WO2000000503A1 - Human az-1 gene, variants thereof and expressed gene products - Google Patents

Abstract

A human AZ-1 gene, mutants, variants and fragments thereof. Protein products encoded by the AZ-1 gene and homologs encoded by the variants of AZ-1 gene acting as tumor suppressors or markers of malignancy progression and tumorigenicity reversion. Identification, isolation and characterization of AZ-1 and AZ-2 genes localized to a tumor suppressive locus at chromosome 10q26, highly expressed in nonmalignant and premalignant cells derived from a human breast tumor progression model. A recombinant full length protein sequences encoded by the AZ-1 gene and nucleotide sequences of AZ-1 and AZ-2 genes and variant and fragments thereof. Monoclonal or polyclonal antibodies specific to AZ-1, AZ-2 encoded protein and to AZ-1, or AZ-2 encoded protein homologs.

Description

HUMAN AZ-1 GENE. VARIANTS THEREOF AND EXPRESSED GENE PRODUCTS

BACKGROUND OF THE INVENTION

Field of the Invention

This invention concerns a novel human AZ-1 gene, mutants, variants and fragments thereof, and protein products encoded by the AZ-1 gene and homologs encoded by the variants of AZ-1 gene acting as tumor suppressors or markers of malignancy progression and tumorigenicity reversion. In particular, this invention concerns identifying, isolation and characterization of novel AZ-1 and AZ-2 genes localized to a tumor suppressive locus at chromosome 10q26, highly expressed in nonmalignant and premalignant cells derived from a human breast tumor progression model. Additionally, the invention concerns findings that AZ-1 and AZ-2 genes exhibit tissue-specific expression profiles and that AZ-1 gene expression in tumor cells is low or absent. The invention further concerns recombinant full length protein sequences encoded by the AZ-1 gene and nucleotide sequences of AZ-1 and AZ-2 genes. The invention also concerns monoclonal or polyclonal antibodies specific to AZ-1, AZ-2 encoded protein and to AZ-1, or AZ-2 encoded protein homologs.

BACKGROUND AND RELATED DISCLOSURE The evolvement of breast cancer is a multistep and cumulative process and understanding of the genetic and phenotypic alterations in successive steps is essential for designing therapeutic interventions and diagnostic assays.

In the human body, the epithelial component of the breast is embedded in the stroma and forms a branching ductal structure that emanates from the nipple, repeatedly bifurcates, and terminates in lobules, alveoli, and end buds. Although stroma accounts for >80% of the breast volume, approximately 95% of the cancers produced in the breast are of epithelial origin. To elucidate the advancement of human breast cancer, a functionally relevant cell culture model is required. The differences in breast tissue compartmentalization, phenotypic characteristics, and the mutagenic frequency between human and rodents underline the need to develop a human breast cell model.

An unconventional spontaneously-transformed HMT-3522 cell lines was described recently in Cancer Research, 56:2039 (1996) where the immortalized human mammary epithelial cells (SI) established from fibrocystic breast tissue was propagated in chemically defined medium described in In Vitro Cell. Dev. Biol._f 23:181 (1987). SI cells were near-diploid and expressed luminal epithelial cell differentiation markers cytokeratin-18 and sialomucin. Genetic changes such as p53 point mutation (Exp. Cell. Res. _r 215:380 (1994)) and c-jjiyc amplification (Cancer Research, 52:1210 (1992)) have already been noted in later passages, i.e., >50 of the nonmalignant Si cells. In passage 118, cells were adapted to grow in epidermal growth factor (EGF) -free medium and a new EGF-independent subline (S2, premalignant) was isolated (ibid).

Alterations in gene expression of EGF receptor, transforming growth factor (TGF)-α and c-erb-B2 were seen in S2 cells. The established S2 cell line underwent cytogenetic evolution and exhibited genomic instability and heterogeneity (Cancer Genetics and Cytoσenetics _r 78:189 (1994)). Nevertheless, S2 cells remained nontumorigenic until passage 238. At that time, it was able to induce tumor growth in nude mice. Following second round of mouse transplantation, another subline (T4-2, tumorigenic) existed as a relatively homogenous malignantly-transformed cell population. One extra copy of chromosome 7p which harbors EGF receptor gene was found in the T4-2 cells (Cancer Res., 56:2039 (1996)).

Detection and suppression of cancerous tissue growth is of extreme interest and importance. It is, therefore, a primary objective of this invention to provide means for identification, detection and suppression of cancerous growth in breast tissue.

All patents, patent applications and publications cited herein are hereby incorporated by reference.

One aspect of the current invention is a human AZ-1 ' gene or a variant, mutant, or fragment thereof.

Another aspect of the current invention is a nucleotide sequence of AZ-1 and AZ-2 genes, identified as SEQ ID NO: 1 and SEQ ID NO: 2. Another aspect of the current invention is a DNA sequence identified as SEQ ID NO: 1 encoding a protein comprising the amino acid sequence identified as SEQ ID NO: 3.

Still another aspect of the current invention is a protein of the amino acid sequence identified as SEQ ID NO: 3.

Still yet another aspect of the current invention is a protein encoded by AZ-1 gene, or by a variant, mutant or fragment thereof, or any protein containing said protein encoded by AZ-1 gene, variant, mutant, or fragment thereof, or any protein which shares homology with AZ-1 encoded protein or AZ-2 encoded protein, variant, mutant or fragment thereof.

Still another aspect of the current invention is a protein encoded by the AZ-1 gene, variant, mutant, analog or fragment thereof acting as a tumor suppressor or a marker of malignancy progression and tumorigenic reversion.

Still yet another aspect of the current invention is a method for diagnosis and detection of progression of human breast cancer by detecting presence and quantity of a protein identified as SEQ ID NO: 3 or a homolog thereof in human breast cells or tissue or by detecting expression of AZ-1 gene, mutant, variant, fragment thereof by in situ hybridization or RT-PCR. Still yet another aspect of the current invention is a diagnostic method for detection of the presence of AZ-1 protein in human breast cancer by treating a biopsy sample of a subject patient with a polyclonal or monoclonal AZ-1 antibodies or detection of the level of AZ-1 message.

Yet another aspect of the current invention is a method for prevention or treatment of human breast cancer by providing a subject in need thereof a therapeutically' effective amount of a protein identified as SEQ ID NO: 3 or homolog thereof, able to act as a tumor suppressor of human breast cancer cells.

Still yet another aspect of the current invention is a tissue targeted gene therapy for treatment of human breast tumor.

Still another aspect of the current invention is an

ELISA kit for detection of expression of AZ-1 gene or a variant thereof by detecting presence or absence of AZ-1 encoded protein with AZ-1 monoclonal or polyclonal antibodies.

Still another aspect of the current invention is a message detection kit for detection of expression of AZ-1 gene or a variant thereof by bDNA technology (Quantigene Gene Expression Assay, Chiron Corp., Emeryville, CA, in situ hybridization or RT-PCR by detecting presence or absence of AZ-1 message.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 is a schematic diagram of malignant transformation of HMT3522 human breast epithelial cells.

Figure 2 shows phenotypic recapitulation of tissue morphology in 3-D reconstituted basement membrane (rBM) culture assay for SI and T4-2 cells.

Figure 3 shows characterization of HMT-3522 progression series. Figure 3A shows comparative genomic hybridization (CGH) performed to determine genomic content and alteration in cells throughout the HMT-3522 progression series. Figure 3B are phase contrast micrographs showing the morphology of mammary epithelial cells (MECs) recapitulated in 3-D rBM assay. Figure 3C shows immunostaining with Ki-67 marker. Figure 3D shows organized cortical F-actin in S-l acini and disorganized F- actin filaments in S-2 and T4-2 tumor colonies. Figure 3E shows E-cadherin staining pattern in SI, S2 and T4-2 cells. Figure 4 is the cDNA sequence (SEQ ID NO: 1) of AZ-1 gene showing a divergent site between nucleotides 429-430. Figure 5 is the cDNA sequence (SEQ ID NO: 2) of AZ-2' gene showing a divergent site between nucleotide 764-765.

Figure 6 is an amino acid sequence (SEQ ID NO: 3) of the protein encoded by AZ-1 gene. Figure 6A shows SPAZI (SEQ ID NO: 7), REGION I (SEQ ID NO: 13), REGION II (SEQUENCE ID NO: 14) and Coiled-Coil (CCD) (SEQ ID NO: 8) domains. Figure 6B is a schematic diagram of all AZ-1 and TACCl domains. Figure 6C shows SPAZI domain homology for AZ-1 (SEQ ID NO: 3), TACCl (A) (SEQ ID NO: 9), TACC2 (B) (SEQ ID NO: 15), TACC3 (SEQ ID NO: 16), and BCK1 (SEQ ID NO: 10) protein. Figure 6D shows homology regions for coiled-coil domain (CCD) in AZ-1 (SEQ ID NO: 8), TACCl (SEQ ID NO: 11), TACC3 (SEQ ID NO: 17) and SB1.8 (SEQ ID NO: 12) genes.

Figure 7 shows Western blot analysis of AZ-1 recombinant proteins.

Figure 8 is a Western blot illustrating recognition of a 64kD protein by AZ-1 antibody.

Figure 9 illustrates differential expression of AZ-1 gene in premalignant and breast tumor cell lines. Figure 9A shows differential display analysis in premalignant S2 and T4 tumor cells. Figure 9B shows Northern blot analysis in premalignant S2 and T4 tumor cells.

Figure 10 shows downregulation of AZ-1 in breast tumor cell lines and biopsies. Figure 10A shows expression of AZ-1 in SI and MCFIOA nonmalignant cell lines and in luminal epithelial and myoepithelial nonmalignant primary cells. Figure 10B shows downregulation of AZ-1 gene in premalignant S2 and malignant T4, HMT 3909, MCF-7, CAMA-1, MDA-MB 468, SKBR-3 , T47D, MDA-MB 231, Hs578T and BT 549 cells. Figure 10C shows downregulation of AZ-1 gene in in situ carcinoma.

Figure 11 shows tissue specific expression of AZ-1 in various normal human tissues.

Figure 12 is a color image showing localization of AZ- 1 gene to chromosome 10q26.

Figure 13 is a color image which shows the association of AZ-1 with cytoskeletal complexes in nonmalignant breast^* cells. Figure 13A is the image taken before, and Figure 13B is the image taken after, treatment with triton X-100 to remove soluble proteins.

Figure 14 is a color image of in situ staining of AZ-1 in normal breast acinus (Figure 14A) and breast duct (Figure 14B) .

Figure 15 illustrates the presence of E-cadherin and β-catenin in AZ-1 immunocomplexes.

Figure 16 shows that ectopically-expressed AZ-1 gene reduces tumorigenicity and invasiveness. Figure 16A shows

Northern blot of AZ-1 and Western blot of AZ-1 in SI, T4-2

(mock) and T4-2+AZ-1 cells. Figure 16B shows the number of colonies of SI, T4-2, T4-2 (mock) cells and reduction in number of colonies in T4-2 cells in the presence of AZ-1 (T4-2+AZ-1) grown in soft agar. Figure 16C shows in vitro invasiveness in SI, T4-2, T4-2 (mock) and T4-2+AZ-1 cells.

Figure 17 shows ectopically-expressed AZ-1 induces tissue morphogenesis in three-dimensional cultures. Figure

17A shows phase contrast images for SI, T4-2 (mock) and T4- 2+AZ-l cells. Figure 17B shows basement membrane staining of T4-2 (mock) and T4-2+AZ-1 cells. Figure 17C shows the colony size in μm for SI, T4-2 (mock) and T4-2+AZ-1 cells.

Figure 18 illustrates upregulation of AZ-1 in morphologically reorganized breast tumor cells. Figure 18A is a schematic diagram of morphology of T4 , SI, T4βl and

T4tyr, phase contrast microscopy (Figure 18B) , AZ-1 message

(Figure 18C) and GADPH message (Figure 18D) .

Figure 19 shows a sequence alignment of AZ-1 and its variant TACC2 cDNAs. Sequence insertions are indicated by dots .

Figure 20 shows an amino acid sequence alignment of AZ-1 and its variant TACC2 proteins. Sequence insertions are indicated by dots.

DEFINITIONS

As used herein,

"CCD" means coiled-coil domain. "ER" means estrogen receptor.

"RACE" means rapid amplification of cDNA ends.

"AZ-1" or "AZ-1 gene" means antizuai-1 gene. AZ-1 refers to the entire AZ-1 genomic sequence, including enhancers, promoter sequences, introns and exons. "AZ-1 mutant", "AZ-1 variant" or "AZ-1 fragment" mean all AZ-1 gene products, such as all potential splice variants derived from the gene and their protein products. Their associated 5 ' and 3 ^• untranslated regions are also included. Also included are all mutant forms of AZ-1, whether spontaneously arising or specifically engineered. All known AZ-1-related genes, such as TACC-1, TACC-2 and TACC-3 are also included under this definition as long as they share about 25% of homology.

"AZ-1 gene encoded protein" means and includes all protein coding sequences encoded by AZ-1 gene, variant or mutant thereof, as well as 5' and 3* untranslated sequences identified here and in other AZ-1 splice variants.

