WO2016115149A1 - Neuropilin-1 as a serum based biomarker - Google Patents

Abstract

Prognostic and predictive methods of diagnostic methods for determining whether a human subject will be highly responsive or less responsive to treatment with VEGF inhibitors, such as tivozanib (AV-951) is disclosed. The methods are based on measurement of circulating levels neuropilin-1 in serum.

Description

NEUROPILIN-1 AS A SERUM BASED BIOMARKER

CROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of, and priority to, U.S. provisional application serial number 62/102,432, filed January 12, 2015, the contents of which are hereby incorporated by reference herein in their entirety.

FIELD OF THE INVENTION

[0002] The field of the invention is molecular biology, genetics, oncology, bioinformatics and clinical diagnostics.

BACKGROUND OF THE INVENTION [0003] Most cancer therapeutics are effective in some subjects, but not others. This results in part from genetic variation among tumors, and can be observed even among tumors within the same subject. Variable subject response is particularly pronounced with respect to targeted therapeutics. Therefore, the full potential of targeted therapies cannot be realized without suitable tests for determining which subjects will benefit from which drugs. According to the National Institutes of Health (NIH), the term "biomarker" is defined as "a characteristic that is objectively measured and evaluated as an indicator of normal biologic or pathogenic processes or pharmacological response to a therapeutic intervention."

[0004] The development of improved diagnostics based on the discovery of biomarkers has the potential to accelerate new drug development by identifying, in advance, those subjects most likely to show a clinical response to a given drug. This would significantly reduce the size, length and cost of clinical trials. Technologies such as genomics, proteomics and molecular imaging currently enable rapid, sensitive and reliable detection of specific gene mutations, expression levels of particular genes, and other molecular biomarkers. In spite of the availability of various technologies for molecular characterization of tumors, the clinical utilization of cancer biomarkers remains largely unrealized because few cancer biomarkers have been discovered. For example, a recent review article states: There is a critical need for expedited development of biomarkers and their use to improve diagnosis and treatment of cancer. (Cho, 2007, Molecular

Cancer 6:25).

[0005] Another recent review article on cancer biomarkers contains the following comments: The challenge is discovering cancer biomarkers. Although there have been clinical successes in targeting molecularly defined subsets of several tumor types - such as chronic myeloid leukemia, gastrointestinal stromal tumor, lung cancer and glioblastoma multiforme - using molecularly targeted agents, the ability to apply such successes in a broader context is severely limited by the lack of an efficient strategy to evaluate targeted agents in subjects. The problem mainly lies in the inability to select subjects with molecularly defined cancers for clinical trials to evaluate these exciting new drugs. The solution requires biomarkers that reliably identify those subjects who are most likely to benefit from a particular agent. (Sawyers, 2008, Nature 452:548-552, at 548). Comments such as the foregoing illustrate the recognition of a need for the discovery of clinically useful biomarkers and diagnostic methods based on such biomarkers.

[0006] There are three distinct types of cancer biomarkers: (1) prognostic biomarkers, (2) predictive biomarkers, and (3) pharmacodynamic (PD) biomarkers. A prognostic biomarker is used to classify a cancer, e.g. , a solid tumor, according to aggressiveness, i.e. , rate of growth and/or metastasis, and refractiveness to treatment. This is sometimes called distinguishing "good outcome" tumors from "poor outcome" tumors. A predictive biomarker is used to assess the probability that a particular subject will benefit from treatment with a particular drug. For example, subjects with breast cancer in which the ERBB2 (HER2 or NEU) gene is amplified are likely to benefit from treatment with trastuzumab (HERCEPTIN®), whereas subjects without ERBB2 gene amplification are unlikely to benefit from treatment with trastuzumab. A PD biomarker is an indication of the effect(s) of a drug on a subject while the subject is taking the drug. Accordingly, PD biomarkers often are used to guide dosage level and dosing frequency, during the early stages of clinical development of a new drug. For a discussion of cancer biomarkers, see, e.g. , Sawyers, 2008, Nature 452:548-552.

[0007] Neuropilin-1 and neuropilin-2 are transmembrane glycoproteins of up to 923 and 926 amino acids respectively, sharing a similar domain structure and amino acid homology of 44%. NRP-1 is known to bind to several members of the VEGF family, as well as class 3 semaphorins, an unrelated class of ligands with biological functions distinct from those of the VEGFs. NRP-1 is found in several isoforms, as a result of alternative splicing. Because NRP- 1 is highly expressed in diverse tumor cell lines and human neoplasms, it has been suggested as a potential therapeutic target in cancer, e.g. , through the use of blocking NRP1 antibodies, or small molecule antagonists of the ligand interaction with the NRP1 extracellular domain. See, Zachary, 2014, Chem Immunol Allergy 99:37-70.

[0008] Tivozanib (also known as AV-951) is a potent and selective small-molecule inhibitor of VEGF receptors 1, 2 and 3. Tivozanib exhibits picomolar inhibitory activity against all three receptors, and it exhibits antitumor activity in preclinical models (Nakamura et al , 2006, Cancer Res. 66:9134-9142). In a global, randomized Phase 3 superiority clinical trial evaluating the efficacy and safety of tivozanib compared to sorafenib in 517 subjects with advanced renal cell carcinoma ("TIVO-1 "), tivozanib yielded positive results (Motzer et al , 2012 ASCO Annual Meeting, Abstract No. 4501).

[0009] Bevacizumab (Avastin®) is an antibody that specifically targets VEGF A, but not other members of the VEGF family of growth factors (Los et al, 2007, The Oncologist,

12:443-450). Tumor expression of neuropilin-1 ("NRP1 ") has been proposed as a biomarker for identification of subjects responsive to treatment with bevacizumab (Avastin®). (US Patent Publication No. 2013/0183304). However, such biomarker assays require obtaining a sample of the tumor tissue itself, for example, via biopsy. [0010] Despite a large amount of pre-clinical and clinical research focused on VEGF- targeted therapy, the mechanisms responsible for the anti-tumor activity of anti-VEGF agents are not fully understood. As with other types of targeted therapy, some, but not all, subjects benefit from tivozanib therapy. The complexity of VEGF biology makes the effectiveness of tivozanib against any given tumor unpredictable. Therefore, there is a need for diagnostic methods based on prognostic and predictive biomarkers that can be used to identify subjects with tumors that are likely (or unlikely) to respond to treatment with tivozanib.

SUMMARY OF THE INVENTION

[0011] The invention is based in part on the identification of lower circulating neuropilin-1 levels in a human subject's serum as a prognostic and/or predictive biomarker useful for identifying whether that human subject is likely to be highly responsive or less responsive to treatment with a VEGF inhibitor. Accordingly, in one embodiment, the invention provides a method of identifying a human subject likely to be responsive to treatment with a VEGF inhibitor. The method includes comprise measuring the serum level of neuropilin-1 in the subject and determining whether said serum level of neuropilin-1 is below a defined threshold indicative that the subject is likely to be responsive to treatment with an inhibitor of VEGF - mediated angiogenesis. If the serum level of neuropilin-1 is below the defined threshold, it is an indication that the human subject is likely to be responsive to treatment with an inhibitor of VEGF-mediated angiogenesis, while a serum level of neuropilin-1 above the defined threshold indicates that the subject is likely to be resistant to treatment with the inhibitor of VEGF - mediated angiogenesis. In one embodiment, the VEGF inhibitor is a VEGFR2 inhibitor or the VEGF-mediated angiogenesis if VEGFR2-mediated angiogenesis.

