KINDLIN-1 AS A MARKER OF SENSITIVITY TO EGFR/RAS PATHWAY INHIBITORS

Abstract

The present invention concerns an ex vivo method for selecting patient having cancer as eligible to EGFR/RAS pathway inhibitor therapy, an EGFR/RAS pathway inhibitor for use for treating cancer in a patient, and a method for prognosis cancer outcome or progression.

Claims

1. An ex vivo method for selecting a patient having cancer as eligible to EGFR/RAS pathway inhibitor therapy, comprising: a) Measuring Kindlin-1 expression level in cancer cells of said patient having cancer; b) Selecting the patient as eligible to EGFR/RAS pathway inhibitor therapy if the measured Kindlin-1 expression level in cancer cells of said patient having cancer is equal to, or above, a reference level.

2. The method according to claim 1, wherein: i. the reference level is determined from the expression level of Kindlin-1 measured in samples of cancer cells sensitive to EGFR/RAS pathway inhibitor therapy, wherein said cancer cells sensitive to EGFR/RAS pathway inhibitor therapy are of the same type as the cancer cells of said patient having cancer, or ii. the reference level is determined from the expression level of Kindlin-1 measured in samples of cancer cells resistant to EGFR/RAS pathway inhibitor therapy, wherein said cancer cells resistant to EGFR/RAS pathway inhibitor therapy are of the same type as the cancer cells of said patient having cancer, and the measured Kindlin-1 expression level in cancer cells of said patient having cancer is above the reference level.

3. The method according to any one of claims 1-2, wherein the cancer is a solid cancer, preferably an epithelial cancer.

4. The method according to any one of claims 1-3, wherein the cancer is breast, lung, colon, pancreas, bladder, or head and neck cancer, preferably a triple negative breast cancer.

5. The method according to any one of claims 1-4, wherein the EGFR/RAS pathway inhibitor therapy is an EGFR inhibitor, a MEK inhibitor, a BRAF inhibitor, or a RAS inhibitor therapy.

6. The method according to claim 5, wherein the EGFR inhibitor is selected from the group consisting of anti-EGFR antibodies, preferably Cetuximab, Panitumumab, Zalutumumab, Nimotuzumab, or Matuzumab; and small molecule inhibitors, preferably Erlotinib, Dabrafenib, Gefitinib, Osimertinib, Pelitinib, AZD3759, afatinib, brigatinib, icotinib, or lapatinib.

7. The method according to claim 5, wherein the MEK inhibitor is selected from the group consisting of PD0325901, CI-1040, Selumetinib, Trametinib, Cobimetinib, Refametinib and TAK-733.

8. The method according to claim 5, wherein the BRAF inhibitor is selected from the group consisting of vemurafenib, PLX-4720 and Dabrafenib.

9. The method according to claim 5, wherein the RAS inhibitor is AMG510.

10. The method according to claim 1, which further comprises administering an EGFR/RAS pathway inhibitor to said patient if identified as having a Kindlin-1 expression level in cancer cells equal to, or above, the reference level.

11. The method according to claim 10, wherein: a. the reference level is determined from the expression level of Kindlin-1 measured in samples of cancer cells sensitive to EGFR/RAS pathway inhibitor therapy, wherein said cancer cells sensitive to EGFR/RAS pathway inhibitor therapy are of the same type as the cancer cells of said patient having cancer, or b. the reference level is determined from the expression level of Kindlin-1 measured in samples of cancer cells resistant to EGFR/RAS pathway inhibitor therapy, wherein said cancer cells resistant to EGFR/RAS pathway inhibitor therapy are of the same type as the cancer cells of said patient having cancer.

12. The method according to claim 11, wherein the EGFR/RAS pathway inhibitor therapy is an EGFR inhibitor, a MEK inhibitor, a BRAF inhibitor, or a RAS inhibitor therapy.

