[0001] This application claims priority to U.S. provisional patent application serial number 61/910,652, filed December 2, 2013, the entire disclosure of which is incorporated herein by reference.

[0002] The present invention relates to methods for the selection of cancer treatment regimens. The invention also relates to kits for the selection of a suitable cancer treatment regimen in a subject with cancer. In particular, the invention relates to methods and kits for the selection of cancer treatment regimens in which a VEGF inhibitor is used for the treatment of cancer, optionally in combination with a chemotherapeutic agent or other combination partner.

BACKGROUND OF THE INVENTION

[0003] Vascular endothelial growth factor (VEGF) is an angiogenic factor known to play a key role in the production of new blood vessels in a range of pathological conditions. In particular, the action of VEGF is known to contribute to blood vessel formation associated with the development and progression of cancer. VEGF signalling occurs through a number of receptor tyrosine kinases.

[0004] VEGF signalling inhibitors are considered to represent promising therapeutic agents for use in the treatment of cancer, particularly in treatment regimens in which these inhibitors are used in combination with chemotherapeutic agents. Unfortunately, the clinical application of VEGF inhibitors to date has been limited due to the fact that not all potential patients exhibit benefits in response to their use. An inability to determine which subjects suffering from cancer will respond favourably to treatment regimens utilising VEGF inhibitors and chemotherapeutic agents has hindered the uptake of these potentially promising treatment regimens.

[0005] In view of the above, it will be recognised that there is an unmet need for biomarkers, methods and kits suitable for the selection of patients likely to respond to VEGF signalling inhibitor based therapy. An improved understanding of the mechanisms underlying tumour sensitivity and resistance to VEGF signalling inhibitors is very important to the aim of identifying potential predictive biomarkers and developing personalized health care therapies. Most previous studies investigating the mechanisms of sensitivity/resistance to VEGF signalling inhibitors have focused on the potential role of factors that may represent alternative pro-angiogenic pathways such as the composition of the tumour micro environment and its paracrine role to support resistance to VEGFi.

SUMMARY OF THE INVENTION

[0006] In a first aspect, the invention provides a method for selecting a suitable cancer treatment regimen in a subject with cancer, the method comprising: • determining the circulating level in the subject of at least one protein selected from the group consisting of: leptin, creatine kinase muscle and brain (CK-MB); interleukin-10 (IL-10); plasminogen activator urokinase (PLAUR); interleukin-6 receptor beta (IL6Rb); TNF receptor superfamily member 6 (TNFSF 6); and TNF related apoptosis inducing ligand receptor 3 (TRAILR3);

• comparing the determined circulating protein value with a reference value; and

• selecting a cancer treatment regimen for the subject comprising provision of a VEGF signalling inhibitor if the determined circulating protein level in respect of leptin or CK- MB is greater than the reference value; and/or the determined circulating protein level in respect of at least one of IL-10; PLAUR; IL6Rb; TNFSF 6 or TRAILR3 is lower than the reference value.

[0007] The protein leptin provides a particularly useful protein for use in such methods, either alone or in combination with one or more of the other listed proteins.

[0008] In a second aspect, the invention provides a method for selecting a suitable cancer treatment regimen in a subject with cancer, the method comprising:

• determining an expression profile in a tumour of the subject for at least 2 genes of the GAHGL gene signature set out in Table 1 ;

• comparing the expression profile determined with a reference expression profile; and

• selecting a cancer treatment regimen for the subject comprising provision of a VEGF signalling inhibitor if the determined expression profile is increased as compared to the reference profile.

[0009] The GAHGL gene signature consists of those genes set out in Table 1 , and is discussed further below. It is so-called because it comprises genes associated with Glucose metabolism, Angiogenesis, Hypoxia, Glutamine metabolism, and Leptin signalling.

[00010] In a third aspect, the invention provides a method for selecting a suitable cancer treatment regimen in a subject with cancer, the method comprising:

• determining an expression profile in a tumour of the subject for at least 2 genes of the LOPS gene signature set out in Table 1 ;

• comparing the expression profile determined with a reference expression profile; and

• selecting a cancer treatment regimen for the subject comprising provision of a VEGF signalling inhibitor if the determined expression profile is decreased as compared to the reference profile.

[00011] The LOPS gene signature consists of those genes set out in Table 1 , and is discussed further below. It is so-called because it comprises genes associated with Lipid metabolism, Oxidative activity, Proliferation, and Serine synthesis.

[00012] Generally, considerations in the present disclosure described as relating to methods of the invention should be considered to be applicable to methods in accordance with any of the first, second, third, eighth, ninth, tenth, or eleventh aspects of the invention, except for where the context requires otherwise.

[00013] In a suitable embodiment, a method of the invention optionally further comprises treating said subject with the selected cancer treatment regimen comprising provision of a VEGF signalling inhibitor. Such methods may also involve the provision of a further anti-cancer agent, such as a chemotherapeutic agent. Such embodiments provide methods for selecting and implementing a suitable cancer treatment regime.

[00014] It will be appreciated that in methods in accordance with the first aspect of the invention, the optional step of treating the subject in such embodiments is only performed in instances in which the circulating protein level is established to differ from the reference value in the required manner.

[00015] In methods in accordance with the second aspect of the invention, the optional step of treating the subject will only be performed in instances in which the expression profile in the tumour is established to be increased as compared to the reference profile.

[00016] In methods in accordance with the third aspect of the invention, the optional step of treating the subject will only be performed in instances in which the expression profile in the tumour is established to be decreased as compared to the reference profile.

[00017] The inventors' findings also lend themselves to the development of kits that may be used in the practice of methods of the invention.

[00018] Accordingly, in a fourth aspect, the invention provides a kit for use in selecting a suitable cancer treatment regimen in a subject with cancer, the kit comprising:

• means for determining the level of at least one protein selected from the group consisting of: subject of at least one protein selected from the group consisting of: leptin, creatine kinase muscle and brain (CK-MB); interleukin-10 (IL-10); plasminogen activator urokinase (PLAUR); interleukin-6 receptor beta (IL6Rb); TNF receptor superfamily member 6 (TNFSF 6); and TNF related apoptosis inducing ligand receptor 3 (TRAILR3) in a sample representative of the subject's circulating blood;

• information regarding a reference value against which the determined circulating level of the protein is compared;

• instructions regarding the selection of a suitable cancer treatment regimen if the determined circulating level of the protein is greater than the reference value; and

• instructions regarding the selection of a suitable cancer treatment if the determined circulating level of the protein is lower than the reference value; wherein

• selecting a cancer treatment regimen for the subject comprising provision of a VEGF signalling inhibitor if the determined circulating protein level in respect of leptin or CK- MB is greater than the reference value; and/or the determined circulating protein level in respect of at least one of IL-10; PLAUR; IL6Rb; TNFSF 6 or TRAILR3 is lower than the reference value.

[00019] In a fifth aspect the invention provides a kit for use in selecting a suitable cancer treatment regimen in a subject with cancer, the kit comprising:

• means for determining an expression profile in a tumour of the subject for at least 2 genes of the GAHGL gene signature set out in Table 1 ;

• information regarding a reference expression profile against which the determined tumour expression profile is compared;

• instructions regarding the selection of a suitable cancer treatment regimen if the determined tumour expression profile is greater than the reference expression profile; and

• instructions regarding the selection of a suitable cancer treatment regimen if the determined tumour expression profile is higher than the reference expression profile; wherein

• only if the determined tumour expression profile is greater than the reference expression profile does the suitable cancer treatment regimen comprise provision of a VEGF signalling inhibitor.

[00020] In a sixth aspect the invention provides a kit for use in selecting a suitable cancer treatment regimen in a subject with cancer, the kit comprising:

• means for determining an expression profile in a tumour of the subject for at least 2 genes of the LOPS gene signature set out in Table 1 ;

• information regarding a reference expression profile against which the determined tumour expression profile is compared;

• instructions regarding the selection of a suitable cancer treatment regimen if the determined tumour expression profile is greater than the reference expression profile; and

• instructions regarding the selection of a suitable cancer treatment regimen if the determined tumour expression profile is lower than the reference expression profile; wherein

• only if the determined tumour expression profile is lower than the reference expression profile does the suitable cancer treatment regimen comprise provision of a VEGF signalling inhibitor.

[00021] In a seventh aspect the invention provides an inhibitor of VEGF signalling for use in the treatment of cancer in a subject selected by a method in accordance with the first, second, third or eighth aspects of the invention, or for use in a method in accordance with the ninth, tenth, or eleventh aspects of the invention.

