In preliminary studies, rhUG inhibited angiogenesis in the ex vivo rat aorta model and showed antitumor activity against human prostate tumor cells (PC-3) in the chick chorioallantoic membrane assay, reducing both tumor volume and vascularity. A recent in vivo pilot study showed that twice daily dosing with rhUG resulted in a statistically significant increase in survival without evidence of toxicity in severe combined immunodeficient mice challenged with a PC-3 cell metastasizing tumor. Thus, rhUG may slow the progression of cancer by inhibiting both tumor cell invasiveness and tumor angiogenesis. It therefore holds the potential to serve as a new weapon in the arsenal of cytostatic, antimetastatic, adjuvant treatment for cancer. In this paper, we will briefly discuss the therapeutic potential of uteroglobin-based strategies for managing prostate cancer.

Prostrate Cancer and Metastasis:
Adenocarcinoma of the prostate is the most commonly diagnosed nonepidermal cancer in men and is the second leading cause of cancer mortality in this group. More than 210,000 men will be diagnosed with prostate cancer in 2002, and it is estimated that 41,800 will die from this disease.1 With prostate cancer as with all solid tumors, once metastasis occurs, there is relentless progression of the disease and it is the metastatic encroachment of cancer on vital organs that leads to the demise of the patient. Deaths due to prostate cancer invariably result from invasion and systemic metastasis to the lymph nodes, liver, lungs, and skeleton. The diagnosis and conventional treatment of many solid tumors employ a strategy of ablation or removal of the primary lesion followed by local salvage therapy alone or in combination with adjuvant systemic therapy when occult or overt metastases is detected.

Currently, however, there are few therapeutic strategies targeted to controlling tumor cell growth and the metastatic process. Recently though, our understanding of the molecular and cellular biology of invasion and metastasis has increased and new targets for therapeutic intervention have been identified. It is now recognized that the malignant progression of cancer is a continuum with a wide window for intervention, and strategies are now being designed to delay the progression of tumor invasiveness as well as slow the growth and invasive spread of existing metastases.2-5 The process of metastasis has been reviewed extensively2-5 and is summarized here briefly using prostate cancer as an example. In order for a prostate epithelial cell to metastasize successfully, it must first upset the normal balance between local proliferation and apoptosis to exhibit a net accumulation of cells. A combination of epigenetic factors and genetic instability probably allows for expansion of phenotypic and, possibly, genotypic variants capable of escaping the confines of local cellular architecture. In prostate adenocarcinoma, these variants are glandular epithelial cells that migrate toward the basement membrane attracted to the hemotactic stromal factors secreted by prostatic fibroblasts on the other side of the basement membrane.

With the appropriate acquired complement of cell surface adhesion molecules, proteolytic enzymes, and motility factors, a variant tumor cell must successfully attach to, degrade, and move through the basement membrane, as well as into and through the interstitial stromal matrix. As tumor cells encounter the microvasculature, especially the tumor-induced eomicrovasculature,
they must intravasate by invasion of the subendothelial basement membrane followed by induction of endothelial cell retraction. Once in the circulation, tumor cells must survive circulatory hemodynamic stress and host defense mechanisms, then specifically adhere to endothelial cells in the microvasculature of a target organ, induce endothelial cell retraction, extravasate from the circulation by invasion of the target organ subendothelial basement membrane and the interstitial stromal matrix. There, they must encounter the appropriate growth milieu in order to resume proliferation (colonization) and re-activate the process of angiogenesis.

Uteroglobin: General Background:
Uteroglobin (UG), also known as Clara cell 10-kd (CC10) protein or blastokinin, is a small (15.8 kd) secreted, cytokinelike homodimer that is secreted exclusively by epithelial cells but has effects on both epithelial cells and fibroblasts.6-8 It was first purified in 1977 from the uterus of a gravid rabbit. Subsequently, UG mRNA has been detected in the tracheobronchial tree, gastrointestinal tract, genitourinary tract, lung, prostate, mammary gland, pituitary gland, and thyroid.6-9 The UG protein has so far been found in the epithelial cells of the lung, tracheobronchial tree, endometrium, uterine tissue, prostate, lung, pituitary gland, and thymus.6-9 It is not found in endothelial cells, fibroblasts, or muscle cells. Very low amounts of UG have
been detected in plasma (not lymphocytes) but it is found in the urine where it is known as urine protein 1.6-8

