EP2473622A1

EP2473622A1 - Prognostic molecular signature of sarcomas, and uses thereof

Info

Publication number: EP2473622A1
Application number: EP10718241A
Authority: EP
Inventors: Frédéric Chibon; Jean-Michel Coindre; Alain Aurias
Original assignee: Universite Victor Segalen Bordeaux 2; Institut Curie; Institut Bergonie; Universite Bordeaux Segalen
Current assignee: Universite Victor Segalen Bordeaux 2; Institut Curie; Institut Bergonie; Universite Bordeaux Segalen
Priority date: 2009-04-22
Filing date: 2010-04-21
Publication date: 2012-07-11
Also published as: US20130065772A1; JP5865241B2; WO2010122243A1; JP2012524537A; CA2759417C; CA2759417A1; US9708666B2; FR2944805B1; FR2944805A1

Abstract

The present invention relates to a prognostic molecular signature for sarcomas, and in particular for genetically complex sarcomas, and to the use thereof for predicting the overall survival of a patient, particularly for predicting the appearance of metastases but also for differentiating prognostically separate sub-groups within a group of tumors having the same histological grade. Said molecular signature can also be used for testing the effectiveness of a treatment or for producing novel therapeutic agents.

Description

PROGNOSTIC MOLECULAR SIGNATURE OF SARCOMES AND

USES

FIELD OF THE INVENTION The present invention relates to a prognostic molecular signature of sarcomas, particularly sarcomas with complex genetics, and its use for predicting metastasis-free survival and overall survival of patients with sarcoma.

It has many applications, particularly in the field of prognosis or diagnosis of sarcoma or to monitor the treatment of patients with sarcoma.

PRIOR ART

Soft tissue sarcomas (STS) in adults are rare and heterogeneous in terms of location, histology, molecular abnormalities, and prognosis. Low-differentiated STS are the most common malignant tumors in adults, representing approximately 50% of pathological diagnoses, and mainly include sarcomas with a complex karyotype, namely leiomyosarcomas (LMS), undifferentiated sarcomas (US), or malignant histiocytofibromas. (MFH), and dedifferentiated liposarcomas (DD-LPS) (FLETCHER et al., World Hearth Organization (WHO), classification of tumors, Pathology and genetics of tumors of soft tissue and bone, Lyon, IARC Press, 2002). At the genetic level, poorly differentiated STS can be divided into two main groups, a group with a complex genomic profile (80%) consisting essentially of US, LMS, pleomorphic rhabdomyosarcomas and pleomorphic liposarcomas, associated with very complex profiles, but recurrent, genomic imbalances (IDBAIH et al., Invest Lab., 85 (2): 176-181, 2005; CHIBON et al., Cancer Genet Cytogenet., 141 (1): 75-78, 2003; DERRE and al., Lab. Invest., 81 (2): 211-215, 2001), and a second group with a simple genetic profile (20%) based on a high level of limited amplifications and consisting exclusively of DD-LPS ( CHIBON et al., Cancer Genet, Cytogenet., 139 (1): 24-29, 2002, COINDRE et al., Pathol Mod, 16 (3): 256-262, 2003). STS are aggressive tumors capable of local and metastatic relapse. Patients with such tumors usually have a poor prognosis, 40 to 50% eventually develop distant metastases, mainly in the lungs, usually within 5 years of diagnosis (WEITZ et al., Clin J. One, 21 (14)). ): 2719-2725, 2003. ZAGARS et al., Cancer, 97 (10): 2530-2543, 2003).

The clinical treatment of STS consists mainly of surgical resection, with adjunctive therapies whose duration and nature depend on surgical margins, tumor histology and histological grade. However, the benefit of Adjunctive therapies like chemotherapy is currently disputed although recent studies tend to demonstrate an effect on local and distant relapses (SMAC, Lancet, 350 (9092): 1647-1654, 1997; FRUSTACI et al., J. Clin Oncol., 19 (5): 1238-1247, 2001; PERVAIZ et al., Cancer, 113 (3): 573-581, 2008). Nevertheless, the effectiveness of chemotherapy is marginal (from 3 to 10% depending on the endpoint, PERVAIZ et al., 2008, cited above); this could result from the selection of patients for whom tumor malignancy is assessed by histological grade. In addition, the management of patients depends essentially on the stage of the disease. Although it provides valuable information on the clinical course of certain types of sarcomas, histological typing has limited predictive value for other types of sarcomas, including unclassified, poorly differentiated and non-translocated sarcomas. To increase the predictive value of histology in terms of prognosis, several grade systems have been generated (BRODERS et al., GynecoL Surg, Obstet., 69: 267-280, 1939, RUSSELL et al., Cancer, 40). (4): 1562-1570, 1977; MARKHEDE et al., Cancer, 49 (8): 1721-1733, 1982; TROJANI et al., Int. J. Cancer, 33 (1): 37-42, 1984; COSTA et al., Cancer, 53 (3): 530-541, 1984). Of these, the National Cancer Institute (NCI) systems (COSTA et al., 1984, supra) and the National Federation of Cancer Control Centers (FNCLCC) (TROJANI et al., 1984, mentioned above) have been widely used even though the second system makes it possible to slightly increase the prediction capacity of distant metastases and has therefore been considered as "gold standard" (GUILLOU et al., J. Clin. 1): 350-362, 1997).

Until now, the histological grade is the best predictor of survival without metastasis and overall survival. The most effective FNCLCC rank has been established for more than 20 years and is still the most commonly used system. It is based on semi-quantitative evaluation of tumor differentiation, necrosis, and mitotic index. However, this system has several limitations: its reproducibility from one pathologist to another is not perfect, it does not apply to all types of sarcomas (COINDRE et al., Cancer, 91 (10): 1914- 1926, 2001) and is not informative for grade 2 cases (accounting for about 40% of cases). However, despite these limitations, for more than 20 years, no study has provided prognostic criteria that could replace this histological grade system.

Over the past decade, prognostic molecular signatures have emerged in a growing number of pathologies. To date, the best example of molecular signature is certainly that of breast cancer in which a predictive expression signature of metastatic relapse was established in 2002 and validated the same year by the same team on an independent group of 295 tumors. (VAN'T VEER et al., Nature, 415 (6871): 530-536, 2002, VAN of VIJVER et al., N. Engl J. Med., 347 (25): 1999-2009, 2002). So far in the field of sarcomas, expression profiles have mainly been established in order to identify new diagnostic markers or to better understand the oncogenesis of these tumors in relation to tumor differentiation (NIELSEN et al. , Lancet, 359 (9314): 1301-1307, 2002, BAIRD et al., Cancer Res., 65 (20): 9226-9235, 2005; FRITZ et al., Cancer Res., 62 (11): 2993- 2998, 2002; MATUSHANSKY et al., Am. J. Pathol., 172 (4): 1069-1080, 2008; SEGAL et al., Am. J. Pathol., 163 (2): 691-700, 2003; LEE et al., J. Cancer, 88 (4): 510-515, 2003; NAKAYAMA et al., Mod Pathol., (7): 749-759, 2007; SINGER et al., Cancer Res., 67 (14): 6626-6636, 2007). Only two studies, on 30 leiomyosarcomas (LEE et al., Cancer Res., 64 (20): 7201-7204, 2004) and on 89 pleomorphic sarcomas (FRANCIS et al., BMC Genomics, 8: 73, 2007) propose a prognostic molecular signature. But these two signatures are composed of many genes (respectively 335 and 244, respectively) without a clear biological link between them. In addition, these two signatures have been established on a relatively small number of specific sarcoma subtypes giving relatively weak meanings. Finally, to date, these two signatures have not been compared to the FNCLCC grade system and have not yet been validated on an independent group, thereby limiting their clinical utility.