[00022] In an eighth aspect the invention provides a method for selecting a suitable cancer treatment regimen in a subject with cancer, the method comprising:

• determining an expression profile in a tumour of the subject for at least 2 genes of the GAHGL gene signature set out in Table 1 ; and

• determining an expression profile in a tumour of the subject for at least 2 genes of the LOPS gene signature set out in Table 1 ;

• comparing the expression profiles determined above with a reference expression profile; and

• selecting a cancer treatment regimen for the subject comprising provision of a VEGF signalling inhibitor if the expression profile determined in respect of the at least two genes of the GAHGL signature is increased as compared to the reference profile and the expression profile determined in respect of the al least two genes of the LOPS signature is decreased as compared to the reference profile.

[00023] In a ninth aspect the invention provides a method of treating cancer in a subject in need of such treatment, the method comprising:

• determining an expression profile in a tumour of the subject for at least 2 genes of the GAHGL signature set out in Table 1 ;

• comparing the expression profile determined with a reference expression profile; and treating the cancer with a VEGF signalling inhibitor if:

• the expression profile in respect of the GAHGL signature is increased as compared to the reference profile.

[00024] In a tenth aspect the invention provides a method of treating cancer in a subject in need of such treatment, the method comprising:

• determining an expression profile in a tumour of the subject for at least 2 genes of the LOPS signature set out in Table 1 ;

• comparing the expression profile determined with a reference expression profile; and treating the cancer with a VEGF signalling inhibitor if:

• the expression profile in respect of the LOPS signature is decreased as compared to the reference profile.

[00025] In an eleventh a method of treating cancer in a subject in need of such treatment, the method comprising:

• determining an expression profile in a tumour of the subject for at least 2 genes of the GAHGL gene signature set out in Table 1 ; and/or

• determining an expression profile in a tumour of the subject for at least 2 genes of the LOPS gene signature set out in Table 1 ; • comparing the expression profile, or expression profiles, determined with a reference expression profile; and

treating the cancer with the VEGF signalling inhibitor cediranib if:

• the expression profile in respect of the GAHGL gene signature is increased as

compared to the reference profile; or

• the expression profile in respect of the LOPS gene signature is decreased as

compared to the reference profile; or

• the expression profile in respect of the GAHGL gene signature is increased as

compared to the reference profile and

• the expression profile in respect of the LOPS gene signature is decreased as

compared to the reference profile.

[00026] In the methods of the invention comparison of the expression profiles determined with the reference expression profile allows the practitioner to determine whether or not the subject will benefit from treatment with a VEGF signalling inhibitor such as Cedirinab. Details of the methods of the eighth, ninth, tenth, or eleventh aspects of the invention, including suitable constituents of expression profiles in respect of the GAHGL and LOPS gene signatures, may be selected with reference to the various parameters or factors set out in respect of the other methods of the invention (such as the methods of the first, second or third aspects of the invention).

[00027] Generally, except for where the context requires otherwise, considerations in the present disclosure described as relating to kits of the invention should be considered to be applicable to kits in accordance with any of the fourth, fifth, or sixth aspects of the invention. Similarly, considerations in the present disclosure described as relating to methods of the invention should be considered to be applicable to kits in accordance with the invention, except for where the context requires otherwise.

[00028] Further details of embodiments of the various methods and kits of the invention are set out elsewhere in the specification.

BRIEF DESCRIPTION OF THE DRAWINGS

[00029] Figure 1 illustrates a volcano plot of soluble biomarkers associated with overall survival (OS) in chemotherapy cediranib treated mCRC patients.

[00030] Figure 2 illustrates a volcano plot of soluble biomarkers associated with overall survival (OS) in chemotherapy placebo treated mCRC patients.

[00031] Figure 3 illustrates a volcano plot of soluble biomarkers associated with overall survival (OS) only in chemotherapy cediranib treated mCRC patients. The soluble biomarkers appearing in Figure 2 were removed from the soluble biomarkers in Figure 1 to result in Figure 3. [00032] Figure 4 illustrates a plot of specific soluble biomarkers associated with overall survival (OS) only in chemotherapy cediranib treated mCRC patients. The biomarkers shown in this Figure are a subset (the proteins having p<0.05) of those shown in Figure 3.

[00033] Figure 5 illustrates a plot of specific soluble biomarkers associated with progression free survival (PFS) only in chemotherapy cediranib treated mCRC patients. This list of biomarkers does not include those that had previously been found to be associated with chemotherapy placebo treatment.

[00034] Figure 6A illustrates a Venn diagram of specific soluble biomarkers associated with OS and/or PFS, and Figure 6B tabulates the names of the markers associated with both PFS and OS.

[00035] Figure 7 shows a Forest plot of the biomarkers associated with both OS (Figure 7A) and PFS (Figure 7B).

[00036] Figure 8 shows a Kaplan-Meier curve for OS based on levels of soluble leptin.

[00037] Figure 9 shows the correlation of leptin levels with body mass index.

[00038] Figure 10 shows a Kaplan-Meier curve for OS based on body mass index.

[00039] Figure 11 shows a Forest plot of expression of significant genes associated with OS in the cediranib treated patients.

[00040] Figure 12 shows the Kaplan Meier curve for overall prognostic significance of the GAGHL genes.

[00041] Figure 13 shows a Kaplan-Meier curve for OS based on the GAGHL signature.

[00042] Figure 14 shows a Kaplan-Meier curve for overall prognostic significance of the LOPS signature.

[00043] Figure 15 shows a Kaplan-Meier curve for OS based on the LOPS signature.

[00044] Figure 16 shows the clinical tral design and work flow for the Horizon II study described herein.

[00045] Figure 17 shows shows a Kaplan-Meier curve for PFS based on levels of soluble leptin.

[00046] Figure 18 shows a Kaplan-Meier curve forPFS based on BMI .

[00047] Figure 19 shows the workflow for analysis of gene expression from the

Horizon II study samples described herein.

[00048] Figure 20 shows the strategy for gene expression analysis for the Horizon II study described herein.

[00049] Figure 21 show a shows a Forest plot for genes associated with PFS in the cediranib treated group.

[00050] Figure 22 shows genes where low expression is associated with PFS or OS benefit from cediranib.

[00051] Figure 23 shows gene where high expression is associated with PFS or OS benefit from cediranib.

[00052] Figure 24 shows a Kaplan-Meier curve for PFS based on the GAHGL signature.

[00053] Figure 25 shows a Kaplan-Meier curve for PFS based on the LOPS signature. DETAILED DESCRIPTION OF THE INVENTION

[00054] The inventors have investigated the predictive potential of serum protein and tumour gene expression in patients with metastatic colorectal cancer (mCRC) receiving either chemotherapy plus placebo or chemotherapy plus the VEGF signalling inhibitor cediranib. Their findings, which underpin the present invention, provide patient selection methods, as well as kits, that may be used in determining suitable cancer treatment regimens in subjects with cancer, with particular reference to whether the subject will gain benefit from treatment with VEGF signalling inhibitors.

[00055] The present invention is based upon the inventors' new and surprising finding that circulating levels of certain proteins (leptin; CK-MB; IL-10; PLAUR; I L6Rb; TNFSF 6; and TRAI LR3), or the profiles of expression of certain groups of genes (specifically the GAHGL gene signature or the LOPS gene signature, constituents of both of which are defined elsewhere in the disclosure), can be used as the basis for selection of suitable cancer treatment regimens.

[00056] In particular, the methods of the first aspect of the invention, and kits of the fourth aspect of the invention, are based upon the inventors' finding that circulating levels of selected proteins (including leptin and CK-MB) are increased in subjects who demonstrate a beneficial response to cancer treatment regimens in which VEGF inhibitors are used, while circulating levels of other selected proteins (including IL-10; PLAUR; IL6Rb; TNFSF 6; and TRAI LR3) are decreased in subjects who demonstrate a beneficial response to cancer treatment regimens in which VEGF inhibitors are used. The inventors have found that, when given treatment regimens of this sort, such subjects exhibit increases in both progression free survival (PFS) and overall survival (OS) that are higher than those in subjects without alterations in these circulating protein levels, illustrating the specific effectiveness of such treatment for these subjects.

[00057] The methods of the second, eighth, ninth, tenth, and eleventh aspects of the invention, and kits of the fifth aspect of the invention, are based upon the inventors' finding that the profile of expression within the tumour of genes of the so-called "GAHGL signature" is increased in subjects who demonstrate a beneficial response to cancer treatment regimens in which both VEGF inhibitors and chemotherapeutic agents are used. The inventors have found that, when given treatment regimens of this sort, such subjects exhibit increases in OS that are higher than those in subjects without increased levels GAHGL gene expression, illustrating the specific effectiveness of such treatment for these subjects. Furthermore, the inventors have found that increased expression of a subset (discussed below) of the genes associated with the GAHGL gene signature is associated with both increases in both OS and PFS.

