WO2006110950A1 - In vivo model for preclinical drug development - Google Patents

Abstract

The invention provides an in vivo model tumour for use in determining the response of a tumour to a drug. Thus, the invention provides a tumour model which is derived from a cell which is capable of producing an imageable tumour when introduced into a non-human animal. The invention provides an in vivo method for assessing an effect of a test compound on the model tumour. Preferred cells used for the tumour model include GIST (gastrointestinal stromal tumour) cells, cells harbouring mutations in c-KIT and/or PDGFRA and particularly FDC-Pl cells. The preferred model is a mouse.

Description

IN VIVO MODEL FOR PRECLINICAL DRU G DEVELOPMENT

Field of the invention

The present invention relates to an in vivo model tumour for use in determining the response of a tumour to a drug.

Background to the invention

Until about 1990, most gastrointestinal sarcomas were considered to be leiomyosarcomas because they resembled smooth muscle histologically. However, clinical oncologists observed a distinctly lower rate of response to standard doxorubicin-based regimens among leiomyosarcomas that arose in the gut than among those that arose in the uterus, trunk, or arms and legs. As early as 1983 immunocytochemical studies of gastrointestinal sarcomas documented their frequent absence of muscle markers that were typical of leiomyosarcomas located elsewhere in the body.

The ability to accurately assess the effects of compounds in vivo is desirable in a wide range of pharmacological and related studies. The ability to assess tissue responses to compounds is particularly important in each stage of drug discovery, development and clinical application. Studies of tissue response may be particularly useful in the early validation of lead compounds during the drug discovery process, in clinical trials of potential therapeutic compounds, and in monitoring the efficacy of therapeutic compounds already used in clinical practice.

Evaluation of tissue response to compounds in vivo may be obtained by administering one or multiple doses of a test compound to an animal, killing the animal, and performing anatomical examination of body organs, histological examination of tissue, and analysis of body fluids. This approach to characterization of pharmacological activity has several significant limitations. Firstly, anatomical and histological studies may not precisely reflect significant changes in the activity of relevant biochemical pathways in normal and abnormal tissue after administration of test compounds. Secondly, because animals must be killed for anatomical and histological studies, serial measurements cannot be performed on an individual animal during and after repeated administration of a test compound. Finally, processing of tissue for histological examination is labour-intensive, time consuming, and expensive, and often represents a serious rate-limiting step in applications such as high-throughput drug discovery.

Thus, it would be of great value to medical research and drug development to be able to rapidly assess the potential response of a malignant tumour in vivo to one or a combination of chemotherapeutic agents. It would also be of great value to be able to predict the outcome of treatment based on quantitative measurements of tumour response after initiation of chemotherapy.

The discussion of documents, acts, materials, devices, articles and the like is included in this specification solely for the purpose of providing a context for the present invention. It is not suggested or represented that any or all of these matters formed part of the prior art base or were common general knowledge in the field relevant to the present invention as it existed before the priority date of this application.

Summary of the invention

In one aspect the present invention provides an in vivo model tumour for use in determining the response of a tumour to a drug.

The present invention provides a model tumour which is derived from a cell which is capable of producing an imageable tumour when introduced into a non-human animal. In one form of the invention the model tumour is suitable for studying drug responses by tumours including, but not limited to, GIST and tumours harbouring mutations in c-KIT and/or PDGFRA. In another form of the invention the model tumour includes a cell harbouring one or more c-KIT and/or PDGFRA mutation(s). In another form of the invention the cell harbours one or more c-KIT mutations, such as V560G and/or D816V. In further embodiments, the c-KIT and/or PDGFRA mutation may be introduced into the model tumour cells. In another form of the invention the mutation recapitulates one or more mutation(s) found in a clinical sample of a GIST. In another form of the invention the model tumour is a model GIST. In particular embodiments, the model tumour is present in a non-human animal.

Sarcomas in the subgroup without muscle or Schwann-cell (i.e., S-100 antigen) markers are termed GISTs. GISTs are uncommon tumours of the gastrointestinal tract. GISTs arise from interstitial cells of Cajal (ICCs) that are part of the autonomic nervous system of the Gl tract. Approximately 70% of GISTs develop in the stomach, 20% in the small intestine, and less than 10% in the esophagus, colon, and rectum. GISTs are typically more cellular than other gastrointestinal sarcomas. They occur predominantly in patients who are 40 to 70 years old but in rare cases may occur in younger persons. It is estimated that between 3000 and 6000 new cases of GIST are diagnosed each year in the US (American Cancer Society).

Most GIST harbour activating mutations in the c-KIT proto-oncogene or in platelet derived growth factor-receptor alpha (PDGFRA) that encode receptor tyrosine kinases. The dependence of GISTs upon constitutive signalling from the KIT-receptor has formed the basis for introduction of molecularly targeted therapies for this. Among these, imatinib (imatinib mesylate, Glivec, STI517) is a highly selective, potent small molecule inhibitor of three type I tyrosine kinases: PDGFR, c-KIT and the ABL kinase. Imatinib has proven to be an effective therapy for chronic myeloid leukaemia (targeting the BCR-ABL fusion protein); GIST (targeting c-KIT); and dermatofibrosarcoma protruberans (targeting PDGFRA). Prior to the introduction of imatinib as a clinically effective treatment for GIST, no effective systemic therapy existed for patients with unresectable GIST.

For these reasons, GIST serves as a very useful paradigm for understanding the molecular basis of cancer as well as providing a model for the development of targeted therapies in solid organ malignancies. In approximately 60% of cases of GIST, mutations in the c-KIT gene are found in the juxtamembrane domain, such as in-frame deletions and point mutations in exon 11. The reported rate of mutation ranges from 21 to 88%.

While GISTs are generally highly sensitive to imatinib, resistance frequently arises, often associated with the acquisition of mutations in c-KIT. Furthermore occasional patients manifest primary resistance to imatinib. Tumours with mutations in c-KIT exon 11 are associated with significantly better response rates (83%) than tumours with wild-type (0%) or exon 9-mutant genotypes (49%). Interestingly however, cells expressing these genotypes are equally sensitive to imatinib in vitro. It is suggested that acquisition of new mutations may be involved in the development of resistance in GIST treated with imatinib. Mutations associated with imatinib resistant GIST have been identified in exons 13 and 17 of the c-KIT gene and exons 18 of PDGFRA.

The present invention also provides an in vivo model for GIST. In one form of the invention the in vivo model for GIST includes a mouse which has a FDC-P1 cell introduced into the mouse. In certain embodiments of the invention the FDC-P1 cell is introduced subcutaneously into the mouse. In one form of the invention the FDC-P1 cell is present in a cell support matrix such as Matrigel. In one form of the invention the mouse is a DBA/2J mouse. The FDC-P1 cell may include a mutant c-KIT and/or PDGFRA gene. In one form of the invention the c-KIT mutation is selected from V560G, D816V or recapitulates one or more mutation(s) found in a clinical sample of a GIST. Examples of such c-KIT and PDGFRA mutations include, but are not limited to, for c-KIT: AY501-502 duplication, V560G, Del557-558, V654A, T6701, N822K, Y823D, D816H, D816V and del815-816; and for PDGFRA: D842V and V561 D. The FDC-P1 cell may include one or more mutations in either or both of a c-KIT and PDGFRA gene.