"HMT-3522" means human mammary tumor cell line 3522.

"Ki-67" means a marker for proliferating cells. "Cadherin" means cell-adherens junction protein.

"E-cadherin" means epithelial cadherin.

"β-catenin" means an adherens junction protein.

"GAPDH" means a metabolic protein GAPDH (glyceraldehyde phosphate dehydrogenase) expressed by a common gene in metabolic pathway. In this application, the message level of GADPH is used as control for RNA loading in Northern blot.

"SI" or "S-l" means a nonmalignant breast cell line.

"S2" or "S-2" means a premalignant breast cell line. "MCFIOA" means a nonmalignant breast cell line.

"Luminal epithelial cells" means normal cells present in the breast tissue. "Myoepithelial cells" means normal cells present in the breast tissue.

"T4", "MCF-7", "CAMA-1", "BT-20", "MDA-MB 468", "SKBR-3", "T47D", "MDA-MB 231", "Hs578T", and "BT 549" means breast tumor cell lines specifically so identified.

"HMT 3909" means a breast tumor cell line contaminated with normal myoepithelial cells.

"T4-2 (mock)" means breast tumor cells transfected with empty expression vector. "Ectopically expressed AZ-1" means artificially expressed or overexpressed AZ-1 gene.

"Upregulation of AZ-1 gene" means observed increased expression of AZ-1 gene.

"Variant" means any variant derived from splicing, exon shuffling, deletion or insertion causing frame shifting. Variant is exemplarized by AZ-2 gene and TACC2 gene where TACC2 gene is a variant of AZ-1 gene.

"Mutant" means AZ-1 gene containing a point mutation, deletion, insertion, or truncation. "Fragment" means a functional domain, such as SPAZI or coiled-coil domain involved in protein-protein interaction.

"Homolog" means any homologous protein sharing a substantial sequence similarity (about 25% or more) with

AZ-1 expressed protein. Exemplary proteins are proteins expressed by TACCl or TACC3 or a variant thereof expressed by TACC2.

"TACCl" means embryonically expressed TACCl gene from the 8pll breast cancer amplicon.

"TACC2" means expressed TACC2 gene from the 10q25-q26 locus.

"TACC3" means expressed TACC3 gene from the 4pl6.3 locus .

"SPAZI" means serine-proline rich AZ-1 domain.

"Coiled-coil" means heptad repeats participating in protein-protein interactions.

"BCK1" means a Saccharomyces cerevisiae yeast gene.

"BLAST" means basic local alignment sequence tool. BLAST is a service available from the National Center for Biotechnology Information which compares and matches a nucleotide or protein sequence against databases at the NCBI. DETAILED DESCRIPTION OF THE INVENTION

This invention relates to a cancer-related gene AZ-1. AZ-1 gene is novel and has never before been described or disclosed. AZ-1 gene and its encoded protein have been found to be present in normal breast cells. The expression of AZ-1 gene in tumor cells, however, has been found to be significantly decreased in ten human breast cancer cell lines and carcinoma cells in situ . The level of the AZ-1 encoded protein is significantly decreased in these cell lines. A protein encoded by AZ-1 gene is believed to act as a protective agent against cancer by suppressing a tumor growth and the detection of its level in breast cells is useful as a marker of malignancy progression and tumorigenic reversion. The invention is useful for diagnosis, prognosis and treatment of breast cancer.

The invention also relates to identification, isolation and sequencing of the human AZ-1 gene and its variant AZ-2 gene encoding proteins acting as tumor suppressors and markers for tumor progression and tumorigenicity reversion.

AZ-1 gene was isolated and sequenced (SEQ ID NO: 1) and the amino acid sequence of AZ-1 encoded protein was deduced (SEQ ID NO: 3) . The protein was additionally identified by specific AZ-1 antibodies. Functional significance of the loss of AZ-1 expression in tumor cells has been investigated in vitro and in vivo and linked to tumorigenic and invasion suppressive roles.

Additionally, the invention relates to a method for treatment, detection and prognosis of breast cancer by providing a patient in need of such treatment a therapeutically effective amount of a recombinant protein, by detecting the level of native protein encoded by AZ-1 gene in the breast tissue biopsy sample or by determining a degree of expression of AZ-1 gene and/or levels of expressed AZ-1 protein.

I. Tumoriσenic Cell Line Progression Model The current invention was developed and tested ori various models of normal or breast tumor cell lines which were developed and are described below. Progression model for tumorigenic cell line T4-2 was developed by malignant transformation of human breast epithelial cells. Presence or absence of AZ-1 gene expression was tested in normal epithelial or myoepithelial cells, nonmalignant SI cell line, premalignant S2 cell line and in T4-2 breast tumor cell line.

Malignant T4-2 breast tumor cell line was derived from nonmalignant fibrocystic breast cells. Malignant transformation of nonmalignant HMT3522-S1 cell line into malignant T4-2 cells is illustrated in Figure 1.

The nonmalignant cells were obtained as a cell mixture from a patient suffering from fibrocystic disease. The cells were cultured to specifically promote the growth of epithelial cells. The epithelial cells were then immortalized as a nonmalignant HMT 3522-S1 cell line. The HMT 3522-S1 cells were repeatedly passaged 20 (S-l 20) , 50 (S-l 50), 110 (S-l 110), 175 (S-l 175) up to 400 passages (S-l 400) in the presence of EGF (+EGF) and were found to be nonmalignant. When, after 110 passages, epidermal growth factor (EGF) was removed from the culture, at about 150 passages, the EGF deprived cells turned into premalignant S-2 cell line. After the S-2 cell line was further propagated up to S-2 215, the S-2 cells at every 10-passage intervals from 150 passages up were injected into nude mice. None of the S-2 cells up to passage 238 were found to produce tumors. After about 8 weeks, approximately 50% of the S2-238 injected nude mice were found to produce tumors. Their tumor nodules were then removed and the cells were further propagated and these cells were again injected into the nude mice. After 8 weeks, more than 90% of nude mice were found to have tumors. The tumor nodules were removed again and the cells were propagated as T4-2 cells. T4-2 cells were tested for AZ-1 gene expression and/or for the presence or absence of AZ-1 encoded protein in assays described below.

Normal cells, cell lines and malignant cell lines were then tested in a series of morphological, structural and adhesion studies and differences between normal S-l and malignant T4-2 cell lines observed in these studies are seen in Figure 2. Figure 2 is a comparative phenotypic recapitulation of normal S-l cell line and malignant T4-2 cell line in 3-D basement membrane culture assay.

As seen in Figure 2, a schematic diagram of S-l and T4-2 cell lines (extreme left) shows S-l cells are organized in a sphere, which corresponds to its morphological organization seen in morphology subset (middle left) . In morphology subset, normal S-l cell are seen to be organized in spherical manner. The malignant T4-2 cells on the other hand, are shown to form disorganized colonies, seen in both schematic and morphology subsets.

The normal S-l cells were seen having a smooth spherical shape basement membrane when tested in a 3-D laminin-rich reconstituted basement membrane culture assay and immunostained with human type IV collagen (middle right) . When the malignant cells were analyzed by human type IV collagen immunostaining, they revealed a disorganized pattern. Staining with E-cadherin (E-cadherin subset, extreme right) shows intact cell-cell adhesion network in SI cells and disrupted organization in T4-2 cells.

Differences between the normal nonmalignant, premalignant and malignant cells which were obtained in progression series of HMT 3522, seen in Figure 1, are further illustrated in Figure 3, which shows characterization of the HMT 3522 progression series. All types of cells, that is the normal S-l (Sl-50, Sl-110 and Sl-175) cells, premalignant S2 (S2-215) cells and malignant T4-2 (T4-2-25) cells obtained as described in Figure 1 were analyzed by comparative genomic hybridization, phase contrast microscopy, and immunostaining with F-actin phalloidin staining, and with E-cadherin.

The HMT-3522 human breast tumor progression series is comprised of a continuum of genetically related cell populations that range in phenotypic behavior from nonmalignant to tumorigenic. In order to identify genes that play a crucial role in the final stages of breast tumor progression, a differential display strategy was utilized to compare the gene expression patterns of the model's tumorigenic cell population (called T4-2) with that of its pre-malignant progenitor (called S2) . One of the genes identified using this approach is abundantly expressed in non-malignant luminal epithelial cells but is expressed at significantly lower levels in a variety of breast tumor cell types. Because of this expression pattern, which is commonly observed with tumor suppressors of the Type II class, this gene is referred to as anti-zuai-1 (or AZ-1, with "zuai" meaning breast cancer in Chinese) .

To determine genomic content and alteration in cells throughout the HMT-3522 progression series, comparative genomic hybridization was performed. Results are seen in Figure 3A. As shown in Figure 3A, genetic alteration, e.g., chromosomal amplification and deletion, were accumulated in HMT 3522 cells during tumorigenic progression.

Figure 3B are phase contrast micrographs showing the morphology of mammary epithelial cells (MECs) as identified at the top of the Figure 3, i.e., nonmalignant SI cells (Sl-50, Sl-110 and Sl-175), premalignant S-2 cells (S2-215) and malignant T4-2 passage 25. In this study, cells were grown embedded in a (3-D) laminin-rich basement membrane (BM) for 10 days. At that time, SI cells formed growth- arrested structures reminiscent of true acini and S2 premalignant and T4-2 tumor cells formed large, irregular colonies.

Figure 3C shows immunostaining of tested cells for Ki- 67, a marker of cell cycle entry. SI passages 50 and 110 were negative, that is, they did not show any Ki-67 immunostaining, while the percentage of Ki-67 positive' cells increased from Sl-175 to the T4-2 tumor cells. These results show a loss of growth-arrest control in response to the BM occurs in progression to malignancy. Sl-175 cells remained organized, but the acini were larger. Figure 3D shows results of staining the cells with F- actin. F-actin phalloidin staining shows organized cortical F-actin in the SI cells while both S2 premalignant and T4-2 tumor colonies are seen to contain only disorganized actin filaments. Figure 3E illustrates E-cadherin studies. E-cadherin immunofluoresence, seen in Figure 3E, shows organized cell- cell contact staining in the SI cells, while in both the S2 premalignant and T4-2 tumor colonies, the cells had punctate, dispersed membrane and intracellular stainings. II. AZ-1 Gene and AZ-1 Encoded Protein

AZ-1 gene has been discovered to be present and expressed in abundance in normal nonmalignant breast cells.

Functional studies indicate that overexpression of the

AZ-1 message in tumorigenic T4-2 cells is sufficient to reduce their tumorigenic phenotype as measured by growth in soft agar, invasion assays and by tumor formation in nude mice. Moreover, overexpression of AZ-1 restores T4-2 cells with a normal polarized phenotype when cultured in a 3- dimensional reconstituted basement membrane. Collectively, these results indicate that AZ-1 gene and/or its protein product is a candidate breast tumor suppressor that may play a role in attenuating cell growth controlling disorganized malignant growth and maintaining appropriate tissue architecture. A. Isolation. Identification and Sequencing of AZ-1

Osns. The AZ-1 gene has been now identified, isolated, sequenced (SEQ ID NO: 1) and its variant AZ-2 gene (SEQ ID NO: 2) nucleotide sequence has been determined. The sequence of AZ-1 gene is shown in Figure 4 as SEQ ID NO: 1. The sequence of AZ-2 gene is shown in Figure 5. Sequences of the related gene TACC2 gene is identified as SEQ ID NO: 5. Other homologs of AZ-1 protein, TACCl (A), TACCl (B) , TACC3, BCKl and SB1.8 are identified as SEQ ID NO: 9, SEQ ID NO: 15, SEQ ID NO: 21, SEQ ID NO: 10 and SEQ ID NO: 12, respectively. 1. Sequences of AZ-1 Gene and Variants

AZ-1 gene is localized to a tumor suppressive locus at chromosome lOq 26 genomic locus. AZ-1 is a novel gene whose sequence gives rise to an AZ-1 protein product (SEQ ID NO: 3) . The AZ-1 protein comprises 571 amino acids and is comprised of 4 distinct structural/sequence domains, two of which, namely coiled-coil and SPAZI domains, represent conserved motifs. The protein, its amino acid sequence and the four domains as seen in Figure 6.

AZ-1 gene, of which cDNA is shown in Figure 4, comprises a nucleotide sequence of 3813 nucleotides. The sequence is identified as SEQ ID NO: 1. The sequence of one of the AZ-1 gene variants, namely, AZ-2 gene is shown in Figure 5. AZ-2 cDNA sequence is identified as SEQ ID NO: 2. AZ-2 cDNA is longer and contains 4148 nucleotides. The divergent site is between nucleotide 764 and 765. The cDNA of AZ-1 gene variant TACC-2 is identified as SEQ ID NO : 5. Two genes which expressed homologous proteins to AZ-1 protein, namely, TACC-1 and TACC-3 gene, are identified as SEQ ID NO: 18 and SEQ ID NO: 20. AZ-1 and AZ-2 genes are diverged at a location marked in Figures 4 and 5 as "T//T", positioned at nucleotides 429 and 430 of AZ-1 gene and at nucleotides 764 and 765 of AZ-2 gene, respectively. AZ-1 and AZ-2 genes share identical sequence downstream of the divergent site indicated above. l. AZ-1 Encoded Protein

AZ-1 sequence encodes a protein of 571 amino-acids. Sequence of AZ-1 encoded protein is identified as SEQ ID NO.:3. The AZ-1 variant AZ-2 encodes a protein of 1219 amino acids. Sequence of AZ-2 encoded protein is identified as SEQ ID NO: 4. Another AZ-1 variant TACCl gene encodes protein identified as SEQ ID NO: 19. Figure 6 shows amino acid sequence of the protein' encoded by AZ-1 gene (SEQ ID NO: 3) .

