WO2005098443A1 - Methods for the diagnosis of cancers or of precancerous lesions - Google Patents

Abstract

The present invention relates to the use of a compound enabling the detection of the substantial lack of expression of the TP53INP1 protein in a biological sample, for the manufacture of a drug intended for the prognosis or the diagnosis of cancers and/or of precancerous lesions, or for the screening of drugs active on cancers.

Description

METHODS FOR THE DIAGNOSIS OF CANCERS OR OF PRECANCEROUS LESIONS

The present invention relates to methods for the diagnosis of cancers or of precancerous lesions. Among human cancers, adenocarcinomas of the lung, breast, prostate, colon, and pancreas represent the most common causes of cancer death in Western societies (Jemal et al.,

CA Cancer J Clin, 2003, 53: 2-26) One of the major causes of death by cancer is the absence of reliable diagnostic methods enabling the detection of cancers early enough, i.e., before they have developed to the point where the available treatments are no longer effective. This is especially crucial for pancreatic ductal carcinoma, a genetic, highly aggressive disease, the fifth leading cause of death by cancer in the Western population (Jemal, et al., 2002, CA Cancer J Clin., 52: 23-47), because most of the time diagnosis is made when cancer is already advanced. The current knowledge of pancreatic carcinoma progression, originally based on histological data, benefited from recent description of genetically and epigenetically defined precursor lesions called pancreatic intraepithelial neoplasias (PanlNs: Kloppel and Luttges, 2001, Verh. Dtsch. Ges. Pathol., 85: 219-228; Hruban et al., 2001, Am. J. Surg. Pathol., 25:

579-586). Numerous alterations of tumor suppressor genes, including K-ras, pi 6, DPC4,

TP53, and BRCA2 genes occur frequently in invasive pancreatic adenocarcinoma (Luttges et al., 2001, Am. J. Pathol, 755:1677-1683; Hansel et al., 2003, Annu. Rev. Genomics Hum.

Genet., 4: 237-256). Occurrence of gene abnormalities in PanlNs is generally not random but associated with early, intermediate or late changes (Maitra et al., 2003, Mod Pathol., 16:902-

912). Transition between the normal state and neoplasia is progressive, although alteration of cell cycle regulation is the hallmark of the neoplastic phenotype, in pancreatic carcinoma as in most human tumors. However, there is a current lack of markers with sufficient sensitivity to detect pancreatic cancer early enough to effect cures (Hingorani et al., 2003, Cancer Biother, 2: 84-86; Rosty and Goggins, Clin. North am, 2002, 16: 37-52). Colorectal cancer (CRC) is one of the most frequently diagnosed neoplasms worldwide and the second most common cause of cancer related death (Parkin et al., Eur J

Cancer, 2001, 37 (suppl 8): S4-66). The progression of normal colonic mucosa to cancer requires several molecular and genetic changes to occur. These changes occur over the course of years, and follow a dedicated course from adenomatous polyp to carcinoma as described by Nogelstein et al. (Ν Engl J Med, 1988, 319: 525-532). Patient survival is dependent on the stage of the disease at the time of diagnosis. Colonoscopy has become an increasingly popular choice for screening because it prevents CRC by the identification and removal of polyps prior to conversion to invasive cancer (Citarda et al., Gut, 2001, 48: 812-815). CRC mortality could be lowered even more through the detection of early cancers and precancerous lesions. It is to be noted though, that the vast majority of polyps do not progress to cancer. Thus, it is of major interest to have a marker reflective of the progression toward CRC. As for gastric cancer, another one of the leading cause of cancer related death in the world, the multistep model of carcinogenesis suggests accumulation of genetic alterations, epigenetic changes and posttranslational modifications. It often metastazises to other organs, including liver, lung, and ovary (Carneiro and Sobrinho, 1996, Human Pathol, 27: 213-214). Multiple factors are known to be related to gastric carcinogenesis, including Epstein-Barr virus (Zur Hausen et al., J Clin Pathol. 2004 May;57(5):487-91) and Helicobacter pylori infections (Suganuma et al., J Cancer Res Clin Oncol. 2004 Dec 23; [Epub ahead of print]), microsatellite instability (Hiyama et al., J Gastroenterol Hepatol, 2004, 19: 756-760). From the molecular point of view, it has now been established that gastric carcinogenesis involved accumulation of mutations in oncogenes and tumor suppressor genes controlling epithelial cell growth and differentiation (King et al. Semin Radiat Oncol 1999; 9:4-11 ; Liu et al. Histopathology 2001; 39: 603-10 ; Feakins et al. Hum Pathol 2003; 34: 1276-82 ; Chetty et al. J Clin Gastroenterol 2003; 37: 23-7 ; Jung et al. Clin Cancer Res 2003; 9: 6052-61). TP53INP1 (Tumor Protein p53 Induced Nuclear Protein 1, formerly named SL?, TEAP or p53DINPl) is a novel gene strongly activated during the acute phase of pancreatitis (Tomasini, et al., 2001, J. Biol. Chem., 276: 44185-44192; Tomasini et al., 2002, Eur. J. Cell. Biol., 81: 294-301; Carrier et al, 1999, hnmunogenetics, 50: 255-270). Overexpression of the two protein isoforms and β encoded by the gene induces cell cycle arrest in Gl and increases p53-mediated cell death through physical interaction with HIPK2, a serine/threonine kinase involved in transcriptional regulation and apoptosis (Tomasini et al, 2003, J. Biol. Chem., 278: 37722-29). The involvement of TP53INP1 in cancer has never been assessed. Despite the description of numerous molecular, in particular genetic, markers of cancer development and/or progression, there are still no markers liable to be used reliably for the detection of cancers, in particular cancers at an early stage of development, and/or of precancerous lesions. Thus, one of the objects of the present invention is to provide the previously unrecognized use of a protein, to detect cancers, in particular cancers at an early stage of development, and/or precancerous lesions. Another object of the invention is to provide a new compound liable to be used for tlxe diagnosis of cancers, in particular of cancers at an early stage of development, and/or of precancerous lesions. The present invention relates to the use of a compound enabling the detection of th_e substantial lack of expression of the TP53INP1 protein in a biological sample, for tke manufacture of a drug intended for the prognosis or the diagnosis of cancers and/or of precancerous lesions, or for the screening of drugs active on cancers. The above mentioned use results from the unexpected finding that the TP53DSEP 1 protein is essentially not expressed in cancerous cells and/or precancerous cells, whereas it i s essentially expressed in the corresponding normal cells. By normal cells it is meant, non-pathological cells, in particular healthy cells. The above mentioned biological samples correspond to tissue samples from an individual which are suspected to contain cancerous cells. In particular, such samples correspond to biopsies or smears of the following tissues: pancreas, colon, lung, prostate, bladder, breast, uterus cervix, ovary, head and neck, oral, skin, kidney, thyroid, stomach., conjonctiva-cornea, thymus, and nervous tissues. The expression "substantial lack of expression of the TP53INP1 protein" means thai; there is substantially no TP53INP1 protein synthesis being carried out in some or all of the tissues which are part of said biological sample. The lack of expression of the TP53USTP1 protein corresponds either to the absence of the TP53INP1 protein or to the absence of the TP53INP1 mRNA. These absences may be linked to epigenetic changes, or to disorders in the transcription or the translation of the TP53INP1 mRNA, or in the degradation of the TP53INP1 protein, for example. Advantageously, the lack of expression of the TP53INP1 protein in tissues, or part of tissues, which naturally express said protein indicates that said tissues are either afflicted by a. cancer or under way of developing a cancer. The expression "precancerous lesions" relates to lesions which are not yet constitute of cancerous cells, said cells being in the process of cancerous transformation. The current identification of precancerous lesions is based on the observation of morphological and histological features with increasing degrees of architectural and nuclear atypia, such as hyperplasia, metaplasia and dysplasia. The precursor lesions of ductal pancreatic adenocarcinoma have been recently codified under the collective term pancreatic intraepithelial neoplasia (PanlN) (Kern et al., Cancer Res, 2001, 61 : 4923-4932). In colon, the so-called polyp is a projection of a growth into the lumen of the colon. The adenomatous polyp is a well-accepted fellow-traveler of CRC. Substantial epidemiological and direct clinical observations support its pivotal role as a precursor lesion for CRC (Fearon and Vogelstein, Cell, 1990, 61 : 759). The expression "drugs active on cancers" means that said drugs are liable to treat individuals afflicted with cancers. According to a preferred embodiment, the present invention more particularly relates to the in vitro diagnosis of cancers. According to another preferred embodiment, the present invention more particularly relates to the in vivo diagnosis of cancers. The invention relates in particular to the use as defined above, of a compound enabling the detection of the substantial lack of expression of the TP53INP1 protein in a biological sample, for the manufacture of a drug intended for the diagnosis of cancers at an early stage. The expression "early stage" relates to newly formed cancers. The cells constituting such cancers are notably characterized by prominent nuclear changes, including changes in size and shape, enlarged nucleoli, and abnormal heterochromatin distribution. In particular, early stage epithelial cancers are dysplastic with abnormal intracellular and intercellular organization. Similarly, early pancreatic cancers (at the Panln-l and- 2 stages) display mild to moderate papillary architectural atypia, nuclear hypercliromatism and pleomorphism, and a beginning of nuclear stratification (Biankin et al., Pathology, 2003, 35: 14-24). The invention more particularly relates to the use as defined above, of a compound enabling the detection of the substantial lack of expression of the TP53INP1 protein in a biological sample, for the manufacture of a drug intended to differentiate malignant tumors, in particular early stage malignant tumors, or precancerous lesions, from healthy tissues and benign tumors. The expression "malignant tumor" relates to a group of clonal cells characterized by morphological changes, such as dysplasia, as compared to the corresponding normal cells, and invasiveness (Hanahan and Weinberg, 2000, Cell, 100: 57-70). The expression "healthy tissues" relates to tissues which are un-afflicted with cancers, in particular early stage cancers, or precancerous lesions. The expression "benign tumor" relates to an abnormal accumulation of hyperplasic cells which does not present the characteristics of malignant tumors. According to a particular embodiment of the invention, the compound involved enables the detection of the substantial absence of the TP53INP1 mRNA and/or of the TP53INP1 protein. According to another particular embodiment of the invention, the compound involved is liable to bind to the TP53INP1 mRNA and/or to the TP53INP1 protein. In this case the substantial absence of binding of said compound corresponds to the absence of the TP53INP1 mRNA and/or protein. According to a more particular embodiment of the invention, the TP53TNP1 mRNA or protein corresponds to the mRNA or to the protein of TP53INP1 isoforms, such as the TP53INPlG.mRNA or protein, or the TP53INPljS mRNA or protein. The sequence of the human TP53INPlα.mRNA corresponds to SEQ ID NO: 1. The sequence of the human TP53INP10! protein corresponds to SEQ ID NO: 2. The sequence of the human TP53INP1/3 mRNA corresponds to SEQ ID NO: 3. The sequence of the human TP53INPlj8 protein corresponds to SEQ ID NO: 4. TP53INP1 and β mRNA are derived from a same TP53PNP1 gene by alternative splicing of the fourth exon of said gene (Tomasini et al., Eu. J. Cell Biol., 2002, 81 : 294-301). According to a preferred embodiment of the invention, the compound is a monoclonal or a polyclonal antibody which binds to the TP53INPlo_ protein and/or to the TP53INP1/3 protein. In another preferred embodiment of the invention, the compound is an anti-TP53INPl antibody fragment, such as a Fab, a F(ab)' or a scFv fragment. According to a particularly preferred embodiment of the invention, the monoclonal antibody is secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Instirut Pasteur, Paris, France) on March 26, 2004, under accession number CNCM 1-3194. The antibody secreted by the hybridoma deposited at the CNCM under accession number CNCM 1-3194 corresponds to antibody E12 of the examples. According to another particular embodiment of the invention, the monoclonal antibody can also correspond to the antibody F8 or Al described in the examples. The general procedure for the preparation of a monoclonal antibody is well known to the man skilled in the art. For example, it can be done according to Galfre et al. (Nature 1977, 266: 550-552), as described in the Examples. Briefly, the TP53INP1 protein is administered several times to a rat; the rat is then sacrificed and its spleen taken. Lymphocytes are purified from the spleen, contacted with myelome cells, and fused to them to obtain hybridomas. The hybridomas are separated from the rest of the unfused cells and cloned. Each cloned hybridoma is then screened for the production of antibodies directed against the TP53INP1 protein. The hybridomas yielding the antibodies with the most interesting properties, such as specificity or affinity are selected. The above mentioned monoclonal antibody according to the invention recognizes both the and β isoforms of TP53FNP1. Furthermore, said monoclonal antibody is more particularly directed against a sequence corresponding to contiguous aminoacids 30 to 110 of of the TP53INP1Q. protein, i.e., to SEQ ID NO: 5. According to another particular embodiment of the invention, the compound comprises at least one nucleotide sequence, said nucleotide sequence containing at least 10 contiguous nucleotides selected from SEQ ID NO: 1 or 3, or from the sequence complementary to SEQ ID NO: 1 or 3, or at least one sequence derived from one of said nucleotide sequence, by insertion, deletion or substitution of at least one nucleotide and presenting at least 90% similarity with the nucleotide sequence from which it derives. The above mentioned nucleotide sequence may correspond for example to a primer for

