of phosphorylated qanciclovir between cells through the interstitial joints appears as one of the best documented theories to explain the mechanism (Wygoda et al., 1997; Vrionis and. Co-workers, 1997). Some tumor cells, such as rat glioblastoma C6 cells for example, have lost the ability to synthesize some of the connexin proteins and therefore the benefit from this important effect will be severely limited. The combination of enzyme and prodrug, linamarase and linamarin, does not require any contact of cells with cells for their surrounding effect, since the evidence of the existence of cyanide in the cell culture medium indicates that it can cross cell membranes. In addition, the expected effect in vi should be greater than in vi tro. The volatile character of the cyanide makes it quite difficult to estimate its annihilating effect for the surrounding cells because the compound dissipates in the medium and in the air more rapidly than in the neighboring cells. On the other hand, within a tightly packaged tumor, the cells closest to the producers will be the first and main target of the released cyanide. The cyanide released into a tumor would also be more effective against neighboring cells because of the increased response effect at low extracellular pH (Hu et al., 1988; Ger eck and Seetharaman, 1996). The cells located outside the tumor, which have extracellular pH values close to 7.4, would be quite resistant to the surrounding effect of cyanide. In the experiments of the present invention, considerations concerning the possible toxic effects have been taken into account. One of the aspects that needed to be considered is related to the possibility of decomposition of linamarin in other parts of the body by enzymes or alternative situations; and the other aspect concerns the putative damage of healthy brain tissue surrounding the tumor. Regarding the first topic, there is a β-glucosidase in mammals that have a broad specificity and a dark biological function, which makes it possible for the organism to hydrolyze the toxic vegetable glycosides found in the diet. In general, the richest source of this cytosolic ß-glucosidase in most vertebrates is the liver and small intestine, except in mice and rats, where the kidney is the richest source of the enzyme (Glew). et al, 1993, Lima Górniak et al., 1993, Ne Mark et al., 1981). In the experiments of the present invention it was confirmed that the cyanogenic glycosides, such as linamarin, are not hydrolyzed by cultured mammalian cells, tumor or normal rat brain tissue and human brain tissue with tumor or normal. Furthermore, it should be considered that the microbial flora of the intestines could be able to metabolize, to a certain extent, some cyanogenic glycosides, such as amygdalin and linamarin, in the intestinal lumen by bacterial β-glucosidases, resulting in release and absorption of hydrogen cyanide. Female hamsters given orally a dose of 108 mg linamarin per kg of body weight show signs of cyanide poisoning that causes death in 18% of animals (Frakes and Sharma, 1986, Hernández et al. , nineteen ninety five) . The toxicity is much lower in animals receiving similar doses intravenously and is practically absent when administered intraperitoneally (Gle et al., 1993, Lima Górniak et al., 1993, Ne Mark et al., 1981). Local administration of linamarin is as safe, and probably safer, than intraperitoneal delivery to avoid gastrointestinal interference and is evident from the in vi tro and in vi ve assays of the present invention that it is well below the dangerous concentrations of the prodrug. An alternative route for the oral administration of cyanogenic glycosides is to pre-treat animals (or patients) with specially extended-spectrum antibiotics to reduce gastrointestinal flora, and that these are less susceptible to the toxicity of the compounds (Newton and collaborators, 1981; Pad aja, 1995; Halstrom and Moller, 1945). In many tropical communities, a potential risk is recognized for greetings from the consumption of leaves and roots of the cassava plant (Maniho t escul en ta Crantz). When mechanical damage to the tissues of this plant occurs, the cyanogenic glucoside (linamarin) is exposed to the action of the endogenous plant linamarase and the a-hydroxynitrile-lyase and hydrocyanic acid are released. The operations of cooking / boiling, desiccation, soaking and fermentation are the usual treatment techniques to deactivate linamarase and decompose linamarin before consumption, the occasional poisoning reported in humans is usually due to leaves or roots of inadequately processed cassava. It is reported that doses of cyanide from 50 to 100 mg are lethal to adults (Newton et al, 1981, Padmaja, 1995, Halstrom and Moller, 1945). This would be equivalent to 3 mg / kg of sodium or potassium cyanide and 500 to 1,000 mg of linamarin. The dose of linamarin potentially necessary to eradicate a human tumor with a diameter of about 4 cm is below the dangerous concentrations of the prodrug and could be administered for a period of one week. The second concern is related to the damage that the release of cyanide could cause in healthy parts of the brain. Cyanide can penetrate rapidly into the plasma membrane and accumulate in mitochondria and membrane elements of the neuronal cell, in a manner proportional to the available concentration. An advantage obtained from applying the present invention is the lower local extracellular pH created within the tumors and the increased susceptibility of such a condition to the cyanide. Physiological extracellular pH values in surrounding normal cells will make them quite resistant to the surrounding effect of cyanide. In one of the experiments in vi that have been performed, a large majority (if not all) of the tumor cells had incorporated the linamarase gene and could therefore transform linamarin into glucose, acetone and cyanide (C6 transfected with pLl i nSp). However, magnetic resonance images reveal a fibrous scar comparable to that previously obtained with the Hs tk and gan ci cl ovi r system (Culver et al, 1992; Ram et al., 1993; Izquierdo et al., 1995;; Izquierdo et al., 1997) or that of the C6 infected with the retroviral lesions, indicating the absence of a devastating effect on the brain. All the surviving animals were healthy and were well several months after treatment, with no apparent signs of neurological damage or any recurrence of the tumor. Glioblastoma is a particularly malignant tumor because of the infiltrating characteristic of its cellular constituents, recurrent tumors are generally not further than 2 cm from the primary site; It can be difficult to remove the infiltrated neoplastic cells without damaging some surrounding brain cells. In the past it has not been possible to cure rat glioblastomas with a volume magnitude greater than 150 mm3 using the Hs tk and ganci cl ovi r system (Izquierdo et al., 1997), which implies that the new approach with linamarase and linamarin It is more effective. When the system of Hs tk and ganci cl ovi r was applied to a small number of patients by other researchers and by the authors of the present invention (Culver et al., 1994, Izquierdo et al., 1996, Roth and Christiano, 1997), he generally recognized a reduction in the size of the genuine tumors in approximately 30% of the treated patients, but a total disappearance of the residual tumors has not been achieved despite several consecutive treatments. In many cases the survival time was increased, but the recurrence eventually killed the patient. The method of the invention has provided results in animal experiments that can be considered as very encouraging.