[00058] Methods in accordance with the second, eighth, ninth, tenth, and eleventh aspects of the invention involve determining an expression profile in respect of at least 2 genes of the GAHGL signature. In suitable embodiments, the number of genes in respect of which an expression profile is determined may be 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , or more. For example, an expression profile may be determined in respect of 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 or 22, or more, genes of the GAHGL signature. In certain embodiments an expression profile may be determined in respect of 23, 24, 25, 26, 27, 28, 29, 30, 31 , 32, or 33 genes of the GAHGL signature.

[00059] Similarly, kits in accordance with the fifth aspect of the invention provide means for determining an expression profile in respect of at least 2 genes of the GAHGL signature. In suitable embodiments, a kit in accordance with the fifth aspect of the invention may comprise means for determining an expression profile in respect of 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , or more genes of the GAHGL signature. For example, a kit in accordance with the fifth aspect of the invention may comprise means for determining an expression profile in respect of 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 or 22, or more, genes of the GAHGL signature. In certain embodiments a kit in accordance with the fifth aspect of the invention may comprise means for determining an expression profile in respect of 23, 24, 25, 26, 27, 28, 29, 30, 31 , 32, or 33 genes of the GAHGL signature.

[00060] The methods of the third, eighth, ninth, tenth, and eleventh aspects of the invention, and kits of the sixth aspect of the invention, are based upon the inventors' finding that the profile of expression of genes within the tumour of genes of the so-called "LOPS signature" decreased in subjects who demonstrate a beneficial response to cancer treatment regimens in which both VEGF inhibitors and chemotherapeutic agents are used. The inventors have found that, when given treatment regimens of this sort, such subjects exhibit increases in OS that are higher than those in subjects without decreased levels of LOPS gene expression illustrating the specific effectiveness of such treatment for these subjects. Furthermore, the inventors have found that decreased expression of a subset (discussed below) of the genes associated with the LOPS gene signature is associated with both increases in both OS and progression free survival PFS.

[00061] Methods in accordance with the third, eighth, ninth, tenth, or eleventh aspects of the invention involve determining an expression profile in respect of at least 2 genes of the LOPS signature. In suitable embodiments, the number of genes in respect of which an expression profile is determined may be 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14, 15 or more. For example, an expression profile may be determined in respect of 16, 17, 18, 19, 20, 21 22, 23, 24, 25, 26, 27, 28, 29, 30 or more, genes of the LOPS signature. In certain embodiments an expression profile may be determined in respect of 31 , 32, 33, 34, 35, 36, 37, 38, 39, or 40 genes of the LOPS signature.

[00062] By the same token, kits in accordance with the sixth aspect of the invention provide means for determining an expression profile in respect of at least 2 genes of the LOPS signature. In suitable embodiments, a kit in accordance with the fifth aspect of the invention may comprise means for determining an expression profile in respect of 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14, 15, or more genes of the LOPS signature. For example, a kit in accordance with the fifth aspect of the invention may comprise means for determining an expression profile in respect of 16, 17, 18, 19, 20, 21 22, 23, 24, 25, 26, 27, 28, 29, 30, or more, genes of the LOPS signature. In certain embodiments a kit in accordance with the fifth aspect of the invention may comprise means for determining an expression profile in respect of 31 , 32, 33, 34, 35, 36, 37, 38, 39, or 40 genes of the LOPS signature.

[00063] Details regarding preferred combinations of genes from either the GAHGL or LOPS gene signatures that may be determined in appropriate embodiments of the methods or kits of the invention are considered in more detail below.

[00064] The methods and kits of the invention are useful in the selection, and optionally implementation, of a suitable cancer treatment regimen for use in a subject with cancer. In a suitable embodiment, the subject with cancer is a subject in whom cancer has already been diagnosed.

[00065] In a suitable embodiment, the subject with cancer has a cancer selected from the group consisting of: colorectal cancer, such as metastatic colorectal cancer; breast cancer; ovarian cancer, such as platinum sensitive relapsed ovarian cancer; prostate cancer; and renal cancer. In a particularly suitable embodiment, the subject is a subject with a cancer selected from the group consisting of: colorectal cancer, such as metastatic colorectal cancer; and ovarian cancer, such as platinum sensitive relapsed ovarian cancer. Definitions

[00066] For the avoidance of doubt, various terms used in the context of the present disclosure will now be further defined in the paragraphs below.

[00067] "leptin" - Leptin is a 16 kDa protein that serves as an adipokine. Leptin has a defined sequence of 167 amino acid residues.

[00068] "creatine kinase muscle and brain" - Creatine kinase muscle and brain (CK- MB) is a cytoplasmic enzyme involved in energy homeostasis and is an important serum marker for myocardial infarction. The encoded protein reversibly catalyzes the transfer of phosphate between ATP and various phosphogens such as creatine phosphate. It acts as a homodimer in striated muscle as well as in other tissues, and as a heterodimer with a similar brain isozyme in heart. [00069] "interleukin-10" - Interleukin- 10 (IL-10) is a cytokine produced primarily by monocytes and to a lesser extent by lymphocytes. IL-10 has pleiotropic effects in immunoregulation and inflammation. It down-regulates the expression of Th1 cytokines, MHC class II Ags, and costimulatory molecules on macrophages. It also enhances B cell survival, proliferation, and antibody production.

[00070] "plasminogen activator urokinase" - Plasminogen activator urokinase (PLAUR) is a serine protease involved in degradation of the extracellular matrix and possibly tumor cell migration and proliferation.

[00071] "interleukin-6 receptor beta" - lnterleukin-6 receptor beta (IL6Rb) is a signal transducer shared by many cytokines, including interleukin 6 (IL6), ciliary neurotrophic factor (CNTF), leukemia inhibitory factor (LIF), and oncostatin M (OSM). This protein functions as a part of the cytokine receptor complex.

[00072] "TNF receptor superfamily member 6" - TNF receptor superfamily member 6 (TNFSF 6) is also known as apoptosis antigen 1 (APO-1 or APT), cluster of differentiation 95 (CD95) or FAS receptor (FasR) is a death receptor on the surface of cells that leads to programmed cell death (apoptosis).

[00073] "TNF related apoptosis inducing ligand receptor 3" - TNF related apoptosis inducing ligand receptor 3 (TRAILR3) is a receptor for the cytotoxic ligand TRAIL involved in apoptosis.

[00074] "circulating level of a protein" - The methods and kits of the first and fourth aspects of the invention involve the determination of a circulating level of at least one protein (such as leptin) selected from a specified group of proteins in the subject in respect of whom the suitable cancer treatment regime is being selected. It will be appreciated that the "circulating level" of the specified protein is the level of the protein (or proteins) in question that is present in the patient's circulation. The circulating level of the protein (such as leptin) may suitably be expressed in concentration values, such as ng per ml.

[00075] The circulating protein level in the subject may be determined in any suitable sample representative of the circulation in the subject. In a suitable embodiment, the circulating protein level is determined in sample selected from the group consisting of: a blood sample; a serum sample; and a plasma sample. It will be appreciated that, due to sample processing, the absolute value of protein present in a given volume of a serum or plasma sample may not be the same as in a blood sample, but all of these remain indicative of the circulating protein level.

[00076] The circulating protein level may be determined by any suitable means known to those skilled in the art. Merely by way of example, the circulating level of a protein (such as leptin) in the subject may be determined by means of an assay in which circulating protein is detected, and the level present determined, by binding of the protein to a specific binding partner. The binding partner may be directly or indirectly labelled. Antibodies or antibody fragments that bind to proteins (such as leptin) represent particularly suitable examples of such binding partners, and these by be used in immunoassays.

[00077] Immunoassays constitute preferred embodiments of the methods of the invention, and antibodies or antibody fragments that may be used in immunoassays represent preferred constituents of kits of the invention.

[00078] Suitable examples of immunoassays of this sort that may be used in the methods or kits of the invention include those selected from the group consisting of: enzyme- linked immunosorbent assay (ELISA); radioimmunoassay (RIA); and multiplex immunoassays (such as Luminex ® assays produced by Luminex Corporation). It will be appreciated that multiplex assays (and the reagents by which these are practiced) are of particular suitability in methods or kits of the invention in which circulating levels of multiple proteins are to be determined.

[00079] Antibodies suitable for use in such embodiments, and indeed kits comprising reagents required for assays such as ELISAs, are readily commercially available.