Molecular Structure:
Uteroglobin is an antiparallel dimer formed by two identical subunits of 70 amino acids each. Two disulfide bonds are formed at cysteine (Cys)3 and Cys69', but the subunits do not easily dissociate after reduction of the bonds because of strong hydrophobic and Van der Waal’s interactions within the hydrophobic pocket.6-8 It has been proposed that the disulfide bridges govern access to the pocket and, when reduced, allow the UG dimer to bind lipids and steroid hormones. Each monomer contains 4 alpha-helical stretches and one beta turn. Uteroglobin is a true secretory protein with a canonical 21- amino acid N-terminal leader peptide. The human UG gene is single copy, about 3 Kb with 3 exons, 2 introns, canonical exon-intron junctions, and typical TATA box (TATAAAA).

The UG mRNA is approximately 600 bases long with a highly conserved 5' untranslated region and a 3' untranslated region including the polyadenylation signal.6-8 The human form of UG was first found in the nonciliated bronchiolar Clara cells,10 and was named CC10 protein based on its apparent molecular weight in sodium dodecyl surfate gels. The human gene is located at 11q12.3-13.1,7 a region exhibiting frequent disruption in a number of human cancers.

The UG gene is under multihormonal control as evidenced in the rabbit oviduct, uterus, and lung, where its expression is controlled by estradiol, progesterone, and corticosteriods, respectively. 6-8 The 5' region of the rabbit gene contains 4 progesterone/ glucocorticoid responsive sites and 2 progesterone/estrogen responsive sites. In humans, only 2 of the progesterone/ glucocorticoid and none of the estrogen-responsive elements have been conserved.

Crystallographic data indicate that the molecular surface topology of UG is similar to that of phospholipase A2 (PLA2).6-8 Uteroglobin binds to PLA2 in vitro and is a noncompetitive inhibitor of in vitro PLA2 activity. We have shown that rhUG suppresses the release of arachidonic acid from growth factor–stimulated cells in culture.11 Uteroglobin and some UG-derived peptides exhibit antiinflammatory activity in some experimental systems, and this activity has been extensively reviewed.6-8 There are small regions of similarity with
amino acid residues in lipocortin I (repeat 3), but UG is genetically distinct from the lipocortins. An inverse relationship between UG protein and leukotriene C4 in the tracheobronchial lavage of children with acute respiratory illness has been demonstrated.6-8 Several activities have been attributed to UG, including antichemotactic effects on macrophages, tolerogenic effects on maternal lymphocytes to spermatozoa; and inhibition of thrombin-induced platelet aggregation.6-8 New clues about its potential function have recently emerged from gene knockout studies (described below).

Uteroglobin Expression in Human Tumors:
Because CC10 is produced in the lung by Clara cells, early expression studies were focused on whether CC10 might be a marker for lung tumors of Clara-cell origin. However, although UG was found in normal bronchiolar epithelium, its expression was undetectable in many human lung carcinomas by in situ hybridization.12 Another study, using an immunohistochemical approach, also reported less frequent positivity for CC10 expression in lung tumor cells compared to normal epithelium.13 Thus, although the main goal of these studies was to evaluate UG as a positive marker for tumors of Clara-cell origin, their data were consistent with what is now understood as evidence that UG expression in the lung is inversely correlated with neoplastic growth. This was also supported by a study showing that in a transgenic simian virus-40 mouse model for lung carcinoma, UG gene expression was lost as the tumor progressed to the undifferentiated state.14

Our laboratory has shown that UG is abundantly expressed in normal prostate epithelial cells, in glands exhibiting benign prostatic hyperplasia, and in some low-grade tumors (Gleason score < 3), but lost or markedly reduced in moderate to highgrade invasive cancers (Gleason score > 3).15 Further support for the loss of UG expression with progression is seen in prostatic intraepithelial neoplasia, where UG expression is decreased compared to non-neoplastic prostatic epithelial cells, and in prostate cancer lymph node metastases, where no UG expression is detected.15 It has also been shown that UG expression is undetectable in human lung, uterine, and multiple prostate tumor cell lines.16,17,18

