FIELD OF THE INVENTION

The present invention generally concerns RNAi-based methods and compositions for treating metastatic cancer.

BACKGROUND OF THE INVENTION

Metastasis is the spread of a cancer from one organ, usually the primary tumor, to another non- adjacent organ. Typically, pancreatic cancer first metastasizes to regional lymph nodes, and later to the liver or to the peritoneal cavity and, rarely, to the lungs, bone or brain. Metastases represent the end-products of a multi-step cell-biological process (Valastyan and Weinberg 2011) termed the invasion metastasis cascade, which involves dissemination of cancer cells to anatomically distant organ sites and their subsequent adaptation to foreign tissue microenvironments. Each of these events is driven by (1) acquisition of genetic and/or epigenetic alterations within tumor cells and (2) cooption of non-neoplastic stromal cells, which together endow incipient metastatic cells with traits needed to generate macroscopic metastases. Recent advances have provided provocative insights regarding these cell-biological and molecular changes, which carry implications concerning the pathogenesis of metastatic progression and the steps of the invasion-metastasis cascade that appear amenable to therapeutic targeting.

An epithelial-mesenchymal transition (EMT) is a biologic process that allows a polarized epithelial cell, which normally interacts with basement membrane via its basal surface, to undergo multiple biochemical changes that enable it to assume a mesenchymal cell phenotype, which includes enhanced migratory capacity, invasiveness, elevated resistance to apoptosis, and greatly increased production of ECM components. Genes that are (abnormally) expressed in the primary tumor can facilitate metastasis by processes including seeding and re-seeding. Some genes relate to colonization at spefic far organ. (Gupta and Massague, 2006). A variety of factors, including Snail, Slug, Twist, Zebl, SIPl, E47, fibronectin 1 (FN1), Vimentin (vim), N-Cadherin, SERPINA3, CD70, IL13RA2 and CD74 , were shown to facilitate metastasis, either via induction of EMT through repression of E-cadherin transcription or via other pathways (Kalluri and Weinberg 2009, Brabletz, 2012, Holtz 2007, Pan and Yang 2011). Ken et al (US patent application 20120302572) grouped in EMT gene signature (EMTGS), genes of which abnormal expression might be associated with EMT, including the following genes: SERPINA3, ACTN1, AGR2, AKAP12, ALCAM, APIM2, AXL, BSPRY, CCL2, CDH1, CDH2, CEP170, CLDN3, CLDN4, CNN3, CYP4X1, DNMT3A, DSG3, DSP, EFNB2, EHF, ELF3, ELF5, ERBB3, ETV5, FLRT3, FOSB, FOSL1, FOXC1, FX YD 5, GPDIL, HMGA1, FIMGA2, HOPX, IFI16, IGFBP2, IHH, IKBIP, IL-11, IL-18, IL6, IL8, ITGA5, ITGB3, LAMB1, LCN2, MAP7, MB, MMP7, MMP9, MPZL2, MSLN, MTA3, MTSS1, OCLN, PCOLCE2, PECAM1, PLAUR, PLXNB1, PPL, PPP1R9A, RASSF8, SCNN1A, SERPINB2, SERPINE1, SFRP1, SH3YL1, SLC27A2, SMAD7, SNAI1, SNAI2, SPARC, SPDEF, SRPX, STAT5A, TBX2, TJP3, TMEM125, TMEM45B, TWIST1, VCAN, VIM, VWF, XBP1, YBX1, ZBTB10, ZEB1, ZEB2.

Some genes of these EMTGS and additional genes are overexpressed in malignant cancers and dounregulation by RNAi potentially can slow metastatic development and progression, including SERPINA3 CD70, IL13RA2 and CD74 (Deng et al 2010). SI OOP is a metastasis- associated gene that facilitates transendothelial migration of pancreatic cancer cells (Barry et al 2013).

Genes that are over-expressed and therefore down-regulated by RNAi agents may suppress the development of metastasis are targets of this invention, including:

ACTN1, AGR2, AXL, CCL2, CD74, CDH1, CDH2, CLDN3, CLDN4, CX3CR1, CXCL1, FOXC1, FXYD5, HMGA1, HOXB7, IHH, IL6, IL8, IL11, IL13RA2, IL18, ITGA5, ITGB3, LAMB1, LCN2, MMP2, MMP7, MMP9, MSLN, MTA1, PCOLCE2, PECAM1, PLXNB1, S100A11, S100A6, S100P, SERPINA3, SERPINB2, SERPINE1, SFRP1, SLC27A2, SMAD7, SNAI1, SNAI2, SPARC, STAT5A, TBX2, TWIST1, VCAN, VEGFA, VIM, XBP1, YBX1, ZEB1 and ZEB2.

K-ras is a GTPase protein encoded by the Kirsten rat sarcoma 2 viral oncogene homolog (K- ras) oncogene (also known as K-ras2 or RASK2; PubMed Gene ID # 3845), which belongs to the family of RAS proteins.

The human RAS family consists of three closely related proto-oncogenes: c-Harvey (H)-RAS, c-Kirsten (K)-RAS, and N-RAS, which share 90% of their peptide sequence. RAS proteins are localized in the inner cell membrane, bind GDP and GTP, and possess an intrinsic GTPase activity, implicated in the regulation of their activity. RAS proteins influence proliferation, differentiation, transformation, and apoptosis by relaying mitogenic and growth signals into the cytoplasm and the nucleolus. In a normal cell, most of the RAS molecules are present in an inactive GDP-bound conformation. K-ras mutations

Genetic alterations in the K-ras signaling pathway are involved in more than 90% of cases of Pancreatic Cancer (PC) (Rejiba et al, and references cited therein); the majority of such mutations are gain-of-function mutations at codon 12 (K-rasG12D). PC is an aggressive disease, being one of the leading causes of cancer-related death in the western world. Mutations in K-ras are also well known in other cancers. For example, K-ras is involved in the development and progression of colorectal cancer (CRC). Mutations are present in about 40% of CRC cases, most commonly at codons 12 and 13. These mutations prevent dephosphorylation and inactivation of the protein, causing it to be permanently switched on, independently of EGFR-mediated signaling. These mutated K-ras proteins are unlikely be affected by inhibition of EGFR, since the mutation causes the encoded K-ras protein to "freeze" in its active state for a much longer duration than its non-mutated counterpart. Hence, the response to anti-EGFR mABs is strongly reduced in tumors with mutated K-ras (Normanno et al). K-ras mutations also occur in 20-30% of lung cancer patients. Anti-K-ras mutations RNAi-based treatment:

RNAi vectors (Brummelkamp et al) and siRNA against mutated K-ras can lead to apoptosis of cancer cells in vitro and in vivo in mice and reduce tumor growth (Wang et al, Morioka et al, Fleming et al, Rejiba et al, WO2010/001325). Moreover, anti mutated K-ras treatment potentially can slow the epithelial-to-mesenchymal transition (EMT) and thereby slow metastasis of PC disease. For example Singh et al observed an association between epithelial differentiation and tumor cell viability, and that EMT regulators in "K-Ras-addicted" cancers represent candidate therapeutic targets. siRNA (short interfering RNA) have been delivered by millimeter- scale depot technology against mutated K-ras (WO2010/001325 to Shemi) and against other non-oncology targets (US2008/0124370 to Marx). SUMMARY OF THE INVENTION

The present invention relates to methods and compositions for treating metastatic cancer that is associated with a primary tumor or a few tumors in the primary organ.

Provided herein is a millimeter-scale drug delivery device (DDD; also referred to herein as a "LODER"), comprising: (a) a biodegradable polymeric matrix; and (b) an RNAi (RNA interference) agent that targets a gene, incorporated within the biodegradable polymeric matrix, for treating a metastatic cancer originating from a primary tumor or tumors, by inserting the DDD into the primary tumor(s). The targeted gene may be any appropriate target. In some embodiments, the gene is a metastasis-facilitating gene (e.g. a gene that promotes metastasis). The RNAi may be, in various embodiments, a DICER substrate, an siRNA or another type of RNAi. As used herein "comprising" is to be interpreted as specifying the presence of the stated features, integers, steps, or components as referred to, but does not preclude the presence or addition of one or more features, integers, steps, or components, or groups thereof. Thus, for example, a drug delivery device comprising a biocompatible and biodegradable polymeric matrix and a given anti-K-rasG12D RNAi agent may contain additional excipients or components, such as non-polymeric components, non-matrix components, and additional agents. Additionally, the term "comprising" is intended to include, as separate embodiments, embodiments encompassed by the terms "consisting essentially of and "consisting of." The phrase "consisting essentially of" limits the scope of a claim to the recited materials or steps, either alone or in combination with additional materials or steps that do not materially affect the basic and novel characteristics of the claimed invention.

In light of the disclosure provided herein, it will be appreciated by those skilled in the art, in light of the present disclosure, that methods of the present invention are suitable for treating any type of metastatic cancer that is associated with a primary tumor or a few tumors in the primary organ. BRIEF DESCRIPTION OF THE DRAWINGS

The following figures are by way of illustrative example and are not meant to be taken as limiting the claimed invention.

Fig 1. Schematic depiction of the treatment described in the invention.

Fig 2. A. siG12D siRNA-containing DDD significantly inhibits growth of pancreatic cancer cells in vitro at 48 hours. B-C. Effect of siG12D LODER on K-ras expression at 48 hours in Panel -LUC cells (B) and original Panel cells (C). In C, vertical axis is the ratio of K-ras/β actin protein level. D. Inhibition of Luciferase expression in the presence of either the siLUC LODER or the siG12D LODER after 72hrs. E-F. Cell viability in the presence of the siG12D- LODER and siGFP-LODER, measured by XTT (E) and LHD (F) assays.

Fig 3. Functionality of the siG12D DDD in vivo. (A) Luciferase activity in mice bearing subcutaneous syngeneic CT26-LUC tumors after intratumoral implantation of 2 siLuc or siGFP LODERs. (B) Tumor weight in the group treated with siLuc LODER in comparison with the siGFP LODER group. (C) Lucif erase activity in the livers of mice that stably express Lucif erase in the liver after implantation of an Empty or siLuc LODERs. * indicates p<0.05 according to the student's t-test Fig 4. siG12D DDD inhibits tumor growth in vivo. A. Images. B. Quantification of percent of necrotic area (vertical axis). Arrows mark the intratumoral locations of the LODER implant.

