METHODS USING AXL AS A BIOMARKER OF EPITHELIAL-TO-MESENCHYMAL TRANSITION

Abstract

The present invention relates to the use of Axl as a biomarker for detecting the occurrence of epithelial-to-mesenchymal transition (EMT) in a subject. More specifically, the invention relates to various methods for detecting the occurrence of epithelial-to-mesenchymal transition (EMT) in a subject by measuring Axl expression and/or activity.

Claims

1.-67. (canceled)

68. A method of diagnosing and treating a disease characterized by proliferative activity, comprising: (a) obtaining a sample from a subject; (b) detecting a level of epithelial-to-mesenchymal transition (EMT) in the sample from the subject; (c) diagnosing the subject with increased risk of a disease characterized by proliferative activity when the level of EMT is increased in the sample as compared to a reference level; and (d) administering an effective amount of an Axl inhibitor to the subject diagnosed with increased risk of a disease characterized by proliferative activity.

69. The method of claim 68 wherein the disease characterized by proliferative activity is metastatic cancer or late stage cancer.

70. The method of claim 69 wherein the metastatic cancer is breast cancer.

71. The method of claim 68 wherein the level of epithelial-to-mesenchymal transition (EMT) is indicated by expression of an Axl biomarker.

72. The method of claim 71 wherein the Axl biomarker comprises a protein.

73. The method of claim 71 wherein the Axl biomarker comprises an mRNA.

74. The method of claim 68 wherein the subject is human.

75. The method of claim 68 wherein the sample is blood, serum, plasma or tissue culture supernatant.

76. The method of claim 68 wherein the Axl inhibitor comprises a small molecule kinase inhibitor.

77. The method of claim 76 wherein the Axl inhibitor is R428.

78. A method of treating a disease characterized by proliferative activity, comprising administering an effective amount of an Axl inhibitor to the subject provided that a sample from the subject has an increased level of epithelial-to-mesenchymal transition (EMT) as compared to a reference level.

79. The method of claim 78 wherein the disease characterized by proliferative activity is metastatic cancer or late stage cancer.

80. The method of claim 79 wherein the metastatic cancer is breast cancer.

81. The method of claim 78 wherein an increased level of epithelial-to-mesenchymal transition (EMT) is indicated by an increased expression of an Axl biomarker.

82. The method of claim 81 wherein the Axl biomarker comprises a protein.

83. The method of claim 81 wherein the Axl biomarker comprises an mRNA.

84. The method of claim 78 wherein the subject is human.

85. The method of claim 78 wherein the sample is blood, serum, plasma or tissue culture supernatant.

86. The method of claim 78 wherein the Axl inhibitor comprises a small molecule kinase inhibitor.

87. The method of claim 86 wherein the Axl inhibitor is R428.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0267] The present invention is further illustrated by way of the following non-limiting examples, and with reference to the following Figures, wherein:

[0268] FIGS. 1(A-D) show that Axl expression is a negative prognostic factor for breast cancer survival. (FIG. 1A(i)) immunochemistry shows weak (60%) Axl expression and (FIG. 1A(ii)) immunohistochemistry shows strong Axl expression. (FIG. 1B) Kaplan-Meier analysis of 8 year clinical follow-up. (FIG. 1C) multivariate analysis. (FIG. 1D(i)-(iv)) shows Axl expression in matched pairs (n=16) of primary and metastatic human breast carcinoma: primary tumors to the left (upper. FIG. 1D(i). Axl negative, lower. FIG. 1D(iii). Axl positive), metastases to the right (upper. FIG. 1D(ii), liver, lower. FIG. 1D(iv), bone). Axl expression tended to be stronger in metastases when compared with corresponding primary tumors (p=0.11, McNemar's test).

[0269] FIGS. 2(A-G) show that Axl is required for breast cancer cell invasiveness. (FIG. 2A) shows FACS analysis and (FIG. 2B) shows Western blot analysis of Axl expression in epi-allelic MB-MDA-231 breast carcinoma cell series. FIGS. 2C(i) and (ii) are results of Western blot analysis that shows Axl is phosphorylated (FIG. 2C(i) lysate and FIG. 2C(ii) Immunoprecipitate (IP)). FIG. 2D and FIG. 2E show Epi-allelic analysis of matrigel invasion assay induced by serum as shown in FIG. 2D or SDF-1, as shown in FIG. 2E. FIG. 2F and FIG. 2G show 3D matrigel analysis of MB-MDA-231/shLuc (upper panels, FIG. 2F(i-iv) and FIG. 2G(i-iii) and MB-MDA-231/shAxl2 (lower panels, FIG. 2F(v-viii) and FIG. 2G(iv-vi)).

[0270] FIGS. 3(A-E) show Axl activity is upregulated by EMT inducers in breast epithelial cell. (FIG. 3A) Flow cytometry analysis of surface levels of Axl on an MCF10a cell line that stably expresses Twist. (FIG. 3B) Extracts from control (wt) and Twist-expressing MCF10a cells were analysed for changes in epithelial (E-cadherin, -catenin) and mesenchymal (N-cadherin) markers. *Conditioned medium was analysed by SDS-PAGE and immunoblotting using an antibody against Gas6. (FIGS. 3C(i) and (ii)) MCF10a cells transduced with retroviral vectors encoding Twist, Zeb2, Slug, Snail or vector control (GFP) were analysed on by flow cytometry for Axl surface expression (left, FIG. 3C(i)) and geometric mean fluorescence (right, FIG. 3C(ii)). (FIG. 3D) Extracts from MCF10a cells transduced with Twist, Zeb2, Slug or Snail retroviral vectors were analyzed by immunoblotting for changes in epithelial (E-cadherin, -catenin) and mesenchymal (Ncadherin, vimentin) markers. (FIG. 3E) Morphology of MCF10a cells transduced with Twist, Zeb2, Slug or Snail retroviral vectors at 72 hours post-seeding.