BRIEF DESCRIPTION OF THE FIGURES Figure 1 Enzymatic action of linamarase. Schematic representation of the decomposition of linamarin by linamarase to form acetone glucose and cyanohydrin and the subsequent spontaneous decomposition of this product that liberates free cyanide and acetone. Figure 2 The recombinant retroviral plasmid pL1 inSp. . Figure 3 Sensitivity to linamarin as a percentage of cell survival. A) Sensitivity of the? CRIP to linamarin, compared to that of a? CRIP clone infected with the retroviral l i n. B) Sensitivity of C6 to linamarin, compared to that of a C6 clone transfected with retroviral l. C) Sensitivity of FLYA13 to linamarin, compared with that of a clone of FLYA13 infected with retroviral lin. Figure 4 Magnetic resonance radiographs enhanced by gadolinium in sagittal projections (ai, b2) of a rat brain tumor, before (ai, a2) and after (bi, b2) a treatment with linamarin.

Figure 5. Gadolinium-enhanced magnetic resonance radiographs on axial (ai, bi), sagittal (a2, b2) and coronal (a3 / b3) projections of a rat brain tumor, before (ai, a2, a3) and after ( bi, b2, b3) of an inoculation of the? CRIPJin of mouse retrovirus producing cells and of a linamarin treatment. Figure 6. Axial T2 (ai, bi) and sagittal Ti (a2, b2) magnetic resonance radiographs of two control animals. Figure 7. Presence of cyanide in tissue homogenates, (a)? CRIPIin, (b) C6, (c) C6l in, (d) normal rat brain, (e) normal human brain, (f) human glioblastoma.