Fig 5. Anti-tumor effect of the siG12D DDD in an orthotopic pancreatic model, as measured by K-ras staining. Images are shown in A; and percentage of K-ras positively-stained cells is shown in B). Percentage of cells staining for K-ras and CDC47 as a function of distance from the LODER is shown in C and D, respectively (vertical axis is the percent of cells staining for K-ras and CDC47, respectively).

Fig 6. K-ras expression in human pancreatic tissue.

Fig 7. siG12D reduces EMT in Panel cells. A. Migration assay. Panc-1 cells were transfected with siG12D or non-targeting control (scrambled) or were treated with TGFP (positive control). The migration ability was assessed using a modified Boyden chamber assay. The percentage of migrated cells relative to untreated cells is presented. B. Scratch assay. Panc-1 cells were mock-transfected or transfected in vitro with scrambled siRNA or siG12D, then were seeded in 6W plates. After 24 hours, the cells were reseeded in 24W plates to form a confluent culture. One day after reseeding, the cells were scratched with a plastic tip. The migration of the cells was followed-up after 7 and 72 hours under a light microscope.

Fig 8. Light microscopy images of migration of siG12D-DDD treated cells in a scratch assay, showing the inhibitory effect of siG12D on cellular migration.

Fig 9. Effect of the siG12D DDD on expression levels of metastasis-facilitating genes. Expression levels are normalized to the untreated cells. Fig 10. A model for metastatic pancreatic cancer.

Fig 11. Clinical results: CT scan showing tumor response nine months post-insertion (bottom panel) compared to the time of insertion (top panel).

Fig 12. Clinical results: comparison of overall survival of patients with LAPC. Figure 13. Plot showing effect of various OMe-modified-siG12D#l-8 (top panel) and #9-15 (bottom panel) on viability of Pancl-luc cells. Higher OD levels indicate higher viability and therefore weaker effect. 0' denotes unmodified siG12D. siRNA against Luc and GFP are included as controls. An asterisk indicates a P-value of < 0.05 relative to unmodified siG12D, by Student's T-test.

Figure 14. Summary of stability and efficacy data with various OMe-modified-siG12D measured at 24h, lw, 2w and 4w (stability), and after 120h (efficacy).

Figure 15. Gemzar® enhances siG12D-DDD effect. Panel cells were transfected using Lipofectamine™with unmodified and modified siG12D and a scrambled (non-targeting) siRNA control molecule or were mock transfected(no siRNA). Cell viability was assessed 72 hrs post- transfection by the Methylene Blue (MB) test. Additional groups were transfected as described above, followed by lhr of treatment with 10μΜ Gemzar®. The numbers 1-4 in the last column relate to the percent of inhibition of cellular growth, wherein the inhibition seen with the unmodified siRNA was assigned the value of 100%. 4, 3, 2, and 1 refer respectively to 75- 100%, 50-75%, 25-50%, and 0-25% of the inhibition seen with the unmodified siRNA.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Provided herein is a millimeter-scale drug delivery device (DDD; also referred to herein as a "LODER"), comprising: (a) a biodegradable polymeric matrix; and (b) an RNAi (RNA interference) agent that targets a gene, incorporated within the biodegradable polymeric matrix, for treating a metastatic cancer originating from one or more primary tumors in the same organ, by inserting the DDD into the primary tumor(s). The targeted gene may be any appropriate target. In some embodiments, the gene is a metastasis-facilitating gene. In other embodiments, the gene is an oncogene. The RNAi may be, in various embodiments, an siRNA, a Dicer substrate RNAi (for example as provided by IDT - INTEGRATED DNA TECHNOLOGIES) or another type of RNAi.

The primary tumor in which the DDD is inserted may be a solid tumor or, in other embodiments, another type of tumor.

The subject receiving the DDD is, in some embodiments, concurrently receiving chemotherapy for the metastatic cancer or has received or will receive the chemotherapy within 6 months of administration of the DDD. Those of skill in the art will appreciate in light of the present disclosure that various types of chemotherapy may be utilized, for example Erlotinib, Gemcitabine, FOLFIRINOX, Abraxane, or for pancreatic cancer; or in other embodiments a substance selected from pyrimidine analogues, EGFR tyrosine kinase inhibitors, and a subset or all the agents included in FOLFIRINOX. Alternatively or in other embodiments in addition, the subject is receiving or has received or will receive within 6 months radiation therapy for the metastatic cancer.

The subject receiving the DDD is, in some embodiments, concurrently receiving, has received in the past, or is about to receive, treatment which is not chemotherapy, including radiation treatment, including external beam radiotherapy external beam radiotherapy (EBRT) and brachytherapy; ultrasound, including High- Intensity Focused Ultrasound (HIFU); surgery; immunotherapy; cryotherapy; hormonal therapy; and monoclonal antibody therapy. The treatment is provided in various modes known to one skilled in the art, including local treatment at the primary tumor or tumors, local treatment at the metastasis and regional and systemic treatment.

In certain embodiments, the chemotherapy that is administered is indicated for a subject that has any performance status from 0 to 4 (In this invention performance status (PS) refers to the ECOG performance status, ranged from 0 to 5, grade 0 = Fully active, able to carry on all pre- disease performance without restriction, and grade 5 = dead. Higher PS refers to lower performance. Not to be confused with the Karnofsky score that runs from 100 to 0, where 100 is "perfect" health and 0 is death). In some embodiments the chemotherapy that is administered is indicated specifically for a subject that has a performance status of 1 or higher. Examples of such regimens are Gemcitabine for pancreatic cancer, which is indicated specifically for patients that are unexpected to tolerate harsh regimens such as FOLFIRINOX. In other embodiments, the patient has a performance status of 2 or higher. In still other embodiments, the patient has a performance status of 3 or higher. In yet other embodiments, the patient has a performance status of 4.

Alternatively, the DDD is administered to a subject for whom chemotherapy for the metastatic cancer is contraindicated, for example due to weakness, advanced age, or high performance status. DDD's described herein have been shown to be well tolerated and are thus suitable for patients unable to tolerate chemotherapy regimens. For example, such patients may have a performance status of 1 or higher and/or be older than 76. In other embodiments, the patient has a performance status of 2 or higher. In still other embodiments, the patient has a performance status of 3 or higher. In yet other embodiments, the patient has a performance status of 4. The RNAi agent contained in the DDD may, in some embodiments, a target gene. In various embodiments, the gene may be Snail, PAK1, PAK2, ZEB1, ZEB2, Twist, Goosecoid, SIX1, FOXC2, Slug, SIP1, Prxxl, E47, FN1, VIMENTIN (VIM), N-CADHERIN, SERPINA3, CD70, IL13RA2, or CD74. In other embodiments, one of the following is selected: Snail, Slug, Twist, Zebl, SIP1, E47, fibronectin 1 (FN1), Vimentin (vim), N-Cadherin, SERPINA3, CD70, IL13RA2 and CD74.

In still other embodiments, the target gene is a mutated K-ras gene, for example a mutation selected from G12D, G12V, G12A, G12S, G12C, G13D, and G13C. Specific sequences of certain embodiments of siRNA's against mutated K-ras are set forth hereinbelow. In yet other embodiments, the DDD comprises, in addition to an RNAi agent against mutated K-ras, another RNAi agent directed against a metastasis-facilitating gene. In various embodiments, the aforementioned gene may be Snail, PAK1, PAK2, ZEB1, ZEB2, Twist, Goosecoid, SKI, FOXC2, Slug, SIP1, Prxxl, or E47, FN1, VIMENTIN (VIM), N- CADHERIN, SERPINA3, CD70, IL13RA2 and CD74. siRNA sequences targeting mutated K-ras

In other the RNAi agent contained within the DDD comprises a duplex region, and the nucleotide sequence of the duplex region of the sense strand consists of a sequence selected from SEQ ID No: 1-7, namely GUUGGAGCUGAUGGCG (SEQ ID No: 1), GUUGGAGCUGUUGGCG (SEQ ID No: 2), GUUGGAGCUGCUGGCG (SEQ ID No: 3), GUUGGAGCUAGUGGCG (SEQ ID No: 4), GUUGGAGCUUGUGGCG

(SEQ ID No: 5), GUUGGAGCUGGUGACG (SEQ ID No: 6), and GUUGGAGCUGGUUGCG (SEQ ID No: 7), either alone or followed by: a sequence selected from: (i) UAGGCAAGAGUGCC (SEQ ID No: 8) and (b) a 5'- fragment of 1-13 nucleotides inclusive of SEQ ID No: 8. "Followed by" in this regard means that the 3'-terminus of the sequence selected from SEQ ID No: 1-7 is connected to the 5'-terminus of SEQ ID No: 8 or a fragment thereof.

For purposes of illustration, the following sense strands contain SEQ ID No: 1 and all or a portion of SEQ ID No: 8: GUUGGAGCUGAUGGCGU (SEQ ID No: 16), GUUGGAGCUGAUGGCGUA (SEQ ID No: 17), GUUGGAGCUGAUGGCGU AG (SEQ ID No: 18), GUUGGAGCUGAUGGCGU AGG (SEQ ID No: 19), GUUGGAGCUGAUGGCGU AGGC (SEQ ID No: 20),

GUUGGAGCUGAUGGCGU AGGCA (SEQ ID No: 21), GUUGGAGCUGAUGGCGUAGGCAA (SEQ ID No: 22),

GUUGGAGCUGAUGGCGUAGGCAAG (SEQ ID No: 23),

GUUGGAGCUGAUGGCGUAGGCAAGA (SEQ ID No: 24),

GUUGGAGCUGAUGGCGUAGGCAAGAG (SEQ ID No: 25),

GUUGGAGCUGAUGGCGUAGGCAAGAGU (SEQ ID No: 26),

GUUGGAGCUGAUGGCGUAGGCAAGAGUG (SEQ ID No: 27),

GUUGGAGCUGAUGGCGUAGGCAAGAGUGC (SEQ ID No: 28), and

GUUGGAGCUGAUGGCGUAGGCAAGAGUGCC (SEQ ID No: 29) Similar sense strains may be readily derived from each of SEQ ID No: 2-7 together with all or a portion of SEQ ID

No: 8.