In the field of sarcomas, it should be noted that the number of studies seeking to correlate molecular alterations to the prognosis is necessarily limited because of the difficulty of obtaining a homogeneous study group of fully annotated tumors. So far, no clear and proven correlation has been established between genetic profiling and survival without metastases.

As a result, tumor progression remains difficult to predict within a group of sarcomas, and the treatments are not as appropriate as they could be. Therefore, there is a definite need to improve the prognosis and diagnosis of sarcomas and therefore to ensure better clinical follow-up of patients.

It is therefore an object of the present invention to provide a more efficient, reliable and reproducible grade system to overcome the disadvantages of the prior art. It is also another object of the present invention to provide the means and kits for implementing such a grade system.

SUMMARY OF THE INVENTION

Assuming that the FNCLCC rank system could represent a phenotypic summary of genomic alterations, the inventors of the present invention have discovered quite unexpectedly that establishing a molecular profile using Emerging technologies, such as DNA microarrays, could allow the identification of alterations / genes causing tumor aggressiveness, with the consequent possibility of defining a more efficient grade-based system. on molecular alterations; which leads to a major breakthrough in the field of sarcoma analysis.

While the number of studies seeking to correlate molecular alterations to the prognosis is limited because of the difficulty of obtaining a homogeneous study group of fully annotated tumors, the inventors have initiated an original project to determine genomic profiles and from 183 primary tumors with complex genetics, untreated and completely annotated, referenced in the GSF database (French Sarcoma Group), an integral part of the European Conticabase database (www.conticabase.org). Cluster analysis was used to identify the molecular alterations associated with the patient's clinical outcome.

This approach, illustrated in the experimental part below, made it possible to identify a particular set of genes, called "pool" or "molecular signature", associated with the complexity of the genome, the tumor aggressiveness, and whose profile of expression has made it possible to establish a reliable prognosis of patients with sarcoma, in particular to predict the occurrence of metastases, but also to distinguish within a group of patients with sarcoma of the same histological grade, -groups with significantly different prognoses.

The present invention therefore relates, in the first place, to a pool of polynucleotides comprising at least two polynucleotides chosen from the polynucleotide sequences SEQ ID NO: 1 to SEQ ID NO: 67. In other words, the polynucleotide pool of the invention may include 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24 , 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49 , 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, or 67 polynucleotides selected from the polynucleotide sequences SEQ ID NO: 1 to SEQ ID NO: 67.

Analysis of the 67 genes identified (SEQ ID NO: 1 to SEQ ID NO: 67) by the Gene Ontology (GO) database further showed that all were involved in the same biological process, ie say the control of the integrity of the chromosome.

The inventors have further demonstrated that these genes can be divided into 5 main groups according to their role in mitosis: control point of mitosis and the cell cycle (12 genes, SEQ ID NOs: 1-12); chromosome biogenesis, condensation, alignment and segregation (26 genes, SEQ ID NOs: 13-38); mitotic spindle and centrosome (12 genes, SEQ ID NOs: 39-50); microtubule engine, kinesin complex (8 genes, SEQ ID NOs: 51-58), and cytokinesis (4 genes, SEQ ID NOs: 59-62); among the last 5 genes grouped according to the experimental results (SEQ ID NOs: 63-67), 3 are known to be involved in chromosomal instability (SEQ ID NOs: 63-65) and 2 are associated with histological grade according to US Pat. study (SEQ ID NOs: 66 and 67). Table 1 below shows the names of each of the genes, their distribution into five main groups, and their respective sequences (references GenBank and SEQ ID NO :).

Table 1

Advantageously, the polynucleotide pool may comprise the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58, SEQ ID NO: 24.

Advantageously, the polynucleotide pool can comprise the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID

NO: 58, SEQ ID NO: 24 and at least one gene whose sequence is selected from the 62 other gene sequences identified within the scope of the invention. In other words, the polynucleotide pool may comprise the polynucleotides of SEQ ID NO:

10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58, SEQ ID NO: 24 and at least one polynucleotide whose sequence is selected from the sequences SEQ ID NO: 1, SEQ ID

NO: 2, SEQ ID NO: 4 to SEQ ID NO: 9, SEQ ID NO: 11 to SEQ ID NO: 23, SEQ ID

NO: 25 to SEQ ID NO: 46, SEQ ID NO: 48 to SEQ ID NO: 57, SEQ ID NO: 59 to SEQ

ID NO: 67.

Alternatively, the polynucleotide pool may consist of polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID

NO: 58 and SEQ ID NO: 24. In other words, the polynucleotide pool may comprise only the polynucleotides of SEQ ID NO: 10, SEQ ID NO:

3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24.

Alternatively, the polynucleotide pool of the invention may consist of polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID

NO: 47, SEQ ID NO: 58, SEQ ID NO: 24 and at least one gene whose sequence is selected from the 62 other gene sequences identified within the scope of the invention. In other words, the polynucleotide pool can consist only of the polynucleotides of sequences SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24 and at least one polynucleotide whose sequence is chosen from the sequences SEQ ID NO : 1, SEQ ID NO: 2, SEQ ID NO: 4 to SEQ ID NO: 9, SEQ ID NO: 11 to SEQ ID NO: 23, SEQ ID NO: 25 to SEQ ID NO: 46, SEQ ID NO: 48 at SEQ ID NO: 57, SEQ ID NO: 59 to SEQ ID NO: 67.

According to another embodiment of the present invention, the polynucleotide pool of the invention may comprise at least one polynucleotide selected from each of the following sets of polynucleotides:

Set 1: SEQ ID NO: 1 to SEQ ID NO: 12; Set 2: SEQ ID NO: 13 to SEQ ID NO: 38;

Set 3: SEQ ID NO: 39 to SEQ ID NO: 50;

Set 4: SEQ ID NO: 51 to SEQ ID NO: 58, and SEQ ID NO: 59 to SEQ ID NO: 62;

Set 5: SEQ ID NO: 63 to SEQ ID NO: 65, and SEQ ID NO: 66 to SEQ ID NO: 67.