[00080] "reference value" - The level of protein (or proteins) in the circulation of the subject is compared to a reference value for the concentration of the corresponding protein (or proteins) in the circulation. It will be appreciated that the reference value should provide information in respect of each of the one or more proteins the circulating levels of which are being assessed.

[00081] A suitable reference value may be determined with reference to the median circulating level of the protein(s) found in control population matched to the subject. The skilled person will be able to select an appropriate control to provide the requisite reference value.

[00082] Merely by way of example, the inventors have found that, in the case of leptin, a suitable reference value for use in the methods or kits of the invention is between approximately 0.066 and 114 ng/ml (values that may be particularly relevant when the sample in question is a serum sample). For example, in an embodiment utilising serum samples a reference value for leptin may be approximately 4.9ng/ml. Indeed, in a suitable example of such an embodiment a reference value in respect of leptin is 4.9ng/ml. As referred to in the claims, circulating levels of leptin above the chosen reference value are indicative that the subject will gain benefit from a treatment regimen comprising provision of a VEGF inhibitor.

[00083] In the case of CK-MB, the inventors have found that a suitable reference value for use in the methods or kits of the invention (particularly in the case of serum samples) is approximately 0.69ng/ml. Indeed, in a suitable example of such an embodiment a reference value in respect of CK-MB is 0.69ng/ml. As referred to elsewhere, circulating levels of CK- MB above the chosen reference value are indicative that the subject will gain benefit from a treatment regimen comprising provision of a VEGF inhibitor.

[00084] With respect to I L-10, a suitable reference value for use in the methods or kits of the invention (in particular with in the case of methods or kits for use in connection with serum samples) is approximately 7pg/ml. Indeed, in a suitable example of such an embodiment a reference value in respect of IL-10 is 7pg/ml. Circulating levels of I L-10 below the chosen reference value are indicative that the subject will gain benefit from a treatment regimen comprising provision of a VEGF inhibitor.

[00085] Turning to PLAUR, the inventors have found that a suitable reference value for use in the methods or kits of the invention (in particular in the case that the sample in question is a serum sample) is approximately 574pg/ml. Indeed, in a suitable example of such an embodiment a reference value in respect of PLAUR is 574.5pg/ml. Circulating levels of PLAUR below the chosen reference value are indicative that the subject will gain benefit from a treatment regimen comprising provision of a VEGF inhibitor.

[00086] For IL6Rb, a suitable reference value for use in the methods or kits of the invention (in particular with reference to serum sample) is approximately 243ng/ml. Indeed, in a suitable example of such an embodiment a reference value in respect of I L6Rb is 243ng/ml. Circulating levels of IL6Rb below the chosen reference value are indicative that the subject will gain benefit from a treatment regimen comprising provision of a VEGF inhibitor.

[00087] In the case of TNFSF 6, the inventors have found that a suitable reference value for use in the methods or kits of the invention (and particularly for serum samples) is approximately 14ng/ml. Indeed, in a suitable example of such an embodiment a reference value in respect of TNFSF 6 is 14ng/ml. As referred to elsewhere, circulating levels of TNFSF 6 above the chosen reference value are indicative that the subject will gain benefit from a treatment regimen comprising provision of a VEGF inhibitor.

[00088] Similarly, for TRAILR3, the inventors have found that a suitable reference value for use in the methods or kits of the invention (and particularly for serum samples) is approximately 14ng/ml. In a suitable example of such an embodiment a reference value in respect of TRAI LR3 is 14ng/ml. Once again, circulating levels of TRAI LR3 above the chosen reference value are indicative that the subject will gain benefit from a treatment regimen comprising provision of a VEGF inhibitor.

[00089] "determining an expression profile in a tumour of a subject" - Methods in accordance with the second, third, eighth, ninth, tenth, and eleventh aspects of the invention involve the determination of an expression profile in a tumour of the subject upon whom the method is to be practice. This expression profile is determined in respect of two or more genes selected from those that make up the gene signatures referred to. [00090] The tumour from which expression values are to be calculated will generally be the cancer (or one of the cancers) to be treated by the cancer treatment regimen.

[00091] The expression profile for the genes to be investigated may be provided by analysis of a sample representative of gene expression in the tumour in question. A suitable sample may include cells of the tumour. For example, a suitable sample may be a tumour biopsy.

[00092] Information allowing the establishment of an expression profile in respect of the genes in a tumour of this sort may be provided by any suitable method. For example, mRNA indicative of gene expression in the tumour may be extracted, and this used to develop the required gene expression profile. The skilled person will be aware of many suitable molecular profiling methods that may be used in the generation of the requisite gene expression profile, including PCR based techniques (such as Taqman or sybr green gene expression assays) or mRNA/cDNA hybridisation techniques (such as gene expression micro array, chips, or NanoString technology). As discussed further in the Examples, a suitable method by which an expression profile for genes in a tumour may be provided my utilise mRNA isolation (for example using a commercially available product such as that sold by Quiagen under the name RNeasy®), followed by quantification of expression of the genes of interest (for example using products commercially available from NanoDrop and NanoString Technology, as described in the Examples) to yield the required expression profile.

[00093] Suitable reagents that may be used in providing the required expression profiles are also considered in the context of the following definition.

[00094] "means for determining an expression profile in a tumour of the subject"

- Kits of the fifth and sixth aspects of the invention provide means for determining expression profiles for a desired number of genes in a tumour, particularly with reference to the ability to determine such expression profiles with regard to genes of the GAHGL or LOPS signatures.

[00095] The skilled person will be aware of many such suitable means for determining an expression profile of the sort required. Merely by way of example, suitable means for determining a gene expression profile may comprise complementary nucleic acids specific to the desired genes in respect of which an expression profile is to be determined. In a suitable embodiment, these complementary nucleic acid sequences may be immobilised on an array, such as a suitably coated chip, allowing the complementary sequences to bind nucleic acid molecules representative of gene expression, and thereby produce a gene expression profile. In another embodiment, the complementary nucleic acid sequences may be provided in the form of complementary primers, that allow the amplification of nucleic acid molecules representative of expression of the genes of interest in respect of which a profile is to be produced.

[00096] The means provided in the kits of the invention may also comprise further reagents that are of use in the method to be employed, including buffers, enzymes, and the like.

[00097] "GAHGL signature" - In the context of the present disclosure, the term "GAHGL signature" is used to refer collectively to the genes set out in Table 1. The term derives from the role of these genes in Glucose metabolism, Angiogenesis, Hypoxia, Glutamine metabolism, and Leptin signalling.

[00098] The methods or kits of the invention respectively involve the determination of, or means for the determination of, at least two genes, and possibly as many as 33 genes from those of the GAHGL signature.

[00099] The inventors have noted that certain genes of the GAHGL signature exhibit expression that is particularly significantly altered as compared to controls. These genes represent a preferred subset of genes expression profiles of which may be determined in association with the methods or kits of the invention. Of particular interest in this respect are the group consisting of: PCK1 ; VEGFR1 ; NR1 H3; NR1 H2; PPARGCIa; PFKFB3; VEGFC; ACACB; and PDHK2. Combinations of two or more genes independently selected by this group may be selected for the determination of an expression profile, or the provision of means for determining an expression profile, in the methods or kits of the invention.

[000100] Merely by way of example, genes, an expression profile of which is to be determined, or means for determining an expression profile are to be provided, may be independently selected from:

• between 2 and 9 of PCK1 ; VEGFR1 ; NR1 H3; NR1 H2; PPARGCIa; PFKFB3;

VEGFC; ACACB; and PDHK2.

• between 2 and 8 of PCK1 ; VEGFR1 ; NR1 H3; NR1 H2; PPARGCIa; PFKFB3;

VEGFC; and ACACB.

• between 2 and 7 of PCK1 ; VEGFR1 ; NR1 H3; NR1 H2; PPARGCIa; PFKFB3; and VEGFC.

• between 2 and 6 of PCK1 ; VEGFR1 ; NR1 H3; NR1 H2; PPARGCIa; and PFKFB3.

• between 2 and 5 of PCK1 ; VEGFR1 ; NR1 H3; NR1 H2; and PPARGCIa.

• between 2 and 4 of PCK1 ; VEGFR1 ; NR1 H3; and NR1 H2.

• between 2 and 3 of PCK1 ; VEGFR1 ; NR1 H3.

• PCK1 and VEGFR1.