PCR (polymerase chain reaction) or PCR related application, such primers corresponding either to a part of SEQ ID NO: 1 or 3 or to a part of the sequence complementary to SEQ ID NO: 1 or 3, to achieve amplification. The above mentioned nucleotide sequence may also correspond to a probe for hybridization experiment. According to a more particular embodiment of the invention, the nucleotide sequence corresponds to SEQ ID NO: 1 complementary sequence or to SEQ ID NO: 3 complementary sequence. According to another particular embodiment said compound is:

- a TP53INP1 protein substrate, or substrate analogue, such as the HIPK2 kinase or fragments of the HIPK2 kinase, or

- an aptamer which binds to the TP53INP1 protein. The "expression" aptamer relates to nucleic acids, in particular ribonucleic acids, which are liable to bind to biological macromolecules such as proteins or sugars for example. According to another particular embodiment of the invention, the detection is carried out with a method selected from: PCR, real-time PCR, RT-PCR, NASBA, Northern Blot, in situ hybridization, chromatin precipitation, ELISA, Western Blot, far Western protein interaction assay, immunoprecipitation, FACS, flow cytometry, cytochemistry, cytofluorescence, immunofluorescence, immunohistochemistry. Such methods are well known to the man skilled in the art. PCR relates to polymerase chain reaction, RT-PCR to reverse transcriptase polymerase chain reaction, NASBA to nucleic acid based sequence amplification, ELISA to enzyme linked immunosorbent assay, and FACS to fluorescence activated cell sorting. According to still another particular embodiment of the invention, the cancers are selected from the group of epithelial cancers comprising pancreas cancer, colon cancer, lung cancer, prostate cancer, bladder cancer, breast cancer, uterine cervix cancer, ovary cancer, head and neck cancer, oral cancer, skin cancer, kidney cancer, thyroid cancer, stomach cancer, conjonctiva-cornea cancer, glioma, thymoma, and neuroblastoma. According to a further particular embodiment of the invention, the precancerous lesions are selected from precancerous lesions leading to cancers selected from the group of epithelial cancers comprising pancreas cancer, colon cancer, lung cancer, prostate cancer, bladder cancer, breast cancer, uterine cervix cancer, ovary cancer, head and neck cancer, oral cancer, skin cancer, kidney cancer, thyroid cancer, stomach cancer, conjonctiva-cornea cancer, glioma, thymoma, and neuroblastoma, or from precancerous lesions arising from inflammatory bowel disease (IBD), such as ulcerative colitis (UC) or Crohn's disease, and endo-brachy-oesophagus (EBO). The present invention also relates to the monoclonal antibody secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Institut Pasteur, Paris, France) on March 26, 2004, under accession number CNCM 1-3194. According to another particular embodiment of the invention, the monoclonal antibody can also correspond to the antibody F8 or Al described in the examples. The present invention further relates to a cancer prognosis or diagnosis kit comprising: - at least one anti-TP53INPl protein antibody, in particular a labelled anti-TP53INPl protein antibody, - a sample of the TP53INP1 protein as a control or a reference. According to a preferred embodiment of the invention, the antibody corresponds to the monoclonal antibody secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Institut Pasteur, Paris, France) on March 26, 2004, under accession number CNCM 1-3194. According to another particular embodiment of the invention, the monoclonal antibody can also correspond to the antibody F8 or Al described in the examples. The invention also relates to a method for the in vitro diagnosis of cancers at an early stage, of cancers, or of precancerous lesions, characterized in that it comprises the following steps: - contacting a tissue sample taken from an individual with a anti-TP53INPl protein antibody, in particular a monoclonal antibody, such as the monoclonal antibody secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Institut Pasteur, Paris, France) on March 26, 2004, under accession number CNCM 1-3194. - washing the tissue sample, - revealing the presence, or the absence, of said antibody, in the washed sample, the substantial presence of said antibody in the washed sample corresponding a healthy tissue, and the substantial absence of said antibody in the washed sample corresponding to a tissue afflicted with a cancer at an early stage, with a cancer, or with a precancerous lesion. According to another particular embodiment of the invention, the monoclonal antibody can also correspond to the antibody F8 or Al described in the examples. In an embodiment of the above mentioned method for the in vitro diagnosis of cancers at an early stage, of cancers, or of precancerous lesions, the following steps are comprised: - contacting a tissue sample taken from an individual with a TP53INP1 protein ligand, such as an anti-TP53INPl protein antibody, in particular a monoclonal antibody, such as the monoclonal antibody secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Institut Pasteur, Paris, France) on March 26, 2004, under accession number CNCM 1-3194. - washing the tissue sample, - revealing the presence, or the absence, of said antibody, in the washed sample, the substantial absence of said antibody in the washed sample as compared to a sample of a corresponding healthy tissue being an indication that said tissue is afflicted with a cancer at an early stage, with a cancer, or with a precancerous lesion. As intended in the above in vitro diagnostic method, the tissue sample has been taken from an area of the tissue which is suspected to carry cancerous or precancerous cells. As such "a sample of a corresponding healthy tissue" particularly relates to a sample taken from a different and presumably healthy area of same tissue. Furthermore, the tissue sample taken from the individual is selected from tissues which under normal or healthy conditions, i.e. non-pathological conditions, express TP53INP1. In an advantageous embodiment of the present in vitro diagnostic method, the tissue samples are prepared in the form of histological sections or tissue arrays suitable for microscopic observations, according to methods well known to the man skilled in the art, and in particular as set forth in the following examples. By way of example: 86% of histological sections of pancreatic ductal adenocarcinoma lack TP53INP1 expression, whereas 0% of normal histological sections or histological sections adjacent to the adenocarcinoma lack TP53INP1 expression; 76% of histological sections of colon carcinoma lack TP53INP1 expression, whereas 0% of normal histological sections or histological sections adjacent to the carcinoma lack TP53 P1 expression; - 36% of histological sections of gastric carcinoma lack TP53INP1 expression, whereas 2% of normal histological sections or histological sections- adjacent to the carcinoma lack TP53INP1 expression. It is also advantageous, in particular when a lack of expression of TP53INP1 is evidenced in a tissue sample, either to proceed to one or several other determinations of the lack of expression of TP53INP1 in independent samples of the same tissue, or to proceed to the detection of the presence or the absence of other cancer or pre-cancer markers. The invention further relates to an anti-cancer drug screening method, characterized in that it comprises the following steps:

- contacting a sample of precancerous or cancerous cells, in which the TP53INP1 protein is substantially absent, with a molecule liable to be an anti-cancer drug, - detecting the substantial presence or the substantial absence of the TP53INP1 protein, in the cell sample which has been contacted with said molecule to screen, with an antibody, in particular the monoclonal antibody secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Institut Pasteur, Paris, France) on March 26, 2004, under accession number CNCM 1-3194.

- selecting the molecule responsible for the presence of the TP53INP1 protein in the cells, as an anti-cancer drug. According to another particular embodiment of the invention, the monoclonal antibody can also correspond to the antibody F8 or Al described in the examples. Description of the figures

Figure IA, Figure IB, Figure IC, Figure ID, Figure IE and Figure IF Figures IA to IF represent immunohistochemical patterns of TP53INP1 protein expression in normal pancreas, chronic obstructive pancreatitis consecutive to pancreatic ductal cancer, intraductal papillary-mucinous tumor (IPMT), liver metastase and PanlN lesions at different stages: epithelium of a normal large duct showing cell cytoplasm and nucleus labelling (Figure IA); invasive ductal adenocarcinoma cells are TP53INP1 -negative (anowhead), whereas a nearby area of chronic obstructive pancreatitis (CP) shows strong positive staining, and the islet of Langerhans (arrow) is negative (Figure IB); strong staining of mucinous cystadenoma (Figure IC); TP53INPl-negative metastase in liver (anowhead) (Figure ID); positive TP53INP1 staining in PanIN-1 lesions and in normal acinar cells (AC), in contrast, PanIN-2 lesions (anowhead) show moderate or no TP53INP1 staining (Figure IE); absence of TP53INP1 staining in PanIN-3 lesions (Figure IF).

Figure 2A and Figure 2B Figure 2A represents a Western blot analysis of TP53INP1 in normal (N) and tumoral (T) tissues from four patients with invasive ductal pancreatic adenocarcinoma (1-4) in the presence (+) or absence of the TP53INP1 antibody blocking peptide. Blots were stripped and reprobed with anti-β-Actin antibody (45 kDa). The two TP53INP1 protein isoforms α (18 kDa) and β (27 kDa) were detected in normal tissues (N). Complete loss (patients 2 and 4) or significant decrease (patients 1 and 3) of TP53INP1 protein were evidenced in tumor samples (T). Specificity was confirmed by pre-incubating TP53INP1 antibody with the conesponding blocking peptide (patient 3).