According to another embodiment of the present invention, the polynucleotide pool of the present invention may be selected from Sets 1 to 5. In other words, the pool of at least two polynucleotides may be made up entirely or part of Set 1, Set 2, Set 3, Set 4 or Set 5. In other words, the pool of the present invention may consist wholly or in part of Set 1, or in whole or in part part of set 2, or part or all of set 3, or all or part of set 4, or all or part of set 5.

According to another embodiment of the present invention, the polynucleotide pool can comprise the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24 sequences. and at least one polynucleotide selected from set 5. This polynucleotide pool may further comprise at least one of the other genes identified within the scope of the invention.

According to another embodiment of the present invention, the polynucleotide pool may consist of polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24 and at least one polynucleotide selected from set 5.

According to another embodiment of the present invention, the polynucleotide pool of the invention comprises the polynucleotides of sequences SEQ ID NO: 1 to SEQ ID NO: 67. It may be for example a pool consisting of sequences SEQ ID NO: 1 to SEQ ID NO: 67. Whatever the embodiment of the invention, advantageously, the polynucleotide pool may advantageously comprise at most 10 polynucleotides. It may be for example a pool comprising at most 10 polynucleotides, comprising the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: And at least one polynucleotide of sequence selected from SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 4 to SEQ ID NO: 9, SEQ ID NO: 11 to SEQ ID NO: 23, SEQ ID NO: SEQ ID NO: 46, SEQ ID NO: 48 to SEQ ID NO: 57, SEQ ID NO: 59 to SEQ ID NO: 67.

Whatever the embodiment of the present invention, advantageously the pool of polynucleotides of the invention is immobilized on a support, for example a solid support or a liquid carrier. In the case where the support is a liquid support, it may comprise beads on which the nucleic acids are fixed. The liquid medium may be a cell culture supernatant, serum, plasma, this list not being limiting. This may be for example the support implemented in Luminex® technology. In the case where the support is a solid support, it is preferably selected from the group comprising a nylon membrane, a nitrocellulose membrane, a glass slide, glass beads, a membrane on a glass support or a chip of silicon, a plastic support. Particularly preferably, the solid support may be a nucleic acid chip, for example a DNA chip, (also called a gene chip, biochip, expression chip). Such chips allow to quantitatively measure an expression variation (differential expression) of two or more polynucleotides of the polynucleotide pool of the invention between (i) 2 experimental conditions: generally a reference condition and a pathological condition or ( ii) several tumors to determine a mean of expression, according to which the tumors can be classified with respect to each other. By way of non-limiting example, it may be an Affimetrix® DNA chip, or a DNA chip from Agilent Technologies.

The genes identified by the present inventors, all of which are involved in the same biological process, may furthermore be potential targets for new therapeutic approaches targeting the early stage of metastatic potential acquisition. In addition, a vital prognosis of the patients on the basis of the expression profile of these genes can be realized very early, or even during the initial diagnosis.

Thus according to a particular embodiment of the present invention, the polynucleotide pool of the invention can be used for detecting, prognosticating, diagnosing soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST ), or to monitor the treatment of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST). According to another particular embodiment of the present invention, the polynucleotide pool of the invention can be used to obtain a compound for the treatment of soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST). In order to identify, generally from the expression microarray data, the expression profile associated with a prognostic group, two main approaches can be used, the supervised top-down approach for select genes directly correlated with poor prognosis (VAN'T VEER et al., 2002, SOTIRIOU et al., J. Natl Cancer Inst, 98 (4): 262-272, 2006) and the approach supervised bottom-up by which the expression profiles associated with a particular biological phenotype are first identified and subsequently correlated with a clinical result (SOTIRIOU et al., N. Engl J. Med., 360 (8) 790-800, 2009). In the context of the present invention, the second "bottom-up" approach has been applied in the sense that the tumor expression profiles have been compared as a function of the biological phenotypes (chromosomal instability, genomic complexity and histological grades), but instead of direct gene selection, biological pathways particularly relevant for the phenotypes tested were first identified and then the genes significantly involved in these pathways were identified. This selection of biological pathway (and not of genes) is the important step which led to the happy results of the present invention in a heterogeneous group such as that of non-translocated sarcomas, and moreover in different types of tumors such as GISTs (gastrointestinal stromal tumors) and breast cancers.

The present invention thus also relates to an in vitro method for selecting a pool of polynucleotides, for example those of the invention, comprising the following steps: a) providing tumor biological samples from patients suffering from sarcoma of soft tissue (STS) or gastrointestinal stromal tumor (GIST); b) detecting and / or quantifying each of the polynucleotides, separately in each of the biological tumor samples; c) comparing the expression profile of the polynucleotide pools obtained in step c) with respect to a biological phenotype, preferably of chromosomal instability, genomic complexity or histological grade, for each of the biological tumor samples; d) selecting the biologically statistically significant (p <10 ^"5) for the tested phenotype; e) select the polynucleotides significantly involved in this biological pathway, the expression of which is indicative of the probability of occurrence of metastases.

By "expression profile" is meant the totality of the results obtained when the expression of a set of polynucleotides is determined. Such a profile facilitates the use of quantitative statistical analytical techniques and allows a quick visual comparison of the results. Preferably, such a profile is obtained on a solid support, such as a DNA chip.

By "biological phenotype", within the meaning of the present invention, is meant the manifestation of a genetic status, ie the set of observable characteristics characterizing a sample from a patient suffering from STS or GIST, who reflect the expression of the information carried by the chromosomes (the genotype).

By "chromosomal instability", within the meaning of the present invention, is meant clonal or non-clonal rearrangements. This instability leads to losses and gains in chromosome arms and unbalanced chromosome rearrangements. The instability of chromosomes inside the nucleus of an individual's cells makes them more vulnerable in terms of neoplasia (the occurrence of cancer). It is in the tumor cells that we find this instability.

For the purposes of the present invention, the term "genomic complexity" is intended to mean the number of imbalances and the nature of the chromosomal fragments involved.

For the purposes of the present invention, the term "histological grade" means a consensual indicator of tumor proliferation, the risk of metastases and the response to adjuvant therapy (chemotherapy). Histological or tumor grade is a decision-making factor for the treatment of a tumor. It is determined by the histological analysis of the tumor and the grading system used is for example that of the FNCLCC. This system adopted by the National Federation of Centers for the Fight Against Cancer (FNCLCC) refers to the following three characteristics:

The histological grade of FNCLCC soft tissue tumors is the sum of the 3 scores "Differentiation", "Mitotic Index" and "Tumor Necrosis": Grade 1 (total score of 2 or 3), Grade 2 (total score of 4 or 5), and Grade 3 (total score of 6 to 8).

The present invention also relates to an in vitro method for the analysis of soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST), said method comprising determining the level of expression of a polynucleotide pool according to the invention in a tumor biological sample.