These studies have recently been validated by DNA microarray studies. Ernst et al19 performed a 12,600 gene DNA microarray
analysis of human prostate cancers and they reported that UG was significantly downregulated in cancer samples. They further confirmed this finding by quantitative real-time reverse transcriptase-polymerase chain reaction. Data soon to be posted on the NIH NCI Cancer Genome Anatomy Project (cGAP) indicate that normal human lung tissue yields a sequence Tag of > 500 per 200,000 whereas two lung adenocarcinomas yields Tag levels of 27 and 8 (J. Jen, MD and D. Sidransky, MD, personal communication, 2002). Beer et al,20 performed a microarray study of human lung cancers. Although not specifically reported in the published paper, they
found that the relative expression levels of UG in normal lung (n = 10), stage 1 adenocarcinoma (n = 67) and stage 3 adenocarcinoma
(n = 19), were 9025, 1267, and 560, respectively (D. Beer, MD, personal communication, 2002).

Uteroglobin as a Tumor Suppressor Antitumor Agent: In Vitro and Ex Vivo Studies:
Our laboratory was the first to show an anticancer activity for recombinant human UG (rhUG).11 Treatment of human prostate tumor cell lines with submicromolar, concentrations of rhUG for 12 hours inhibited the invasiveness of 3 prostate tumor cell lines and what is now known as a bladder cancer cell line by up to 90% (Figure 1).11 In addition, transient transfection of the human UG gene under control of a heterologous promoter inhibited invasiveness of PC-3 cells (Figure 2). Although it may not be the rate-limiting step in metastasis, tumor cell invasiveness is a fundamental characteristic of the malignant, metastatic phenotype. Thus, decreasing the invasiveness of prostate tumor cells may decrease metastatic potential and may have therapeutic value in controlling prostate cancer metastasis.

Szabo et al16 pursued the question of UG in lung cancer further by overexpressing a transfected CC10 (UG) gene in a small-cell lung cancer cell line that does not express detectable amounts of endogenous CC10. Tumor cells overexpressing the transgenic CC10 (UG) exhibited a marked reduction of invasiveness and decreased anchorage-independent growth.16 They concluded that CC10 expression antagonizes the neoplastic phenotype. For a tumor cell to traverse the basement membrane, some dissolution of the protein matrix is necessary. A tight balance must prevail, both between adhesion and proteolytic dissolution as well as between proteolytic enzymes and their inhibitors.21 Too much digestion of the basement membrane and the cell will be unable to gain the traction necessary for motion; too little and it will remain at the primary site. This balance is achieved by the expression of both proteases and protease inhibitors, but local invasion occurs when the balance favors increased protease activity. Currently, low-molecularweight matrix metalloproteinase (MMP) inhibitors are being developed for use in the clinic as antimetastatic agents.21 With
respect to UG, the study by Szabo et al16 showed that enforced overexpression of CC10 (UG) resulted in decreased expression
of MMP-2 and MMP-9 in lung tumor cells.

A study by Peri et al17 used human uterine (endometrial) tumor cells with enforced expression of transgenic rhUG. This resulted in a marked antineoplastic effect manifested as decreased proliferative capacity and loss of anchorage-independent growth potential. Additional transfection studies by Zhang et al18 will be discussed below with respect to potential mechanism of action.

Potential Antiangiogenic Properties of Uteroglobin:
In addition to its inhibition of tumor cell invasiveness, it appears that UG may possess another activity that is antagonistic to neoplastic growth. Using the standard rat aorta assay, we found that exposure of the aortic slices to 30 µg/mL or 60 µg/mL rhUG resulted in a decrease in the diameter and density of the halo network of new sprouts (Figure 3). The cells comprising the sprouts that grew in the presence of rhUG appeared stunted and markedly less interconnected as a network than the controls. Interestingly, rhUG did not interfere with the differentiation and tubule formation of pure cultures of endothelial cells, a process termed morphological angiogenesis (not shown). This suggests that the nonendothelial cellular components of true angiogenesis (mesenchymal pericytes,
smooth muscle cells), may be targets for rhUG’s activity. Thus, rhUG may have an antiangiogenic activity in addition to its direct
effects on tumor cell invasiveness.