In the experiments of the invention, the following materials and methods were used a) Construction of a recombinant plasmid encoding linamarase The plasmid pCas 5 (Hughes et al., 1992) was opened with the restriction enzyme Sacll and filled with Klenow, then cut with Sal I to generate the Blunt blunt-Sal I fragment. fragment was inserted in the pBabe Puro (Morgenstern and Land, 1990), opened by the restriction enzymes Sna Bl / Sal I (the Sna Bl gives a blunt end, so that the insert was oriented by the enzyme Sal). The recombinant (Figure 2) is called Pll inSp and can be selected for its puromycin resistance (1-2.5 μg / ml puromycin) under the SV40 promoter; Linamarase is under the LTR promoter. The prokaryotic residue allowed a growth of the recombinant plasmid in bacteria before transduction of mammalian cells. b) Cell lines and culture Cells: C6 (ATCC CLL 107) and L9 (rat glioblastoma, Weizsaecker et al., 1981), MGU 373 (U-373 MG, ATCC HTB-17, human glioblastoma), NIH-3T3 (ATCC CRL6331, CRE (ecotropic packaging line, Danos and Mulligan, 1988), CRIP (amphotropic packaging line, Danos and Mulligan, 1988), Ag9181 (human fibroblast) and FLYA13 (Cosset et al., 1995) were grown in Eagle medium modified by Dulbecco (DMEM) Some of them were supplemented with 10% heat-deactivated fetal calf serum: C6, L9, MGU 373, Hs683, Ag981) and the rest of them were with calf serum . The cells were cultured at 37 ° C, with 6% C02 and a relative humidity of 97%. The pH of the DMEM was artificially lowered to pH 6.6 in some experiments by decreasing the concentration of NaHCO3 from 3.7 g / 1 to 0.7 g / 1. c) Transfections and infections The retroviral plasmid PllinSp was introduced into ecotropic CRE? packaging cells using the AVET technique for DNA delivery (Cotten et al., 1992); Wagner and collaborators, 1992; von Rüden et al., 1995) and clones were selected by puromycin (48 h later). The same plasmid was introduced into ecotropic CRE? Packaging cells. Transfections were performed essentially as described previously (von Rüden et al., 1995). 4 μg of pLli nSl in 150 μl of HBS was used to transfect 200, 000 cells for 4 hours at 37 ° C. After that, the medium containing the transfection complex was replaced by a fresh culture medium, which contained 10 μm of dexamethasone. 48 h later, for selection, 2 mg / ml puromycin was added to the cells. Two weeks later, the colonies were isolated using cloning rings. The viral supernatants from turf collected at 48 h after transfection were used to infect amphotropic CRIP (? CRIPlin) packaging cells. After selection with puromycin (1.5 μg / ml), colonies were isolated using cloning rings. The clonal lines were analyzed for the vector title, using NIH 3T3 cells. The best titles were 3-6 x 105 u.f.c./ml. The cell lines shown in Table 1 were infected with? CRIPlin supernatants (U373MG, Hs683, Ag9181, Sk-N-MC and FLYA13) or? CRElin supernatants (C6 and 3T3), to obtain l in + cell lines . After selection with puromycin, the colonies were isolated and sensitivity to linamarin was estimated. d) Marking of cells The cells were marked essentially as described previously (Tamura et al., 1997). 2 mg of Dil were dissolved (1, 1 '-dioctadecyl-3, 3, 3', 3 '-tetramet-il-indocarbocyanine perchlorate) in 1 ml of ethanol and 50 μl of the solution was added to 5 ml of cells in a culture tray of 6 ml. cm. After incubation for one hour at 37 ° C, the cells were trypsinized, washed twice with PBS, and resuspended in DMEM at a concentration of 5 x 10 4 cells / μl. e) Analysis of the sensitivity to linamarin and of the toxicity of cyanide in a culture Cells were seeded in a flat side tube (provided with a screw cap to prevent the release of HCN), with a density of 10 5 cells / tube. After incubating for 24 h, increasing concentrations of linamarin were added. Two days later, the percentage of surviving cells was determined to evaluate the toxicity of the drug. The cells were incubated for 2.5 h with the mitochondria substrate MTT (3- [4,5-dimethyl-thiazol-2-yl] -2,5-diphenyl-tetrazolium bromide, 200 μg / ml) which can be transformed into formazan only by living cells. Then the medium was removed and 3 ml of DMSO (dimethyl sulfoxide) was added to dissolve the formazan. Samples were measured by spectrophotometry at 540 nm after 10 minutes. The toxicity of the cyanide was analyzed as described above, except for the addition of increasing amounts of NaCN to the cultures. f) Measurement of linamarase activity and the presence of cyanide The activity of linamarase and the presence of cyanide were determined by Lambert's method (Lambert et al., 1975) with the following modifications: 1 ml of the succinimide and N-chlorosuccinimide mixture reagent was added to the flattened tubes, which contained the cells with different amounts of linamarin (48 h later) cultured in 1 ml of DMEM without phenol red. Then 1 ml of the barbituric acid and pyridine mixture reagent was added, mixed and the absorption was read at 592 nm after 10 minutes (calibration curves were made with NaCN). g) Tissue homogenised materials Tissues were shredded and homogenized in five volumes of a medium containing sucrose (0.25 M), ethylenediamine tetraacetate (1 mM) and Tris HCl buffer (0.01 M, pH 7.2) supplemented with the protease inhibitor phenylmethylsulfonyl fluoride (PMSF) 1 mM (all the processes were carried out at 1-4 C). The homogenized material was centrifuged (with 600 xg, for 10 minutes). The supernatant fraction was stored at -70 ° C (Glew, R. et al., 1976). The protein content was measured by the BCA Protein Analysis. The experiment reproduced in Figure 7 was carried out at 37 ° C, with incubation for 4 hours with 1 mM linamarin, 0.2 mg protein, pH 7.4, in a total volume of 1 ml. The presence of cyanide was determined by the Lambert method (Lambert et al., 1975). h) Inoculation in the brain of a rat of cells transfected with the C6 lin and administration of linamarin The processes applied were essentially those previously described (Izquierdo et al., 1995, Izquierdo et al., 1997). Male Wistar rats weighing 250-300 grams were anesthetized with a mixture of ketamine (50 mg / ml), valium (5 mg / ml) and atropine (1 mg / ml) in a volume ratio of 5: 4: 1 in a dose of 0.3 ml / 100 g of body weight before placing them in a device is tereotaxic. C6 glioma cells or the same cells transfected with the plasmid pL1 inSp and selected for puromi-crna resistance were injected at a concentration of 10 5 cells / μl in a complete PBS (with calcium and magnesium) supplemented with 0.1% glucose. With the help of the manipulation arm of the stereotaxic apparatus a total of 10 μl was introduced, during a 5 minute interval, into the fronto-parietal lobe of the right cerebral hemisphere (at 4 mm to the right of the bregma and at a depth of 4.5 mm from the skull) using a Hamilton syringe of 10 μl capacity, connected to a 26 gauge needle; The needle was kept in place 3 minutes before and after the injection. The? CRIPlin-producing cells were inoculated at the same coordinates as tereotaxic than C6 cells except for a very wide tumor in which an additional tumor site had also been reached, the cells were delivered at different heights throughout the tumor. All rats received tetracycline in drinking water (approximately 75 mg / kg) and dexamethasone (1 mg / 500 ml) for three days before and one week after surgical intervention. Treatment with linamarin was initiated after visualization by magnetic resonance of the tumor and stimulation of its size. The prodrug was administered by an infusion cannula in the brain placed at the site of tumor inoculation and fixed to the skull using dental cement; a small stainless steel screw acts as an anchor to hold the cannula and dental cement. The cerebral infusion cannula is then connected subcutaneously (s.c.) to an osmotic pump (alzet® model 2001) that delivers 1 μl / h for 7 or 8 days. The osmotic pump was filled with the appropriate amount of linamarin in PBS. j) Magnetic resonance imaging. Sagittal and coronal radiographs of the skull were taken with a small Cpflex coil around the animal. Representations were obtained in images of sagittal and axial TI projections using a spin echo, with TR 650, TE 17.0 / 1 and acquisition 2, 3 mm slides. The field of vision 50x100. The 256x512 matrix, high resolution sequence after gadolinium enhancement (0.2 ml / kg) with M magnetization transfer. k) Estimation of tumor volumes Tumor volumes were estimated as follows. If [O / L] is a parameterization of the tumor interval in the sagittal section and C (x), O < _ x. < _ L, designates the area of the corresponding coronal section, then you have the exact formula for the volume of the tumor Vsjf - t. which provides the approximation and -rJ. I? i In practice, however, it was necessary that the volume be estimated with the knowledge of only one section of each type. These tumors are usually adapted to the geometry of the head of the rat: they tend to be quite symmetrical (elliptical) in the coronal sections, while they have. some long and irregular sagittal configurations. Having been given two "well-taken" sections, the coordinates were introduced: coronal view sagittal view A plausible model for the tumor is given by the volume generated by providing ho sections for Co in each position x with a relation S (x) / d0. Therefore, the area ratio is [S (x) / dQ] 2 r where CQ = area of the given coronal section D0 = diameter of that coronal section in the z-direction (sagittal) S (x) = length of the sagittal section in the z direction at the x position. N was chosen taking into account the configuration of the sagittal image. It is possible to obtain similar formulas by changing sagittal by axial. However, taking into account the normalized geometries of these tumors, the axial section seems to be less significant than the sagittal. In general, it was expected to have an estimate from below using the axi - coronal approach.

