USE OF eIF-5A TO KILL MULTIPLE MYELOMA CELLS The present application is a divisional application of Australian Application No. 2006325752, which is incorporated in its entirety herein by reference. 5 RELATED APPLICATIONS This application claims priority to U.S. provisional application 60/749,604, filed on December 13, 2005 and 60/795,168, filed on April 27, 2006, both of which are incorporated by reference in their entirety. 0 FIELD OF THE INVENTION The present invention relates to apoptosis-specific eukaryotic initiation factor ("eIF-5A") and the use of polynucleotides encoding the same to kill multiple myeloma cells, as well as other cancer cells. The present invention relates to the use of apoptosis-specific eIF-5A or referred to 5 as "apoptosis-specific elF-5A" or "eIF-5A " as well as the use of the eIF-5A2 isoform to inhibit multiple myeloma, kill multiple myeloma cells, and to inhibit and/or kill other cancer cell growth. BACKGROUND OF THE INVENTION D Apoptosis is a genetically programmed cellular event that is characterized by well defined morphological features, such as cell shrinkage, chromatin condensation, nuclear fragmentation, and membrane blebbing. Kerr et al. (1972) Br. J. Cancer, 26, 239-257; Wyllie et al. (1980) Int. Rev. Cytol., 68, 251-306. It plays an important role in nonnal tissue development and homeostasis, and defects in the apoptotic program are thought to contribute to a wide range 5 of human disorders ranging from neurodegenerative and autoimmunity disorders to neoplasms. Thompson (1995) Science, 267, 1456-1462; Mullauer et al. (2001) Mutat. Res, 488, 211-231. Although the morphological characteristics of apoptotic cells are well characterized, the molecular pathways that regulate this process have only begun to be elucidated. 1 Another key protein involved in apoptosis is a protein that encoded by the tumor suppressor gene p53. This protein is a transcription factor that regulates cell growth and induces apoptosis in cells that are damaged and genetically unstable, presumably through up-regulation of Bax. Bold et al. (1997) Surgical Oncology, 6, 133-142; Ronen et al, 1996; Schuler & Green 5 (2001) Biochen. Soc. Trans., 29, 684-688; Ryan et al. (2001) Curr. Opin. CellBioL, 13, 332 337; Z6rnig et al. (2001) Biochein. Biophys, Aca, 1551, FI-F37. la Alterations in the apoptotic pathways are believed to play a key role in a number of disease processes, including cancer. Wyll ie et at (1980).nt. Rev. (vtol, 68, 251-306; Thompson (1995) Science, 267, 1456-1462; Sen & D'Incalci (1992) FEBSLleters, 307, 122-127; McDonnell et a, (1995) Seminars in Cancer and Biology 6, 53-60. 5 investigations into cancer development and progression have traditionally been focused on cellular proliferation. However, the important role that apoptosis plays in tumorigenesis has recently become apparent, In fact, much of what is now known about apoptosis has been learned using tumor models, since the control of apoptosis is invariably altered in some way in tumor cells. Bold et at (1997) Surgical Oncology, 6, 133-142. 10 Cytokines also have been implicated in (he apoptotic pathway. -Biological systems require cellular interactions for their regulation, and cross-talk between cells generally involves a large variety of cytokines. Cytokines are mediators that are produced in response to a wide variety of stimuli by many different cell types. Cytokines are pleiotropic molecules that can exert many difTerent effects on many different cell types 15 but are especially important in regulation of the immune response and hematopoietic cell proliferation and differentiation. The actions of cytokines on target cells can promote cell survival proliferation, activation, differentiation, or apoptosis depending on the particular cytokine, relative concentration, and presence of other mediators. Deoxyhypusine synthase (DHS) and hypusine-containing eukaryotic translation 20 initiation Factor-5A (elF-SA) are known to play important roles in many cell ular processes ing cell gypusine, a unique amino acid, is found in all examined eukaryotes and archaebacteria, but not in eubacteria, and eiF-5A is the only known hypusine-containing protein. Park (1988) .1 Bol Chem., 263, 7447-7449; Scimann & Klink (1989) Systmn, App. Microbial, i. 103~107; Bartig et alt (1990) 25 System. App!. MicrobioL, 13, 112-116; Gordon et al, (1987a) J Iol (hem., 262, 16585 16589. Active elF-5A is formed in two post-translational steps: the first step is the formation of a deoxyhypusine residue by the transfer of the 4-aminobutyl moiety of spermidine to the ri-amino group of a specific lysine of the precursor eIF-5A catalyzed by deoxyhypusine synthase; the second step involves the hydroxylation of this 4-aminobutyl 30 moiety by deoxyhypusine hydroxy lase to form hypusi ne. The amino acid sequence of elF-5A is well conserved between species, and there is strict conservation of the amino acid sequence surrounding the hypusine residue in elF 5A, which suggests that this modification may be important for survival. Park et a. (1993) Bioaors. 4, 95-104. This assumption is further supported by the observation that inactivation of both isofonrs of elF-5A found to date in yeast, or inactivation of the DHS gene, which catalyzes the first step in their activation, blocks cell division. Schnier et a. (1991) Mo6 Cell. Blot., 11. 3105-3114; Sasaki et at. (1996) FES Left,. 384, 151-154; Park et at. (1998)1 Biol Chem., 273, 1677-1683, However, depletion of e[f-5A protein 5 in yeast resulted in only a small decrease in total protein synthesis suggesting that eaF-5A may be required for the translation of specific subsets of mRNA's rather than for protein global synthesis. Kang et al. (1993), "Effect of initiation factor eJF-5A depletion on cell proliferation and protein synthesis," in Tuite, M. (ed.), Protein Synthesis and Targeting in Yeast, NATO Series HT. The recent finding that ligands that bind efF-SA share highly 10 conserved motifs also supports the importance of eF-5A, Xu & Chen (2001),. Riol Chem 276, 2555-2561 In addition, the hypusine residue of modified elF-5A was found to be essential or sequence-specific binding to RNA, and binding did not provide protecion. from ribon ucl eases. In addition, intracellular depletion of elf-5A results in a significant accumulation 15 of specific mRNAs in the nucleus, indicating that elF-5A may be responsible for shutling specific classes of mRNAs from the nucleus to the cytoplasm. Liu & Tartakoff (1997) Supplement to Molecular Biology of the Cell., 8 426a. Abstract No. 2476, 37th American Society for Cell Biology Anual Meeting. The accumulation of eF-5A at nuclear pore associated intranuclear filaments and its interaction with a general nuclear export receptor 20 further suggest that elF-5A is a nucleocytoplasmic shuttle protein, rather than a component of polysonies. Rosorius et al. (1999)1 Cei Science, .112, 2369-2380. 'The first cDNA for efF-5A was cloned from human in 1989 by Smit-McBride et at, and since then cDNAs or genes for elF-5A have been cloned from various eukaryotes including yeast, rat, chick embryo, alfadfa, and tomato. Smit-McBride et al. (1989)1 .iot 25 (hem., 264, 1578-1583; Schnier et at. (1991) (y east); Sano, A. (1995) in Imahori. M et al. (eds), Polyamines, Basic and Clinical Aspects, VNU Science Press, The Netherlands, 81 88 (rat); Rinaudo & Park (1992) PA SEJ,/I./. 6, A453 (chick embryo); Pay eta]. (1991) P/ant oi. Biol, 17, 927-929 (alfalfa): Wang et at. (2001), Bio Chen., 276, 17541 17549 (tomato). 30 Multiple myeloma is a. progressive and fatal disease characterized by the expansion of malignant plasma ells in the bone marrow and by the presence of osteolytic lesions. Multiple myeloma is an incurable but treatable cancer of the plasma. cell. Plasma cells are an important part of the immune system, producing immunoglobulins (antibodies) that help fight infection and disease. Multiple myeloma is characterized by excessive numbers 3 of abnornal plasma cells in the bone marrow and overproduction of intact monoclonal immunoglobulins (IgO, IgA, JgD. or IgE; "M-proteins") or Bence-Jones protein (free monoclonal light chains). Hypocalcaemia, anemia, renal damage, increased susceptibility to bacterial infection, and impaired production ofnormal immunoglobulin are common 5 clinical manifestations of multiple myeloma. Multiple myeloma is often also characterized by diffuse osteoporosis, usually in the pelvis, spine. ribs, and skull. Conventional therapies for of multiple myeloma include chemotherapy, stem cell transplantation, high-dose chemotherapy with stem cell transplantation, and salvage therapy. Chemotherapies include treatment with Thalomiid@(uthalidoniide), bortezoniib, 10 Aredia@P. (pamidronate), steroids, and Zometa@ (zoledronic acid). However many chemotherapy drugs are toxic to actively dividing non-cancerous cells, such as of the bone marrow, the lining of the stomach and intestines, and the hair follicles. Therefore, chemotherapy may result in a decrease in blood cell counts,. nausea, vomiting, diarrhea, and loss of hair. 15 Conventional chemotherapy, or standard-dose chemotherapy, is typically the primary or initial treatment for patients with of Tmultiple myelomia. Patients also may receive chemotherapy in preparation for high-dose chemotherapy and sten cell transplant, Induction therapy (conventional chemotherapy prior to a. stem cell transplant) can be used to reduce the tumor burden prior to transplant. Certain chemotherapy drugs are more 20 suitable for induction therapy than others, because they are less toxic to bone marrow cells and result in a greater yield of stem cells from the bone marrow. Examples of chemotherapy drugs suitable for induction therapy include dexamethasone, thalidomide/dexamethasone, VAD (vincristine, Adriamycin@i} (doxorubicin), and dexamethasone in combination), and Vd. (pegylated liposomal doxorubicin (Doxi@,i 25 Caelyx@), vincristine, and reduced schedule dexarethasone in combination). The standard treatment for of multiple myeloma is melphalan in combination with prednisone (a corticosteroid drug), achieving a response rate of 50%. Unfortunately, melphalan is an alkylating agent and is less suitable for induction therapy. Corticosteroids (especially dexamethasone) are sometimes used alone for multiple myeloma therapy, 30 especially in older patients and those who cannot tolerate chemotherapy. Dexamethasone is also used in induction therapy, alone or in combination with other agents. VAD is the most commonly used induction therapy, but DVd has recently been shown to be elective in induction therapy. Bortezomib has been approved recently for the treatment of multiple neloma, but it is veiy toxic. However, none of the existing therapies offer a significant 4 potential ibr a cure. Thus, there remains a need for a suitable therapy to kill nnltiple yeloma cells. The present invention provides this need. SUMMARY OF INVENTION 5 The present invention provides a method of inhibiting cancer cell growth and/or killing cancer cells. The present invention also provides a method of inhibiting or slowing down the ability of a cancer cell to metastasize. Inhibiting cancer growth includes a reduction in the size of a tumor, a decrease in the growth of the turor, and can also encompass a complete remission of the tumor. The cancer can be any cancer or tumor. 