In certain embodiments, the aforementioned 5' fragment of SEQ ID No: 8 is at least 3 nucleotides in length; thus, the duplex region of the antisense nucleotide is at least 19 nucleotides in length.

In other embodiments, the duplex region is 19 nucleotides in length (thus comprising both one of SEQ ID No: 1-7 with a 3-nucleotide fragment of SEQ ID No: 8). More specific embodiments of such sequences are those selected from GUUGGAGCUGAUGGCGUAG (SEQ ID No: 9), GUUGGAGCUGUUGGCGUAG (SEQ ID No: 10), GUUGGAGCUGCUGGCGUAG (SEQ ID No: 11), GUUGGAGCUAGUGGCGUAG (SEQ ID No: 12), GUUGGAGCUUGUGGCGUAG (SEQ ID No: 13), and GUUGGAGCUGGUGACGUAG (SEQ ID No: 14), and GUUGGAGCUGGUUGCGUAG (SEQ ID No: 15).

In other embodiments, the duplex region is 16 nucleotides in length (thus consisting of one of SEQ ID No: 1-7 alone). In other embodiments, the duplex region is 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, or 29 nucleotides in length (thus comprising one of SEQ ID No: 1-7, together with a fragment of SEQ ID No: 8 that is 1, 2, 4, 5, 6, 7, 8, 9, 10, 11, 12, or 13 nucleotides, respectively). In other embodiments, the duplex region is 30 nucleotides in length (thus comprising one of SEQ ID No: 1-7, together with SEQ ID No: 8 in its entirety).

Reference herein to the "duplex region of the sense strand" indicates the portion of the sense strand that exists in a duplex structure after hybridization to the antisense strand. Similarly, reference to the "duplex region of the antisense strand" indicates the portion of the antisense strand that exists in a duplex structure after hybridization to the sense strand. In various embodiments, the duplex region of the sense strand may be either the entire sense strand or a fragment thereof. Similarly, the duplex region of the antisense strand may be either the entire antisense strand or a fragment thereof.

In certain preferred embodiments, the duplex region is perfectly complementary. Thus, the duplex region of the antisense strand is complementary to the duplex region of the sense strand. For purposes of illustration, the following antisense strands have a sequence perfectly complementary to SEQ ID No: 1-7, respectively: CGCCAUCAGCUCCAAC (SEQ ID No: 30), CGCCAACAGCUCCAAC (SEQ ID No: 31), CGCCAGCAGCUCCAAC (SEQ ID No: 32), CGCCACUAGCUCCAAC (SEQ ID No: 33), CGCCACAAGCUCCAAC (SEQ ID No: 34), CGUCACCAGCUCCAAC (SEQ ID No: 35), and CGCAACCAGCUCCAAC (SEQ ID No: 36).

Reference herein to the "nucleotide sequence of the duplex region" relates to the sequence of the nucleotides found in the duplex region. For these purposes, modified ribonucleotides are considered to have the same identity as the ribonucleotide from which they were derived (the "parent ribonucleotide"), provided that they retain the base pairing specificity of the parent ribonucleotide. Thus, cytosine containing a 2'-OMe modification, for example, is still considered cytosine, since it retains the ability to pair with guanine.

In still other embodiments, a DDD of the present invention comprises more than one, for example 2, 3, 4, or 5, of the 16-30 nucleotide anti-K-ras RNAi agents described herein, in other words, RNAi agents whose sense strand comprises (a) a sequence selected from SEQ ID No: 1-7, optionally in combination with (b) either SEQ ID No: 8 or a 1-13 nucleotide fragment thereof. In even more specific embodiments, a DDD of the present invention comprises more than one, for example 2, 3, 4, or 5, of the 19 nucleotide RNAi agents described herein, in other words, RNAi agents whose sense strand consists a sequence selected from SEQ ID No: 9-15. In still other embodiments, the DDD comprises an RNAi agent against a metastasis-facilitating gene in addition to the multiple anti-K-ras RNAi agents.

In still other embodiments, genes that are over-expressed and therefore down-regulated by RNAi agents may suppress the development of metastasis are targets of this invention, including:

ACTN1, AGR2, AXL, CCL2, CD74, CDH1, CDH2, CLDN3, CLDN4, CX3CR1, CXCL1, FOXC1, FXYD5, HMGA1, HOXB7, IHH, IL6, IL8, IL11, IL13RA2, IL18, ΓΓΟΑ5, ITGB3, LAMB1, LCN2, MMP2, MMP7, MMP9, MSLN, MTA1, PCOLCE2, PECAM1, PLXNB1, S100A11, S100A6, S100P, SERPINA3, SERPINB2, SERPINE1, SFRP1, SLC27A2, SMAD7, SNAI1, SNAI2, SPARC, STAT5A, TBX2, TWIST1, VCAN, VEGFA, VIM, XBP1, YBX1, ZEB1 and ZEB2.

In still other embodiments, at least one of the sense strand and the antisense strand further comprises a 3'-overhang 1-6-nucleotides (nt) in length, in other words, a region that does not hybridize with the other strand. In more specific embodiments, the 1-6-nt 3'-overhang is present on both the sense strand and the antisense strand.

In other embodiments, the overhang present on at least one of the sense strand and the antisense strand is a 2-nt 3'-overhang. In still more specific embodiments, the overhangs present both the sense strand and the antisense strand are 2-nt 3'-overhangs. In other embodiments, the 3'-overhang present on at least one of the sense strand and the antisense strand consists of dTdT (a dimer of deoxythymidine). In still more specific embodiments, 3'-overhangs are present on both the sense strand and the antisense strand, each of which consists of dTdT.

In certain embodiments, the sense and antisense strands of an RNAi agent of methods and compositions of the present invention each contain only cytosine, guanine, adenine, and thymidine in the duplex region and the dTdT overhangs.

DDD's containing multiple RNAi agents

Also provided herein is a DDD that comprises (a) an RNAi agent against mutated K-ras and (b) another RNAi agent directed against a metastasis-facilitating gene. In various embodiments, the aforementioned gene may be Snail, PAK1, PAK2, ZEB1, ZEB2, Twist, Goosecoid, SIX1, FOXC2, Slug, SIP1, or Prxxl, or may be E47, FN1, VIMENTIN (VIM), N- CADHERIN, SERPINA3, CD70, IL13RA2 or CD74.

Suitable nucleotides and nucleotide analo ues

The term "nucleotide" as used herein refers to RNA nucleotides cytosine, guanine, adenine, and thymidine and to other similar nucleotide analogues suitable for use in RNAi agents. Such nucleotide analogues may generally be used as the building blocks of the RNAi agents described herein, except where indicated otherwise. Except where stated explicitly, nothing in this application is intended to preclude the utilization of non-classical nucleotide analogues, for example nucleotide analogues whose base or backbone has been chemically modified to enhance its stability or usefulness as an a RNAi agent, provided that they pair with the appropriate complementary bases, as previously mentioned. In certain embodiments, nucleotide analogues that may be used in methods and compositions of the present invention include derivatives wherein the sugar is modified, as in 2'-0-methyl, 2'-0-allyl, 2'-deoxy-2'- fluoro, and 2',3'-dideoxynucleoside derivatives; nucleic acid analogs based on other sugar backbones, such as threose; locked nucleic acid derivatives; bicyclo sugars; hexose, glycerol and glycol sugars; nucleic acid analogs based on non-ionic backbones, such as "peptide nucleic acids", these nucleic acids and their analogs in non-linear topologies, including as dendrimers, comb-structures, and nanostructures, and these nucleic acids and their analogs carrying tags (e.g., fluorescent, functionalized, or binding) to the ends, sugars, or nucleobases.

One non-limiting example of non-classical nucleotide analogues suitable for use in methods and compositions in accordance with the present invention include locked nucleic acid (LNA) nucleotide analogues. Certain embodiments of LNA nucleotide analogues are bicyclic nucleic acid analogs that contain one or more 2'-0, 4'-C methylene linkages, which effectively lock the furanose ring in a C3'-endo conformation. This methylene linkage "bridge" restricts the flexibility of the ribofuranose ring and locks the structure into a rigid bicyclic formation. Because of its unique structural conformation, oligonucleotides comprising LNA nucleotide analogues demonstrate a much greater affinity and specificity to their complementary nucleic acids than do natural DNA counterparts. LNAs typically hybridize to complementary nucleic acids even under adverse conditions, such as under low salt concentrations. LNA nucleotide analogues are commercially available, and are described, inter alia, in United States Pat. Nos. US 6,130,038, US 6,268,490, and US 6,670,461.

Another non-limiting example of non-classical nucleotide analogues suitable for use in methods and compositions in accordance with the present invention are peptide nucleic acid (PNA) nucleotide analogues. In certain embodiments of PNA nucleotide analogues, the negatively charged sugar-phosphate backbone of DNA is replaced by a neutral polyamide backbone composed of N-(2-aminoethyl) glycine units, such as in the illustrative example below, wherein B represents a nucleoside base. The chemical configuration of PNA typically enables the nucleotide bases to be positioned in approximately the same place as in natural DNA, allowing PNA to hybridize with complementary DNA or RNA sequence. PNA nucleotide analogues are commercially available, and are described, inter alia, in the PCT Applications having the Publication Nos. WO 92/20702, WO 92/20703 and WO 93/12129.

Another non-limiting example of non-classical nucleotide analogues suitable for use in methods and compositions of the present invention are glycol nucleic acid (GNA) nucleotide analogues (Zhang, L et al (2005) "A simple glycol nucleic acid". J. Am. Chem. Soc. 127:4174-4175). Certain embodiments of GNA nucleotide analogues have an acyclic propylene glycol phosphodiester backbone and have one of the structures below, wherein B represents a nucleoside base:

(f?)-GNA {S GHA

Another non-limiting example of non-classical nucleotide analogues suitable for use in methods and compositions of the present invention are threose nucleic acid (TNA) described in Wu et al, "Nucleotide Analogues" Organic Letters, 2002, 4(8):1279-1282. Certain embodiments of TNA nucleotide analogues have the structure below, wherein B represents a nucleoside base:

Another non-limiting example of non-classical nucleotide analogues suitable for use in methods and compositions in accordance with aspects of the present invention are bicyclic and tricyclic nucleoside analogs (Steffens et al, Helv Chim Acta (1997) 80:2426-2439; Steffens et al, J Am Chem Soc (1999) 121: 3249-3255; Renneberg et al, J Am Chem Soc (2002) 124: 5993-6002; and Renneberg et al, Nucl Acids Res (2002) 30: 2751-2757).