[0012] In one embodiment, the invention provides a method of treating a human subject with a VEGF-associated condition in need thereof. The method includes the step of administering an effective amount of a VEGF inhibitor to the human subject that has a serum level of neuropilin-1 below a defined threshold thereby to reduce VEGF-mediated angiogenesis associated with the condition in the subject. In one embodiment, the VEGF-mediated angiogenesis reduced is VEGFR2-mediated angiogenesis. In one embodiment, the condition is colorectal cancer and the threshold score is, for example, 300 pg/mL or about 300 pg/ml (298.5 pg/mL). In another embodiment, the condition is renal cell carcinoma and the threshold score is, for example, 185 pg/mL or about 185 pg/ml (185.5 pg/mL).

[0013] In certain embodiments, the invention involves determining whether a human subject with cancer or a tumor has a cancer or tumor type that is likely to respond to treatment with tivozanib and/or whether a particular subject is likely to benefit from treatment with tivozanib. In particular, the inventors have surprisingly found that, while subjects with high serum levels of neuropilin-1 may benefit from treatment with tivozanib, subjects with low serum levels of neuropilin-1 are more likely to benefit from such treatment and may show superior response to treatment with tivozanib than those subjects with high levels of neuropilin- 1.

[0014] Accordingly, the invention provides a method of identifying a human cancer in a subject as responsive to treatment with tivozanib. The method includes the steps of measuring a serum level of neuropilin-1 in the subject and determining whether the serum level of neuropilin-1 is below a defined threshold indicative that the cancer is likely to be responsive to tivozanib. According to this method, when the serum level of neuropilin-1 is below the defined threshold, it is indicative that the cancer is likely to be responsive to tivozanib, whereas when the serum level of neuropilin-1 is above the defined threshold, it is indicative that the cancer is likely to be resistant to tivozanib. According to one embodiment, the method also includes a threshold determination analysis to generate a defined threshold. In a further embodiment, the method includes a step of measuring the serum level of one or more additional biomarkers selected from Leptin, IGFBP2, Kallikrein5, endoglin, PSAT, IGFBP1 , IP-10 and MIG/CXCL9. In yet a further embodiment, the method further comprises the step of performing a threshold determination analysis for neuropilin-1 and the one or more additional biomarkers, thereby generating a defined threshold neuropilin-1 score and a defined threshold score for the one or more additional biomarkers. In one embodiment, the human cancer is colorectal cancer while in another embodiment, the cancer is renal cell carcinoma.

[0015] In another embodiment, the invention provides a method of determining whether to treat a cancer in a human subject with tivozanib rather than with bevacizumab (Avastin®). The method involves measuring a serum level of neuropilin-1 in the subject and determining whether the serum level of neuropilin-1 is below a defined threshold indicative that the subject is more likely to benefit from treatment with tivozanib than with bevacizumb. In one embodiment, the cancer is colorectal cancer, while in another embodiment, the cancer is renal cell carcinoma.

[0016] In one embodiment, the invention provides a method for treating a cancer in human subject in need thereof. The method includes the step of administering an effective amount of tivozanib to the human subject, where the subject has been determined to have a serum level of neropilin-1 below a defined threshold thereby to treat the cancer. According to this method, a serum level of neuropilin-1 below the defined threshold indicates that the cancer is likely to be responsive to tivozanib, whereas a serum level of neuropilin-1 above the defined threshold indicates that the cancer or tumor is likely to be resistant to tivozanib. In a further embodiment, the cancer is colorectal cancer or renal cell carcinoma. In yet another embodiment, the serum level of neuropilin-1 below the defined threshold indicates that the cancer or the tumor is more likely to be responsive to treatment with tivozanib than with bevacizumab. [0017] In certain embodiments, the invention provides a method of assessing the likelihood that a subject with cancer is likely to respond to treatment with tivozanib. In certain embodiments, the subject may have renal cell carcinoma, or colorectal cancer. In a certain embodiment, the invention provides a method of assessing the likelihood that a subject with cancer is likely to respond more favorably to treatment with tivozanib than to treatment with bevacizumab.

[0018] In certain embodiments, the invention provides a method of identifying colorectal cancer as likely to be sensitive to treatment with tivozanib. In another embodiment, the invention provides a method of identifying colorectal cancer as likely to respond more favorably to tivozanib than it is likely to respond to treatment with bevacizumab .

[0019] In certain embodiments, the invention provides a method of assessing the likelihood that a subject with colorectal cancer is likely to respond to treatment with tivozanib. In a certain embodiment, the invention provides a method of assessing the likelihood that a subject with colorectal cancer is likely to respond more favorably to treatment with tivozanib than to treatment with bevacizumab. In certain embodiments, the invention provides a method for assessing a subject with colorectal cancer in order to decide whether to treat the subject with (1) bevacizumab + FOLFOX6; or (2) tivozanib + FOLFOX6.

[0020] In certain embodiments, the invention provides a method of assessing the likelihood that a subject with colorectal cancer is likely to respond to treatment with tivozanib. In a certain embodiment, the invention provides a method of assessing the likelihood that a subject with colorectal cancer is likely to respond more favorably to treatment with tivozanib than to treatment with bevacizumab. In certain embodiments, the invention provides a method for assessing a subject with colorectal cancer in order to decide whether to treat the subject with (1) bevacizumab + FOLFOX6; or (2) tivozanib + FOLFOX6. [0021] In one embodiment, the invention provides a method for determining whether to treat a human subject having a cancer with tivozanib in combination with FOLFOX6 rather than bevacizumab in combination with FOLFOX6. The method includes the steps of measuring a serum level of neuropilin-1 in the subject and determining whether said serum level of neuropilin-1 is below a defined threshold indicative that the cancer is more likely to be responsive to treatment with tivozanib in combination with FOLFOX6 than treatment with bevacizumab in combination with FOLFOX6. According to the method, when a serum level of neuropilin-1 is below the defined threshold, it is indicative that the cancer is more likely to be responsive to treatment with tivozanib in combination with FOLFOX6 than to treatment with bevacizumab in combination with FOLFOX6. In one embodiment, the cancer is colorectal cancer, whereas in another embodiment, the cancer is renal cell carcinoma.