[0271] FIGS. 4(A and B) show that tissue engineered breast tumours require Axl expression. (FIG. 4A(i)-(v)) In vivo imaging of MDA-MB-231/GFP-Luc tumor tissue engineering implants in NODSCID mice. Temporal tumor growth was monitored by in vivo optical imaging of luciferase bioluminescence from MDA-MB-231/GFP-Luc cells (FIG. 4A(i)). Tumor cell number (total photon) and extent of radial infiltration (signal diameter) measurements are from control implants (solid line) and cell implants expressing MDA-MB-231/GFP-Luc-shAxl2 (dashed line) show Axl-dependence of tumor growth and colonization within poly-lactic acid tissue engineering scaffolds (FIG. 4A(ii)). Appearance of bilateral scaffolds upon excision (FIG. 4A(iii)). Immunohistochemistry analysis of tissue engineered tumors with anti-human Axl at 28 days post-implantation (left panel, vector control; right panel, shAxl2) (FIG. 4A(iv)). Tumor tissue engineering implants with wildtype control (FIG. 4A(iv), left panel) and shAxl2-expressing cells (FIG. 4A(iv), right panel) were analyzed by immunohistochemistry with anti-human Axl (FIG. 4A(v)). Black squares demarcate colonization and radial spread of MDA-MB-231/GFP Luc cells. *p<0.05, **p<0.005 (paired t-test) compared with control. N=6 mice/group. (FIG. 4B(i)-(v)) Temporal in vivo imaging of tricellular implants comprising primary human microvascular cells (EC), vascular smooth muscle cells (SMC) and MDA-MB-231/GFP Luc cells (FIG. 4B(i)). Tumor growth (total photon) and radial spread (diameter) measurements in control (solid line) and shAxl2 expressing MDA-MB-231 cell implants (broken line) analyzed in the presence of a tissue engineered vasculature (FIG. 4B(ii)). Excised tumors (28 days) are highly vascularized by engineered human microvessels (FIG. 4B(iii)). Immunohistochemistry analysis shows that the engineered anti-human CD31-staining vessels contain intra-lumenal red blood cells (inset) indicative of patency and perfusion (FIG. 4B(iv): left panel, vector control; FIG. 4B(iv): right panel, shAxl-2). *p<0.05, **p<0.005, ***p<0.0005 (paired t-test) compared with control. N=7 mice/group. (FIG. 4B(v)) Intrascaffold vessel diameter (FIG. 4B(v), left) is unaffected while microvascular density (FIG. 4B(v), right) is slightly enhanced in tissue engineered MDA-MB-231 tumors inhibited by shAxl2 expression.

[0272] FIGS. 5(A-D) show in vivo epi-allelic analysis reveals a distinct Axl expression threshold required for breast tumour formation. (FIG. 5A(I-III)) Temporal in vivo imaging of bioluminescence from subcutaneous epi-allelic MDA-MB-231/GFP-Luc xenografts in NOD/SCID !2mnull mice comprising graded Axl expression by Axl-targeting shRNAs shAxl278 FIG. 5A(I), shAx280 FIG. 5A(II) and shAxl2 FIG. 5A(III) compared to an ineffective Axl-targeting shRNA (shAxl279) control. (FIG. 5B) is a graph that shows correlation with surface Axl levels demonstrated a threshold of Axl expression required for tumor formation (FIGS. 5C(i) and (ii)) Bar graphs show mean changes in photons (FIG. 5C(i)) and tumour diameters (FIG. 5C(ii)) based on optical imaging analysis of tumours. Epi-allelic MDA-MB-231/GFP-Luc tumor growth (total photon, FIG. 5C(i)) and radial infiltration (signal diameter, FIG. 5C(ii)) measurements were normalized to shAxl279 (ineffective shRNA). (FIG. 5D) Tumor growth (28 day measurements) plotted versus Axl knockdown reveal the therapeutic threshold (80% reduced expression). *p<0.05, **p<0.005 (paired t-test) compared with control. N=6 mice/group.

[0273] FIGS. 6(A-D) show that Axl is required for metastasis of breast carcinoma cells. (FIG. 6A(I)-(III)) In vivo monitored primary tumour and metastasis, FIG. 6A(I), Orthotropic growth (upper panel, FIG. 6A(I)(i)) and metastasis to lymph node over time (lower panel, FIG. 6A(I)(ii)). FIG. 6A(II), Primary tumour growth monitored by in vivo optical imaging (GE Explore Optix) of wild type (solid line) and Axl RNAi implants (broken line). (FIG. 6B) Detection of metastasis in different organs. (FIG. 6C(1) and (II)) Ex vivo detection of breast cancer cells. FIG. 6C(I), Images of excised organs are shown. FIG. 60(11), Histological analysis of the same tissues confirmed metastasis (arrow) to different organs. (FIG. 6D(i)-(iii)). (FIG. 6D(i)) Mann-Whitney (log-rank) test between control (vector) and shAxl2 orthographically injected mice shows an increasing of survival in MDA-MB-231-D3H2LN/GFP-Luc-shAxl2 tumor-bearing mice (P=0.013). FIG. 6D(ii) is a table that shows the mean and median days alive for control (vector) and shAxl2 orthographically injected mice. FIG. 6D(iii) is a table that shows the test statistics.

[0274] FIGS. 7(A-D) show that Axl is essential for post-immune response recurrence and metastasis of syngeneic breast carcinoma cells in BALB/c mice. (FIG. 7A) 4T1-GFP-Luc mouse breast carcinoma cells expressing mouse Axl-targeting shRNA (shmAxl2) or human Axl-targeting shRNA (shAxl279) were analyzed by flow cytometry for mouse Axl surface expression or isotype control. (FIGS. 7B(i) and (ii)) Temporal in vivo imaging of bioluminescence from orthotropic (mammary fat pad) injected 4T1-GFP-Luc cells expressing either a mouse Axl-targeting shRNA (4T1-GFP-LucshmAxl2) or negative control human-specific shRNA (4T1-GFP-Luc-shAxl279) into BALB/c mice (FIG. 7B(i)). Quantification of whole-body bioluminescence (total photon) in control (4T1-GFP-Luc shAxl279, solid line) and Axl-knockdown (4T1-GFP-Luc-shmAxl2, grey line) injected BALB/c mice over an 8 week period (FIG. 7B(ii)). *p<0.05 (t-test), N=7 mice/group. (FIG. 7C) Survey of spontaneous metastasis (at 8-weeks post-orthotopic implantation) to different organs monitored by ex vivo bioluminescence detection of 4T1-GFP-Luc cells in excised organs from control or Axl-knockdown tumor bearing BALB/c mice. FIGS. 7D(i) and (ii) shows individual excised organs that were imaged and total light emission quantified.

[0275] FIG. 8 shows the sequence for the vector L383 pCSI Puro2AGFP2ALuc2.