[000101] Another selection of preferred genes from among the GAHGL signature may be made on the basis of statistical association with beneficial increases both OS and PFS as compared to controls. In this respect, two or more preferred genes may be independently selected from the group consisting of: VEGFR1 ; NR1 H2; PPARGCI a; ACACB; PDHK2; MDH1 ; ME2; MAGL; VEGFA; PFKFB2; and SLC22A8.

[000102] "LOPS signature" - In the context of the present disclosure, the LOPS signature is used to refer collectively to the genes set out in Table 1. The term derives from the role of these genes in Lipid metabolism, Oxidative metabolism, Proliferation, and Serine/Glycine metabolism.

[000103] The methods or kits of the invention respectively involve the determination of, or means for the determination of, at least two genes, and possibly as many as 40 genes from those of the LOPS signature.

[000104] The inventors have noted that certain genes of the LOPS signature exhibit expression that is particularly significantly altered as compared to controls. These genes represent a preferred subset of genes expression profiles of which may be determined in association with the methods or kits of the invention. Of particular interest in this respect are the group consisting of: SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN; VEGFB; ACACA; MLYCD; NDUFA4; TIGAR; PCNA; and PDP2. Combinations of two or more genes independently selected by this group may be selected for the determination of an expression profile, or the provision of means for determining an expression profile, in the methods or kits of the invention.

[000105] Merely by way of example, genes, an expression profile of which is to be determined, or means for determining an expression profile are to be provided, may be independently selected from:

• between 2 and 1 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN;

VEGFB; ACACA; MLYCD; NDUFA4; TIGAR; PCNA; and PDP2.

• between 2 and 13 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN;

VEGFB; ACACA; MLYCD; NDUFA4; TIGAR; and PCNA.

• between 2 and 12 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN;

VEGFB; ACACA; MLYCD; NDUFA4; and TIGAR.

• between 2 and 11 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN;

VEGFB; ACACA; MLYCD; and NDUFA4.

• between 2 and 10 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN;

VEGFB; ACACA; and MLYCD.

• between 2 and 9 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN;

VEGFB; and ACACA.

• between 2 and 8 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; FASN; and VEGFB. • between 2 and 7 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; HMGCR; and FASN.

• between 2 and 6 of SHMT2; NDUFB6; SLC25A10; DHCR7; LDHB; and HMGCR.

• between 2 and 5 of SHMT2; NDUFB6; SLC25A10; DHCR7; and LDHB.

• between 2 and 4 of SHMT2; NDUFB6; SLC25A10; and DHCR7.

• between 2 and 3 of SHMT2; NDUFB6; and SLC25A10.

• SHMT2 and NDUFB6.

[000106] Another selection of preferred genes from among the LOPS signature may be made on the basis of statistical association with beneficial increases both OS and PFS as compared to controls. In this respect, two or more preferred genes may be independently selected from the group consisting of: SHMT2; NDUFB6; SLC25A10; DHCR7; HMGCR; FASN; VEGFB; MLYCD; NDUFA4; GSK3B; PDK1 ; PDP1 ; MTOR; SREBF2; SLC1A5; NDUFB9; LDHA; SMHT1 ; and ACSL3.

[000107] "VEGF signalling inhibitor" - In a suitable embodiment, a VEGF signalling inhibitor to be used in a cancer treatment regimen selected through use of a method or kit of the invention may be selected from the group consisting of: cediranib; Sunitinib; Sorafenib; pazopanib; Tivozanib; vandetanib; soraninib; axitinib; cabozantinib; Bevacizumab; aflibercept; and ramucirurlmab.

[000108] Of the examples set out above, cediranib constitutes a preferred VEGF signalling inhibitor that may be used in cancer treatment regimens selected through use of the methods of the invention.

[000109] "Cediranib" - Cediranib (lUPAC name 4-[(4-fluoro-2-methyl-1 H-indol-5- yl)oxy]-6-methoxy-7-[3-(pyrrolidin-1-yl)propoxy]quinazoline) is a compound that exhibits the ability to inhibit VEGF signalling, an activity that is believed to occur as a result of inhibition of VEGF receptor tyrosine kinsases. Cediranib is generally suitable for oral administration to subjects in which it is wished to inhibit VEGF signalling. Cediranib is also alternatively known as 4-[(4-fluoro-2-methyl-1 - -indol-5-yl)oxy]-6-methoxy-7-[3-(pyrrolidin-1- yl)propoxy]quinazoline.

[000110] "cancer treatment regimen" - A cancer treatment regimen the selection of which may be indicated by the methods or kits of the invention comprises provision to the subject of a VEGF signalling inhibitor. Examples of VEGF inhibitors that may be employed in such treatment regimens are set out above. Suitably the VEGF inhibitor may be provided as the only (or primary) anti-cancer agent in a treatment regimen. Details of treatment regimens using VEGF inhibitors in this manner will be known to those skilled in the art, and known regimens of this type may be used in accordance with the present invention.

[000111] Alternatively, a suitable treatment regimen may comprise provision of a VEGF signalling inhibitor, and also provision to the subject of a further anti-cancer agent, such as a chemotherapeutic agent and/or an epidermal growth factor receptor (EGFR) inhibitor. Examples of suitable chemotherapeutic agents are considered elsewhere in the present specification.

[000112] Examples of cancer treatment regimens in which both VEGF inhibitors and chemotherapeutic agents are provided to a subject requiring treatment for cancer will be well known to those skilled in the art. Such known cancer treatment regimens may be eminently suitable for selection by virtue of the methods of the invention.

[000113] Suitably cancer treatment regimens that involve administration of both the VEGF signalling inhibitor and a chemotherapeutic agent may involve administration of these agents at substantially the same time. Alternatively, suitable treatment regimens may involve sequential administration of the VEGF signalling inhibitor and chemotherapeutic agent. In embodiments in which the VEGF signalling inhibitor and chemotherapeutic agent are both to be administered at the same time, the VEGF signalling inhibitor and chemotherapeutic agent may be administered as part of the same formulation, or in separate formulations.

[000114] As referred to above, in suitable embodiments, a cancer treatment regimen may comprise, the provision of anti-cancer agents such as EGFR inhibitors (for example Erbitux) as combination partner, in addition to the provision of a VEGF signalling inhibitor and/or a chemotherapeutic agent.

[000115] "chemotherapeutic agent" - A chemotherapeutic agent suitable for use in a cancer treatment regimen selection of which is indicated by a method or kit of the invention may be one that is conventionally used for treatment of the cancer in question. For example, in the case that the cancer to be treated is colorectal cancer or ovarian cancer, a suitable chemotherapeutic agent for used in a cancer treatment regimen selected in accordance with the invention may be one that is conventionally used for treatment of colorectal and/or ovarian cancer.

[000116] Merely by way of example, a chemotherapeutic agent for use in a cancer treatment regimen selected in accordance with the methods of the present invention may suitably be selected from the group consisting of: folinic acid (Leucovirorin); fluorouracil; oxaliplatin; irinotecan; and capecitabine.

[000117] The invention will now be further illustrated, with reference to the following non-limiting Examples.

EXAMPLES

[000118] The following abbreviations are used throughout the Examples:

OS: overall survival;

PFS: progression free survival HR: hazard ratio

CI : confidence interval

BMI: body mass index

GAHGL: Glucose metabolism, Angiogenesis, Hypoxia, glutamine metabolism, Leptin signalling

LOPS: Lipid metabolism, Oxidative activity, Proliferation and Serine synthesis.

FFPE: Fixed formalin paraffin embeddedPlac: placebo

Ced: Cediranib

[000119] Materials and methods

[000120] Patients and samples

[000121] Eligible patients enrolled in the phase I I I double-blind HORIZON II study were ≥18 years old with histologic/cytologic confirmation of metastatic (stage IV) CRC; a World Health Organization (WHO) performance status of 0/1 ; and a life expectancy of ≥12 weeks (Reference 1). Patients must not have received prior systemic therapy for mCRC; any adjuvant (or neoadjuvant) therapy with oxaliplatin or 5-FU must have been received >12 months or >6 months, respectively, before study entry. Patients were initially randomly assigned 1 : 1 : 1 to receive once-per-day cediranib 30 mg, cediranib 20 mg, or placebo in combination with FOLFOX CAPOX (Figure 16, 19). Because recruitment to the cediranib 30 mg arm was discontinued (Reference 1), correlation with clinical outcome was performed only on patients treated with cediranib 20 mg or placebo in combination with FOLFOX/CAPOX.