[0022] In yet another embodiment, the invention provides a method of treating cancer in a human subject in need thereof. The method includes the step of administering an effective amount of a combination of tivozanib and FOLFOX6 to the human subject where the human subject has been determined to have a serum level of neuropilin-1 below a defined threshold. Further, in this method, when the human subject has a serum level of neuropilin-1 below the defined threshold, it is indicative that the cancer is more likely to be responsive to treatment with the combination of tivozanib and FOLFOX6 than with a combination of bevacizumab and FOLFOX6. In one embodiment, the cancer is colorectal cancer while in another embodiment, the cancer is renal cell carcinoma. [0023] In certain embodiments, the invention comprises measuring a subject's serum level of neuropilin-1, comparing the subject's serum level of NRP1 with a predetermined threshold serum level of NRP1; and, if the subject's serum level of NRP1 is below the predetermined threshold, treating the subject with the anti-cancer drug known as tivozanib. Accordingly, the invention provides a method for predicting quantitatively whether a human subject will be more likely to benefit from treatment with tivozanib, and may show superior response to treatment with tivozanib [i.e. , "NRP1 low" vs. "NRP1 high"]. Suitable methods of measurement include, for example, "capture-sandwich" assays like ELISA or a variation of it such as a Myriad RBM® assay. Other platforms for measurement of circulating levels of protein may be used, such as, for example, a Luminex® microbead/microsphere assay. [0024] In certain embodiments, measurements of circulating serum levels of NRP1 alone will be measured and used to determine whether to treat the subject with tivozanib. In other embodiments, the circulating levels of one or more additional serum factors related to angiogenesis may be measured in addition to NRP1, and a determination whether to treat the subject with tivozanib based thereon. Suitable factors for measurement of circulating serum levels in combination with NRP1 include one or more of the following proteins: Leptin,

IGFBP2, Kallikrein5, endoglin, PSAT, IGFBP1, IP-10 and/or MIG/CXCL9. Thus, in some embodiments, the method includes measuring the circulating serum levels of NR l and at least one other protein selected from the group consisting of Leptin, IGFBP2, Kallikrein5, endoglin, PS AT, IGFBPl, IP- 10 and MIG/CXCL9. In some embodiments, the method includes performing a threshold determination analysis, thereby generating a defined threshold. The threshold determination analysis can include a receiver operator characteristic curve analysis.

[0025] In another aspect, the invention provides a method of treating a subject having cancer. The method comprises: (a) determining whether the subject is likely to be responsive to tivozanib by measuring, in a sample from the serum of the subject, the relative circulating level of NR l; and (b) administering to the subject a therapeutically effective amount of tivozanib if the foregoing steps yield a result indicating that the cancer is likely to be responsive to tivozanib. In another aspect, the method comprises (a) measuring in a serum sample from subject, the neuropilin-1 and at least one additional protein selected from the group consisting of: Leptin or IGFBP2 or Kallikrein5 or endoglin or PS AT or IGFBPl or IP- 10 or MIG/CXCL9 and determining a relative NRPl score; and (b) administering to the subject a therapeutically effective amount of tivozanib if the foregoing steps yield a result indicating that the cancer is likely to be responsive to tivozanib.

BRIEF DESCRIPTION OF THE DRAWINGS

[0026] FIGURE 1 illustrates the difference in baseline pretreatment NRPl serum levels between Responders (Resp) and Non-Responders (Non-resp). See, Example 1 : Exploratory Serum Biomarker Analysis.

[0027] FIGURE 2 illustrates the difference in progression-free survival (PFS) observed between subjects with high levels of NRPl ("NRPl high") and subjects with low levels of NRPl ("NRPl low"). See, Example 2: Neuropilin-1 Serum Levels in Renal Cell Carcinoma.

[0028] FIGURE 3 illustrates the difference in overall survival (OS) observed between subjects with high levels of NRPl ("NRPl high") and subjects with low levels of NRPl ("NRPl low"). See Example 2: Neuropilin-1 Serum Levels in Renal Cell Carcinoma.

[0029] FIGURE 4 illustrates the median time to event (i.e. , progression-free survival time or "PFS") for subjects with colorectal cancer in the tivozanib arm of the study, broken down into subjects with low NRPl (i.e. , NRPl <298.5 pg/ml; n = 56)("NRP1 low") and high NRPl (i.e. , NRPl >298.5 pg/ml; n = 53) ("NRPl high"). Subjects in the NRPl low group exhibited statistically significant improvement in median PFS over those subjects in NRPl high group; 17.9 months, v. 7.3 months, respectively. See, Example 3: Neuropilin-1 Serum Levels in Colorectal Cancer: Treatment Of Subjects With Colorectal Cancer with Tivozanib and

FOLFOX6 vs. Bevacizumab and FOLFOX6.

[0030] FIGURE 5 illustrates the median time to event (i.e. , progression-free survival time or "PFS") for subjects with colorectal cancer in the bevacizumab (Avastin®) arm of the study, broken down into subjects with low NRPl (i.e. , NRPl <298.5 pg/mL; n = 56)("NRP1 low") and high NRPl (i.e. , NRPl >298.5 pg/mL; n = 53) ("NRPl high"). Subjects in the NRPl low group exhibited improvement in median PFS over those subjects in NRPl high group; 11.2 months v. 7.5 months, respectively. (Logrank test p value = 0.046). See, Example 3:

Neuropilin-1 Serum Levels in Colorectal Cancer: Treatment Of Subjects With Colorectal Cancer with Tivozanib and FOLFOX6 vs. Bevacizumab and FOLFOX6.

[0031] FIGURE 6 illustrates the median time to event (i.e. , overall survival time or "OS") for subjects with colorectal cancer in the tivozanib arm of the study, broken down into subjects with low NRPl (i.e. , NRPl < 298.5 pg/mL; n = 56)("NRP1 low") and high NRPl (i.e. , NRPl > 298.5 pg/mL; n = 53) ("NRPl high"). Subjects in the NRPl low group exhibited statistically significant improvement in median OS over those subjects in NRPl high group; >25 months vs. 16.2 months, respectively. (Logrank p value =1.72 e-4.) See, Example 3: Neuropilin-1 Serum Levels in Colorectal Cancer: Treatment Of Subjects With Colorectal Cancer with Tivozanib and FOLFOX6 vs. Bevacizumab and FOLFOX6.

[0032] FIGURE 7 illustrates the lack of a statistically significant difference in

progression-free survival (PFS) in subjects receiving tivozanib + FOLFOX6 (n=177) compared to NRPl high subjects receiving Avastin® + FOLFOX6 (n = 88). Median PFS for subjects in the tivozanib + FOLFOX6 arm was 9.76 months, while median PFS for subjects receiving Avastin® + FOLFOX6 was 9.53 months. (Logrank test p value p = 0.345). See, Example 3: Neuropilin-1 Serum Levels in Colorectal Cancer: Treatment Of Subjects With Colorectal Cancer with Tivozanib and FOLFOX6 vs. Bevacizumab and FOLFOX6. Accordingly, without testing for NRPl levels (NRPl low v. NRPl high) there is no significant difference detected for PFS between the two treatment arms. [0033] FIGURE 8 illustrates statistically significant increase in progression-free survival (PFS) in NRP1 low subjects receiving tivozanib + FOLFOX6 (n=53) compared to NRP1 low subjects receiving Avastin® + FOLFOX6 (n = 29). Median PFS for subjects in the tivozanib + FOLFOX6 arm was 17.9 months, while median PFS for subjects receiving Avastin® + FOLFOX6 was 11.2 months. (Logrank test p value p = 0.013). See, Example 3: Neuropilin-1 Serum Levels in Colorectal Cancer: Treatment Of Subjects With Colorectal Cancer with Tivozanib and FOLFOX6 vs. Bevacizumab and FOLFOX6.