[0276] FIG. 9(A-E). FIG. 9, Panel A(i) and (ii), shows flow cytometry analysis of MCF10a cells transduced with Slug or Ha-Ras (pBABE puro H-Ras V12, Addgene) constructs, analysed using co-expression of GFP; Ha-Ras expression was selected by puromycin treatment for 48 hours. Slug and Ha-Ras expression in MCF10a cells led to a strong increase in surface expression of the cancer stem cell marker CD44.

[0277] FIG. 9, Panel B, shows that Axl surface expression in MCF10a cells encoding Slug, Ras. Axl surface expression correlates with the presence of CD44 and mesenchymal traits both in Slug and Ha-Ras induced EMT.

[0278] FIG. 9, Panel C, shows mesenchymal morphology of Slug or Ha-Ras expressing MCF10a cells at 72 hours post seeding sorted by FACS for CD44 high (CD44+) and low (CD44) CD44-expressing sub-populations. CD44 cells show epithelial morphology, while CD44+ MCF10 cells demonstrate elongated mesenchymal morphology.

[0279] FIG. 9, Panel D(i) and (ii), show Western blot analysis of CD44 and CD44+MCF10a cells transduced with retroviral vectors encoding Slug(FIG. 9D(i), Ras(FIG. 9D(ii)). CD44 MCF10a cells retained epithelial junctional and cytoskeletal protein expression. In contrast, CD44+ cells showed strong mesenchymal marker expression (vimentin, N-cadherin) and loss of E-cadherin, demonstrative of EMT.

[0280] FIG. 9, Panel E, shows growth of the CD44+ and CD44 Slug and Ha-Ras expressing MCF10a cells in 3-D matrigel. CD44+, Axl-expressing MCF10a cells are invasive, consistent with a mesenchymal phenotype.

[0281] FIG. 10(i)-(iii) shows that MCF10a cells constitutively express the Axl ligand, Gas6 (Top (FIG. 10(i), Western blot, total lysate), and secrete (Middle (FIG. 10(ii), Western blot, conditioned medium) Gas6 that becomes cell associated. (Bottom (FIG. 10(iii), anti-Gas6 flow cytometry analysis) on Slug- and Snail-induced Axl expression.

[0282] FIG. 11(i)-(iii) shows that MDA-MB-213 cells constitutively express the Axl ligand, Gas6. (Top FIG. 11(i), Western blot, total lysate that is predominantly cell-associated; Middle FIG. 11(ii), Western blot, conditioned medium; Bottom FIG. 11(iii), anti-Gas6 flow cytometry analysis). Axl knockdown reduces cell-associated Gas6 while increasing levels in conditioned medium (Gas6*) without affecting overall expression.

EXAMPLES

[0283] Materials and Methods

[0284] Plasmids and Antibodies

[0285] All shRNAs were expressed from a modified human U6 promoter in the LTR of the retroviral vectors RRI-Red/L087 (Genbank: EU424173) used to transform MDA-MB-231 cells via retroviral infection. RRI-Red also expresses Puro2AmRed1 resulting in puromycin resistance and red fluorescence in successfully transformed cells. All Axl cDNA nucleotides are numbered as in Genbank: BC032229.

[0286] The following sequences were used: Axl2; (hairpin in small letters)

TABLE-US-00002 (SEQIDNO:1) GACATCCTCTTTCTCCTGCGAAGCCCATctggtcATGGGCTTCGCAGGAG AAAGAGGATGTC, shAxl278; (SEQIDNO:2) ACGGGTCTCCTTCTTTCGCCGttggatccctggtcggatccaaCGGCGAA AGAAGGAGACCCG, shAxl279; (SEQIDNO:3) GCTTCAGGCGATTTCCCCGGCGttggatccctggtcggatccaaCGCCGG GGAAATCGCCTGAAGC, shAxl280; (SEQIDNO:4) ATGCACGCCCAGCCGCACAGCGttggatccctggtcggatccaaCGCTGT GCGGCTGGGCGTGCAT shLuc; (SEQIDNO:5) GATTATGTCCGGTTATGTAAACAATCCGGctggtcCCGGATTGTTTACAT AACCGGACATAATC.

[0287] The retroviral expression vector L383 pCSI Puro2AGFP2ALuc2 (see FIG. 8) was made in several stages by cloning the coding sequences of puromycin-N-acetyl-transferase, EGFP, and firefly lucif erase into CRUS-retroviral expression vector (Bl et al., 2007). Each open-reading frame was separated from the next by a linker encoding the 2A region (XXSGLRSGQLLNFDLLKLAGDVESNPGP) from foot-and-mouth disease virus. This sequence is cleaved co-translationally resulting in the production of approximately stoichiometric amounts of each protein (Lorens 2004). Plasmids expressing hSnail, hSlug and hZEB2 were constructed by cloning the appropriate fragments from constructs BC012910, BC015895 and BC060819 (Open Biosystems) respectively into the CRU5-IRES-GFP retroviral vector (Lorens et al 2000). Two antibodies against human Axl were used; mouse monoclonal anti-human Axl (MAB154, R&D Systems) and goat anti-human Axl (M-20, Santa Cruz). In addition, the following antibodies were employed; rabbit anti-human pAxl (Y779, R&D), rat anti-human Snail (SN9H2, Cell Signaling), mouse anti-human Slug (L40Cb, Cell Signaling), rabbit anti-human E-cadherin (24E10, Cell Signaling), rabbit anti-human N-cadherin (ab18203, Abcam), actin, mouse anti-human b-catenin (L54E2, Cell Signaling), mouse anti-human Gas6 (R&D Systems), rabbit anti-human Twist (Twist2C1a, Abcam).

[0288] Cell Culture, Retroviral Transductions and Cell Proliferation Assay

[0289] All cells were cultured at 37 C., 5% CO.sub.2. Phoenix A cells (Dr. Gary Nolan, Stanford), MDA-MB-231 human breast epithelial carcinoma cells (American Type Culture Collection, Rockville, Md.), human dermal microvascular endothelial cells (HMVEC), and pulmonary artery smooth muscle cells (PASMC) (Cambrex, Walkersville, Md., USA) were maintained as previously described (Holland 2005). The clonal cell line MDA-MB-231-DH3L2N (Xenogen Corporation, Alameda, Calif., USA) was cultured in Minimum Essential Medium with Earl's Balanced Salts Solution MEM/EBSS medium supplemented with 10% FBS, 1% nonessential amino acids, 1% L-glutamine, and 1% sodium pyruvate. Phoenix A cells were transfected using the calcium phosphate method (Swift, 1999). Approximately 30 hours after transfection, the medium was changed to growth medium for the cells to be infected, supplemented with 10% FBS. Infectious supernatant was collected 48 hours after transfection. Target cells were exposed to supernatant containing 5 mg/ml protamine sulphate over night. Infected cells were selected with(1 g/ml puromycine. Cell proliferation assay was performed to analyze the proliferation potential of the different Axl knock down cells compared to the control cell line using MTS assay from Promega. Cells were seeded in 96-well tissue culture plates either untreated or coated with either 20 l Collagen (from Rat tail, Roche), Fibronectin (Sigma) or Matrigel (BD Biosciences). 2000 cells in 100 l medium were seeded in triplicates in 96-well plates and assayed every 24 hours using the MTS assay from Promega according to the manufacture's instructions.