10 including but not limited to colon cancer, colorectal adenocarcinoma, bladder carcinoma, cervical adenocarcinoma, and lung carcinoma, The methods of the present invention involve the administration of eVF-5A, preferably human elF-5A1 to a patient (a inammal, preferably a human) having said cancer. The elf-5A2 isoform may also be used, although eIF-SAI is preferred. The efF-5A may be delivered to a subject in need thereof by any 15 suitable method know in the art, It may be delivered as naked DNA, such as DNA in biologically suitable medium and delivered through IV or subcutaneous injection or any other biologically suitable delivery mechanism. Alternatively, the elF-5A may be delivered in a vector such as an adenovirus vector, Alternatively, the DNA may be delivered in liposomes or any other suitable"carrier" that provided for delivery of the 20 DNA to the target cancer cells. The elF-SA iay also be delivered directly to the site of the tumor. One skilled in the art would be able to determine the dose and length. of treatment regimen for delivery of eIF-5A. elF-SAl and elF-5A2 is known and has been described in earlier co-pending applications, such as 09/909,796 (U S. 6,867,237); 10/141,647 (allowed); I0200.148; 25 10/277,969; 10/383,614; 10/792,893; 11/287,460; 10/861.980; 11/134,445; 11/184,982; 11/293,391; 60/749,604; and 60/795,168, which are all herein incorporated by reference. Since e[F-5As are highly conserved among species, any elF-5A may be used in the present invention, human, rat, mouse, dog etc. Preferably, a human elF-5A would be used for treatment of humans, etc. The elF-5 A also includes mutant elF-5As, as long as the mutant 30 is capable of up-regulating or increasing expression of elF-5A and hence inhibit the growth of cancer or kill cancer cells. The present invention also provides for a method of activating MAPKISAPK signaling pathway in a cell by providing a nucleotide encoding elf-5Al to said cells. 'The eF-5A polynucleotide and elF-5A1 protein is as described above. s The Present invention also provides pharmaceutical compositions useful for killing myelonma cells comprising polynucleotides encoding eiF5A.. The eIF5A. maybe eFS Al, eF5A2 or a mutant eIF5AL Preferably the eTF5A is elF5Al. The composition may further comprise a delivery vehicle. The delivery vehicle may be, but is not limited to, a 5 vector, plasniid, liposome, or dendrimer. The present t invention also provides the use of elF5A (preferably eF5A 1) to make a medicament to kill multiple myeloma cells in a subject having multiple myeloma. The present invention further provides a method of killing mul tiple myclorma cells comprising administering to the myeloma cells a composition comprising a polynucleotide 10 encoding eTFSA , wherein the composition kills the multiple myeloma cells. The eTF5Al may be a mutant, wherein the mutant has had the conserved lysine changed to another amino acid and wherein said mutant is unable to be hypusinated. Compositions usefiu in the methods of treatment are as described herein. The present invention further provides a method of killing multiple myeloma cells 15 wherein a composition comprising siRNA directed against elF-5Al is provided in addition to a composition comprising polynucleotides encoding elFA . The siRNA down regulates endogenous expression of elF-5A1, and thus down regulates expression of L.-6, which in turn causes in apoptosis in myeloma cells. The composition comprising the elF 5A I siRNA may be administered intravenously or administered within a delivery vehicle 20 such as a pasnmid, vector, liposome or dendrimer. BRIEF DESCRI1ON OF THE FIGURES Figure I shows that elF5A.1 expression is increased by genotoxic stress and. nitric oxide. 25 A) Northern blot (top panel) and Westem blot (bottom panel) analysis of eFSA expression in normal colon fibroblasts treated with 0.5 pg/ml Actinomycin D for(), 1, 4, 8, and 24 hours. The Westem blot was probed with antibodies against etFSA, p53, and f@ actin. B) Northern blot (top panel) and Westem blot (bottom panel) analysis of eIF5A expression in RKO cells treated with 3 mM sodium nitroprusside for 0, 2, 4, 8, and 24 30 hours. The Western blot was probed with antibodies against eIF5A and Il-actin. The RKO cells do not express eIF5A2 at significant levels. Further, elF-5A2 is only one amino acid longer in size than elF-5A 1, but it runs higher on an SDS PAGE gel, so that if eF5A2 were expressed, it would be seen as a separate band from elF-SA L. 6 Figure 2 shows suppression of eIF5A . expression has no effect on cell proliferation. A) Westem blots of cell lysates isolated from HT-29 cells 72 hours afler transfection with either elF5AI siRNA or control siRNA. Western blots from two independent experiments are shom. The blots were probed with antibodies against elFSA and f-actin. B) The 5 metabolic activity of cells transfected with e[F5A I. siRNA was measured using an XTT cell proliferation assay. HT-29 cells were seeded on a 96-well plate 24 hours before transfection with either control siRNA or eF5AI siRNA. Twenty-four hours after transfection, the cells were either left untreated or treated with Actinomycin D (LO pg/ml) for 48 hours before measuring metabolic activity. Values are means for two experiments 10 performed in quadruplicate and were nonnalized to the value obtained for the 0 hour control which was set at 1 C) The proliferative ability of IIT-29 cells transfected with control or elF5AI siRNA was compared to that of cells incubated with 50 M GC7 for 72 hours. Cell proliferation was measured by BrdU incorporation. Values are means + SE for n =4 and were normalized to the value for the GC7 (+) serum sample which was set at 15 L Asterisks (*) denote values considered significantly different by paired Student mest (p <0.01). D) XTT and E) BrdU incorporation cell proliferation assays of HT-29 cells transfected with either control siRNA or etF5A siRNA from the day of transfection (day 0) through to day 5 after transfection. Values for day 0 were set at L 20 Figure 3 shows that elF5AI regulates expression of p53 in response to Actinomycin. D. RKO cells were transfected with either control siRNA (C) or eF5A siRNA (5A- ). Seventy-two hours after transfection. the cells were treated for 0, 4, 8, or 24 hours with 0.5 pg/m Actinomycin D. A) Western blot of cell lysates blotted with antibodies against eIF5A, p53, or 0-actin. The result is representative of three independent experiments. B) 25 Plot of the relative intensities of p53 in Westem blots that were normalized to the corresponding actin bands. Values are means i SE for a minimum of n =3. Figure 4 shows that overexpression of human eWF5AT induces apoptosis independently of p53. RKO cells (A) or RKO-E6 cells (B) were transfected with pHM6-LacZ, pHM6 30 eIFSA, or pHM6-eIF5AA37 (a 37 amino acid truncation of the C-terminus). Forty-eight hours after transfection, the cells were fixed and labeled using the TUNEL method. The nuclei were stained with loesch.t 33258, and the labeled cells were viewed by fluorescence microscopy. Cells stained bright green were scored as apoptotic. Hoescht stained nuclei were used to determine the total cell number. Values are means + SE for in 7 4 (A) or a= 3 (B). Asterisks (*) denote significant difference from the control (pHM6 LacZ) by paired Student Q-test (p <0.02). C) Western blot of RKO and RKO-E6 cell lysates harvested 0. 4. 8 and 24 hours after treatment with 0.5 pg/mi Actinomycin D. Blots were probed with antibodies against p53 and p-actin, 5 Figure 5 shows that eIF5At adenovirus constructs induce apoptosis in colon cancer cells. A) Western blot of cell lysate isolated from HT-29 cells seventy-two hours after infection with Ad-LacZ (L), AdelF5AI (5A), or Ad-eiFSA(K50A) (K50A). B) 2-D gel electrophoresis of cell lysate isolated front -TT-29 cells after treatment with GC7 or DFO 10 for seventy-two hours or seventy-two hours after infection with adenovirus constructs followed by Western blotting with an antibody against elFSA. 7 pg of protein was separated except for lysate from Adenovirus-infected cells overexpressing eOF5A for which 0.3 g of protein was separated. C) Top panel: TUNEL staining of HT-29 cells forty-eight hours after infection with adenovirus constructs. Bottom panel: Hoechst 15 stained nuclei of cells in the same field. All photographs were taken at 400 X magnifcation. The results are representative of three independent experiments. D) Percent apoptosis of FIT-29 cells forty-eight hours after infection wait adenovirus constructs. Values are means + SE for n = 3. E) XTT cell proliferation assay of HT-29 cells seven days after infection with adenovirus constructs. Values are means -+ SE for n = 20 4. Figure 6 shows Annexin V and PT staining of HT-29 cells infected. with eFAI adenovirus constructs. A) Histograms of Annexin-FITC and propidium iodide (PI) labeling of IT-29 cells fort-eight after infection with adenovirus constructs. B) Percent apoptosis of HT-29 25 cells, as determined by Annexin V labeling and flow cytometiy analysis, twenty-four forty-eiglit and seventy-two hours after infection with adenovirus constructs or after treatment with the apoptosis-inducing agents, Actinonycin D or Brefeldin A. Values are means + SE for a = 2 (#- of events > 5000). 30 Figure 7 shows immunofluorescent localization of efFA1. The subcelhdar localization of eWF5A protein in HT-29 cells stimulated with IFN-y and TNF-a (A) or Actinomycin D (B) was determined by indirect immunofluoresceace. Hoechst 33258 was used to stain the nuclei. A) H.-29 cells were either untreated or primed with IFN-y for 16 hours before stimulation with TNF-g for O min. [(-) TNF-af, 10 min. or 30 min. Top panel: 8 inIUunofIuorescent detection of e]FSAI; middle panel: Hoechst-stained nuclei of cells in the same field; bottom panel: merged images. B) HT-29 cells were citrher untreated or treated with Actinomycin D for 0.5 ir, 1.5 hrs. or 4 irs. Top panel: inimunolluorescent detection of elF5A1; middle panel: Hoechst-stained nuclei of cells in the same field; 5 bottom panel: merged images. All photographs were taken at 400 X magnification. The results are representative of three independent experiments. Figure 8 shows that elF5AI regulates expression of p53 in response to Actinomycin D. RKO-E6 cells do not express p533. A) RKO or RKO-E6 cells were treated with 0.5 pg/mI 10 Actinomycin D for L 4, 8, or 24 hours. Cell lysate was harvested and analyzed for p53 expression using Western blotting. B) RKO cells were transfected with either control siRNA (C), elF5A siRNA (5A-1) or a second eIF5A siRNA. targeting a different region of ceF5A (5A-2). Seventy-two hours after transfection, the cells were treated for 0, 4, 8, or 24 hours with 0.5 pg/ml Actinornycin D. A) Western blot of cell lysates blotted with 15 antibodies against eIF5A, p53, or W-actin. The result is representative of three independent experiments. Figure 9 shows that deoxyhypusinated eIF5A 1 accumulates during apoptosis. This figure provides a 2-D gel electrophoresis of cell lysate isolated from IT-29 cells after treatment 20 with Actinomycin D (A) or an agonist Fas antibody (B) for times ranging from I to 24 hours followed by Western blotting with an antibody against elFSA. Figure 10 shows that infection with .Ad-eIF5.A or .Ad-eJF5AI(K50A) (Ad-elF5AIM) induces apoptosis in HT-29 colorectal adenocarcinoma cells. This figure provides the 25 percent apoptosis of HT-29 cells, as determined by Annexin V labeling and flow cytometry analysis, twenty-four, forty-eight, and seventy -two hours after infection vi th adenovirus constructs or after treatment with the apoptosis-inducing agents, Actinomycin D or Brefeldin A, Values are means + SE for n= 2 (H of events > 5000). 