[0034] FIGURE 9 illustrates the lack of a statistically significant increase in progression- free survival (PFS) in NRP1 high subjects receiving tivozanib + FOLFOX6 (n=56) compared to NRP1 high subjects receiving Avastin® + FOLFOX6 (n = 26). Median PFS for subjects in the tivozanib + FOLFOX6 arm was 7.33 months, while median PFS for subjects receiving Avastin® + FOLFOX6 was 7.46 months. (Logrank test p value p = 0.992). See, Example 3: Neuropilin-1 Serum Levels in Colorectal Cancer: Treatment Of Subjects With Colorectal Cancer with Tivozanib and FOLFOX6 vs. Bevacizumab and FOLFOX6. DETAILED DESCRIPTION OF THE INVENTION

[0035] The circulating serum level of neuropilin-1 in a subject can be used as a prognostic biomarker for classifying human tumors or cancers according to their likelihood of responding to treatment with the anti-tumor drug tivozanib, a VEGF tyrosine kinase inhibitor (TKI). Such classification of tumors is useful for identifying human subjects who are suitable candidates for treatment with tivozanib in a clinical setting. The methods of the present invention, in which quantifying serum levels of neuropilin-1 in a subject may surprisingly be used as a prognostic biomarker for treatment of such subject with angiogenesis inhibitors broadly, such as inhibitors of VEGF and inhibitors of VEGF -receptors. This is surprising in that previously only intratumoral analyses of NRP1 in solid tissue samples has been used as putative prognostic biomarkers for bevacizumab (Avastin®). Serum biomarker assays are advantageous over intratumoral tissue assays in that they are less invasive, drawing blood rather than requiring a tissue biopsy. Thus, when available, serum biomarkers are more easily tested multiple times, while the analysis of a biomarker found in or on the tumor itself is usually a one-time test result. Further, serum biomarker assays, when available, have the advantage of being more sensitive and accurate and have a wider dynamic range. [0036] Additionally, the methods described herein take advantage of the inventor's discovery that neuropilin-1 is surprisingly useful as a predictive biomarker in colorectal cancer as to whether the anti-tumor drug tivozanib is a better choice for treatment than bevacizumab (Avastin®). In a specific embodiment, serum expression levels of neuropilin-1 can be used to predict whether a particular subject is more likely to have a better response to treatment with tivozanib in combination with FOLFOX6, compared to treatment with bevacizumab (Avastin® in combination with FOLFOX6.

Definitions

[0037] As used herein, "AV-951" and "tivozanib" mean N-{2-chloro-4-[(6,7-dimethoxy-4- quinolyl)oxy]-phenyl}-N'-(5-methyl-3-isoxazolyl)urea, which has the following chemical structure, including salts and polymorphs thereof:

[0038] As used herein, "threshold neuropilin-1 score" means the median neuropilin-1 score for a particular cancer at which the classifier gives the most desirable balance between the cost of false negative calls and false positive calls.

[0039] As used herein, "median neuropilin-1 score" for a particular cancer means the numerical value calculated for a population of subjects, below which a subject is classified as being 'NRP1 LO' and therefore being likely to benefit from treatment with tivozanib.

Additionally, the cancer or tumor of an 'NRP1 LO' subject is more likely to be responsive to treatment with tivozanib. Neuropilin-1 scores higher than the threshold neuropilin-1 score level will be interpreted as indicating that the subject is 'NRP1 ΗΓ and therefore less likely to benefit from treatment with tivozanib. Additionally, the cancer or tumor of an 'NRP1 ΗΓ subject is less likely to be responsive (i.e. , may be resistant) to treatment with tivozanib.

[0040] As will be recognized by one skilled in the art, the median neuropilin-1 score will vary depending upon known parameters. For example, as described herein, in a clinical trial of subjects with advanced colorectal cancer (CRC), the inventors determined the median NRPl level to be 300 pg/mL, more specifically, 298.5 pg/mL. Accordingly, for purposes of the present invention, a serum concentration of 300 pg/mL (298.5 pg/ml) may be considered appropriate for use as a 'median neuropilin-1 score' for subjects with CRC. A subject having CRC with serum NRPl below 300 pg/mL (298.5 pg/ml) may therefore be considered to have low levels of NRPl (i.e. , 'NRPl LO'); a subject having CRC with serum NRPl equal to or above 300 pg/mL (298.5 pg/mL) may therefore be considered to have high levels of NRPl (i.e. , 'NRPl ΗΓ). In a clinical trial of subjects with renal cell carcinoma (RCC), the inventors determined the median NRPl level to be 185 pg/mL, more specifically 185.5 pg/mL.

Accordingly, for purposes of the present invention, a serum concentration of 185 pg/mL (185.5 pg/mL) may be considered appropriate for use as a 'median neuropilin-1 score' for subjects with RCC. A subject having RCC with serum NRPl below 185 pg/mL (185.5 pg/mL) may therefore be considered 'NRPl LO'; a subject having RCC with serum NRPl equal to or above 185 pg/mL (185.5 pg/mL) may therefore be considered 'NRPl HI. ' [0041] As used herein, "receiver operating characteristic" (ROC) curve means a graphical plot of false positive rate (sensitivity) versus true positive rate (specificity) for a binary classifier system. In construction of an ROC curve, the following definitions apply:

False negative rate: FNR = 1 - TPR

True positive rate: TPR = true positive / (true positive + false negative)

False positive rate: FPR = false positive / (false positive + true negative)

[0042] As used herein, "response" or "responding" or "responsive" to treatment means, with regard to a treated subject, that the treated subject exhibits: (a) an increase in the period of progression-free survival ("PFS"); or (b) an increase in the period of overall survival, compared to untreated subjects. In the case of a solid tumor, a "response" or "responding" or

"responsive" to treatment means that the tumor of a treated subject displays: (a) slowing of growth, (b) cessation of growth, or (c) regression (i.e., decrease in size).

[0043] As used herein, "threshold determination analysis" means analysis of a dataset representing a given cancer or tumor type, e.g., human renal cell carcinoma, to determine a threshold neuropilin-1 concentration, e.g. , an optimum threshold neuropilin-1 score, for that particular cancer or tumor type. In the context of a threshold determination analysis, the dataset representing a given cancer or tumor type includes (a) actual response data (response or non-response), and (b) a threshold determination for each cancer or tumor from a group having such cancer or tumor type. In the case of a preclinical trial, the group may be tumor-bearing mice or other non-human subject. In the case of clinical trials or therapeutic treatment, the group may be human subjects having such cancer type or tumor type.

[0044] Treatment in accordance with the methods and compositions described herein may improve or ameliorate one or more the characteristics or symptoms of a VEGF-associated condition. As used herein, "treat," "treating" and "treatment" mean the treatment of a disease in a subject, e.g. , in a human. This includes: (a) inhibiting the disease, i.e. , arresting its development; and (b) relieving the disease, i.e. , causing regression of the disease state.

[0045] As used herein, "VEGF-associated condition" means any disease or disorder associated with VEGF that can be treated with a VEGF inhibitor described herein. Exemplary VEGF-associated conditions include, but are not limited to, cancer, including colorectal cancer, kidney cancer, and breast cancer; rheumatoid arthritis; diabetic retinopathy; age-related macular degeneration (AMD); angiosarcoma; pulmonary emphysema; proteinuria; and preeclampsia.