[0290] Immunostaining, Flow Cytometry and Cell Sorting

[0291] The MDA-MB-231 cells were trypsinated using standard procedures and washed in PBS-0.2% BSA before staining with anti-Axl (#MAB154, R&D Systems) at a final concentration of 5 mg/ml in PBS-0.2% BSA for 40 minutes at room temperature. Cells were washed twice in PBS-0.2% BSA) and incubated with secondary antibody (Goat anti-mouse APC (Allophyocyanin, crosslinked, Molecular Probes) at a final concentration of 0.2 mg/ml for 30 minutes at room temperature in the dark. Cells were washed twice with PBS-2% BSA and resuspended in 300 ml PBS-0.2% BSA before analysis on a FacsCalibur Flow Cytometer (BD Biosciences). Data analyses were carried out using the FlowJo software (Tree Star, Inc., Ashland, Oreg., USA).

[0292] Cells expressing high levels of GFP, RFP and low Axl (shRNA) were isolated by FACS Aria SORP with laser 488 nm, 532 nm, 638 nm and 407 nm to establish stable, homogenous populations of cells.

[0293] Protein Extracts, SDS-PAGE, Immunoblotting and Immunoprecipitation

[0294] Cells were lysed in RIPA buffer (PBS with 1% (v/v) Nonidet P-40 (NP-40), 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) SDS) supplemented with protease inhibitor (Complete Mini, EDTA-free, Roche #13457200) and 0.2 mM PMSF. SDS-PAGE and immunoblotting were carried out according to standard procedures. For immunoprecipitation, cells were lysed in NP-40 buffer (10%glycerol, 1% NP-40, 50mM Tris pH 7.4, 0.2 M NaCl, 2.5 mM MgCl2) supplemented with protease inhibitor and PhosSTOP phosphatase inhibitor cocktail (Roche). Extracts were incubated with antibody coubled to protA/G beads for 1 hour at 4 oC, and the beads were washed in NP-40 buffer four times before elution.

[0295] Invasion Assay and 3D Matrigel Assay

[0296] The boyden chamber chemoinvasion assay (Albini et al 2004) was carried out using Becton Dickinson (BDFalcon cell culture inserts (8 m), Falcon multiwell 24 well plate, and growth factor reduced matrigel from BD. Inserts were coated on the inside with serum free medium diluted matrigel to a final concentration of 30 g matrigel each insert. Matrigel work is handled at 4 C. until solidification for 30 minutes at 37 C. 5105 cells, resuspended in serum free cell medium with 0.1% BSA added on top of each insert. FBS enriched cell medium functions as a chemoattractant. After 20 hours incubation in 37 C. with 5% CO2 the cells inside the chamber were removed by a cotton swab. Cells in the other side of the membrane were fixated, and then stained with DAPI. Pictures were taken using a fluorescent microscope and the cells counted using ImageJ rsb.info.nih.gov/nih-imageJ, Wayne Rasband,National).

[0297] The 3D assays were modified from Sandal et al 2007; 30000 cells were seeded out on gel and cultures were allowed to grow for 10-12 days before they were analysed.

[0298] Clinical Samples

[0299] The present series of breast cancers was selected from the population based Norwegian Breast Cancer Screening Program (Hordaland County), which started in 1996 with two-view mammography done every 24 months. Briefly, 95 invasive interval cancers occurred during the first two screening intervals (1996-2001), and these were matched by size with 95 screen-detected tumors from a total of 317 invasive tumors during the first two rounds (median diameter 15.6 and 15.7 mm, respectively. After matching, the mean tumor size for screen detected and interval cases were 25.1 and 23.1 mm, respectively, and the corresponding mean age in these groups was 62 and 59 years. In addition to age and tumor diameter (by pathologic examination), basic characteristics, such as breast density, histologic type, histologic grade, lymph node metastases, and distant metastases at diagnosis were recorded. The median time from the last mammogram to the diagnosis of interval cancer was 17.1 months. Last date of follow-up was Nov. 31, 2004, and median follow-up time (of survivors) was 72 months. During the follow-up period, 31 patients died of breast cancer.

[0300] Immunohistochemistry

[0301] Tissue microarray slides were used in the present study. The tissue microarray technique is tissue conserving and has been validated in several studies. Immunohistochemistry was performed on 5 pm thick sections of formalin-fixed, paraffin-embedded tissues. Antigen retrieval was performed by boiling for 10 min at 750 W and 20 min at 350 W in TRS (Target Retrieval Solution; DakoCytomation, Denmark, AS) buffer, ph 6.0, in a microwave oven. A DakoCytomation Autostainer was used for staining.

[0302] The slides were incubated overnight at room temperature with a polyclonal antibody against Axl (H-124; cat. #20741), dilution 1:200 (Santa Cruz, USA). Immunoperoxidase staining was carried out using the DakoCytomation Envision Kit (DakoCytomation, Denmark AS) with diaminobenzidin tetrachloride peroxidase as substrate prior to counterstaining with Mayer's haematoxylin (DakoCytomation, Denmark AS).

[0303] Evaluation of Staining

[0304] The staining was predominantly cytoplasmatic, although there was some concentration of staining in the cytoplasmatic membrane. Staining was recorded by a semiquantitative and subjective grading system, considering the intensity of staining and the proportion of tumor cells showing a positive reaction. All three cores from each case were evaluated. Intensity was recorded as 0 (no staining) to 3 (strong staining); the percentage of membranous staining area was recorded as 0 (no tumor cells positive), 1 (<10%), 2 (10%-50%), and 3 (>50% of tumor cells). A staining index (SI) was calculated as the product of staining intensity and area. Immunohistochemical registration was done blinded for patient characteristics and outcome.