[000122] Regarding the samples used for gene expression analysis, 354 fixed formalin embedded biopsy samples were centrally available in AstraZeneca bio-bank from this trial. Two FFPE sections were generated from patient biopsies (Figure 20). One section was stained for H&E coloration and the second section used for RNA extraction. After histological assessment of the specimens by a pathologist, viable tumour areas were marked on the H&E slide and then superimposed with the unstained section in order to replicate the marking (Figure 20). Tumour tissues were then macro-dissected in order to exclude necrotic parts and normal tissue of the analysis and RNA was extracted. 95 samples were excluded from the analysis after histological assessment (n=71 ; lack or low amount of tumour) and RNA quality control (n=24). As described in CONSORT diagram (Figure 6 A), genes expression were quantified in 259 mCRC patients out of 1076 initially randomized and allocated to receive chemotherapy plus placebo (95/358 ; 26.5%), chemotherapy plus cediranib 20 mg (106/502 ; 21.1 %) and chemotherapy plus cediranib 30mg 58/216 ; 26.9%). The 259 analyzed samples were included in the clustering analyses. Because cediranib 30 mg was discontinued in this trial (Reference 1), only the patients treated with chemotherapy plus placebo and chemotherapy plus cediranib 20 mg were used for correlation with clinical outcome.

[000123] Serum protein analysis

[000124] Blood samples from consenting patients were collected into SST (serum separated tubes and centrifuged within 1 hour for 15 minutes at 3000g. The obtained serums were transferred into vials immediately stored at -80°C and not repeatedly freeze/thawed through the whole process. The serum samples were shipped and analyzed centrally at Rules-Based Medicine (Myriad RBM, Austin, TX). 207 have been quantified by using a luminex bead-based multiplex immunodetection methodology as previously described. Myriad RBM's multi-analyte profiles (MAPs) have been validated to Clinical Laboratory Standards Institute (formerly NCCLS) guidelines based upon the principles of immunoassay. This study used the OncologyMAP version 1. Each assay is developed as a single test to establish the sensitivity and dynamic range necessary for that analyte. Once the assay is optimized, it is ready to be incorporated into a multiplex. Multiplexed calibrators (8 levels per analyte) and controls (3 levels per analyte) are developed to monitor the assays' performance. Key performance parameters such as lower limit of quantification, precision, cross-reactivity, linearity, spike-recovery, dynamic range, matrix interference, freeze-thaw stability, and short-term sample stability are established for every assay. 582 samples from eligible consenting patients were analysis as described in CONSORT diagram (Figure 16). Analyzed proteins were selected based on their relevant in angiogenesis, tumour progression. Some "hypothesis-free proteins" have been included based on the availability of the assays developed by Myriad RBM.

[000125] Gene expression analysis from biopsy samples

[000126] Samples have been randomize and processed in batch of 24 samples. Total RNA was extracted from FFPE samples using RNeasy FFPE kit from Qiagen (cat. number #73504). Briefly, using a scalpel, one 5μΜ tumour section was scraped from a slide into a microcentrifuge tube containing deparaffinization solution from Qiagen (cat. number #19093). Samples were vortexed for 10s, centrifuged and incubated at 56°C with proteinase K overnight. The next day, samples were incubated with DNAse I for 15 min and RNA was purified using Qiagen columns as described in the kit's instructions in 28uL elution volume. RNA was quantified by NanoDrop and samples with a yield < 5ng/uL and/or a 260/280 ratio < 1.6 were not included in any further analysis. 100 ng RNA was used to quantify gene expression using NanoString Technology.

[000127] Table 2 describes the list of genes analyzed and the housekeeping genes used for normalization. Analyzed genes were selected based on their relevant link with leptin, cell metabolism and angiogenesis.

RESULTS

[000128] Brief summary of results shown in Figures 1-7. [000129] The results shown in Figures 1-7 concern the identification of serum biomarker associated with clinical outcome in mCRC patients related with chemotherapy plus cediranib.

[000130] Briefly, the concentrations of 207 serum proteins were measured by multiplex analyses in 582 patients with mCRC. Each quantified protein is represented as a dot. For each protein, patients were dichotomized into high (above 1.0) and low categories (below 1.0) based on the median baseline serum concentrations. Cox regression models were used to compare OS of patients (high vs. low categories as described above) in response to chemo-ced (Figure 1) and chemo-plac treatment (Figure 2). HRs and Pvalues associated with these analyses are represented by the volcano plots for all the serum proteins. Proteins of interest that show a significant HR (p<0.05) appear as dots below the horizontal line labelled as "p-value<0.05" in Figures 1 , 2, and 3. Figure 3 illustrates volcano plots that show the proteins only associated with OS in patients treated with cediranib. Proteins found already associated with response to chemo-plac were excluded in these plots. A magnification on the protein found with a p<0.05 is shown (Figure 4). Figure 5 shows that identical analyses led to the identification of serum proteins correlated with PFS only in patients treated with chemo-ced. Figure 6A shows a Venn diagram highlighting the number of proteins associated with PFS and OS, PFS exclusively and OS exclusively in patient treated with chemo-ced. Figure 6B sets out details of the main biological functions of the proteins found associated with PFS and OS. Forest plots show the impact of serum concentrations and treatments on OS (Figure 7A) and PFS (Figure 7B).

[000131] Brief summary of results shown in Figure 8-10

[000132] The results shown in Figures 8-10 illustrate that high serum leptin concentrations and obesity improve OS of patients treated with chemo-ced.

[000133] Figure 8 shows Kaplan Meyer analyses showing the impact of high versus low serum concentrations and OS in response to chemotherapy ± cediranib. The table insert shows the Hazard Ratio (HR), 95% confidence interval (95%CI) and p-value for each of the comparators indicated. Figure 9 shows linear regression analysis showing the correlation between serum concentrations and BMI in HORIZON II trial. Figure 10 sets out Kaplan Meyer curves showing the impact of high versus low BMI on OS in response to chemotherapy ± cediranib. The table insert shows the Hazard Ratio (HR), 95% confidence interval (95%CI) and p-value for each of the comparators indicated.

[000134] Brief summary of results shown in Figure11

[000135] Figure 11- illustrates that genes involved in GAHGL and LOPS are associated with benefit to cediranib on OS and PFS in mCRC patients.

[000136] Baseline gene expression was quantified from diagnostic fixed formalin CRC biopsies using NanoString technology. Figure 11 shows a forest plot which represents the impact of treatment (chemo-plac vs. chemo-ced) in patient groups defined as high and low concentrations (relative to the median for each gene) on OS for the significant genes. The biological functions of the genes are as follows:

[000137] Lipid metabolism: ACSL3; SREBF2; SCD; MLYCD; ACACA; FASN; HMGCR; and DHCR7

[000138] Oxidative metabolism: FH, NDUFB9; ME3; NDUFB5; NDUFC2; PDP1 ; PDP2;

NDUFA4; SLC25A10; and NDUFB6

[000139] Proliferation: CCND2; MKI67; and PCNA

[000140] Serine/glycine metabolism: SHMT1 ; and SHMT2

[000141] Glucose metabolism: PCK1 ; PFKB3; PKHK2; HK1 ; SLC2A4; and PFKFB2 [000142] Angiogenesis/vasculature: VEGFR1 ; VEGFC; VEGFR2; VEGFA; TIE1 ; and NRP2

[000143] Hypoxia: CAV1 ; HIF1A; and CA9

[000144] Glutamine metabolism: GLS2; GLUD2; and GLUD1

[000145] Leptin signalling: LEPR

[000146] An asterisk (*) notation indicates genes that are associated with PFS benefit (as in Figure 21).

[000147] Brief summary of results shown in Table 1-3 and Figures 12-15

[000148] The results set out in Figures 12-15 illustrate that gene expression signatures predict OS benefit in mCRC patients treated with chemo-ced.

[000149] Patients were stratified using independent hierarchical clustering based on the level of expression of genes shown in Table 2. This revealed two clusters of patient identified by differential co-expression of genes mainly involved in GAHGL (shown in Tablel). The levels of expression for the most significant genes that have been used to define these clusters are shown in Table 3. The patients could therefore be separated based on hierarchical clustering of co-expressed genes mainly involved in GAHGL into two clusters of patients: high vs. low expression of GAHGL. Figure 12 shows a comparison of the OS in patients with high and low expression of GAHGL showing the absence of prognostic value of this signature independently of treatments. Figure 13 sets out Kaplan Meyer curves and Cox regression analyses comparing the effect of chemo-plac and chemo- ced in patients with high and low GAHGL expression on OS. The table insert shows the Hazard Ratio (HR), 95% confidence interval (95%CI) and p-value for each of the comparators indicated. A second independent hierarchical clustering analysis based on the level of expression of LOPS genes shown in Table 2. also revealed two clusters of patient identified by differential co-expression of genes mainly involved in GAHGL (shown in Tablel). The levels of expression for the most significant genes that have been used to define these clusters are shown in Table 4. This panel of genes also allowed stratification of the patients based on hierarchical clustering of co-expressed genes involved in LOPS showed two clusters of patients: high vs. low expression of LOPS. Figure 14 sets out a comparison of the OS in patients with high and low expression of LOPS showing the absence of prognostic value of this signature independently of treatments. Figure 15 shows Kaplan Meyer curves and Cox regression analyses comparing the effect of chemo-plac and chemo-ced in patients with high and low LOPSL expression on OS. The table insert shows the Hazard Ratio (HR), 95% confidence interval (95%CI) and p-value for each of the comparators indicated.