Figure 6A identifies amino acids in the sequence and separates them into four domains. The SPAZI domain contains amino acids 1-107 (SEQ ID NO: 7) . REGION I (SEQ ID NO: 14) contains amino acids 109-248 (SEQ ID NO: 13) . REGION II contains amino acids 250-360. The last and largest coiled-coil domain (CCD) contains amino acids 362- 571 (SEQ ID NO: 8) .

Figure 6B is a schematic diagram comparing AZ-1 gene with TACC-1 gene. As seen, there are different degrees of homology in the SPAZI REGION I and coiled-coil domain of AZ-1 and TACC-1. There is a deletion in TACC-1 in the region corresponding to REGION II of AZ-1 gene. Figure 6C shows point of homology between AZ-1 (SEQ ID NO: 7) , TACC-1 (A) (SEQ ID NO: 9), TACC-1 (B) (SEQ ID NO: 15), TACC-3 (SEQ ID NO: 16) and BCKl (SEQ ID NO: 10) in the SPAZI domain.

Figure 6D shows points of homology between AZ-1 (SEQ ID NO: 8), TACC-1 (SEQ ID NO: 11), TACC-3 (SEQ ID NO: 17) and SB1.8 (SEQ ID NO: 12) in coiled-coil domain. Sequence analysis of the full-length AZ-1 cDNA clone reveals an open reading frame that translates to a protein product of 571 amino acids (Figure 6A) . At its N-terminus AZ-1 contains a novel serine and proline-rich domain, called a SPAZI domain, that is evolutionarily conserved and is predicted to exhibit an Ig-domain like fold. The C- terminal region of AZ-1 displays a series of heptad repeats consistent with the presence of an extensive, but discontinuous, coiled-coil domain.

Using this protein product sequence as the query for a PSI-BLAST database search (Nucleic Acids Res. , 25:3389 (1997)), AZ-1 was found to share significant sequence similarity, particularly at its N- and C-termini, to TACCl gene (Genbank locus AF049910) cDNA (SEQ ID NO: 18) , the unpublished product of a gene cloned from the breast cancer amplicon 8pll. A second, unpublished gene product, TACC2 (AF095791) cDNA (SEQ ID NO: 5) , seems to be a splice variant of AZ-1 gene since, apart from two insertions, it is identical to AZ-1 at both the nucleic acid and protein levels (Figures 19 and 20) .

Inspection of the alignment of AZ-1 and TACCl summarized in Figure 6B suggests that AZ-1 can be partitioned into four domains. At its N-terminus, AZ-1 exhibits serine- and proline-rich "SPAZI" domain. SPAZI domain is shown in Figure 6C. The SPAZI domain of AZ1 is compared to corresponding TACCl (A) , TACC2 (B) , TACC3 and BCKl domains. The serine and proline residues, which are distributed throughout this protein region, each comprise 18% of the AZ-1 sequence content for an overall proline/serine content of 36%. SPAZI domains are present once in AZ-1, twice in TACCl and once in the Saccharomyces cerevisiae gene product BCKl, a member of the MAPKKK (mitigen activator protein kinase kinase kinase) family of serine/threonine kinases. In many instances, the abundant serine and proline residues are conserved in 2 or more of these sequences; 2 serine residues in the second half of the motif are invariant. Fold recognition studies, using the GenTHREADER program (J. Mol. Biol., 287:797 (1999)), indicate that the SPAZI domain is likely to display an Ig-like beta-sandwich fold. For each SPAZI domain, at least one protein having a known, immunoglobulin-like (Ig-like) structure was reported. The most reliable prediction with estimated probability of correct match = 0.59, was for the BCKl SPAZI domain which matched an Ig-like domain in human Cd2 , T lymphocyte adhesion glycoprotein, PDB (Protein Data Bank) identifier lhnf. Based on these structural predictions, the SPAZI domain seems to be a new member of the hnf or C2- set superfa ily.

The SPAZI domain of AZ-1, as seen in Figure 6B, is followed by two domains, referred to as REGION I and REGION II, that are defined by virtue of their relationship to TACCl. REGION I of AZ-1 shows 20% identity with parallel amino acids of the TACCl sequence. REGION II corresponds to those sequences of AZ-1 that are absent from TACCl gene ' product. Fold recognition analyses of REGION I and REGION II predict that they too have Ig-like folds. AZ-l*s REGION I matches known immunoglobulin structures, namely, PDB lpfc, a fragment of an IgGl with estimated probability of match = 0.32, and PDB lpsk, an antibody Fab fragment, with estimated probability of match = 0.51. Analysis of the third region present only in AZ-1 indicates a beta barrel fold, namely, PDB lhtp, H-protein with estimated probability of match = 0.27. The fourth and C-terminal region of AZ-1 displays a series of heptad repeats consistent with the presence of an extensive, but discontinuous, coiled-coil domain seen in Figure 6D. Structural studies have demonstrated that coiled-coil domains, like the one found in AZ-1, form amphipathic helices that associate with other like domains to form superhelical bundles comprised of anywhere from 2 to 4 helices. The seven structural positions of a heptad repeat are named a-g. Positions a and d, occupied by hydrophobic residues, form the helix interface whereas the remainder are hydrophilic and form the solvent-exposed part of the helix surface.

Apart from the closely related TACCl coiled-coil domain, the highest scoring sequence from a PSI-BLAST search is the human SB1.8/DXS423E protein, a putative homologue of the Saccharomyces cerevisiae SMC-1 protein that is essential for proper chromosomal segregation during mitosis (PIR locus 154383) . Alignments generated using the Multicoil program (Protein Sci.. 6:1179 (1997)) indicate three major regions where the characteristic heptad repeats fall into register in all three proteins (Figure 6D) . These regions in AZ-1, TACCl and SB1.8 correspond to regions that the Multicoil program predicts to form dimers (probability >0.90) . The two d positions towards the end of the coiled coil are occupied by notably charged residues E and K.

Cellular localization predictions generated using the PSORT program indicate that AZ-1 contains two putative' nuclear localization sequences (NLSs) . One NLS is positioned N-terminally in the SPAZI domain, while the second NLS starts at amino acid 122 in AZ-l's REGION I shown in Fig. 6A as underlined. 2. AZ-1 Gene Cloning

AZ-1 gene was identified, sequenced and cloned using methods known in the art.

Specifically, the nucleotide sequence of the 180 bp differential display cDNA fragment was determined and compared to existing Genbank sequences. The sequence of the 180 bp fragment was identical to three ESTs. All three sequences contained the 180 bp sequence plus additional 5' and/or 3* sequences; two of these clones exhibited polyadenylation sites. None displayed significant open reading frames, thereby indicating that the 180 bp fragment resided in the 3' untranslated region of the gene product. This information indicated that the remainder of the gene product would be positioned 5 ' to the isolated fragment and could be cloned by performing multiple rounds of 5 • RACE (rapid amplification of cDNA ends) , according to protocols from Life Technologies, Inc. Grand Island, NY) .

Primers corresponding to the 180 bp differential display fragment were used to initiate the 5 ' RACE cloning procedure according to the manufacturer's instructions. Ultimately the protocol was repeated 12 times to obtain 3.8 kb of AZ-1 sequence. In each cycle, 500-800 bp of additional overlapping 5 '-end sequence was obtained. Sequencing of the 5 ' RACE PCR products was conducted using cycle sequencing (Amersham Life Science, Cleveland, OH) . Upon final analysis, the 3.8 kb of AZ-1 sequence contained a candidate translational start codon consistent with the Kozak consensus rules and an inframe stop codon (£_≥H, 34:471 (1983); Nature _r 308:241 (1984)).

To confirm the composition of the 3.8 kb AZ-1 sequence, and to generate a cDNA containing the entire AZ-1 open reading frame (ORF) , primers corresponding to each end of the AZ-1 gene product were utilized in long template PCR' (Boehringer Mannheim Corp. Indianapolis, IN) . In two independent experiments, each using distinct pools of total SI RNA, the RT-PCR resulted in the amplification 3.8 kb gene products whose sequences were identical to the AZ-1 sequence originally derived using 5' RACE technology. The resulting cDNAs were subcloned into pCR 2.1 (Invitrogen, Carlsbad, CA) for further amplification and use.

Other AZ-1 constructs were also prepared. AZ-1 coding region sequences were amplified in polymerase-chain reactions using AZ-1-specific primers supplemented with EcoRI and Xhol restriction sites.

Forward primer: 5' CTGAATTCATGGACCTGGACTCTGCCCTCCAG 3' (SEQ ID NO: 22) .

Reverse primer: 5' GCCTCGAGTTAGGGCTGCTGGAACAGAAGGCC 3' (SEQ ID NO: 23) .

Amplified fragments, once digested with the appropriate enzymes, were ligated into the retroviral expression vector pLXSN (Clontech Inc., Palo Alto, CA) .

Cycle sequencing was performed to verify the sequence fidelity of each AZ-1 construct.

Probable sequence similarities between protein encoded by AZ-1 and other proteins were determined using BLAST computer-driven algorithm that calculates the sequence similarities. Based on the BLAST search results, the sequence of AZ-1 cDNA fragment did not match the cDNA sequences of any known proteins.

To determine whether any one of the GTGs or CTGs located upstream of the first ATG of the AZ-1 open reading frame can be used as an alternate initiator, long and short AZ-1 constructs, pCR-LAZl, (nucleotides 403-3325) and pCR- SAZ1 (nucleotides 1610-3325) were generated. The major products of the in vitro synthesized proteins, using the rabbit reticulocyte lysates (Promega, Inc. , Madison, WI) of these constructs exhibited a similar size predicated for the short construct (data not shown) suggesting that the first ATG in the deduced open reading frame was likely the translation start site. Based on these results, AZ-1 was predicted to encode a protein spanning 571 amino acids in length.

The BLAST search discovered that a splice variant of AZ-1, TACC2, has been cloned. TACC2 appeared to be isolated as a homolog of TACCl which was located at 8pll, a breast cancer amplicon. The alignment differences between AZ-1 and TACC2 are shown in Figures 19 and 20. TACC2 contains 31 additional upstream residues but no defined translation start site has been indicated. Two sequence insertions were located in the TACC2 protein (Figure 20) . The shorter insertion of DTFR residues was also noted in a small fraction of cDNAs transcribed from the premalignant S2 cell line RNA. The presence of these insertions seems to be characteristic of the premalignant cells and may serve as a marker for detection of pre alignancy.

B. Antibody Characterization of Protein Encoded by

AZ-1 Gene This invention demonstrated that polyclonal antibodies directed against AZ-1 specific peptide specifically recognized the protein encoded by the cloned AZ-1 cDNA.

Polyclonal or monoclonal anti-AZ-1 antibodies were prepared using methods known in the art by raising the antibodies against either AZ-1 fusion proteins (full-length or N-terminus, as described below) or immunogenic AZ-1 peptides (amino acids 1-20 or amino acids 131-145) . The studies performed during the development of this invention demonstrated the results obtained from using an affinity- purified polyclonal antibody directed against the AZ-1 peptide (amino acids 131-145) , hereafter called the anti- AZ-1 antibody. This AZ-1 antibody specifically recognized the protein encoded by the cloned AZ-1 cDNA. For these studies, different AZ-1 cDNA fragments, e.g., full length (pET-full length, nucleotides 1610-3325), N-terminus (pET-NT, nucleotides 1610-2692) and C-terminus (pET-CT, nucleotides 2693-3325) were subcloned into the pET28a+ bacterial expression vector (Novagen, Madison, WI)" and were expressed in bacteria as a fusion protein containing an N-terminal T7 tag and a polyhistidine epitope. Bacterially-expressed AZ-1 fusion proteins isolated by His-Bind chromatography were analyzed by SDS- PAGE and Western blot hybridization. For each expressed fusion protein, antibody against T7 tag recognized a band corresponding to the protein size predicted for AZ-1 peptide and the N-terminally added T7 and his tag fragments (Figure 7) . On a parallel blot, the affinity-purified AZ-1 antibody (raised against AZ-1 amino acids 131-145) also recognized similar sizes of the full length and N-terminus AZ-1 bacterially expressed products (data not shown) . Altogether, these experiments demonstrate that the anti-AZ- 1 antibody recognizes the protein encoded by the AZ-1 cDNA. Protein encoded by AZ-1 gene in the breast epithelial cells was also characterized by this AZ-1 antibody. Figure 8 shows the Western blot analysis of AZ-1 protein in the SI cell lysate with the AZ-1 antibody. Arrow indicates a 64- kd protein recognized by the AZ-1 antibody. To further test its specificity, the antibody was preincubaed with either 15 μg "+" or 30 μg "++" or AZ-1 immunogenic peptide before use in hybridization. In the presence of the antigenic peptide, the 64 kDa band was effectively competed away. On the other hand, the intensities of two minor bands (of higher molecular weight) , were not diminished by the peptide competition. The results confirm that the 64 kDa band corresponds to the cellular AZ-1 protein. III. Functionality of AZ-1 Gene

AZ-1 gene was found in abundance in the breast cells and its expression appears to be correlated with breast cell malignancy. The function of AZ-1 gene and its protein was, therefore, investigated. A. Function of AZ-1 Gene in Breast Cells Function of the AZ-1 gene was investigated in normal nonmalignant SI breast cells, in premalignant S2 cells and in breast tumor cells T4-2 obtained as shown in Figure 1. Results show that a novel gene AZ-1 is expressed in' nonmalignant breast cells and the expression is downregulated in malignant breast cells.