Serum Data

[0046] Serum Samples. A serum tissue sample collected from a human subject prior to treatment may be used as a baseline, and may be analyzed using any suitable assay for the measurement of serum proteins, so that a baseline neuropilin-1 serum level can be determined. The tissue sample is optimally obtained from the subject and analyzed using conventional instruments and procedures. For example, NRP1 serum level may be determined using the Myriad RBM® platform (Myriad RBM, Inc., Salt Lake City, Utah). Other recognized medical procedures for obtaining and analyzing serum protein levels can be used by the skilled person to obtain and analyze serum NRP1 levels for practicing the methods described herein. The serum tissue sample should be large enough to provide sufficient protein for measuring individual biomarker serum levels. In some embodiments, the sample is in the form of circulation tumor cells (CTCs) in a blood sample. Serum protein levels can be analyzed using an NRP1 ELISA kit. Such kits are commercially available, for example, the Human Neuropilin-1 Quantikine ELISA Kit (R&D Systems, Minneapolis, MN); and Human Soluble Neuropilin-1 ELISA Kit (A viscera Science, Santa Clara, CA).

[0047] ELISA. Performing a neuropilin-1 protein ELISA, requires at least one antibody against an NRP1 protein, i.e. , the detection antibody. In an exemplary embodiment, subject protein from a sample to be analyzed is immobilized on a solid support such as a polystyrene microtiter plate. This immobilization can be by non-specific binding of the subject, e.g. , through adsorption to the surface. Alternatively, immobilization can be by specific binding, e.g. , through binding of NRP1 protein from the sample by a capture antibody (anti-NRPl antibody different from the detection antibody), in a "sandwich" ELISA. After the NRP1 is immobilized, the detection antibody is added, and the detection antibody forms a complex with the bound NRP1. The detection antibody is linked to an enzyme, either directly or indirectly, e.g. , through a secondary antibody that specifically recognizes the detection antibody.

Typically between each step, the plate, with bound NRP1, is washed with a mild detergent solution. Typical ELISA protocols also include one or more blocking steps, which involve use of a non-specifically-binding protein such as bovine serum albumin to block unwanted nonspecific binding of protein reagents to the plate. After a final wash step, the plate is developed by addition of an appropriate enzyme substrate, to produce a visible signal, which indicates the quantity of NRP1 in the sample. The substrate can be, e.g., a chromogenic substrate or a fluorogenic substrate. ELISA methods, reagents and equipment are well-known in the art and commercially available.

[0048] It is understood that the expression levels of other marker proteins, e.g. , Leptin or IGFBP2 or Kallikrein5 or endoglin or PSAT or IGFBP1 or IP-10 or MIG/CXCL9, as well as other marker proteins can be measured, for example, by ELISA using detecting antibodies specific for each additional marker protein. In certain methods of the present invention, these marker proteins may be used in conjunction with NRP1.

[0049] Other biomarkers, e.g. , Leptin or IGFBP2 or Kallikrein5 or endoglin or PSAT or IGFBP1 or IP-10 or MIG/CXCL9, as well as other serum-based protein biomarkers can be measured, for example, median biomarker serum protein levels can be determined and biomarker serum protein levels can be measured and compared to the median. In the case of some cancers or tumors, if desired, protein biomarkers may be measured, other than by measuring serum protein levels, using known methods of measuring protein levels, such as immunohistochemistry (IHC). Additionally, one skilled in the art may identify non-protein biomarkers, such as DNA expression levels, RNA levels, etc. to be used in methods of the present invention. The selection of additional biomarkers, if any, to be used in combination with serum levels of NRPl, and the method of measuring levels of such biomarkers, are design choices, and can be accomplished by a person of skill in the art, without undue

experimentation.

Data Interpretation

[0050] Neuropilin-1 score levels preferably are interpreted with respect to a threshold neuropilin-1 score. A neuropilin-1 score lower than the threshold neuropilin-1 score will be interpreted as indicating that a cancer or tumor is likely to be responsive (sensitive) to tivozanib, and hence, a subject with such cancer or tumor may benefit from treatment with tivozanib. Neuropilin-1 scores higher than the threshold neuropilin-1 score level will be interpreted as indicating that a cancer or tumor is likely to be non-responsive (resistant) to tivozanib, and hence, a subject with such cancer or tumor is not as likely to benefit from treatment with tivozanib.

[0051] When discussing or interpreting neuropilin-1 score data, the skilled person will be aware that the serum level "score" and serum level for a particular biomarker may be directly related, or inversely related, depending on the biomarker, the assay method and units of measurement employed. It is contemplated that a given threshold neuropilin-1 score will vary depending on cancer or tumor type. In the context of the present invention, the term "cancer type" or "tumor type" takes into account (a) species (mouse or human); and (b) organ or tissue of origin. Optionally, cancer type or tumor type further takes into account cancer

categorization based on gene expression characteristics, e.g. , HER2-positive breast tumors, or non-small cell lung cancers expressing a particular EGFR mutation. [0052] For any given cancer type or tumor type, an optimum neuropilin-1 score can be determined (or at least approximated) empirically by performing a threshold determination analysis. Preferably, threshold determination analysis includes receiver operator characteristic (ROC) curve analysis.

[0053] ROC curve analysis is an established statistical technique, the application of which is within ordinary skill in the art. For a discussion of ROC curve analysis, see generally Zweig et at , 1993, "Receiver operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine," Clin. Chem. 39:561-577; and Pepe, 2003, The statistical evaluation of medical tests for classification and prediction, Oxford Press, New York.

[0054] Neuropilin-1 scores and the optimum neuropilin-1 score may vary from cancer type to cancer type (or tumor type to tumor type). Therefore, a threshold determination analysis preferably is performed on one or more datasets representing any given cancer type or tumor type to be tested using the present invention. The dataset used for threshold determination analysis includes: (a) actual response data (response or non-response), and (b) a neuropilin-1 score for each serum sample from a group of human cancers or tumors or mouse cancers or tumors. Once a neuropilin-1 score threshold is determined with respect to a given cancer type or tumor type, that threshold can be applied to interpret neuropilin-1 scores from cancers of that cancer type, or tumors of that tumor type.

[0055] The ROC curve analysis is performed essentially as follows. A sample with a neuropilin-1 score below threshold is classified as a responder, and a sample with a neuropilin- 1 score above the threshold is classified as a non-responder. Whether the neuropilin-1 score is higher or lower will depend on the type of assay and the units of measure employed. For every neuropilin-1 score from a tested set of samples, "responders" and "non-responders"

(hypothetical calls) are classified using that neuropilin-1 score as the threshold. This process enables calculation of TPR (y vector) and FPR (x vector) for each potential threshold, through comparison of hypothetical calls against the actual response data for the data set. Then an ROC curve is constructed by making a dot plot, using the TPR vector, and FPR vector. If the ROC curve is above the diagonal from (0, 0) point to (1.0, 0.5) point, it shows that the neuropilin-1 score test result is a better test than random.

[0056] The ROC curve can be used to identify the best operating point. The best operating point is the one that yields the best balance between the cost of false positives weighed against the cost of false negatives. These costs need not be equal. The average expected cost (C) of classification at point x,y in the ROC space is denoted by the expression

C = (1-p) alpha*x + p*beta(l-y) wherein: alpha = cost of a false positive, beta = cost of missing a positive (false negative), and p = proportion of positive cases.