Figure 2B represents an experiment of colony formation assay with the Mia-PaCa-2 pancreatic cell line. Cells were transfected with the empty pcDNA4-V5 vector (A), and the conesponding vectors harboring either TP53INPla (B) or TP53INPlβ (C). Transfected cells were selected with zeocine for 10 days and stained with crystal violet. The number of colonies is significantly higher in A than in B or C Figure 3A and Figure 3B

Figures 3A and 3B represent the characterization of TP53INP1 monoclonal antibody (clone E12). Figure 3A shows the reactivity of the anti-TP53INPl with the TP53INP1 species in Western blots with native recombinant protein either TP53INPlα or TP53INPlβ. Figure 3B shows data obtained by Western blots on the total lysates from Cos-1 cells expressing either TP53INP1 isoform after transient transfection with the Myc-tagged TP53INPl or TP53INPlβ pcDNA3 expression vectors. Figure 3B upper and lower panels show the controls for the presence of the two overexpressed TP53INP1 isofo ms and the same amount of proteins, by using anti-Myc and anti-β Actin antibodies, respectively. Anti-TP53INP1 (anti- E12) strongly detected TP53INPlα and in a faint manner TP53INPlβ (middle panel).

Figure 4

Figure 4 represents the protein sequences of TP53INP1 and β isoforms, as well as schematic representation of the regions used for epitope mapping of monoclonal anti- TP53INP1 antibodies (boxes below the sequences). The sequences of human TP53INPlα (164 aa) and TP53INPlβ (240 aa) were divided into three and four fragment regions, respectively. Fragments 1 and 2 conespond to aa 2 to 42 and aa 33 to 110 of both TP53INPlα and TP53INPlβ, respectively. Fragment 3 conesponds to aa 101 to 164 of TP53INPlα (and includes aa 101 to 157 of TP53INPlβ).

Figure 5A, Figure 5B, Figure 5C, Figure 5D, Figure 5E and Figure 5F Figures 5A to 5F represent immunohistochemical patterns of TP53INP1 protein expression in normal and cancerous colon as revealed by using the monoclonal anti-TP53INPl (clone E12). Figure 5 A shows a strong staining in all epithelial cells of normal colon. Figure 5B shows that the invasive colon carcinoma is TP53INP1 -negative (on the right), whereas the adjacent normal colon displays strong staining (on the left). Figures 5C and 5D show invasive colon carcinoma with a complete loss of TP53INP1 expression in cancer cells, or a heterogeneous pattern of expression, respectively. Figures 5E represents polyps with low displasia and normal TP53INP1 expression and Figure 5F represents polyps with high displasia and no TP53INP1 expression. Figure 6

Figure 6 represents a Western blot analysis of TP53LNP1 in normal (N) and tumoral (T) tissues from four patients with invasive colon cancer (1-4). Blots were stripped and subsequently reprobed with anti-β-Actin antibody (45 kDa). The two TP53INP1 protein isoforms α (18 kDa) and β (27 kDa) were detected in normal tissues (N). A drastic decrease in TP53ENPl expression was observed in all but one (patient 3) cases.

Figure 7

Figure 7 represents a genetic model for the progression of colorectal tumorigenesis adapted from Fearon and Vogelstein (Cell, 1990, 61 : 759) showing that loss of TP53INP1 is an early event. Colorectal tumors progress through a series of clinical and histopathological stages, ranging from single crypt lesions (abenant crypt foci: ACF) to small benign tumors (adenomatous polyps: early, intermediate, and late adenoma) and to malignant cancers (adenocarcinoma). Several documented (APC, KRAS, DCC, DPC4, TP53) or putative (JV18) markers of the progression are represented.

Figure 8A, Figure 8B, Figure 8C and Figure 8D

Figures 8A-8D represent pictures of immunohistochemical analyses of TP531NP1 expression in gastric carcinoma. Figure 8 A: TP53INP1 was strongly expressed in normal gastric mucosa or intestinal metaplasia foci (inset).

Figure 8B: TP53FNP1 expression decreased in gastric carcinoma (anow).

Figure 8C: Well differentiated tubular carcinoma exhibited moderate alteration of TP53INP1 expression. Figure 8D: Poorly differentiated carcinoma showed wealdy staining with TP53INP1 antibody.

Figure 9

Figure 9 represent the percentage of TP53INP1 positivity (vertical axis) in cell cytoplasm and in nucleus of normal gastric mucosa (left bar) and gastric cancer tissues (right bar) (the star symbol (*) represents pθ.0001). Figure 10 A, Figure 10B and Figure IOC

Figures 10A and 10B show representative patterns of TP53INP1 -positive (Figure 10 A) and negative (Figure 10B) carcinoma in TUNEL staining. Arrows indicate TUNEL-positive nuclei (original magnification x 20). Figure 10C represents the apoptotic index xlOO (vertical axis) in TP53rNPl-positive and - negative cancer tissues (the star symbol (*) represents pθ.0001).

Figure IIA, Figure IIB and Figure IIC

Figure I IA represents survival curves for TP53INP1 -positive and negative gastric cancers. The 60-months survival rates are 75% and 50%, respectively. The difference between the values is highly significant (p=0.0006).

Figure 1 IB represents survival curves for TP53INP1 -positive and negative "well or moderately differentiated adenocarcinoma (intestinal-type), the survival of TP53MP 1 -positive patients was not significantly different from that of TP53INP1 -negative group (p=0.1110). Figure IIC represents survival curves for TP53INPl-positive and negative poorly differentiated adenocarcinoma (diffuse-type), the 60-months patient survival rates are 60% and 40% for TP53INP1 -positive and -negative gastric cancers, respectively. This difference is statistically significant (p=0.019).

Figure 12A, Figure 12B, Figure 12C and Figure 12D

Figures 12A-12D show representative immunohistochemical patterns of TP53INP1 protein expression in gastroenteropancreatic normal and tumoral tract. Figure 12A, strong TP53INP1 immunoexpression in normal gastric mucosa. Figure 12B, TP53INP1 expression in gastric endocrine cells that exhibit endocrine granules using Grimelius silver staining (anows).

Figure 12C, in normal pancreas, acini are TP53INP1 positive whereas Langherans islets (L) are negative.

Figure 12D, a TP53INP1 positive pancreatic endocrine tumor: a normal TP53INP1 negative pancreatic islet is entrapped. EXAMPLES