30 Figure II shows that infection with Ad-eTFSAI or Ad-eIF5A2 induces apoptosis in HFTB 9 bladder carcinoma cells. This figure provides percent apoptosis of HTB-9 cells, as determined by Annexin V labeling and flow cytometry analysis, twenty-four, forty-eight, and seventy-two hours after infection with adenovirus constructs or after treatment with 9 the apoptosis-inducing agents, Actinomycin D or Brefeldin A. Values are means i SE for n = 2 (# of events > 5000). Figure 12 shows that infection with Ad-elF5Al or Ad-eIF5A2 induces apoptosis in HTB 5 4 bladder carcinoma cells. This figure provides percent apoptosis of HTB-4 cells, as derin.eed by Annexin V labeling and flow cytometry analysis, twenty-four, forty-eight, and seventy-two hours after infection with adenovirus constructs or after treatment with the apoptosis-inducing agents, Actinomycin D or Brefeldin A. Values are means + SE fr n =2 (# of events > 5000). to Figure 13 shows that infection with Ad-elF5A, Ad-eiF$AI(K50A) (Ad-eIF5AIM}, or Ad-eF5A2 inhibits growth of HTB-4 bladder carcinoma cells. This figure provides the results of an XTT cell proliferation assay of HTB-4 cells twenty-fur, forty-eight, and sevenwt-two after infection with adenovirus constructs. Values are means t SE for n = 2, 15 Figure 14 shows that infection with Ad-eIF5A, Ad-elF5A1(K50A) I Ad-elFSA I M}, or Ad-elF5A2 inhibits growth of ITB-9 bladder carcinoma cells. This figure provides the results of an XIT cell proliferation assay of HTB-9 cells twenty-four, forty-eight, and seventy-two after infection with adenovirus constructs. Values are means i SE for n = 2. 20 Figure 15 shows that infection with Ad-eIF5Al, Ad-elF5A1(K50A) {Ad-eIF5AIM), or Ad-eIF5.A2 inhibits growth of HTB-1 bladder carcinoma cells. This figure provides the results of an XTT cell proliferation assay of HTB-1 cells twenty-four, forty-eight, and seventy-two after infection with adenovirus constructs. Values are means +SE For n =2. 25 Figure 16 shows that infection with Ad-elF5Al, Ad-eIF5A(K50A) {Ad-elF5AIM), or Ad-eIFSA2 inhibits growth of lUMUC-3 bladder carcinoma cells. This figure provides the results of an XTT cell proliferation assay of UMUC-3 cells twenty-four, forty-eight, and seventy-two after infection wi th adenovirus constructs. Values are means + SE for n = 2. 30 Figure 17 shows that infection with Ad-elF5Ai, Ad-elF5At(K50A) {Ad-eIF5A I M}, or Ad-eWFA2 inhibits growth of HeLa cervical adenocarcinoma cells. This figure provides the results of an XT cell proliferation assay of HeLa cells twenty-four, forty-eight, and seventy-two after infection with adenovirus constructs. Values are means i SE for n =2. t M Figure 18 shows that infection with Ad-e[[5AP, Ad-elFSAI(K50A) (Ad-eFSA I M), or Ad-elF$A2 induces PARP cleavage in HTB-4 cells, This figure provides the results of a Westem blot of cel lysate isolated from HTB-4 cells forty-eight or seventy-two hours 5 after infection with adenovirus constructs using PARP, elFT5A and f-actin antibodies. Figure 19 shows that infection with Ad-eFSAI, Ad-eiF5Ai(K50A) (Ad-eIF5A1M), or Ad-eIFSA2 does not induce PARP cleavage in HTB-1 cells. This figure provides the results of a Western. blot of cell lysate isolated from HTB-1 cells forty-eight or seventy 10 two hours after infection with adenmovirus constructs using PARP, eIF5A and 0-actin antibodies. Figure 20 shows that over-expression of elFSAl in A549 lung carcinoma cells activates MAPK/SAPK signaling pathways. A549 cells were infected with Ad-eiF5A I (5A) at 15 increasing inultiplicities of infection (MOI). Cells were infected with Ad-LacZ (Lac) for comparison, Forty-eight hours after infection, the cells were treated with EOF for 30 minutes and the cell lysate was harvested for Western blot analysis. Figure 21 shows that over-expression of eiF5Al in A549 lung carcinoma cells activates 20 MAPKISAPK signaling pathways. A549 cells were infected with Ad-e[F5AI (5A) or with Ad-LacZ (L) at 50 MOL Cells were treated with EGF for 30 minutes either 6, 24, 48, or 72 hours after infection and the cell lysate was harvested for Western blot analysis. Figure 22 shows that over-expression of a mutant eIF5A I that is incapable of being 25 hypusinated activates MAPK/SAPK signaling pathways, A549 cells were infected with Ad-eiF5A.1 (K50A) (M) at increasing multiplicities of infection (MOI). Cells were infected with Ad-eIF.A (5A) and Ad-LacZ (Lac) for comparison. Cells were incubated with or without the MEK inhibitor U31026. Forty-eight hours after infection, the cells were treated w ith EGF for 30 minutes and the cell lysate was harvested for Western blot 30 analysis. Figure 23 shows that over-expression of eFSAJ or mutant elF5Ai(K.50A) in. A549 lung carcinoma cells increases expression of p53. A549 cells were infected with Ad-elFSAI. I I (5A) or Ad-eIF5Al(K50A) (M) at increasing mudtiplicities of infection (MOI). Cells were infected with Ad-LacZ (Lac) for comparison. Cells were incubated Xi th or without the MEK inhibitor U1026. Forty-eight hours after infection, the cells were treated with EGF for 30 minutes and the cell lysate was harvested for Western blot analysis. 5 Figure 24 shows that over-expression of elF5Al or mutant eiF5A1(K50A) reduces the capacity of A549 lung carcinoma cells to invade through MatrigelR"I extracellular matrix. Twenty-four hours after infection with Adenovirus constructs, A549 cells were seeded ontO MatrigeO
M
-coated transwells in serum-free media. Serum-containing media was 10 placed in the bottom well to act as a chemoattractant. Twenty-four hours after plating. cells which had invaded through to the bottom surface of the transwell were stained with crystal violet and the average number of cells per field was determined by counting cells under a light microscope. A minimn of six fields per sample were counted. 15 Figure 25 shows the results of experiment 11: Lung tumor load was significantly reduced by eIF5Ai in B16F1 0-bearing C57BL/6 mice. 200,000 B16FTO cells were injected into the tail vein of 6-week old C57.BL/6 mice, Plasmid DNA bearing either the LacZ gene (as a DNA control) or efF5AI was injected into the tail vein on days 2.4, 7, 11, 16,21, 26, and 31. Three different concentrations of DNA were used: I X (3.3 mg/kg), 0.1 X (0.3 20 mg/kg), mid 2X (6.6 mg/kg). Mice which received PBS instead of B16F1O cells were used as a negative control while 816F10-bearing mice that received injections of PBS rather than plasmid DNA were used as a positive control. When the mice became moribund, they were sacrificed and the lungs were removed and photographed 25 Figure 26 shows the result of experiment. : Lung weight was significantly reduced by eIF5AL in B16F10-bearing C57BL/6 mice. 200,000 B16FO cells were injected into the tail vein of 6-week old C57BL/6 mice. Plasmid DNA bearing either the LacZ gene (as a DNA control) or eIF5AI was injected into the tail vein on days 2, 4. 7, 11, 16, 21, 26, and 31. Three different concentrations of DNA were used: IX (3.3 ng/kg) 0, IX (0.3 mg/kg), 30 and 2X (6,6 mg/kg). Mice that received PBS instead of BI6FlO cells were used as a negative control while B16F10-bearing mice that received injections of PBS rather than plasmid DNA were used as a positive control. When the mice became moribund, fthe were sacrificed and the lungs were removed and weighed. 12 Figure 27 shows the results of experiment 11: VEGF expression was significantly reduced by ePF5A in B1610-bearing C57BL/6 mice. 200,000 B16F10 cells were injected into the tail vein of 6-week old C57BL/6 mice. Plasmid DNA bearing either the LacZ gene (as a DNA control) or eF5AI was injected into the tail vein on days 2. 4, 7, I1, 16. 21, 26. 5 and 31 Three different concentrations of DNA were used: IX (3.3 mg/kg), 0.1X (0,3 mg/kg), and 2X (6.6 mg/kg). Mice that received PBS instead of BI6F 10 cells were used as a negative control while B161 0-bearing mice that received injections of PBS rallier than plasmid DNA were used as a positive control. When the mice became moribund, they were sacrificed and the lngs were removed and weighed.. Once lung weight was 10 determined, the lungs were frozen and later used for VE F ELISAs. Lung tissue was ground in Iysis buffer and the amount of VEGF present in the lysate was detennined by ELISA. Figure 28 shows the results of experiment III: Anti-tumor efficacy of eIFSA] gene therapy 15 was improved by coniplexing DNA with DOTAP, 50,000 B16F110 cells were injected into the tail vein of 6-week old C57BL/6 mice (Day 0). Plasmid DNA bearing either the LacZ gene (as a DNA control) or elFSA I complexed with DOTAP prior to injection into the tail vein on days 7, 14, and 21. Tumor-free mice that received injections of DOTAP without plasmid DNA were used as a control for the effects of DOTAP. Mice were sacrificed at 20 Day 25 and the lungs were removed aid photographed. Figure 29 shows the results of experiment III: Anti-tumor efficacy of elFSA1 gene therapy was improved by complexing DNA with DOTAP. 50,00 B16F1.0 cells were injected into the tail vein of 6-week old C57.BL6 mice (Day 0). Plasmid DNA bearing either the LacZ 25 gene (as a DNA control) or eJF5AI complexed with DOTAP prior to injection into the tail vein on days 7, 14, and 21. Tumor-free mice that received injections of DOTAP without plasmid DNA were used as a control for the effects of DOTAP. Mice were sacrificed at Day 25 and die hngs were removed and weighed. 30 Figure 30 shows the results of experiment IV: Ad-elF5Al injection induces apoptosis in 1316F0 and B161710 tumors. 500,0001316F0 or B16F10 cells were subcutaneously injected into the right flank of C57:BL/6 mice. I 110 pFu Ad-5A in 50 p1 of PBS was injected into tumors when the tumors reached about 4mm in diameter (10-12 days after 13 B16 cell injection). Mice were sacrificed after 48 hours and tumors were excised, Exed, and embedded in paraffin. Two sections for each cell tumor type (Ad-5Ai1 and Ad-5Al 2) were stained by TUNEL (Prom ega) according to the manufacturers protocol. Negative control slides (Ad-SAT-neg) in which the TdT enzynme was left out of the TUNEL reaction 5 were included for each cell type. Figure 31 shows the results of experiment V: Tumor growth was signifcantly retarded by injection with Ad-eIF5Al. C57BL/6 bearing subcutaneous B16-F)0 were injected with either 1 .X 109 pft of Ad-elF5A or Ad-LacZ or an equivalent volume of buffer. The first 10 day of treatment occurred when tumors reached an approximate diameter of 8 mm and was designated Day 0. Mice were injected every day for the first three days and then every other day thereafter until the mouse was sacrifced. Mice were sacrificed when the tumor size exceeded 15 to 16 nun in one dimension. There were three mice in each group. Group I mice (I-L. 1-2, 1-3) were treated with Ad-elF5Al, Group 2 mice (2-1, 2-2, 2-3) 15 were injected with Ad-LacZ and Group 3 (3-1, 3-2, 3-3) mice received only buffer. Tumor size was measured every day using calipers and used to estimate tumor volume. Figure 32 shows the results of experiment V: Survival was signifi cantly increased by iniection with Ad-elFSAL. C57BL/6 beating subcutaneous B16-F10 were injected with 20 either 1 X 1o9 pfu of Ad-eF5A1 or Ad-LacZ or an equivalent volume of buffer. The first day of treatment occurred when. tumors reached an approximate diameter of S mim and was designated Day 0. Mice were injected every day for the first three days and then eveiy other day thereafter until the mouse was sacrificed. Mice were sacrificed when Pie tumor size exceeded .15 to 16 mm in one dimension. There were three mice in each group. 