In Vivo Studies in Transgenic Mice:
Although it is not yet published in an independent peer-reviewed publication, several published papers from a National Institutes of Health laboratory describe unpublished results regarding cancer in UG -/- knockout mice.18,22 The reports indicated that after 1.5 years, 16 out of 16 knockout mice exhibited malignancies compared to 0 out of 25 UG +/+ controls.

In Vivo Antitumor Effects of Recombinant Human Uteroglobin in the Chick Chorioallantoic Membrane Assay:
The effect of rhUG on tumor growth was further studied using a modified chick chorioallantoic membrane assay (CAM). We modified this assay by inoculating fertilized white Leghorn chick eggs CAM with human prostate tumor cells (PC-3) stably transfected with green fluorescent protein. The eggs were then divided into a control group, and 2 treatment groups. Recombinant human uteroglobin was then delivered in a continuous fashion to the vicinity of the tumor inoculation using an osmotic minipump. The pumps dispensed either saline (control group), 1 µg/mL rhUG, or 30 µg/mL rhUG, at a rate of 0.5 µL/hour for 7 days. The rate of delivery of rhUG to the CAM was 12 or 360 ng/day, respectively, for a total 7-day dose of approximately 0.1 or 2.5 µg.

On day 20, one day before hatching, the portion of CAM with tumor nodules was removed and evaluated for color, surface area, 3-dimensional shape, and vascularity (Figure 4). Visual examination revealed that nodules formed in both the salinetreated controls and the low-concentration group exhibited a similarly large surface area, with a 3-dimensional, bulky, highly vascular mass that extended down into the CAM. Due to better photographic quality, representative nodules from the low-concentration group were shown instead of the saline controls. In contrast, nodules that formed in the rhUG (30 µg/mL) higher concentration group exhibited a marked reduction in surface area and were flatter. These lesions also appeared to exhibit diminished vascularity (Figure 4). The reduced vascularization observed in the treated groups, when combined with results from the ex vivo antiangiogenesis assay (described above), further supports the conclusion that rhUG may be antiangiogenic.

In Vivo Antitumor Effects of Recombinant Human Uteroglobin in a Murine Model:
In conjunction with the Developmental Therapeutics Branch of the National Cancer Institute, a pilot in vivo murine (severe combined immunodeficient [SCID] mice) study was conducted to determine the therapeutic potential of rhUG as an anticancer agent. The study consisted of 60 SCID mice: 20 for the control group and 10 for each of 4 treatment groups. Based on preliminary pharmacokinetic data, the test doses selected were 2.5 mg/kg, 5 mg/kg, 10 mg/kg, and 20 mg/kg, twice daily through subcutaneous injections. Initially, the animals received a tumor challenge of 2.5 million human prostate cancer cells (PC-3). Treatment with rhUG or vehicle began on day 1 and continued through day 35 when the first control animal succumbed to the tumor. Severely moribund or suffering animals deemed to be imminently near death were sacrificed. All surviving animals were sacrificed on day 52. At the time of death or sacrifice, all animals bore visible metastatic tumor lesions.

The results of these studies demonstrated two significant findings. First, no toxicity was observed in any treatment group. In addition, a survival benefit was seen in the treatment groups; median survival time increased by 6-7 days in rhUG 20 mg/kg compared to controls (Figure 5A and 5B). At days 40 and 44, 28% and 56% of control mice had died from their tumor compared to 0% and 10% in the treated groups, respectively. The increased survival at rhUG 20 mg/kg achieved strong statistical significance (P = 0.02). Lower doses exhibited a trend toward dose-dependent increased survival, but the differences were not statistically significant. The outcome of this experiment was likely adversely affected by ending treatment at day 35, allowing unchecked tumor growth to occur for 10-17 days before tumorbearing animals died. Nevertheless, these interesting results, achieved with less than maximally tolerated doses, lend optimism for an even greater response with dose regimens that include longer treatment schedules.