[0057] False positives and false negatives can be weighted differently by assigning different values for alpha and beta. For example, if it is decided to include more subjects in the responder group at the cost of treating more subjects who are non-responders, one can put more weight on alpha. In this case, it is assumed that the cost of false positive and false negative is the same (alpha equals to beta). Therefore, the average expected cost of classification at point x,y in the ROC space is: C' = (l-p)*x + p*(l-y).

The smallest C can be calculated after using all pairs of false positive and false negative (x, y). The optimum neuropilin-1 score threshold is calculated as the score of the (x, y) at C.

[0058] In addition to predicting whether a subject with a cancer or tumor is likely to benefit from treatment with tivozanib, a neuropilin-1 score provides an approximate, but useful, indication of how likely a cancer or tumor is to be responsive or non-responsive. In general, the lower the neuropilin-1 score, the more likely a cancer or tumor is to be responsive to tivozanib, and the higher the lower the neuropilin-1 score, the more likely a cancer or tumor is to be resistant to tivozanib.

[0059] In some embodiments, serum samples may include measurement of levels of one or more proteins used as internal standards. The choice of proteins to be used as internal standards can be affected by the type of tissue sample to be tested. For any given type of tissue sample, there is wide latitude in the number and identity of internal standards. Choice of internal standards is within skill in the art.

[0060] In some embodiments, the set of materials for analysis of serum samples, e.g., ELISA tests, form part of a diagnostic test kit that contains other components such as buffers, reagents and detailed instructions for using the set, e.g., in the methods described herein. Such a diagnostic test kit can enhance the convenience and reproducibility in the performance of the methods described herein. Depending on the particular platform technology used, the kit may take advantage of availability of platform technologies that are well developed and currently available from commercial vendors, such as the Myriad RBM® platform described above.

[0061] In an exemplary ELISA-based embodiment, a basic diagnostic test kit includes an ELISA assay for neuropilin-1. In some embodiments, the kit includes antibodies for ELISA assays for one or more additional proteins, such as Leptin, IGFBP2, Kallikrein5, endoglin, PS AT, IGFBP1, IP- 10 and/or MIG/CXCL9 for development of the score used in the methods described herein. In some embodiments, a more elaborate test kit contains not only antibodies for ELISA assays, but also buffers, reagents and detailed instructions for measuring the serum levels of the members of a neuropilin-1 score, using ELISA-based technology. In some embodiments, the kit includes a test protocol and all the consumables, i.e., reagents and single- use components (except test samples), needed to perform the methods disclosed herein.

Treatment of Subjects

[0062] The invention includes a method of treating a human subject with a VEGF- associated condition in need thereof. The method includes administering an effective amount of a VEGF inhibitor to the human subject that has a serum level of neuropilin-1 below a defined threshold thereby to reduce VEGF-mediated angiogenesis associated with the condition in the subject. In one embodiment, the VEGF-mediated angiogenesis reduced is VEGFR2- mediated angiogenesis.

[0063] Generally, a therapeutically effective amount of active component, e.g. , a VEGF- inhibitor, is in the range of 0.1 mg/kg to 100 mg/kg, e.g., 1 mg/kg to 100 mg/kg, 1 mg/kg to 10 mg/kg. The amount administered will depend on variables such as the type and extent of disease or indication to be treated, the overall health of the patient, the in vivo potency of the active component, the pharmaceutical formulation, and the route of administration. The initial dosage can be increased beyond the upper level in order to rapidly achieve the desired blood- level or tissue-level. Alternatively, the initial dosage can be smaller than the optimum, and the daily dosage may be progressively increased during the course of treatment. Human dosage can be optimized, e.g., in a conventional Phase I dose escalation study designed to run from 0.5 mg/kg to 20 mg/kg. Dosing frequency can vary, depending on factors such as route of administration, dosage amount, serum half-life of the antibody, and the disease being treated. Exemplary dosing frequencies are once per day, once per week and once every two weeks. [0064] The optimal effective amount of the compositions can be determined empirically and will depend on the type and severity of the disease, route of administration, disease progression and health, mass and body area of the subject. Such determinations are within the skill of one in the art. Examples of dosages of VEGF inhibitors which can be used for methods described herein include, but are not limited to, an effective amount within the dosage range of any of about 0.01 μg/kg to about 300 mg/kg, or within about 0.1 μg/kg to about 40 mg/kg, or with about 1 μg/kg to about 20 mg/kg, or within about 1 μg/kg to about 10 mg/kg. For example, when administered subcutaneously, the composition may be administered at low microgram ranges, including for example about 0.1 μg/kg or less, about 0.05 μg/kg or less, or 0.01 μg/kg or less.

[0065] The invention includes a method of treating a cancer subject, comprising:

(a) determining whether the subject is likely to be responsive to tivozanib by measuring, in a serum sample from the subject, the serum protein level of neuropilin-1 and optionally one or more of the following genes: Leptin, IGFBP2, Kallikrein5, endoglin, PSAT, IGFBP1, IP-10 and/or MIG/CXCL9; and

(b) administering to the subject a therapeutically effective amount of tivozanib if the foregoing steps yield a result indicating that the cancer is likely to be responsive to tivozanib.

[0066] Information on tivozanib dosage level, dosing schedule, and potential side effects is known in the art. See, e.g. , Nosov et al , 2012, J. Clin. Oncology 30: 1678-1685, the full disclosure of which is hereby incorporated by reference herein.

[0067] Tivozanib may be administered in combination or in sequence with any approved or known therapy regimen. For example, subjects with colorectal cancer are frequently treated with a combination of chemotherapy agents known as "FOLFOX." FOLFOX comprises a combination of folinic acid (or leucovorin) ("FOL"), fluorouracil (or 5-FU) ("F") and oxaliplatin (or Eloxatin® Sanofi S.A.) ("OX"). FOLFOX6 constitutes the sixth variation of the FOLFOX treatment regimen. Other treatment regimens approved for colorectal cancer include "FOLFIRI," which is a combination of folinic acid, fluorouracil and irinotecan hydrochloride (or Camptosar® Pfizer, Inc.) ("IRI"); and "XELOX," which is a combination of capecitabine (or Xeloda® Genentech, Inc.) ("XEL") and oxaliplatin. [0068] In a modified FOLFOX6 ("mFOLFOX6") regimen, chemotherapy is administered in conjunction with tivozanib according to a 28 day cycle. Tivozanib is administered orally at a dose of 1.5 mg daily beginning on Day 1 of each Cycle for 21 days, followed by 7 days off treatment, subject to dose reduction to 1.0 mg daily for Grade 3 drug-related adverse events. Hypertension events must be treated with anti-hypertensive drugs prior to dose reduction. In the mFOLFOX regimen, oxaliplatin is administered at 85 mg/m² via intravenous bolus in 500 mL of D5W over 2 hours on days 1 and 15 of the 28 day cycle. Folinic acid is administered at 400 mg/m² via intravenous bolus in 500 mL of D5W over 2 hours on days 1 and 15 of the 28 day cycle, which may be concurrently with oxaliplatin via a separate IV line. Fluorouracil bolus is administered at 400 mg/m² via intravenous over 5-15 minutes, or infused per institutional guidelines on days 1 and 15. Fluorouracil is additionally administered at 2400 mg/m² via infusion pump over 46-48 hours or infusion per institutional guidelines on days 1-3 and 15-17 of the 28 day cycle.