13. A method for prognosis cancer outcome or progression in a patient suffering from EGFR/RAS-driven cancer, comprising: a) Measuring Kindlin-1 expression level in cancer cells of said patient suffering from EGFR/RAS-driven cancer; b) Determining that the cancer is likely to progress or have a poor outcome if the measured Kindlin-1 expression level in cancer cells of the patient is equal to, or above, a first reference level; or determining that the cancer is unlikely to progress or have a poor outcome if the measured Kindlin-1 expression level in cancer cells of the patient is below a second reference level.

14. The method for prognosis cancer outcome or progression according to claim 13, wherein the patient is suffering from a bladder or a head and neck cancer.

15. The method for prognosis cancer outcome or progression according to claim 13, wherein: a. the first reference level is determined from the expression level of Kindlin-1 measured in samples of cancer cells of aggressive EGFR/RAS-driven cancers, and wherein said cancer cells are of the same type as the cancer cells of said patient having cancer, and b. the second reference level is determined from the expression level of Kindlin-1 measured in samples of cancer cells of non-aggressive EGFR/RAS-driven cancers, and wherein said cancer cells are of the same type as the cancer cells of said patient having cancer.

16. The method for prognosis cancer outcome or progression according to claim 13, which further comprises administering an EGFR/RAS pathway inhibitor or a Kindlin-1 inhibitor to said patient if identified as having a cancer likely to progress or having a poor outcome.

Description

FIGURES

[0083] FIG. 1A depicts a scatter plot representing the differential mRNA expression levels of Kindlin-1 in EGFR/RAS driven breast cancer cells versus non EGFR/RAS driven cells. Mean ±SE represented. Statistical analysis were performed using the Mann-Whitney test (****p<0,0001 ).

[0084] FIG. 1B depicts western blot results performed in a subset of breast cancer cell lines to compare Kindlin-1 expression with the activation of EGFR/RAS pathway at a protein level.

[0085] FIG. 2 depicts western blot performed in a subset of lung and pancreatic cancer cell lines to compare Kindlin-1 expression with the activation of EGFR/RAS pathway at a protein level.

[0086] FIG. 3 depicts a scatter plot showing Kindlin-1 protein expression in low versus high EGFR expression. Each point represents the H-score from a single tissue sample ranging from total absence of Kindlin-1 in the epithelial compartment (H-score 0), to very strong Kindlin-1 staining (H-score 300). Mean ±SE represented.

[0087] FIG. 4 depicts Box and whisker plots representing the differential mRNA expression levels of Kindlin-1 in EGFR/RAS driven tumors versus non EGFR/RAS driven tumors in pancreas, lung, bladder, head and neck and colon human tumors. Mean ±SE represented. Statistical analysis were performed using the Mann-Whitney test (****p<0.0001 ; ***p<0.001; **p<0.01 ).

[0088] FIG. 5 depicts Kaplan-Meier plots showing metastasis free survival of triple negative breast (TNBC), lung, pancreatic, bladder, head and neck and colon cancer patients for the expression of Kindlin-1 taking into account the activation of EGFR pathway and/or KRAS, HRAS, BRAF, EGFR mutational status. Statistical analysis were performed by a Log-rank test.

[0089] FIG. 6 depicts a scatter plot representing the differential mRNA expression levels of Kindlin-1 in low sensitive breast cancer cells versus high sensitive cells to different EGFR/RAS pathway inhibitors compared to a standard drug used in breast cancer treatment (palbociclib). Mean ±SE represented. Statistical analysis were performed using the Mann-Whitney test (**p<0.01; *p<0.05; ns: not significant).

[0090] FIG. 7 depicts a scatter plot representing the differential mRNA expression levels of Kindlin-1 in low sensitive cancer cells versus high sensitive cells to different EGFR/RAS pathway inhibitor, respectively in lung cancer, pancreas cancer, bladder cancer, head and neck cancer and colon cancer. Mean ±SE represented. Statistical analysis were performed using the Mann-Whitney test.

[0091] FIG. 8A depicts western blot of Kindlin-1 protein in 27 triple-negative breast cancer PDX models.

[0092] FIG. 8B depicts Pearson correlation was calculated between tumor growth inhibition (TGI) and Kindlin-1 protein expression in 15 PDX models (R= 0.58, p<0.02).