25 Group I mice (1-1, 1-2, 1-3) were treated with Ad-eIF5AI; Group 2 mice (2-1, 2-2, 2-3) were injected with Ad-LacZ; and Group 3 (3-1. 3-2, 3-3) mice received only buffer. Figure 33 shows that elf-5Al increases the accumulation and phosphorylation of p53 tumor suppressor protein in A549 lung carcinoma cells. 30 Figure 34 shows that elF-SAl increases p 5 3 mRNA levels in A549 lung carcinoma cells. Figure 35 shows that the increase in p53 levels is dependent upon p53 transcriptional activity in A549 lung carcinoma cells. 14 Figure 36 shows that TNFR1 mRNA levels are upregulated by infection with elF-5AT in A549 lung carcinoma cells. 5 Figure 37 shows that elF-5Al induced apoptosis in A549 lung carcinoma cells and the use of a MEK inhibitor increases the amoun t of apoptosis induced by eIF-5A1. Figure 38 shows a hypothetical model of elF-5Al s effect on MAPK/SAPK pathways and apoptosis in. A549 Iung carcinoma cells. 10 Figure 39 shows that elF-5A1 is better than elF-5A2 in suppression of tumor growth (in mouse xenograft model (B1I6-FO melanoma cells)). Figure 40 shows that elF-5A l is better than eLF-5A2 in prolonging mouse survival (in 15 mouse xenograft model (B 16-FO melanoma cells)). Figure 41 shows that elf-5A increases the number of dead or dying cells as compared to cells being treated with the control with or without IL-6. 20 Figure 42 provides the sequence of elF-SA2, Figure 43 provides the location and sequences of elF-5Al siRNAs. Figure 44 provides a nucleotide alignment of huIman elF-5Al with human elF-5A2. 25 Figure 45 provides an amino acid alignment of huanan elF-5Al with human elF-5A2. Figure 46 A and 46 B provide the location and sequences of elF-5Al siRN As. 30 DETAiLED DESCRIPTION OF THE INVENTION Eukaryotic translation initiation factor 5AL (elFSAI) has been hypothesized to function as a nucleocytoplasmic shuttle protein involved in facilitating translation of subsets of mRNAs involved in cell proliferation, However, eF5AI has also been identified as a regulator of apoptosis (Taylor et a., Invest Ophthahnol Vis 15 Sci.;45(10):3568-76 (2004)) and a pro-apoptotic protein capable of regulating expression of p53 (Li at a], (2004) 1 Mol Chem.: 279:49251-49258: and Figure 23 of this document t). The tumor suppressor protein p53 plays a central role in mediating cell cycle arrest and apoptosis in response to stress and DNA damage. Although over-expression of elF5AI is 5 capable of up-regulating p53 expression in cancer cell lines (Li el al (2004) .1 Bol Chem,; 279:49251-49258; and Figure 23 of this document), over-expression of eIF5AL is also capable of inducing apoptosis in p53-deficient cell lines (Taylor et at (2006) Journal of Molecular and Cellular Biology, Feb 9 2006 (decision pending)) indicating that eIF5AI over-expression induces apoptosis by multiple mechanisms. 10 Expression of eJF5AI is con-elated to apoptosis. Western blots of normal colon fibroblasts treated wilh the topoisoimerase inhibitor Actinomycin D (Figure I A) demonstrate that eF5A. protein expression is up-regulated in parallel with p53 twenty four hours after initiation of Actinomyvcin D treatment. RKO cells undergoing apoptosis resulting from exposure to the nitric oxide (NO)-donor, SNP, also upregulate elF5AI 15 protein expression (Figure 1 B). RNA transcript levels for elFSA I do not increase under these conditions indicating that elF5A I protein is accumulating as a result of post transcriptional regulation of miRNA (Figure l.A and 1 B). 'These results indicate that eF5AI may be involved in apoptosis resulting from genotoxic stress or nitric oxide. A requirement for elF5A1 in cell proliferation has long been proposed. partly due 20 to reports of DHS and D111 inhibitors inducing cell cycle arrest and apoptosis and the conclusion that hypusinated elF5 Al must be required to sustain cell growth. However, it was determined that specific suppression of eIF5AI expression through the use of siRNAs had no effect on cell growth. Suppression of etEF5A1 expression by >90% in IIT-29 cells had no effect on proliferation of the cells over a period of 5 days (Figure 2A-E). The DHS 25 inhibitor GC7, however, had a profound effect on the growth of these cells (Figure 2C), These results suggest that. eJF5AI may not be required for cell growth. Also of interest is the fact that suppression of eF5AI was able to partially protect HT-29 cells from Actinomycin D-induced cytotoxicity (Figure 2.B), further supporting eF5Al.'s involvement in apoptosis resulting from genotoxic stress. 30 The abil ity of eIF5AI siRNAs to protect cells from apoptosis prompted the examination of the effect e]FSAI protein suppression had on p53 expression. RKO cells have a functional p53 protein that does not accumulate except under conditions of stress, Suppression of elF5Al by siRNA was able to inhibit the accumulation of p53 protein by 69 %4 after twenty fours of Actinomycin D treatment (Figure 3A and 3$), indicating that 16 efF5AI is required for proper expression of p53 during genotoxic stress. A second siRNA against eLF5A I having a different sequence was also able to prevent p53 upregulation in response to Actinomycin D (Figure 8B) indicating that this effect is not a non-specific effect of the siRNA, However, eIF5AI was capable of inducing apoptosis in both .RKO 5 cells (with functional p53) and RKO-E6 cells (without functional p53; Figure8A) (Figure 4A and 4B), indicating that elF5A1 can induce apoptosis by p53-independent mechanisms as wxelL Adenovirus constructs expressing either el5AL or eIF5A I containing a point mutation in the conserved lysine (K)(position 50) that is required for the hypusine 10 modification {eF5Al(K50A)} were constructed. The point mutation caused the lysine to be an alanine (A). Infection of HT-29 cells with either construct induced apoptosis in these cells (Figure SC and 5D and Figure 6) and greatly decreased cell viability (Figure SE). Two dimensional gel electrophoresis was used to determine which form of .F5Al was accumulating as a result of infection with Ad-eJF5Al or Ad-elF5A1(K50A) (Figure 5B). 15 As expected, unmodified e15Al accumulated after Ad-elF5AI(K50A) infection. Infection with Ad-el F5AI resulted in a dramatic increase in the accumulation of both unmodilied and deoxyhypusine-modi fed eFISAI, indicating that DHS and D111 activities were insufficient to hypusinate most of the eIF5AT protein being generated by the virus (Figure 5B). These results strongly suggest that the unmodified eIF5AI and perhaps the 20 deoxyhypusine-modilied elF5A I are the forms leading to apoptosis of the cells. It is not yet clear whether hypusinated elF5Al shares this ability. A nucleocytoplasmic shuttling function has been proposed for elF5AI, however, elF5Al has been reported to be expressed predominantly in the cytoplasm (Shi et al, Exp ('el Res; 225:348-356 (1996b)). A nucleocytoplasniic shuttle protein would be expected 25 to have a nuclear localization as well. Since a function for elF5Al during apoptosis was found, the localization of eJF5Al changes during apoptosis were studied. Apoptosis was induced in H T-29 cells by two di fTerent mechanisms, death receptor activation via treatment with TN- y and TNF-a and genotoxic stress via treatment with Acti.nomy cin D. Indirect immunofluorescence revealed that el FSA1 localization was cytoplasmic in 30 untreated, growing cells (Figure 7). H however, treatment with apoptotic stimulators vex quickly stimulated the movement of eAl from a predominantly cytoplasmic localization to a predominantly nuclear localization (Figure 7A and 7B). This shift in eIFSAI localization occurred veryv quickly, within 10 minutes for lFNyTTNF-c-treated 17 cells (Figure 7A) and within 90 minutes for cells treated with Actinomycin D (Figure 7B). These results indicate that elFSAI has a nuclear function during apoptosis induced by both death receptor activation and genotoxic stress. in order to clarify whether it is the unmodified, deoxyhypusine-modified, or 5 kypusine-modified form of eF5AI that is involved in apoptosis, the fions of eIF5At accumulating during apoptosis was examined by 2-D gel electrophoresis. HT-29 cells were induced to undergo apoptosis by simulation with either Actinomycin D (genoloxic stress) or incubation with an agonistic antibody against Fas (death receptor pathway). Coll lysates were collected after various ime points raging from 1 hour to 24 hours and then 10 analyzed by 2-D gel electrophoresis and Western blotting with eIF5A antibody (Figure 9). In untreated cells, the eIFSAI protein was predominantly in the hypusinated form in agreement with what is reported in the literature. Both apoptotic inducers stimulated an increase in the presence of the deoxybypusinated form of eJFAI beginning I hour after treatment but disappearing 24 hours after treatment. An accumulation of unmodified 15 elEFAl was also observed 2 hours afler treatment. These results are consistent with the deoxyhypusinated and/or unmodified eIF5AI being involved in regulating apoptosis. The ability of elF5AI, e[F5AI.(K50A)(nutant elF5A1), and. eF5A2 to induce apoptosis and to inhibit proliferation was examined in a variety of cancer cell lines. Both eIFSAI and etF5Ai(K50A) were able to induce apoptosis in the colon carcinoma cell line 20 IHT-29 (Figure 5, 6, and 10) and the bladder cancer cell lines HTB-9 (Figure I1) and 13TB 4 (Figure 12). In addition, infection with an adenovirus expressing erF5A2 can induce apoptosis in the bladder cancer cell lines HTB-9 (Figure 11) and HTB-4 (Figure 12) Ad eIEFAT, Ad-eIF5Ai(K50A), and Ad-eIF5A2 were all able to inhibit growth of the bladder cancer cell lines HT3-4 (Figurel13). HTB8-9 (Figure 14), 382 HTB-1 (Figure 15), and 25 UMUC-3 (Figure 16), These viral constructs were also able to inhibit growth of HeLa cervical adenocarcinoma cells (Figure 17). Apoptosis in HTB-4 cells was also observed by PARP cleavage in response to Ad-eIF5At, Ad-eIF5AI(K50A), and Ad-eWFSA2 infection (Figure 18). PARP, which is normally in volved in DNA repair, DNA stability and other cellular events, is cleaved by members of the caspase family during early 30 apoptosis. Detection of caspase cleavage of PARP has been shown to be hallmark of apoptosis. Lazebnik Y, et al. Nature 371, 346-347 (1994). These results indicate that eIF5A 1 and eIF5A2 may in fact have redundant functions during apoptosis. Furthermore. the mutant form of eJF5A I was able to arrest growth and induce apoptosis a. variety of cell lines. The ability of the wild-type eiF5A1 (and eIF5A2) to induce apoptosis could be is related to the accumulation of deoxyhypusinated and unmodified eIF5AI that occurs (Figure5B) when Ad-elF5A is introduced into a cell due to limiting activities of DHS and DH.H. A recent report has shown that in order for exogenous elF5AI to be over-expressed in a cell in the hypusinated form, both DHS and DHH have to be over-expressed in the 5 same cell (Park et aL 2006, PAS; 103(1): 51-56). These results suggest that the ability of DHS and DHH inhibitors to inhibit cell growth. may not be due a. decrease in hypusinated eIF5AI required for cell proliferation, but instead may be due to an accumlation of unmodified (DHS inhibitors) or deoxyhypusine-modified (DIH i inhibitors) eIF5AI, which