The present invention also provides a method of preparing an in vivo model tumour, wherein the method includes the step of introducing into a non-human animal a cell that produces an imageable tumour when introduced into the non- human animal. In one form of the invention, the method includes the step of introducing a cell into a mouse. In another form of the invention, the method of involves preparing an in vivo model of GIST, wherein the method includes the step of introducing a FDC-P1 cell into a mouse. In certain embodiments of the invention the FDC-P1 cell is introduced subcutaneously into the mouse. In one form of the invention the FDC-P1 cell is present in a cell support matrix such as Matrigel. In one form of the invention the mouse is a DBA/2J mouse. In certain embodiments the FDC-P1 cell harbours a mutant c-KIT and/or PDGFRA gene. In one form of the invention the c-KIT mutation is selected from V560G, D816V or recapitulates one or more mutation(s) found in a clinical sample of a GIST. Examples of such c-KIT and PDGFRA mutations include, but are not limited to, for c-KIT: AY501-502 duplication, V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V and del815-816; and for PDGFRA: D842V and V561D. The FDC-P1 cell may include one or more mutations in either or both of a c-KIT and PDGFRA gene.

In a further aspect the present invention provides a use of an FDC-P1 cell in a model tumour. In one form of the invention the model tumour is imageable. In another form of the invention the model tumour is a model GIST wherein the FDC-P1 cell is provided to a non-human animal in order to produce a model GIST. In certain embodiments the non-human animal is a mouse. In other embodiments the FDC-P1 cell harbours a mutant c-KIT and/or PDGFRA gene. In one form of the invention the c-KIT mutation is selected from V560G, D816V or recapitulates one or more mutation(s) found in a clinical sample of a GIST. Examples of such c-KIT and PDGFRA mutations include, but are not limited to, for c-KIT: AY501-502 duplication, V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V and del815-816; and for PDGFRA: D842V and V561 D. The FDC-P1 cell may include one or more mutations in either or both of a c-KIT and PDGFRA gene.

In a further aspect the present invention provides an in vivo method for assessing an effect of a test compound on the model tumour of the present invention, including the steps of: providing the tumour in a non-human animal; treating the tumour with a test compound; and detecting a change in a biological function of the tumour. In one form of the invention, the ability of the test compound to change a biological function of the tumour is indicative of the test compound being an antitumour compound

In one form of the invention the test compound is a drug useful for treating GIST or a small molecule kinase inhibitor of VEGFR, PDGFRA, c-KIT and/or FLT-3. In another form of the invention the test compound is selected from the group consisting of: imatinib, SU011248; AMN107; AMG706 and BMS354825. The test compound may also be a combination of one or more compounds.

In one form of the invention the step of detecting a change in a biological function of a tumour is performed by one or more of: measuring the size of the tumour; histological examination of the tumour; imaging the tumour by computed tomography (CT scan); imaging the tumour by positron emission tomography (PET); or any standard method for detecting the presence of a tumour known to the skilled addressee.

In one form of the invention the step of detecting a change in a biological function of the tumour is performed by supplying the non-human animal with a radioactively-labeled tracer which is capable of being detected by PET and detecting the radioactive label with PET. In one form of the invention the tracer is labeled with a positron emitting isotope. In one form of the invention the biological function is glucose utilization. In certain embodiments, the radioactive tracer is fluorine-18 fluoro-2-deoxyglucose (¹⁸FDG or F18-FDG).

In a further aspect, the method for assessing an effect of a test compound on the model tumour, as herein described, is used to identifying an antitumour compound. In another aspect, the present invention provides a compound identified by such methods.

Description of the figures Figure 1 shows the rapid kinetics of PET response to imatinib. A patient was treated with imatinib (400mg/d). Left panel: untreated; right panel: 24h later.

Figure 2 shows the use of ¹⁸FDG-PET to measure response in an in vivo model for GIST. A- Longitudinal and transaxial images of mice bearing flank tumours of FDC-P1 cells harbouring c-Kit mutations as indicated. Imatinib

100mg/kg given p.o., 2X daily beginning at time zero. B- Changes in tumour: background ratio in an independent experiment carried out V560G tumour bearing mice as described in A, except that scans were undertaken 4h and 24h after the first dose of imatinib. C- Quantification of GLUT-1 immuno-staining in

V560G tumours removed at the indicated times post imatinib. Percentages of cells positive for GLUT-1 are presented as mean ±SEM.

Figure 3 shows effects of imatinib on glucose uptake in FDC-P1 cells carrying mutations in c-KIT. A- FDC-P1 V560G and D816V cells were treated with 0.5 μM imatinib and 2DOG uptake measured at 2, 4, 8 and 24 hr. Results are shown as percentage of control untreated cells. B- Experiment as in A, but showing viable cell number as determined by flow cytometry for DNA content.

C- Lineweaver-Burk analysis of 2DOG uptake by FDC-P1-V560G cells treated with or without 0.5 mM imatinib for 4 hr. [3H]-2DOG (1 μCi, 100 μM) uptake was determined for 3 min at 37°C, over a range of 2DOG concentrations (0.1 to

5 mM). D- Low glucose conditions sensitise FDC-P1 V560G cells to imatinib.

Cells were treated with or without imatinib in low glucose conditions (L/-, L/+)

(glucose free-RPMI1640, 2.5% FCS) or normal glucose conditions (N/-, N/+) (glucose free-RPMI1640, 2.5% FCS, 11 mM glucose) for 8 hr subjected to flow cytometry for DNA content.

Figure 4 shows the protein sequence of human c-KIT.

Figure 5 shows the nucleotide sequence of the pRUFneo vector.

Detailed description of the invention The present invention relates to an in vivo model tumour for use in determining the response of a tumour to a drug.

The present invention provides a model tumour which includes a cell introduced into a non-human animal. An advantage of this model is that the cell is capable of producing an imageable tumour. In one form of the invention the model tumour is suitable for studying drug responses by tumours including, but not limited to, GIST and tumours harbouring mutations in c-KIT and/or PDGFRA. In one form of the invention the model tumour includes a cell harbouring one or more c-KIT and/or PDGFRA mutation(s).

A "non-human animal" according to the present invention may be any animal that can be manipulated in order to study the antitumour effects of test compounds. In one form of the invention the non-human animal is a rodent. In another form of the invention the non-human animal is a rat or a mouse. It is certain embodiments of the invention the cell and the non-human animal are syngeneic in order to facilitate tumour growth and obviate the need to use an immunologically deficient animal.

As used herein, the term "imageable tumour" means that the tumour is capable of being visualized. For example, in specific embodiments of the present invention, the tumour may be visualized by, for example, an unaided eye, microscope, computed tomography or PET. Such tumours are most likely to occur as discrete tumours rather than dispersed tumours such as leukaemias.

The term "mutation", when used in reference to a nucleic acid, is intended to mean the replacement of a nucleotide with another (for example: A with C, G or T; C with A, G or T; G with A, C or T; T with A, C or G), it is also intended to encompass deletions and insertions of sequence. The term "mutation", when used in reference to a peptide, polypeptide or protein, is intended to mean replacement of an amino acid with another, it is also intended to encompass deletions and insertions of sequence. Many of the mutations referred to herein are written in the form X###Y, where X represents the original nucleotide or amino acid residue, ### represents the position in the sequence, and Y represents the new nucleotide or amino acid.