[0305] Statistics

[0306] Comparisons of groups were performed by Pearson 2 test. In all statistical analyses, cut-off values for staining index (SI) categories were based on median values. Univariate survival analyses (using death from endometrial carcinoma as end point; death from other causes were censored) were performed using the product-limit procedure (Kaplan-Meier method), with the time of primary operation as the entry date.

[0307] The log-rank (Mantel-Cox) test was used to compare survival curves for different categories of each variable. Variables with impact on survival in univariate analyses (P0.15) were examined by log-log plot to determine how these variables could be incorporated in Cox' proportional hazards regression models.

[0308] Mouse Strain and Animal Care

[0309] In the present study we used female mice of PrkdcSCID/B2mnull (abbreviated as NOD/SCID/B2mnull), severe combined immunodeficient mice (GADES Institute, Norway), aged 8-10 weeks and weighting between 20-25 g. They were kept under standard conditions in a 12 h light/dark cycle and allowed at least 7 days to acclimatize to their new environmental condition prior to onset of experiment. This investigation, designed to minimize the number of animals and suffering, was carried out in accordance with the Norwegian Regulation on Animal Experimentation, the European Convention for the Protection of Vertebrate Animals used for scientific purposes and the guidelines of the Norwegian Animal Research Authority.

[0310] Bioluminescence Imaging (BLI)

[0311] BLI was performed using a eXplore Optix (GE Healtcare) camera mounted in a specimen box. Imaging and quantification of signals was done using eXplore Optix software. For in vivo imaging, animals received via intraperitoneal injection (i.p) the substrate D-luciferin (Biosynth), 150 mg/kg in PBS (Phosphate Buffered Saline) and anesthetized with isoflurane. Mice were placed into warmed stage inside of camera box with continuous exposure to 1-2% isoflurane and imaged for different views depending on the tumor model. Region of interest were identified and were quantified as total photon/sec-1 using eXplore Optix software (GE Explore Optix). In vivo background bioluminescence was in the range of 2-310 photon counts. For ex vivo imaging, 150 mg/kg D-luciferin was injected into the mice just before necropsy. Tissues of interest were excised, placed into plates and imaged.

[0312] In Vivo Tumor Models

[0313] Tissue Engineering

[0314] 1106 MDA-MB 231 cells which express GFP-Luc biomarker and different RNAs interference that regulate Axl expression, were suspended in a 1:1 mixture of F-12 Kaighn's (Invitrogen) :Matrigel (BD Biosciences) and seeded in 66 mm Poly-lactic acid (PLLA) scaffolds.

[0315] Females NOD/SCID/B2mnull were anesthetized by exposure to 1-2% isoflurane (Isoba vet.-Schering-Plough A/S) during the implantation procedure and on subsequent imaging days. Two scaffolds were implanted subcutaneously (s.c) in each mouse according to Nor et al. Lab Invest.; 81 (4) 453; 2001 Scaffolds with cells that express Axl were implanted on the left side of each mouse while on the right side scaffolds with cells in which Axl was knockdown. After the surgical procedure, anesthetized mice were placed in Imaging System and imaged for both left and right sides 10-15 min after intraperitoneal injection of 150 mg/kg D-luciferin (Biosynth). Tumor development was monitored in vivo once a week by imaging for 4 weeks.

[0316] Xenograft Assay

[0317] 1106 MDA-MB-231 cells infected with shRNA vectors and GFP-Luc were suspended in 100 l of F12K medium+10% FBS/Matrigel (1:1 ratio, BD Biosciences) and injected with a 29-gauge insulin needle subcutaneously into both flanks of female NOD/SCID/B2m. At the left flank positive cells on Axl expression were injected, while in right flank cells negative in Axl expression were injected.

[0318] For tricellular implant MDA-MB-231 cells infected with shRNA vectors and GFP-Luc were mixed with Human dermal microvascular endothelial cells (HMVEC), and pulmonary artery smooth muscle cells (PASMC) in ratio of 1:2:2. Tumor growth was monitored weekly in vivo by imaging for 4 consecutive weeks.

[0319] Mammary Fat Pad Spontaneous Metastasis Model

[0320] Subline of human MDA-MB-231 cells (called MDA-MB 231 DH3L2NXenogen) which express GFP-Luc biomarker and different RNAs interference that regulate Axl expression were injected into mammary pad of mice. NOD/SCID/B2mnull mice were anesthetized by exposure to 1-3% isoflurane and injected with 50 l of 2106 MDA-MB-231 DH3L2N cells suspended in MEM/EBSS medium/Matrigel (1:1) into the abdominal mammary fat pad. 10-15 min after D-luciferin (Biosynth) injection, mice were placed in the eXplore Optix Imaging System and imaged from the ventral view. Tumor growth and metastasis spread was monitored every second week by bioluminescent imaging for up to 9 weeks. The lower part of each animal was covered before reimaging, to minimize the bioluminescence from the primary tumor so that the signals from the metastatic regions could be observed in vivo.

[0321] Tissue Collection

[0322] At the end of each experiment the tumor tissues implants and different organs were retrieved from the mice and preserved in 10% Paraformaldehyde (Sigma-Aldrich) for further analysis. Tissues was prepared for histopathology (paraffin embedding, sectioning and staining) and analyzed by microscope evaluation.

[0323] Statistical Analysis of Animal Model Results

[0324] The mean bioluminescence (photons/sec-1), tumor diameter and corresponding standard errors were determined for each experiment. Regression plots were used to describe the relationship between bioluminescence, cell number and tumor diameter. Statistical analyses were based on paired t-test.

[0325] Results

[0326] Axl Expression is a Strong Prognostic Factor for Overall Survival of Breast Cancer Patients

[0327] In order to assess the role of Axl in breast cancer pathogenesis, we investigated Axl expression in tumors from a series of breast cancer patients identified during a Norwegian Breast Cancer Screening Program which started in 1996 entailing bi-annual two-view mammography (Wang et al 2001). Briefly, 95 invasive interval cancers occurred during the first two screening intervals (1996-2001), and these were matched by size with 95 screen-detected tumors from a total of 317 invasive tumors during the first two rounds (median diameter 15.6 and 15.7 mm, respectively (Collett et al., 2005). After matching, the mean tumor size for screen detected and interval cases were 25.1 and 23.1 mm, respectively, and the corresponding mean age in these groups was 62 and 59 years. In addition to age and tumor diameter (by pathologic examination), basic characteristics, such as breast density, histologic type, histologic grade, lymph node metastases, and distant metastases at diagnosis were recorded. The median time from the last mammogram to the diagnosis of interval cancer was 17.1 months. Clinical parameters were monitored, last date of follow-up was Nov. 31, 2004, and median follow-up time (of survivors) was 72 months. During the follow-up period, 31 patients died of breast cancer.