in turn trigger cell cycle arrest and/or apoptosis in the cell. It is also interesting to note 10 that DHS has been found to be over-expressed in certain cell lines (Clement et aL 2006, FEBS; 273(6):1102-14) and that DHS has been identified as one of a signature set of amplified genes in cancer metastases (Ramaswan er at 2003, Nat Genet; 33, 49-A4). One possible interpretation of this information is that certain cancer cells over-express DHS in order to prevent apoptosis triggered by accumulation of unmodified or 15 deoxyhiypusinated eF5Al (which appear to accumulate when cells are triggered to undergo apoptosis, see Figure 9). By over-expressing DHS, cancer cells may be able to reduce apoptosis caused by genotoxie stress and other stressors by keeping eIF5AI in a hypusinated and thereby. "safe" form. In order to elucidate which signaling pathways are affected by over-expression of 20 eIFSAI, activation of the mitogen activated protein kinase (MAPK)/stress activated protein kinase (S APK) MAPRK/S APK) pathways was examined in response to Ad-elF5AI or Ad-eIF5AI(K50A) infection in A549 lung carcinoma cells. The three major MAPK pathways are the ERK MAPK pathway., the p38 MAPK pathway, and the JNK SAPK pathway. The ERK .M.APK pathway is mainly triggered in response to mitogenic stimuli 25 such as growth hormones like epidermal growth factor (EGF) and supports the growth and survival of a broad range of tumors. The p38 MAPKI pathway is activated in response to cellular stresses, UV light, growth factor withdrawal, and pro-inflammatory cytokines. Activation of p38 via phosphonation leads to the phosphotylation of transcription factors such as p53, which can in turn lead to increased activity or stability of p53. Activation of 30 p38 is involved in both pro-apoptotic and anti-apoptotic pathways as well as inflammation. The JNKSAPK.pathwav mediates responses to cellular stresses including UV light, DNA damage and pro-inflammatory cytokines and results in the phosphorylation and increased activity of transcription factors such as c-jun. Activation of the JNK pathvay can lead to numerous cellular responses including growth, transformation and apoptosis. The JNK 19 pathway appears to be a prime effector pathway for EGF-induced growth in A549 cells (Bost el al 1997, JBC: 272:33422-33429). Infection of A549 cells with Ad-eff5Al induces the activation of all three of these pathways. Infection with increasing amounts of Ad-elFSAl in A549 cells stimulated with EGF for 30 minutes. resulted in increasing 5 phosphorylation/activation of ERK and its downstream target, p9ORSK, while amounts of unphosphorylated ERK remained unchanged (Figure 20). A strong upregulation of phosphotylatedlactivated p38 and phosphorylated/activated JNK was also observed in response to Ad-eIF5AI infection in a dose responsive manner (Figure 20). A time course of Ad-elFAl infection revealed that activation of ERK, p38, and JNK pathways was 10 sustained from 24 to 72 hours after infection (Figure 21). A dose-responsive increase in activation of these pathways in response to infection with the unhypusinated mutant eIFAl, Ad-eF5AI(K50A) (Figure 22) was found. It is of interest to note that inhibiting the activation of ERK using a MEKI inhibitor, U1026, also inhibited the activation of JNK (Figure 22), confirming that in these cells, JNK is activated in response to ERK 15 activation (Bost et al. 1997.C; 272:33422-33429). In contrast, the MEKI inhibitor has consistently resulted in increased p38 activation in response to either Ad-eWFAI or Ad elFSAI(K50A) infection indicating that activation of the ERK pathway negatively regulates activation of the p38 pathway in response to elF5Al (Figure 22) Figure 38 provides a hypothetical model of elF-5Al's effect on MAPKJSAPK 20 pathways and apoptosis in A549 I wig carcinoma cells. Over-expression of elF5A1 (either wild-type or a mutant incapable of being hypusinated)(K50A) induces apoptosis in A549 cells. Over-expression of eF5Al is also accompanied by an increase in MAPK/SAPK pathways, including p38, JNK and ERK. Over-expression of elE5Al induces an increase in p53 protein levels, including increases in phosphorylated forms of p53 that are 25 associated with increased stability and activity of p53. This increase in p53 is dependent on the activity of the MEK/ERK pathway and on p53 transcriptional activity., However, inhibition of the MEK/ERK pathway, and to a lesser extent inhibition of the JNK pathway, results in a potentiation of the apoptosis induced by eFSA . These results indicate that eJF5AI may have use as a therapeutic in order to induce apoptosis in cancer cells as well 30 as to increase the cancer cell-killing abilities of the MEK inhibitor class of anti-cancer drugs. The effect of Ad-elF5Ai and Ad-elF5A(K50A) on the expression and phosphoiylation of p53 can be seen in Figure 23. Over-expression of both wild-type and mutant eTF5AI led to an increase in the overall expression of p53 protein, in a dose 20 responsive manner (Figure 23). An increase in p 5 3 phosphorylated on seine 15 was also observed for both forms of eIF5AI (Figure 23). An increase in p53 phosphorylated on seine 46 was observed in a dose-responsive manner for Ad-elF5A1(K50A), p53 can be phosphorylated on seine 15 by several kinases including ATM, ATR, and p38. and this 5 modification reduces the ability of MDM2 to bind p53 and target it for ubiquination and thereby promotes the accumulation and functional activity of p53. p53 is phosphorylated at seine 46 by various kinases including PKCdeita and p38 and this modification increases the aftinity of p53 for pro-apoptotic promoters and is believed to be important for p53-induced apoptosis. 10 Invasion and metastasis of tumors is a complex process that requires a tumor cell to adapt its ability to adhere, to degrade the surrounding extracellular matrix, to igrate and proliferate at a secondary site, and finally to promote angiogenesis to sustain increased growth. Basement membranes are contiguous sheets of extracellular matrix (ECM) that surround every organ and act as a barrier to macromolecules and cells. The invasiveness 15 of tumor cells can be measured by coating transwells with an 8 pm membrane with reconstituted ECM (Matrigel "") and staining cells that are able to penetrate this layer and reach the other side of the membrane in response to chemnotactic stimulation. A549 lung carcinoma cells are highly invasive and are able to secrete proteins such as matrix metaloproteases, which are gelatinases capable of digesting components of the ECM. The 20 effect of eIF5A1 over-expression on the invasiveness of A549 cells was examined infection of Ad-elF5Al or Ad-eF5AI (K50A) significantly decreased the number of cells that invaded through Matrigel'-coated transwells (Figure 24). The results presented thus far suggested the possibility that treatment of tumors with eJF5AI may be of therapeutic benefit. Thus, a model of experimental metastasis in 25 mice was used. In Experiment 11, experimental metastasis was initiated by injecting mice with the highly invasive mouse melanoma cell line .B16F10 (day 0). Plasmid DNA encoding either the LacZ gene (as a DNA control) or e15A1 was injected into the tail vein day on days 2, 4, 7, I1 16, 21. 26, and 31. Three different concentrations of DNA were used: IX (3.3 mg/kg). 0, X (0.3 mg/kg), and 2X (6.6 mg/kg). When the mice 30 became moribund, they were sacrie flod and the lungs were removed and photographed (Figure 25). A dose-responsive decrease in tumor burden is observed wh1ien mice are treated wit plasmid DNA encoding elFSAi (Figure 25 and 26). Although the 2X dose appeared to be less effective than the IX dose. there was a significant decrease in tumor 21 burden observed between the 0. 1X dose and the IX dose. The ability of tumors to grow macroscopically is dependent on the formation of new blood vessels (angiogenesis). Vascular endothelial growth factor (VEGF) is a cytokine that is secreted by many tumor cells and is a key factor in promoting tumor angiogenesis. The cell lysate from B16F10 5 bearing lungs isolated in Experiment 1I were analyzed for VEGF expression by ELISA (Figure 27). There was a significant, dose-responsive decrease in VEGF levels in mice that received eIFSA plasmid DNA, indicating that eIF5Ai may regulate VEGF. In Experiment 111, a model of experimental metastasis using tail vein injected B16F1l0 cells was again used, but this time the plasmid DNA was complexed with 10 DOTAP in order to increase the half-life of the DNA in serum and increase uptake of the plasmid into the lung tumors. Injections of DNAIDOTAP complexes occurred on Days 7, 14, and 21 Mice were sacrificed when they became moribund and the tungs were removed and photographed (Figure 28). There was an average 60% reduction in Iung~ weight in mice treated with the eIF5AI plasmid DNA/DGTAP complexes (Fig. 29) when 15 compared to mice that received the LacZ plasmid DNA/DOTAJP complexes. In order to detennine whether eLF5AL treatment induces apoptosis in melanoma tumors. mice bearing either B16FO or B1 610 subcutaneous tumors were injected intra~ tumorally with Ad-eIF5Al (Experiment IV). Forty-eight hours later the tumors were excised, paraffin embedded and sectioned. TUNEL staining of the sectioned tissues 20 revealed that Ad-elF5Al induced apoptosis of turnor cells (Figure 30). Experiment V was designed to determine whether the ability of intra-tumorally-injected Ad-eIFSAt to induce apoptosis would result in reduced tumor growth of subcutaneous BI 6F1 tumors and increase survival. B16F10 cells were injected subcutaneously into the flank of C57BL/6 mice. When the tumors had reached an approximate diameter of 8 mm. they were injected 25 intra-tumorally with either Ad-LacZ, Ad-e&F5A1, or buffer. Injections were repeated daily for a total of three days and then every other day thereafter until the tumor exceeded 15 or 16 mm in diameter, at which point the mouse was sacrificed. Tumor dimensions were measure daily using calipers and used to estimate tumor volume. A delay in tumor growth was clearly observed when mice were treated with Ad-eIF5AI (Figure 31). Mice 30 that received buffer-only injections lived a maximum of 4 to 6 days after the initiation of treatment while mice that received Ad-LacZ injections only survived 4 days aler beginning treatment (Figure 32). Mice that received Ad-e1F5AI injections lived at least 8 days after treatment and as much as 25 days after treatment demonstrating that treatment 22 with Ad-eIF5Ai can result in dramatic improvement in survival of tumor-bearing mice (Figure 32). Accordingly, the present invention provides a method of inhibiting cancer growth. The present invention also provides a method of inhibiting or slowing down the ability of 5 a cancer cell to metastasize. iNhibiting cancer growth includes a reduction in the size of a tumor, a decrease in the growth of the tumor, and can also encompass a complete remission of the tumor, Inhibiting cancer growth also means killing cancer cells. The cancer can be any cancer or tunor, including but not limited to colon cancer, colorectal adenocarcinoma, bladder carcinoma, cervical adenocarcinoma, and lung carcinoma. 