The present invention also provides an in vivo model for GIST. In certain embodiments, the in vivo model for GIST includes a mouse which has a FDC-P1 cell introduced optionally subcutaneously into the mouse. In one form of the invention the FDC-P1 cell is present in a cell support matrix such as Matrigel. The mouse may be a DBA/2J mouse. The FDC-P1 cell optionally includes a mutant c-KIT and/or PDGFRA gene. In one form of the invention the c-KIT mutation is selected from V560G, D816V or recapitulates one or more mutation(s) found in a clinical sample of a GIST. Examples of such c-KIT and PDGFRA mutations include, but are not limited to, for c-KIT: AY501-502 duplication, V560G, DeI557-558, V654A, T670I, N822K, Y823D, D816H, D816V and del815-816; and for PDGFRA: D842V and V561 D. The FDC-P1 cell may include one or more mutations in either or both of a c-KIT and PDGFRA gene.

A significant challenge in the introduction of targeted therapies for GIST into clinical practice has been the difficulty associated with defining objective measures of clinical response. A feature of targeted therapies is that standard RECIST response criteria may not apply. Although the reduction in volume may frequently take many months or not occur at all, patients can obtain rapid symptomatic benefit and very significant improvements in survival. This has led to problems in assessing response for the purposes of planning clinical trials and establishing justification for public funding for this expensive medication and most importantly in obtaining a practical measure of benefit to guide ongoing patient care. In Figure 1 , Inventors demonstrate a high avidity of GIST for the glucose analogue, ¹⁸FDG on positron emission tomography (PET) and the striking observation that PET responses to imatinib occur within days of commencing treatment and robustly predict clinical benefit. The FDG-PET changes occur long before measurable changes in dimensional parameters on conventional imaging. A second implication of the extremely rapid metabolic response observed in patients treated with imatinib is that glucose metabolism may be a primary determinant of response rather than a result of the response.

A significant limitation to the progress of developing new anti-GIST therapies has been the lack of useful models for preclinical testing of lead compounds. Until the present invention there have been no good GIST models. The GIST-882 cell line was generated from a primary human GIST, but subsequently proved to be unstable and failed to respond to imatinib. An animal model of GIST has been generated through knock-in of an c-KIT exon 11 V558D allele. After a period of several months, the mice develop patchy hyperplasia and neoplastic lesions histologically indistinguishable from human GIST, but predominantly arising in the distal small bowel. The location and patchy distribution of these lesions renders them difficult to evaluate by either calliper measurements or imaging. Furthermore, due to the long lead-time for tumour development, this model is unsuitable for rapid assessment of novel therapies.

Therefore, there remains a pressing need for a readily manipulated in vivo model for preclinical drug testing that can provide convenient read-outs for therapeutic response. Ideally, a useful model should be adaptable so that novel mutations identified can be tested against a panel of potential inhibitors, and the read-out should be sensitive and rapid. An additional advantage of basic and preclinical studies in an in vivo model is that pharmacologic parameters relevant to drug development, such as absorption, biodistribution, and preliminary estimation of toxicity can be generated in parallel with assessment of therapeutic effect.

In particular embodiments, the present invention provides a model which may be used for developing new anti-GIST therapies.

As discussed above, a large percentage of GISTs contain c-KIT mutations. Gain-of-function mutations in the c-KIT protein have been associated with a range of human cancers, in particular, amino acid substitution within the intracellular juxtamembrane domain (V560G) or the kinase catalytic domain (D816V) have been described in various cancer cell lines and human tumours. c-KIT juxtamembrane mutations such as V560G are commonly observed in GIST while the kinase domain mutation is commonly observed in mastocytosis. Expression of either mutation in factor-dependent murine haemopoietic lines or murine bone marrow cells results in constitutive receptor phosphorylation and kinase activity, growth-factor independence and tumourigenicity in mice.

Surprisingly, the GIST model of the present invention may employ a murine haemopoietic cell line (FDC-P1 ) optionally expressing mutant or wild-type c-KIT

(WT-KIT). Although FDC-P1 is a haemopoietic cell line, Inventors have been able to grow the c-KIT mutant lines as tumour grafts which renders them amenable to rapid assessment of response using, for example, PET. Thereby providing a model system in which to study c-KIT mutations in GIST and their effect on imatinib sensitivity in vivo. Similarly, mutations of PDGFRA may also be analysed using FDC-P1 cells.

Preparing the model tumour in a non-human mammal includes introducing a cell capable of developing into a tumour, into the mammal. In order to facilitate tumour growth and obviate the need to use an immunologically deficient mammal the cell and the mammal may be syngeneic. Immunologically deficient mammals require particular handling conditions to reduce the likelihood of infections. To facilitate handling of the non-human mammal, the non-human mammal should not be immunologically deficient.

In certain embodiments of the present invention the non-human mammal is a rodent. In one form of the invention the non-human mammal is selected from the group consisting of a rat and a mouse. In another form of the invention the non-human mammal is a mouse.

The cell capable of developing into a tumour may be introduced into any part of the non-human mammal. Possible routes of entry of the cell into the non- human mammal, but are not limited to: intraperitoneally, subcutaneously and intravenously. In one form of the invention the cell is introduced subcutaneously.

In another aspect the present invention provides a method of preparing an in vivo model of a tumour, wherein the method includes the step of introducing a cell into a mouse. In certain embodiments the present invention provides a method of preparing an in vivo model of GIST, wherein the method includes the step of introducing a FDC-P1 cell into a mouse, optionally subcutaneously. In one form of the invention the FDC-P1 cell is present in a cell support matrix such as Matrigel. The number of cells introduced may be between approximately 5x10⁵ and approximately 5*10⁷ cells in PBS:Matrigel (at approximately 1 :1). The mouse is may be a DBA/2J mouse. In certain embodiments the FDC-P1 cell includes a mutant c-KIT and/or PDGFRA gene. In one form of the invention the c-KIT mutation is selected from V560G, D816V or recapitulates one or more mutation(s) found in a clinical sample of a GIST. Examples of such c-KIT and PDGFRA mutations include, but are not limited to, for c-KIT: AY501-502 duplication, V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V and del815-816; and for PDGFRA: D842V and V561 D. The FDC-P1 cell may include one or more mutations in either or both of a c-KIT and PDGFRA gene.

In a further aspect the present invention provides a use of an FDC-P1 cell in a model tumour. In one form of the invention the model tumour is a model GIST wherein the FDC-P1 cell is provided to a non-human mammal in order to produce a model GIST. In certain embodiments the non-human mammal is a mouse. In other embodiments the FDC-P1 cell includes a mutant c-KIT and/or PDGFRA gene. In one form of the invention the c-KIT mutation is selected from V560G, D816V or recapitulates one or more mutation(s) found in a clinical sample of a GIST. Examples of such c-KIT and PDGFRA mutations include, but are not limited to, for c-KIT: AY501-502 duplication, V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V and del815-816; and for PDGFRA: D842V and V561 D. The FDC-P1 cell may include one or more mutations in either or both of a c-KIT and PDGFRA gene. A study comparing FDC-P1 cells carrying WT-KIT or mutant c-KIT showed increased susceptibility to imatinib by V560G-KIT expressing cells, and resistance by D816V-KIT expressers, as compared with cells expressing WT-KIT. In addition, imatinib blocked Kit ligand-induced phosphorylation of downstream signalling molecules including ERK, AKT and STAT3. Furthermore, imatinib inhibited constitutive activation of signalling pathways induced by V560G-KIT but had no affect on activation of these pathways by D816V-KIT.