Example 1

Monoclonal Antibodies (itiAbs) elicited against TP53INP1

Full-length human TP53INPlα cDNAs (SEQ ID NO: 1) were subcloned into either pGEX-5X-2 (Pharmacia) or pQE-31-5 (Qiagen) fusion vectors allowing the synthesis of glutathione-S-transferase or 6-Histidine, respectively, upstream from the N-terminus of TP53INPlα. pcDNA4-V5 (Invitrogen), pTP53INPl -V5 and pTP53INPlβ-V5 have been previously described (Tomasini, et al, 2001, J. Biol. Chem., 276: 44185-44192, 20O1). All constructs were controlled by sequencing. The TP53INPlθ! protein was produced and purified either by means of a GST- TP53P Pl fusion protein or by means of a 6-HIS-TP53INPlα fusion protein. For the GST-TP53INPlα protein, the TP53INP10! coding sequence was PCR amplified and subcloned in the BamRVSall sites of the pGEX-5X-2 expression vector to yield pGEX-5X-2-TP53INPl. The construct was controled by sequencing. E. coli strain BL21 (Stratagene) was transformed with the bacterial expression plasmid pGEX-5X-2-TP53INPl. 50Oml bacteria culture (2XYT) were grown at 37°C to an O.D. (600 nm) of about 0.6 - 0.8 and expression of protein was induced by adding isopropyl-D- thiogalactopyranoside (IPTG) at 25°C to a final concentration of 0.1 mM. After 3 h, cells were collected by centrifugation for 10 minutes at 3,000 x g and 4°C. The cell pellet was resuspended in 30ml cold PBS for washing. After centrifugating lOminutes at 3,000 x: g, the pellet was resuspended in lOml cold STE and 100 μl lysozyme was added to a concentration of 10 mg ml^"1. After 15 min at 0°C 100 μl DTT 1M and 1,4ml of Sarkosyl 10% were added. After cells were lysed by sonication, and the extract was cleared by centrifugation at 1 l,000g for 20 min, 4ml of Triton 10% and enough STE to reach. 20ml were added. After an incubation step of 30 minutes at room temperature, affinity chromatography was performed by adding 1ml of glutathione-Sepharose 4B beads to the extract and by further incubating 1 hour at room temperature. The whole sample was transfened to a column and washed with 3x50ml PBS. GST-TP53INP1 was eluated with lOxlml of elution buffer (20mM of reduced glutathion in 50mM Tris-HCl pH 9). Collected fractions were separated by 12.5 % SDS- PAGE and stained with Coornassie-blue for determination of the elution peak. For the 6-HIS-TP53INP1 a. protein sequence, TP53INPlα coding sequence was PCR amplified and subcloned in the I?ύ.7«HI/HmdIII sites of the pQE-31 expression vector to yield pQE31- TP53INP1. The construct was controled by sequencing. E. coli strain Ml 5 was transformed with the bacterial expression plasmid pQE31- TP53INP1. 1 L bacteria cultures were grown to an O.D. (600 nm) of about 0.6 and expression was induced by adding isopropyl-D-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. After 4 h, cells were collected by centrifugating for 30 min at 4,000 x g. The cell pellet was resuspended in 3ml/g of lysis buffer 1 (50 mM Tris-Cl, pΗ 8, 1 mM EDTA pΗ 8, 100 mM NaCl) with 4μl/g PMSF 1 OOmM and lmg/ml of lysozyme and the suspension was incubated on ice for 20 min. 4mg/g of deoxycholic acid (lmg/ml) were added to the suspension, which was then sonicated six times at 50 % for 5 s and centrifuged for 15 min at 15,000 x g and 4 °C. The suspension was incubated 30minutes at 37°C and 20 μl/g DNase (1 mg/ml) were then added. The resulting suspension was further incubated 30 minutes at room temperature. After a centrifugation step of 15minutes at 4,000xg and 4°C, the pellet was resuspended in 5ml/g of lysis buffer 2 (100 mM NaΗ₂PO₄, lOmM Tris-HCl, 8 M Urea). After 30 min at room temperature, the cells were lysed by sonication, and the extract was cleared by centrifugation at 10,000g and 4°C for 30minutes. The supernatant was added to 400μl of a 50 % sluny Ni-NTA-agarose for 2 h under constant rotation. The whole sample was transfened to a column and washed 5x10ml Λvith wash buffer (100 mM NaH₂PO₄, lOmM Tris-HCl, 8M Urea, lOmM imidazol). To remove urea, the column was washed with 50ml TBS and the solubilized protein was then elute with TBS containing 50mM EDTA. Rat immunization with human TP53INPlα recombinant protein (18 kDa) (SEQ ID NO: 2) and hybridoma selection were done according to Galfre et al. (1977 Nature, 266:550- 2), and in accordance with institutional guidelines. Briefly, two LOU/N rats were immunized over 4 months, with ^' 50 μg of either recombinant TP53INPloGST or 6-HIS-TP53INPlα, using 4 successive injections. The first injection was i.p. in complete Freund's adjuvant, the two following were i.p. in incomplete Freund's adjuvant, and the last one was i.v. Fusion was made three days after the last boost, according to the method of Galfre et al. Spleens were removed and dilacerated, and cell suspensions were treated using Tris- buffered ammonium chloride to lyse red blood cells. Splenocytes, as well as myeloma cells, were then rinsed twice using DMEJVI medium without serum. Fusions of spleen cells (1.5xl0⁸) with X63.Ag8.653 myeloma cells (1.5xl0⁷) were made using polyethylene glycol 1500 (PEG). After fusion, cells were plated into 96-well microtiter plates (6 plates per fusion), and hybrids selected in HAT medium. Supernatants were screened on day 13 using an ELISA assay with TP53INPlo;-coated plates. Selected hybrids were expanded and cloned (0.3 cells by wells). Their isotype was determined using the mAb-based rat Ig isotyping kit (BD Pharmingen). As a result: - for the fusion using the spleen of GST-TP53INPlα: injected rat, the screening was done with the 6-HIS TP53INPlc_ recombinant protein, 33 positive hybrids were obtained, cloning was done for 10 hybrids, and 6 hybridoma clones were obtained at the end (2 IgGl and 4 IgG2a) - for the fusion using the spleen of 6-HIS-TP53JNP1G! injected rat, the screening was done with the GST-TP53INPlα recombinant protein, 31 positive hybrids were obtained, cloning was done for 10 clones, and 2 hybridoma clones were obtained at the end (both IgG2a, including E12). All 8 hybridoma clones secrete monoclonal antibodies recognizing both TP53INPlθ! and TP53INPl]3. The anti-TP53INPl monoclonal antibody, clone E12, was selected based on its reactivity in ELISA with TP53P Pl , in Western blots with native recombinant protein, and in immunofluorescence studies on transfected cells. Briefly, the mouse monoclonal anti-Myc (9E10, Santa Cruz Biotechnology, Inc) was used for immunofluorescence and immunohistochernistry studies, respectively. For indirect immunofluorescence, Cos-1 cells were transiently transfected using 1 μg plasmid DNA and 3 μl FuGene transfection reagent (Roche Diagnostics, Meylan, France) following the manufacturer's recommendations. Twenty-hours after transfection, cells were grown on lab-Tek chamber slides (Nun Inc.) in medium containing 1:2000 Calpain Inhibitor I (ALLN, Sigma-Aldrich) for one day. Cells were trien fixed in 4% formaldehyde in IX PBS for 15 min at room temperature, and permeabilized in 0.2% Triton X-100 in PBS for 5 min. After PBS washes, cells were blocked in 3% fetal calf serum/PBS for 20 min. Cells were incubated with either anti-TP53INPl (1:10), or anti-Myc antibody at a final concentration of 1.6 μg/ml for 1 h at room temperature, rinse several times, and then incubated with 1.6 μg of either goat anti-rat IgG-FITC (sc-2011 from Santa Cruz Biotechnology, Inc) or Alexa Fluor 594 goat anti-mouse IgG (from Molecular Probes, Inc) secondary antibodies, respectively. After several washes in PBS, slides were mounted in Mowiol. The reactivity of antibody E12 with TP53INP1 species in Western blots with native recombinant protein is presented in Figure 3A. Furthermore, as shown in Figure 3B, anti- E12 strongly detected TP53ιNPlα and to a lesser extent TP53__NIPlβ in Cos-1 cells expressing either TP53INP1 isoform. Mass spectrometry analysis of t_he bands revealed the specific TP53INP1 peptide profile. While TP53INPlα and β have a predicted molecular weight of 18kDa and 27kDa, respectively, the two protein isoforms migrate at a higher molecular weight (approximately 35kDa and 45kDa, respectively) for epitope-tagged constructs (Figure 3A) and for transiently transfected products (Figures 3B). Anti-El 2 antibody was next used in indirect immuno fluorescence studies on Cos-1 cells transiently transfected with Myc-tagged-TP53INPl constructs. Λ.s expected, anti-E12 recognized the TP53INPl protein mostly located in the cytoplasm, and TP53INPlβ foimd quite exclusively located within the nucleus, in the nucleoplasm, and in discrete nuclear areas conesponding to PML-nuclear bodies as recently shown (Tomasini, et al., 2003, J. Biol. Chem., 278: 37722-37729). Identical protein localizations were found using anti-Myc antibody for TP53INPlα and β, respectively. Taken together, these results clearly demonstrated that the monoclonal anti-E12 TP53INP1 antibody recognize the two TP53INP1 isoforms and is suitable for ELISA, Western blot, and immunofluorescen.ee studies. The epitope specificity of three of the eight obtained monoclonal antibodies was also studied. Indeed, whereas all eight of the monoclonal antibodies recognize recombinant TP53INP1 produced in a bacterial expression system, only three of them (E12, F8, and Al) recognized recombinant TP53INP1 produced in an eukaryotic expression system. In order to determine which part of TP53INP1 is recognized by the different mAbs, the sequences of human TP53INPlα (164 aa) and TP53INPlβ (240 aa) were divided into three and four regions, respectively. Region Fl and region F2 respectively conespond to amino acids 2 to 42 (SEQ ID NO: 5) and amino acids 33 to 110 (SEQ ID NO: 6) of both TP53INPlα and TP53INPlβ, region F3 conesponds to amino acids 101 to 164 (SEQ ID NO: 7) of TP53INPlα (and includes amino acids 101 to 157 of TP53INPlβ), and region F4 conesponds to amino acids 152 to 240 (SEQ ID NO: 8) of TP53INPlβ. The nucleotidic sequences encoding these fragments were cloned into the procaryotic expression vector pQE30 (Qiagen) and transformed in E. coli. Bacterial protein lysates were blotted onto nitrocellulose membranes and hybridized with the different mAbs. B12 and F8 were shown to recognize region F2, whereas Al recognizes region F3 (Figure 4)- Thus, the E12 antibody conesponds to an IgG2a isotype (as detected using the mAb- based rat isotyping kit, BD Pharmingen) that recognizes an epitop e localized in a sequence common to the two TP53INP1 isoforms and β (fragment F2, amino acids 33 to 110, SEQ ID NO: 6). Monoclonal antibody F8 recognizes an epitope localized in a sequence common to the two TP53INP1 isoforms a and β (fragment F2, amino acids 33 to 110, SEQ ID NO: 6). Monoclonal antibody Al recognizes an epitope localized in a sequence common to the two TP53LNP1 isoforms a and β (fragment F3, amino acids 101 to 164, SEQ ID NO: 7). The results gathered for the monoclonal antibodies obtained lierein are summarized in the following table: Clone Full Name + Western Blot Western Blot Epitope Name isotype Results on bacteriae Results on eultaryotes mapping Al A13A1 + + F3 IgG2a B7 A14B7 + - IgGl Bl l A12B11 + - IgG2a C3 A21C3 + - IgG2a D9 A15D9 + - IgG2a E12 A25E12 + + F2 IgG2a F5 A13F5 + IgGl F8 A15F8 + + F2 IgG2a The anti-TP53INPl antibody from clone E12, was used for all the studies described in this work.

Example 2

Immunostaining analysis of pancreatic tissues Patients, samples, anatomopathology and statistical analyses. Archival resection specimens from 71 patients were studied. The samples included 5 normal pancreas (from patients with gastric cancer who underwent total gastrectomy with distal pancreatectomy and splenectomy without cancer invasion of the pancreas), 3 intraductal papillary mucinous tumors (IPMT), 9 mucinous cystic neoplasms, 43 pancreatic adenocarcinomas, 4 liver and 4 lymph node metastases (from which 2 primary pancreatic adenocarcinoma were also analyzed), and 5 carcinomas of the ampulla of Nater. Tissues were either formalin-fixed and paraffin-embedded, or flash frozen in liquid nitrogen. Differentiation and pathological staging were assessed by histological examination according to criteria defined by Kloppel et al. (1984, Dtsch. Med. Wochenschr., 109: 702-708) and Hermreck et al, (Am. J. Surg., 1974, 127: 653-657), respectively. Chronic pancreatitis cases were obtained from 31 of the above mentionned patients. All cases of histologically confirmed chronic pancreatitis were obstmctive chronic pancreatitis due to either mucinous cystic neoplasms (9 cases) or IP lT (1 case) or carcinoma of the ampulla of Vater (5 cases) or pancreatic adenocarcinoma (16 cases). Pancreatic intraepithelial neoplasias (PanlΝ) were classified according to the new nomenclature of pancreatic duct lesions (Hruban et al., Am. J. Surg. Pathol., 2001, 25: 579- 586). The mean age of patients with pancreatic adenocarcinoma (23 males, 20 females^ was 67.2 years (range, 51-85). The clinical and follow-up data were obtained from medical records, after coding according the recommendations of the local Committee for Bioetliics. The following parameters were considered: age, gender, tumor size, histological grade, clinical stage and location within pancreas, lymph node and metastasis status, and survival. Statistical analysis was done by χ testing. Immunostaining. Anti-TP53IΝP1 monoclonal antibody was purified using a G- sepharose column and used as primary antibody (6 μg/ l, overnight incubation) for immunostaining of paraffin-embedded sections from the 71 patient samples. Detection was done using Rat ABC Staining system (sc-2019, Santa Cruz Biotechnology, Inc.) according to the manufacturer's instructions. Slides were then counterstained with hematoxylin, and mounted using Eukitt solution. For negative control experiments, anti-TP53INPl was either replaced by saline or pre-incubated with recombinant TP53INP1 protein (lOμg ml). Immunostainings were independently assessed by three of the Inventors.