Example 1 Construction of recombinant retroviruses carrying the linamarase gene A therapeutic retrovirus was engineered by inserting the linamarase gene (ca 5, Hughes et al., 1992) in Moloney murine leukemia vector pBabepuro (Morgenstern and Land, 1990). The resulting recombinant retroviral plasmid carries the cas 5 gene expressed under the control of the LTR promoter (long terminal repeat) at 5 'and the selection gene pa c (puromycin-N-acetyl transferase) under the SV40 promoter (Figure 2 ). The estimated size of the retroviral plasmid was 6.9 kilobases (kb) and this was designated pLlinSp (plasmidid, LTR, linamarase, SV40 promoter, puromycin-N-a cetyl-transferase). The recombinant pLI i nSpL can be selected for its puromycin resistance (1-2.5 μg / ml puromycin) under the SV40 promoter; l i namara sa is under the LTR promoter.

Example 2 Determination of sensitivity to linamarin of retrovirus-infected cells and of the toxicity of sodium cyanide to cells a) Sensitivity to linamarin as a percentage of survival of cells i) The sensitivity of? CRIP to linamarin was compared to that of a clone of? CRIP infected with retroviral lin, in increasing concentrations of linamarin and at two pH values. The initial number of cells was 105 in all cases. After incubating for 24 hours, increasing concentrations of linamarin were added. Two days later, the percentage of surviving cells was determined to evaluate the toxicity of the drugs. The clonal lines were analyzed for the vector titer, using NIH 3T3 cells. The best titles were 3-6 x 105 u.f.c./ml. The results of the experiment are shown in Figure 3a. Some clones showed marked toxicity to the presence of linamarin, with estimated concentrations that resulted in a 50% cytotoxicity (CI5o) of 7-60 μg / ml. ii) The sensitivity of C6 to linamarin was compared with that of a C6 clone infected with retroviral linas, at increasing concentrations of linamarin and at two pH values. The initial number of cells was 105 in both cases and the IC50 estimated for this clone of C 61 in is d.e. 345 μg / ml at pH 7.4 and 202 μg / ml at pH 6.6. The results of the experiment are shown in Figure 3b. iii) The sensitivity of FLYA13 to linamarin was compared with that of a clone of FLYA13 infected with retroviral lin, at increasing concentrations of linamarin and at two pH values. The IC50 estimated for this FLYA131in clone is 195 μl / ml at pH 7.4 and 77 μl / ml at pH 6. 6 The results of the experiment are shown in Figure 3c. The sensitivity of other cell lines, whose CI5o was determined in the same way, is shown in Table 1. b) Cyanide toxicity Next, the toxicity of sodium cyanide on the cells was evaluated. The chemical compound itself is more toxic by an order of magnitude (ten times more) than linamarin in both CRIP and C6. This result was interpreted as showing that only 10% of the linamarin applied to the medium ends in the form of cyanide. It was also found that the cell lines of glioblastoma tumors (C6, L9, or U373MG) as well as an explant of glioblastoma from a human patient, are 10 to 500 times more resistant to the toxic effect of exogenous sodium cyanide than the? CRIP (and therefore the 3T3, from which the CRIPs have been derived), a property that can be partially related to the pH conditions maintained in vi tro and in vi vo.