For the purposes of the present invention, the term "tumor biological sample" is intended to mean a tissue sample optionally derived from (i) a primary tumor (ii) from the center of a tumor (iii) a site in the tumor other than that the center and (iv) any tumor located outside the tumor tissue per se of a patient with STS. Said biological tumor sample may come for example from a surgical procedure or a tumor resection performed on the patient's STS, a biopsy where a part of the tumor tissue is collected from the patient's STS for later analysis; a blood sample, for example whole blood, plasma or serum, containing tumor cells derived from the primary tumor or tumor proteins produced by the tumor cells derived from the primary tumor.

The level of expression of a polynucleotide pool of the present invention can be determined by any method known in the art. For example, the level of expression of at least two polynucleotides involved in the molecular signature of the invention in samples obtained from patients with STS can be determined by measuring the level of mRNA corresponding to the polynucleotide and / or the protein encoded by the polynucleotide. The RNA can be isolated from the samples by methods well known to those skilled in the art, for example by that described in AUSUBEL et al. (Curr Mol Mol Biol., 1: 4.1.1-4.2.9 and 4.5.1-4.5.3, John Wiley & Sons, Inc., 1996). The methods for detecting the level of mRNA expression that can be used to practice the present invention are well known in the art and include, but are not limited to, expression chips, northern blotting, PCR. RT-PCR, RT-PCT with Taqman probes or micro-fluid maps, and, in general, hybridization techniques (ie non-covalent bonds of two single-stranded polynucleotides, wholly complementary or sufficiently complementary to hybridize with each other, and form a double-stranded structure).

Advantageously, when the polynucleotide pool comprises at most 10 polynucleotides, the level of expression of a polynucleotide pool of the present invention can be determined by routine quantitative PCR. It may further be possible to use RNAs derived from paraffin blocks containing tissue or organ samples, or biological samples.

According to the invention, a particularly efficient method for detecting the level of mRNA transcripts expressed from a plurality of the polynucleotides described involves hybridization of labeled mRNA to an oligonucleotide chip (also called a DNA chip, chip with genes, expression chips). Such a method makes it possible to simultaneously determine the level of transcription of a plurality of polynucleotides to generate polynucleotide expression profiles. The oligonucleotides used in this hybridization method are generally fixed on a support, for example a solid support or a liquid support. In the case where the support is a liquid support, it may comprise beads on which the nucleic acids are fixed. The liquid medium may be a cell culture supernatant, serum, plasma, this list not being limiting. This may be for example the support implemented in Luminex® technology. Examples of solid carriers include, but are not limited to, membranes, filters, slides, paper, nylon, fibers, magnetic or non-magnetic beads, gels, polymers and any solid carrier known to the art. skilled person. Any solid support on which oligonucleotides can be immobilized, either directly or indirectly, either covalently or non-covalently, can be used. A particularly advantageous solid support consists of a nucleic acid chip, in particular a DNA chip. These chips contain a particular oligonucleotide probe at a predefined location of the chip. Each predefined location may contain more than one molecule of the particular probe. Because the oligonucleotides are at specific locations on the support, the hybridization patterns and intensities (which together form a unique expression pattern) can be interpreted in terms of the level of expression of particular polynucleotides.

The oligonucleotide probes are preferably of sufficient length to specifically hybridize only to complementary transcripts of the polynucleotides of the invention. For the purposes of the present invention, the term "oligonucleotides" is intended to mean a single-stranded nucleic acid. Generally oligonucleotide probes consist of 16-20 nucleotides, and in some cases up to 25 nucleotides, or even up to 500 nucleotides or more. Once the probes are contacted with mRNA or a copy of the cDNA, the presence of mRNA or hybrid cDNA is detected by methods known in the art. For example, the oligonucleotide probes are labeled with one or more markers to enable detection of hybridized probe / target polynucleotide complexes. The markers may comprise compositions detectable by spectroscopic, biochemical, photochemical, bioelectronic, immunochemical, electrical, optical or chemical means. Examples of labels include, but are not limited to, radioisotopes, chemiluminescent compounds, labeled binding proteins, heavy metal atoms, spectroscopic markers, such as fluorescent and dye labels, bound enzymes, mass spectrometry and magnetic markings. For example, it is Cy ³ / Cy ⁵ or Alexa labeling for biochips, FAM (6-carboxyfluorescein) or TAMRA (6-carboxy-tetramethyl-rhodamine) labeling for Taqman probes. Oligonucleotide probe chips for expression monitoring can be prepared and used according to methods well known in the art, as described for example in LOCKHART et al. (Nature Biotechnol., 14: 1675-1680, 1996, McGALL et al., Proc Natl Acad Sci USA 93: 13555-13460, 1996 US 6,040,138 Such biochips are commercially available from for example, from Affimétrix (Santa Clara, California).

It is also possible to detect the expression of a protein encoded by two or more of the polynucleotides involved in the molecular signature of the invention. This can be accomplished by methods well known in the art, such as, for example, the use of a probe which is detectably labeled, or which can be subsequently labeled. Generally, the probe is an antibody that recognizes the expressed protein. The expression level of the protein in the sample is then determined by an immunoassay method using the antibodies, for example dot blotting, western blotting, ELISA, immunohistochemistry, FACS, etc.

According to a particular mode of implementation, the method of the invention makes it possible to establish the prognosis of a patient suffering from an STS or a GIST, in particular makes it possible to determine the risk of / predicting the - occurrence of metastases.

By "predicting the occurrence of metastases", within the meaning of the present invention, it is intended to determine a relative value making it possible to quantify the probability of the occurrence of metastases of one or more tissues or organs, in a patient suffering from STS or a GIST. Preferably, the prediction of the occurrence of metastases is expressed by a statistical value, including a p value, calculated from the expression values obtained for each of the polynucleotides tested. According to another particular mode of implementation, the method of the invention makes it possible to establish the prognosis of a patient suffering from an STS or a GIST, in particular to distinguish subgroups of good or bad prognosis from a group of soft tissue sarcomas (STS) or gastrointestinal stromal tumor (GIST) initially considered to be of the same histologic grade.

For the purposes of the present invention, the term "good prognosis" refers to the indication of patients not likely to relapse, that is to say the occurrence of metastases, during their treatment or in the following 5 to 6 years. their treatment, a survival with no significantly long-term metastases. Thus in the context of the present invention; STS or GIST patients may be considered to belong to a "good prognosis" subgroup when they under-express the genes of the molecular signature of the invention and are prone to developing metastases in less than 20% of all types of sarcomas, and in particular in none of the GIST cases. On the other hand, the term "poor prognosis" refers to patients who may have a relapse (metastases) during treatment or within 5 to 6 years of treatment. Thus, in the context of the present invention, patients with STS or GIST can be considered to belong to a subgroup of poor prognosis when they over-express the molecular signature genes and are subject to develop metastases in at least 50% of cases. Advantageously, the determination of the level of expression of the polynucleotide pool in the method of the invention is carried out on a nucleic acid chip, also called a biochip, DNA chips, gene chip, expression chip. Such chips make it possible to quantitatively measure and rapidly visualize a variation in the level of expression, or differential expression, of two or more polynucleotides between (i) 2 experimental conditions, for example a reference experimental condition and a pathological one, at from a biological sample of a patient or (ii) several tumors to determine a mean of expression, according to which the tumors can be classified with respect to each other. As a non-limiting example, Affimetrix ™ DNA chips or DNA chips from Agilent Technologies can be used.