[000150] Brief summary of results shown in Figures 16-18

[000151] Figure 16 is a CONSORT diagram of the serum samples used in HORIZON II trial. Figure 17 sets out Kaplan Meyer analyses showing the impact of high versus low serum concentrations and PFS in response to chemotherapy ± cediranib. The table insert shows the Hazard Ratio (HR), 95% confidence interval (95%CI) and p-value for each of the comparators indicated. Figure 18 shows Kaplan Meyer curves showing the impact of high versus low BMI on PFS in response to chemotherapy ± cediranib. The table insert shows the Hazard Ratio (HR), 95% confidence interval (95%CI) and p-value for each of the comparators indicated.

[000152] Brief summary of results shown in Figures 19-20

[000153] Figure 19 is a CONSORT diagram of the tumour biopsy samples used in HORIZON II trial. Figure 20 shows a schematic view of the methodology to identify genes associated with clinical outcomes in HORIZON II trial.

[000154] Brief summary of results shown in Figure 21-

[000155] Figure 21 shows a forest plot that represents the impact of treatment (chemo- plac vs. chemo-ced) in patient groups defined as high and low concentrations (relative to the median for each gene) on PFS for the significant genes. The biological functions of the genes are as follows:

[000156] Lipid metabolism: HMGCS1 ; MVD; PPARd; PPARG; MVK; FASN; CAMKK2;

HMGCR; APOC1 ; DHCR7; ACSL3; PPARA; SREBF2; and MLYCD

[000157] Oxidative metabolism: NDUFB6; PDP1 ; NDUFB9; SLC25A10; and NDUFA4

[000158] Proliferation: CCND2; MKI67; and PCNA

[000159] Serine/glycine metabolism: PSAT1 ; SHMT1 ; and SHMT2

[000160] Glucose metabolism: PDHK1 ; and PFKFB2

[000161] Angiogenesis/vasculature: VEGFR1 ; VEGFR3; and VEGFA

[000162] Hypoxia: EPAS1

[000163] An asterisk (*) notation indicates genes that are associated with OS benefit (see Figure 11). Figure 22 and 23 show Venn diagrams showing the number and the name of genes associated with OS and/or PFS. [000164] Brief summary of results shown in Figure 24 and 25

[000165] Figure 24 shows Kaplan Meyer curves and Cox regression analyses comparing the effect of chemo-plac and chemo-ced in patients with high and low GAHGL expression on PFS. Figure 25 shows Kaplan Meyer curves and Cox regression analyses comparing the effect of chemo-plac and chemo-ced in patients with high and low LOPSL expression on PFS.

[000166] Markers identified in mCRC patients treated with chemo-ced.

[000167] In order to identify biomarkers associated with clinical outcome in VEGFi treated patients, we have retrospectively correlated the concentration of serum proteins with progression free survival (PFS) and overall survival (OS) in 582 mCRC patients enrolled in the Phase III, randomized, double blind trial, HORIZON II (Reference 1). For each protein, patients were dichotomized into high and low categories as described above, based on the median baseline serum concentrations. Correlation with OS was performed using a two separated Cox proportional hazards model comparing sensitivity of patients (high vs low categories as described above) in response to 1) chemo-ced (Figure 1) and 2) chemo-plac treatment (Figure 2). In Figures 1 , 2, and 3, HRs and p-values associated with this analysis are represented as volcano plots for all the serum proteins. In particular, the proteins of interest showing a significant HR (p<0.05) appear as dots below the horizontal line labelled as "p-value<0.05". In order to focus our analysis on the identification of biomarkers related with cediranib benefit (Figure 3), we excluded the proteins that were already associated with OS in chemo-plac (Figure 2) from the volcano plot in Figure 1. This model revealed a number of biomarkers significantly associated with OS in patients receiving chemo-ced but not chemo-plac (Figure 4). The same approach led to the identification of several biomarkers correlated with PFS in response to chemo-ced (Figure , 5). Venn diagram (Figure 6A) highlighted seven markers associated with both PFS and OS. HR, CI and pvalues are represented by the forest plots for OS (Figure 7A) and PFS (Figure 7B). Patients who had high concentration of Leptin (LEP) and creatine kinase muscle and brain (CK-MB) had longer PFS and OS than did those with low concentrations. Patients with low concentrations of interleukin 10 (IL-10), plasminogen activator urokinase (PLAUR), interleukin 6 receptor beta (IL6Rb), TNF receptor superfamily member 6 (TNFRSF 6) and TNF related apoptosis inducing ligand receptor 3 (TRAILR3) had longer PFS and OS than did those with high concentration. Figure 6B described the biological functions of these proteins.

[000168] High serum leptin and obesity predict improved overall survival in cediranib treated patients with mCRC.

[000169] The association between leptin concentrations and clinical outcome in cediranib treated patients was of a particular interest for several reasons. First, leptin was the most significant protein associated with prolonged OS only in cediranib treated patients. Second, leptin has been shown to modify many aspects of tumour biology including angiogenesis (Park EMM, 201 1). Finally, leptin is an adipokine strongly correlated with obesity and obesity has been associated with the risk to develop many types of cancers (Vucenic, ann N Y, 2012) including CRC. The impact of leptin status and treatments on OS is represented by Kaplan Meier curves and HR, CI, pvalues show the comparisons between the different arms (Figure 8). Interaction between treatments and leptin status was significant for OS (Pinteraction<0.001) and PFS (Pinteraction=0.001) (Figure 17). Patients with high leptin concentration showed an improved OS in response to chemo-ced when compared to low leptin patients (HR=0.55, CI=0.40-0.75, p<0.001) but also when compared to patient treated with chemo-plac (HR=0.66, CI=0.47-0.93, p=0.016). Importantly, serum leptin concentration was not associated with OS in patients treated with chemo-plac (HR=0.90, Cl=0.66-1.23, p=0.52).

[000170] Baseline tumour mRNA profiling reveals genes associated with longer survival in mCRC patients treated with cediranib.

[000171] Obesity and leptin have been linked with biological functions including angiogenesis and metabolism. Aerobic glycolysis or glycolytic pathway, also known as "Warburg effect" and angiogenesis are hallmarks of cancer. Therefore we investigated whether expression of genes involved in angiogenesis and tumour metabolism (Table 2) was correlated with clinical response to cediranib in HORIZON II trial (Figure 1 1). After histological assessment of 354 available fixed formalin and paraffin embedded diagnostic biopsies, viable tumour areas were marked and macro-dissected in order to exclude necrotic regions and non tumour tissue and RNA was extracted (Figure 20). 95 samples were excluded from the analysis because of poor tissue quality, lack of tumour tissue or low RNA yield/quality (Figure 19). Gene expression was analyzed in 259 samples representing 26.5% of the patients treated with chemotherapy plus placebo (n=95/358), 21.1 % of the patients treated with chemotherapy plus cediranib 20mg (n= 106/502) and 26. 9% of the patients treated with chemotherapy plus cediranib 30mg (n=58/216) (Figure 19) in the whole cohort. We performed two different types of statistical analyses to identify predictive genes for response to cediranib.

[000172] We stratified the patient population into two groups for each gene based on high and low gene expression (relative to median) gene and compared the OS and PFS in the patients treated with chemo-plac versus chemo-ced using a Cox regression model. The genes significantly associated with OS (Figure 1 1) and PFS (Figure 21) are listed on the forest plots showing the HR, CI and Pvalues. High expression of genes involved in Glucose metabolism, Angiogenesis, Hypoxia, Glutamine metabolism, Leptin signaling (GAHGL) was correlated with an improved OS in patients treated with chemo-ced compared to chemo-plac (Figure 11). However, low expression of genes involved in Lipid metabolism, Oxidative activity, Proliferation and Serine synthesis (LOPS) was associated with longer survival in patients treated with chemo-ced compared to chemo-plac. Many of these genes (stars) were found also associated with PFS as illustrated by Venn diagrams (Figure 22 and 23).