Another non-limiting example of non-classical nucleotide analogues suitable for use in methods and compositions of the present invention are phosphonomonoester nucleic acids which incorporate a phosphorus group in the backbone, for example analogues with phosphonoacetate and thiophosphonoacetate internucleoside linkages (US Pat. App. No. 2005/0106598; Sheehan et al, Nucleic Acids Res (2003); 31(14):4109-18). In other embodiments, a cyclobutyl ring replaces the naturally occurring furanosyl ring.

In other embodiments of non-classical nucleotide analogues suitable for use in methods and compositions of the present invention, the base is modified. A representative, non-limiting list of modified nucleobases includes 5-methylcytosine (5-me-C), 5-hydroxymethyl cytosine, xanthine, hypoxanthine, 2-aminoadenine, 6-methyl and other alkyl derivatives of adenine and guanine, 2-propyl and other alkyl derivatives of adenine and guanine, 2-thiouracil, 2- thiothymine and 2-thiocytosine, 5-halouracil and cytosine, 5-propynyl ( ^" --C- --CH

uracil and cytosine and other alkynyl derivatives of pyrimidine bases, 6-azo uracil, cytosine and thymine, 5-uracil(pseudouracil), 4-thiouracil, 8-halo, 8-amino, 8-thiol, 8-thioalkyl, 8-hydroxyl and other 8-substituted adenines and guanines, 5-halo particularly 5-bromo, 5-trifluoromethyl and other 5-substituted uracils and cytosines, 7-methylguanine and 7-methyladenine, 2-F- adenine, 2-amino-adenine, 8-azaguanine and 8-azaadenine, 7-deazaguanine and 7-deazaadenine and 3-deazaguanine and 3-deazaadenine. Further modified nucleobases include tricyclic pyrimidines such as phenoxazine cytidine (lH-pyrimido(5,4-b)(l,4)benzoxazin-2(3H)-one), phenothiazine cytidine (lH-pyrimido(5,4-b)(l,4)benzothiazin-2(3H)-one), G-clamps such as a substituted phenoxazine cytidine (e.g. 9-(2-aminoethoxy)-H-pyrimido(5,4-b)(l,4)benzoxazin- 2(3H)-one), carbazole cytidine (2H-pyrimido(4,5-b)indol-2-one), and pyridoindole cytidine (H- pyrido(3',2':4,5)pyrrolo(2,3-d)pyrimidin-2-one). Modified nucleobases may also include those in which the purine or pyrimidine base is replaced with other heterocycles, for example 7- deaza-adenine, 7-deazaguanosine, 2-aminopyridine and 2-pyridone. Further nucleobases include those disclosed in U.S. Pat. No. 3,687,808; those disclosed in The Concise Encyclopedia Of Polymer Science And Engineering, pages 858-859, Kroschwitz, J. I., ed. John Wiley & Sons, 1990; those disclosed by Englisch et al., Angewandte Chemie, International Edition, 1991, 30, 613; and those disclosed by Sanghvi, Y. S., Chapter 15, Antisense Research and Applications, pages 289-302, Crooke, S. T. and Lebleu, B., ed., CRC Press, 1993. Certain of these nucleobases are known to those skilled in the art as suitable for increasing the binding affinity of the compounds of the invention. These include 5-substituted pyrimidines, 6-azapyrimidines and N-2, N-6 and 0-6 substituted purines, including 2-aminopropyladenine, 5-propynyluracil and 5-propynylcytosine. Modified nucleobases and their use are described, inter alia, in United States Pat. Nos. US 3,687,808, US 4,845,205; US 5,130,302; US 5,134,066; US 5,175,273; US 5,367,066; US 5,432,272; US 5,457,187; US 5,459,255; US 5,484,908; US 5,502,177; US 5,525,711; US 5,552,540; US 5,587,469; US 5,594,121, US 5,596,091; US 5,614,617; US 5,645,985; US 5,830,653; US 5,763,588; US 6,005,096; US 5,681,941; and US 5,750,692.

Another non-limiting example of non-classical nucleotide analogues suitable for use in methods and compositions of the present invention are polycyclic heterocyclic compounds in place of one or more of the naturally-occurring heterocyclic base moieties. A number of tricyclic heterocyclic compounds have been previously reported. Such compounds are routinely used in antisense applications to increase the binding properties of the modified strand to a target strand. The most studied modifications are targeted to guanosines hence they have been termed G-clamps or cytidine analogs. Representative cytosine analogs that make 3 hydrogen bonds with a guanosine in a second strand include l,3-diazaphenoxazine-2-one (Kurchavov, et al., Nucleosides and Nucleotides, 1997, 16, 1837-1846), l,3-diazaphenothiazine-2-one, (Lin, K.-Y.; Jones, R. J.; Matteucci, M. J. Am. Chem. Soc. 1995, 117, 3873-3874) and 6,7,8,9-tetrafluoro- l,3-diazaphenoxazine-2-one (Wang, J.; Lin, K.-Y., Matteucci, M. Tetrahedron Lett. 1998, 39, 8385-8388). Incorporated into oligonucleotides these base modifications were shown to hybridize with complementary guanine and the latter was also shown to hybridize with adenine and to enhance helical thermal stability by extended stacking interactions.

Other examples of non-classical nucleotide analogues suitable for use in methods and compositions of the present invention are described in US Application Publication No. 2003/0207804 and 2003/0175906.

In other embodiments, an RNAi agent used in methods and compositions of the present invention is conjugated to a cholesterol moiety.

In still other embodiments, an RNAi agent used in methods and compositions of the present invention is conjugated to an a-tocopherol moiety. 2'-QMe Modifications

In more specific embodiments, an RNAi agent used in methods and compositions of the present invention is chemically modified with a modification selected from 2'-OMe and 2'-0- methoxy on at least one residue of at least one strand thereof. In other, more specific embodiments, the RNAi agent is chemically modified with 2'-OMe or 2'-0-methoxy on at least one residue of each of its 2 strands. In still more specific embodiments, the modification is 2'- OMe. In yet more specific embodiments, the 2'-OMe modification is present on at least one residue of each of its 2 strands. In even more specific embodiments, four 2'-OMe modifications are present on each strand of the RNAi agent. Alternatively, four 2'-OMe modifications are present on one strand of the RNAi agent, while the other strand contains no modified bases (deoxyribonucleotides naturally occurring in DNA are not considered modified bases for these purposes). In still other embodiments, eight 2'-OMe modifications are present on one strand of the RNAi agent, while the other strand contains four 2'-OMe modifications.

In yet other embodiments, an RNAi agent used in methods and compositions of the present invention is chemically modified with 2'-F.

In more particular embodiments, the sequences of the sense and antisense strands of the RNAi agent of a DDD of methods and compositions of the present invention are:

5'- GUoUoGGAGCUGAUGGCGoUoAGdTdT (SEQ ID No: 37) and

5'-CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 38), respectively. As used herein, "o" preceding a residue denotes a 2'-0-Methyl modification of said residue.

Alternatively, the sequences of the sense and antisense strands of the RNAi agent are: 5'-GUoUGGAGCoUGAoUGGCGoUAG (SEQ ID No: 39) and

5'-CoUACGCoCAUoCAGCUCoCAAC (SEQ ID No: 40), respectively.

Another aspect of the present invention provides an RNAi agent, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'-GUoUoGGAGCUGAUGGCGoUoAGdTdT (SEQ ID No: 37) and 5'-CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 38), respectively.

Yet another aspect of the present invention provides an RNAi agent, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'-GUoUGGAGCoUGAoUGGCGoUAG (SEQ ID No: 39) and

5'-CoUACGCoCAUoCAGCUCoCAAC (SEQ ID No: 40), respectively. Another aspect of the present invention provides a pharmaceutical composition comprising one of the above RNAi agents. Millimeter-scale implant/matrix drug delivery devices

The drug delivery device of the present invention may be a cylinder, in other embodiments a sphere, and in other embodiments, another shape suitable for an implant.

In certain preferred embodiments, a biodegradable matrix used in methods and compositions of the present invention is a polymeric biodegradable matrix. "Polymeric" as used herein refers to the characteristic of comprising a polymer. In more specific embodiments, the biodegradable matrix is held together by the structure of the polymer.

In other embodiments, a biodegradable matrix used in methods and compositions of the present invention is a biocompatible biodegradable matrix. "Biocompatible" as used herein refers to the characteristic of not inducing toxic effects, either as a result of reactivity with the human immune system or body tissues, or the production of toxic degradation by-products. In more specific embodiments, the biodegradable matrix is a polymeric, biocompatible biodegradable matrix.

In yet other embodiments, a DDD of the present invention is biostable. "Millimeter-scale", as used herein, refers to a device whose smallest diameter is at least 0.24 mm (fitting within the inner diameter of a 25-gauge needle). In certain preferred embodiments, each of the dimensions (diameter, in the case of a sphere or cylinder; and height and/or width or length, in the case of a cylinder, box-like structure, cube, or other shape with flat walls) is between 0.3 - 10 mm, inclusive. In more preferred embodiments, each dimension is between 0.5 - 8 mm, inclusive. In more preferred embodiments, each dimension is between 0.6 - 1.3 mm, inclusive.

In more preferred embodiments, the device is a cylinder, having a diameter 0.68 mm +/- 0.45 mm and a length of 4.5 +/- 1.5 mm. In other embodiments, the cylinder has a diameter of 0.1- 2.5 mm, inclusive, more preferably 0.2-2.0 mm, more preferably 0.3-1.5 mm, more preferably 0.38-0.98 mm, more preferably 0.48-0.88 mm, more preferably 0.53-0.83 mm, more preferably 0.58-0.78 mm, more preferably 0.63-0.73 mm, more preferably 0.68 mm. In other preferred embodiments, the cylinder has a length of 0.5-3.5 mm, inclusive, more preferably 0.7-3.0 mm, more preferably 0.9-2.5 mm, more preferably 1.0-2.0 mm, more preferably 1.1-1.5, more preferably 1.3 mm. In other embodiments, the cylinder has a diameter of 0.38-0.98 mm, inclusive and a length of 0.5-3.5 mm, inclusive. In more preferred embodiments, the diameter is 0.48-0.88 mm, and the length is 0.7-3.0 mm. In more preferred embodiments, the diameter is 0.53-0.83 mm, and the length is 0.9-2.5 mm. In more preferred embodiments, the diameter is 0.58-0.78 mm, and the length is 1.0-2.0 mm. In more preferred embodiments, the diameter is 0.63-0.73 mm, and the length is 1.1-1.5 mm. In more preferred embodiments, the diameter is 0.68 mm, and the length is 1.3 mm. In still other embodiments, the device is a suitable size for administration with a endoscope and a 19-gauge needle.