According to a particular mode of implementation, the method of the invention can be used to detect, prognose, diagnose a soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST), or to monitor the treatment of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST), comprising performing a method of the invention on the nucleic acids of a biological sample said patient.

The present invention also relates to an in vitro method for predicting the occurrence of metastases in a patient with tissue sarcoma soft tissue (STS) or a gastrointestinal stromal tumor (GIST) comprising the following steps: a) providing a previously collected tumor biological sample of said patient to be tested; b) determining in said tumor biological sample the level of expression of a pool of polynucleotides of the invention; c) comparing the level of expression obtained in step b) with the level of expression of the same polynucleotide pool measured in a biological control sample; a deregulation of the expression level of the oligonucleotide pool relative to its corresponding expression level measured in a biological control sample being predictive of the occurrence of metastasis.

By "deregulation of the level of expression" is meant the over-expression or the under-expression of two or more polynucleotides of a polynucleotide pool according to the invention measured in a tumor biological sample of a patient suffering from a STS or test GIST, relative to the corresponding expression measured in a biological control sample as defined below. In particular, a higher level of expression in the tumor biological sample of a patient with a STS or a GIST to be tested compared to that of a biological control sample is the indication of a patient susceptible to develop metastases, which is considered an indication of a poor prognosis. Conversely, a lower level of expression in the tumor biological sample of a patient with a STS or a GIST to be tested compared to that of a biological control sample is the indication of a non-patient. likely to develop metastases, that is to say assimilated to the indication of a good prognosis. For the purposes of the present invention, the term "biological control sample" is intended to mean (i) a tissue sample derived from a tumor of another patient suffering from STS or GIST than that to be tested, or (ii) a tissue sample of a healthy subject, namely an individual who has no pathology or pathological symptoms diagnosed by a physician. Thus the tumors can be classified with respect to each other, depending on the level of expression of the genes of the molecular signature of the invention in each case.

The present invention also relates to an in vitro method for assessing the prognosis of a patient with soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST), comprising the steps of: a ) provide a pre-collected tumor biological sample from the patient with STS or gastrointestinal stromal tumor (GIST) to be tested; b) determining in said tumor biological sample the level of expression of a pool of polynucleotides of the invention; c) comparing the level of expression obtained in step b) with the level of expression of the same polynucleotide pool measured in a biological control sample, where a deregulation of the level of expression of the oligonucleotide pool with respect to its level Corresponding expression levels measured in a biological control sample can identify a subgroup of good prognosis or a subgroup of poor prognosis.

The present invention also relates to an in vitro method for screening candidate compounds for the treatment of soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST) comprising the steps of: a) bringing into contact a tumor biological sample previously collected with a test compound; b) determining in said tumor biological sample the level of expression of a pool of polynucleotides of the invention; c) comparing said level of expression obtained in step b) with that of the same tumor biological sample not brought into contact with the test compound, where a decrease in the level of expression in the biological tumor sample in presence of the test compound relative to that of the tumor biological sample in the absence of the test compound is an indication of a candidate compound for the treatment of STS or GIST.

The present invention also relates to an in vitro method for monitoring the anti-metastasis efficacy of a treatment of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST ), comprising the steps of: a) providing a previously collected tumor biological sample of said treated patient to be tested; b) determining in said tumor biological sample the level of expression of a pool of polynucleotides of the invention; c) comparing said level of expression obtained in step b) with that of a biological control sample or a biological tumor sample of said patient before treatment, where a decrease in the level of expression of the biological tumor sample after treatment compared to that of the biological control sample or the tumor biological sample before treatment is an indication of an antimetastase efficacy of the therapeutic treatment.

The present invention relates, fourthly, to a kit ("kit") comprising a pool of polynucleotides of the invention. According to the invention, this kit or kit can be used, for example, for the in vitro prediction of the occurrence of metastases and / or for the evaluation of the prognosis of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST) and / or for monitoring the anti-metastasis efficacy of a therapeutic treatment of a patient with soft tissue sarcoma (STS) or gastrointestinal stromal tumor -intestinal (GIST).

According to the invention, this kit or kit may further comprise the means for detecting and / or quantifying the expression of a nucleotide pool of the invention. These means may be for example one of those defined above or exemplified below.

The present invention relates, fifthly, to a nucleic acid chip, in particular to a DNA chip, comprising or consisting of a pool of polynucleotides of the invention. This DNA chip may for example be as defined above, in particular as regards the support. Advantageously, a nucleic acid chip of the invention may comprise "probes", for example cDNA fragments or oligonucleotides (for example 60 to 80 bases, or more), etc., attached to a solid support. . These "probes" specifically bind the "targets" by hybridization, for example the complementary genes present in the biological samples to be tested. This hybridization requires the association by non-covalent linkages of single-stranded nucleic acid sequences that are completely complementary or sufficiently complementary to hybridize with each other and form a double-stranded structure.

BRIEF DESCRIPTION OF THE FIGURES - FIG. 1 represents 3 types of genomic profile (a) amplified

(16%) (a) arm (23%) and (d) rearranged (61%).

Figure 2 shows the Kaplan-Meier-free survival curves of different groups of sarcomas according to the CINSARC signature.

Figure 3 represents progression-free survival / Kaplan-Meier metastasis curves of three tumor groups according to the CINSARC signature.

Figure 4 shows progression-free survival / metastasis (% of cases without metastases versus years after treatment) of Kaplan-Meier in a group of sarcomas (tumor group in which the signature was defined) according to signature by means of the nucleotide pool consisting of the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24. Curve A shows a curve of survival of patients with good prognosis, presenting about 80% of cases without metastases at 5 years. Curve B shows a survival curve of patients with poor prognosis, presenting about 50% of cases without metastases at 5 years.

Figure 5 represents progression-free survival / metastasis (% of cases without metastases versus years after treatment) curves of Kaplan-Meier from a group of sarcomas (tumor group independent of signature identification) according to the signature by means of the nucleotide pool consisting of the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24. Curve A shows a curve survival of patients with good prognosis, presenting about 90% of cases without metastases at 5 years. Curve B shows a survival curve of patients with a poor prognosis, presenting about 50% of cases without metastases at 5 years.

EXAMPLES

Example 1: Pool of the present invention

Patients and samples

The French Sarcoma Group (GSF) database as part of the Conticabase (www.conticabase.org) contains adult soft tissue sarcoma data processed in 11 centers with patient description, primary tumors , treatments, monitoring and tumor samples. This database contained approximately 3800 cases at the time of the study. All cases were reviewed by the subgroup of pathologists and classified according to the WHO 2002 classification using histology, immunohistochemistry and cytogenetics and molecular genetics when necessary. For this study, soft tissue sarcomas without recurrent chromosomal translocations were selected and for which a frozen tissue sample of the untreated primary tumor was available. Finally, the biological samples from 183 patients described in Table 2 below were studied.