The invention also provides an in vivo method of assessing the antitumour effect of a test compound on a model tumour. The tumour may include a c-KIT and/or PDGFRA mutation. The c-KIT mutation may be V560G, D816V or another mutation identified in a clinical sample of a GIST. Examples of such c-KIT and PDGFRA mutations include, but are not limited to, for c-KIT: AY501-502 duplication, V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V and del815-816; and for PDGFRA: D842V and V561 D.

The test compound screened may be of any type. It may be naturally occurring or a man-made (synthetic) compound. For example, the compound may be a small molecular weight compound, peptide, protein, antibody or derivative thereof. In one form of the invention the test compound is a drug useful for treating GIST or a small molecule kinase inhibitor of VEGFR, PDGFRA, c-KIT and/or FLT-3. In another form of the invention the test compound is selected from the group consisting of: imatinib, SU011248; AMN107; AMG706 and

BMS354825. The test compound may also be a combination of one or more compounds.

By producing a series of c-KIT and PDGFRA mutations, a large number of mutant FDC-P1 -derived tumours can be produced. It is envisaged that such a series of mutants could be used in the model GIST described above to test various test compounds to determine which compound is most likely to be clinically effective against tumours bearing analogous mutation. Surprisingly, the Inventors have found that glucose utilization is rapidly affected by inhibition of receptor tyrosine kinases such as, but not limited to, c-KIT and PDGFRA. Such inhibition may be achieved by agents such as, but not limited to, imatinib. Accordingly, glucose utilization provides an example of a biological function that may be monitored to rapidly detect a response to a test compound by a model tumour.

Based on this finding, PET has been used to show that glucose utilization by the model GIST is a determinant of response to imatinib. Moreover, the rapid

¹⁸FDG PET response of GIST to the inhibition of c-KIT allows rapid assessment of novel therapies directed at GIST. Inventors have therefore developed an in vivo model of c-KIT- and/or PDGFRA-induced neoplasia that is responsive to antitumour agents such as imatinib and can be used to evaluate a biological function, such as glucose utilization, in response to novel targeted therapeutics.

While glucose utilization is exemplified here, it is envisaged that other biological functions may be monitored in order to identify an antitumour effect of a test compound, these are discussed below. In specific embodiments, this model allows the evaluation of many novel mutations in receptor tyrosine kinases such as, but not limited to, c-KIT and PDGFRA and their response to novel targeted therapeutics in vivo.

The present invention provides an in vivo method for assessing an antitumour effect of a test compound on the model tumour of the present invention. The method includes the steps of providing the tumour in a non-human mammal; treating the tumour with a test compound; and detecting a change in a biological function of the tumour; wherein the model tumour optionally harbours a c-KIT and/or PDGFRA mutation.

The invention described herein is of great value in clinical practice. It enables a researcher to rapidly assess the potential response of a tumour to one or a combination of chemotherapeutic agents. It also allows the researcher to predict the outcome of treatment based on quantitative measurements of tumour response after initiation of chemotherapy.

In one form of the invention the step of detecting a change in a biological function of a tumour is performed by one or more of: measuring the size of the tumour; histological examination of the tumour; imaging the tumour by computed tomography (CT scan); imaging the tumour by positron emission tomography (PET); or any standard method for detecting the presence and status of a tumour known to the skilled addressee.

The method of the present invention may be used to identify an antitumour agent from a panel of test compounds. It is anticipated that the antitumour agent will have an antitumour effect on the tumour. This effect may be assessed and/or detected as a change in a biological function of the tumour. A change in a biological function of the tumour may include, but are not limited to, gross changes such as tumour growth; histological changes; and metabolic changes. Biological functions may be monitored by many different methods to detect changes, these methods include, but are not limited to: measuring the size of the tumour to monitor tumour growth; histological analysis of the tumour to detect changes in tumour cell appearance; CT scan of the tumour to monitor growth of the tumour; and imaging the tumour by PET to monitor a biological function of the tumour.

The step of treating the tumour with the test compound may be performed by directly applying the test compound to the tumour. However, in order to mimic possible treatment regimens of humans, the test compound may be introduced into the tumour-containing non-human mammal systemically. Possible routes for systemic introduction of the test compound include, but are not limited to: oral and intravenous.

Measuring the size of the tumour may be performed using caliper measurements from outside the non-human mammal or by surgically removing the tumour and then measuring the tumour. Histological examination of the tumour generally requires removal of part or all of the tumour before microscopic examination by standard techniques.

CT, also known as a CAT scan, can be used to monitor the tumour in vivo. This method uses x-ray equipment to obtain image data from different angles around the non-human mammal, and then uses computer processing of the information to show a cross-section of tissues, organs and the tumour.

In certain embodiments, the biological function is monitored by imaging the tumour by PET. The biological function may include any function of a living cell. In one form of the invention the biological function monitored allows differentiation between the tumour and non-tumour in the non-human mammal. PET allows changes in a metabolic activity to be monitored.

As used herein, a "metabolic activity", refers to any cellular function which can be monitored within a cell. These functions include, but are not limited to, metabolism (use and production of energy); protein and/or DNA synthesis/breakdown; nutrient uptake; substrate or substrate analogue uptake; and metabolite production. A "substrate analogue" is intended to encompass compounds that cannot be metabolized by a cell. Since the methods of the present invention seek to identify a test compound with antitumour effect, an antitumour effect will be indicated if the metabolic activity to be monitored changes in response to exposure to the test compound. Any metabolic activity may be monitored as long as there is an isotope-labeled tracer available for its detection by imaging technologies such as PET, as well understood by the person skilled in the art. In one form of the invention the tracer is labeled with a positron emitting isotope. In certain embodiments, the metabolic activity is glucose utilization which may be monitored using a radioactive isotope of 2-deoxyglucose.

In one form of the invention the step of detecting a change in a biological function of the tumour is performed by PET imaging the tumour to monitor metabolic activity. PET is an imaging technique that relies on changes in cellular biochemistry. Most conventional medical imaging techniques, such as X-ray, ultrasound, CT and magnetic resonance imaging (MRI), depend on changes in the anatomy or structure of organs. PET is able to very rapidly image changes in metabolic activity, long before there is a change in anatomy. Biologically active natural compounds such as oxygen, carbon and glucose labeled with radioactive positron emitters are given intravenously and react in the body identically to their non-radioactive counterparts. The normal and abnormal distribution can be imaged with a PET scanner. The small amounts used do not have a pharmacological effect and there are no known biological hazards associated with PET imaging.

In one form of the invention the step of imaging the tumour by PET is performed by supplying the tumour with a radioactively-labeled tracer which is capable of being detected by PET and detecting the radioactive label with PET. In one form of the invention the tracer is labeled with a positron emitting isotope. Exemplary positron emitting isotopes useful in PET include, but are not limited to, ¹⁸F, ¹⁵0, ¹³N and ¹¹C.