[0093] FIG. 8C depicts graphs showing relative tumor growth after Selumetinib treatment of two PDX models highly expressing Kindlin-1 (#965 and #138) and 2 PDX models with a low Kindlin-1 expression (#73 and #408).

[0094] FIG. 9A depicts box and whisker plot representing the differential mRNA expression levels of Kindlin-1 in head and neck cancer patients non responding versus head and neck cancer patients responding to Cetuximab monotherapy (p=0.01).

[0095] FIG. 9B depicts box and whisker plot representing the differential disease free survival between patients expressing low versus high levels of Kindlin-1 (p=0.2) in head and neck cancer patients. Mean ±SE represented. Statistical analysis were performed using the Mann-Whitney test.

[0096] FIG. 10 depicts graphs representing cell viability at 48h and 72 h as a function of the drug concentration M, for MDA MB 468 control cells (w) or stably overexpressing Kindlin-1 (k1).

[0097] FIG. 11 depicts box and whisker plot representing the differential mRNA expression levels of Kindlin-1 in patients with a long progression free survival (PFS) versus patients with a short progression free survival (p=0.004) on a cohort of EGFRwt/KRASwt NSCLC patients treated with erlotinib (Byers et al., 2012 cohort). Statistical analysis were performed using the Mann-Whitney test.

[0098] FIG. 12A depicts box and whisker plot representing the differential mRNA expression levels of Kindlin-1 in head and neck cancer patients with short progression free survival versus long progression free survival following Cetuximab treatment (p=0.001), on a cohort of HNSCC patients treated with cetuximab (Bossi et al., 2016 cohort). Mean ±SE represented. Statistical analysis were performed using the Mann-Whitney test.

[0099] FIG. 12B depicts box and whisker plot representing the differential mRNA expression levels of Kindlin-1 in metastatic CRC patients with a progressive disease versus patients with disease control treated with Cetuximab (p=0.044) on a cohort of Metastatic CRC patients treated with cetuximab (Khambata-Ford et al., 2007 cohort). Mean ±SE represented. Statistical analysis were performed using the Mann-Whitney test.

EXAMPLES

Example 1: Materials and Methods

Gene Set Enrichment Analysis (GSEA)

[0100] Microarray expression of cancer cell lines from the Cancer Cell Line Encyclopedia (CCLE) dataset publicly available from cBioPortal (www.cbioportal.org/) were analyzed by GSEA v4.0.2 software (58 breast, 174 lung, 40 pancreas, 24 bladder, 30 head and neck, 44 colon cancer cell lines). Gene sets used for the analysis were obtained from the Molecular Signatures Database (MSigDB 7.0; http://software.broadinstitute.org/gsea/msigdb/). We compared expression profiles of cell lines with a high Kindlin-1 expression versus cell lines with a low Kindlin-1 expression. The optimal cutoff point to categorize cell lines into high versus low Kindlin-1 expression groups was determined performing a hierarchical clustering on Kindlin-1 expression level. GSEA plots showed specific gene sets significantly enriched in cancer cell lines with a high mRNA expression of Kindlin-1.

Cell Culture

[0101] Human breast cancer cell lines DU4475, MDA-MB-231, MDA-MB-468, HCC-1954, HCC-1569, MDA-MB-415, MDA-MB-361, BT549, MDA-MB-436, MCF7 and HS578T; human lung cancer cell lines H727, H358, H1975, A549, H69, H1650 and CALU1 ; and human pancreatic cancer cell lines HPAFII, ASPC1, MIAPACA2 and PANC1 were purchased from ATCC (Manassas,VA,USA), maintained at 37° C. with or without 5% CO.sub.2 and grown in DMEM, RPMI 1640, MEM Alpha, McCoy’s or Leibovitz’s L-15 medium supplemented with 10% FBS and 1% antibiotics (50 .Math.g/mL penicillin, 50 .Math.g/mL streptomycin, 100 .Math.g/mL neomycin).