First, functionality studies were directed to identifying determinants of tumor progression by differential display.

The HMT-3522 breast culture model has the potential to provide significant insight into the molecular basis of tumor progression. By comparing the gene expression pattern of the model's tumorigenic cell population (HMT-3522-T4-2) with the nonmalignant SI cell line and with that of its premalignant progenitor (HMT-3522-S2) , genes that play a crucial role in the final stages of tumorigenic conversion were identified. Results of these studies are seen in Figure 9. Figure 9 illustrates differential expression of AZ-1 in premalignant and tumor breast cells where Figure 9A shows differential display of gene messages in premalignant (S2) and tumor (T4-2) cells on a sequencing gel. The arrow indicates a band representing a more intense AZ-1 cDNA fragment in S2 cells than in T4-2 cells, in the absence (-) , or in the presence (+) of a reverse transcription reaction (RT).

A PCR-based differential display seen in Figure 9A strategy was used to screen for genes whose expression varies between S2 and T4-2 cell populations. Using this approach, a 180 bp partial cDNA that was reproducibly present at higher levels in the pre-malignant cell samples than in their tumorigenic counterparts was detected.

To confirm the expression pattern observed in the different display experiments, the cDNA fragment was isolated and used as a probe in Northern blot analyses of total RNA derived from S2 and T4-2 cell cultures seen in Figure 9B.

Figure 9B shows Northern blot analysis where the top panel shows that the 4.4-kb AZ-1 message was highly abundant in S2 cells. In contrast, the similar RNA was greatly reduced in the T4-2 cells. In the bottom panel, the GAPDH" probe was used as a control for the amounts of RNA used. Two additional transcripts, 7-kb and 10-kb sizes, present in much less intensity, were also observed. The minor bands may correspond to RNA splice variants or unprocessed RNA species. Consistent with the differential display results, the tumor cell samples displayed a dramatic, more than 10- fold, reduction in the expression of the 4.4 kb message in comparison with the pre-malignant S2 cells.

Using the 180 bp cDNA fragment as starting material, 5 ' RACE technology was used to recover a full-length cDNA clone. Additional probes derived from the complete cDNA sequence were used to establish the expression pattern of these gene product in the HMT-3522 human breast cell series and in other human breast cells. Results, shown in Figures 10A-C, indicate a consistent expression pattern using probes derived from different regions of AZ-1 cDNA sequence.

Figure 10 shows expression of AZ-1 in nonmalignant breast epithelial cells and downregulation of AZ-1 in breast tumor cells and biopsies. AZ-1 expression patterns in nonmalignant and malignant breast cells were shown by Northern blot analysis.

Figure 10A shows expression of AZ-1 gene in nonmalignant breast cells, namely, in SI and MCFIOA nonmalignant breast epithelial cell lines, and in primary luminal epithelial (luminal epi) and myoepithelial (myoepi) cells. As seen in Figure 10A, the 4.4-kb message of AZ-1 gene was present in all nonmalignant breast epithelial cells examined.

Results seen in Figure 10A show that an abundant and specific 4.4 kb message corresponding to AZ-1 expression was detected not only in the non-malignant human epithelial cell lines, HMT-3522-S1 and MCFIOA, but also in primary cultures of human luminal epithelial and myoepithelial cells.

Figure 10B shows the presence or absence of AZ-1 observed in premalignant S2 and in ten malignant breast epithelial cell lines. T4-1 (T4) , HMT3909, MCF-7, CAMA-1, BT-20, MDA-MB-468, SKBR-3, T47D, MDA-MB-231, Hs578t, and' BT549 cell lines. The 4.4-kb message of AZ-1 gene was absent in ten breast tumor cell lines examined. The low level of AZ-1 message in HMT3909 cells is probably due to the presence of some contaminating nonmalignant myoepithelial cells (personal communication, Ole Petersen, unpublished data) . The RNA from premalignant cells (S2) was used as a positive control and, as expected, shows higher levels of AZ-1 expression.

Results seen in Figure 10B show that the 4.4 kb message was significantly reduced in the 10 of the 11 breast carcinoma cell lines which were examined.

Gene message level was also examined in in situ ductal breast carcinoma cells. RNAs were isolated from in situ carcinoma in the breast and normal tissue from reduction mammaplasty. Three out of four samples taken from breast cancer patients exhibited a lower level of AZ-1 mRNA than the normal tissue. One sample exhibited a higher AZ-1 message level, presumably from a patient at the premalignant stage. Collectively, the gene expression profiles obtained above for the 4.4 kb transcript are consistent with a role for the identified gene product as a Class II tumor suppressor in breast epithelia.

In order to determine whether AZ-1 gene is tissue specific, that is if it is solely expressed in the breast cells or also in other tissues, expression of AZ-1 in normal human tissues was studied. Results are seen in Figure 11.

Figure 11 shows Northern analysis of multiple human tissue RNA blot (Clontech, Inc., Palo alto, CA) probed with

AZ-1 cDNA fragment. The 4.4-kb AZ-1 message was shown to be expressed in heart, brain, lung, kidney, and pancreas, whereas it was low or absent in placenta, liver, and skeletal muscle. The β-actin probe was used to indicate the amounts of RNA loaded. On a separate multiple human tissue RNA blot from the same source, similar AZ-1 message was shown to be expressed in prostate, testes, and colon. The message was absent or in very low abundance in spleen, thymus, ovary, small intestine, and peripheral blood' leukocyte (data not shown) .

B. Chromosomal Localization of AZ-1 Gene

Mapping of AZ-1 gene placed AZ-1 gene to chromosome 10q26. Localization of AZ-1 gene to chromosome 10q26 is seen in Figure 12.

Figure 12 shows the chromosomal localization of AZ-1 gene by fluorescence in situ hybridization (FISH) . The arrows indicate AZ-1 gene is located at chromosome 10q26. Deletions at 10q26, such as by loss of heterozygosity (LOH) correlated with the occurrence of a variety of human cancers, including brain tumors, endometrial carcinoma, and gliobastomas .

C. AZ-1 Association with Cytoskeletal Complexes Association of AZ-1 with cytoskeletal complexes and its localization in nonmalignant HMT-3522 SI cells were also studied. Results are shown in Figure 13.

Figure 13 shows subcellular localization of AZ-1 protein. The immunostaining and fluorescent images were analyzed by confocal microscopy. Figure 13A shows immunofluorescence images of AZ-1 in SI cells grown on tissue culture plastic detected by the AZ-1 polyclonal antibody. AZ-1 was found to be present primarily in the cytoplasm and it existed as punctate, occasionally revealed in intense aggregates. Figure 13B shows that upon treatment with detergent such as Triton X-100, known to remove soluble components in the cytoplasm and nucleus, AZ-1 staining pattern remained, albeit having somewhat lower intensity. These results indicates that AZ-1 may be associated with cytoskeletal complexes.

In situ staining of AZ-1 in human breast tissues is shown in Figure 14. In these studies, localization of AZ-1 in normal human breast tissues was determined by AZ-1 antibody. Figure 14, top panel shows the control without the antibody. Figure 14, middle panel shows that AZ-1 is primarily present in the myoepithelial cells and in low abundance in the luminal epithelial cells of breast acini.' Bottom panel shows AZ-1 to be present primarily in the myoepithelial cells of breast ductal tissues. Some AZ-1 protein could also be observed in the luminal epithelial cells. D. AZ-1 Interaction with E-Cadherin and β-Catenin

AZ-1 interaction with the cell functional proteins, E- cadherin and β-catenin was also explored.

E-cadherin and β-catenin proteins function at cell-cell junctional complexes called adherens junctions. The adherens junctions are localized to sites of cell-cell contact along the lateral surface of epithelial cells. At this site, the adherens junctions interact with the active cytoskeleton and are believed to be crucial for maintaining the integrity of the cell structure. Loss of E-cadherin function correlates with increased cell invasion in many cell types, including the breast cells.

Interactions between adherens junction proteins E- cadherin (E-Cad) , β-catenin (β-Cat) and AZ-1 were investigated by coimmunoprecipitation shown in Figure 15. AZ-1 polyclonal antibody was used to immunoprecipitate AZ-1 protein in SI and T4-2 cell lysates. The immunocomplexes were analyzed by SDS-PAGE and detected by Western analysis with AZ-1, β-catenin and E-cadherin antibodies. Left panel shows that the 64-kd AZ-1 protein was immunoprecipitated with AZ-1 antibody. The rabbit IgG (IgG) was used as a control. The same blot when probed with either β-catenin or E-cadherin antibody also detected the presence of both antigens. E-cadherin (lower panel) was used as a control for the amounts of cell lysates loaded. These results suggested the plausible interactions of AZ-1 with adherens junctional complexes, either through E- cadherin or β-catenin. IV. Suppression and Reversion of Tumorigenicity of T4-2 Cells In Vitro and In Vivo

To determine the regulatory function of AZ-1 gene in suppression of tumorigenicity, T4-2 cancer cells were investigated in both in vitro and in vivo assays.

AZ-1 tumor suppression function in vivo and in vitro was assayed using a retroviral gene delivery system to introduce a full length AZ-1 transgene into the HMT-3522 T4- 2 tumor cells. The expression pattern of AZ-1 in non-malignant and tumorigenic cells suggests that AZ-1 may function as a Class II tumor suppressor and that, as such, AZ-1 may affect changes in cellular phenotype by virtue of its expression level. A. Tumorigenicity Suppression In Vitro

Overexpression of AZ-1 gene in the HMT-3522 tumor cells (T4-2) was tested in order to confirm that such overexpression is sufficient to attenuate their tumorigenic phenotype. Results are seen in Figure 16. Figure 16 illustrates the findings that ectopically- expressed AZ-1 reduces invasiveness in vitro.

In Figure 16A, two pooled populations of cells containing stably incorporated DNA were screened for AZ-1 expression by performing Northern analysis of total cellular RNA. In both cases, exogenous expression of the AZ-1 gene in T4-2 cells resulted in the accumulation of AZ-1 message at levels comparable to those observed in the nonmalignant SI cells. The levels observed in the AZ-1 overexpressed T4- 2 cells were 2 to 3-fold higher than AZ-1 mRNA and protein expression in the mock-infected T4-2 cells.

Because the transcript derived from the AZ-1 transgene comigrates on gels with the endogenous AZ-1 gene product (at 4.4 kb) , the expressed species were further characterized using transcript specific probes. Results obtained in these studies show that the increased expression observed in the T4-2+AZ-1 cells is entirely attributable to expression from the AZ-1 transgene (data not shown) . To test the potential tumor suppressor function of the AZ-1 gene product, assays of anchorage-independent growth, a generally accepted indicator of tumorigenicity in vitro, were performed on the AZ-1-overexpressing T4-2 cells and their unmodified counterparts. Results are seen in Figure' 16B where equal numbers of SI, T4-2, mock-infected T4-2 or T4-2+AZ-1 cells were embedded in semi-solid-agar and, after 4 weeks in culture, the number of viable colonies (>40 μm) was counted. As expected, non-malignant SI cells did not support growth in soft agar, whereas their tumorigenic T4-2 cell counterparts (both naive and mock-transfected) exhibited a markedly higher capacity for colony formation and anchorage- independent growth. Consistent with a role for the AZ-1 gene product in tumor suppression, the T4-2 cells overexpressing AZ-1 have shown an 80% diminished potential for growth capacity in soft agar assay, whereas the mock transfectants has around 10% reduction in its growth capacity. Ectopically expressed AZ-1 significantly suppressed tumorigenicity of T4-2 cells in vitro.

As an additional test of the tumorigenic cell phenotype, the capacity of the T4-2+AZ-1 cells to invade through basement membrane-coated filters in a modified Boyden chamber assay was performed. Using this approach, HMT-3522-S1 cells were found to be largely non-invasive (Invasion Metastasis _r 9:192 (1989)), whereas the tumorigenic T4-2 cells displayed the capacity to migrate effectively through the matrix-coated filters. AZ-1 overexpression diminished the tumor-like behavior of the T4-2 cells, in this case, by attenuating their invasive tendencies to 18% of that displayed by the mock-transfected T4-2 cells (Figure 16C) .

B. Tumorigenicity Suppression In Vivo

In vivo tumorigenicity of T4-A2+AZ-1 cells was examined by injecting the cells subcutaneously into the rear flanks of nude mice. After 6-8 wks post-injection, the mice were inspected for palpable tumors. Results are seen in Table 1.