In other embodiments, the volume of the device is between 0.1 mm 3 and 1000 mm 3.

In other embodiments, the w/w agent:polymer load ratio above 1:100. In more preferred embodiments, the load is above 1:20. In more preferred embodiments, the load is above 1:9. In more preferred embodiments, the load is above 1:3 In other embodiments, the device is a technology described in US Patent Application Pub. No. 2011/0195123, the contents of which are incorporated herein by reference.

The DDD in some embodiments comprises degradable polymers, wherein the release mechanism includes both bulk erosion and diffusion; or in some embodiments, non degradable, or slowly degraded polymers are used, wherein the predominant release mechanism is diffusion rather than bulk erosion, so that the outer part functions as membrane, and its internal part functions as a drug reservoir, which practically is not affected by the surroundings for an extended period (for example from about a week to about a few months). Combinations of different polymers with different release mechanisms may also optionally be used. The concentration gradient at the surface is preferably maintained effectively constant during a significant period of the total drug releasing period, and therefore the diffusion rate is effectively constant (termed "zero mode" diffusion). The term "constant" refers to a diffusion rate that is preferably maintained above the lower threshold of therapeutic effectiveness, but which may still optionally feature an initial burst and/or fluctuate, for example increasing and decreasing to a certain degree. In other embodiments, there is an initial burst of less than 10% of the total amount of drug, which may be considered negligible. In other embodiments, there is an initial burst of about 20% of the total amount of drug. In other embodiments, the design enables initial a strong burst of 30% or more of the total amount of drug. The diffusion rate is preferably so maintained for a prolonged period, and it can be considered constant to a certain level to optimize the therapeutically effective period, for example the effective silencing period. These embodiments are described in US Patent Application Pub. No. 2011/0195123. In other embodiments, a DDD used in methods and compositions of the present invention further comprises another active agent. In certain embodiments, the additional active agent is another RNAi agent, for example ALN-VSP (Alnylam Inc, Cambridge, MA). Other examples of additional active agents that may be used are chemotherapy drugs, for example mAB-based agents, cytotoxic drugs, and kinase inhibitors.

Biodegradable matrices

In certain embodiments, the biodegradable matrix present in the drug delivery device comprises poly(lactic acid) (PLA). In other embodiments, the biodegradable matrix comprises poly(glycolic acid) (PGA). In other embodiments, the biodegradable matrix comprises both PLA and PGA (known as poly(lactic-co-glycolic acid) or PLGA). In other embodiments the matrix comprises a natural polymer comprising β-( 1-4) -linked D-glucosamine and N-acetyl-D- glucosamine. In a more specific embodiment, the matrix comprises Chitosan (NCBI PubChem ID #71853).

Methods for making PLGA matrices that incorporate RNAi agents are well known to those skilled in the art. Exemplary methods are described in described in US Patent Application Pub. No. 2011/0195123, the contents of which are incorporated herein by reference— for example in Examples 1.1 and 1.2 thereof.

In more specific embodiments, the PLA/PGA ratio is at least 75:25, more preferably between 75:25 and 95:5 inclusive. In other embodiments, the ratio is between 75:25 and 85:15 inclusive. In yet other embodiments, the ratio is no more than 25:75, more preferably between 25:75 and 5:95 inclusive. In other embodiments, the ratio is between 25:75 and 15:85 inclusive.

In other embodiments, the polymer comprises both PLA and PEG (poly(ethylene glycol)).

In other embodiments, tri-block PLA-PCL-PLA is used to form the matrix. PCL denotes poly- caprolactone.

In other embodiments, Poly(D,L-lactide) (DL-PLA), poly(D,L-glycolide), or poly(D,L-lactide- co-glycolide) is used, each of which is considered a separate embodiment.

Design of biodegradable controlled drug-delivery millimeter- scale DDD containing PLA, PGA, PEG, and/or PCL to have a specified release profile is well within the ability of those skilled in the art in light of the present disclosure. For example, general principles applicable to nano-scale and micro-scale particle carriers containing PLA, PGA, PEG, and/or PCL are described inter alia in Makadia and Siegel; and Park et al.

In another embodiment, the polymer is a polymer described in paragraphs 0076-0078 of US Patent Application Pub. No. 2011/0195123, the contents of which are incorporated herein by reference.

In other embodiments, the biodegradable matrix further comprises an additive for modulating hydrophilic -hydrophobic interactions; in other embodiments for enabling dispersion of the drug and eliminating aggregation; in other embodiments for preserving the drug in hot- temperature or cold-temperature storage conditions, for example 55°C and -20°C, respectively; in other embodiments for facilitating creation of cavities in the implant that affect to drug diffusion from the matrix. Hydrophilic -hydrophobic interactions may cause aggregation of the active substance in cases of hydrophilic active substances, such as siRNA, incorporated within a hydrophobic polymer, resulting in aggregation during production or subsequently when the device is implanted into the body of a subject and it is subjected for example to hydrolysis. Non-limiting examples of such additives are open monosaccharides, for example mannitol; disaccharides such as trehalose; sorbitol; and other cyclic monosaccharides such as glucose, fructose, galactose and disaccharides such as sucrose. The above additives that are chiral may be in the form of the D-enantiomer, the L- enantiomer, or a racemic mixture. In other embodiments, more than one additive is present. In other embodiments, the biodegradable matrix further comprises an additive for preserving the drug within the implant after implantation against low pH. The microenvironment in the implant interior tends to be acidic. Unlike chemotherapy, the pH should preferably be maintained above a threshold. For example while doxorubicin is stable in an acidic environment, with minimal hydrolytic degradation within a pH range of 3 to 6.5, RNAi agents can degrade at pH <3. In more specific embodiments, this additive may be selected from bicarbonate and carbonate, for example sodium bicarbonate and sodium carbonate. In some embodiments, one or more additives against decrease of pH are included in the matrix to control the degradation of the polymer, typically degradation by hydrolysis, thereby to control the release. Such an additive is designed with or without regard to the drug. Non limiting examples include magnesium carbonate, calcium carbonate, calcium hydroxyapatite, sodium bicarbonate and in general salts. In some embodiments, the salt distribution within the polymeric matrix is non-homogenous. In some embodiments, the salt distribution is lower toward the matrix surface and higher toward the center. For example, in some embodiments the concentration of salt in the center is at least 20% higher than the concentration at the matrix edge.

Suitable release profiles

In preferred embodiments, a device of the present invention is designed to release the active agent in a controlled fashion. It will be apparent to those of skill in the art, in light of the knowledge in the art taken together with the information provided herein, that the PLA:PGA ratio, the design of the polymer end group, for example a capped terminal end group and an uncapped carboxylic end group, the composition and additives, the molecular weight (MW) of the polymer, and the surface-to-volume ratio of the implant may be adjusted to achieve a particular release profile. For example, deviating the PLA:PGA ratio from 50:50, increasing the MW, or reducing the surface-to-volume ratio can increase the release time.

In other embodiments, the DDD of the present invention is designed with a particular release profile. One relevant parameter is the time point at which 90% of the active agent has been released. In some embodiments, a DDD of the present invention releases 90% of the active agent over a time period between 3-24 months inclusive. In other embodiments, the time point of release of 90% of the active agent is between 3-12 months inclusive. In other embodiments, the time point is between 2-24 months inclusive. In other embodiments, the time point is between 2-15 months inclusive. Another relevant parameter is the percent of active agent released at a given time point. For example, in some embodiments, 80-99% inclusive of the active agent is released at the 3-month timepoint. In other embodiments, 80-99% inclusive of the active agent is released at the 2-month timepoint or the 4-month, 6-month, 9-month, 12- month, or 24-month timepoint, each of which is considered a separate embodiment.

Alternatively or in addition, in some embodiments no more than 30-50% of the active agent of a DDD of the present invention is released during the first 3 weeks. Dosage and drug percentage

A DDD of the present invention may, in certain embodiments, contain at least 10 μg siRNA. In more specific embodiments, the amount is between 10-2000 μg (inclusive) siRNA per device.

In still more specific embodiments, the amount is between 100-1500 (inclusive) μg, more preferably 150-1000 μg, more preferably 200-470 μg, more preferably 330-420 μg, more preferably 350-400 μg, more preferably 375 μg siRNA per device. In certain embodiments, 2-8

DDD, containing a total of 0.7-3.5 mg siRNA, are implanted into a tumor of a human subject. In certain embodiments 8-24 DDD are implanted into a tumor of a human subject, to improve distribution of the released drug within the tumor. In certain embodiments the number of DDDs per tumor is X, where X is between 8-24 inclusive, and the drug load of each DDD is 3.0/X mg +/-20% (for example X=10 and each DDD has 0.3mg +/-20%) In certain embodiments, the drug content of a device of the present invention is at least 10% by weight. Alternatively, the drug content of the device is at least 20%. In still other embodiments, the drug content of the device is at least 11%, at least 12%, at least 13%, at least 14%, at least 15%, at least 16%, at least 17%, at least 18%, at least 19%, at least 21%, at least 22%, at least 23%, at least 24%, at least 25%, at least 26%, at least 28%, or at least 30%, or at least 33%, or at least 36%, or at least 40%.

In another embodiment of a millimeter- scale DDD of the present invention, (a) the biodegradable matrix comprises PLGA, a sugar, (non-limiting embodiments of which are mannitol, trehalose, sucrose, and sorbitol), a pH modulating additive (non-limiting embodiments of which are sodium bicarbonate, magnesium carbonate and calcium carbonate); and (b) the nucleotide sequence of the sense strand of the RNAi agent is GUUGGAGCUGAUGGCGUAGdTdT (SEQ ID No: 41), and the sequence of the antisense strand is CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 42). In other embodiments, an additional siRNA, for example against a metastasis-facilitating gene, is also present within the DDD. Some exemplary formulations of drug delivery devices of methods and compositions of the present invention are as follows, by weight: 16-32 % siRNA, 7-11% Mannitol, 60-77% PLGA, and 0.5%-3% sodium bicarbonate and other excipients. In more specific embodiments, the DDD is 20% +/- 5% siRNA by weight.