Western Blotting

[0102] Cells were lysed using RIPA buffer (50 mM Tris-HCl, pH 8; 150 mM NaCl; 0.5% triton; 0.5% deoxycholic acid) containing protease inhibitors (1:1000 orthovanadate, 1:1000 apoprotinine, 1:200 PMSF). Protein extracts were loaded on a polyacrylamide gel, transferred to a nitrocellulose membrane and incubated overnight at 4° C. with primary antibodies for Kindlin-1 (1:1000), pMEK½ (1:1000, 9154P, Cell Signaling, Danvers, MA) or MEK½ (1:1000, 9126S, Cell Signaling). GAPDH (1:500, Clone V18, Santa Cruz Biotechnology) or b-actin (1:16000, Clone AC15, Sigma-Aldrich) were used as loading controls. The signals were detected according to the ECL Western Blotting Analysis System procedure (GE Healthcare, Buckinghamshire, UK).

Human Breast, Lung, Pancreas, Urinary Tract, Head and Neck and Colon Tumors

[0103] Publicly available data set from the TCGA were analyzed for different human cancer types (n= 1084 breast tumors, 566 lung tumors, 184 pancreatic tumors, 411 urinary tract tumors, 488 head and neck tumors and 524 colon tumors obtained from cBioPortal (www.cbioportal.org/). Kindlin-1 and EGFR mRNA expression and/or amplification and the mutational status of KRAS, NRAS, HRAS, EGFR and BRAF were examined for each set of cancers.

[0104] Tumor sections of 62 breast tumors (tissue microarray) and 96 lung adenocarcinomas and adjacent normal breast tissues from patients treated at the Institut Curie were obtained from the Pathology Department of Hospital Curie. For the semi-quantitative analysis of Kindlin-1 and EGFR protein expression, the H-scored method assigned a score of 0-300 to each patient, based on the percentage of cells stained at different intensities. For lung tumors, EGFR and KRAS mutational status had been determined.

[0105] A series of 18 head and neck patients undergoing Cetuximab monotherapy at the Hôpital Curie was also analyzed at the RNA level.

Breast Cancer Patient Derived Xenografts (PDX)

[0106] 27 triple negative breast cancer PDX models were obtained from the Laboratoire d′Investigation Pre-clinique (Institut Curie) as previously described. Informed consent was obtained from patients before xenograft establishment. When tumors reached a volume of 60 to 200 mm.sup.3, mice were randomly assigned to the control or treated groups. Each group of treatment consisted of 3 or 4 mice. Selumetinib was purchased from MedChemExpress. It was administrated orally at a dose of 50 mg/kg 5x/week. Experiments complied with the current laws of France and were approved by Institut Curie ethical committee.

[0107] Tumor growth inhibition (TGI) was calculated using the following formula [1-(V.sub.ft-V.sub.0t)/(V.sub.fc-V.sub.0c))]*100 where V.sub.ft = final volume of the treated group (at the end of the treatment); V.sub.0t = initial volume of the treatment group (at the beginning of the treatment); V.sub.fc= final volume of the control group (at the end of the treatment); V.sub.0c = initial volume of the control group (at the beginning of the treatment).

Statistical Analysis

[0108] Statistical analyses were performed with Prism, version 5 and 8.3.1 (GraphPad Software Inc.). Survival distributions were estimated by the Kaplan Meier method using PASW Statistics (version 18.0; SPSS Inc.)

Results

Kindlin-1 Expression Is Associated With the Activation of EGFR/RAS Pathway In Breast Cancer Cell Lines

[0109] To identify signaling pathways associated with Kindlin-1 overexpression in human breast cancer, we screened the microarray expression of 58 breast cancer cell lines from the Cancer Cell Line Encyclopedia (CCLE) dataset. A gene set enrichment analysis (GSEA) was used to detect the pathways differentially enriched in cell lines with a high Kindlin-1 expression (19 cell lines) versus cell lines with a low Kindlin-1 expression (39 cell lines). GSEA plots demonstrated that EGFR/RAS pathway is enriched in breast cancer cell lines with a high mRNA expression of Kindlin-1. We highlighted several enriched gene sets from the EGFR/RAS pathway: MEK upregulation, EGFR upregulation, KRAS signaling and KRAS dependency.