10 The methods of the present invention involve the administration of a polynucleotide encoding elF-5A. preferably human elF-5Al to a patient (a mammal, preferably a human), preferably elF-SA I accession number NM 001970 (See figure 44) having said cancer. The eF~5A2 isoform may also be used (i.e. accession number NM 020390, although elF-5AT is preferred. The elf-SA may be delivered by any suitable 15 method know in the at. It may be delivered as naked DNA, such as DNA in biologically suitable medium and delivered through IV or subcutaneous injection or any other biologically suitable delivery mechanism. Atematively, the eTF-5A may be delivered in a plasmid, vector such as an adenovirus vector or any suitable expression vector, Alternatively, the DNA may be delivered in liposomes or any other suitable 20 'carrier" or "vehicle" that provides for delivery of the DNA (or plasmid or expression vector) to the target tumor or cancer cells. See for example, Luo, Dan, et al., Nature Bioechnologv, Vol, 18, Jan. 2000, pp. 33-37 for a review of synthetic DNA delivery systems. Although the present inventors have earlier shown that elF-5AL is non toxic to normal tissue (see pending application 1 1/293,391, filed November 28, 2005, which is 25 incorporated herein by reference in its entirely), a delivery system (as compared to direct administration of the eiF5A polynucleotides/plasid/expression vector) is preferred. A preferred delivery system provides an effective amount of elF-5A to the tumor or group of cancer cell, as well as preferably provides a targeted delivery to the tumor or group of cancer cells. Thus, it is preferable to deliver the eF-5A nucleotidesipiasmidlexpressi on 30 vector via a vehicle of nanometer size such as liposomes, dendrimers or a similar non toxic nano-particle. Further the vehicle preferably protects Ie elF-5A nucleotides/plasmid/expression vector from premature clearance or from causing an imnmue response while delivering an effective amount of the elF-5A nucleotides/plasiidlexpression vector to the tumor or group of cells. Exemplary vehicles 23 may range from a simple nano-particle associated with the elF-A nucleotides/plasmid/expression vector to a more complex pegylated vehicle such as a pegy lated liposome having a ligand attached to its surface to target a specific cell receptor. Liposomes and pegylated liposomes are known in the art. In conventional 5 liposomes, the molecules to be deli veTed (i.e. small drugs, proteins, nucleotides or plasmids) are contained within the central cavivy of the liposome. One skilled in the art would appreciate that there are also "stealth", targeted, and cationic liposomes useful for molecule delivery. See for example, Hortobagyi. Gabriel N et al., J. Clinical Oncology. Vol. 19, issue 14 (July) 2001:3422-3433 and Yu, Wei, et al, Nucleic Acidv Research. 10 2004, 32(5);e48. Liposomes can be injected intravenously and can be modified to render their surface more hydrophilic (by adding polyethylene glycol ("pegylated") to the bilayer, which increases their circulation time in the bloodstream. These are known as "stealth" liposomes and are especially useful as carriers for hy drophilic (water soluble) anticancer drugs such as doxorubicin and mitoxantrone. To further the specific binding properties of 15 a drug caring liposome to a target cell. such as a tumor cell, specific molecules such as antibodies, proteins, peptides, etc, may be attached on the liposome surface. For example, antibodies to receptors present on cancer cells may be used to target the liposome to the cancer cell. In the case of targeting multiple myeloma, folate, 11-6 or transferrin for example. may be used to target the liposomes to multiple mnyeloma cells. 20 Dendrimers are also known in the art and provide a preferable delivery vehicle. See for example Marjoros, istvan, J., et al, "PAMAM Dendriner-Based Mutifunctional Conjugate for Cancer Therapy: Synthesis, Characterization, and Functionality Biomacromolecules, Vol. 7, No. 2, 2006; 572-579, and Maijoros, Jstvan J, et al, J Med Chem, 2005. 48, 5892-5899 for a discussion of dendrimers. 25 The elf-5A may also be- delivered directly to the site of the tumor. One skilled in the art would be able to determine the dose and length of treatment. regimen for delivery of elF-5A. Another embodiment of the present invention provides a method of inducing cell death in multiple myeloma cells. Multiple myeloma is a type of bone marrow cancer that 30 produces high levels of inflammatory cytokines, which can lead to bone lesions and tumor progression. Cytokines IL-1 B and IL-6 act as growth factors for the mnyeloma cells. An adenovirus vector construct containing polynucleotides encoding elF-5Al (the full length coding region) was administered to a multiple myeloma cell line. KAS 61 cells. The KAS 6/1 cell line was created at the Mayo Clinic and reported in Westendorf, 24 J., et atl. Leukemia (1996) 10.. 866-876. The cell line was created directly from isolates from a patient with aggressive multiple myeloma. The present inventors have shows that when an adenovirus construct containing eTF-5A is administered to KAS 6/1 cells, there was an increase in the number of dying or dead cells (leaving fewer viable cancer 5 cells)(indicated as "WT" in figure 41) as compared to cells having been treated with a conriTol vector alone (indicated as "CTL'. in figUre I). See figure 41. Approximately 90% of the cancer cells treated with Factor 5A died, in comparison to approximately 25% of the untreated cells. Accordingly, one embodiment of the present invention provides a method of killing iveloma cells by providing polynucleotides encoding elF-5A 1 to up 10 regulate expression of eff-sAL and cause the multiple myeloma cells to die. In addition. IL-6 was also administered along with the control (indicated as "CTL + IL-6") and the elF-5A construct (indicated as "WT+ Il-6" in figure 41). IL-6 is a cvtokine that acts as growth factor for myeloma. cells. The results show that even when IL-6 was co-administered with the efF-5A construct, an increase in apoptosis was still 15 achieved. See figure 4 1 This shows that Factor 5A1 is not only able to kill myeloma cells, but also eliminate myeloma cells in the presence of 1L-6. This finding is of interest in that it has proven to be very difficult to induce apoptosis in myeloma cells in the presence of IL-6 with standard therapies such as dexamethasone, Further, an adenovirus construct with a mutant elF-5A (K50A) (unable to be 20 hypusinated due to changing the conserved lysine at position 50 to another amino acid) was also administered alone ("MUT") or with 1L1-6 ("MUT + [L-6"). The results show that the mutant efF-5AI was also able to increase apoptosis as compared to the control cells, even in the presence of IL-6. See figure 41. Accordingly, one embodiment of the present invention provides a method of killing myeloma cells by administering a mutant 25 elF-5Al, said mutant being unable to by hypusinated, 'The mutant elF-5Al causes an increase in expression of unhypusinated elf-5AL, which increases cell death in the myel oma cells. Since RL- acts as a growth factor for iyeloma cells. down regulating expression of IL-6 would also provide amethod of killing myeloma cells. The present inventors have 30 show that siRNA against elF-5Al (See figures 43 and 46A and 46B for siRNA constructs) not only down regulates the expression of endogenous elF-5AL but also down regulates the expression of various pro-inflammatory cytokines. Th us, one embodiment of the present invention provides a method of killing myeloma cells by administering siRNA against apoptosis-specific elF-5A to down regulate endogenous expression of IL-1B, 25 which in turn down regulates expression of IL-6, which in turn provides less IL-6 to act as a growth factor for nyelorna cells. .By down regulating the expression of IL-6, there is less IL-6 available, which is necessary for the continued growth and survival of myeloma cells, 5 In another embodiment, polynucleotides encoding elF-5As are administered to provide an increase of apoptosis in the myeloma cells in conj unction with the siRNA against elF-5A. The polynucleotides encoding elF-SAl are preferably administered in a vector, such as an adenovirus vector, such that the siRNA does not inhibit expression of the exogenous e[F-5A1. For instance, the siRNA targets the 3' UTR, but the 10 polynucleotides encoding exogenous eWf5AI preferably contain the entire open reading frame (ORF) and thus have no 3'UTR to be targeted by the siRNA. Suitable siRNA constructs have been previously described in co-pending application I I /287,460; 1I/ 34,445: 11/184,982; and 11/293,39.1 which are all herein incorporated by reference in their entireties. See also figures 43 and 46A and 46B. The elF-5Al is expressed in the 15 myelonia cells and causes cell death. In addition, the elF-5Al siRNA decreases expression of endogenous expression of elF-5A, which in turn decreases expression of IL 6, and in turm increases cell death in the neloma cells. in this method a mutant elF-5A as described above may also be used in the vector construct. The present invention also provides a combination therapy to kill multiple 20 myelo.ma cells. Compositions comprising polynucleotides encoding elFSA, preferably eJF5A.i may be administered in conjunction with standard therapies. The eIF5A compositions may be administered before, during or after conventional therapies. The elF5A may be administered in as a pharmaceutical composition or may be administered within a delivery vehicle as discussed above. 25 EXAMPLES EXAMPLE 1: In vitro Experiments Chemicals N l-guanyl ,7-diaminoheptane (GC7; Biosearch Technologies), an inhibitor of 30 DHS, was used at a concentration of 50 pM. Actin omycin D (Calbiochem) was used at 0.5 or 1,0 pg/mL Sodium nitroprusside and desferrioxamine were purchased from Sigma and used at a. concentration of 3 mM and 500 yiM, respectively. Brefeldin A was also acquired from Sigma and used at a concentration of 4 nM. 26 Cell Culture and Treatment The human colon adenocarcinoma cell line, HT-29, was used for cell proliferation and eFSA localization studies and was a kind gifl from Anita Antes (ULniversity of 5 Medicine and:Dentist of New Jersey). HT-29 cells were maintained in RPMI 1640 supplemented with 1 mM sodium pyruvate, .10 mM HEPES, and 10% fetal bovine serum (FBS), All other cell lines were obtained from the American Type Culture Collection. CCD1 12Co is a normal colon fibroblast cell line. RKO is a human colorectal carcinoma cell, line (CRL-2577) containing a wild-type p53. The RKO-E6 cell line (CRL-2578) was 10 derived from the RKO cell line. It contains a stably integrated human papilloma virus E6 oncogene and therefore lacks appreciable functional p53 tumor suppressor protein, RKO, RKO-E6, A549, and the cell hie CCD1 12(o,. were grown in Modified Eagle Minmnum Essential Medium with 2 mM Iiglutamine and Earle's Balanced Salt Solution adjusted to contain 1.5 g/L sodium bicarbonate, 0. 1 mMvi non-essential amino acids, I mI sodium 15 pyruv ate and supplemented with 10 % FBS. Cells were maintained at 37*C in a humidified environment containing 5 % (02. Cloning and Construction of Plasmids Human eJFSA was cloned by RT-PCR from total RNTA isolated from RKO cells 20 using the GenElute Mammal ian RNA miniprep kit (Sigma) according to the manufacturer's protocol for adherent cells. The primens used were: forward., 5' CGAGT'TGGAATCGAAGCCTC-3'; and reverse, 5'GGITCAGAGGA TCACTGCTG-3' The resulting 532 base pair product was subcloned into pGEM-T Easy (Promega) and sequenced. The resulting plasmid was used as a template for PCR using the primers: 25 forward. 