Example 3 Intensity and surrounding effect of cyanide In order to determine the surrounding effect in this system, experiments were performed in which linamarase producing cells were mixed in a ratio of initial cells to non-producing fifty / fifty cells. One of the mixtures was the line of packaging cells? CRIP infected with the retrovirus carrying l in (? CRIPlin) cells plus normal C6 cells, marked in red with Dil (1, 1 'perchlorate -dioctadecyl-3, 3, 3 ', 3' -tetramethyl-indocarbocyanin (Tamura et al., 1997) In the presence of 2.5 mg / ml of linamarin, both cell types, (? CRIPlin and C6 labeled), died, while 86% of the C6 alone they will survive at linamarin concentrations of 10 mg / ml Lower concentrations of linamarin (100 μg / ml) only partially affect the viability of any of the cell types in this mixture, despite the fact that? alone they would be very sensitive to this concentration (Figure 3a) .These results were interpreted as showing that the C6 cells, which are more resistant to cyanide than the? CRIP, are catching enough cyanide to decrease the free concentration in the medium. Mixtures of? CRIP infect All carriers with the retro-virus carriers or with either CRIP or 3T3 marked red ("CRIPlin /? CRIP red or? CRIPJin / 3T3 red) were devastated at the lowest concentration of linamarin (100 μg / ml) due to the greater sensitivity of the cells marked from red to the cyanide released by the producer cells. The results indicate that the surrounding effect in this system is very effective, being proportional to the amount of cyanide released by the producer cells and the lethal dosage for the cyanide in neighboring cells.

Example 4 Curability of a rat glioblastoma in vivo Based on the encouraging in vi tro data, the in vivo sensitivity of the system was evaluated using an intracerebral tumor model. Two types of experiments were performed: in the first type, C6 cells already transfected with pL1 i nSp and selected for puromycin resistance were inoculated into the brain of the animal and allowed to form a tumor. The glioblastoma would subsequently be treated with linamarin. In the second type of experiments, brain tumors were induced by ordinary C6 cells and CRIPlin retrovirus producing cells were inoculated into the growing tumor, 3 or 5 days later the tumors were exposed to linamarin. A total of 106 cells transfected with Cßl in were inoculated, in the first type of experiments, into the white matter of the right cerebral hemisphere (4 mm to the right of the bregma and 4.5 mm of depth from the skull) of three rats Wistar. A few weeks later (from 2 to 4), an intracranial tumor was clearly visible by magnetic resonance imaging. (volume magnitude around 70 mm3) (Figure 4 ai, a2). One day later, linamarin was administered through a direct infusion cannula into a brain tumor, connected by a catheter tube to an osmotic minipump. A total of 2.4 mg of linamarin was supplied over a period of 8 days. The release was carried out at a constant rate or flow rate of 1 μl per hour (12.5 μg of linamarin / μl).