[0110] Next, we analyzed the correlation between Kindlin-1 expression and the EGFR/RAS altered status in these 58 breast cancer cell lines from CCLE. RNA expression levels of Kindlin-1 were examined in EGFR/RAS-driven breast cancer cells versus non EGFR/RAS-driven cells. Scatter plots (FIG. 1A) showed that high Kindlin-1 mRNA expression is associated to the activation of the EGFR/RAS pathway in breast cancer (p<10.sup.-4).

[0111] To test whether high kindlin-1 expression is associated with the activation of EGFR/RAS pathway at the protein level, we evaluated phosphorylation levels of MEK, one of the key downstream effectors of the EGFR/RAS pathway in a panel of breast cancer cell lines (n=11). We found that Kindlin-1 protein levels highly correlated with the phosphorylation levels of MEK (FIG. 1B).

Kindlin-1 Expression Is Associated With EGFR/RAS Pathway Activation in Lung, Pancreas, Bladder, Head and Neck and Colon Cancers

[0112] Since Kindlin-1 overexpression has also been described in other solid tumors and its up regulation suggested decreasing patient survival in different cancer types. We further wanted to evaluate whether Kindlin-1 could be associated to the EGFR/RAS pathway not only in breast but also in a broader range of epithelial cancers. First, as we did with breast cancer cell lines, the microarray expression of 174 lung, 40 pancreatic, 24 bladder, 30 head and neck and 44 colon cancer cell lines from the Cancer Cell Line Encyclopedia (CCLE) dataset were examined by a gene set enrichment analysis (GSEA) to detect the signaling pathways differentially enriched in cell lines with a high Kindlin-1 expression versus cell lines with a low Kindlin-1 expression. GSEA plots demonstrated that EGFR/RAS pathway is also enriched in broad range of cancer cell lines with a high mRNA expression of Kindlin-1 , including lung, pancreas, bladder, head and neck and colon cancer cell lines.

[0113] We then examined protein expression in several lung and pancreatic cancer cell lines. As we observed in breast cancer cell lines, we found that Kindlin-1 protein levels highly correlated with phosphorylation levels of MEK, indicating an activation of EGFR/RAS pathway specifically in these cell lines expressing high levels of Kindlin-1 (FIG. 2).

Kindlin-1 Expression Is Associated With the Activation of EGFR/RAS Pathway In Human Breast Cancers

[0114] To further determine whether Kindlin-1 expression is associated with the EGFR/RAS pathway in human tumors, we analyzed Kindlin-1 and EGFR transcripts in a cohort of breast cancer patients for whom expression and mutational profiles are publicly available from the TCGA project (n=1084). We found that Kindlin-1 mRNA expression is correlated with EGFR mRNA expression (R.sup.2=0.45, p=0.02) supporting our results in cell lines. In addition, as found in cell lines, Kindlin-1 mRNA expression was associated not only with EGFR upregulation but also with the EGFR amplification and activating mutations in major drivers of the EGFR/RAS signaling axis (KRAS, HRAS, NRAS and BRAF genes) (p<0.0001).

[0115] We next performed an immunohistochemical analysis to evaluate Kindlin-1 and EGFR expression at a protein level in human breast tumors. We analyzed a tissue microarray (TMA) consisting of 62 human breast tumors. We compared the expression of both proteins in the same series of patients. Once again, we found a higher Kindlin-1 expression in those tumors highly expressing EGFR (FIG. 3).