Tissue Microarray Analysis (TMA). Paraffin blocks from pancreas ca_ncer specimens were provided by the Department of Pathology at the Vancouver Hospital and Health Sciences Centre. Using matching Hematoxylin and Eosin (H&E) stained slides, ajreas with adenocarcinoma in donor paraffin blocks were marked and identification of the paraffin blocks appointed to map of the tissue microanay (TMA). TMA was constructed using a tissue puncher (Beecher Instruments, Silver Spring, MD) in triplicate cores, each 0.6 min in diameter, from 49 individual tumors, and cores were anayed in recta-linear pattern. Sections were dewaxed, rehydrated in the graded ethanol, steamed in buffered citrate (pH=6) for 30 minutes. After washing three times with PBS, slides were incubated with 3% H2O2 in PBS for 15 minutes to block endogenous peroxidase activity. After washing three times with PBS, slides were then incubated in 3% BSA (Bovine serum albumin, Promega, Madison, WI) for 30 minutes. Incubation with 6 μg/ml of Anti-TP53INP1 monoclonal antibody in 1% BSA was ca ied out overnight at 4C°. The next day the primary antibody was carefully washed 3 times with PBS. Biotin conjugated goat anti rat Ig with concentration of 5 μg/ml (Pharmingen, BD bioscience, San Diego, Ca, USA), was used as secondary antibody and streptavidin-HRP from LSAB+ kit (Dako, Carpinteria, CA) followed by Chromogen Nova-red (Vector Laboratories, Burlingame, CA) was applied for 5 minutes. Counterstaining was performed with hematoxylin. After ethanol rehydration, slides were covered by a coverglass with xylene based mounting media, Cytoseal (Stephen Scientific, Riverdale, NJ). Negative control slides were processed in an identical fashion to that above mentioned, with the substitution of 1 % BSA for the primary antiserum. The staining intensity in each section was evaluated and scored by one pathologist and graded by sum of the intensity in 4 point visual score scheme (0 - 3 representing negative to strong staining). All comparisons of staining intensity and percentages were made at 200X magnification. Using identical microscopic and camera settings (RT color-SPOT high resolution digital camera from Diagnostic Instrument. Inc. Tampa, FL, USA, mounted on Olympus System light microscope model BX51) digital images were taken from representative area reflecting the overall staining.

Mutation analysis by direct sequencing. Genomic DNAs were purified using QIAamp DNA Mini kits (Qiagen). Each exon of TP53INP1 gene was amplified by PCR (sequences for primers are available upon request) followed by sequencing reaction using PRISM dye terminator kit (Applied Biosystems) and automated sequencer ABI373.

TP53INP1 is expressed in normal pancreas, in chronic obstructive pancreatitis, and in benign tumors. To determine the level and distribution of TP53INP1 expression in pancreas tissue and in chronic obstructive pancreatitis associated or not with cancer, immunohistochemical analysis of whole sections was done using the monoclonal anti- TP53INP1 antibody. In normal tissues, the epithelial layer of the large ducts, the main duct and its major branches showed immunoreactivity to anti-TP53INPl, mostly located to the cytoplasm (Figure IA, Table 1). Thirty out of the 31 cases (97%) of chronic obstractive pancreatitis areas showed an enhanced cytoplasmic TP53PNP1 expression in both acinar cells and large interlobular ducts (Figure IB). No staining was observed in the Langerhans islets in any of the analyzed samples (Figure IB). These data are consistent with previous results demonstrating that induction of acute pancreatitis in mice resulted in rapid and strong induction of both TP53INP1 mRNAs in pancreas (Tomasini, et al., 2001, J. Biol. Chem., 276^': 44185-44192). In benign pancreatic tumors, significant expression of TP53INP1 protein was detected in 100% of intraductal papillary-mucinous tumors (IPMT, Table 11, and mucinous cystadenomas (Figure IC, and Table 11. Taken together, these data demonstrate that TP53INP1 is overexpressed in chronic obstructive pancreatitis, associated or not with cancer. TP53INP1 is also expressed in benign tumors of the pancreas such as IPMT with no or slight dysplasia.

TP53INP1 protein is lost in primary ductal pancreatic adenocarcinomas and in metastases. Immunohistochemistry and TMA analysis of pancreatic samples from patients with adenocarcinomas revealed that TP53INP1 protein expression was significant in only 14% of tumor samples (12/87). It was not detectable in the remaining (86%) cancer specimens (Table 1 and Figure B). In addition, TP53INP1 expression was detected in none of the metastases of pancreatic adenocarcinoma, from liver (Table 1, Figure 1D1 or lymph nodes (Table 11. Loss of TP53INP1 expression in pancreatic adenocarcinomas is significantly associated with high pathological grade (p=0.04), metastasis (p=0.06), and .the sex of patients (Table 21. The exact mechanisms remain to be elucidated since no TP53INP1 mutation nor hypermethylation in the promoter was evidenced in patient samples. This study suggests that loss of TP53TNP1 may be important in pancreas cancer progression and that the detection of this altered TP53INP1 expression in precursor lesions of invasive carcinoma might be useful for the differential diagnosis of benign and malignant intraductal lesions of the pancreas.

TP53INP1 inactivation occurs early in pancreatic adenocarcinoma progression. The cunent multistep progression model for adenocarcinoma within pancreatic ducts from normal epithelium follows the sequence hyperplasia - dysplasia - in situ adenocarcinoma- invasive adenocarcinoma. The precursor lesions have been identified as intraductal epithelial proliferations or lesions termed pancreatic intraepithelial neoplasia (PanlNs) (Klδppel and Luttges, 2001, Verh. Dtsch. Ges. Pathol., 85: 219-228; Hruban et al.,. 2001, - Am. J. Surg. Pathol., 25: 579-586; Biankin et al., 2003, Pathology, 55:14-24). Increasing evidence suggests that PanlNs represent true neoplasms of the pancreatic ductal epithelium, accumulating histologic and genetic abnormalities in their progression toward invasive cancer (Maitra et al., 2003, Mod Pathol, 16:902-912; Wilentz et al., 1998, Cancer Res., 58: 4740-4744). Deregulation of genes encoding proteins involved in cell cycle regulation and cell signalling pathways have been described and classified as early events (K-ras, PSCA, apomucin MUCl and 5, Cyclin Dl, and pi 6) or late events (p53, DPC4, BRCA2) in the progression model. Understanding the molecular pathogenesis of the precursor lesions of invasive tumors is an important challenge for early detection of this lethal neoplasm. Most invasive adenocarcinomas show inactivation of pi 6 and harbour mutations of the TP53 gene but these abenations appear late in the progression model. The inventors have shown herein that TP53INP1 is lost in 91% of invasive ductal carcinoma. To investigate its putative role in cancer progression, the expression of TP53INP1 was documented in PanlNs, the precursors of ductal carcinomas. In the 43 pancreatic adenocarcinoma samples, only the in situ component in the vicinity of invasive ductal carcinoma were analyzed. Of the 69 PanlNs observed (Table 1), 24 were graded as PanlN- 1, 29 as PanIN-2, and 16 as PanIN-3 according to the cunent grading scale (Hruban et al., 2001, Am. J. Surg. Pathol., 25: 579-586). The inventors have found that TP53INP1 was expressed in all early PanIN-1 lesions (Figure IE) as in normal duct epithelium, and that, in contrast, TP53INP1 expression was not detected in 45% of PanIN-2 (Figure 1E1 and in 100% of PanIN-3 (Figure IF) lesions. Inverse conelation between TP53INP1 expression and the grade of PanlN lesions adds further support to the progression model proposed for pancreatic carcinoma. Thus, it is shown that loss of TP53INP1 expression is concomitant with the occunence of dysplasia. Indeed, complete loss of TP53INP1 protein was observed after transition from PanIN-2 to PanIN-3 stages. Therefore, TP53INP1 appears as one of the important targets of alteration in the stepwise progression of PanlNs towards cancer. By comparison, loss of pi 6 expression appears as a common and early event in the multistep model of pancreatic adenocarcinoma. Frequency of loss of pl6 expression increased with dysplasia progression (Maitra et al, 2003, Mod Pathol., 76:902-912; Biankin et al., 2003, Pathology, 55:14-24) to reach approximately 95% of invasive pancreatic adenocarcinomas. But, contrary to TP53INP1, pi 6 expression is lost in PanlN- 1 lesions in the absence of dysplasia (Hansel et al., 2003, Annu. Rev. Genomics Hum. Genet., 4: 237-256; Maitra et al., 2003, Mod Pathol., 76:902-912; Biankin et al., 2003, Pathology, 55:14-24; Wilentz et al., 1998, Cancer Res., 58: 4740-4744). Therefore, TP53INP1 is the first described gene whose expression is present in all PanIN-1 lesions but completely abolished in PanIN-3 lesions.

Example 3 Immunoblot analysis of pancreatic tissues

To further document the loss of TP53TNP1 protein expression, Western blotting experiments were performed using a commercial polyclonal anti-TP53INPl serum. Briefly, the polyclonal anti-P53DINPl (the TP53INP1 former name) and mouse monoclonal anti-β- actin antibodies were purchased from ψ ProSci, Inc. (distributed by ALEXIS platfonn) and Sigma- Aldrich, respectively. The Western blotting analysis was done on 4 patients from which non-cancerous and neoplasic pancreatic samples were available. Approximately 200 mg of frozen normal or cancer tissues were grinded in liquid nitrogen and thawed in ice-cold buffer [50mM Hepes, 1 mM EGTA, 1 mM EDTA, 150 mM NaCl, 10% glycerol, 1 % TritonX- 100, supplemented with protease inhibitors], then homogenized for 30 min at 4°C and centrifuged. Supernatants were collected and assessed for protein concentration (Bradford method). Total protein extracts (50-100 μg) were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and blotted onto membranes (Hybond-P, Amersham Pharmacia Biotech). Protein extract incubation in the presence of the blocking peptide p53DINPl (PSC-3045P, ψ ProSci, Lie) was used as a control. The filters were first probed with either monoclonal anti- TP53INP1 or polyclonal p53DINPl antibodies, stripped and then subsequently reprobed with anti-β-actin mouse monoclonal antibody (Sigma- Aldrich). Incubation with the conesponding horseradish peroxidase-conjugated immunoglobulin was then done followed by enhanced chemoluminescence immunoblot detection system (ECL, Amersham) according to the manufacturer's instructions. As shown on Figure 2A, TP53INP1 proteins were detectable as two bands of approximately 18 and 27 kDa, the expected sizes of the and β isoforms, respectively f Figure 2A1. Western blot analyses also confirmed the loss of TP53INP1 in pancreatic tumors. Example 4

Colony formation assays

To get further insight into the involvement of TP53INP1 in pancreatic tumorigenesis, the consequences of overexpressing each of the two TP53INP1 isoforms in the human pancreatic cancer cell line MIA-PaCa-2 were monitored. Briefly, MIA-PaCa-2 cells (10⁵) were plated into 10-cm Petri dishes and transfected

24 hours later with the indicated expression vectors (i.e., pTP53INPlα-V5 and pTP53INPlβ-

V5) using Fugene reagent (Roche Diagnostics) following the manufacturer's recommendations. The transfected cells were selected with zeocine (0.2 mg/mL, Invitrogen) for 10 days and stained with crystal violet to assess colony numbers. 116 + 6 colonies were thus scored in control experiments (transfection with the empty vector), compared to 41 + 6, and 17 + 5 for TP53INPla and TP53INPlβ, respectively (Figure 2B). Thus, TP53INP1 expression significantly inhibits pancreatic cancer cell proliferation in vitro.