In one form of the invention the metabolic activity is glucose utilization. In certain embodiments, the tracer is a radioactively-labeled form of 2-deoxyglucose, for example, ¹⁸F-labeled deoxyglucose (¹⁸FDG). When ¹⁸FDG is used, the resultant PET images are often referred to as FDG-PET scans.

However, there are a wide variety of possible radioactive tracers for use in PET, depending on the metabolic activity being analysed. For applications involving the imaging of cancers, an exemplary tracer is fluoro-2-deoxyglucose to measure glucose uptake and/or metabolism since these processes respond quickly to antitumour treatments. Other studies may include monitoring blood flow into a tumour by following H₂ ¹⁵O. Accordingly, the present invention is not intended to be limited to the use of fluoro-2-deoxyglucose. Any positron emitting variant of a compound can be used, as long as the metabolic activity to be monitored will change if the test compound has an antitumour effect. Some of the more common tracers used in clinical PET imaging, and the examples of their application, are listed in Table 1 below.

Table 1: Tracers commonly used in clinical PET analyses

Tracer Common use

¹⁸FDG glucose uptake and metabolism.

H₂ ¹⁵O blood perfusion

¹⁸F-thymidine DNA synthesis

C¹⁵O₂ blood pool imaging

¹¹C-tyrosine amino acid uptake and protein synthesis

¹³NH₃ myocardial perfusion

¹¹C-raclopride neurotransmitter receptor targeting

¹⁸F-DOPA neurotransmitter uptake

¹¹C-acetate myocardial perfusion and metabolism

The validity of measuring changes in metabolic rate to assess the efficacy of chemotherapy agents is supported by the finding that serial measurements of untreated malignant tumours in humans conducted over a 10-day period before the initiation of chemotherapy remain within a narrow range and demonstrate high reproducibility between measurements. Subsequent changes in glucose metabolic rate as a result of administration of chemotherapeutic agents may thus be accurately quantified.

However, although PET imaging has demonstrated very high accuracy in the measurement of glucose metabolism in normal and malignant tissue, it suffers from a disadvantage that diminishes its utility for routine monitoring of chemotherapy as well as its use with model systems. The spatial resolution of PET is relatively low, and there is a consequent decrease in accuracy in imaging of tumours and metastases that are smaller than approximately 1 cm.

The development of micro-PET or small animal PET has facilitated the use of PET in animal models of disease. Accordingly, the present invention provides a use of small animal PET to monitor a metabolic activity of a model tumour in response to treatment with a test compound. It is envisaged that the test compound will be introduced into the non-human mammal as a pharmaceutical composition. The test compound may be administered by any suitable route, including, but not limited to: orally, rectally, percutaneously, or parenterally.

The present invention also provides a use of PET to assess an antitumour effect of a test compound on a tumour, including the steps of providing a tumour in a non-human mammal; treating the tumour with a test compound; and imaging the tumour by PET to monitor a metabolic activity of the tumour.

In one form of the invention the tumour is a model tumour. A model tumour may respond to treatment in a similar manner to the tumour upon which the model tumour is based. A model tumour suitable for PET analysis in an animal may, but not necessarily must, develop as a discrete mass, since imaging metabolic changes in dispersed tumours is difficult.

Generally, it is envisaged that, according to specific embodiments of the present invention, a test compound can be tested for antitumour effect against GIST in the following manner. Two groups of non-human mammals, in this case mice, each harbouring a model GIST, wherein the model GIST may include a FDC-P1 cell optionally harbouring one or more mutations in c-KIT and/or PDGFRA, are either treated with the test compound, or left as a control group. Following treatment, all mice are given a positron-emitting isotope of a tracer, for example ¹⁸FDG if glucose uptake and/or metabolism is being monitored. The mice are then subjected to PET analysis to determine whether the tracer, in this case ¹⁸FDG, has been taken up by the model GIST. In the control group it is expected that ¹⁸FDG will be taken up by the model GIST and the tumours will show as dark regions on the PET analysis. The treated mice may show reduced darkening of the tumours compared to the control mice. A reduction in darkening on PET analysis indicates that ¹⁸FDG is not being taken up as readily following treatment with the test compound as compared to the untreated mice. Therefore, in this example, the reduced darkening indicates a change in the metabolic activity of the model tumour and that the test compound is having an antitumour effect on the model GIST.

By repeating the above methodology for a large panel of tumours and test compounds it is possible to generate a database in which mutations in c-KIT and/or PDGFRA can be correlated to test compounds which give the best antitumour effect. Such a database could be invaluable in the clinical setting. For example, a patient with a tumour (which was being treated with imatinib) is no longer responding to imatinib treatment. A sample of the non-responsive tumour can be taken and its c-KIT and/or PDGFRA genes sequenced by standard methods. During the sequencing a mutation is found in one or both genes. The mutation predicts an altered polypeptide product of the respective gene. By accessing the database, a "best course of treatment" can be determined by comparing the patient's mutation(s) with those in the database and correlating that with test compounds which gave the best antitumour effect on tumours bearing that mutation(s).

Specific embodiments and applications of the present invention will now be discussed in detail by reference to the accompanying examples. This discussion is in no way intended to limit the scope of the invention.

Examples

Example 1 - DEVELOPMENT AND ANALYSIS OF A MODEL GIST FDG-PET has an important role in the early assessment of response of GIST to imatinib that predicts subsequent response on standard CT criteria. Furthermore, the Inventors have observed dramatic PET responses within 24 h of commencing imatinib treatment (Figure 1). To investigate the mechanism underlying this dramatic metabolic response to imatinib, the Inventors subsequently sought to develop and characterize an in vivo model in which this effect was recapitulated.

As discussed above, FDC-P1 murine haemopoietic cell lines expressing either V560G-KIT or D816V-KIT had previously been compared to WT-KIT with respect to imatinib sensitivity and downstream signalling through c-KIT. These studies demonstrated increased susceptibility to imatinib of V560G-KIT expressing cells, and resistance of D816V-KIT expressers, compared with cells expressing WT-KIT. In addition, imatinib blocked Kit ligand-induced phosphorylation of downstream signalling molecules including ERK, AKT and STAT3. Furthermore, imatinib inhibited constitutive activation of signalling pathways induced by V560G-Kit but had no affect on activation of these pathways by D816V-Kit.

It was surprisingly found that, although FDC-P1 is a haemopoietic cell line, Inventors have been able to grow the FDC-P1 c-KIT mutant lines as tumour grafts which renders them amenable to rapid assessment of response using PET.

Female DBA/2J mice were inoculated subcutaneously on each flank with 5MO⁶ cells in PBS:Matrigel (Becton-Dickinson) (1 :1). Once the tumors had reached a volume of approximately 200 mm³ the mice were randomised into two groups of 7-8 animals (day 0). Mice were fasted for 3 hr then anaesthetised in a container into which 2.5% isoflurane in 1 :1 O₂ and air was delivered (flow rate 200 ml/min). Anaesthetised mice were injected via the tail vein with 300 mCi ¹⁸FDG and anaesthesia maintained for a further 20 min before the animals were removed into husbandry cages where they were allowed recover. 1.5 h after tracer injection the mice were again anaesthetised and scanned for 5 min on a Phillips A-PET prototype small animal PET scanner.