Potential Mechanisms of Uteroglobin's Anticancer Activity:
Although our understanding of uteroglobin’s mechanism of action is not yet well elucidated, at present it appears that two of its known activities may contribute to its function. These include (1) binding to fibronectin and altering the signaling pathways triggered by a cell’s interaction with extracellular matrix proteins16,22 and (2) inhibiting the release and activity of proinvasive arachidonic acid metabolites.11,17 Both of these may additionally result in downstream downregulation of matrix metalloproteinases. Therefore, the extracellular matrix effects and the intracellular signal transduction effects on arachidonate may be a complementary and possibly synergistic dual mechanism of action against tumor biology.

Before a tumor cell can invade, it must first attach to the basement membrane. The basement membrane and its underlying interstitial stroma comprise the extracellular matrix. The basement membrane consists of collagen, glycoproteins such as laminin and fibronectin, and proteoglycans, which intercalate in a dense matrix, thus preventing any passive traversal of this membrane.23 Attachment is mediated via interactions with various cell adhesion molecules, frequently subclassified into integrin and nonintegrin varieties. Integrins have been implicated in metastasis because many cancer cells exhibit alterations in expression and abnormal distribution of these receptors on their cell surfaces. 23 Binding of the integrin-fibronectin receptor with receptor-binding fragments resulted in a decreased invasive capability in vitro and in some lung colonization assays.24 This is exactly what Szabo et al16 showed with uteroglobin, where overexpression of CC10 (UG) resulted in decreased adhesion of lung tumor cells to fibronectin. Further evidence of the UG-fibronectin relationship has been reported in a very intriguing study using transgenic mice, which indicated that one of UG’s principle natural functions may be to bind to fibronectin and modulate fibronectin structure and function in the extracellular matrix protein.22

A complementary second mechanism of action for UG involves its inhibition of arachidonic acid metabolism. The integrins and proteases are components of a complex signaling network that propagates the invasive phenotype.25 These pathways intersect with a G protein-coupled kinase/phosphatase cascade5 and an eicosanoid-mediated signaling cascade, both of which have been shown to have a key role in modulating the invasiveness and metastatic potential of tumor cells. Inhibitors of arachidonic acid metabolism, especially cyclooxygenase inhibitors, have demonstrated anticancer activity in human studies of cancers of the colon, esophagus, stomach, and rectum.26 These compounds inhibit metastases in experimental animals27 and inhibit invasiveness of some tumor cells in vitro.27,28 Our laboratory was the first to show that rhUG inhibits PLA2 activity in prostate tumor cells and that specific inhibitors of PLA2 partially inhibit prostate tumor cell invasiveness in vitro.11,27 These results suggest that the inhibition of arachidonic acid release by UG may be a component of UG’s anti-invasive mechanism.

Further support for this concept emerged in a study by Peri et al17 using human uterine (endometrial) tumor cells. Although normal uterine cells produce UG, cancerous uterine cells do not. Enforced expression of transgenic rhUG in these tumor cells resulted in inhibition of PLA2, which in turn resulted in decreased synthesis and secretion of the autocrine tumor growth factor known as platelet-activating factor. UG transfectants exhibited markedly decreased proliferative capacity and loss of anchorageindependent growth potential.

Additionally, a study by Zhang et al18 using uterine tumor cells and an inducible rhUG expression vector showed that rhUG inhibited
both anchorage-independent growth and extracellular matrix invasion. These authors have attributed the anti-invasive activity of rhUG to a receptor-based mechanism.18,29 However, although several cell surface uteroglobin binding proteins have been found, an unequivocal receptor for UG has yet to be identified and there is conflicting data in the literature about this subject. The concept that a UG receptor is necessary for rhUG’s antiinvasive action does not fit well with data on the secretion and binding of UG to fibronectin, unless the putative UG receptor is also the fibronectin-integrin receptor complex. Perhaps UG is bound to fibronectin fragments, which in turn can be found associated with specific integrin receptors. It is noteworthy that other laboratories,11,16,17 have reported that UG exhibits marked anti-invasive activity against cell lines that apparently lack the binding proteins thought to be the putative UG receptor.