5'GCCAAOCTTAATGGCAGATGATTTGO-3 and reverse, 5 CCTGAATTCCAGTTATTTTGCCATGG-3 and the PCR product was su bcloned into the finzdII and EcRI sites of pHM6 (heimaggl utinin {HA] tagged; Roche Molecular Biochemicals) to generate the pHM6-eiF5A vector. A C-terminal truncated construct of etF5A (pHM6-eF5AA37) was generated by PCR using the following primers: forward, 30 5-GCCAAGCTTAATGGCAGATGATGTT(-3'; and reverse, 5 GCCGAATrTCCCTC AGGCAGAGAAG-3'. The resulting PCR product was subcIoned into the pHM6 vector. The pHM6-LacZ vector (Roche Molecular 27 Biochemicals) was used to optimize transfection and as a control for the effects of transfection on apopiosis. Northern Blotting 5 RKO cells were grown to confluence on 6-well plates and treated for 0, 1, 4, or 8 hours with L 0 g/ml Actinomycin D, Total RNA was isolated from the cells using the GenElute Mammalian RNA miniprep kit (Sigma), and 5 pg of RNA was fractionated on a L2 % agarose/fonmaldehy de gel. The membrane was probed with a 2 P-labelled cDNA homologous to the 3-untranslated region (3-TUTR) of eFW5A according to established 10 methods. The eIF5A 34UTR cDNA that was used for Northern blotting was cloned by RT-PCR from RKO cells using the following primers: forward, 5 GAGGAATTCGCTGTTFCAATCAAGGC-33 and reverse, 5' TTiTAAGCTTTGTGTCGGGGAGAGAGC-3. The r-actin cDNA that was used as a loading control for Northern blotting was cloned by RT-PCR using the following primers: 15 forward. S'-GATGATATCGCCGCGCTCGT-3'; and reverse, 5 GCTAGATGGGCACAGTGTCGGTG-3. Transfection of Plasmids and Detection of Apoptosis RIO and RKO-E6 cells were transiently transfected with plasmid DNA using 20 Lipofectamine 2000 (lIwitrogen) according to the manufacturer's recommended protocol, Forty-eight hours after tran section, apoptotic cells con gaining fragmented DNA were detected by terminal deoxynucleotidyl transferase-iediated dUTP-digoxigenin nick end labeling (TjUNEL) using a DNA Fragmentation Detection Kit (Oncogene Research Products) according to the manufacturer's protocol. For fluorescence microscopy 25 analysis, cells were. transfected on 8-well culture slides, fixed with 4% formaldehyde and then labeled by TUNEL and stained with Hoescht 33258 according to the methods described by Taylor ei al (2004). Transfection of siRNA 30 All siRNAs were obtained from Dharmacon. The elFSA siRNA, which targets the 3'UTR of the eIfEA iRNA (Accession No. BC085015), had the following sequence: sense strand, Y-GCCUGACUCCUCCUACACAdTdT-3'; and antisense strand, 3 dTdTCGACCUGAGGAGAUOLGUU-5'. A second siRNA (5A-2) directed against 28 eLF5A had the following sequence: sense strand, S& AGGAALJGAC UUCCAGCUGAdTdT-3': and anti sense strand, 3 dTdTUCCUUACUGAAGOUCGACLT-St The control siRNA that was used had the reverse sequence of the elF5A-specific siRNA and had no identity to any known human 5 gene product, T he control siRNA had the following sequence: sense strand, ACACA UCOCUCCUJC AGGUIC~dTdT-3'; and anti sense strand, 3' dTdTJGUGAGGAGGAGUCCAGC-5. Cells were transfected with siRNA"' using Lipofectamine 2000 and used in proliferation studies or for Western blotting. 10 Western Blotting Protein for Western blotting was isolated using boiling lysis buffer [2% ,SDS 50 mM Tris-U CI (pH 7.4)[. Protein concentrations were determined using the Bicinchoninic Acid Kit (Sigma). For Western blotting, 5 pg of total protein was fractionated on a. 12 % SDS-poiy acrvlamide gel and transferred to a polyvinylidene difluoride membrane. The 15 primary antibodies used were anti.-eIfSA (BD1) Transduction Laboratories; mouse igG) and anti--actin (Oncogene; mouse IgM). both at a. dilution of 1:20,000 in 5 % milk The secondary antibodies were anti-mouse IgG conjugated to horseradish peroxidase (HRP; Sigma) and anti-mouse TgM-H RP (Oncogene). Antibody-proiein complexes were visualized using the enhanced chemiluinnescence method (ECL, Amersham Biosciences). 20 Following detection of eiF5A, the blots were stripped according to the protocol provided by the ECL Plus Western blotting detection system and reprobed with anti-$-actin antibody to confinn equal loading. Western blotting of lysate from A549 cells used for MAPK/SAPK pathway analysis was performed using lysate collected in MAPK lysis buffer (10 mM Tris-pH 7.4, 25 2 % SDS, 10% glycerol). Ten micrograms of lysate was separated on 10% SDS-PAGE gels and transferred to a PVDF membrane. The membrane was blocked for one hour in 5 % non-fat skim milk in PBS, washed with PBS and incubated with prinary antibody at 1:1000 in 5% BSA/PBS-T overnight at 4*C0 with shaking. The MAPKSAPK antibodies (P-p38, p38, P-JNK. JNK, P-ERK, ERK, p90RSK) were purchased from Cell Signaling. 30 The p53 antibodies used in the A549 study were also obtained from Cell Signaling and used in a similar fashion. 29 2-D Gel Electrophoresis For 2-D gel elecrrophoresis, HT-29 cell lysate was harvested in cold lysis buffer (7M4 Urea, 21 Thiourea, 30mM. Tris, 4% CHAPS, protease inhibitor cocktail), sonicated and cleared of debris by centrifugation. Protein concentration was determined using the S Bradford method. The first dimensional isolelectric focusing was performed with the Ettan IPOphor Isoelectric Focusing System (Amersham.Biosciences) according to (he manufacture's instruction. Inmobiline DryStrips (7 cm pH 4-7; Amersham Biosciences) were rehydrated in rehydration bufflr (8M Urea, 2% CHAPS, 0. 2%DTT, 0.5% pH 4-7 IPG buffer, 0.002% Bromophenol blue) along with cell lysate at room temperature for 12 10 hours, The isoelectric focusing was performed at 500 V for 30 min, 1000 V for 30 mM. and 5000 V for I hour and 40 min. The protein on the IPO strip gel was then separated by SDS-PAGE and transferred to a PVDF membrane (Amersham Biosciences). Western blotting wvas performed using eUF5A antibody (BD Biosciences). 15 Generation of Adenoviras Adenoviruses (Adenovirus 5 serotype, ElE3-deleted) expressing human eJFSA or euF5A bearing a single point mutation (K50-A50) [eIF5A(K50A)j that prevents hypusniation were constructed using the AdMaxIN Hi-IQ system (Microbix Biosystems fie, Toronto, Canada). The site-specific mutation was created in the eIF5A cDNA using 20 PCR. The eLF5A cDNAs were amplified by PCR using plasmid DNA as template and lighted into the Smal site of the adenovirus shuttle vector pDC516(io). The sequence of the PCR primers were: forward, 5'~CCCAAGCTTAATOGC AGA TGA'TTG-3; and reverse, 5'-CCTGA ATTCCAGT'TAfTTTTGCCATGG-3. The adenovirus genomic plasmid vector pBHGfrt(dcl)EI,3FLP and the shuttle vectors were propagated inK col 25 DH5ct and purified using Qiagen EndoFree Plasmid Mega Kit. 5 pg each of the adenovirs genomic plasmid pBHIGfrt(del)E 1.3FLP and shuttle vector, pDC516(io) elFSA or pDC51 6(io)-eIF5A(K5OA), were transfected using the CaCI 2 method recommended by Microbix Biosystems Inc. into 60-80% confluent 2934Q cells (Microbix Biosystems) in 60 mn culture plates. Plaques appeared after 7 to 10 days incubation at 30 37C, and the resulting adenoviral particles [Ad-elFSA and Ad-eIF5A(K50A)] were amplified in 2934Q cells. Pure, high titer adenovirus stocks were prepared by CsCI gradient ultracentrifugation according to the protocol provided by Microbix Biosystems Inc, An adenovirus vector expressing LacZ (Ad-LacZ; serotype 5; E,E3-deleted) was 30 purchased from Qbiogene (Cali fria, USA) and employed as a control and reporter in these experiments. The Ad-LacZ adenovirus was amplified and purified in the same manner as the Ad-eIFSA and Ad-eF5A(K0A) viruses. 5 Adenoviros Infection and Annexin V lIbeling HT~29, HTB-9, or ITB-4 cells were seeded at IX I0 cells per plate in 100 mm tissue culture plates and infected with adenovirus the following day at 3000 infectious units per cell in 5 ml of RPMI 1640 + 2% FBS. Additional media was added to the cells aerfour hours and te concentration of FBS brought to- 0%. Twent-four, fory-eight, 10 or seventy-two hours after infection the cells were detached by trypsinization, washed and stained with Annexin V-ITC according to the manufacturers' protocol (BD Biosciences). The cells were sorted by flow cytometry (Coulter Epics XL-MCL) with a 488 TnM argon laser source and filters for fluorescein detection and the data analyzed by WinMfDI 2.8. HT-29 cells were infected with 3000 infectious units per cell and experiments with 15 A549 cells were performed using 1500 infectious units per cell. Proliferation Assays IT-29 cells were transfected with siRNA on 96-well plates using Lipofectamine 2000 (Invitrogen). Metabolic activity of proliferating cells was measured with the XTT 20 Cell Proliferation Kit (Roche Applied Science). The BrdU Cell Proliferation Kit (Roche Applied Science) was used to measure DNA synthesis following (he manufacturer's protocol. For XTT assays performed following adenovirus infection, 5000 cells were plated per well in a 96-well plate. The next day cells were infected with Ad-lacZ, Ad eLF5A1 and Ad-eiF5Al (K50A) (Ad-eIF5AIM) respectively with untreated cells as 25 negative control and Actinomycin D treated cell as positive control. XIT substrate was added and A4751mn was measured with A690nm as the reference. Indirect Immn fluorescence HT-29 cells were cultured on poly-L-lysine-coated glass coverslips. Subconfluent 30 cells were incubated for 16 hours with 200 Lnits of interferon gamma (IFN-y; Roche Applied Science) followed by TNF-cz (1.00 ng/ml; Leinco Technologies) for times varying from 10 minutes to 8 hours. Alternatively, cells were treated with 1.0 gg/ml Actinomycin D for increasing lengths of time from 30 minutes to 16 hours. The treated cells were fixed 31 with 3 % fonaldehyde (methanol-free; Polysciences Inc.) for 20 minutes. washed twice for 5 minutes with PBS and once for 5 minutes with P.BS containing 100 mM glycine, and peneabilized with 0.2% Triton X-100 in PBS for 4 minutes. Cells were then labeled for inmuno fluorescence using a standard protocol. The primary antibody was anti-elF5A 5 (BD Transduction Laboratories; mouse gG) incubated at a dilution of 1:250 for 1 hour. The secondary antibody was anti-mouse igG-AlexaFluor 488 (Molecular Probes) used at a dilution of 1:200 for I hour. Following antibody labeling, the nuclei were stained with Hoescht 33258, and the labeled cells were observed by fluorescent microscopy. 10 Matrigef' Cell Invasion Assays A549 cells were infected with adenovirus at 1500 infectious units per cell and incubated for 24 hours. The cells were then detached with trypsin, washed with serumi free media, and plated in serum-free media at 30,000 cells per well on a transwell (Falcon 8.0 psm cell culture insert) that had been precoated with 15 gtg of Matrige M Basement 15 Membrane Matrix (BD Biosciences). Media containing 10 % FBS was placed in. the bottom well ([he well of the 24-well plate in which the transwell is resting) and the cells were incubated for a further 24 hours. After the incubation, the media was removed from the upper chamber of the transwell, the transwell was removed and placed into the well of a 24-well plate containing 500 nicroliters of crystal violet. The transwell was incubated 20 20 minutes in the dye and then washed repeatedly by dunking the transwell in a. beaker of vater. A pre-wetted cotton swab was used to scrape cells from the top surface of the rranswell. Cells that had migrated to the bottom surface of the transwell were viewed by ligJt microscopy, photographed, and the number of migrated cells per field were counted. See Figure 24, 25 Statistics Student's r-test was used for statistical analysis. Significance was determined by a confidence level above 95% (P<0.05). 