TABLE 1

¹ Two injection sites per mouse ² Lump > 10 mm³

As seen in Table 1, the non-malignant SI cells failed to give rise to tumors, while the T4-2 cells (naive or mock infected) gave rise to obvious tumor growth in 90% of the injected sites. Mice injected with T4-2 cells overexpressing AZ-1 gave a diminished tumorigenic response with only 4 of the 28 inoculated sites giving rise to detectable tumors. Moreover, the sizes of the T4-2+AZ-1 tumors were much smaller than those observed with the mock- transfected T4-2 cells. These observations indicate that AZ-1 overexpression in human T4-2 cells is sufficient to reduce the tumorigenic behavior of these cells, both in vivo and in vitro. C. AZ-1 Upregulation of Expression Upregulation of AZ-1 expression or overexpression in phenotypically-reverted T4-2 cells was also investigated and AZ-1 overexpression was found to be sufficient to restore normal tissue architecture to tumorigenic MEC cells in culture.

According to PEAS, 89:9064 (1992), the behavioral phenotype of non-malignant and tumorigenic primary cultures or immortalized cell lines can be effectively reproduced in the context of a 3-dimensional reconstituted basement membrane assay.

Consequently, an additional test of AZ-l's tumor suppressor function was designed to investigate whether overexpression of AZ-1 in T4-2 cells would be sufficient to induce such phenotypic reversion of the HMEC tumor cells in the 3D basement membrane assay. These studies are illustrated in Figure 17. SI cells, the mock-infected cells and AZ-1 expressing-' T4-2 cells were embedded in 3-dimensional basement membrane gels. After 10 days in culture, the cell colonies were measured for size and tested with immunofluorescence microscopy. Results show that the SI cells formed polarized, growth-arrested acinar structures with organized endogenous basement membranes. The mock infected T4-2 cells continued to grow and formed large irregular colonies, as seen in Figure 17B, that failed to deposit an organized (polarized) endogenous basement membrane. Tumorigenic T4-2 cells, grown under the same conditions, formed large disorganized cell colonies that continued to grow.

When the overexpression of AZ-1 gene was induced in T4- 2 cells, the T4-2+AZ-1 colonies underwent phenotypic reversion. They adopted sizes comparable to the SI cell colonies and were capable of depositing an organized basement membrane at the basal perimeter of the acinar structures. These results, seen in Figure 17, indicate that AZ-1 overexpression was sufficient not only to reduce the size of the tumor colonies, but also to facilitate the structural reorganization that is required to give rise to polarized, organotypic acinar structures of malignant cells.

When the colony size was measured in the three above groups, as seen in Figure 17C, the size of T4-2+AZ-1 colonies reverted to the size of the normal nonmalignant cells.

The acinus-like structure of T4-2+AZ-1 cells in the 3D culture were reminiscent of the morphologically-reverted T4- 2 cells in the presence of an inhibitor anti-βl integrin (A2BII) or an EGFR specific inhibitor (tyrphostin AG1478) . Accordingly, studies were performed to examine whether the AZ-1 message level was modulated in the reverted T4-2 cells. Phenotypic reversion of T4-2 cells in the 3D rBM assay is dependent on the establishment of bi-direction reciprocal cross-talk between at least three intracellular signal mediators, including βl integrin, EGFR and MAP kinase (ibid) (EUAS_/ 95:14821 (1998)). Functional inhibition of any one" of these elements abrogates the signaling activity of the other two and results in the reduction of total βl integrin and EGFR protein levels. The AZ-1 gene product might also play a role in the observed cross-talk phenomenon, and its expression might be modulated in the presence of the previously described reverting agents (J. Of Cell. Biol._f 137:231 (1997), £NA£, 95:14821 (1998)).

To test this, total RNA was extracted from 3D rBM cultures of SI cells and T4-2 cells treated with or without inhibitors of either βl integrin (mAb AIIB2) or EGFR (Tyrphostin) . Results are seen in Figure 18. Northern blot analysis seen in Figure 18 revealed that AZ-1 levels, while significantly down-modulated in untreated T4-2 cells (Figure 18C) , were restored to Sl-like levels in cultures treated with either inhibitors of βl integrin (T4βl) or EGFR (T4tyr) (Figure 18C, two right panels) . When T4-2 cells were cultured in 3D rBM in the presence of functional inhibitors of βl integrin or EGFR, the T4-2 cells become "phenotypically reverted", that is they became reorganized to form Sl-like organotypic spheres. Collectively, these findings suggest that AZ-1 expression is coupled to βl integrin and EGFR activity. By virtue of its connections with βl integrin and EGFR, AZ-1 may provide essential cellular information that dictates cellular structure and phenotype.

In this regard, the 3D rBM assay served as another assay in testing tumor suppression, not only with respect to inhibition of cell growth, but also with respect to restoration of the appropriate tissue polarity and architecture.

D. Detection of AZ-1 Protein in Breast Tumor

Biopsies In order to determine the degree of malignancy in relation to the presence or absence of AZ-1 expressed protein and to determine whether the protein may be useful for detection of malignancy and tumorigenic progression, biopsies were obtained from 19 patients with confirmed stages of breast cancer progression. Results are seen in" Table 2.

TABLE 2

Table 2 shows in situ staining of AZ-1 protein in breast tumor biopsies. Nineteen in situ breast carcinoma biopsies with diagnosed mucinous carcinoma (stage 1) , infiltrating ductal carcinoma (stage 1, stage 2 and stage 3) , medullary carcinoma (stage 3) and metaplastic carcinoma (stage 3) , were investigated to determine relative abundance of AZ-1 protein by immunostaining with AZ-1 antibody. The malignancy state of the carcinomas was graded by a scale of 1 to 3. Number 1 indicates more differentiated and an early stage of malignancy, whereas number 3 indicates samples at a more advanced stage.

The results show that AZ-1 protein was present in 60% (4/7) of breast tumors at early stage 1 of malignancy whereas it was nearly absent (1/12) in the biopsies taken at more advanced state. Specifically, 1 out of 6 samples in the malignancy stage 2 showed the presence of AZ-1 protein. None of the six samples of the advanced stage 3 showed the presence of AZ-1 protein.

V. Methods for Detection. Diagnosis. Treatment,

Prophylaxis and a Kit for Diagnosis.

The current invention further concerns methods for detection of breast cancer, for its treatment and' prophylaxis and kits suitable for diagnostic detection of breast cancer growth.

A. A Method For Treatment of Breast Cancer

The proteins encoded by AZ-1 gene or its variants AZ-2 or TACC2 genes were found to be present in large amounts in the normal nonmalignant epithelial breast cells. Its level was found to be significantly decreased or nonexistent in the breast malignant cells.

AZ-1 gene or its variants, which express the protein, are thus actively expressing the protein in nonmalignant cells. Such expression, however, is decreased or absent in malignant breast cells.

The presence of protein expressed by AZ-1 gene, therefore, affects tumorigenicity of the breast tissue and is believed to act, and the findings described herein support its function, as the tumor suppressor.

The method for treatment of breast tumor, thus, comprises providing the subject patient with either the tumor suppressing protein directly targeted to the breast tissue or with genetic material able to express such protein. The protein may be delivered to a patient encapsulated within liposomes or formulated for target delivery using any other targeting means known in the art and used for targeted delivery of drugs to specific organs and tissue.

The second mode for treatment provides the subject patient with genetic material able to express AZ-1 protein. This is typically achieved through gene therapy.

A method for treatment comprises in vivo and ex vivo therapeutic approaches as well as in vivo gene therapy and ex vivo methods.

The gene therapy according to the invention utilizes two approaches. One approach comprises genetic modifications of the tumorigenic cells of a subject to be treated. Such modifications may be induced in the cells in vivo by, for example, developing and transferring a genetic material for expression of the specific protein, or the genetic manipulation may be performed in the subject's own* cells or other mammalian cells outside of the body, under ex vivo conditions. The resulting protein then may be imported and delivered to the cells or tissue of the treated subject. In vivo gene therapy consists of transferring the genetic material directly into the subject's cells.

Ex vivo gene therapy consists of removing cells from the subject and inserting the genetic material into these cells in vitro, prior to replacing the cells in the treated subject.

In the in vivo treatment, cDNA and the expression vectors are prepared. The plasmid encoding the AZ-1 protein is prepared and used to transfect subject's cells. In the cells, the plasmid is transcribed into mRNA and the protein is expressed.

In vivo , thus, the genetic material encoding the protein is transferred directly into the subject's cells or tissue. To ensure the efficiency of the method and expression of the AZ-1 protein at suitably high levels, the coding DNA sequence is engineered to be flanked by an appropriate regulatory sequence such as a viral promoter for ensuring high level expression or tissue specific promoter for ensuring specific organ target.

The transfer of genetic material according to the invention is designed to incorporate AZ-1 gene into breast cells by integrating it into chromosome 10q26. The genetic treatment is intended to permanently alter patient's genetic apparatus ensuring continuous long-term expression of AZ-1 gene. The method for transfer of genetic material in vivo utilizes, for example, adenovirus vectors, herpes simplex vectors, receptor mediated endocytosis and liposomes, among others, as well as nonviral systems or replication incompetent viruses. Genetic material transfer is achieved, for example, by direct injection, electroporation, particle bombardment, receptor-mediated endocytosis or using liposomes targeted for specific cells or tissues. In the ex vivo approach, the genetic material, that is" AZ-1 cDNA, are built into the expression vector, for example plasmid, cloned and transferred into cells, preferably subject's tumorigenic breast cells, grown in culture, that is grown in vitro and extracorporally. These cells are then transformed and expanded by cell culture in vitro and only then introduced into the subject. To avoid immune system rejection the autologous cells are normally used. For this pvirpose, the cells are collected initially from the subject to be treated, grown in culture, transformed and reintroduced by implantation into the same subject.

The ex vivo transfer of genetic material is achieved by and utilizes, for example, retrovirus vectors, adeno- associated virus vectors, and to a lesser degree, adenovirus vectors, herpes simplex vectors and liposomes. The ex vivo transfer of genetic material involves essentially four steps:

(a) cloning a dual-function genetic material into a vector, such as retroviral vector;

(b) transfecting the subject's cells where the targeted AZ-1 protein synthesis is to occur with the genetic material encoding recombinant AZ-1 protein;

(c) verifying the expression of the AZ-1 protein in the cells; and (d) reimplanting these cells in the patient.

The methods used for in vivo gene therapy applicable to the breast cancer therapy according to the invention are known in the art and are described, for example, in Molecular Biotechnology: Therapeutic Applications and Strategies _r Sunil Maulik, Solil D. Patel, WILEY-LISS, A JOHN WILEY & SONS, Inc., New York, (1997); Medical Genetics, pp. 252-257, George H. Sack, McGraw-Hill, New York (1999); Hum Molecular Genetics , 551-588, T. Strachan, A.P. Read, Bios Scientific Publishers, WILEY-LISS, A JOHN WILEY & SONS,

Inc., New York (1998); and Clinical Trials of Genetic

Therapy with Antisense DMA nd DMA Vectors . Ed. Enc

Wickstrom, Marcell Decker, Inc., New York (1998), all the above hereby incorporated by reference.

The administration of the genetic material, such as DNA, to the subject is done by means of a composition comprising the cDNA expressing the AZ-l protein and a pharmaceutically-acceptable carrier and/or other agents such as recombinase enzymes, a lipid agent, a lipid and protein agent, and the like.

Typically, the carrier may comprise solid, liquid or gaseous carriers. Examples of carriers are aqueous solutions, including water, buffered, aqueous solutions and the like.

While it is possible for the cDNA to be administered alone it is preferable to administer it as a pharmaceutical formulation.

In the preferred embodiment of the invention, the DNA of the above composition comprises AZ-1 gene cDNA sequence (SEQ ID NO: 1) encoding the AZ-1 protein.

The delivery of the cDNA into the cell may be conducted by a variety of techniques discussed above. These encompass providing the DNA enveloped by a lipid layer (liposomes) , further complexed with a protein and a lipid or a dendrimer.

The complexing the cDNA encoding the protein with lipid, lipid-protein, or dendrimer is especially applicable to n vivo transfection since less cell lethality is encountered, the DNA is protected from DNase degradation and the method is compatible with intracorporeal injection or administration.

One concern about the direct intravenous delivery of genetic material in vivo is the ability of the polynucleotide to survive in circulation long enough to arrive at the desired cellular destination.

In this respect, the coating or masking of the DNA is of extreme utility. The utilization of liposomes, a lipoproteic, or a dendrimer coating is extremely useful. In addition, a successful liposome system uses the cationic lipid reagent dioleyloxytrimethalammonium (DOTMA) . DOTMA may be mixed with phosphatidyl ethanolamine (PE) to form the reagent LIPOFECTIN^R. When this reagent is utilized to carry the polynucleotides the liposomes are mixed with the DNA and" readied for administration.