Conclusion:
Tumor cell invasion may not be the rate-limiting step for the development of metastasis, but it represents a necessary and critical
component of the metastatic cascade that is intimately connected to both colonization and angiogenesis. Our studies of the process of invasion have identified at least two targets for intervention by rhUG in prostate cancer: (1) the invasive motility of the tumor cells
themselves and (2) the process of angiogenesis. Additional results in an in vivo mouse study indicate that replacement therapy with
rhUG has therapeutic potential for epithelial cell cancers including prostate cancer and justify further exploration of UG as a biologic
agent for therapeutic purposes. The determination of therapeutic potential of UG may represent an innovative and promising approach
to controlling tumor growth and metastasis. In addition to the therapeutic potential of UG, there is evidence that the loss of UG expression in the progression of prostate and lung cancer may have prognostic potential in predicting the metastatic potential of an individual cancer and the need for UG replacement therapy.

Summary of Potential for Recombinant Human Uteroglobin as a Therapeutic Antitumor Agent:

1. UG is produced and secreted by many types of epithelial cells, including lung, prostate, pancreas, urogenital tract, and gastrointestinal tract.
2. UG is naturally found in the circulation (at low levels) and in the urine.
3. Expression of UG is lost as tumors of these epithelial tissues progress to the point of acquiring invasiveness and metastatic potential. UG double-knockout mice (UG -/-) are prone to tumor formation.
4. Pharmacological and transgenic reconstitution of rhUG to epithelial tumor cells (to date shown in prostate, lung, and uterine cancer) have direct effects on tumor cells, markedly inhibiting their invasiveness and reversing several key characteristics of the
5. Recombinant human UG may exert additional indirect effects on tumor progression through inhibition of angiogenesis.
6. Recombinant human UG treatment resulted in increased survival in the in vivo SCID mouse metastasis model using human prostate tumor cells (PC-3), at nontoxic doses.
7. Recombinant human UG is a small (15.8 kd), stable, nonglycosylated (ie, no posttranslational modification required for activity) protein that can be produced at relatively high yield in Escherichia coli and other recombinant production systems. It appears to be biochemically compatible with formulation for multiple routes of administration.

References:

1. Cancer facts and figures. American Cancer Society 2002.
2. Kohn EC. Invasion and metastasis: biology and clinical potential. Pharmacol Ther 1991; 52:235-244.
3. Aznavoorian S, Murphy AN, Stetler-Stevenson WG, et al. Molecular aspects of tumor cell invasion and metastasis. Cancer 1993; 71:1368- 1383.
4. Kohn EC, Liotta LA. Molecular insights into cancer invasion: strategies for prevention and intervention. Cancer Res 1995; 55:1856-1862.
5. Liotta LA, Stetler-Stevenson WG, Steeg PS. Cancer invasion and metastasis: positive and negative regulatory elements. Cancer Invest 1991; 9:543-551.
6. Miele L, Cordella-Miele E, Mantile G, et al. Uteroglobin and uteroglobin- like proteins: the uteroglobin family of proteins. J Endocrinol Invest 1994; 17:679-692.
7. Zhang Z, Zimonjic DB, Popescu NC, et al. Human uteroglobin gene: structure, subchromosomal localization, and polymorphism. DNA Cell Biol 1997; 16:73-83.
8. Miele L, Cordella-Miele E, Mukherjee AB. Uteroglobin: structure, molecular biology, and new perspectives on its function as a phospholipase A2 inhibitor. Endocr Rev 1987; 8:474-490.
9. Peri A, Cordella-Miele E, Miele L, et al. Tissue-specific expression of the gene coding for human Clara cell 10-kD protein, a phospholipase A2- inhibitory protein. J Clin Invest 1993; 92:2099-2109.
10. Plopper CG, Hill LH, Mariassy AT. Ultrastructure of the nonciliated bronchiolar epithelial (Clara) cell of mammalian lung. III. A study of man with comparison of 15 mammalian species. Exp Lung Res 1980; 1:171-180.
11. Leyton J, Manyak MJ, Mukherjee AB, et al. Recombinant human uteroglobin inhibits the in vitro invasiveness of human metastatic prostate tumor cells and the release of arachidonic acid stimulated by fibroblast-conditioned medium. Cancer Res 1994; 54:3696-3699.
12. Broers JL, Jensen SM, Travis WD, et al. Expression of surfactant associated protein-A and Clara cell 10 kilodalton mRNA in neoplastic and non-neoplastic human lung tissue as detected by in situ hybridization. Lab Invest 1992; 66:337-346.
13. Linnoila RI, Jensen SM, Steinberg SM, et al. Peripheral airway cell marker expression in non-small cell lung carcinoma. Association with distinct clinicopathologic features. Am J Clin Pathol 1992; 97:233-243.
14. Sandmoller A, Halter R, Suske G, et al. A transgenic mouse model for lung adenocarcinoma. Cell Growth Differ 1995; 6:97-103.
15. Weeraratna AT, Cajigas JA, Schwartz A, et al. Loss of uteroglobin expression in prostate cancer: relationship to advancing grade. Clin Cancer Res 1997; 3:2295-2300.
16. Szabo E, Goheer A, Witschi H, et al. Overexpression of CC10 modifies neoplastic potential in lung cancer cells. Cell Growth Differ 1998; 9:475-485.
17. Peri A, Bonaccorsi L, Muratori M, et al. Uteroglobin reverts the transformed phenotype in the endometrial adenocarcinoma cell line HEC-1A by disrupting the metabolic pathways generating platelet-activating factor. Int J Cancer 2000; 88:525-534.
18. Zhang Z, Kundu GC, Panda D, et al. Loss of transformed phenotype in cancer cells by overexpression of the uteroglobin gene. Proc Natl Acad Sci U S A 1999; 96:3963-3968.
19. Ernst T, Hergenhahn M, Kenzelmann M, et al. Decrease and gain of gene expression are equally discriminatory markers for prostate carcinoma. Amer J Pathology 2002; 160:2169-2180.
20. Beer D, Kardia S, Huang C, et al. Gene-expression profiles predict survival of patients with lung adenocarcinoma. Nat Med 2002; 8:816-824.
21. Stetler-Stevenson WG, Hewitt R, Corcoran M. Matrix metalloproteinases and tumor invasion: from correlation and causality to the clinic. Semin Cancer Biol 1996; 7:147-154.
22. Zhang Z, Kundu GC, Yuan CJ, et al. Severe fibronectin-deposit renal glomerular disease in mice lacking uteroglobin. Science 1997; 276:1408- 1412.
23. Akiyama SK, Olden K, Yamada KM. Fibronectin and integrins in invasion and metastasis. Cancer Metastasis Rev 1995; 14:173-189.
24. Knutson JR, Iida J, Fields GB, et al. CD44/chondroitin sulfate proteoglycan and alpha 2 beta 1 integrin mediate human melanoma cell migration on type IV collagen and invasion of basement membranes. Mol Biol Cell 1996; 7:383-396.
25. Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 1992; 69:11-25.
26. Levy GN. Prostaglandin H synthases, nonsteroidal anti-inflammatory drugs, and colon cancer. Faseb J 1997; 11:234-247.
27. Reich R, Martin GR. Identification of arachidonic acid pathways required for the invasive and metastatic activity of malignant tumor cells. Prostaglandins 1996; 51:1-17.
28. Attiga F, Fernandez P, Weeraratna A, et al. Inhibitors of prostaglandin synthesis inhibit prostate tumor cell invasiveness and reduce the release of matrix metalloproteinases. Canc Res 2000: 60;4629-4637.
29. Kundu GC, Mantile G, Miele L, et al. Recombinant human uteroglobin suppresses cellular invasiveness via a novel class of high-affinity cell surface binding site. Proc Natl Acad Sci U S A 1996; 93:2915-2919.