[000173] To determine how these genes were correlated with each other and identify some cluster of co-expressed genes that define patient population associated with differential response to chemotherapy ± cediranib, we performed hierarchical clustering analyses aiming to stratify the mCRC patient population based on these sub-clusters of co- expressed genes involved in common biological functions. Interestingly we found two clusters of co-expressed genes involved in GAHGL (Table 3) and LOPS (Table 4). We examined the effect of these signatures on OS (Figure 12 , 13, 14 , 15) and PFS (Figure 24, 25) in response to chemotherapy ± cediranib. Importantly, independent of the treatments received, these signatures were not prognostic on OS (Figure 12). However we found a significant interaction between treatments and GAHGL (Pinteraction=0.028) and LOPS (Pinteraction=0.002) signatures on OS. These data indicate that patients with high expression of GAHGL gene signature showed improved clinical outcomes in response to cediranib addition to chemotherapy on OS when compared to chemotherapy plus placebo (HR= 0.50; CI=0.30-0.83, p= 0.007). No benefit on OS was observed in response to cediranib treatment in patients with low expression of this signature (HR= 1.03; Cl=0.62- 1.71 , p= 0.900). This signature was not associated with PFS benefit (Pinteraction=0.204; Figure 24). A strong significant interaction between treatments and LOPS signature (p=0.002) was observed (Figure 15). Patients with a low gene expression of this signature showed improved OS in response to cediranib (Figure 15), particularly when compared to patients treated with chemotherapy plus placebo (HR= 0.42; CI=0.26-0.68, p< 0.001). LOPS gene signature was also predictive on PFS (Pinteraction=0.026; Figure 25).

[000174] Overall, this data show that baseline expression of gene involved in angiogenesis and tumour metabolism is able to segregate patients with mCRC into subgroups and to predict response to cediranib based therapy.

DISCUSSION

[000175] Link between obesity/body composition and sensitivity to VEGFi:

[000176] The correlation between leptin, an adipocytes derived cytokine, and hypoxic tumour gene expression in our study suggested the existence of a potential endocrine action of the adipose tissue on the tumour biology that may explained the predictive value observed for leptin in cediranib treated patients.

[000177] Link between leptin and tumour biology (mainly metabolism):

[000178] In HORIZON II, high leptin concentrations and obesity were found associated with improved PFS and OS in response to chemo-ced when compared to chemo-plac. However, the interaction between leptin status and treatments on OS was highly significant (Pinteraction<0.001 on OS and Pinteraction=0.001 on PFS) compared to the interaction between BMI and treatment (Pinteraction=0.024 on OS and Pinteraction=0.028 on PFS). This indicates that leptin may be a stronger predictive marker than BMI on clinical outcomes of mCRC patient treated with chemo-ced. Leptin binds to leptin receptor (LEPR) and can regulate hematopoiesis and proliferation. In HORIZON I I patients, we have not found any correlation between leptin concentrations and the expression of genes involved in angiogenesis and vasculature (data not shown). However we found serum leptin concentrations associated with the expression of glycolytic/hypoxic genes such as HIF1a, SLC16A3, CA-9 and PDP1.

[000179] Our data show LEPR clustered with a large number of genes involved in angiogenesis/vasculature and hypoxic tumour metabolism. In vitro, we found LEPR correlated with lactate release and showed that short term leptin treatment in Lovo cells induces an increased glycolytic activity indicative of a hypoxic-like phenotype. Our data suggest that leptin can switches the metabolic utilization of glucose from a mitochondrial oxidative phosphorylation to an aerobic glycolytic pathway as shown by an inhibition of the PDC activity, the rate determining step for glucose oxidation. In line with this data, we showed that leptin is able to regulate the expression of genes involved in lipogenesis and fatty acid oxidation in cancer cells potentially through an activation of AMPK signaling (Reference 2). This hypothesis is particularly strengthened by the negative correlation we observed between the expression of PDP1 , a PDC activator, and serum leptin concentrations.

[000180] Link between tumour metabolism and VEGFi sensitivity:

[000181] Among GAHGL and LOPS signatures, we found the expression of genes involved in glucose and lipid metabolism correlated with tumour shrinkage in preclinical models and associated with PFS and/or OS in response to cediranib based treatment in mCRC patients. In line with this idea, in house data showed similar inhibition of vessel density in sensitive (SW620) vs less responsive (H526) models following cediranib treatment (data not shown, Farren et al). Moreover, Tomaso et al. (Reference 3) published "lack of rebound" revascularization as mode of escape after cediranib therapy in patients with GBM. These observations suggest that intrinsic sensitivity and/or adaptive resistance to VEFGi are likely to be very complex mechanisms to dissect, involving not only vasculature components but also some tumour cells autonomous components.

[000182] Based on preclinical and clinical data, our results revealed that high expression of GAHGL genes (including glycolytic gene) was associated with better sensitivity to cediranib whereas high expression of LOPS genes (including mitochondrial dependent genes) was linked to a lack of cediranib response. The aerobic glycolysis is inefficient in terms of ATP production as it provides only 4 ATP/mol glucose when oxidative phosphorylation generates 36 ATP/mol glucose. Although the aerobic glycolysis is has been considered as one of the hallmarks of cancer and have been considered to confer advantage to tumour cells for decades, more and more evidences showed that complete catabolism of pyruvate - the end product of glycolysis - to lactate and C02 may be counterproductive because it may limit the availability of upstream precursors critical for biomass production and proliferation (Reference 4).

[000183] The idea of the metabolic requirements of the tumour cells as an intrinsic mechanism for sensitivity to VEGFi suggests the existence of different capacity of metabolic plasticity to adapt under VEGFi treatment pressure between tumours. Interestingly, we observed an increase expression of several glycolytic/hypoxic genes (including CA-9) in the less sensitive xenograft H526 but not in the highly sensitive models A498 treated with cediranib.

[000184] References

[000185] Reference 1 : Hoff PM, Hochhaus A, Pestalozzi BC, Tebbutt NC, Li J, Kim TW, et al. Cediranib plus FOLFOX/CAPOX versus placebo plus FOLFOX/CAPOX in patients with previously untreated metastatic colorectal cancer: A randomized, double-blind, phase I II study (HORIZON II). J Clin Oncol 2012;30(29):3596-603.

[000186] Reference 2: Luo Z, Zang M, Guo W. AMPK as a metabolic tumor suppressor: Control of metabolism and cell growth. Future Oncol 2010;6(3):457-70.

[000187] Reference 3: di Tomaso E, Snuderl M, Kamoun WS, Duda DG, Auluck PK, Fazlollahi L, et al. Glioblastoma recurrence after cediranib therapy in patients: Lack of "rebound" revascularization as mode of escape. Cancer Res 2011 ;71 (1): 19-28.

[000188] Reference 4: Vander Heiden MG, Lunt SY, Dayton TL, Fiske BP, Israelsen WJ, Mattaini KR, et al. Metabolic pathway alterations that support cell proliferation. Cold Spring Harb Symp Quant Biol 201 1 ;76:325-34.

[000189] TABLE 1

[000190] TABLE 2

0191] TABLE 2 (continued)

0192] TABLE 2 (continued)

0193] TABLE 2 (continued)

[000194] TABLE 2 (continued)

[000195] TABLE 2 (continued)

[000196] TABLE 3

[000197] TABLE 3 (continued)

[000198] TABLE 3 (continued)

[000199] TABLE 3 (continued)

[000200] TABLE 3 (continued)

[000201] TABLE 3 (continued)

[000202] TABLE 3 (continued)

[000203] TABLE 3 (continued)

[000204] TABLE 3 (continued)

[000205] TABLE 3 (continued)

[000206] TABLE 3 (continued)

[000207] TABLE 3 (continued)

Genes

[000208] TABLE 3 (continued)

[000209] TABLE 3 (continued)

[000210] TABLE 3 (continued)

[000211] TABLE 3 (continued)

[000212] TABLE 4

Genes

3] TABLE 4 (continued)

4] TABLE 4 (continued)

[000215] TABLE 4 (continued)

Genes

[000216] TABLE 4 (continued)

Genes

[000217] TABLE 4 (continued)

Genes

[000218] TABLE 4 (continued)

Genes

[000219] TABLE 4 (continued)

Genes

[000220] TABLE 4 (continued)

Genes

[000221] TABLE 4 (continued)

Genes

[000222] TABLE 4 (continued)

Genes

[000223] TABLE 4 (continued)

Genes

[000224] TABLE 4 (continued)

Genes

[000225] TABLE 4 (continued)

Genes

[000226] TABLE 4 (continued)

Genes