Example 5

Immunostaining analysis of colonic tissues Human colon cancers undergo a multistage carcinogenesis pathway from adenomatous polyps to carcinoma (reviewed in Refs: Graber et al., 1999, Gastroenterology, 116: 210-212; Boland et al., 1998, Cancer Detect. Prev., 22: 377-382). The inventors have studied the expression of TP53INP1 in various colonic tissue samples of patients with colonic cancers and/or polyps. Colonic tissue samples, obtained from archival collection from 71 patients at surgery (50 cancers, 21 adenomatous polyps) The colonic cancer patients comprised 27 males and 20 females with a mean age of 67.2 (range, 51-85) years. Adjacent normal colon was obtained from patients with cancer and/or polyps. All surgical specimens were fixed in 10% buffered formalin at 4°C overnight, embedded in paraffin and cut into 3 μm-thick slices. After deparaffinization, these sections were pretreated with microwaves in 10 mM citrate buffer, pH 6.0, for 10 min. Extremely sensitive immunohistochemical staining was done using a rat ABC Staining System (Santa Craz Biotechnology, Inc.). This procedure incorporated a signal amplification method based on the peroxidase catalyzed deposition of a biotinylated compound, followed by a biotinylated secondary reaction, avidin and biotinylated horseradish peroxidase. The primary monoclonal antibody against human TP53INP1 (E12 clone) was used at a dilution of 1:200. The chromogen used for the colon by diaminobenzidine (DAB) reagent. After counterstaining with hematoxylin, the sections were observed under a microscope. Negative control experiments were performed using without primary antibody or by coincubating antibody with recombinant protein. The results are presented in Figures 5A-5F, 6, 7 and in Table 3. TP53INP1 is expressed in the normal colon mucosa. In normal colon, TP53INP1 was detected in superficial epithelial cells as well as in crypt cells. The staining was seen in nuclei and in cell cytoplasm (Figure 5A1. ^f The inventors also analyzed the TP53INP1 expression status in patients' samples with normal colonic-adenoma-carcinoma sequence. In all cases, it was found that TP53P P1 was expressed with high level in 100% of normal colon tissues and normal colonic mucosa adjacent to the tumor (Table 3, and Figures 5 A and 5B). Tissue macrophages and lymphoid follicles were negative. Essentially, all tumors had no TP53TNP1 expression. 76% of patient samples (38 of 50) had no TP53INP1 expression in primary tumors, as compared with the normal colon (Table 3, Figures 5B and 5Q. In the remaining cases (12 of 50 tumor samples), TP53INP1 expression was extremely weak with a heterogeneous pattern and sometimes detected in a small fraction of the invasive component (Figure 5D). The loss of TP53INP1 was confirmed by Western blot analysis (Figure 61. To assess if TP53INP1 expression is reduced in the normal colonic epithelium- adenoma-carcinoma sequence, the Inventors studied TP53INP1 down-expression in precursor lesions (adenomas). It was found that TP53INP1 expression was significantly decreased in high grade adenomatous polyps compared with polyps of low grade (Table 31. Thus, there is a complete absence of TP53INP1 in polyps with high displasia (Figure 5F1 as compared to polyps with low displasia (Figure 5E). Taken together, these results demonstrate that the loss of TP53INP1 is associated with the progression of colorectal tumors and is involved in their early stage development, as presented in Figure 7. In conclusion, the present results strongly suggest that the loss of TP53INP1 expression could be used as a predictor of progression in patients with colon adenomas and also be clues to the understanding of cancer progression. The functional mechanisms of this gene should also be clarified to develop new strategies for cancer therapies. Example 6

In situ hybridization of TP53INP1 mRNA Non-radioisotopic RNA in situ hybridization was done on 36 cases of pancreatic cancers of which 35 had normal adjacent tissue. The riboprobes were obtained as follows: The pBluescript KS (+/-) containing the 796 bp TP53INPlα cDNA was linearized and labeled with digoxigenin-UTP by in vitro transcription for the antisense (positive) or sense (negative contol) using the Dig RNA labeling kit according to the nanufacturer's instructions (Roche Diagnostics, Meylan, France). Sections from the colon tissues were processed according to Ugolini et al, (Oncogene, 2001, 20: 5810-5817). Semi-quantitative analysis was done by estimation of a staining score [number of mRNA TP53INP1 positive cells x 20/(cell density)]. A strong expression was detected in epithelial cells of normal exocrine pancreas. Down expression of TP53INP1 mRNA was evidenced in the majority of cancer samples, as exemplified by a staining score of 0.673 in cancer cells versus a score of 5.3 in normal cells.

In some cases, the level of mRNA was found to be similar to that to adjacent normal pancreatic cells.

Example 7

TP53INP1 expression in gastric cancer

Patients and Specimens^'. One hundred and forty-two patients with gastric cancer were enrolled in this study. The areas adjacent to cancer lesions were used as non-maligant gastric tissue. The patients underwent operation at the Cancer Research Institute Hospital, Kanazawa University. The histological classification was defined using the Japanese classification of gastric carcinoma (Borchard, Hepatogastroenterology 1990, 37: 223-32). Intestinal type was defined as either papillary or well to moderately differentiated tubular adenocarcinomas. Diffuse type was defined as poorly differentiated adenocarcinoma, signet-ring cell carcinoma, or mucinous adenocarcinomas. The series included 104 men and 38 women, and the mean age was 63.1±10.6 years. There were 76 and 66 cases of differentiated and undifferentiated type, respectively. Imimmohistochemistry. A standard avidin-biotin-peroxidase complex method (ABC) was used for immunostaining. Deparaffinized sections were treated by microwaving at a high power for 5 min two times in a 10 mM citrate buffer to retrieve antigenicity. After washing with PBS, the sections were immersed in 3%> hydrogen peroxide in methanol for 20 min to block any endogenous peroxidase activity. Then the ABC staining system kit (Santa Cruz Biotechnology Inc, Santa Cruz, CA, USA) was used for detection. Sections were incubated with 10% normal serum for 1 hour to inhibit nonspecific antibody binding. Then, sections were incubated overnight at 4°C with 6μg/ml of rat anti-human monoclonal antibody raised against to TP53INP1 as described above. After washing with PBS, detection was done by successively incubating the sections with biotinylated goat anti-rat IgG for 30 min, and avidin-biotin-HRP for 30 min. After extensive washings with PBS, sections were stained with 3-diaminobenzidine for 2~10 min. Then, sections were counterstained with hematoxylin, dehydrated and mounted in Eukitt. Nuclei were lightly counterstained with Mayer's hematoxylin. TP 53INP1 -positive cells were counted in fields chosen at random (100* magnification), and the percentage of the number of positive cells per 1,000 cells was expressed as TP 53INP1 -positive index (%). Normal IgG was used as a negative control.

Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling

(TUNEL). TUNEL-positive epithelial cells were detected on the sections using ApopTag Plus peroxidase in situ apoptosis detection kit (Chemicon hitemational, Inc., Temecula, CA, USA). Briefly, after pretreatment with 20 μg/ml of proteinase K and 3% hydrogen peroxide, sections were incubated with a labeling mixture for 1 hour at 37°C. Then 55 μl of anti- digoxigenin-peroxidase were deposited on sections and incubated for 30 minutes. The reaction products were visualized by 3,3-diaminobenzidine substrate. Nuclei were counterstained with methyl green for 10 minutes. After washing with n-butanol, the sections were dehydrated, and mounted. Apoptotic index (%) conesponding to the number of labeled nuclei per 1,000 nuclei was calculated.

Statistical Analyses. Experimental results were expressed as mean ± SEM. Difference between the means was evaluated by the Mann- Whitney U test. P < 0.05 was considered as statistically significant.

TP53INP1 was expressed in non-malignant gastric tissues and its expression was reduced in gastric cancer tissues. In the non-neoplastic gastric mucosa, TP53INP1 was mainly located in the cytoplasm of epithelial cells (Figures 8 A and 8B). Some nuclei were also stained for TP53INP1 (Figure 9). Similar patterns were observed for intestinal metaplasia samples (Figure 8A, insert). To determine if TP53INP1 is differentially expressed in gastric carcinomas or if it is stage-related, immunohistochemical analysis were performed on 142 samples (76 cases of intestinal type, and 66 cases of diffuse type). All the cancer samples had accompanying non-malignant tissues, 98% of them were positive with TP53INP1 expression (Table 41 and thus could be used as internal control. In contrast, for the cancer samples, the expression of TP53INP1 protein was seen in 91 cases (64%), the other 51 samples (36%) were found TP53INPl-negative (Figure 8B, anow). Overall, TP53INP1 expression in gastric cancers was significantly lower (both in cell cytoplasm and nucleus) than in non-malignant gastric tissue (p<0.0001, Table 4; Figure 91. However, the expression of TP53INP1 was markedly decreased in well-differentiated tubular carcinoma (Figure 1C1, and was sparse or completely lost in poorly differentiated-type cancer (Figure ID). It was next examined whether TP53INP1 expression is associated or not with development and progression of gastric carcinoma. The clinical details of the cohort of patients and the statistical analysis are listed in Table 5. Only two non significant associations were observed, i.e., age and gender. TP53INP1 negativity was associated with gastric body and autrum tumor location (p=0.0193), with poorly differentiated adenocarcinomas (diffuse type) (pO.OOOl). With regard to the depth of invasion, the positive TP53INP1 expression rate was 100% intramucosal tumors (5/5), 81.8% when mucosa was invaded (18/22), 76.5% in muscularis propria (26/34), 54.3% in subserosa (25/46), and 48.6% in serosa (17/35). These results clearly show that alteration of TP53INP1 expression was conelated to the staging of the tumors. The difference was statistically significant when TI tumors were compared to the other stages (p=0.0111, Table 51. In addition, TP53INP1 was significantly expressed in node- negative patients (p=0.0037), and significantly associated with lymphatic invasion-negative patients (p=0.0010). Taken together, these results indicated that loss of TP53INP1 expression was significantly associated with poorly differentiated histology, deep invasion, lymph node invasion, and metastasis. TP53INP1 and apoptosis. TP53INP1 modulates the cell cycle anest and programmed cell death (Tomasini et al., J Biol Chem 2003, 278: 37722-9). To investigate whether the modulation of TP53INP1 expression is associated with differences in apoptotic activity, TUNEL assays were done on sections derived from gastric cancer samples for which the TP53INP1 status was known. Tunel-positive nuclei were clearly seen in TP 53 MPI -positive (Figure 10A) and negative (Figure 10B1 cancer lesions. As shown in Figure IOC, the apoptotic index in the TP53INP1 -positive group (7.48%±2.66%) was significantly higher than that foimd in the TP53INP1 -negative group (4.16%±2.41%). TP53INP1 expression and prognosis. On univariate analysis, patient survival according to pathological stage was significantly different between TP53INPl-positive and TP53INP1 -negative groups. Those patients with TP53INP1 -positive expression had significantly better survival than those without TP53INP1 expression (p=0.0006, Figure 11 A). Survival for TP53INP1 -positive patients with poorly differentiated adenocarcinoma was significantly longer than that of TP53TNP1 -negative patients (p=0.0199, Figure 4B1, whereas the survival of TP53INP1 -positive patients in well or moderately differentiated adenocarcinoma was not significantly different from that of TP 53 INPl -negative patients (p=0.1110, Figure 4C). Taken together, the results indicate that alteration of TP53EMP1 expression was associated with a poor prognosis. Nevertheless, no prognostic value for TP53INP1 expression was evidenced from the multivariate analysis (Table 61. Histological type, apoptotic index, metastasis, and lymph node invasion were the most important independent prognostic factors, TP53INP1 could not be considered as an independent prognostic marker.