For PET scanning and analysis, attenuation correction, either measured or estimated was not performed. Scans were acquired in a 3D volume mode and rebinned into two dimensions using published algorithms (Tanaka and Kudo, 2003, Phys Med Biol. 48:1405-22; Chiang et ai, 2004, Nucl Med Commun. 25:1103-7). Image analysis was performed by determining the maximum and average tracer uptake in a defined region representing the tumor and background respectively. A tumoπbackground ratio was then determined by dividing the maximum count within a tumor by the average count within the background region. Mice were scanned on day 0, day 1 (V560G only) and day 2.

Both V560G and D816V tumours were avid for ¹⁸FDG at baseline, consistent with the high FDG uptake observed clinically in GIST (Fig. 2A), which was unaffected by imatinib treatment in D816V-KIT expressing tumours. However, V560G-KIT expressing tumours showed a substantial reduction in FDG signal at 24 h and by 48 h, FDG uptake was almost completely abolished in these cells. In a follow-up experiment, cohorts of V560G bearing mice were imaged at 4 or 24 h post imatinib treatment and sacrificed immediately after imaging for ex vivo analysis of biomarkers for effects on metabolism (Glut-1 expression), proliferation (BrdU incorporation) and apoptosis (activation of caspase 3). Quantification of FDG uptake (PET to background ratios- Fig. 2B) revealed that even at the 4 hour time point when no change in tumour volume was apparent, FDG uptake was reduced by approximately 30% compared with untreated control tumours (P=O.002). As in the earlier experiment, at the 24 h time point, FDG uptake was reduced by about 60% compared with controls (P<0.001).

Tumours from imaging experiment in Fig. 2B were examined by immunohistochemistry for effects of imatinib on expression of the glucose transporter, GLUT-1 (Fig. 2C). This analysis revealed that imatinib significantly and rapidly down-regulated total and membrane-associated GLUT-1 staining in V560G tumours at both 4 h and 24 h (Fig. 2C P<0.01). No significant change in proliferation (BrdU incorporation) or apoptosis (activated-caspase 3) was evident at 4 h in either c-KIT mutant. However, in V560G tumours at 24 h, BrdU incorporation was markedly reduced and caspase 3 activation increased. No effect of imatinib on GLUT-1 , activated-caspase 3 or BrdU incorporation was observed in the D816V tumours.

Example 2 - IN VITRO BIOCHEMICAL STUDIES OF MODEL GIST

To further examine the basis of the rapid reduction in FDG uptake in the V560G tumours, the Inventors investigated the effect of imatinib on glucose uptake in vitro. As seen in vivo, exposure of V560G cells in vitro to imatinib resulted in the rapid inhibition of 2-deoxy-glucose (2D0G) uptake with a 50% reduction observed at 2 h (Fig. 3A). This reduction was not due to drug induced cell death as no change in cell viability was observed until 8 h of drug exposure (Fig. 3B). No effect of imatinib on glucose uptake or cell viability was observed in the D816V expressing cells.

To establish the dose dependence of glucose uptake in the presence of imatinib, a kinetic analysis of 2-DOG was performed in V560G cells. Transporter number was reduced in the presence of imatinib (Vm_aχ: control 36.4 nmol/10⁶ cells/min; imatinib treated 28.3 nmol/10⁶ cells/min) while affinity of remaining transporters for 2DOG was not reduced (K_m: control 1.72 mM; imatinib treated 1.22 mM: Fig. 3C). These findings suggest that reduction in glucose uptake in the imatinib treated cells is due to a reduction in the number of glucose transporters at the cell membrane rather than a drop in the rate of glucose transport. Inventors then investigated directly the effect of glucose levels on imatinib induced apoptosis in the V560G cells. Treatment of cells with imatinib in medium containing normal levels of glucose (11 mM) resulted in a sub-GI population of 23% at 8 h (Fig. 3D). In contrast, the cells cultured in low glucose (0.25mM) medium were more susceptible to the apoptotic effects of imatinib (56% sub-G1) suggesting that ambient glucose availability modulates imatinib-induced apoptosis in V560G tumours.

In summary, the kinetics of altered glucose uptake and loss of cell viability suggest that changes in metabolic processes, involving decreased levels of glucose transporters precede and may contribute to the cytoreductive effect of imatinib both in vitro and in vivo.

Example 3 - GLUCOSE METABOLISM IN THE PRESENCE OF IMATINIB IN MODEL GIST To determine if reduced glucose transport and or metabolism play a primary role in the response to imatinib, FDC-P1- C-KIT-V560G cell lines stably over-expressing GLUT-1 , hexokinase or GLUT-1 and hexokinase are generated. Studies have demonstrated that GLUT-1 is the major glucose transporter expressed in FDC-P1 cells and that it is rapidly down regulated following treatment with imatinib (Fig. 2). Inventors use retrovirus-mediated gene transfer to express pBabepuro, pBabepuro-GLUT1 , MSCV and/or MSCV-hexokinase. Cells infected with pBabepuro or pBabepuro-GLUTI are selected in 10μg/ml puromycin. Cells infected with MSCV and/or MSCV-hexokinase are then sorted by flow cytometry.

The cell lines are characterized for expression and localization of GLUT-1 by immuno-fluorescence, glucose transport by 2DOG uptake assays (see preliminary data) and for hexokinase activity using a spectrophotometric assay in which glucose-6-phosphate formation is coupled to NADPH production. The response to O.δμM imatinib is determined using these assays at 2, 4, 8, 16 and 24 h. The effects on cell cycle progression using flow cytometry following BrdU labelling and effects on apoptosis by flow cytometry following staining with Annexin-V and propidium iodide are also able to be examined. The effect of expression of GLUT-1 /hexokinase on expression of apoptotic markers including BCL-2, BCL-X and BAX is also determined. It has been postulated that there is a role involving altered conformation of BAX in response to inhibition of signalling through AKT.

Once the biochemical changes in response to imatinib are established in vitro, it is important to determine the effects on sensitivity to imatinib on FDG uptake in tumours in vivo. For these experiments, each cell line is grown as a subcutaneous tumour graft on the flanks of DBA2/J mice. Each mouse is implanted with lines expressing GLUT-1 , hexokinase or GLUT-1 /hexokinase on one flank and a control empty vector line on the other. When tumours reach approximately 150 mm³, a baseline FDG-PET scan is performed and treatment with imatinib or vehicle begun. FDG-PET scans will then be repeated at 4, 24 and 48 h post-treatment and volume change evaluated by calliper measurements. In addition, the same cell lines are implanted into a parallel cohort of mice for examination of biochemical, cell cycle and apoptotic parameters in response to imatinib. Three mice are sacrificed at each scanning time point and tumours excised and fixed for IHC to confirm GLUT-1 expression and localization. In addition, the tumours are examined for BrdU incorporation, expression of apoptotic markers including BAX conformation and for activated caspase-3. Over-expression of GLUTI+hexokinase allows C-KIT-V560G-FDC-P1 tumours to continue to accumulate FDG and diminish the rapid cell cycle and/or apoptotic response to imatinib of parental V560G tumours.