The DNA may be conveniently enveloped by a lipid layer (liposomes) , encapsulated by a lipid and a protein layer, or is complexed to dendrimer. The choice of the foregoing preparations will vary depending on the cell type used, the in vitro , ex vivo or in vivo conditions and the inherent limitations of each transfection method. Preferred conditions for enveloping the cDNA with a lipid layer are as follows. The cDNA is admixed with a lipid such as dioleophosphatidyl ethanolamine , dipalmitoylphosphatidylethanolamine (dipalmitoyl PtdEtn) , palmitoyloleoylphosphatidylethanolamine (palmitoyloleoyl PtdEtn) , dioleoylphosphatidylcholine (PtdCho) , dimyristoylphospatidylethanolamine (dimyristoyl PtdEtn) , diphytanoylgeycero-phosphatidylethanolamine (diphytanoyl

PtdEtn) , N-monomethyl PtdEtn, and N-dimethyl PtdEtn in a proportion of about 1 μg: Inmole to lμg: 500 nmoles, in an aqueous solution. Other components and proportions are permissible when this technology is applied to the in vivo method. The pH of the solution may be adjusted to about 8 to 10, and more preferably about 9. In addition to the above, ingredients such as a buffer and other known components may also be added to this composition. The amounts in which these components may be added are standard in the art and need not be further described herein.

In addition to the above, efficient transfer of genetic material requires the targeting of the genetic material encoding the AZ-1 protein to the breast cells. This can be attained by procedures based upon receptor mediated endocytosis according to J. Biol. Chem. , 262:4429(1987) or la B-iα , Chen , 263:14,621 (1988)). This technology utilizes a cell-specific ligand-polylysine complex bound to the DNA polynucleotide sequence through charge interactions. This complex is taken up by the target cells. The successful transfection of a similar hepatoma cell line resulting in stable expression of enzymatic activity following directed targeting was reported in Biochem. Pharmacol.. 40:253 (1985)). EHAS, (USA), 87:3410 (1990) an PNAS_f (USA) 88:4255 (1991) utilized a transferrin-polycation to attain the delivery of a plasmid into a human leukemic cell line and observed expression of the encoded luciferase gene. These proteins require attachment to the polynucleotide via, for example, a polylysine linker.

Moreover, in many receptor-mediated systems as chloroquine or other disrupters of intracellular trafficking may be required for high levels of transfection. Adenovirus, for instance, has been used to enhance the delivery of polynucleotides in receptor-mediated systems (PNAS (USA) , 88:8850 (1991)).

Alternatively, the genetic material may be masked through association with lipids. In one embodiment, the DNA is encased in standard liposomes as described, for example, in U.S. Patent No. 4,394,448, the relevant portion of the specification of which is hereby incorporated by reference. In another embodiment, the DNA is incubated with a synthetic cationic lipid similar to those described in U.S. Patent No. 4,897,355. The above-described synthetic cationic lipid effectively mask the DNA when associated therewith. The methods described in the above and below references are hereby incorporated by reference.

The cell recognition element is a molecule capable of recognizing a component on the surface of a targeted cell, covalently linked with a DNA-associating moiety by conventional methods. Cell recognition components include antibodies to cell surface antigens, ligands for cell surface receptors including those involved in receptor- mediated endocytosis, peptide hormones, etc. Specific ligands contemplated by this invention include carbohydrate ligands such as galactose, mannose, mannosyl 5-phosphate, fucose, sialic groups, N-acetylglucosamine or combinations of these groups as complex carbohydrates such as those found on glycolipids of the blood groups or on various secreted proteins. Other ligands include folate, biotin, various peptides that can interact with cell surface or intracellular receptors such as the chemoattractants peptide containing N-formyl peptides that contain a cystine residue or that interact with cell surface protein such as the human immunodeficiency virus GP-120, and peptides that interact with CD-4.

Other ligands include antibodies or antibody fragments. The specificity of the antibodies can be directed against a variety of epitopes that can be expressed on cell surfaces including histocompatibility, macromolecules, autoimmune antigens, viral, parasitic or bacterial proteins. Other protein ligands include hormones such as growth hormone and insulin or protein growth factors such as GM-CSF, G-CSF, erythropoietin, epidermal growth factor, basic and acidic fibroblast growth factor and the like. Other protein ligands would include various cytokines that work through cell surface receptors such as interleukin 2, interleukin 1, tumor necrosis factor and suitable peptide fragments from such macromolecules.

The membrane-permeabilizing element of this system is a molecule that aids in the passage of a polynucleotide across a membrane. The liposomes, synthetic cationic lipids, lipid-proteins, and dendrimer described above as DNA- asking components also may function as membrane- permeabilization components.

Additional membrane-permeabilizing components that will facilitate delivery of the genetic material of this invention also include polycations that neutralize the large negative charge on polynucleotides. Polycations of this invention include polylysine, polyarginine, poly (lysine- arginine) and similar polypeptides, and the polyamines and the polycationic dendrimers. The membrane-permeabilizing component that facilitates transfer of the protein or DNA of this invention may be an amphiphathic cationic peptide. A phipathic cationic peptides are peptides whose native configuration is such that the peptide is considered to have a cationic face and a neutral, hydrophobic face. In a preferred embodiment, the peptide is a cyclic peptide. Examples of the amphipathic cationic cyclic peptides of this invention are gramicidin S, and tyrocidines. The peptide may also contain some or" all of the amino acids in the D configuration as opposed to the naturally occurring L configuration.

The membrane permeabilizing elements, i.e., the cyclic peptide and optional phospholipid and polyamine, may be added to the composition simultaneously or consecutively. Preferably, the cyclic peptide is added first, and the phospholipid or polyamine is added later. The molar ratio of added cyclic peptide to added polyamine is preferably from about 1:1 to about 1:3. The molar ratio of added cyclic peptide to added phospholipid is preferably from about 1: 1 to about 1:20.

The subcellular-localization element of this system is a molecule capable of recognizing a subcellular component in a targeted cell, covalently linked with a DNA-associating moiety by conventional methods. Particular subcellular components include the nucleus, ribosomes, mitochondria, and chloroplasts. In a preferred embodiment of this invention, the subcellular-localization component is a nuclear- localization component. The nuclear-localization components include known peptides of defined amino acid sequences, and longer sequences containing these peptides.

For the conventional therapy_: the patient is provided with the recombinant AZ-1 protein. Using either in vitro or ex vivo methods described above and in examples, the AZ-1 recombinant protein is prepared and delivered in the conventional way using pharmaceutically acceptable delivery vehicles and routes. This type of delivery needs to assure that the protein is properly protected from the destruction by digestive proteases when administered orally, or destroyed in the body before it reaches the target breast cells or tissue.

In this mode, the AZ-1 protein may be delivered orally, intravenously, intramuscularly, intraperitoneally, subcutaneously, as aerosol, or using any other mode of delivery known in the art.

In order to avoid the major drawbacks of delivery of proteins, such as instability in the proteolytic environment" of the GI tract and poor absorbability through the mucosa, AZ-1 protein containing compositions are preferably prepared as sterile solutions and administered to patients by daily injection. In this particular instance, the protein is prepared ex vivo, isolated, purified and administered to a subject. However, this form of drug delivery could cause pain and inconvenience to patients and thus could be poorly accepted. Many novel delivery systems have been developed to address these problems (Trends Biotech;, 16(8): 343-9, (1998), J. Phar aceut. Sci_r 87(11): 1331:1334, (1998)). Examples of such deliveries of proteins include, but are not limited to combinations of:

1. Targeted delivery of the protein to cells or tissues to be treated or administered orally, nasally, as injectable, etc. , as alternate sites of delivery to a site where the target protein should be delivered.

2. Formulations containing the AZ-1 protein for sustained release. 3. Formulations for administration of concentrated AZ-1 protein into cells at the mucosal surface.

4. Formulations modified for enhanced absorption of the protein into breast cells.

5. Formulation for inhibition of proteolysis including protease inhibitors, chemical modification of the peptide molecule to produce prodrugs and analogs (Nippon Rinsho. Jap. J.

Ωl±I , Med., 56(3): 601-7 (1998) and genetic engineering of proteolytically resistant forms. A number of novel delivery systems have been approved by the FDA (J. Pharmaceut. Sci.. 87:1331-4 (1998)). In many cases, a given formulation incorporates elements of several of the above mentioned approaches. Development and optimization of an oral drug delivery systems tends to be specific for each protein or peptide drug. (J. Pharmaceut. Sci. _f 85:1282-5 (1996)).

Examples of formulations which prevents proteolytic degradation of the AZ-1 protein, sustained release and- enhanced absorption follow.

The AZ-1 protein may be delivered via microparticles or microsphere. Gelatin capsules coated with various concentrations of sodium alginate, for example, 20% w/v, and cross-linked with appropriate concentrations of calcium chloride, are resistant to the harsh environment of the stomach and deliver the drug to the distal gastrointestinal tract where drug absorption occurs. (J. Biomaterials Sci., Polymer Ed.. 7(1): 39-48 (1995). The protein can also be incorporated into biodegradable microparticles to reduce the effect of gut secretions and to enable the absorption of the protein in an unaltered form. The uptake of microparticulates through the gut wall is accepted as a true biological phenomenon but the mechanism and route of uptake have not been established.

Lipid delivery vehicles enhance microparticle uptake and the selective transport of microspheres across M cells (J. Anatomy _r 189 (Pt3) : 487-90 (1996)). Microparticles and microspheres also allow sustained release. Gelatin nanoparticle-poly (lactic-co-glycolic acid) (PLGA) microsphere composites can be prepared by encapsulating protein-loaded gelatin nanoparticles in PLGA microspheres. This encapsulation is conducted by using a phase separation method and a solvent extraction method. Protein release experiments described in . Pharmaceut. Sci. , 86(8): 891-5 (1997), indicate that this composite system possesses sustained release characteristics. This system also demonstrates the capability of preventing the denaturation of the AZ-1 protein. Poly (vinyl alcohol) (PVA) hydrogel nanoparticles may be prepared by using a water-in-oil emulsion technology plus cyclic freezing-thawing process. The PVA hydrogel nanoparticles prepared by this method are suitable for the AZ-1 protein drug delivery since formation of the hydrogel does not require crosslinking agents or other adjuvants and does not involve any residual monomer. The PVA hydrogel nanoparticles swell in an aqueous solution and the swelling degree increases with the increase of temperature.

Another route of delivery for AZ-1 protein is through mucoadhesives. Mucoadhesives are polymeric delivery systems for use to concentrate protein and peptide pharmaceuticals at the mucosal surface. Some of the mucoadhesive polymers were found to display other important biological activities, i.e., inhibition of proteolytic enzymes and or modulation of the permeability of usually tight epithelial tissue barriers. Rather than being just adhesives, mucoadhesive polymers may therefore be considered as a novel class of multifunctional macromolecules with a number of desirable properties for their use as biologically active drug delivery adjuvants which are particularly useful in the context of peptide and protein drug delivery.

Carbopol (polyacrylic) polymers with strong bioadhesive properties also can inhibit lumenal degradation of peptide or proteins, offering multiple advantages for their uses in oral drug delivery (J. Pharm. Pharmacol., 48(1): 17-21,

(1996), Pharmaceut. R££, 12(9): 1293-8 (1995). The mucoadhesive polymers carbomer 934P and chitosan hydrochloride are able to enhance the intestinal absorption of some agents such as buserelin in vivo, and are therefore suitable excipients in peroral delivery system for AZ-1 protein.

Mucoadhesives include monolithic type devices in which the drug is dispersed throughout the polymer and protein- polymer conjugates where the drug is covalently bound to the polymer. Advanced delivery systems include systems containing mucoadhesive polymers providing an intimate contact to the mucosa, thereby reducing the drug degradation between delivery system and absorbing membrane. They also may contain controlled release systems which provide a simultaneous release of protein and inhibitor or inhibitor prodrug or the immobilization of enzyme inhibitors on delivery systems (J. Controlled Release. 52(1-2): 1-16 (1998); J. Med. Chem_f 41(13): 2339-44 (1998).

For treatment of the breast cancer, the patient will be provided either with the gene therapy or with the AZ-1 protein formulated as described above.

Treatment regimen would be 1-3 times a day, daily, 1-2 times a week or as needed and will be continued until the normal levels of AZ-1 protein are detectable, until the AZ-1 gene expression is restored and until the tumorigenicity reversion is achieved and confirmed by immunostaining morphological micrography or any other means.

B. A Method For Prophylaxis of Breast Cancer

The method for prophylaxis of cancer growth in the breast cells is achieved in the same way as described above for the treatment.

A patient to be prophylactically treated would have either a family history of the breast cancer or the low level of expression of AZ-1 gene, or the low level of AZ-1 protein will be detected in the biopsy, although there will not yet be any visible, palpable observable or detectable tumor growth.

For prophylaxis, the dosages of the protein will typically be lower than for treatment, treatment will be administered 1-2 times weekly and the patient will be monitored weekly or monthly for AZ-1 gene expression or for the levels of AZ-1 protein in the breast tissue biopsies.

C. A Method For Detection of Breast Cancer Detection and diagnosis of the breast cancer comprises determining a level of expression of AZ-1 message or detecting a level of protein expressed in breast biopsies.