EXAMPLES [0069] The invention is further illustrated by the following examples. The examples are provided for illustrative purposes only, and are not to be construed as limiting the scope or content of the invention in any way.

Example 1: Exploratory Serum Biomarker Analysis

[0070] In order to identify serum factors that could serve as a prognostic biomarker for a subject's treatment with inhibitors of the VEGF pathway, pre-treatment baseline serum levels of 99 serum proteins were measured in serum samples from 50 subjects (25 'Responders' and 25 'Non-responders' ) subsequently treated with tivozanib and analyzed for correlation with a positive response to treatment with tivozanib. For the purpose of the analysis, the 'Responders' were subjects who exhibited both: (1) maximum percent tumor reduction (MPR), which was a reduction in tumor size from between -40.6 % to -82.3 %; and (2) progression-free survival (PFS), measured by independent assessment, from 156 to 838 days. 'Non-responders' were subjects who exhibited both: (1) MPR varying from a reduction in tumor size of -7.8% to an increase in tumor size of 174.2%; and (2) PFS, measured by independent assessment, from 49 to 144 days. [0071] All 50 serum samples were profiled on Rules Based Medicine (RBM) Human Oncology Map 1.0 platform, which measures the serum levels of 99 proteins. Table 1 shows the results of this analysis for 10 serum factors that were found to have the most significant differences in Responders v. Non-Responders.

TABLE 1

[0072] In this experiment, NRPl was the serum factor which was most significantly different between Responders and Non-Responders with a p-value < 0.01 (p= 0.00329). The difference in NRPl serum levels between Responders and Non-Responders is shown in Figure 1. Several other serum factors exhibited differences between Responders and Non-Responders with a p-value 0.05, notably Leptin (p=0.0130) and Insulin-like Growth Factor Binding Protein-2 (IGFBP2) (p = 0.0254).

[0073] Because of the lack of a control in this analysis, the role of NRPl as a potential prognostic or response indicator remained unclear, prompting further analysis of NRPl serum levels.

Example 2: Neuropilin-1 Serum Levels in Renal Cell Carcinoma

[0074] TIVO-1 was a global, randomized Phase 3 superiority clinical trial evaluating the efficacy and safety of investigational drug tivozanib compared to sorafenib in 517 subjects with advanced RCC (Motzer et al, 2012 ASCO Annual Meeting, Abstract No. 4501). All subjects in TIVO-1 had clear cell RCC, had undergone a prior nephrectomy, and had not been treated previously with either a VEGF or mTOR therapy. TIVO-1 demonstrated a significant improvement in progression-free survival (PFS) in subjects receiving tivozanib vs. sorafenib. [0075] From the 517 subjects in the TIVO-1 trial, pretreatment serum samples from 240 subjects who were treated with tivozanib were analyzed retrospectively by RBM. Progression- free survival (PFS) and overall survival (OS) probabilities were analyzed for 120 subjects with high levels of NRP1 {i.e. , NRP1 >185.5 pg/ml) ("NRP1 high") vs. 120 subjects with low levels of NRP1 {i.e., NRP1 <185.5 pg/ml) ("NRP1 low"). [0076] Median PFS in the NRP1 low group of subjects was 14.8 months; while median PFS for the NRP1 high group of subjects was 8.5 months. The Logrank test p value was p = 0.016. (See Figure 2). Median OS in the NRP1 low group of subjects was not reached at 30 months; median OS in the NRP1 high group of subjects was 21.3 months. The Logrank test p value was p = 0.000107. (See Figure 3). Example 3: Neuropilin-1 Serum Levels in Colorectal Cancer: Treatment Of Subjects

With Colorectal Cancer with Tivozanib and FOLFOX6 vs. Bevacizumab and FOLFOX6.

[0077] In another clinical trial, 265 subjects were randomly assigned (2: 1) to receive treatment with either tivozanib and a modified FOLFOX6 regimen (n= 177) or bevacizumab and FOLFOX6 (n= 88). In the modified FOLFOX6 ("mFOLFOX6") regimen, chemotherapy is administered in conjunction with tivozanib according to a 28 day cycle. Tivozanib is administered orally at a dose of 1.5 mg daily beginning on Day 1 of each Cycle for 21 days, followed by 7 days off treatment, subject to dose reduction to 1.0 mg daily for Grade 3 drug- related adverse events. Hypertension events must be treated with anti-hypertensive drugs prior to dose reduction. In the mFOLFOX regimen, oxaliplatin is administered at 85 mg/m² via intravenous bolus in 500 mL of D5W over 2 hours on days 1 and 15 of the 28 day cycle.

Folinic acid is administered at 400 mg/m² via intravenous bolus in 500 mL of D5W over 2 hours on days 1 and 15 of the 28 day cycle, which may be concurrently with oxaliplatin via a separate IV line. Fluorouracil bolus is administered at 400 mg/m² via intravenous over 5-15 minutes, or infused per institutional guidelines on days 1 and 15. Fluorouracil is additionally administered at 2400 mg/m² via infusion pump over 46-48 hours or infusion per institutional guidelines on days 1-3 and 15-17 of the 28 day cycle.

[0078] In the bevacizumab arm, bevacizumab is administered as an intravenous infusion at a dose of 5 mg/kg every 2 weeks beginning on day 1 of cycle 1. Subjects receive 2 treatments of bevacizumab in each treatment cycle, on day 1 and day 15. Bevacizumab will be administered prior to the start of the mFOLFOX6 chemotherapy regimen or per institutional guidelines.

[0079] In a prospectively defined retrospective analysis of progression-free survival (PFS) and overall survival (OS) subjects in the tivozanib arm of the study, those subjects with low NRPl (i.e. , NRPl < 298.5 pg/ml; n = 56)("NRP1 low") exhibited statistically significant differences in survival over those subjects with high NRPl (i.e. , NRPl > 298.5 pg/ml; n = 53) ("NRPl high") in both PFS (17.9 months v. 7.3 months (Logrank test p value= 3.56 e-6) [See Figure 4] and in overall survival (>25 months v. 16.2 months (Logrank test p value = 1.72 e-4) [See Figure 6]. In the bevacizumab arm, those subjects with low NRPl (i.e., NRPl < 298.5 pg/ml; n = 26)("NRP1 low") exhibited a statistically significant difference in progression-free survival (PFS) over those subjects with high NRPl (i.e., NRPl > 298.5 pg/ml; n = 29) ("NRPl high") (11.2 months v. 7.5 months) (Logrank test p value = 0.046) [See Figure 5]. The above results demonstrate that serum levels of NRPl can serve as a prognostic biomarker for responsiveness of a subject to both VEGF pathway inhibitors, tivozanib and bevacizumab (Avastin®).