[0328] There were no significant associations between Axl expression (divided in two groups by median staining index; FIG. 1A) and important clinico-pathologic features such as histologic grade, tumor diameter, expression of estrogen and progesterone receptors, and axillary lymph node status. Also, Axl expression was not associated with HER2, E-cadherin, markers of basal differentiation (Cytokeratin 5/6, P-cadherin), EZH2, or tumor cell proliferation by Ki-67 expression.

[0329] However, univariate survival analysis (Kaplan-Meier method, log-rank test), Axl expression was significantly associated with reduced patient survival (p=0.035; FIG. 1B). In multivariate analysis (proportional hazards method), including basic prognostic factors like tumor diameter, histologic grade and lymph node status in addition to Axl expression (step one), Axl expression status remained as an independent negative prognostic factor in the final model (p=0.021), in addition to histologic grade and lymph node status (see FIG. 1C). Thus, Axl expression is a strong prognosticator of poor clinical outcome in breast cancer patients.

[0330] We then investigated Axl expression in patient biopsies of matched pairs (n=16) of primary and metastatic breast carcinomas. Axl expression tended to be further elevated in metastases when compared with corresponding primary human breast carcinomas (p=0.11, McNemar's test; FIG. 1d; metastases to the right, liver (upper) and bone (lower)), suggesting that Axl expression is a strong prognosticator of poor clinical outcome in breast cancer patients and associated with metastatic spread.

[0331] Axl is Required for Breast Cancer Cell Invasiveness

[0332] The strong correlation of Axl expression in early breast carcinomas with poor survival indicates an important role for Axl in overall disease pathogenesis. As breast cancer-related mortality invariably results from complications of metastatic disease, we assessed whether Axl expression was required for malignant breast carcinoma cell invasiveness. Axl is expressed in several highly metastatic human breast carcinoma cell lines including MB-MDA-231. The Axl ligand, Gas6 is often co-expressed, leading to autocrine activation (Holland et al., 2005). In order to effectively correlate Axl expression levels in MB-MDA-231 cells with specific cellular behaviors, we developed an epi-allelic series of Axl-targeting shRNAs that reduce Axl expression in a dose dependent manner, using a recently developed FACS-based RNAi approach, CellSelectRNAi (FIG. 2A; Micklem et al., in preparation). This Axl shRNA collection was used to create an epi-allelic Axl MB-MDA-231 cell series with graded total (FIG. 2B) and phosphorylated (FIG. 2C) Axl protein levels.

[0333] Malignant carcinoma cells exhibit mesenchymal cell invasiveness in three-dimensional extracellular matrix protein gels (Matrigel) that correlates with in vivo metastatic potential (Bissell). Epi-allelic analysis demonstrated a dose-dependent requirement for Axl expression for MB-MDA-231 cell invasion in response to serum (FIG. 2D) or the SDF-1 chemokine (FIG. 2E), an important factor in breast carcinoma metastasis. In contrast, Axl knockdown had no effect on MB-MDA-231 cell proliferation and only a modest effect on two-dimensional (lateral epithelial wound healing) migration. This indicated a specific requirement for Axl in three-dimensional growth and invasiveness. We therefore evaluated the effect of Axl knockdown on MB-MDA-231 cells in a 3D-Matrigel assay. Normal breast epithelial cells self-organize into polarized spheroid acinar structures in 3D-Matrigel, while malignant MB-MDA-231 cells proliferate forming large disorganized colonies with invasive, stellate outgrowths that reflect aggressive tumors (Bissell). Knockdown of Axl expression strongly reversed the malignant phenotype of MB-MDA-231 cells in 3D-Matrigel, creating small round colonies without malignant outgrowths (FIGS. 2F,G). Together, these data suggest that Axl signaling is required to maintain the mesenchymal-like invasiveness of metastatic breast carcinoma cells.

[0334] Axl is Upregulated by EMT-Inducing Transcription Factors in Breast Epithelial Cells

[0335] The acquisition of mesenchymal invasiveness, the ability to migrate and invade ECM, is the functional hallmark of EMT. The EMT-inducing transcription factor Twist is required for metastasis of breast carcinoma cells (Yang et al 2004). We therefore investigated if Twist expression upregulates Axl in breast epithelial cells. Twist expression in normal breast epithelial cells (MCF10A) induces EMT (FIG. 3A; Glackin et al., in preparation). Strikingly, Axl is also strongly upregulated in Twist-expressing MCF10A cells (FIG. 3A-B). Further, as Gas6 is constitutively expressed by normal breast epithelial cells, this Twist-induced Axl expression establishes an autocrine activation loop, evidenced by increased cell-associated Gas6 and tyrosine phosphorylated Axl (FIGS. 3A, C). To determine if other EMT inducing transcription factors similarly upregulate Axl expression, we analyzed MCF10A cells expressing ZEB2, Snail and Slug. Each of these EMT transcription factors induced mesenchymal transition in MCF10A cells and upregulated Axl expression. These results suggest that EMT induction leads to Axl expression and can establish autocrine signaling.

[0336] Axl Expression is Necessary for Tumor Formation in Experimental Tissue Engineered Breast Tumors

[0337] In order to evaluate the requirement for Axl for malignant growth in vivo in we used a tissue engineering approach comprising MB-MDA-231 cells that express a GFP-luciferase construct for efficient in vivo optical imaging (CSI; Tiron et al., unpublished results), seeded with Matrigel into poly-lactic acid tissue engineering scaffolds and implanted subcutaneously into immunocompromised NOD-SCID mice. Growth within engineered tumor microenvironments is associated with tumor cell mesenchymal characteristics as tumor cells colonize the scaffold (Mooney). MB-MDA-231 cells readily form tumors this biomimetic microenvironment, displaying aggressive colonization of the scaffold (FIG. 4A). Axl knockdown strongly inhibited tumor formation, lateral spread and malignant morphology (FIG. 4A).