Example 4 - AKT IN THE PRESENCE OF IMATINIB IN MODEL GIST

It has been previously shown that AKT-directed glucose metabolism promotes growth factor independent survival in an IL-3 dependent cell line possibly by preventing conformational change in the pro-apoptotic protein, BAX. It was shown that myristoylated-constitutively activated-AKT (Myr-AKT) inhibited cell death upon withdrawal of IL-3. This was associated with translocation of the glucose transporter, GLUT-1 , to the cell membrane and increased glucose metabolism. It has also been found that the response of the FDCP1-C-KIT-V560G cell line to imatinib is manifested by down-regulation of AKT activity. Inventors observed a rapid (within 4 h) reduction in FDG uptake when mice bearing grafts of this cell line were treated with imatinib. This was accompanied by reduction of GLUT-1 expression and membrane localization. Furthermore, glucose transport was down-regulated in these cells within 2 h of imatinib treatment in vitro. Since AKT is a downstream target of c-KIT signalling, it is plausible that signalling events downstream of AKT are a primary mechanism of regulation FDG uptake and glucose metabolism in GIST and the FDC-P1 model GIST system.

It has been shown that several signalling pathways are activated downstream of c-KIT including STAT-3, MAPK-ERK and PI3K-AKT pathways. Of these, the PI3K-AKT pathway is a well-recognised regulator of glucose metabolism in response to insulin stimulation. Furthermore, as indicated above, in the presence of glucose, activated AKT can maintain growth factor independent survival of IL3 dependent cells. To determine whether the GIST model FDC-P1 c-KIT mutant cell lines have a similar requirement, constitutively-active or dominant-negative forms of Akt are inducibly expressed in the c-KIT mutant lines using the Tet-ON system, as well understood by the person skilled in the art. The AKT mutants used are myristylated-AKT, a constitutively membrane associated and activated isoform and K179A, a mutant form of the protein with amino acid substitutions in the kinase domain that renders the protein catalytically inactive. Vectors harbouring these forms of AKT are transfected into FDC-P1-G560V and -D816V cells, and lines with regulated expression selected. The cells are characterized in vitro for the effects of expression of Myr-AKT or DN-AKT on glucose uptake and cell survival.

The cells are implanted and grown as tumour grafts in DBA2/J mice and when volumes reach 150mm³, expression of the AKT isoforms is induced by ingestion of doxycycline. 24 h later, mice bearing both Myr-Akt-V560G and empty vector control-V560G tumours are randomised into imatinib and vehicle control treatment groups. Following a baseline FDG-PET scan, treatment is initiated and mice are imaged at 4, 24 and 48 h. Parallel groups of 12 mice are treated as for the PET study and 3 mice sacrificed at each time point for IHC analysis of GLUT-1 , BrdU labelling, and markers of apoptosis and for hexokinase activity. Downstream consequences of AKT activation including phosphorylation of S6K, mTOR, p27, FKHRL1 and BAD, and the effect of imatinib on phosphorylation of these substrates are also examined. Myr-AKT expression renders the V560G-C-KIT tumours refractive to imatinib treatment and allow the tumours to continue to accumulate FDG and activate downstream signalling pathways.

Example 5 - INTRODUCTION OF MUTATIONS INTO C-KIT AND PDGFRA IN MODEL GIST

FDC-P1 cell lines expressing c-KIT mutations in exons 9, 11 , 13 or 17 and PDGFRA exon 12 or 18 are prepared and analysed as described herein (mutations listed in Table 2, the protein sequence of human c-KIT is shown in Figure 4 as SEQ ID NO:1 ). These sites are chosen because they are common in untreated GIST (exon 11 and exon 9) or are situated in sites known to be associated with imatinib resistance (exons 13 and 17 c-KIT and exon 18 PDGFRA). Table 2: Summary of c-KIT and PDGFRA mutants

Mutations lmatinib Sensitivity c-KIT Exon 9

AY501-502 duplication Sensitive Extracellular domain c-KIT Exon 11 V560G oensiuve

Juxtamembrane domain Del557-558 c-KIT Exon 13 V654A Acquired resistance

Kinase domain I T670I to imatinib

N822K c-KIT Exon 17 Y823D Acquired and intrinsic

Activation loop D816H resistance to imatinib

D816V

PDGFRA Exon 18

Intrinsic resistance to Activation loop D842V imatinib

PDGFRA Exon 12

V561D Sensitive Juxtamembrane domain

The mutations listed in Table 2 are exemplary and not intended to be limiting; it is envisaged that many other mutations could be introduced into c-KIT and/or PDGFRA as required. Retroviral transfer using pRUFneo and selection methods used previously to generate the V560G (exon 11) and D816V (exon 17) c-KIT expressing lines can be employed. Expression vectors carrying each of the mutations listed in table 1 are used for sub-cloning these mutations into pRUFneo for generation of FDC-P1 cell lines. While pRUFneo is used in this example, it is not the only possible vector for using in the methods described. The skilled addressee will be aware of many suitable vectors. The nucleotide sequence of pRUFneo is shown as SEQ ID NO:2 in Figure 5. FDC-P1 cells are readily available to the skilled addressee from the American Type Culture Collection under accession number CRL-12103. There are many methods available for introducing specific mutations into DNA, the skilled addressee would be aware of these methods and be able to apply them in order to produce any mutation into c-KIT and/or PDGFRA.

Example 6 - DEFINING BASELINE FDG-PET UPTAKE AS A FUNCTION OF MUTATION STATUS Once phenotypic characterization is complete, each mutant line is implanted into DBA2/J mice and grown as grafts for studies in vivo. FDG-PET studies are performed and the response to imatinib evaluated as described above. FDG-uptake is correlated with resistance to imatinib as defined clinically or using in vitro studies. However, it remains possible that different mutations may vary in their ability to induce FDG-uptake, particularly in response to imatinib. Consistent with this possibility Inventors have observed two clinical cases with acquired resistance to imatinib that displayed mutations in exon 17 of c-KIT (D816H and del815-816) that had clearly documented progression that was not associated with FDG-uptake in the progressing lesions whilst the patients remained on imatinib.

It is possible to evaluate the impact of mutations in c-KIT and PDGFRA on other biological processes that can be imaged by PET and assayed ex vivo, including cellular proliferation and amino acid metabolism. In addition to changes in glucose uptake and metabolism, responses to imatinib are characterised by inhibition of cell cycle progression and cell growth. PET tracers may also be used to image these processes. Inventors have established tracers and protocols to image amino acid transport (fluoro-ethyl tyrosine, FET) and uptake of a thymidine analogue (fluorine-L-thymidine, FLT). These function as surrogate markers for cell growth and proliferation, respectively. Inventors have also shown that uptake of these tracers can be evaluated using the small animal-PET scanner and that uptake can be modulated by systemic cancer therapies. Accordingly, other PET tracers provide useful biological information regarding the response of the transformed FDC-P1 cell to imatinib.

Example 7 - TESTING COMPOUNDS FOR USE AGAINST GIST While the use of imatinib represents a major therapeutic breakthrough in the treatment of GIST, its use is not without significant problems. As with many single agent therapies, the emergence of drug resistance is now a major clinical issue. Furthermore, some patients develop intolerance to the drug and therefore must discontinue the only effective therapy for GIST. One approach to overcoming such problems is the use of other small molecule inhibitors of c-KIT. Indeed SU011248, a small molecule kinase inhibitor of VEGFR, PDGFR, c-KIT and FLT-3 may find use in the treatment of imatinib-refractory GIST.