Example 8

TP53INP1 expression for predicting metastasis in gastroenteropancreatic neuroendocrine tumors

Endocrine tumors of the gut and pancreas are relatively rare neoplasms. They constitute a heterogeneous group of tumors, which includes different histopathological and prognostic classes. Their histological identification by pathologists is generally easy regardless of the tumor site. For a long time, only metastasis allowed to affirm the malignancy (Heitz et al. Hum Pathol. 1982;13:263-271). More recently, Kloppel et al. (Capella et al. Virchows Arch. 1995;425:547-560; Kloppel et al. World J Surg 1996;20:132-141) proposed a prognostic classification based on location of the lesion, tumor size, mitotic index and functional status. However, these prognostic criteria are often not reliable in predicting the biological behavior and the prognosis of these tumors, e.g., some tumors with favorable features may present fatal outcome whereas others are still classified as "of uncertain malignancy". Thus, the research for new and useful prognostic molecular markers in endocrine gastroenteropancreatic tumors is of great importance.

Study Group. The study group comprised 127 patients who had undergone surgery for endocrine tumors of the gastrointestinal tract or pancreas at Hδpital Edouard Heniot (Lyon, France). The complete clinical and follow-up data for 79 patients were obtained from medical records, after coding according to the recommendations of the local Committee for Bioethics. The following parameters were considered: age, gender, location of the tumor, tumor size, histological grade, clinical stage, duration of survival, duration of follow-up, status at the end of the follow-up period, and cause of death. Tumors were classified according to the WHO classification (Solcia et al. Histological typing of endocrine tumours. 2^nd edition, New York : Springer Verlag, 2000) as follows: benign well-differentiated tumors (OMS la), well-differentiated tumors of uncertain malignancy (OMS lb), well differentiated endocrine carcinomas (OMS 2), and poorly differentiated endocrine carcinomas (OMS 3). The hormone profile of each endocrine tumor was assessed by immunohistochemical examination of the expression of the following peptides and hormones: insulin, glucagon, somatostatin, pancreatic polypeptide, gastrin, calcitonin, serotonin, and vasoactive intestinal peptide. In Table 1 are listed the antibodies used in this study.

Tissue samples. A tissue microanay (TMA) constructed from all of the surgical specimens was used for the TP53INP1 immunohistochemical expression analysis. TMA technique was found to yield results comparable to conventional staining method (n=48). Immunostaining analysis. Anti-TP53INP1 antibody (Example 11 was used as primary antibody (6 μg/ml, overnight incubation) for immunostaining of paraffin-embedded sections. Detection was done using Rat ABC Staining system (sc-2019, Santa Cruz Biotechnology, Inc.) according to the manufacturer's instructions. Slides were then counterstained with hematoxylin (Vector Laboratories, Burlingame, CA), and mounted using Eukitt solution. For negative control experiments, anti-TP53INPl was either replaced by saline or pre-incubated with recombinant TP53INP1 protein (10 μg/ml). To evaluate the histological neuroendocrine-type of the TP53INP1 positive cells in the gastrointestinal epithelium, cytoplasmic argyrophilia by Grimelius' staining was checked on the same slides prior to TP53INP1 immuno detection. Statistical analysis The Fisher Exact test was used to determine association between the expression of TP53INP1 and selected variables. The results were considered as significant for/?<0.05.

Clinical and histopathological data. The series included 127 patients with endocrine tumors from either gastrointestinal tract (63 patients) from which 4 cases of multiple endocrine neoplasia syndrome, or pancreas (64 patients). The main data are summarized in Table 8. Gastrointestinal endocrine tumors (GIETs) from 63 patients (25 male and 38 female) composed the first group. The mean patient age was 62.8 years (range, 39-87 years). Malignant lesions were located all along the gastrointestinal tract (stomach: n=6, duodenum: n=9, Vater ampulla region: n=5, small intestine: n=35, and colon: n=8). The mean tumor size was 28.2 mm (range 4-100) according to OMS classification (Solcia et al. Histological typing of endocrine tumours. 2^nd edition, New York : Springer Verlag, 2000). Tumors were considered as benign (2 cases), of uncertain malignancy (3 cases), and well differentiated (54 cases) or undifferentiated (4 cases) carcinomas. Of the 63 patients, lymph node invasion was identified in 52 patients (82.5%) alone or associated with either hepatic metastasis (23 patients: 36.5%) or carcinomatosis (6 cases: 9.5%). Hormone status was studied by immunohistochemistry in 37 cases (58.7%). No secretion was evidenced in 14 of the 37 cases (38%>). Serotonin was detected in 17 cases, gastrin in 3 cases, somatostatin and calcitonin in lease. One tumor had multiple secretions. The second group encompassed 64 patients with pancreatic endocrine tumors (PETs). It included 39 male and 25 female. The mean age of the patients at resection was 50.2 years (range, 22-81 years). Pancreatic lesions ranged from 9 to 120 mm (mean: 34.3 mm). According to the OMS classification, pancreatic tumors were classified as benign (OMS la, n=15), of uncertain malignancy (OMS lb, n=13), well differentiated carcinoma (OMS b, n=32), and undifferentiated carcinoma (OMS 3, n=4). At the time of surgery, metastases were present in either lymph nodes (32 cases, 50%) or liver (14 cases, 21.8%), or simultaneously in the two organs (14 cases, 21.8%), or with peritoneal dissemination (1 case). Information on the functional status of the tumors was available in 56 cases (87.8%) (Table 8). Among these PETs, there were 16 nonfunctional tumors (28.5%), 9 somatostatinomas (16%>), 8 glucagonomas (14.3%), 6 pancreatic peptide producing tumors (10.7%>), 5 insulinomas (8.9%), 1 calcitonin producing tumor (1.8%), and 1 gastrinoma (1.8%). Ten lesions (18%) were found to produce multiple secretions (from 2 to 4 peptides).

TP53INP1 is expressed in the normal mucosa of the gastrointestinal tract, including the neuroendocrine cells. Whole sections from normal or diseased organs from the gastrointestinal tract were subjected to immunohistochemical analysis using the monoclonal anti-TP53INPl antibody as described (Example 11. A strong TP53INP1 expression was detected in all cases of normal gastric (Figure 12A), intestinal, and colonic mucosas. TP53INP1 staining was present in epithelial cells of the mucosa and in neuroendocrine cells in witch we have colocalised endocrine granules by using the Grimelius argyrophilic staining (Figure 12B1. No specific TP53INP1 was seen in stromal cells, fibroblasts, smooth-muscle cells, neural tissue, or endothehal cells. At the cellular level, TP53INP1 staining was predominantly cytoplasmic, and minimal nuclear and/or membranous reactivity was observed.

TP53INP1 is not expressed in the normal pancreas islet cells. Sections from normal or diseased pancreas (acute or obstructive chronic pancreatitis associated or not with cancer) were analyzed by immunohistochemistry using the same monoclonal anti-TP53INPl antibody No staining was observed in the Langerhans islets (Figure 12C). TP53INP1 staining was restricted to the epithelial exocrine component of the pancreas, and was enhanced upon stress.

TP53INP1 modification of expression in gastroenteropancreatic neuroendocrine tumors. To determine if TP53INP1 could be differentially expressed in gastroenteropancreatic endocrine tumor cells by comparison with their normal counterparts, a retrospective immunohistochemical analysis was undertalcen. h agreement with results obtained for normal endocrine cells, positive tumor staining was mainly seen in the cell cytoplasm, sometimes associated with membranous or nuclear staining whatever the tumor origin. As described above, two main and opposite features of TP53INP1 expression were identified in the normal neuroendocrine cells from the gastrointestinal tract and pancreas, i.e., TP53INP1 was strongly expressed in the gastrointestinal neuroendocrine cells whereas no expression was evidenced in the pancreatic ones. In tumors, loss of TP53INP1 expression was evidenced in the large majority of gastrointestinal tumors (Table 91. More specifically, 53 out of the 63 gastrointestinal tumors (84.1%_), and 45 out of 64 pancreatic tumors (70.3 %>) did not express TP53INP1. Surprisingly, 19 of the 64 pancreatic neuroendocrine lesions (29.6%) were TP53INP1 positive (Figure 12D and Table 9). Taken together, these results clearly show that modifications in TP53INP1 expression occuned in the gastroenteropancreatic neuroendocrine tumors. TP53INP1 expression in gastroenteropancreatic endocrine tumors is associated with good prognostic factors. The relationships between TP53INP1 expression and the clinicopathological variables were investigated in tumors (Table 91. No conelation was found between TP53INP1 immunoexpression and the hormonal status of the gastrointestinal and pancreatic tumors (p=0.6 and p=0.58, respectively). In gastrointestinal tumors, no conelation was observed between tumor size and TP53INP1 labeling. However, a significant conelation was found between TP53INP1 positive expression and OMS histological low grades (p=0.003): four out of the 5 (80%) lesions classified as OMS 1 were TP53INP1 positive whereas only 6 out of 49 (12.2%) OMS 2 tumors and none of the OMS 3 tumors were found positive. TP53INP1 immunodetection was also able to predict metastasis (p= 0.005). The majority of TP53INP1 negative tumors (47 out of the 53 cases, 88.6%) were associated with metastatic dissemination. A significant conelation between TP53INP1 expression and distant metastasis was shown (p=0.005). In pancreatic tumors, TP53INP1 was more specifically expressed in small tumors: thirteen of 21 tumors of less than 2 cm (61.9%) were TP53HMP1 positive whereas only 6 of 43 (13.9%) large tumors were labeled (pO.OOl). TP53INP1 expression was also conelated with histological grade. The percentage of positive tumors was 80% (12/15) in OMS la, 53% (7/13) in OMS lb, and 0% in OMS 2 and 3 (32 and 4 cases, respectively). TP53INP1 positive expression significantly conelated with low grade proliferation (pO.OOl). TP53TNP1 expression was inverse conelated with metastasis. Finally, none of the TP53DS1P1 positive cases was associated with metastasis (pO.OOl). Extra pancreatic metastasis location was found for the 32 (71,1%) TP53INP1 negative tumors. Altogether, these results clearly indicate that TP53INP1 positive expression is of clinical importance since it was associated with a benign behaviour.