[0338] To ascertain whether Axl influences the ability of breast cancer cells to attract and co-opt blood vessels, we developed a tri-cellular implant approach comprising MB-MDA-231 cells seeded together with primary human microvascular endothelial (HuMVEC) and vascular smooth muscle cells (vSMC) to create tumor vasculature. Implants of human EC-vSMC cells readily form perfused intrascaffold human microvasculature in NOD-SCID mice within a two-week period (Hegen et al., in preparation). As shown in FIG. 4B, MB-MDA-231 cells form aggressive, highly vascularized tumors in this tri-cellular implant model. The engineered human tumor vasculature is evenly distributed and perfused with intralumenal red blood cells. In contrast, Axl knockdown, blocked tumor formation, without affecting development of a perfused human microvasculature. This indicates that Axl is required for tumor formation even in the presence of a microvasculature that likely obviates the need for induced angiogenesis.

[0339] A Distinct Axl Expression Threshold is Required for Breast Tumor Formation

[0340] In order to evaluate the level of Axl expression needed to form a tumor, we conducted an in vivo epi-allelic analysis of Axl in subcutaneous MDA-MB-231tumors. The Axl epi-allelic MDA-MB-231cell series (FIG. 1) was injected subcutaneously and temporally monitored for tumor formation by bioluminescent scanning. This approach revealed an Axl dose response for tumor growth (FIG. 5). Correlation with surface Axl levels demonstrated a threshold of Axl expression required for tumor formation (FIGS. 5B,C). This dose response is congruent with the dose dependent effects of Axl inhibition on invasiveness (FIG. 2).

[0341] Axl is Essential for Metastasis of Breast Carcinoma Cells

[0342] In order to evaluate the requirement of Axl for breast cancer metastasis, we orthotopically injected MDA-MB-231-D3H2LN, a rapidly growing and highly metastatic in vivo MDA-231 isolate, into the mammary fat pad. Using whole body bioluminescent imaging we temporally monitored spontaneous metastasis development over a 9-week period. Control MDA-MB-231-D3H2LN cells generated large orthotopic mammary tumors that became necrosing within 5-6 weeks (FIG. 6A). Axl knockdown in MDA-MB-231-D3H2LN reduced the rate of primary mammary tumor formation but also grew substantial primary mammary tumors (FIG. 6A). Spontaneous metastasis was initially detected in the thoracic sentinel lymph node of control MDA-MB-231-D3H2LN injected mice at 4 weeks (FIG. 6A). In contrast no metastases were detected in MDA-231DHLN-AxlshRNA implanted mice. Upon sacrifice at 9 weeks, excised organs were scanned individually for bioluminescence due to the presence of metastatic cells. MDA-MB-231-D3H2LN injected mice had formed extensive spontaneous metastasis in all mice, including lymph node, lung, ovaries and kidneys. In contrast, MDA-MB-231DHLN-AxlshRNA cells did not form detectable metastasis (apart from a single lesion in the kidney of one mouse). Histological analysis of tissue biopsies from the organs of MDA-MB-231-D3H2LN injected mice confirmed the presence of multiple micro- and macrometastases in all organs as predicted by the bioluminescent total photon measurements (FIG. 6C). No micro- or macrometastases were observed in tissue biopsies from MDA-MB-231 DHLN-AxlshRNA injected mice, confirming that the lack of observed bioluminescence was due to inhibited metastasis formation. These results show that Axl is essential for breast carcinoma metastasis.

[0343] We evaluated the functional contribution of Axl to overall survival of NOD-SCID mice with orthotopically injected MDA-MB-231-D3H2LN/GFP-Luc control or Axl knockdown cells. Overall survival was significantly increased in MDA-MB-231-D3H2LN/GFP-Luc-shAxl2 tumor-bearing mice (P=0.013, log-rank test; FIG. 6d). These results together with our clinical observations support the conclusion that Axl is an important to the development of metastatic disease and overall patient survival.

[0344] In order to validate our results in a different metastatic model we transduced the highly metastatic mouse breast carcinoma 4T1 cell line with the CSI-construct and selected for GFPluciferase expression by FACS (4T1-GFP-Luc). The 4T1 cells are dependent on Twist expression for metastasis and exhibit high levels of Axl expression (FIG. 7a). We developed a retroviral vector that expresses a mouse Axl-targeting shRNA (shmAxl2) that effectively suppresses mouse Axl surface levels in 4T1 cells (FIG. 7a). A mismatched human Axl-targeting shRNA (shAxl279) had no effect on mouse Axl expression. Similar to Twist knockdown in 4T1 cells and results with MDA-MB-231 cells, tissue culture expansion of the Axl-knockdown 4T1 cells was not significantly affected (data not shown).

[0345] When introduced into the mammary gland of female normal BALB/c mice, the syngenic 4T1 tumor cells display a biphasic growth pattern, due to a rigorous immune response that leads to tumor regression associated with leukocyte infiltration and necrosis, followed by re-growth at the primary site that coincides with extensive metastasis to multiple organs. 4T1-GFPLuc cells show only monophasic growth in immunocompromised NOD-SCID mice (data not shown). We injected 4T1-GFP-Luc cells expressing either a mouse Axl-targeting shRNA (4T1-GFP-Luc-shmAxl2) or negative control human-specific shRNA (4T1-GFP-LucshAxl279) into the mammary fat pad of female BALB/c mice and quantified tumor growth and metastasis by temporal whole-body in vivo optical imaging (FIG. 7b). The control 4T1-GFP-Luc-shmAxl2 cells displayed rapid primary growth, reaching a maximum after one week. This was followed by a precipitous regression that was sustained for five weeks (FIG. 7b). At week 6, recurrence at the primary site and multiple distant metastases were observed that subsequently grew rapidly, causing moribundity and lethality in all mice by week 8. The 4T1-GFP-Luc cells expressing the mouse Axl-targeting shRNA (4T1-GFP-Luc-shmAxl2) initially followed a similar course: a rapid primary tumor growth, slightly attenuated by Axl suppression, followed by regression (FIG. 7b). However, the subsequent recurrence of the primary tumor and emergence of rapidly growing distant metastasis was completely absent (FIG. 7b). Indeed, all mice injected with the 4T1-GFP-Luc-shmAxl2 cells remained healthy at the time of sacrifice (8 weeks). The splenomegaly associated with the leukemoid reaction characteristic of the 4T1 model was also reduced in mice bearing Axl-knockdown tumors (data not shown). Upon sacrifice at 8 weeks, individual excised organs were imaged and total light emission quantified, confirming the presence of metastases at common dissemination sites in all mice bearing 4T1-GFP-Luc-shAxl279 tumors (FIGS. 7c-d). In contrast, no bioluminescent tumor cells were detected in the organs from mice with 4T1-GFP-LucshmAxl2 tumors (FIGS. 7d-e). These results support the conclusion that Axl is an essential regulator of breast tumor metastasis.