Figure 4 shows the gadolinium-enhanced magnetic resonance radiographs in sagittal (ai, bi) and axial (a2, b2) projections of a rat brain tumor, before (ax, a2) and after (bi, b2) of the treatment with linamarina A brain tumor (magnitude of volume 70 mm3, arrow ax a2) of cells transfected with C61in of glioblastoma, had been eradicated completely (arrow bi b2) after treatment for a week with constant local release of 12.5 μg of linamarin per hour. The magnetic resonance image taken three weeks after the treatment (Figure 4 bi, b2) demonstrates a total absence of the tumor mass and instead a residual hypointense image at the injection site, interpreted as a fibrous scar or as a small residual of hematoma with internal hemosiderin. One of the animals, which had a tumor of similar size, was treated with only 1 mg of linamarin. The animal finished the treatment but finally died as a consequence of the tumor, indicating a lower dosage limit for the prodrug. In the second type of experiments, 106 C6 cells were inoculated into the right hemisphere of 10 more Wistar rats and allowed to develop into the tumor form. Three of the animals with larger tumors (volume estimation is shown in the materials and methods section) with approximate magnitudes of volumes of 400 mm3, 600 mm3 and 1,000 mm3 (Figure 5 ai, a2, a3), selected for subsequent inoculation with retrovirus-producing cells CRIPLIN (7 x 106 cells for the two minor and 107 cells for the major) and for local treatment with linamarin. The murine retroviral producing cells (? CRIPlin) had a titer of 5 x 105 u.f.c. (colony forming units) and inoculated into a single site (the same one used for C6 delivery) in all tumors except the largest one, where a small additional craniotomy was made at a distant site within the tumor, in order to increase the potential access of retroviruses to all neoplastic cells. A few days later (three for the larger tumor and five for the others), 6 mg of linamarin was administered in 200 μl at a steady flow rate of 1 μl per hour. (supply of the osmotic pump). Total tumor eradication was achieved for the major glioblastomas (Figure 5 b3, b2, b3) and the minor ones; the animal with the tumor of intermediate magnitude, on the other hand, also finished the treatment but died a month later. The examination of the brain of the dead animal showed that a hemorrhage within the tumor is still probably the cause of death. Figure 5 shows gadolinium-enhanced magnetic resonance radiographs in axial (ai, bi), sagittal (a / b2) and coronal (a3 / b3) projections of a rat brain, before (ax, a2, a3) and after (bx, b2, b3) of the inoculation of retroviral producing cells? mouse CRIPlin and linamarin treatment. As can be observed, a very large brain tumor (with a volume magnitude of approximately 1,000 mm3, arrows ai, a2, a3) of C6 cells from rat glioblastoma, had been completely eradicated as shown on the three MRI radiographs (arrows bi, b2 and b3). These images show the appearance of the live brain after therapeutic inoculation of retrovirus-producing murine cells within the tumor (? CRIPlin) and treatment for one week, three days later, with constant local release of 30 μg of linamarin per hour. Two control animals with tumors having volume sizes of 220 and 140 mm3, which had been treated only with the prodrug (6 mg) under the same conditions as described above, died shortly thereafter. The cured animals were well, without any apparent neurological damage or recurrence. Some of the treated animals showed no apparent neurological damage for a period of up to six months, but they developed a recurrence later. This was observed only in very large tumors (larger than 500 mm3) in approximately 50% of cases. In all cases with recurrence, the tumor mass had crossed the interhemispheric fissure and reached the ventricles. At this point, the tumor cells could be distributed via the cerebrospinal fluid to parts of the brain quite distant from the original inoculum. Of course, some of the recurrent tumors were multiple and were far from the original site. Figure 6 shows axial T2 (ai, bi) and sagittal axial magnetic resonance radiographs (a2, b2) of two control animals that had been treated in the right hemisphere with 5 x 10 6 cells producing retroviruses in mice and then with 4 (a) and 6 (b) mg of linamarin. The image of the left hemisphere can be considered, as a witness of untreated brain tissue. The toxicity to normal brain tissue was estimated by direct inoculation of? CRIPL? N producing cells, followed by treatment with linamarin (2.5, 4 or 6 mg) in the absence of tumor cells. The magnetic resonance of the treated animals does not reveal any large damage, but a small scar similar to that normally present in brains after the tumor had been eradicated (Figure 6). We analyzed two types of relaxation time images (in Tesla): Ti and T2 in order to visualize lesions other than tumor. In general, Ti-weighted images mainly provide anatomical information, images weighted by T2 are more sensitive to early pathological changes in most tissues. No injury related to neoplasm, edema, ischemia, infection, demyelination or trauma was observed, it was observed that the animals exhibited only small abnormalities in the MRI (magnetic resonance image) and did not exhibit large lesions. These results can be interpreted as proof that the treatment itself does not produce any appreciable damage to the animals' brains.

Example 5 Toxicity experiments The activities of ß-glucosidase and linamarase were estimated in various tissues from rats and humans. It was expected to obtain ß-glucosidase activity but not linamarase activity in most of the tissues analyzed. The tissues tested from a rat were: normal brain, glioblastoma C6, liver, kidney and intestines. The normal brain and cells from an explant of glioblastoma were tested on humans. The tissues were comminuted and homogenized in five volumes of: 0.25 M sucrose, 1 M ethylenediamine and 0.01 M Tris-HCl pH 7.2, and 1 mM phenylmethylsulfonyl fluoride as a protease inhibitor; at 1 - 4 ° C. The homogenized material was centrifuged (10 minutes, 600 xg) and the supernatants were stored at -70 ° C. The protein content was measured with the BCA Protein Analysis (Pierce). Enzyme determinations of the glucosidase and linamarase activities were carried out at 37 ° C. The glucosidase activity was measured as p-nitrophenol release (yellow, absorbance at 400 nm) from p-nor trofenyl-β-D-glucoside. Linamarase activity was estimated as cyanide release measured by the Lambert method (Lambert et al., 1975). In this analysis, flat side tubes with screw caps were used in order to avoid the release of cyanide, 0.1, 0.2, 0.5 and 1 mg of tissue proteins were estimated. The amounts of linamarin were 1 M, 5 mM, 10 mM and 20 mM7 The reaction was measured at two pH values, pH 6 and physiological pH 7.4. The positive control was constituted by transfected C dl i n cells and transfected CRIPIin cells. All human tissues gave negative results in terms of linamarase activity. In rats only the kidney gave low but detectable levels of cyanide production. Figure 7 shows the presence of cyanide in tissue homogenates, (a)? CRIPlin, (b) C6, (c) Cdlin, (d) normal rat brain, (e) normal human brain, (f) human glioblastoma .