Table 2

DNA extraction and analysis by CGH (Comparative Genomic Hybridization) on a DNA chip

Genomic DNA from frozen tumor tissues was isolated using a standard phenol-chloroform extraction protocol and analyzed on a spectrophotometer (Nanodrop). Thus, after digestion with DpnII (Ozyme, Saint-Quentin en Yvelines, France) and purification on a column (Qiagen PCR Purification Kit, Qiagen), 1.5 μg of tumor DNA and 1.5 μg of normal DNA were labeled in using BioPrime DNA labeling System Kit (Invitrogen, Cergy Pontoise, France) with Cy5-dCTP or Cy3-dCTP (Perkin Elmer), respectively. The labeled normal and tumor DNAs were mixed and precipitated together with 100 μg of human Cot-1 DNA (Invitrogen), resuspended in 72 μl of hybridization buffer (50% formamide, 40 mM NaH ₂ PO ₄ , 0.1 % SDS, 10% dextran sulfate, 2X SSC). Prehybrid probes were plated and introduced into wet chambers (Corning) and hybridization was performed at 37 ° C for 48 h.

In order to establish the genomic profiles, BAC (Bacterial Artificial Chromosome) chips composed of 3803 BAC clones were made with an average of 1 Mb between the clones. The BAC clones have been deposited in triplicate.

Washings after hybridization were performed as follows: washing at 65 ° C in 0.5X SSC, 0.03% SDS, followed by washing at 45 ° C in the same solution.

The slides were scanned (Scanarray 4000XL, Packard Bioscience) and analyzed with the GenePix Pro 5.1 image analysis software. Standardization, subdivided filtration, cluster analysis and graphical representation were performed using the CGH DNA Chip Analysis Platform (CAPWeb). Clones with more than 50% missing values have been eliminated. Cy5-Cy3 ratios greater than 2 were considered amplifications, ratios above 1.2 and below 0.8 were considered gains and losses, respectively. CGH analysis on a DNA chip (calculation of genomic alterations) was performed by the VAMP interface (LA ROSA et al., Bioinformatics, 22 (17): 2066-2073, 2006).

RNA Extraction and Expression Analysis Total RNA was extracted from frozen tumor sample with the TRIzol reagent (Life Technologies, Inc.). The RNA was then purified using the RNeasy® Min Elute ™ Cleanup Kit (Qiagen) according to the manufacturer's instructions. The quality of PARN was verified on the Agilent 2100 Bioanalyzer (Agilent Technologies).

The samples were then analyzed on a Human U 133 Plus 2.0 genome (Affimetrix®), according to the manufacturer's instructions. All DNA microarray data were normalized simultaneously using the GCRMA algorithm (WU et al., J. Am., Stat Assoc., 99: 909-917, 2004). Hierarchical grouping analyzes were performed using the dChip software (http://biosunl.harvard.edu/complab/dchip/). For the Welch, Willcoxon and SAM tests, p-values were adjusted using the Benjamini-Hochberg (R-multitest package) procedure.

The analysis in the Gene Ontology database (GO; http://www.geneontology.org/) was performed to find statistical enrichment at the boundaries of the GO. Statistical analysis

Chi-square tests (χ ² test) were performed to evaluate the link between different tumor characteristics, genomic alterations, expression profiles and clinical outcome. The reciprocal influence among the different predictors was determined by multivariate analysis using an upstream logistic regression test. All factors were included in the logistic regression analyzes, regardless of their univariate P-values, but only those with a P <5% value were retained in the final models. Survival without metastases was obtained by the Kaplan-Meier method and compared with the logarithmic rank test. All statistical tests were double-sided and the significance level was p = 0.05. All statistical analyzes (logistic regression model) were performed using SAS version 8.

RESULTS

Genomic profile of the 183 poorly differentiated sarcomas

The genomic profiling of the 183 poorly differentiated sarcomas was performed by CGH analysis on a BAC chip containing 3803 clones. Three main recurring patterns were identified, based on both the number and type of alterations identified, among 174 genomic profiles that can be interpreted in fine (Figure 1). A first group of 28 tumors (16%) with simple genetics, called "amplified" profile, based on co-amplifications and corresponding almost exclusively to dedifferentiated liposarcomas. A second group of 40 tumors (23%), called the "arm" profile, with some alterations (less than 30), mainly involving an alteration of the entire chromosome or of an entire chromosome arm. A third group of 106 tumors (61%), referred to as "rearranged" profile, characterized by a high level of chromosomal complexity with more than 30 to 85 alterations.

It remains to be seen whether the genomic profile is associated with the clinical outcome. Supervised group analysis according to the genomic profile (profile

"Arm" vs "rearranged" profile) did not significantly predict the occurrence of metastases (p = 0.17). Interestingly, a positive correlation was found between the "rearranged" profile and the histological grade 3, in the 183 sarcomas of the study (p = 0.001), and in the subgroup of 117 sarcomas of members with complex genetics profiles with "arms" and "rearranged" (p = 2,2XlO ^"4). the histological grade is an indirect assessment of tumor aggressiveness, it was shown that although no correlation with poor clinical outcome n ' has been obtained, the genomic complexity is associated with tumor aggressiveness. It remains to be demonstrated whether the expression of the genes associated with the genomic complexity and / or the tumor grade could be predictive of the occurrence of metastases.

Expression profiles and establishment of the prognostic molecular signature The gene expression profiles of the 183 sarcomas in the study were reconsidered to test the hypothesis of a correlation between specific gene expression in complex genome tumors. and the occurrence of metastases.

To do this, the 183 samples were first grouped according to a previously established signature composed of 70 genes selected as being related to chromosomal instability (CARTER et al., Nat., Genet., 38 (9 ): 1043-1048, 2006). But this led to a prediction of trend but not significant survival without metastases.

Also in a second step, the goal was to establish a set of sarcoma specific genes associated with the level of imbalances and able to predict the future of a patient. In three supervised analyzes, the expression profiles of tumors classified into two groups were analyzed according to i) the number of CGH imbalances, less than 20 imbalances vs more than 35 imbalances, ii) the FNCLCC 3 histological grade vs the tumor grade 2 , and iii) Carter's signature. From the first two comparisons, 118 clones corresponding to 86 genes and 92 clones corresponding to 73 genes were significantly differentially expressed between the stratified tumors either by the CGH imbalances (differential expression factor (= number of times the gene was is more expressed)> 3, or not, p <0.01) or by grade (differential expression factor> 2, p <0.01), respectively. These genes were then analyzed by the Gene Ontology database to determine pathways associated with CGH imbalances and histologic grade. Interestingly, these pathways are extremely similar in groups determined by CGH imbalances and those determined by histological grade comparisons, and are primarily involved in chromosome integrity and control of mitosis (Table 3). Among the genes of signing Carter, 22 genes that have not yet been identified in the first two comparisons were significantly expressed (p <10 ^"5) differentially between the two groups of sarcomas.