Kindlin-1 Expression Is Associated With the Activation of EGFR/RAS Pathway In Human Lung, Pancreas, Bladder and Head and Neck Cancers

[0116] Next, we examined Kindlin-1 mRNA expression in a series of human lung (n=566), pancreatic (n=184), bladder (n=411), head and neck (n=488) and colon (n=524) tumors using publicly available data from the TCGA project. KRAS, HRAS, NRAS, BRAF activating mutations and EGFR mutations and/or amplification in those tumors were interrogated from the cbioportal database (FIG. 4). We found that Kindlin-1 expression was higher in KRAS-driven lung and pancreatic carcinomas versus the wild type ones. In the case of tumors from the bladder, those tumors with activating mutations of KRAS, HRAS, NRAS or BRAF and EGFR amplifications had increased levels of Kindlin-1 mRNA. Head and neck tumors with EGFR amplification and overexpression also showed higher Kindlin-1 expression levels. Finally, colon tumors with KRAS mutations and EGFR overexpression tended to have higher levels of Kindlin-1, even though the statistical significance was not reached (FIG. 4).

[0117] To corroborate these results, we investigated the expression of Kindlin-1 protein in a series of 96 lung adenocarcinomas with respect to their EGFR and KRAS mutation status. Immunohistochemical analyses showed a higher expression of Kindlin-1 in those tumors mutated for EGFR/RAS pathway.

[0118] Altogether, our findings suggest that Kindlin-1 expression is associated with EGFR/RAS pathway activation in a broad range of epithelial cancers.

Kindlin-1 Expression Identifies EGFR/RAS-Driven Cancer Patients With Worst Survival Outcome

[0119] We then investigated the role of Kindlin-1 expression in patient survival using the TCGA data sets that are large enough to stratify patients based on EGFR and KRAS status. For breast cancers, Kindlin-1 expression was found to discriminate a subset of triple-negative breast cancer tumors, (the only subtype reported as EGFR-driven) with a worst clinical outcome (Log-rank test, p=0.003).

[0120] In the other cancer types, since EGFR overexpression/amplification and KRAS, EGFR, HRAS and BRAF mutations are usually mutually exclusive, we took into account all these drivers to define the EGFR-RAS axis high or low status. Kaplan-Meier curves are represented for each cancer type (FIG. 5). We found that high expression of Kindlin-1 is associated with a decreased metastasis-free survival in patients harboring EGFR-RAS driven tumors. Thus, Kindlin-1 expression identifies RAS/EGFR-driven cancer patients with the worst survival outcome.

Kindlin-1 Expression Is Associated With Sensitivity to EGFR/RAS Pathway Inhibition in Breast Cancer Cells

[0121] We next wanted to evaluate whether Kindlin-1 could be predictive of the response to inhibitors of the EGFR/RAS pathway in breast cancer. IC.sub.50 values of different drugs were obtained from Genomics of Drug Sensitivity in Cancer database (www.cancerrxgene.org/) for the 58 breast cancer cell lines from CCLE. We demonstrated that Kindlin-1 mRNA expression was significantly increased in cell lines more sensitive to MEK inhibitors (Trametinib, p=0.0063; Selumetinib, p=0.0035; CI-1040, p=0.044 and PD0325901, p=0.0034), EGFR inhibitors (Cetuximab, p=0.017; Erlotinib, p=0.015 and Pelitinib, p=0.011), BRAF inhibitor (PLX-4720, p=0.005), and KRAS inhibitor (KRAS Inhibitor-12, p=0.016) compared with resistant cell lines. Moreover, this effect was specific to EGFR/RAS/MAPK inhibitors as there was no correlation between Kindlin-1 mRNA expression and Palbociclib, a CDK inhibitor used in breast cancer therapy (FIG. 6).

Kindlin-1 Expression Is Associated With Sensitivity to EGFR/RAS Pathway Inhibitors in a Broad Range of Epithelial Cancers