[0346] Autocrine Regulation by Gas6

[0347] The Axl ligand Gas6 is often coexpressed with Axl, consistent with autocrine activation. Cell-associated Gas6 and phosporylated Axl levels indicative of autocrine signaling are reduced upon Axl knockdown in MDA-MB-231 cells (FIG. 11). In MCF10A cells, Gas6 was constitutively expressed and became cell associated on EMT-induced Axl expression (FIG. 10). These results suggest that EMT program induction leads to Axl expression and can establish autocrine signaling in breast epithelial cells. As Axl, often coexpressed with Gas6, is detected in many metastatic cancers, autocrine Axl signaling may be a frequent consequence of EMT in many tumor types. EMT-inducing transcription factors such as Snail, Slug, and Twis potently induce Axl expression, suggesting that Axl could participate in a positive feedback loop that sustains the malignant mesenchymal phenotype of tumor cells. This notion is consistent with our observation of elevated Axl expression in metastatic lesions.

[0348] CD44+ Phenotype is Associated with Axl Expression

[0349] MCF10a cells were transduced with Slug or Ha-Ras (pBABE puro H-Ras V12, Addgene) constructs. Slug transduced cells were analysed by flow cytometry using co-expression of GFP; Ha-Ras expression was selected by puromycin treatment for 48 hours. As shown by flow cytometry (FIG. 9, panel A (i and ii)), Slug and Ha-Ras expression in MCF10a cells led to a strong increase in surface expression of the cancer stem cell marker CD44. Flow cytometry analysis of MCF10a cells transduced with Slug or Ha-Ras further showed a strong increase in surface Axl expression. Slug or Ha-Ras expressing MCF10a cells were sorted by FACS for CD44 high (CD44+) and low (CD44) CD44-expressing sub-populations. The CD44 cells showed epithelial morphology while the CD44+ MCF10 cells demonstrated elongated mesenchymal morphology (FIG. 9, Panel C). Western blot analysis of these cells (FIG. 9, Panel D) demonstrated that CD44 MCF10a cells retained epithelial junctional and cytoskeletal protein expression. In contrast, CD44+ cells showed strong mesenchymal marker expression (vimentin, N-cadherin) and loss of E-cadherin, demonstrative of EMT. Axl expression correlated with the presence of CD44 and mesenchymal traits both in Slug and Ha-Ras induced EMT (FIG. 9, Panels B, D). Growth of the CD44+ and CD44 Slug and Ha-Ras expressing MCF10a cells in 3-D matrigel (FIG. 9, Panel E) demonstrated that the CD44+, Axl-expressing MCF10a cells are invasive, consistent with a mesenchymal phenotype. These results demonstrate that Axl is upregulated by Slug and Ha-Ras expression in MCF10a cells and correlates with mesenchymal and cancer stem cell traits.

[0350] Various modifications and variations of the described aspects of the invention will be apparent to those skilled in the art without departing from the scope and spirit of the invention. Although the invention has been described in connection with specific preferred embodiments, it should be understood that the invention as claimed should not be unduly limited to such specific embodiments. Indeed, various modifications of the described modes of carrying out the invention which are obvious to those skilled in the relevant fields are intended to be within the scope of the following claims.

REFERENCES

[0351] Camp, R. L., V. Neumeister, and D. L. Rimm, A Decade of Tissue Microarrays: Progress in the Discovery and Validation of Cancer Biomarkers. J Clin Oncol, 2008.

[0352] Collett, K., et al., A basal epithelial phenotype is more frequent in interval breast cancers compared with screen detected tumors. Cancer Epidemiol Biomarkers

[0353] Elston, C. W. and I. O. Ellis, Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology, 1991. 19(5): p. 403-10.Prev, 2005. 14(5): p. 1108-12.

[0354] Gupta P B, Mani S, Yang J, Hartwell K, Weinberg R A. The evolving portrait of cancer metastasis.Cold Spring Harb Symp Quant Biol. 2005; 70:291-7.

[0355] Holland S. J., Friera A. M., Franci C., Chan E., Atchison R., McLaughlin J., Swift S. E., Pali E., Yam G., Wong S., Lasaga J., Shen M., Yu S., Xu W., Hitoshi Y., Payan D. G, Nor J. E., Powell M. J, and Lorens J. B. (2005) The Receptor Tyrosine Kinase Axl Regulates Angiogenesis and Tumor Growth. Cancer Research 65:9294-9303.

[0356] Kononen, J., et al., Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med, 1998. 4(7): p. 844-7.

[0357] Sun, W., Fujimoto, J. & Tamaya, T. Coexpression of Gas6/Axl in human ovarian cancers. Oncology 66, 450-7 (2004).

[0358] Thiery J P. Epithelial-mesenchymal transitions in tumour progression. Nat Rev Cancer. 2002 June; 2(6):442-54

[0359] Vajkoczy P, Knyazev P, Kunkel A, Capelle H H, Behrndt S, von Tengg-Kobligk H, Kiessling F, Eichelsbacher U, Essig M, Read T A, Erber R, Ullrich A Dominant-negative inhibition of the Axl receptor tyrosine kinase suppresses brain tumor cell growth and invasion and prolongs survival. Proc Natl Acad Sci U S A. (2006)103:5799-804.

[0360] Yang J, Mani S A, Weinberg R A Exploring a new twist on tumor metastasis. Cancer Res 2006 66:4549-52.

[0361] Yang J, Mani S A, Donaher J L, Ramaswamy S, Itzykson R A, Come C, Savagner P, Gitelman I, Richardson A, Weinberg R A. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell. 2004 117:927-39.

[0362] Yang J, Weinberg R A. Epithelial-mesenchymal transition: at the crossroads of development and tumor metastasis. Dev Cell. 2008 June; 14(6):818-29.

[0363] Wang, H., et al., Mammography screening in Norway: results from the first screening round in four counties and cost-effectiveness of a modeled nationwide screening. Cancer Causes Control, 2001. 12(1): p. 39-45.

[0364] Holland et al. R428, a selective small molecule inhibitor of Axl kinase, blocks tumour spread and prolongs survival in models of metastatic cancer. Cancer Research; 70(4), Feb. 15, 2010.