These results were selected all significant genes belonging to the tracks over-represented significantly from the first two comparisons (p <10 ^"5; Table 3) and 22 genes signature Carter defined below above.

Table 3 a) from the Welch test

Number of probes / input clones: 92 (b) after the Welch test

This final set of genes, called by the inventors CINSARC (Complexity INdex SARComas), consists of 67 genes all involved in the control of the genome.

Example 2 Prediction of the Occurrence of Metastases in Sarcomas Using CINSARC

The correlation of the CINSARC expression signature with the occurrence of metastases was evaluated in the entire study group (183 sarcomas). Cluster analysis allowed the classification of tumors into three subgroups (subgroups 1, 2, 3), with a significant difference in the occurrence of metastases (Figure 2). Multivariate analysis showed that subgroup 3 tumors have a triple risk of metastasis compared to subgroup 1 tumors (Kaplan-Meier analysis, HR = 3.01, 95% CI [1.8- 5.2], p <10 ^" ) A multivariate analysis taking into account other standard prognostic factors, such as histological type, FNCLCC tumor grade, tumor size, localization, vascular or bone, sex and age, also showed a risk of metastasis three times higher for subgroup 3 compared to subgroup 1 (Cox model, HR = 3.1, 95% CI [1.8 5.4], p <10 ^"3). These results showed that the CINSARC signature is an independent prognostic factor strongly associated with the development of metastases. After this validation of the CINSARC signature as an independent prognostic factor, 6 specific subgroups of sarcomas were also tested by unsupervised cluster analysis (Figure 2). Of the 117 sarcomas Complex genetic members, univariate analysis divided the tumors into two subgroups and demonstrated a risk of triple metastasis for subgroup 2 vs. subgroup 1 (Kaplan-Meier analysis, HR = 3.1; 95% CI [1.6 to 6.0]. p <10 ^"3) similarly, among the 52 leiomyosarcoma, three subgroups of different significant clinical outcome (P = OOl O ₅₎ have been found (it It is interesting to note that subgroup 2 consists almost exclusively of LMS developed in the inner trunk instead of the outer trunk for the other two subgroups.) Also when only the LMS of the external trunk are taken into consideration by an analysis. by unsupervised group, the 36 patients are divided into two subgroups with a sextuple difference in metastatic risk (Kaplan-Meier analysis, HR = 6, 95% CI [2,1-16,9], p <10 ^{~ 3} ).

The performance of the CINSARC signature was also analyzed for patients with the same histologic grade (Figure 2). In Grade 3 tumors (100 cases), a triple metastatic risk was observed in tumors of subgroup 2 vs tumors of subgroup 1 (Kaplan-Meier analysis, HR = 3, 95% CI [ 1.6-5.6], p <10 ^"3 ) and within grade 2 tumors (40 cases) with arm or rearranged profiles (ie all except DD-LPS), patients were equally distributed in two groups of different clinical outcome (Kaplan-Meier analysis, HR = 2.6, 95% CI [1-7.5], p = 0.05), survival without metastases was not significantly different in two groups of dedifferentiated liposarcomas grouped according to the CINSARC signature Thus the CINSARC signature of the present invention made it possible to separate into two groups having a probability of occurrence of different metastases, tumors considered with the same metastatic potential according to the FNCLLC grade system ( This result is perhaps the most important, as it demonstrates clearly that the CINSARC signature may be a more effective system than the one currently used to determine therapeutic strategies.

In addition, for the first time in the field of sarcomas, a gene expression profile attributes a better clinical prognosis than that obtained with the FNCLLC grade system. Thus, in the whole group of different histotypes, the CINSARC signature allowed the identification of a subgroup of tumors with poor prognosis, whereas the FNCLLC grade system failed to separate these tumors from distinct prognoses ( data not shown).

Example 3: Prediction of the occurrence of metastases in other cancers using CINSARC The predictive value of CINSARC in other sarcomas was tested and a series of 32 GISTs were analyzed (YAMAGUCHI et al., J; Clin Oncol., 26 (25): 4100-4108, 2008). As shown in Figure 3, the CINSARC signature allowed hierarchical unsupervised cluster analysis leading to two GIST groups with a different prognosis (p <10-3). Interestingly, this classification is location-independent even though the GISTs of the small intestine and those of the stomach form two separate groups in each different prognostic group.

Since the CINSARC signature is composed exclusively of genes involved in chromosome integrity and the expression is associated with chromosomal imbalances, the CINSARC signature could also have a prognostic value for highly rearranged tumors, such as breast carcinomas. Consequently, two series of breast cancer (78 and 295 cases) from the Netherlands Cancer Institute (VAN'T VEER et al., 2002, cited above, VAN de VIJVER et al., 2002, cited above) have been compiled by signing CINSARC, and again two groups of patients with different clinical outcome very significant (p <10 ^"3) were obtained (Figure 3).

As demonstrated in the study, the CINSARC signature is a powerful independent predictor tool for better assessment of the occurrence of metastases as well as the attribution to patients of a better clinical prognosis compared to the FNCLCC grade system. This new molecular-grade system should improve the clinical monitoring of patients. In addition, this biological significance of the CINSARC signature genes defines them as potential targets for new therapeutic approaches targeting the early stage of metastatic potential acquisition.

The fact that the CINSARC signature is associated with the occurrence of metastases through such heterogeneous tumor groups (from sarcomas to carcinomas) is sufficiently encouraging to consider, instead of the current histological grade system, the use of this profile. expression to identify patients at high risk for metastases and target complementary chemo-therapeutic strategies.

Current therapeutic strategies combine surgical resection and chemotherapy / radiotherapy in adjuvant or neoadjuvant situations. However, only sarcomas with high metastatic potential should benefit from such treatment. This is currently the case for GISTs whose adjuvant treatment with imatinib is being validated for tumors at high risk of recurrence. However, currently used systems are imperfect. The use of the CINSARC signature could improve the selection of these patients and thus increase the benefit of adjuvant therapies.

There is therefore a strong interest in the use of the CINSARC signature as a major decision criterion for the eligibility of adjuvant therapy, in particular concerning GISTs (gastrointestinal stromal tumors) for which targeted therapy already exists. (Gleevec).

Example 4 Prediction of the occurrence of metastases in sarcomas using a pool of CINSARC polynucleotides The correlation of the expression signature of the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID

NO: 24 of CINSARC with the occurrence of metastases was evaluated on two sets of sarcomas (Figures 4 and 5). Cluster analysis allowed the classification of tumors into two subgroups (subgroups A and B), with a significant difference in the occurrence of metastases. Analysis by the closest centers method showed that subgroup B tumors had a higher risk of metastasis compared to subgroup A tumors. These results showed that the CINSARC 5 gene signature is a independent prognostic factor strongly associated with the development of metastases.