[0122] We next evaluated whether Kindlin-1 could be used as a predictive biomarker of the response to inhibitors of the EGFR/RAS pathway not only in breast but also in other epithelial cancers. Microarray expressions of 120 lung, 41 pancreatic, 21 bladder, 54 colon and 12 head and neck cancer cell lines from the CCLE dataset were analyzed using the IC.sub.50 values obtained from Genomics of Drug Sensitivity in Cancer database as we did for breast cancer. We demonstrated that Kindlin-1 mRNA expression was significantly increased in lung cancer cell lines sensitive to MEK inhibitors (Trametinib, p=0.03; PD0325901, p=0.0018) and EGFR inhibitors (AZD3759, p=0.0098; Erlotinib, p=0.02; Gefitinib, p=0.0098 and Osimertinib, p=0.0032); pancreatic cancer cell lines more sensitive to MEK inhibitors (Trametinib, p=0.046; Selumetinib, p=0.07; PD0325901, p=0.05) and BRAF inhibitor Dabrafenib (p=0.07); urinary tract cancer cell lines more sensitive to MEK inhibitors (Trametinib, p=0.014 and Refametinib, p=0.038) and colon cancer cell lines more sensitive to MEK inhibitors (Trametinib, p=0.03; PD0325901, p=0.02 and Refametinib, p=0.001). Kindlin-1 mRNA expression had also a tendency to be increased in head and neck cancer cell lines more sensitive to EGFR inhibitors (Cetuximab, p=0.09 and Erlotinib p=0.07) and to the MEK inhibitor CI-1040 (p=0.1) (FIG. 7).

Kindlin-1 Expression Is Associated With Sensitivity to Selumetinib in Breast Cancer in Vivo

[0123] To further investigate whether Kindlin-1 expression could have a biomarker relevance, a cohort of 27 triple negative breast cancer patient derived xenografts (PDX) was used to test the correlation between Kindlin-1 expression and the response to MEK inhibitors. Kindlin-1 protein expression was assessed by western blots (FIG. 8A). The PDX were then treated with 50 mg/Kg of the Selumetinib, 5 days per week during 5 weeks. Tumor growth inhibition (TGI expressed in %) was evaluated with regard to the Kindlin-1 expression in each PDX model (FIG. 8B). We found a statistically significant correlation (Pearson r= 0.58; p=0.02) between Kindlin-1 protein expression and tumor growth inhibition (FIG. 8B). Selumetinib treatment had an important effect inhibiting tumor growth in PDXs highly expressing Kindlin-1 protein, not observed in PDXs with a lower Kindlin-1 expression (FIG. 8C).

Kindlin-1 Expression Is Predictive of Cetuximab Response in Head and Neck Cancer Patients

[0124] To further investigate the relevance of Kindlin-1 as a sensitive biomarker of EGFR/RAS pathway inhibitors response, a cohort of 18 head and neck cancer patients treated with Cetuximab monotherapy at the Curie Hospital was analyzed. We found that those patients with a good response to the treatment had higher Kindlin-1 expression levels than those that were not well responding, thus wherein the cancer progressed (FIG. 9A). Moreover, there was a tendency in patients with higher Kindlin-1 expression towards a delay in the progression of the disease (FIG. 9B).

Example 2

Materials and Methods

Co-Immunoprecipitation

[0125] Cells were lysed using NP40 buffer (50 mM Tris-HCl, pH 7.5; 150 mM NaCl; 0.5% NP40) containing protease inhibitors (1:1000 orthovanadate, 1:1000 apoprotinine, 1:200 PMSF). Protein extracts were incubated with 1 mg antibodies for Kindlin-1, EGFR or normal rabbit IgG (GTX35035, GeneTex, Irvine, CA, USA) and 10 ml Sepharose Protein A beads (Rockland, Limerick, PA, USA) with rotation at 4° C. overnight. Beads were washed with NP40 buffer three times and immunoprecipitates were resolved by western blotting as previously described.

Immunofluorescence

[0126] Cells were plated on fibronectin coated coverslips. After serum starvation overnight, cells were treated with 100 ng/ml EGF for the indicated time, fixed in 4% paraformaldehyde, permeabilized and immunostained with primary antibodies (anti-Kindlin-1, 1:700; anti-EGFR, 1:250, A-10, Santa Cruz Biotechnology, Santa Cruz, CA, USA) followed by alexa fluor-conjugated secondary antibodies (A11031 and A11034, Invitrogen, Carlsbad, CA, USA). Cells were then counterstained with DAPI and imaged with the fluorescence Eclipse Ti microscope from Nikon (Melville, NY, USA).