30 EXAMPLE 2: In vivo Experiments Mice and establishment of tumors C57BL/6 mice were purchased from Charles River, Quebec, Canada at 54 weeks of age. Mice were allowed one week to acclimate before experimentation began. BI 6F1 0 32 murine melanoma cells were purchased from ATCC and cultured in DMEM-1 0%FBS. The cell monolayer was trypsinized and neutralized with MEM-10%FBS. Cells were washed with PBS twice and cell viability was determined by trypan blue staining. For experimental metastasis experiments (Experiments 1 and 111). melanoma tumors were 5 established in the lung by tail vein injection of B1 6F0 cells into 6-week old mice. B16F10 cells were diluted to 1x I04 viable cells/mi in PBS. 200 1 of cells was injected into each mouse via tail vein. For subcutaneous tumor experiments (Experiments IV and V), melanoma tumors were established by subcutaneous injection of 500,000 B16F1 0 cells into the right flank of 10 to 14-week old mice. At the end of all experiment (when mice 10 became moribund or tumor exceeded a pre-determined size), the mice were euthanized by C02 inhalation. EXAMPLE 3: Experiment H Construction and purification of plasmid DNA 1$ The pCpG-lacZ expression vector lacking CpG dinucleotides was purchased from InvivoGen. San Diego, USA. An IA-tagged eIEFA cDNA was subcloned into the pCpG lacZ vector by first digesting the plasmid with Ncol and Nhel and isolating a 3. kb of pCpG vector backbone (thereby removing LacZ coding sequence) and ligating with a PCR amplified cDNA of elFiAl containing an HA tag (pCpG- lA5Al), The PCR primers 20 were elF5Al for: HA-5Al for: 5('-CTCCATGGAT(TACCCATACGACGTCCC-3'; and efFSAI rev' S'-CGCGCTAGCCAGlTATTTGCCATCGCC-3'. The pCpG-lacZ and pCpG-HA5A I were amplified in E. coli G1 15 cultured in LB or 2XYT medium containing 25 pg/ml of zeocin, The plasmids were extracted and purified by QIAGEN Endofree Plasmid Gina kit. The DNA concentration was measured by UV 25 absorption at 260nm and agarose gel electrophoresis, Tail vein injection of plasmid DNA (Experiment 1.): Plasmid DNA dissolved in Ix PBS (around 200 pl based on body weight) were injected into the tail vein of mice at days 2. 4, 7, 11, 16t 21. 26, 31. Plasmid DNA 30 concentration was 660 ng/pl for 2x (6.6mg/kg), 330 ng/pl for Ix (3.3mg/kg), and 33 ng/pl for 0. Ix (033 mg/kg).
Body and hung weights: Body' weights were measured before tail vein injection or every Monday and Friday. Mice were. euthanized with CQ 2 when they reached morbidity (lethargic. respiratory distress) and lungs were removed, weighed, photographed, frozen, and stored 5 at 470*C. See Figures 25-27. VEGF ELISA: Harvested lung tissues were washed with PBS, frozen in dry ice, and stored at 704C. Protein lysates were isolated from ground lung tissues. 50pg of lung tissue proteins 10 were used to determine VECF concentration in mouse lung using Mouse VEGF immunoassay kit (R&D Systems, Inc, Minneapolis, US A). EXAMPLE 4: Experiment III injection of plasmid DNA/DOTAP complexes IS 6 weeks old mice were tail vein injected with B16F10 melanoma cells (50,000 cells in 200p1 PBS/mouse), plasmid DNA, and DNA carrier complex containing 50 pg of endotoxin-free kit puri fed plasmid DNA, and 80 pg of DOTAP were tail vein injected into mice at days 7, 14, and 21. Mice vere sacrificed at day 25. Lungs were removed, weighed, and photographed. See Figures 28-29. 20 EXAMPLE 5: Experiment IV injection of Adenovirus constructs and TUNEL 1 0-week old C57BL/6 mice were injected subcutaneously into the right flank with either 500,000 B16FO or B11610 cells. lx109 pft Ad-5Al in u5. of PBS was injected 25 into tumors when the tumors reached about 4mm in diameter (10-12 days after Bl6 cell infection:). Mice were sacrificed after 48 hours and tumors were excised, fixed, and embedded in paraffin. Two sections for each cell tumor type (Ad-SA1-1 and Ad-5AI-2) were stained by TUNIEL (Promnega) according to the manufacturers protocol. Negative control slides (Ad-5Al-neg) in vhich the TdT enzyme was left out of the TUNEL reaction 30 were included for each cell type. See Figure 30. 34 EXAMPLE 6: Experiment V Establishment of tumors I 4-week old C57BL/6 mice were injected subcutaneously with 500,000 B16F10 cells (in 100ul of PBS) on the right flank. The progress of tumor formation was checked 5 daily until the tumor size reached around 8 nun in diameter. Injection with Adenovirus Treatment began when tumor size reached a diameter of 8 nn. Mice were injected with either 1 X 10? plaque forming units (pfu) of Ad-eLF5Al or Ad-LacZ. Injections were 10 distributed over three sites in the tumor. Mice were injected every day for the first three days and then every other day thereafter until the mouse was sacrificed. Mice were sacrificed when the tumor size exceeded 15 to 16 mm in one dimension. Buffer only mice received only the bufferin which the adenovirus was suspended (10 mM Tris-HCl pH 7.4, 10 mM NgCI 2 , 10% glycerol). The tumor dimensions were measured every day using 15 calipers and the tumor volume was calculated using the equation below: Tumor volume (mnm)= L * W" *0.52 Where L= length, W = width (always shorter dimension) See Figures 31 and 32. 20 EXAMPLE 7: A549 lung carcinoma cells were infected with either an adenovirus expressing LacZ (Lac) or eIF5A (5A). Four hours after infection, the media was replaced with media containing either DMSO, 10 tiM ofie p38 inhibitor SB203580 (Calbiochem), 10 pM of the JNK inhibitor H (Calbiochem), 10 pM of the MEK inhibitor U1026 25 (Cal biochem), or 30 pM of the p53 inhibitor Piflithrin-a (Calbiochem). Forty-eight hours later, the cells were treated with EGF for 30 minutes and the cell lysate was harvested. Western blots were performed on the lysale using antibodies directed against either total p53 (p5 3 ), or p53 phosphorylated on serine 15 [P-p53(serl.5)], or p53 phosphorylated on 37 IP-p53(ser37)]. See Figure 33. 30 The results shown in Figure 33 shows that Ad-eIF5AI induces accumulation of p33 by 48 hours after infection. The figure also shows that Ad-elFSAl induces phosphorylation of p53 by 48 hours after infection; accumulation and phosphorylation of p53 is inhibited by inhibitor of TMIEK; accumulation and phosphorylation of p53 is inhibited by inhibitor of p53 activity; elFSAI stimulates MEK-dependent phosphoiylation of p53; and etF5AT stimulates p53-dependent accumulation of p53. EXAMPLE 8: 5 A549 lung carcinoma cells were infected with either an adenovirus expressing LacZ (Ad-LacZ) or elF.5A (Ad-elF5A.l). Forty-eight hours later, the total RNA. was isolated from the cells. The levels of p53 and bax mRNA transcript levels were determined by Real Time PCR using GAPDH as a reference gene. The p53 primers were: 5ctCGCTGCTCAGA TAGCGATGGTC~3 t (5tprimer) and i'~ 1o CTTCrTOGCTOGGGAGAiGAG-3' (3 t -primer) [These p53 primer sequences were obtained from: Li et at (2004). A Novel eJFSA /complex Functions as a Regulator of p53 and p53-dependent Apoptosis, J Bio (Chem. 279 4925 1-49258 See Figure 34, which shows that over-expression of eLF5AI induces accumulation of p53 at mRNA level, yet no effect on bax was seen. 15 EXAMPLE 9: A549 Jung carcinoma cells were infected with either an adenovirus expressing LacZ (Ad-LacZ) or eLFSAI (Ad-eIPF5A). Four hours after infection, the media was replaced with media containing either DMSO, 10 pM of the MEK inhibitor U 026 20 (Calbiochem), or 30 gM of the p53 inhibitor Pifithrin-a (Calbiochem). Forty-eight hours later, the total RNA was isolated from. the cells. The levels of p53 transcript levels were determined by Real Time PCR using GAPDH as a reference gene. The p53 primers were:5 t -CGCTGCTC AGATAGCGATG(GTC-3' ('-primer) and 5' CTTCTTTCGCTGGGGAGAGGAG-3' (Y-primer) [These p53 primer sequences were 25 obtained from: Li et at (2004). A Novel eTF5A /complex Functions as a Regulator of p53 and p53-dependent Apoptosis, J Biol Chein. 279 49251-492581. See Figure 35, which shows that the eFSA I-dependent accumulation of p53 is dependent on p53 transcriptional acti vity, and the upregulation of p53 protein that occurs in response to elF5AI results in increased transcription of p53 mRNA. 30 EXAMPLE 10 A549 lung carcinoma cells were infected with either an adenovirus expressing LacZ (Ad-LacZ) or eIF5A1 (Ad-eIF5AI). Four hours after infection, the media was replaced with media containing either DMSO, 10 pM of the MEK inhibitor CI 026 (Calbiochem), or 30 gM of the p53 inhibitor Pifithrin-a (Calbiochem). Fony-eighthours later, the total RNA was isolated from the cells. The levels ofp 5 3 transcript levels were determined by Real Time PCR using GAPDH as a reference gene. The TNFRI primers were: TNFRI-F 5' ATCTCTTCTTGCACAGTGC3' and TNFRI-R 5' 5 CAATGGAGTAGAGCTTCAC 3' See Figure 36. which shows that TNFRI mRNA levels are upreaulated by infection with Ad-eIF5AL and that this accumulation of TNFR.1 mRNA is partially dependent on MEK. This accumulation of TNFRI1 mRNA is dependent on p5 3 transcriptional aclivitv. 10 EXAMPLE H1 .A549 lung carcinoma cells were infected with either an adenovirus expressing LacZ (Lac) or elFiAt (5A). Four hours aRer infection, the media was replaced with media containing either DMSO, 10 pM of the p38 inhibitor SB203580 (Calbiochem), 10 pM of the JINK inhibitor 11 (Calbiochem), 10 pM of the MEK inhibitor U1 026 15 (Calbiochem), or 30 pM of the p53 inhibitor Pifthrin-a (Calbiochem). Forty-eight hours later, the cells were harvested and the percentage of cells undergoing apoptosis was determined by Annexin/Pi staining (BD Bioscience)'followed by analysis by flow cvtometiy. See Figure 37, Figure 37 shows that Ad-eIFSA I induces apoptosis by 48 hours after infection: 20 that inhibition of JNK increases apoptosis induced by elFSAI; that inhibition of MEK increases apoptosis induced by e.IF5A I, and that inhibition of p53 decreases apoptosis induced by eIF5AL EXAMPLE 12 25 Mice were injected with 50,000 B16-FO melanoma cells sub-cutaneously. When the tumors reached a size of around 5 x5nm (65niim) intra-tumoral injections were initiated. 0 I pfi of either Ad-lacZ (group 2), Ad-eJF5AI (group 3), or Ad-elF5A2 (group 4) diluted in 50-100 PI of PBS/I 0% glycerol buffer or buffer only (group 1) were injected into the tumors in three sites per tumor every other day. The tumor size was 30 measured eveiy other day until every the other day until the sacrifice of mice when tumor size reached 10% of the body weight See Figure 39. 37 EXAMPLE 13 Mice were injected with 50,000 B1 6-FO melanoma cells sub-cutaneously. When the tumors reached a size of around 5 x5mm (65mm') intra-tumoral injections were initiated. 1x1:i pfu of either Ad-lacZ (group 2), Ad-eIF5Al (group 3), or Ad-eIF5A2 5 (group 4) diluted in 50-100 pl of PBS/I 0% glycerol buffer or buffer only (group 1) were injected into the tumors in three sites per tumor every other day. The mice were sacrificed when tumor size reached 10% of body weight. See Figure 40 10 EXAMPLE 14: Multiple myeloma On day 0. KAS 6/1 multiple myeloma cells were infected with 3000 IFUcell wild type or mutant elF~a (unable to be hypusinated -- conserved lysine is mutated) adenovirus vector constrct for 4 hours. Three replicates with or without IL-6 present in the post infection culture media were set up, In addition KAS cells were plated for controls (were 15 not infected). On day 2 and 4, MTT and annexhit/P assays were performed. The supernatants were harvested. The results are shown in figure 4 38