Additional Features In other embodiments, a DDD of the present invention is non homogeneous. In other embodiments, a DDD of the present invention is multi-layered. In other embodiments, a DDD of the present invention comprises a biodegradable polymeric component attached to a bio- stable polymeric component and/or to a metallic component. In other embodiments, a DDD of the present invention is coated. A coating can be designed in some embodiments for a number of characteristics, including modulating the release rate or preventing protein stickiness during long-term storage. The coating in other embodiments comprises the same material or a similar material used to form the matrix. In other embodiments, the coating is a PLGA matrix containing no siRNA, or in other embodiments consists of PLA, or of PGA. In other embodiments, the coating further comprises another material, for example PEG.

Therapeutic Methods Another aspect of the present invention provides a method of treating a patient having metastatic cancer, comprising the step of administering to the primary tumor(s) a millimeter- scale DDD, wherein the DDD comprises:

A. A biodegradable matrix; and

B. An RNAi (RNA interference) agent incorporated within the biodegradable matrix, wherein the RNAi agent comprises a duplex region, and the nucleotide sequence of the duplex region of the sense strand consists of a sequence selected from SEQ ID No: 1-7, namely GUUGGAGCUGAUGGCG (SEQ ID No: 1), GUUGGAGCUGUUGGCG (SEQ ID No: 2), GUUGGAGCUGCUGGCG (SEQ ID No: 3), GUUGGAGCUAGUGGCG (SEQ ID No: 4), GUUGGAGCUUGUGGCG (SEQ ID No: 5), GUUGGAGCUGGUGACG (SEQ ID No: 6), and

GUUGGAGCUGGUUGCG (SEQ ID No: 7), either alone or followed by: a sequence selected from: (i) UAGGCAAGAGUGCC (SEQ ID No: 8) and (b) a 5'- fragment of 1-13 nucleotides inclusive of SEQ ID No: 8.

In more specific embodiments, the DDD is any of the embodiments thereof mentioned herein, each of which is considered to be a separate embodiment. In still more specific embodiments, the biodegradable matrix of the DDD comprises PLGA, mannitol (for example D-mannitol), and sodium bicarbonate; the nucleotide sequence of the sense strand of the RNAi agent of the DDD is GUUGGAGCUGAUGGCGUAGdTdT (SEQ ID No: 41), and the sequence of the antisense strand of the RNAi agent is CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 42). In more specific embodiments, the DDD contains 16-32% siRNA by weight, 7-11% Mannitol, and 60-77% PLGA, + 0.5%-3% sodium bicarbonate and other excipients. In more specific embodiments, the DDD is 20% +/- 5% siRNA by weight. In other embodiments, an additional siRNA, for example against a metastasis-facilitating gene, is also present within the DDD.

Another aspect of the present invention provides a method of treating a patient having metastatic cancer, comprising the step of administering to the primary tumor(s) a pharmaceutical composition comprising an RNAi agent, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'- GUoUoGGAGCUGAUGGCGoUoAGdTdT (SEQ ID No: 37); and

5'-CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 38), respectively. In other embodiments, the aforementioned method further comprises administration of a chemotherapy drug such as gemcitabine. In other embodiments, the method further comprises radiation treatment.

Another aspect of the present invention provides a method of treating a patient having metastatic cancer, comprising the step of administering to the primary tumor(s) a pharmaceutical composition comprising an RNAi agent, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'-GUoUGGAGCoUGAoUGGCGoUAG (SEQ ID No: 39); and

5'-CoUACGCoCAUoCAGCUCoCAAC (SEQ ID No: 40), respectively.

In other embodiments, the aforementioned method further comprises administration of a chemotherapy drug such as gemcitabine. In other embodiments, the method further comprises radiation treatment.

In other embodiments, the metastatic cancer treated by a method of the present invention is selected from pancreatic tumor, colon tumor, and lung tumor. In more specific embodiments, the cancer is selected from pancreatic carcinoma, pancreatic ductal adenocarcinoma, small-cell lung carcinoma, and colorectal cancer. In even more specific embodiments, the cancer is pancreatic ductal adenocarcinoma. In some embodiments, the tumor is an inoperable tumor; alternatively, it is an operable tumor.

In some preferred embodiments, a device is implanted intratumorally. In still other embodiments, the device is implanted into the vicinity of the tumor. In more specific embodiments, in the case of a well-defined solid tumor, several devices are spaced within the tumor volume. In yet other embodiments, several devices are implanted along a needle cavity within the tumor. In still other embodiments, the device or devices are implanted such that they are not in a direct contact with the perimeter of the tumor. Alternatively, in the case of a poorly defined solid tumor, the device is inserted into an area believed to contain tumor cells. In still other embodiments, a method of the present invention further comprises the step of administering an anti-cancer agent to the patient. In more specific embodiments, the anticancer agent comprises a pyrimidine analogue, non-limiting examples of which are 5-azacytidine, 5-aza-2'-deoxycytidine, 5-fluoro-uracil, 5-fluoro-deoxyuridine (floxuridine), and 5-fhiorodeoxyuridine monophosphate. In more specific embodiments, the anti-cancer agent is an inhibitor of ribonucleoside-diphosphate reductase large subunit (EC 1.17.4.1), non-limiting examples of which are motexafin gadolinium (CHEBI: 50161); hydroxyurea; gemcitabine (2',2'-difluorodeoxycytidine); elacytarabine (CP-4055; an ara-C-5'elaidic-acid-ester) and CP- 4126, (CO 1.01; a gemcitabine-5'elaidic-acid-ester; Adema AD et al, Metabolism and accumulation of the lipophilic deoxy nucleoside analogs elacytarabine and CP -4126. Invest New Drugs. 2011 Oct 15. [Epub ahead of print]), and those described in WO2011062503, the contents of which are incorporated herein by reference. In even more specific embodiments, the anti-cancer agent comprises gemcitabine. In alternative embodiments, the anti-cancer agent is gemcitabine. In yet other embodiments, the anti-cancer agent is an EGFR tyrosine kinase inhibitor. In yet other embodiments, the anti-cancer agent comprises a thymidylate synthase inhibitor. In more specific embodiments, the anti-cancer agent comprises leucovorin (Folinic acid; 2-[[4-[(2-amino-5-formyl-4-oxo-l,6,7,8-tetrahydropteridin-6- yl)methylamino]benzoyl] aminojpentanedioic acid). In yet other embodiments, the anti-cancer agent comprises irinotecan. In yet other embodiments, the anti-cancer agent comprises oxaliplatin. In still other embodiments, the anti-cancer agent comprises FOLFIRIN (5-fluorouracil, leucovorin, and irinotecan in combination). In still other embodiments, the anti-cancer agent is FOLFIRINOX (5-fluorouracil, leucovorin, irinotecan, and oxaliplatin in combination), or any combination of a subset of the four agents in FOLFIRINOX. In one embodiments the agent is Modified FOLFIRINOX, administrated as follows: Oxaliplatin (85mg/m ) - IV for 2 hours, immediately followed by Irinotecan (180 mg/m 2 ) - IV for 90 min. Leucovorin - 400mg/m 2 , followed by a Fluorouracil continuous IV infusion of 2,400mg/m (over 46 hours) every two weeks. In yet other embodiments, the anti-cancer agent comprises an EGFR tyrosine kinase inhibitor. In more specific embodiments, the anti-cancer agent is Erlotinib.

In some embodiments the Gemcitabine is administrated as follows 1,000 mg/m2 IV is given on a weekly basis (Dayl, day7 & day21), (cycle = 4 weeks).

In other embodiments, the anti-cancer agent is a Protein-bound paclitaxel, including an agent based on nanoparticle albumin-bound (nab) technology, including Abraxane In some embodiments, the anti-cancer agent is administered to the patient after administration of the DDD. In more specific embodiments, the anti-cancer agent may be administered to the patient up to 10 days after administration of the DDD. Alternatively, the anti-cancer agent is administered to the patient simultaneously with administration of the DDD. In still other embodiments, the anti-cancer agent is administered to the patient before administration of the DDD. In yet other embodiments, the DDD is administered during ongoing administration of the anti-cancer agent. In still other embodiments, the anti-cancer agent is administered to the patient intratumorally, by a method such as injection and controlled release, or including in other embodiments administration from the same DDD. In still other embodiments, a method of the present invention further comprises the step of administering radiation therapy to the patient. In some embodiments, the radiation is administered to the patient after administration of the DDD. In more specific embodiments, the radiation may be administered to the patient up to 10 days after administration of the DDD. Alternatively, the radiation is administered to the patient simultaneously with administration of the DDD. In still other embodiments, the radiation is administered to the patient before administration of the DDD. In yet other embodiments, the DDD is administered during ongoing administration of the radiation.

Pharmaceutical compositions

In another embodiment, the present invention provides a pharmaceutical composition for treating a patient having metastatic cancer, said pharmaceutical composition comprising a millimeter- scale DDD, said DDD comprising:

A. A biodegradable matrix; and

B. An RNAi (RNA interference) agent incorporated within the biodegradable matrix, wherein the RNAi agent comprises a duplex region, and the nucleotide sequence of the duplex region of the sense strand consists of: a sequence selected from SEQ ID No: 1-7, namely GUUGGAGCUGAUGGCG (SEQ ID No: 1), GUUGGAGCUGUUGGCG (SEQ ID No: 2), GUUGGAGCUGCUGGCG (SEQ ID No: 3), GUUGGAGCUAGUGGCG (SEQ ID No: 4), GUUGGAGCUUGUGGCG (SEQ ID No: 5), GUUGGAGCUGGUGACG (SEQ ID No: 6), and GUUGGAGCUGGUUGCG (SEQ ID No: 7), either alone or followed by: a sequence selected from: (i) UAGGCAAGAGUGCC (SEQ ID No: 8) and (b) a 5'- fragment of 1-13 nucleotides inclusive of SEQ ID No: 8.