Other c-KIT inhibitors in various stages of clinical development may be tested against this model GIST. These agents include, but are not limited to: 1) SU011248 (Pfizer); 2) AMN107 (Novartis); 3) AMG706 (Amgen) and 4) BMS354825 (Bristol Myers Squibb). To evaluate these agents the maximum tolerated dose (MTD) of each compound may be determined in the DBA2/J mice using standard methods. Briefly, four drug doses are administered to groups of 2 mice by the appropriate route and schedule and weight loss assessed daily. The MTD, which is defined as that dose which causes ~10% weight loss, this is then confirmed in a further cohort of 5 mice.

Example 8 - USE OF THE MODEL GIST FOR PREDICTING CLINICAL OUTCOME

A possible application of the model GIST described herein is in the preparation of a panel of c-KIT and/or PDGFRA mutants against which the effectiveness of a cohort of test compounds could be determined. This would provide a database which could be used to match a imatinib-refractory GIST with a treatment likely to treat the GIST, once its c-KIT and/or PDGFRA were sequenced.

For example, a patient being treated with imatinib for GIST presents to a clinic with imatinib-refractory GIST. A biopsy of the imatinib-refractory GIST is taken and the C-KIT and PDGFRA genes of the GIST are sequenced. The resultant sequence indicates a mutation at a frequently mutated site in c-KIT. From the database produced as above, it is known that a particular drug is effective in treating imatinib-refractory GIST having that c-KIT mutation. The database has prevented a significant amount of time being wasted trying to identify a suitable drug with which to treat the patient. This may lead to a more rapid and thorough recovery for the patient.

Claims

1. An in vivo model tumour for use in determining the response of a tumour to a drug, wherein the model tumour is derived from a cell which is capable of producing an imageable tumour when introduced into a non- human mammal.

2. The model tumour of claim 1 , wherein the model tumour is suitable for studying drug responses by tumours selected from the group consisting of: GIST and tumours harbouring mutations in c-KIT and/or PDGFRA.

3. The model tumour of claim 1 , wherein the model tumour includes a cell harbouring one or more c-KIT and/or PDGFRA mutation(s).

4. The model tumour of claim 1 , wherein the cell harbours a c-KIT mutation.

5. The model tumour of claim 4, wherein the cell harbours a V560G and/or D816V c-KIT mutation.

6. The model tumour of claim 1 , wherein the cell harbours a mutation which recapitulates one or more mutation(s) found in a clinical sample of a GIST.

7. The model tumour of claim 1 , wherein the model tumour is a model GIST.

8. The model tumour of claim 7, wherein the model GIST includes a FDC-P1 cell which is introduced into the mouse.

9. The model tumour of claim 8, wherein the FDC-P1 cell is introduced subcutaneously.

10. The model tumour of claim 8, wherein the mouse is a DBA/2J mouse.

11. The model tumour of claim 9, wherein the FDC-P1 cell includes a mutant c-KIT and/or PDGFRA gene.

12. The model tumour of claim 11 , wherein the mutant c-KIT gene has one or more mutations selected from the group consisting of: V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V, del815-816 and a mutation which recapitulates one or more mutation(s) found in a clinical sample of a GIST.

13. The model tumour of claim 11 , wherein the mutant PDGFRA gene has one or more mutations selected from the group consisting of: D842V and V561 D.

14. A method of preparing an in vivo model tumour, wherein the method includes the step of introducing into a non-human animal a cell that produces an imageable tumour when introduced into the non-human animal.

15. The method of claim 14, wherein the model tumour is a model of GIST.

16. The method of claims 14 or 15, wherein the non-human animal is a mouse.

17. The method of claim 16, wherein the method includes the step of introducing a FDC P1 cell into a mouse.

18. The method of claim 17, wherein the FDC P1 cell is introduced subcutaneously into the mouse.

19. The method of claim 18, wherein the mouse is a DBA/2J mouse.

20. The method of claim 18, wherein the FDC P1 cell harbours a mutant c- KIT and/or PDGFRA gene.

21. The method of claim 20, wherein the mutant c-KIT gene has one or more mutations selected from the group consisting of: V560G, Del557 558, V654A, T670I, N822K, Y823D, D816H, D816V, del815 816 and a mutation which recapitulates one or more mutation(s) found in a clinical sample of a GIST.

22. The method of claim 20, wherein the mutant PDGFRA gene has one or more mutations selected from the group consisting of: D842V and V561 D.

23. A non-human animal including a tumour produced according to the method of any one of claims 14 to 22.

24. A method of preparing an in vivo model of a tumour, wherein the method includes the step of introducing a cell into a mouse.

25. A method of claim 24, wherein the model tumour is a model of GIST, and wherein the method includes the step of introducing a FDC-P1 cell into a mouse.

26. The method of claim 25, wherein the FDC-P1 cell is introduced subcutaneously into the mouse.

27. The method of claim 25, wherein the mouse is a DBA/2J mouse.

28. The method of claim 25, wherein the FDC-P1 cell harbours a mutant c-KIT and/or PDGFRA gene.

29. The method of claim 28, wherein the mutant c-KIT gene has one or more mutations selected from the group consisting of: V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V, del815-816 and a mutation which recapitulates one or more mutation(s) found in a clinical sample of a GIST.

30. The method of claim 28, wherein the mutant PDGFRA gene has one or more mutations selected from the group consisting of: D842V and V561D.

31. Use of an FDC-P1 cell in a model tumour, wherein the model tumour is imageable.

32. The use of claim 31 , wherein the model tumour is a model GIST.

33. The use of claim 31 , wherein the FDC-P1 cell harbours a mutant c-KIT and/or PDGFRA gene.

34. The use of claim 33, wherein mutant c-KIT gene has one or more mutations selected from the group consisting of: V560G, Del557-558, V654A, T670I, N822K, Y823D, D816H, D816V, del815-816 and a mutation which recapitulates one or more mutation(s) found in a clinical sample of a GIST.

35. The use of claim 33, wherein the mutant PDGFRA gene has one or more mutations selected from the group consisting of: D842V and V561 D.

36. An in vivo method for assessing an antitumour effect of a test compound on a model tumour, the method including the steps of: providing the tumour in a non-human mammal; treating the tumour with a test compound; and detecting a change in a biological function of the tumour.

37. The method of claim 36, wherein the ability of the test compound to change a biological function of the tumour is indicative of the test compound being an antitumour compound.

38. The method of claim 36, wherein the test compound is selected from the group consisting of: a drug useful for treating GIST; and a small molecule kinase inhibitor of VEGFR, PDGFRA, c-KIT and/or FLT-3.

39. The method of claim 36, wherein the test compound is selected from the group consisting of: imatinib, SU011248; AMN107; AMG706 and BMS354825.

40. The method of claim 36, wherein the step of detecting a change in a biological function of a tumour is performed by one or more of: measuring the size of the tumour; histological examination of the tumour; imaging the tumour by computed tomography (CT scan); imaging the tumour by positron emission tomography (PET).

41. The method of claim 36, wherein the step of detecting a change in a biological function of the tumour is performed by supplying the non-human mammal with a radioactively-labeled tracer which is capable of being detected by PET and detecting the radioactive label with PET.

42. The method of claim 41 , wherein the biological function is glucose utilization.

43. The method of claim 41 , wherein the radioactive tracer is fluorine-18 fluoro-2-deoxyglucose.

44. The method of any one of claims 36 to 43, when used to identify an antitumour compound.

45. A compound identified by a method according to claims 44.