TP53INP1 expression and outcome. Follow-up was available for 74 patients (32 gastrointestinal group, 42 in pancreatic group). Data are summarized in Table 10. Twenty three patients (31%) died of disease. Among the 45 alive patients, 10 (22 %) had recunences or metastasis. Gastrointestinal TP53INP1 negative tumors had a significantly decreased survival with 55% (16 of 29) of patients having a fatal outcome compared with TP53INP1 positive tumors (3.4%). In the pancreatic group, none of the 9 patients with a TP53INP1 positive tumor died of disease, only one locally recuned. In contrast, 6 of 29 patients (20.6%.) who had a TP53INP1 negative tumor died of disease, and 7 patients (24%) were alive with metastasis. The sensibility of the TP53INP1 immunodetection test to predict the absence of metastasis was 90% and 100%> in gastrointestinal and pancreatic group, respectively. The specificity of such a test was 45% and 59%), respectively. Taken together, these results indicate that TP53INP1 could be used as a predictor of survival in the pancreatic neuroendocrine tumors.

Table 1. TP53TNP1 protein expression in normal and diseased pancreas Tissues Loss of TP53INP1 (n)^a value" (X2 test)

Normal pancreas/adjacent normal pancreas 0% (33)

Chronic pancreatitis 3% (31) 0.197

Intraductal papillary mucinous tumors (IPMT) 0% (3)

Mucinous cystic neoplasms 0% (9)

Pancreatic ductal adenocarcinoma 86% (87) O.OOOl

Pancreatic intraepithelial neoplasia (PanlNs):

PanIN-1 0% (24)

PanIN-2 45% (29) 0.0004

PanIN-3 100% (16) 0.0001

Metastasis of pancreatic adenocarcinoma: Liver 100% (4) 0.0001 Lymph node 100% (4) 0.0001

"number of samples; tumor versus the conesponding adjacent non-neoplastic tissue.

Table 2. Conelation between TP53INP1 expression and clinicopatho logical parameters in pancreatic adenocarcinomas

Patient and tumor characteristics ³1 expression (%)^a P value (χ²) Age (yr)

<65 1/13 (8) 0.24 >65 6/26 (23)

Gender

Male 1/22 (5) 0.03 Female 6/21 (29)

Pathological grade

Well differentiated 4/11 (36 ) 0.04 Moderately/Poorly differentiated 2/32 (9)

Clinical stage

I/II 4/17 (24) 0.46 III/IV 2/15 (13)

Tumor size (cm)

<2 4/18 (22) 0.35 2> 1/7 (14)

Tumor location

Head/Body 5/26 (19) 0.58 Tail 1/9 (10)

Nodal involvement

Negative 3/10 (30) 0.27 Positive 4/27 (15)

Metastasis

Negative 7/27 (26) 0.06 Positive 0/11 (0)

Survival (mo)

<6.4 1/14 (7) 0.19 >6.4 4/16 (25) percentage of TP53INP1 expression Table 3. TP53INP1 protein expression in normal and diseased colon

Tissues Loss of TP53INPl (n)^a value" (7.2 test)

Normal colon/adjacent normal colon 0% (47)

Adenomatous polyps: Light dysplasia 7% (14) 0.06 Moderate dysplasia 50% (12) O.0004 Severe dysplasia 50% (6) 0.0001

Colon cancer 76% (50) 0.0001 ^anumber of samples; tumor versus the conesponding adjacent non-neoplastic tissue.

Table 4. TP53INP1 expression in gastric cancer

TP53INP1 Non-malignant Gastric cancer p value gastric tissue (n=142) (n=142)

Positive 139 (98%) 91 (64%) 0.0001 Negative 3 (2%) 51 (36%) ^'

Table 5. Conelation between TP53INP1 expression levels and clinicopatho logic features in gastric cancer

TP53INP1 -positive TP53rNPl -negative p value (n=91 of 142 patients) (n=51 of 142 patients)

Age (years)

< 60 31 18 NS >60 60 33

Sex

Male 68 36 NS

Female 23 15

Location

Cardia 20 22

Body 58 26 0.0193

Autrum 13 3

Histological type

Differentiated^a 62 14

Undifferentiated⁰ 29 37 O.0001

Tumor invasion

Tla+Tlb 23 4 0.0111

T2+T3+T4 68 47

Lymph node metastasis

Positive 5 11 0.0037

Negative 86 40

Metastasis

Positive 0 5 0.0024

Negative 91 46

Lymphatic invasion

Positive 50 42 0.0010

Negative 41 9

NS: not significant

"Differentiated type conesponds to well and moderately differentiated tubular and papillar tumors (intestinal type)

"Undifferentiated type includes poorly differentiated and signet-ring cell carcinomas (diffuse type)

Table 6. Multivariate survival analysis using the Cox regression model

Factor Reference Odds Ratio TP53INP1 + vs - 1.250 0.3680 Age < 60 vs >60 1.388 0.1077 Gender Male vs female 1.158 0.5217 Location Body+cardia vs autmm 1.253 0.4847 Histological type Poor vs well+moderately 2.043 0.0026 Tumor invasion T2+T3+T4 vs TI 1.061 0.8485 Stage III+IV vs II+I 1.269 0.2612 Apoptotic index <4% vs > 4% 2.244 0.0008 Metastasis + vs - 18.688 0.0007 Lymphatic invasion + VS - 0.721 0.2124 Lymph node invasion + s - 3.121 0.0032

Table 8. Clinical histopathological and immunophenotypic features of the study groups

* Available in 79 cases Table 9. TP53INP1 expression in gastrointestinal and pancreatic tumors according to clinicopathological prognostic factors

pos = positive expression neg = negative expression Table 10. Clinical outcome of patients according to their tumor location and TP53INP1 status

AW = Alive and Well without disease AWD = Alive With Disease DOD = Dead of Disease DOC = Dead of Other Cause LFU = Lost of Follow Up pos = positive expression neg = negative expression

Claims

1. The use of a compound enabling the detection of the substantial lack of expression of the TP53INP1 protein in a biological sample, for the manufacture of a drug intended for the prognosis or the diagnosis of cancers and/or of precancerous lesions, or for the screening of drugs active on cancers.

2. The use according to claim 1, of a compound enabling the detection of the substantial lack of expression of the TP53INP1 protein in a biological sample, for the manufacture of a drag intended for the diagnosis of cancers at an early stage.

3. The use according to claim 1 or 2, of a compound enabling the detection of the substantial lack of expression of the TP53INP1 protein in a biological sample, for the manufacture of a drug intended to differentiate malignant tumors, in particular early stage malignant tumors, or precancerous lesions, from healthy tissues and b enign tumors.

4. The use according to any of claims 1 to 3, wherein the compound enables the detection of the substantial absence of the TP53INP1 mRNA-. and/or of the TP53TNP1 protein.

5. The use according to any of claims 1 to 4, wherein the compound is liable to bind to the TP53INP1 mRNA and/or to the TP53INP1 protein.

6. The use according to any of claims 1 to 5, wherein the TP53INP1 mRNA or protein conesponds to the mRNA or to the protein of TP53INP1 isoforms, such as the TP53INPlα; mRNA or protein, or the TP53INP 1 β mRNA or protein.

7. The use according to any of claims 1 to 6, wherein the compound is a monoclonal or a polyclonal antibody which binds to the TP53INP Ice protein and/or to the TP53TNPljS protein.

8. The use according to any of claims 1 to 7, of the monoclonal antibody secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Institut Pasteur, Paris, France) on March 26, 2004, under accession number CNCM 1-3194.

9. The use according to any of claims 1 to 6, wherein the compound comprises at least one nucleotide sequence, said nucleotide sequence containing at least 10 contiguous nucleotides selected from SEQ ID NO: 1 or 3, or from the sequence complementary to SEQ ID NO: 1 or 3, or at least one sequence derived from one of said nucleotide sequence, by insertion, deletion or substitution of at least one nucleotide and presenting at least 90% similarity with the nucleotide sequence from which it derives.

10. The use according to claim 9, wherein the nucleotide sequence conesponds to SEQ ID NO: 1 complementary sequence or to SEQ ID NO: 3 complementary sequence.

11. The use according to any of claims 1 to 10, wherein "the detection is canied out with a method selected from: PCR, real-time PCR, RT-PCR, NASBA, Northern Blot, in situ hybridization, chromatin precipitation, ELISA, Western BLot, far Western protein interaction assay, immunoprecipitation, FACS, flow cytometry, cytochemistry, cytofluorescence, immunofluorescence, immunohistochemistry.

12. The use according to any of claims 1 to 11, wherein the cancers are selected from the group of epithelial cancers comprising pancreas cancer, colon cancer, lung cancer, prostate cancer, bladder cancer, breast cancer, uterine cervix cancer, ovary cancer, head and neck cancer, oral cancer, skin cancer, kidney cancer, thyroid cancer, stomach cancer, conjonctiva- cornea cancer, glioma, thymoma, and neuroblastoma.

13. The use according to any of claims 1 to 12, wherein the precancerous lesions are selected from precancerous lesions leading to cancers selected fromt the group of epithelial cancers comprising pancreas cancer, colon cancer, lung cancer, prostate cancer, bladder cancer, breast cancer, uterine cervix cancer, ovary cancer, head and neck cancer, oral cancer, skin cancer, kidney cancer, thyroid cancer, stomach cancer, conjonctiva_-conιea cancer, glioma, thymoma, and neuroblastoma, or from precancerous lesions arising from inflammatory bowel disease (IBD), such as ulcerative colitis (UC) or Crohn's disease, and endo-brachy-oesophagus (EBO).

14. The monoclonal antibody secreted by the hybridoma deposited under the Budapest Treaty at the CNCM (Collection Nationale de Culture de Microorganismes, Institut Pasteur, Paris, France) on March 26, 2004, under accession number CNQV 1-3194.

15. A cancer prognosis or diagnosis kit comprising:

- at least one anti-TP53rNPl protein antibody, in particular a labelled anti-TP53ENPl protein antibody, - a sample of the TP53INP1 protein as a control or a reference.

16. A cancer prognosis or diagnosis kit according to claim 15, wherein the antibody conesponds to the monoclonal antibody of claim 14.

17. A method for the in vitro diagnosis of cancers at an early stage, of cancers, or of precancerous lesions, characterized in that it comprises the following steps: - contacting a tissue sample taken from an individual with a anti-TP53INPl protein antibody, in particular a monoclonal antibody, such as the monoclonal antibody of claim 14, - washing the tissue sample, - revealing the presence, or the absence, of said antibody, in the washed sample, the substantial absence of said antibody in the washed sample as compared to a sample of a conesponding healthy tissue being an indication that said tissue is afflicted with a cancer at an early stage, with a cancer, or with a precancerous lesion.

18. An anti-cancer drug screening method, characterized in that it comprises the following steps:

- contacting a sample of precancerous or cancerou-s cells, in which the TP53INP1 protein is substantially absent, with a molecule liable to be an anti-cancer drug, - detecting the substantial presence or the substantial absence of the TP53INP1 protein, in the cell sample which has been contacted with said molecule to screen, with an antibody, in particular the monoclonal antibody of claim 14,

- selecting the molecule responsible for the presence of the TP53INP1 protein in the cells, as an anti-cancer drug.