In more specific embodiments, the DDD is any of the embodiments thereof mentioned herein, each of which is considered to be a separate embodiment. In still more specific embodiments, the biodegradable matrix of the DDD comprises PLGA, mannitol (for example D-mannitol), and sodium bicarbonate; the nucleotide sequence of the sense strand of the RNAi agent of the DDD is GUUGGAGCUGAUGGCGUAGdTdT (SEQ ID No: 41), and the sequence of the antisense strand of the RNAi agent is CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 42). In other embodiments, an additional siRNA, for example against a metastasis-facilitating gene, is also present within the DDD.

In another embodiment, the present invention provides a pharmaceutical composition for treating a patient having metastatic cancer, said pharmaceutical composition comprising an RNAi agent, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'- GUoUoGGAGCUGAUGGCGoUoAGdTdT (SEQ ID No: 37); and 5'-CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 38), respectively.

In other embodiments, the aforementioned pharmaceutical composition further comprises a chemotherapy drug such as gemcitabine. In other embodiments, the pharmaceutical composition is indicated as an adjunct for radiation.

In another embodiment, the present invention provides a pharmaceutical composition for treating a patient having metastatic cancer, said pharmaceutical composition comprising an RNAi agent, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'-GUoUGGAGCoUGAoUGGCGoUAG (SEQ ID No: 39); and

5'-CoUACGCoCAUoCAGCUCoCAAC (SEQ ID No: 40), respectively.

In other embodiments, an additional siRNA, for example against a metastasis-facilitating gene, is also present within the DDD.

In other embodiments, the aforementioned pharmaceutical composition further comprises a chemotherapy drug such as gemcitabine. In other embodiments, the pharmaceutical composition is indicated as an adjunct for radiation. In another embodiment, the present invention provides use of a millimeter- scale DDD in the preparation of a medicament for treating a patient having metastatic cancer, wherein said DDD comprises:

A. A biodegradable matrix; and B. An RNAi (RNA interference) agent incorporated within the biodegradable matrix, wherein the RNAi agent comprises a duplex region, and the nucleotide sequence of the duplex region of the sense strand consists of: a sequence selected from SEQ ID No: 1-7, namely GUUGGAGCUGAUGGCG (SEQ ID No: 1), GUUGGAGCUGUUGGCG (SEQ ID No: 2), GUUGGAGCUGCUGGCG (SEQ ID No: 3), GUUGGAGCUAGUGGCG (SEQ ID No: 4), GUUGGAGCUUGUGGCG

(SEQ ID No: 5), GUUGGAGCUGGUGACG (SEQ ID No: 6), and GUUGGAGCUGGUUGCG (SEQ ID No: 7), either alone or followed by: a sequence selected from: (i) UAGGCAAGAGUGCC (SEQ ID No: 8) and (b) a 5'- fragment of 1-13 nucleotides inclusive of SEQ ID No: 8. In other embodiments, an additional siRNA, for example against a metastasis-facilitating gene, is also present within the DDD.

In more specific embodiments, the DDD is any of the embodiments thereof mentioned herein, each of which is considered to be a separate embodiment. In still more specific embodiments, the biodegradable matrix of the DDD comprises PLGA, mannitol (for example D-mannitol), and sodium bicarbonate; the nucleotide sequence of the sense strand of the RNAi agent of the DDD is GUUGGAGCUGAUGGCGUAGdTdT (SEQ ID No: 41), and the sequence of the antisense strand of the RNAi agent is CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 42).

In another embodiment, the present invention provides use of an RNAi agent in the preparation of a medicament for treating a patient having metastatic cancer, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'- GUoUoGGAGCUGAUGGCGoUoAGdTdT (SEQ ID No: 37); and

5'-CUACGCCAUCAGCUCCAACdTdT (SEQ ID No: 38), respectively.

In other embodiments, an additional siRNA, for example against a metastasis-facilitating gene, is also present within the DDD. In other embodiments, the aforementioned RNAi agent and a chemotherapy drug such as gemcitabine are both used in the preparation of a medicament for treating a patient having metastatic cancer.

In another embodiment, the present invention provides use of an RNAi agent in the preparation of a medicament for treating a patient having metastatic cancer, wherein the sequences of the sense and antisense strands of the RNAi agent are:

5'-GUoUGGAGCoUGAoUGGCGoUAG (SEQ ID No: 39); and

5'-CoUACGCoCAUoCAGCUCoCAAC (SEQ ID No: 40), respectively.

In other embodiments, the aforementioned RNAi agent and a chemotherapy drug such as gemcitabine are both used in the preparation of a medicament for treating a patient having metastatic cancer.

In other embodiments, the metastatic cancer treated using a pharmaceutical composition of the present invention is selected from pancreatic tumor, colon tumor, lung tumor, brain, liver, kidney, prostate, melanoma, endometrial carcinoma, gastric carcinoma, renal carcinoma, biliary carcinoma, cervical carcinoma, and bladder carcinoma. In more specific embodiments, the cancer is selected from pancreatic carcinoma, pancreatic ductal adenocarcinoma, small-cell lung carcinoma, and colorectal cancer. In even more specific embodiments, the cancer is pancreatic ductal adenocarcinoma. In some embodiments, the tumor is an inoperable tumor; alternatively, it is an operable tumor.

In still other embodiments, a pharmaceutical composition of the present invention further comprises an anti-cancer agent. Alternatively, the pharmaceutical composition is indicated for administration in combination with an anti-cancer agent. In more specific embodiments, the anti-cancer agent is a pyrimidine analogue, non-limiting examples of which are mentioned herein. In even more specific embodiments, the anti-cancer agent comprises gemcitabine. In alternative embodiments, the anti-cancer agent is gemcitabine. In yet other embodiments, the anti-cancer agent is an EGFR tyrosine kinase inhibitor. In yet other embodiments, the anticancer agent comprises a thymidylate synthase inhibitor. In more specific embodiments, the anti-cancer agent comprises leucovorin (Folinic acid; 2-[[4-[(2-amino-5-formyl-4-oxo-l,6,7,8- tetrahydropteridin-6-yl)methylamino]benzoyl] aminojpentanedioic acid). In yet other embodiments, the anti-cancer agent comprises irinotecan. In yet other embodiments, the anticancer agent comprises oxaliplatin. In still other embodiments, the anti-cancer agent comprises FOLFIRIN (5-fluorouracil, leucovorin, and irinotecan in combination). In still other embodiments, the anti-cancer agent is FOLFIRINOX (5-fluorouracil, leucovorin, irinotecan, and oxaliplatin in combination). In yet other embodiments, the anti-cancer agent comprises an EGFR tyrosine kinase inhibitor. In more specific embodiments, the anti-cancer agent is Erlotinib.

In some embodiments, the anti-cancer agent is administered to the patient after administration of the DDD. In more specific embodiments, the anti-cancer agent may be administered to the patient up to 10 days after administration of the DDD. Alternatively, the anti-cancer agent is administered to the patient simultaneously with administration of the DDD. In still other embodiments, the anti-cancer agent is administered to the patient before administration of the DDD. In yet other embodiments, the DDD is administered during ongoing administration of the anti-cancer agent.

In yet other embodiments, a pharmaceutical composition of the present invention is indicated for administration with radiation therapy. In some embodiments, the radiation is administered to the patient after administration of the DDD. In more specific embodiments, the radiation may be administered to the patient up to 10 days after administration of the DDD. Alternatively, the radiation is administered to the patient simultaneously with administration of the DDD. In still other embodiments, the radiation is administered to the patient before administration of the DDD. In yet other embodiments, the DDD is administered during ongoing administration of the radiation.

Also provided herein are the following sequences that can be applied in siRNA design against the indicated targets, where the (typically 2 different) siRNA design contain the sense and anti- sense strands, in a precise manner as follows, per target. While only the sense strand is shown, the antisense strand is complementary, as depicted below: siRNA seq (SLUG):

1. sense: Sense: 5'- GUU UGC AAG AUC UGC GGC A dTdT -3' (SEQ ID No: 83).

Antisense 5'- UGC CGC AGA UCU UGC AAA C dTdT -3' (SEQ ID No: 84).

In other embodiments, the below siRNA comprises the following sense strand a) can be linked to additional nucleotides at the 5' end and/ or at the 3' end, provided that (b) the overall strand length is 19-30 nucleotides, inclusive, c) the antisense strand is designed in full complimentarity in the overlapped region within the duplex, provided that d) the antisense need not be the same length as the sense strand is fully complementary, and the length of the antisense is in the range of 19-30 inclusive, e) and in some embodiments the 3' end includes a dTdT end; and f) some embodiments one or more nucleotides of the siRNA are modified: SLUG = SNAI2 siRNA seq: siRNA seq:

sense: 5'- GUU UGC AAG AUC UGC GGC A -3' (SEQ ID No: 69). sense 5'- CUG GUC AAG AAG CAU UUC A -3' (SEQ ID No: 70). TWIST sense 5'- GUC UGC AGC UCU CGC CCA A -3' (SEQ ID No: 71). sense 5'- G GUG UGC GUC CAG CCG UUG -3' (SEQ ID No: 72).

Zebl sense: 5'- GAC UCG AGC AUU UAG ACA C -3' (SEQ ID No: 73). sense 5'- CAG GUG UAA GCG CAG AAA G -3' (SEQ ID No: 74).

Zeb 2: sense 5'- GUC AUU AGA AGA GGC GUA A -3' (SEQ ID No: 75). sense 5'- CAU UAG AAG AGG CGU AAC A -3' (SEQ ID No: 76). N-Cadherin Cadherin = CDH2 NM 001792.3 sense 5'- GUA CAA UAU GAG AGC AGU G -3' (SEQ ID No: 77). sense 5'- CAG CUG GAC UUG AUC GAG A -3' (SEQ ID No: 78). Fibronectin (FN1) - targets splice variants 1, 3, 4, 5, 6, 7:

NM_002026.2; NM_212482.1 ; NM_212478.1; NM_212476.1; NM_212474.1; NM_054034.2 sense 5'- CAC UGA UUG CAC UUC UGA G -3' (SEQ ID No: 79). Vimentin (vim) NM 003380.3 1. sense 5'- GUC UAA CGG UUU CCC CUA A -3' (SEQ ID No: 80).

2. sense 5'- GAG AAA UUG CAG GAG GAG -3' (SEQ ID No: 81).