Table 1. Sensitivity of different cell lines to linamarin Cell line% of IC50 cells μg / ml of linamarin survivors clones lln with 1 mg / ml best clone mean value of linamarin C6 100a 345 407 Glio a rat U-373 MG 80 145 236 Human Glioblastoma Hs 683 80 230 288 Human Glioma Sk-N-MC 80 183 215 Human Neuroblastoma FLYA13 88a 195 382 Human Amphotropic packaging cells? CRIP 85 32 Anfotropic murine packaging cells NIH 3T3 90a N.D. N.D. Murine fibroblasts Ag 9181 89a N.D. N.D. Human fibroblasts N.D .: not determined a In these cell lines the concentration of linamarin was increased to 5 mg / ml without any substantial drop in survival values. In the other cell lines it was not determined. b The mean value was determined with at least 4 different clones. Clones with IC 50 greater than 500 μg / ml were not considered.

References Abaza, M.S. et al., 1997, J. Cancer Res., 88 (11) 1094-1099 Andreansky, S.S., et al., 1996, Proc. Nati Acad. Sci USA, Vol. 93, Oct., 11313-11318 Bilbao, G. et al., 1997, FASEB J. Vol. 11, 624-634 Bonnekoh, B., et al., 1995, J Invest Dermatol, 104, 313- 317 Braas, G. et al., 1996, Bioseparation, 6, 211-228 Chen, S.H., et al., 1994, Proc. Nati Acad. Sci., USA, 91, 3054-3057 Cirielli, C, et al., 1997, J Neuro-Oncology, 31, 217-223 Cosset, FL., Et al., 1995, J. of Virology, 69, 7430-7436 Corten , M., et al., 1992, Proc. Nati Acad. Sci., USA, 89, 6094-6098 Culver, KW, et al., 1992, Science, 256, 1550-1552 Culver, K., et al., 1994, Human gene therapy, 5, 343-379 Danos, 0., and Mulligan, 1988, Proc. Nat. Acad. Sci. USA, 85, 6460-6464 Deonarain, 1994, Gene Therapy, 1, 149-151 Descamps, V., and collaborators, 1996, J. Mol. Med. 74, 183-189 Fields, M.A., and Gerhardt, KE., 1994, Electrophoresis, 15, 654-661 Felgner, P.L., et al., 1987, Proc. Nati Acad. Sci. USA, 84, 7413-7417 Felgner, J.H., et al., 1994, J. Biol. Chem. 269, 2550-2561 Frakes, R.A., and Sharma, R.P., 1986, Fd. Chem. Toxic. 24, 417-420 Frehner, M., and Conn, EE, 1987, Plant Physiol., 84, 1296-1300 Gerweck, LE, and Seetharaman, K., 1996, Cancer Research, 56, 1194-1198 Glew, RH, et al., 1993, of ß-glucosidases: biochemistry and molecular biology, compiler A. Zsen, Am. Soc. Symp., series 533 Gmelin, RH, et al., 1973, Phytochemistry, 12 (2), 457-461 Guenzburg, W.H., and Salmons, B., 1995. Molecular Med Today, Reviews, 410-417 Guenzburg, W.H., et al., 1996, Cytokines and Molec. Therapy, Vol. 2, 177-184 Halstrom, F., and Moller, K.D., 1945, Acta Pharmacol. Toxicol 1, 18 Hernández, T., et al., 1995, Natural Toxins, 3, 114-117 Hu, JJ, et al., 1988, Biochem, Pharmacol., 37, 2259-2266 Hug, P. Y Sleight, RG, 1991 , Biochimica et Biophysica Acta, 1097, 1-17 Hughes, MA, and Dunn, MA, 1982, Plant Molecular Biology, 1, 169-181 Hughes, MA , et al., 1992, Archives of Biochemistry and Biophysics, 295, 273-279 Itoh-Nashida, T., et al., 1987, J. Biochem., 101, 847-854 Izquierdo, M., et al., 1995, Gene Therapy, 2, 66-69 Izquierdo, M., et al., 1996, Gene Therapy, 3, 491-495 Izquierdo, M. et al., 1997, Acta Neurochirurgica, supl. 68, 111-117 Jensen, SR, and Nielsen, BJ, 1973, Acta Chem. Scand., 27 (7), 2661-2 Kakes, P., 1985, Planta, 166, 156-160 Kovesdi, I., and collaborators, 1997, Curr. Opin. Biotechnol., Oct. 8 (5), 583-589 Kramm, CM, et al., 1995, Brain Pathology, 5, 345-381 Lambert, JL, et al., 1975, Analytical Chemistry, 47, 916-918 Latchmah, DS , 1994, Molecular Biotechnol., Vol. 2, 179-195 Lichtor T., and Dohrmann, GJ , 1986, Neurosurgery, 19, 896-899 Lima Górniak, S., et al., 1993, J. of Ethnophaimacology, 38, 85-88 Maduh, E.U. , et al., 1991, Journal of applied toxicology, 11, 97-101 Mao, C.H., and Anderson, L., 1967, Phytochemistry, 6 (4), 473-483 Mkpong, O.E., et al., 1990, Plant Physiol., 93, 176-181 Morgenstern, J.P. and Land, H., 1990, Nuc. Acids Res. 18, 3587-3596 Nartey, F., 1968, Phytochemistry, 7 (8), 1307-1312 Newmark, J., et al., 1981, Proc. Nati Acad. Sci. USA, 78, 6513 Newton, GW, et al., 1981, West J. Medicine., 134, 97-103 Nishi, T., et al., 1996, Cancer Res, 56, 1050-1055 Qxtoby, E., et al., 1991, Plant Molecular Biology, 17, 209-219 Padmaja, G., 1995, Crit. Rev. in Food and Nutl., 35, 299 Peltékian, E., et al., 1997, J Neuroscience Methods, 71, 77-84 Ra, Z. et al., 1993, Cancer Research, 53, 83-88 Roth, J.A. , and Cristiano, R.J., 1997, Journal of the National Cancer Institute, 89, 21 Runnebaum, I.B., 1997, Anticancer Res, 17, 2887-2890 Rüden von, T. and collaborators, 1995, BioThechniques, 18, 484-489 Sambrook, J. et al., 1989, Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY San, H. and collaborators, 1993, Hum. Gene Ther., 4, 781-788 Selmar, D. et al., 1987 a), Plant Physiol., 83, 557-563 Selmar, D., et al., 1987 b), Analyt. Biochemistry, 166, 208-211 Stevens, DL, and Strobel, GA, 1968, J. Bacteriol., 95 (3), 1094-1102 Tamura, M., et al., 1997, Human Gene Therapy, 8, 381-391 Vile, RG, and Russell, SJ, 1995, British Med Bulletin, Vol. 51, No. 1, 12-30 Vrionis, F.D., et al., 1997, Gene Therapy, 4, 577-585 Wagner, E., et al., 1992, Proc. Nati Acad. Sci. USA, 89, 6099-6103 Weizsaecker et al., 1981, J. Neurology, 224, 183-191 Wygoda, MR, et al., 1997, Cancer Research, 57, 1699-1703 Yoshida, J., and Mizzumo, M., 1994, J. Neuroonc, 19, 269-274 Yung, WKA, 1994, Curr Opinion Neurol, 7, 501-505 It is noted that in relation to this date, the best method known to the applicant to carry out the aforementioned invention, is the conventional one for the manufacture of the objects or products to which it refers.