The level of AZ-1 expression is determined by, for example, in situ hybridization of AZ-1 RNA using a complimentary DNA probe or by RT-PCR assay using gene specific primers according to Example 5. The level of AZ-1 expressed protein is determined, for example, by detecting with polyclonal or monoclonal AZ-1 antibodies specifically prepared (Example 8) against AZ-1 protein or peptide using indirect immunofluorescence assay as described in Example 18.

D. Kits for Detection and Diagnosis of Breast Cancer Kits for detection and diagnosis of breast cancer are based on two approaches outlined in Section C, namely, on detecting the AZ-1 protein or on detecting AZ-1 gene' message.

1. The Kit for Detection of AZ-1 Protein

The kit for detection of AZ-1 protein in breast biopsies (either cryosections or parafilm embedded blocks) for diagnosis of breast cancer comprises:

(a) polyclonal or monoclonal anti AZ-1 antibodies;

(b) secondary antibody, e.g. FITC or Texas Red conjugated;

(c) permeabilizing/fixing reagent (optional) ; (d) blocking/hybridization solution;

(e) means of tabulating detected AZ-1 protein in the breast tissue samples and correlating the level with the presence, absence, or the stage of breast tumorigenicity. Examples of results using this kit are shown in Figure

14 and Table II following the method described in Example

18.

2. The Kit for Detection of AZ-1 Message

The kit for detection of AZ-1 message in breast biopsies for diagnosis of breast cancer comprises components needed for in situ hybridization or for detection of the message by RT reverse transcription PCR.

(A) In situ hybridization kit comprises:

(a) means for preparing intact breast biopsies, e.g., cryosection and mounting condition;

(b) AZ-1 gene specific cDNA probe;

(c) β-actin cDNA as a positive control;

(d) means for labeling of cDNA probes;

(e) means for hybridization/blocking; and (f) means for tabulating and correlating the detected AZ-1 message levels with the malignant stages of breast cancer patients. (B) RT-PCR kit comprises:

(a) means for preparing intact breast biopsies, e.g., cryosection and preservation of fresh tissues; (b) means for RNA extraction from breast" biopsies;

(c) AZ-1 gene specific cDNA primers;

(d) means for RT-PCR amplification of RNA extracts; (e) means for detection of amplified cDNA product, e.g., by ethidium bromide staining/agarose gels; (f) means for quantification of detected cDNA products; and (g) means for correlation of obtained values to a degree of tumorigenicity.

UTILITY

The current invention is useful for diagnostic and therapeutic purposes for detection and treatment of human breast cancer. For diagnostic purposes, the level of AZ-1 gene encoded protein is determined in the breast biopsy from the patient. When the amount of AZ-1 protein is high, then there are no tumor cells present in the biopsy. When the

AZ-1 encoded protein is absent or at a low level, then there are breast tumor cells present. The protein is useful also as a marker for the malignancy progression. The protein is also useful as a diagnostic marker for tumorigenic reversion in cancer patients undergoing conventional cancer therapy.

The invention is also useful for therapy, particularly gene therapy of breast tumors wherein the specifically targeted AZ-1 gene is introduced into the breast cells, using, for example, retrovial delivery system for gene therapy or the AZ-1 protein is administered in tissue targeted formulation. EXAMPLE 1

Cell Separation This example describes the procedure used for cell separation.

Human breast luminal epithelial and myoepithelial cells were purified from organoids after these had spread out to form monolayers in primary culture or had been passaged once. Cells were trypsinized and resuspended in N-(2-' hydroxyethyl) -piperazine-N' -2-ethanesulfonic acid (HEPES) buffer with 0.5% (W/V) bovine serum albumin (BSA, fraction V, A4919, Sigma) and filtered through a 100 mM nylon mesh (Millipore, Hedehusene, Denmark) to remove residual cell clumps. The cell suspension was incubated 30 minutes at 4°C with the primary mAb, 115D8, directed against sialomucin (provided by Jo Hilgers, Amsterdam, The Netherlands) or J5 directed against common acute lymphoblastic leukemia antigen (CALLA) or CD10 antigen (Coulter Clone, Struers Kebo Lab., Albertslund, Denmark) diluted 1:100 and 1:10, respectively in HEPES/BSA. The cells were then washed twice in HEPES/BSA and incubated 15 minutes at 4°C with goat anti-mouse IgG microbeads (AH Diagnostics, Arhus, Denmark) diluted 1:5 in HEPES/BSA and washed twice in HEPES/BSA. Cell separation was carried out by use of the MiniMACS magnetic cell separation system obtained from AH Diagnostics according to the kit instructions.

EXAMPLE 2 Cell Culture and Human Luminal Epithelial Cells This example describes cell culture conditions used for culturing human luminal epithelial cells.

The HMT-3522 human mammary epithelial cells were grown in H14 medium consisting of DMEM/F12 medium (GIBCO/BRL, St. Louis, MO) and additives including 250 ng/ml insulin, 10 μg/ml transferrin, 2.6 ng/ml sodium selenite, 10^"10 M estradiol, 1.4X 10^"6 M hydrocortisone, and 5 μg/ml prolactin. The S-l and MCFIOA cells were propagated in H14 medium as monolayers on plastic in the presence of lOng/ml epidermal growth factor (EGF) whereas the S2 and T4-2 cells were cultured as monolayers on flasks coated with collagen Type I (Vitrogen 100, Celtrix Laboratories, Palo Alto, CA) in the absence of EGF. HMT3909 and MCF-7 cells were cultured as monolayers on collagen type I in DMEM/F12 medium supplemented with 1.4 X 10^"6 M hydrocortisone and 2 μ M glutamine, respectively. Breast tumor cell lines, e.g., CAMA-1, BT-20, MDA-MB468, SKBR3 , T47D, MDA-MB231, Hs578T and BT549, were cultured as monolayers in DMEM/F12 with 5%' bovine serum.

Human breast luminal epithelial cells were purified organoids grown as monolayers in primary culture. Three dimensional (3D) cultures were prepared by growing SI, T4-2 cells, and T4-2 transfectants to confluence as monolayers, followed by trypsinization and embedding (8.5 X 10⁵ cells/ ml) as single cells into a commercially prepared reconstituted basement membrane (Matrigel, Collaborative Research, Waltham, MA) from Englebreth-Holm-Swarm mouse tumors .

EXAMPLE 3 Probe Mapping on Metaphase Chromosome This example describes probe mapping on metaphore chromosome and primers and procedure used for chromosome localization of AZ-1 gene.

U4(CGTATGCACTACTGTATTTCCTTTC) (SEQ ID NO: 24) and L3 (GGGCAAGGGCCAAGGTCCAGCAATG) (SEQ ID NO: 25) primers were used to generate 199-bp genomic DNA to screen for Pl/BAC/PAC clones to determine location of AZ-1 on human chromosome by fluorescence in situ hybridization (FISH) .

Pl/BAC/PAC clone was used to determine the location of AZ-1 on human chromosomes. DNA was extracted from an overnight culture using alkaline lysis technique. Probe DNA was labeled with digoxigenin-11-dUTP by nick translation. Hybridization was carried out in the presence of human Cot 1 DNA to suppress the background signal and hybridized to metaphase chromosomes overnight. The hybridized signal was detected by anti-digoxigenin conjugated with FITC. The location of the probes was determined by digital image microscopy following FISH and localized by the fractional length from the p-terminus (FLpter) described previously in Human Genet. 83: 335 (1989). EXAMPLE 4 RNA Extraction, Quantification and Northern Blot Analysis This example describes conditions and procedures used for RNA isolation and Northern blot analysis. Total RNA was extracted from cells cultured as* monolayers or in 3D rBM or cells from normal tissues or in situ carcinoma with TRIzol reagent (Life Technologies, Inc. Grand Island, NY) . For Northern blots, total RNA (20 μg/lane) was resolved on denaturing agarose gels and transferred to Hybond-N⁺ membranes (Amersham) . Resulting blots were hybridized with ³²P-labeled cDNA probes and analyzed by autoradiography. A GAPDH probe was used to normalize variations in loading.

EXAMPLE 5 Differential Display

This example describes conditions used for differential display and RACE cDNA amplification.

Differential display was performed using the RNA image protocol (GenHunter Corp., Nashville, TN) following manufacturer ^• s instructions .

The total RNA (DNA-free) from S2 and T4-2 cell lines was reverse transcribed and the cDNA products were amplified by polymerase chain reaction using the anchored (H-T_nM, M=A,C,G) and arbitrary primers (H-AP and H-T_nA) provided in the kit. Amplified products were resolved on 6% acrylamide gels and differential expression of the amplified species was evaluated by autoradiography.

The expression patterns of these two cell lines were observed on the sequencing gel. The differences in the intensity of bands representing differential gene expression were further confirmed by agarose gel electrophoresis analysis of the reamplified cDNA fragments that had been eluted from the gel. To confirm observed differential expression patterns, cDNA fragments of interest were excised from the gel, subject to a second PCR amplification and analyzed on agarose gels. Gene identification of the cDNA products and differences in the message levels were then verified by northern blot analysis.

One of the DNA fragments from the PCR products of H-T_nA (5-AAGCTTTTTTTTTTTA) SEQ ID NO: 28 and H-AP1 (5'- AAGCTTGATTGCC) SEQ ID NO: 29 primers showed a significantly higher intense band in S2 than in T4-2 cells on both^' sequencing and agarose gel analyses. Northern blot analysis using this cDNA fragment as a probe confirmed its gene product was greatly more abundant in the S2 cells than in T4-2 cells. Sequence analysis revealed it was novel. This gene has been named AZ-1.

EXAMPLE 6 Rapid Amplification of 5' cDNA Ends (RACE. This example describes methods used for rapid amplification of 5' cDNA end. 5' RACE system (Life Technologies, Inc. Grand Island, NY) was performed according to the manufacturer's protocol to extend the 5* end of the cDNA length. The procedure was repeated approximately twelve times to map a total length of 3.8 kb of AZ-1 sequence. In each run, 500-800 bp of additional 5 '-end sequence was obtained. To further determine the contiguousness and accuracy of AZ-1 sequence, 3.8 kb AZ-1 cDNA fragments were prepared from two separate reverse transcription products using Expand Long Template PCR System (Boehringer Mannheim Corp., Indianapolis, IN). A complete match of the sequence of these two cDNA clones confirmed AZ-1 cDNA sequence.

EXAMPLE 7 Sequencing of 5' RACE PCR Products This example describes sequencing of 5 ' RACE PCR product.

Sequencing of the 5 ' RACE PCR products was conducted by thermo sequenase radiolabeled terminator cycle sequencing kit (Amersham Life Science, Cleveland, OH) . AZ-1 cDNA clones described in Example 6 were sequenced by the sequencing facility in the University of California at Berkeley. EXAMPLE 8 Preparation of Anti AZ-1 Antibodies This example describes procedures used for preparation of antibodies against AZ-1 peptides, AZ-1 N-terminal (fragment 1-368) , and full length AZ-1 protein. Polyclonal Antibodies:

AZ-1 polyclonal antibody was raised against an immunogenic peptide AZ-l-A (residues 121-135, KPAKKKKTPLKTVKK) (SEQ ID NO: 26) in rabbits by Animal Pharm Services, Inc. (Healdsburg, CA) . The immunoglobulin G fraction of the antiserum was further purified by AZ-1 peptide-linked affinity chromatography.

AZ-1 polyclonal antibodies are raised against immunogenic peptides AZ-l-B (residues 1-20, MPLRRPKMKKTPEKLDNTPA) (SEQ ID NO: 27) , and purified His- tagged fusion protein containing residues 1-368 of AZ-1 protein fragment in rabbits by ImmunoVision Inc., (Daly City, CA) . The immunoglobulin G fractions of the antisera are further purified by AZ-1 peptide or purified protein- linked affinity chromatography. The antibody is further purified by affinity chromatography and analyzed by enzyme- linked immunosorbent assay (ELISA) . Monoclonal antibody:

AZ-1 monoclonal antibody is raised against purified His-tagged full length AZ-1 fusion protein in mice by

ImmunoVision Inc. The antibody is further purified by affinity chromatography and analyzed by enzyme-linked immunosorbent assay (ELISA) .

EXAMPLE 9 Reversion Assays

This example describes reversion assays. The βl-integrin function-blocking mAb AIIB2 (C. Damsky, UCSF) was introduced into the cell-embedded substratum at a concentration of 100 μg/ml ascites protein (which corresponds to 4-10 μg/ml purified rat IgGl) at the time of Matrigel gelation. Tyrphostin AG 1478 (Calbiochem) dissolved in dimethyl sulfoxide was added to the medium at a concentration of 100 nM on alternate days. Control cultures were treated with mouse IgG and vehicle only for AIIB2 antibody and inhibitor experiments, respectively.

EXAMPLE 10 Immunoblotting and Immunoprecipitation

This example describes immunoblotting and immunoprecipitation methods.