MT Cell Viability Assay

[0127] 15 000 cells were seeded in a 96-well plate and treated with Cetuximab (Erbitux 5 mg/ml, Merck Europe). After 48 and 72 hours cell viability was assessed using the CellTiter-Glo 2.0 Assay according to the instructions of the manufacturer (Promega, Madison, WI, USA). Luminiscence was recorded on a 96-well microplate reader (Promega Glomax Discover, Promega).

Human Tumors

[0128] Raw data from different studies deposited in the GEO repository (http://www.ncbi.nlm.nih.gov/geo/) was also analyzed: 20 NSCL patients treated with Erlotinib (GEO accession number GSE33072), 40 HNSCC patients treated with Cetuximab (GEO accession number GSE65021) and 68 metastatic CRC patients treated with Cetuximab (GEO accession number GSE5851).

Results

Kindlin-1 Directly Interacts and Colocalizes With EGFR in Cancer Cells and Has an Impact on the Activation of EGFR Pathway

[0129] Due to the correlation found between Kindlin-1 and EGFR at mRNA and protein levels, it was hypothesized that these two proteins might interact in cancer cells. To address this question, the inventors chose BT20 triple negative breast cancer cells which present a high expression of Kindlin-1 and EGFR. Possible interaction at the endogenous level was examined by co-immunoprecipitation experiments. The inventors found that Kindlin-1 is able to immunoprecipitate EGFR and vice versa. Then, to further characterize this association, the subcellular localization of both proteins was examined by immunofluorescence. In absence of EGF stimulation, EGFR mainly localizes at the plasma membrane of BT20 cells while Kindlin-1 exhibits a dot-like staining predominantly at the perinuclear region. However, after 15 min EGF stimulation, promoting EGFR activation, EGFR is not anymore present at the cell surface. The receptor is internalized and colocalizes with Kindlin-1 in some ventral structures that should be further characterized.

[0130] As endogenous Kindlin-1 and EGFR proteins interact in cancer cells, the inventors wanted to study if Kindlin-1 could have an impact in the activation of the EGFR pathway. Therefore Kindlin-1 was depleted in different cancer cells lines (MDA-MB-468, H1975, BT20 and H358) and the activation of EGFR pathway checked using as a readout, the levels of phosphorylated ERK, one of the key effectors of the pathway. The inventors found that silencing Kindlin-1 drastically decreased the phosphorylation levels of ERK suggesting the involvement of Kindlin-1 in the EGFR pathway.

Kindlin-1 Overexpression Renders Sensitive Cells More Resistant to Cetuximab

[0131] The possibility that Kindlin-1 ectopic expression could reverse the sensitivity to the EGFR inhibitor Cetuximab was tested. The inventors hypothesized that ectopic expression of Kindlin-1 could bypass the action of Cetuximab, maintaining the EGFR pathway activated and therefore rendering sensitive cells more resistant. To confirm our hypothesis, MDA-MB-468 cells stably overexpressing Kindlin-1 was generated (named k1). Control- and Kindlin-1-expressing MDA-MB-468 cells were treated with increasing concentrations of Cetuximab (1-20 .Math.M) and a MT cell viability assay was performed after 48 h and 72 h of treatment. As expected, cells expressing ectopic Kindlin-1 became more resistant to the drug as determined by the calculation of the half maximal inhibitory concentration or IC50 (control cells: 7.3 .Math.M (48 h); 6.1 .Math.M (72 h); Kindlin-1 cells: 14.8 .Math.M (48h); 11.3 .Math.M (72 h)) ( see FIG. 10).

Kindlin-1 Expression Predicts Cetuximab/Erlotinib Response in Head and Neck, Lung And Colorectal Cancer Patients

[0132] To further investigate the relevance of Kindlin-1 as a sensitive biomarker of the response to EGFR inhibitors, raw data from different studies deposited in the GEO repository were also studied: 20 NSCL patients treated with Erlotinib as well as 40 HNSCC and 68 metastatic CRC patients treated with Cetuximab were analysed. In all the three cohorts, a higher Kindlin-1 mRNA expression was found in patients with a better response and a longer patient relapse-free survival (FIGS. 11 and 12).