Claims (23)

CLAIMS Having described the invention as above, the content of the following claims is claimed as property:

1. Use of a nucleic acid molecule encoding a glycosidase together with a non-toxic cyanogenic glycoside substrate, which after degradation by glycosidase produces hydrogen cyanide, for the preparation of a medicament intended for the separate administration of the two components that They interact in the treatment of brain tumors in animals or humans.

2. The use according to claim 1, wherein the glycosidase is a glucosidase.

3. The use according to claim 2, wherein the glucosidase is a β-glucosidase.

4. The use according to claim 3, wherein the β-glucosidase is linamarase.

5. The use according to any of claims 1 to 4, wherein the nucleic acid molecule is a DNA molecule.

6. The use according to any of claims 1 to 5, wherein the nucleic acid molecule is a DNA or RNA molecule that is carried by a viral vector.

7. The use in accordance with the claim 6, in which the vector is a retroviral vector.

8. The use in accordance with the claim 7, in which the virus title is _ > l * - * 6 colony forming units.

9. The use according to claim 7, wherein the medicament contains packaging cells that have been infected with retroviral vectors.

10. The use according to claim 9, wherein the medicament contains > 109 cells.

11. The use in accordance with the claim 6, in which the viral vector is an adenoviral vector

12. The use in accordance with the claim 6, in which the viral vector is a Herpes Simplex vector.

13. The use according to claim 5, wherein the DNA molecule is carried by a plasmid.

14. The use according to claim 13, wherein the DNA plasmid is contained in a liposome.

15. The use according to claim 1, wherein the substrate is a cyanogenic monosaccharide.

16. The use according to claim 3 and claim 15, wherein the monosaccharide is glucose.

17. The use according to claim 16, wherein the substrate is linamarin.

18. Case for the treatment of brain tumors, wherein a first container contains the preparation with the nucleic acid molecule encoding the glycosidase, and wherein a second container contains the cyanogenic substrate.

19. The kit according to claim 18, characterized in that the glycosidase is linamarase and the substrate is linamarin.

20. The kit according to claim 18 or 19, characterized in that the first container contains a preparation of retroviral vectors.

21. The kit according to claim 20, characterized in that the first container contains a preparation of producer cells that have been infected with vectors * retrovi icos.

22. The case of. according to claim 18 or 19, characterized in that the first container contains a preparation of adenoviral vectors.

23. The kit according to claim 18 or 19, characterized in that the first container contains plasmid DNA.