[0080] Additional retrospective analysis of subjects with low NRPl levels showed statistically significant increased PFS in NRPl low subjects receiving tivozanib + FOLFOX6 (n=53) compared to NRPl low subjects receiving bevacizumab (Avastin®) + FOLFOX6 (n = 29). Median PFS for subjects in the tivozanib + FOLFOX6 arm was 17.9 months, while median PFS for subjects receiving bevacizumab (Avastin®) + FOLFOX6 was 11.2 months. (Logrank test p value p = 0.013) [See Figure 8]. This result demonstrates that low levels of serum NRPl can serve as a predictive biomarker for higher efficacy of Tivozanib + FOLFOX6 compared to bevacizumab (Avastin®) + FOLFOX6 in terms of PFS response in subjects with CRC. [0081] Analysis of the undifferentiated subjects in the intent-to-treat population [i.e. , no differentiation between NR l low and NRPl high; as well as analysis of subjects with NRPl high levels showed no statistically significant difference in PFS between subjects receiving tivozanib + FOLFOX6 compared to NRPl low subjects receiving bevacizumab (Avastin®) + FOLFOX6. For all subjects, median PFS for subjects in the tivozanib + FOLFOX6 arm

(N=177) was 9.76 months, compared to median PFS for subject in the bevacizumab (Avastin®) + FOLFOX6 arm (N=88) of 9.53 months (Logrank test p value p = 0.345). [See Figure 7]. For the NRPl high population, median PFS for subjects in the tivozanib + FOLFOX6 arm was 7.33 months, while median PFS for subjects receiving bevacizumab (Avastin®) + FOLFOX6 was 7.46 months. (Logrank test p value p = 0.992) [See Figure 9]. Thus, low levels of NRPl (NRPl low) is a predictive biomarker for better efficacy of tivozanib over bevacizumab in subjects with low levels of NRPl.

INCORPORATION BY REFERENCE

[0082] The entire disclosure of each of the patent documents and scientific articles cited herein is incorporated by reference for all purposes.

EQUIVALENTS

[0083] The invention can be embodied in other specific forms with departing from the essential characteristics thereof. The foregoing embodiments therefore are to be considered illustrative rather than limiting on the invention described herein. The scope of the invention is indicated by the appended claims rather than by the foregoing description, and all changes that come within the meaning and range of equivalency of the claims are intended to be embraced therein.

Claims

What is claimed is: 1. A method of treating a human subject with a VEGF-associated condition in need thereof, the method comprising:

(a) administering an effective amount of a VEGF inhibitor to the human subject that has a serum level of neuropilin-1 below a defined threshold level thereby to reduce VEGF- mediated angiogenesis associated with the condition in the subject.

2. The method of claim 1, wherein the VEGF-mediated angiogenesis is VEGFR2- mediated angiogenesis.

3. The method of claim 1, wherein the condition is colorectal cancer.

4. The method of claim 3, wherein the threshold score is 300 pg/mL.

5. The method of claim 1 , wherein the condition is renal cell carcinoma.

6. The method of claim 5, wherein the threshold score is 185 pg/mL.

7. A method of treating a cancer in a human subject in need thereof, the method comprising:

(a) administering an effective amount of tivozanib to the human subject, wherein the subject has been determined to have a serum level of neuropilin-1 below a defined threshold thereby to treat the cancer,

wherein a serum level of neuropilin-1 below the defined threshold indicates that the cancer is likely to be responsive to tivozanib, and a serum level of neuropilin-1 above the defined threshold indicates that the cancer or tumor is likely to be resistant to tivozanib.

8. The method of claim 7, wherein the cancer is colorectal cancer.

9. The method of claim 7, wherein the cancer is renal cell carcinoma.

10. The method of any of claims 7-9, wherein a serum level of neuropilin-1 below the defined threshold indicates that the cancer or the tumor is more likely to be responsive to treatment with tivozanib than with bevacizumab.

11. A method of treating cancer in a human subject in need thereof, the method comprising: administering an effective amount of a combination of tivozanib and FOLFOX6 to the human subject, wherein the human subject has been determined to have a serum level of neuropilin-1 below a defined threshold,

wherein a serum level of neuropilin-1 below the defined threshold is indicative that the cancer is more likely to be responsive to treatment with the combination of tivozanib and FOLFOX6 than with a combination of bevacizumab and FOLFOX6.

12. The method of claim 11, wherein the cancer is colorectal cancer.

13. The method of claim 11, wherein the cancer is renal cell carcinoma.

14. A method of identifying a human subject likely to be responsive to treatment with a VEGF inhibitor, comprising:

(a) measuring a serum level of neuropilin-1 in the subject; and

(b) determining whether said serum level of neuropilin-1 is below a defined threshold indicative that the subject is likely to be responsive to treatment with an inhibitor of VEGF - mediated angiogenesis,

wherein a serum level of neuropilin-1 below the defined threshold is indicative that the subject is likely to be responsive to treatment with the inhibitor of VEGF-mediated

angiogenesis, and a serum level of neuropilin-1 above the defined threshold is indicative that the subject is likely to be resistant to treatment with the inhibitor of VEGF-mediated angiogenesis.

15. The method of claim 14, wherein the VEGF inhibitor is a VEGFR2 inhibitor.

16. A method of identifying a human cancer in a subject as responsive to treatment with tivozanib, comprising:

(a) measuring a serum level of neuropilin-1 in the subject; and

(b) determining whether said serum level of neuropilin-1 is below a defined threshold indicative that the cancer is likely to be responsive to tivozanib,

wherein a serum level of neuropilin-1 below the defined threshold is indicative that the cancer is likely to be responsive to tivozanib, and a serum level of neuropilin-1 above the defined threshold is indicative that the cancer is likely to be resistant to tivozanib.

17. The method of claim 16, further comprising performing a threshold determination analysis, thereby generating a defined threshold.

18. The method of claim 16, further comprising measuring a serum level of one or more additional biomarkers, wherein the one or more additional biomarkers are selected from the group consisting Leptin, IGFBP2, Kallikrein5, endoglin, PSAT, IGFBP1, IP-10 and MIG/CXCL9.

19. The method of claim 18, further comprising performing a threshold determination analysis for neuropilin-1 and the one or more additional biomarkers, thereby generating a defined threshold neuropilin-1 score and a defined threshold score for the one or more additional biomarkers.

20. The method of claim 16, wherein the cancer is colorectal cancer.

21. The method of claim 16, wherein the cancer is renal cell carcinoma.

22. A method of determining whether to treat a cancer in a human subject with tivozanib rather than bevacizumab, the method comprising:

(a) measuring a serum level of neuropilin-1 in the subject; and

(b) determining whether said serum level of neuropilin-1 is below a defined threshold indicative that the subject is more likely to benefit from treatment with tivozanib than with bevacizumab.

23. The method of claim 22, wherein the cancer is colorectal cancer.

24. The method of claim 22, wherein the cancer is renal cell carcinoma.

25. A method of determining whether to treat a human subject having a cancer with tivozanib in combination with FOLFOX6 rather than bevacizumab in combination with FOLFOX6, the method comprising:

(a) measuring a serum level of neuropilin-1 in the subject; and (b) determining whether said serum level of neuropilin-1 is below a defined threshold indicative that the cancer is more likely to be responsive to treatment with tivozanib in combination with FOLFOX6 than treatment with bevacizumab in combination with

FOLFOX6;

wherein a serum level of neuropilin-1 below the defined threshold is indicative that the cancer is more likely to be responsive to treatment with tivozanib in combination with FOLFOX6 than to treatment with bevacizumab in combination with FOLFOX6.

26. The method of claim 25, wherein the cancer is colorectal cancer.

27. The method of claim 25, wherein the cancer is renal cell carcinoma.