This result is important, as it clearly demonstrates that the signature of five CINSARC genes may be a more efficient system than the one currently used to determine therapeutic strategies.

Claims

A pool of polynucleotides comprising at least two polynucleotides selected from the polynucleotide sequences SEQ ID NO: 1 to SEQ ID NO: 67.

2. Pool of polynucleotides according to claim 1, said polynucleotide pool comprising the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24.

The polynucleotide pool of claim 1, said polynucleotide pool comprising at least one polynucleotide selected from each of the following sets of polynucleotides:

Set 1: SEQ ID NO: 1 to SEQ ID NO: 12; Set 2: SEQ ID NO: 13 to SEQ ID NO: 38; Set 3: SEQ ID NO: 39 to SEQ ID NO: 50;

Set 4: SEQ ID NO: 51 to SEQ ID NO: 58, and SEQ ID NO: 59 to SEQ ID NO: 62;

Set 5: SEQ ID NO: 63 to SEQ ID NO: 65, and SEQ ID NO: 66 to SEQ ID NO: 67.

The pool of polynucleotides according to claim 3, said pool comprising the polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58 and SEQ ID NO: 24 and at least one polynucleotide of Set 5.

The pool of polynucleotides according to claim 1, said pool comprising at least two polynucleotides selected from:

Set 1: SEQ ID NO: 1 to SEQ ID NO: 12; or Set 2: SEQ ID NO: 13 to SEQ ID NO: 38; or Set 3: SEQ ID NO: 39 to SEQ ID NO: 50; or Set 4: SEQ ID NO: 51 to SEQ ID NO: 58, and SEQ ID NO: 59 to SEQ ID

NO: 62; or

Set 5: SEQ ID NO: 63 to SEQ ID NO: 65, and SEQ ID NO: 66 to SEQ ID NO: 67.

Polynucleotide pool according to any one of the preceding claims, said polynucleotide pool comprising at most ten polynucleotides.

The polynucleotide pool of claim 1, said polynucleotide pool consisting of polynucleotides of SEQ ID NO: 10, SEQ ID NO: 3, SEQ ID NO: 47, SEQ ID NO: 58, SEQ ID NO: 24.

The pool of polynucleotides according to claim 1, said polynucleotide pool comprising the polynucleotides of SEQ ID NO: 1 to SEQ ID NO: 67.

9. Pool of polynucleotides according to any one of claims 1 to 3, immobilized on a solid or liquid support.

10. Use of a polynucleotide pool according to any one of claims 1 to 5, for obtaining a compound for the treatment of a soft tissue sarcoma (STS) or a gastrointestinal stromal tumor. (GIST).

11. An in vitro method for analyzing a soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST), said method comprising determining the level of expression of a pool of polynucleotides according to the any of claims 1 to 5 in a biological sample.

The method of claim 11, wherein said determining the level of expression predicts the occurrence of metastases.

The method according to claim 11, wherein said determining the level of expression makes it possible to distinguish subgroups of good or poor prognosis from a group of soft tissue sarcomas (STS) or a gastrointestinal stromal tumor ( GIST) of the same histological grade.

The method of any one of claims 11 to 13, wherein said determining the level of expression is performed on a nucleic acid chip.

15. Use of a method according to claim 11 for detecting and / or prognosis and / or diagnosis of soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST), or for monitoring the treatment of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST).

16. In vitro method for predicting the occurrence of metastases in a patient with soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST) comprising the steps of: a) providing a tumor biological sample previously collected from said patient to be tested; b) determining in said tumor biological sample the level of expression of a pool of polynucleotides according to any one of claims 1 to 4; c) c) comparing the level of expression obtained in step b) with the expression level of the same polynucleotide pool measured in a biological control sample, a deregulation of the level of expression of the oligonucleotide pool with respect to its corresponding expression level measured in a biological control sample being predictive of the occurrence of metastasis.

17. An in vitro method for assessing the prognosis of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST), comprising the steps of: a) providing a tumor biological sample previously collected from the patient with an STS to be tested; b) determining in said tumor biological sample the level of expression of a pool of polynucleotides according to any one of claims 1 to 4; c) comparing the level of expression obtained in step b) with the level of expression of the same polynucleotide pool measured in a biological control sample, where a deregulation of the level of expression of the oligonucleotide pool with respect to its level Corresponding expression levels measured in a biological control sample can identify a subgroup of good prognosis or a subgroup of poor prognosis.

18. An in vitro method for screening candidate compounds for treating soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST) comprising the steps of: a) contacting a tumor biological sample with a test compound; b) determining in said tumor biological sample the level of expression of a pool of polynucleotides according to any one of claims 1 to 4; c) comparing said level of expression obtained in step b) with that of the same tumor biological sample not brought into contact with the test compound, where a decrease in the level of expression in the biological tumor sample in presence of the test compound relative to that of the tumor biological sample in the absence of the test compound is an indication of a compound useful in the treatment of STS or GIST.

19. In vitro method for monitoring the anti-metastasis efficacy of a treatment of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST), comprising the following steps : a) providing a previously collected tumor biological sample of said treated patient to be tested; b) determining in said tumor biological sample the level of expression of a polynucleotide pool according to any one of claims 1 to

4; c) comparing said level of expression obtained in step b) with that of a biological control sample or a biological tumor sample of said patient before treatment, where a decrease in the level of expression of the biological tumor sample after treatment compared to that of the biological control sample or the tumor biological sample before treatment is an indication of an antimetastase efficacy of the therapeutic treatment.

20. The in vitro method for selecting a polynucleotide pool according to any one of claims 1 to 4 comprising the following steps: a) providing tumor biological samples from patients with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor

(GIST); b) detecting and / or quantifying each of the polynucleotides, individually in each of the biological tumor samples; c) comparing the expression profile of the polynucleotide pools obtained in step c) with respect to a biological phenotype, preferably of chromosomal instability, genomic complexity or histological grade, for each of the biological tumor samples; d) selecting the statistically significant way (P <10 ^"5) for the tested phenotype; e) select the polynucleotides significantly implicated in this biological pathway, and the expression of which is indicative of the probability of occurrence of metastases.

21. A kit comprising a pool of polynucleotides according to any one of claims 1 to 5.

22. Kit for in vitro prediction of metastasis, prognosis evaluation of a patient with soft tissue sarcoma (STS) or gastrointestinal stromal tumor (GIST) and / or monitoring the antimetastase efficacy of a therapeutic treatment of a patient with soft tissue sarcoma (STS) or a gastrointestinal stromal tumor (GIST), said kit comprising means for detecting and / or quantifying the expression of a nucleotide pool according to any one of claims 1 to 5.

23. A nucleic acid chip comprising a polynucleotide pool according to any one of claims 1 to 4, immobilized on a solid or liquid support.