Cells grown as monolayers were lysed in situ in RIPA buffer [1% Nonidet P-40, 0.5% deoxycholate, 0.2% SDS, 150 mM sodium chloride, 50 mM Tris-HCl (pH 7.4) containing 2 mM sodium fluoride, 1 mM sodium orthovanadate, 10 μg/ml E64, and 1 mM Pefabloc] . Cells grown in 3D rBM cultures for 10 days were isolated as colonies with ice-cold PBS/EDTA [0.01 M sodium phosphate (pH 7.2) containing 138 mM sodium chloride and 5 mM EDTA] and thereafter were lysed in RIPA buffer. Protein lysates were resolved upon 7.5% SDS-PAGE gels, electrotransferred to immobilon-P blots (Millipore Corp.) And the blots were then subjected to Western analyses and enhanced chemiluminescence (ECL) (Amersham Corp. , Arlington Heights, IL) detection. For reprobing, the blots were stripped by incubating in 2% SDS, 62.5 mM Tris-HCl, (pH 6.7), 2-mercaptocthanol, at 50°C for 30 minutes.

For immunoprecipitation of AZ-1, the RIPA lysates were first precleared by incubating with rabbit IgG and protein A coupled to Sepharose 4B beads (Pharmacia LKB Biotechnology, Inc., Piscataway, NJ) before immunoprecipitation. The precleared lysates were incubated with affinity-purified polyclonal anti AZ-1 antibodies, monoclonal antibody, or normal rabbit IgG antibodies (negative control) together with protein A coupled to Sepharose-4B beads. Immunoprecipitates were washed five times in RIPA buffer and dissolved in equal volume of 2X SDS sample buffer (0.125M Tris-HCl, 4% SDS, 20% glycerol, 0.02% bromophenol blue, 4% β-mercaptoethanol) for SDS-PAGE analysis and subsequent Western analyses.

EXAMPLE 11 AZ-1 Plasmid Construct This example describes preparation of plasmid AZ-1 constructs . pCR-LAZl, pCR-SAZl, pET- full length AZ1, pET-NT, pET- CT, and pLXSN-AZl constructs were used. Full length AZ-1 open reading frame (nucleotides 1610-3325) was inserted into Sacl-Sall sites of pET28 a(+) vector (Noragen) , and EcoRI-' Xhol sites of pLXSN vector (Clontech Inc.) to generate pET- full length AZ1 and pLXSN-AZl constructs.

N-terminus (nucleotides 1610-2692) and C-terminus (nucleotides 2693-3325) of AZ-1 protein sequences were inserted into Sacl-Sall sites of pET28 vector to generate pET-NT and pET-CT constructs. Two AZ-1 cDNA sequences, positioned at nucleotides 403-3325 and nucleotides 1610-3325 were inserted into pCR 2.1 vector (Invitrogen) to generate pCR-LAZl and pCR-SAZl constructs. EXAMPLE 12

Transfection Assay

This example describes methods used for transfection.

PLXSN vector and pLXSN-AZl were transfected into PT60 cells provided in the Retro-X system (Clontech Laboratories, Inc., Palo Alto, CA) and the stable virus-packaging PT60 cells were generated by selection with 500 μg/ml G418

(Genetisen; Gibco Inc.). The retrovirus particles collected from the growth media selection of the stably transfected

PT67 cells were then used to infect T4-2 cells and stable transfectants were selected in 50 μg/ml G418.

EXAMPLE 13 Jn Vitro Transcription and Translation This example describes transcription and translation methods. The CR-LAZl and pCR-SAZl constructs were used to generate in vitro translated product by a TNT coupled reticulocyte lysate system (Promega, Madison, WI) .

EXAMPLE 14 Soft Agar Assay This example describes anchorage-independent growth assay.

SI, T4-2, T4-2 (mock) and T4-2 + AZ1 cells were seeded at 1 X 10^s cells/well in 0.35% soft agar on 12-well plate for 4 weeks and the size of the colony was measured by eyepiece. Colonies greater than 40 μm was scored as positive and counted. Four repeats were performed on each cell and the experiments performed in triplicate. EXAMPLE 15

In Vivo Tumorigenicity

This example describes assay used for testing in vivo tumorigenicity.

SI, T4-2, T4-2 (mock) and T4-2+AZ-1 cells propagated as monolayers were trypsinized and dispersed in DMEM:F12 medium at a concentration of 2.5 X 10⁷/ml. An aliquot of 100 μl (2.5 X 10⁶ cells) was subcutaneously injected into each flank of 4-6 week old BalbC nu/nu mice. The size of nodule on the flank was measured by a caliper and recorded at 6-8 weeks after injection.

EXAMPLE 16 In Vitro Invasion Assay This example describes assay used for testing in vitro invasion. 8 μM Falcon cell culture PET inserts (Becton Dickinson Labware, Franklin Lakes, NJ) were coated with 10 μl of 1:2 dilution (50 μg/filter) of matrigel in DMEM/E12. 1x10^s of SI, T4-2, T4-2 (mock), T4-2-AZ1 cells resuspended in 200 μl H14 medium were grown on top of coated insert in 24 well plate. After 18-24 hours, the cells migrated through the matrigel was fixed in glutaldehyde, stained with toluidine blue and counted. Four repeats were performed for each cell line and the experiment was repeated three times.

EXAMPLE 17 Morphogenesis Assessment and Criteria

This example describes assays used for assessment of morphogenesis .

Morphology was assessed in situ by examining the degree of colony organization visually by phase contrast microscopy, and by measuring colony diameter using an eyepiece equipped with a micrometer spindle. Polarity was indicated by the presence of a basally organized basement membrane (BM) as determined by collagen IV and β4-integrin immunostaining.

EXAMPLE 18 Indirect Immunofluorescence This example describes the procedure used for measurement of indirect immunofluorescence.

Cells or breast tissues were permeabilized in situ (0.5% Triton X-100 in 100 mM NaCl/300 mM sucrose/10 mM PIPES, pH 6.8/5 mM MgCl₂ containing 1 mM Pefabloc Sc (AEBSF) (Boehringer Manheim)/10 g/ml leupeptin/10 μg/ml aprotinin/10 μg trypsin inhibitor type 11/250 μM NaF) , fixed in 2% paraformaldehyde, and immunostained in the presence or absence of AZ-1 antibody using essentially the assay as described in J. of Cell Biology, 137:231 (1997).

EXAMPLE 19 Homology Search and Secondary

Structure Prediction Programs This example describes methods and programs used for homology search and secondary structure prediction.

The gapped BLAST (National Center for Biotechnology Information, NCBI) , BEAUTY + BLAST (Baylor College of Medicine) , and TFASTA (University of Wisconsin) were the used sequence homology search programs.

Website (http: //catt. poly.edu/rjps/matches.html) was used to predict coiled coil structure and (http: //s/rec3.unil.ch/software/COILS-form. html) and (http://nightingale.lcs.mit.edu/cgi-bin/multicoil) were used to calculate the probability that AZ-1 sequence adopts a coiled-coil conformation.

EXAMPLE 20 AZ-1 Recombinant Proteins

This example describes production of AZ-1 recombinant proteins.

Different AZ-1 cDNA fragments, e.g., full length (pET- full length, nucleotides 1610-3325) , N-terminus (pET-NT, nucleotides 1610-2692) and C-terminus (pET-CT, nucleotides 2693-3325) were subcloned into pET 28 bacterial expression vector (Novagen, Madison, WI) and were expressed in bacteria as a fusion protein containing an N-terminus T7 tag and a polyhistidine epitope. The expressed proteins in the solubilized bacterial cell lysates were purified by His Bind column chromatography and following the manufacturer's procedures. The AZ-1 recombinant proteins were eluted with IX elute buffer (150 mM imidazole, 500 mM NaCl, 20 mM Tris- HCl, pH 7.9) .

Claims

WHAT IS CLAIMED

1. A cDNA sequence consisting of a sequence encoding a protein acting as a tumor suppressor or a marker of malignancy progression or reversion, wherein said cDNA' sequence encoding said protein is identified as SEQ ID NO: 1, or a variant, mutant or fragment thereof.

2. The cDNA sequence of claim 1 wherein the variant is identified as SEQ ID NO: 2.

3. The cDNA sequence of claim 1 wherein the variant is identified as SEQ ID NO: 5 (TACC2) .

4. The cDNA sequence of claim 1 wherein encoded protein is identified as amino acid sequence SEQ ID NO: 3 or a variant thereof identified as SEQ ID NO: 4 or SEQ ID NO: 6.

5. A recombinant protein consisting of an amino acid sequence identified as SEQ ID NO: 3, or a variant, mutant or fragment thereof.

6. The protein of claim 5 wherein the variant is identified as SEQ ID NO: 4 or SEQ ID NO: 6.

7. The protein of claim 5 acting as a tumor suppressor in breast cells.

8. The protein of claim 5 acting as a marker for malignancy of breast cells.

9. The protein of claim 5 acting as a marker for malignancy progression and tumorigenic reversion.

10. A tumor suppressing agent comprising a protein consisting of an amino acid sequence SEQ ID NO: 3 or a variant, mutant or fragment thereof.

11. The agent of claim 10 wherein the protein variant is identified as the amino acid sequence SEQ ID NO: 4 or SEQ ID NO: 6.

12. A marker for detection of malignancy and for" determination of malignancy progression and tumorigenic reversion comprising a protein consisting of an amino acid sequence SEQ ID NO: 3 or a variant, mutant or fragment thereof .

13. The marker of claim 12 wherein the protein variant is identified as the amino acid sequence SEQ ID NO: 4 or SEQ ID NO: 6.

14. A malignant T4-2 breast tumor cell line derived from nonmalignant fibrocystic breast cells HMT 3522-S1.

15. The cell line of claim 14 obtained by repeated passaging immortalized nonmalignant HMT 3522 S-l cells in epidermal growth factor deprived culture, injected into nude mice and repeatedly passaged therein.

16. A method for diagnosis of breast cell malignancy comprising steps (a) detecting in a patient's tissue a degree of expression of AZ-1 gene or a level of a protein encoded therein; and

(b) correlating the degree of expression of AZ-1 gene with breast cells malignancy wherein the high expression of AZ-1 gene and high level of the protein encoded therein is correlated with nonmalignancy and the low or nonexistent AZ- 1 gene expression and a low or nonexistent level of the protein encoded by AZ-l gene is correlated with malignancy.

17. The method of claim 16 wherein the diagnostic detection is repeated every month to determine malignancy progression.

18. The method of claim 17 wherein the diagnostic detection is run in parallel with a cancer treatment.

19. The method of claim 18 wherein the diagnostic detection determines a treatment efficacy.

20. The method of claim 19 wherein the cancer treatment comprises administration of a protein, variant, mutant or fragment encoded by a DNA consisting of a nucleotide sequence identified as SEQ ID NO: 1 or a variant, mutant or fragment thereof.

21. The method of claim 20 wherein the protein has an amino acid sequence SEQ ID NO: 3 and the variant has an amino acid sequence SEQ ID NO: 4 or SEQ ID NO: 6.

22. The method of claim 21 wherein the variant has the amino acid sequence SEQ ID NO: 4.

23. A method for therapy and prevention of breast cell malignancy comprising steps:

(a) detecting in a patient's tissue a degree of expression of AZ-1 gene level of a protein encoded by AZ-1 gene; and (b) correlating the degree of expression of AZ-1 gene with breast cells malignancy wherein the high expression of AZ-1 gene and high level of the protein encoded by AZ-1 gene is correlated with nonmalignancy and the low or nonexistent AZ-1 gene expression or low or nonexistent levels of the protein encoded by AZ-1 gene is correlated with malignancy.

24. The method of claim 23 wherein the therapy and prevention detection is repeated every month to determine malignancy progression.

25. The method of claim 24 wherein the therapy and prevention detection is run in parallel with a cancer treatment.

26. The method of claim 25 wherein the therapy and prevention comprises gene therapy or administration of the protein encoded by AZ-1 gene.

27. The method of claim 26 wherein the cancer therapy and prevention treatment comprises administration of a protein encoded by a DNA consisting of a nucleotide sequence identified as SEQ ID NO: 3 or a variant, mutant or fragment thereof.

28. The method of claim 27 wherein the variant has an amino acid sequence SEQ ID NO: 4 or SEQ ID NO: 6.

29. The method of claim 27 wherein the fragment of AZ- 1 protein comprises a SPAZI or coiled-coil domain.

30. A method for gene therapy of breast cell malignancy by providing a patient in need thereof a genetically engineered material inducing expression of AZ-1 gene in breast cells.

31. The method of claim 30 wherein the genetically engineered material is targeted to the breast cells.

32. The method of claim 31 wherein the genetically engineered material is encapsulated in liposomes.

33. A method for detection of AZ-1 gene expression comprising steps: (a) obtaining a biopsy sample from breast tissue;

(b) contacting said sample with AZ-1 antibodies; and

(c) detecting formation of AZ-1 protein antibody complex.

34. The method of claim 33 wherein the antibodies are polyclonal.

35. The method of claim 34 wherein the antibodies are monoclonal.

36. A kit for detection of breast malignancy or premalignancy comprising: a means for detecting expression of AZ-1 gene, or a means for detecting a presence of AZ-1 gene expressed protein.

37. The kit of claim 36 wherein the means for detection is a RT-PCR based system or in situ hybridization to AZ-1 RNA.

38. The kit of claim 36 wherein the means for detection of the AZ-1 expressed protein is a polyclonal or monoclonal antibody.