Sanil = Snai 2 NM 003068.4

Sense 5'- CAG AUG UCA AGA AGU ACC A -3' for 1-2 (SEQ ID No: 82).

In various embodiments, each strand of any of the above siRNA's contains a dTdT overhang, as shown hereinabove in the aforementioned sense and antisense SLUG sequences.

EXPERIMENTAL DETAILS SECTION Example 1 : Schematic depiction of the described treatment.

Fig. 1 shows a schematic depiction of the described treatment of metastatic cancer. Local and prolonged delivery of RNAi-based treatment within a primary tumor is administered, often in parallel to systemic administration of drugs including chemotherapy drugs

Example 2: LODER-derived siG12D significantly inhibits growth of pancreatic cancer cells in vitro:

Methods

DDD's containing RNAi molecules were produced in a biological-class hood in a clean room, as follows:

Step 1: Preparation of siRNA/D-Mannitol/ Sodium Bicarbonate mixture: siRNA was added to the pre-weighed D-Mannitol and Sodium Bicarbonate, and they were dissolved in RNase-free sterile water.

Step 2: Freezing: The liquid was placed into glass vials, frozen in dry ice, and lyophilized for 48 hours.

Step 3: PLGA Preparation: PLGA was dissolved in Ethyl Acetate.

Step 4: Combining PLGA with D-Mannitol/Sodium Bicarbonate/siRNA: The PLGA solution was poured into the glass vial containing the lyophilized D-Mannitol/Sodium Bicarbonate/siRNA in fractions and stirred until homogenization Step 5: Solvent evaporation. The solution was poured into a Teflon-covered dedicated glass dish and left to evaporate inside a dedicated container for 3-5 days.

Step 6: Release of film: The film was released from the glass dish using tweezers and a scalpel. Step 7: Excision of individual DDDs: Individual DDDs were excised using a dedicated puncher. Each DDD was of a cylindrical shape, with a length and diameter of -1.3 mm and -0.6 mm, respectively.

In other embodiments, steps 5-7 are replaced by direct injection of solution into a tube, for example a PTFE tube, where the injection amount per tube is calculated for example to comprise siRNA in the range of 200 μg-500 μg. The tube is put in a drying chamber, for example drying chambers of the group including heat/humidity chamber, vacuum chamber (Desiccator), hood and centrifuges.

Results

The described DDD's have been demonstrated to enable prolonged release of siRNA while protecting the siRNA embedded in the LODER from degradation (WO 2010/001325 to Shemi). The silencing potential of the siRNA sequence set forth in SEQ ID No: 41-42, directed against K-ras ^siG12D (siG12D), was determined; this sequence was used as the siG12D siRNA sequence throughout the Examples below. Panel cells, human pancreatic carcinoma cells that carry the siG12D mutation in the K-ras gene, were mock-transfected or transfected with siG12D or control scrambled siRNA (si-scr). The relative K-ras mRNA levels were assessed by Real-Time PCR, with Hypoxantine PhosphoRybosyl Transferase (HPRT) and Ubiquitin C (UBC) used as endogenous controls. As shown in Figure 2A, siG12D specifically reduced K- ras mRNA levels, both wt and mutant (data not shown) with a half-maximal inhibitory concentration (IC ₅₀ ) of 67pM. We confirmed the site-specific, siG12D-directed, cleavage of the K-ras message, by Rapid Amplification of cDNA Ends (RACE).

Next, we assessed the effect of the LODER-derived siG12D on K-ras protein levels using Pancl-LUC cells, which are a sub-clone of Panel cells, stably expressing the lucif erase reporter gene. Pancl-LUC cells were incubated in the presence of the siG12D LODER or with a LODER containing a control siRNA (siGFP) and transfection reagent (Lipofectamine). In parallel, two groups of cells were directly transfected with free siG12D or siGFP for comparison. Cells were lysed after 48 or 72 hrs, and total K-ras protein level was assessed by Western blot analysis. Figure 2B-C shows that the siG12D LODER significantly decreased the level of K-ras protein (by 2.7-fold) at only 48 hrs after treatment. This decrease was augmented in a time-dependent manner (by 9-fold after 72 hrs). Similar results were obtained using original Panel cells. These results show that the LODER-driven siRNA efficiently silenced target gene expression.

Inhibition of K-ras expression leads to the inhibition of growth of Panel pancreatic cancer cells. In order to examine the potential growth-inhibiting effect of siG12D, we measured both cell viability and cell death of Panel cells treated with the siG12D LODER. Figures 2E- F demonstrate a time-dependent decrease in cell viability, which correlated with an increase in cell death in the presence of the siG12D LODER compared to the siGFP LODER.

Panel -LUC cells were further used to: i) confirm the ability to follow cell growth by monitoring Luciferase activity and ii) to demonstrate the functionality of the LODER-derived siRNA. Pancl-LUC cells were incubated with the siG12D LODER or the siLUC LODER (positive control) or the siGFP LODER (negative control). The cells were also transfected in parallel with free siG12D, siLUC or siGFP for comparison to LODER-embedded siRNA (data not shown). 72 and 96 hours later, cells were harvested and Luciferase expression was measured. Figure 2D demonstrates a significant time-dependent inhibition of Luciferase expression in the presence of either the siLUC LODER or the siG12D LODER after 72hrs (50% and 31% reduction, respectively) and 96 hrs (65% and 32% reduction, respectively) when compared to the siGFP LODER. These results show that LODERs are able to efficiently release functional siRNA in a manner comparable to transfection with free siRNA. In addition, the results confirm the feasibility of using luciferase expression to monitor Pancl-LUC cell growth for in vivo assessment.

Example 3: Functionality of the siRNA LODER in vivo: For the in vivo assessment of LODER-siRNA effects, we initially used the murine colon cancer CT26 cells. These cells were stably transfected to constitutively express the luciferase gene (CT26-LUC cells). CT26-LUC cells were injected subcutaneously into BALB/c mice. When tumor volume reached ~\ cm , two siLuc or control (siGFP) LODERs were implanted into the tumors. As seen in Figure 3A, luciferase expression increased by approximately 3-fold in the siGFP LODER-implanted mice, while in mice implanted with the siLuc LODERs, the luciferase levels barely increased. Figure 3B reveals that there was no significant effect on tumor weight, at sacrifice, in the group treated with the siLuc LODER in comparison with the siGFP LODER group, indicating that the lower Luciferase levels were not caused by an antitumor effect; rather, the siLuc released from the LODER inhibited luciferase expression, as measured by its activity in vivo.

We also assessed the ability of the LODER-driven siRNA to inhibit a target gene in normal tissue. We implanted an empty DDD or siLuc DDD into the livers of transgenic mice, expressing the luciferase gene (MUP-Luc) in their liver. The Luciferase levels were measured using a CCCD (a Cooled Charge Coupled Device) camera. Figure 3C reveals that siLuc LODERs led to a significant decrease in the luciferase levels compared to the empty LODER. These results further confirm that siRNA is released from the LODER and silences target gene in vivo.

To study the toxic potential of the LODER-driven siRNA, we compared the effect of the siG12D LODER (16, 32 and 64 μg siG12D/kg weight) to that of the empty LODER and to sham-operated or untreated animals. LODERs were implanted into the pancreas of normal mice and rats, both males and females of both species. In parallel, the same and higher doses of siG12D were intraperitoneally injected into normal mice and rats (16, 160 and 320 μg siG12D/kg weight). The duration of this study was 2 weeks in mice and 8 weeks in rats. The results demonstrated that the siG12D LODER had no effect on animal mortality, behavior, on body and liver weight. Hematology and biochemistry tests failed to reveal any statistically significant differences between the tested groups. Gross and histopathology analyses revealed that all noted changes were of minimal severity, and typical in untreated mice or rats of the same age and strain. Based on the lack of adverse reaction following siRNA treatment, the maximum dose of 0.32mg siRNA/kg body weight was considered as not constituting an acute toxicity risk.

Example 4. LODER-driven siG12D inhibits tumor growth in vivo: We next aimed to assess the ability of LODERs to inhibit growth of tumors derived from human pancreatic tumor cell lines in vivo. Towards that aim we used two human pancreatic tumor cells: Panel and Capanl, both constitutively expressing the luciferase (LUC) gene, referred to as Pancl-LUC and Capanl-LUC. These cell lines bear mutated K-ras genes: K- ras G12D (G12D) and K-ras G12V (G12V), respectively, both mutations leading to the constitutive activation of the K-ras protein. To evaluate the therapeutic potential of the LODER-driven K- ras oncogene silencing, mice bearing subcutaneous tumors originating from Panel or Capanl cells were randomly assigned to treatment groups when the tumors reached an average volume of 1cm . Mice bearing Capanl-LUC-derived subcutaneous tumors were divided into 4 groups: untreated and intratumoral implantation of siG12D LODERs, siG12V LODERs or empty LODERs. Each siRNA-loaded LODER contained 2 μg of siRNA. In order to further validate the effect of siG12D LODERs on tumor growth, we used a syngeneic mouse tumor model whereby the Panc02 mouse pancreatic cell line that bears a G12D mutation was subcutaneously injected into mice of matching strain (Corbett et al., 1984). In vitro, siG12D inhibited Panc02 cell growth to a similar extent as Panel cells. In vivo, in the subcutaneous syngeneic model, implanted siG12D LODERs eliminated almost entirely the tumor tissue by one week after the implantation (Figure 4A). In some siG12D LODER-treated cases, the estimated size of necrotic areas revealed that, it exceeded over 90% of total tumor tissue area (Figure 4B).

In an effort to determine the long-term potential effect of the siRNA-LODERs, we analyzed the remaining siRNA content of LODERs isolated from tumors after implantation. The results revealed that the LODER's structure protected the siG12D from degradation for greater than a 70 day period. These results demonstrate that the siG12D LODER effectively inhibited tumor growth in vivo.

Example 5: The anti-tumor effect of the siG12D LODER in an orthotopic pancreatic model

In order to extend the results of the subcutaneous tumor model experiments, we established an orthotopic pancreatic cancer model. For this purpose, Pancl-LUC or Capanl-LUC cells were injected into the tail of the mouse pancreas via an incision. Tumor development was assessed by in vivo Luciferase activity. When tumors were detected, mice were stratified and divided into treatment groups, keeping a similar average level of Luciferase activity between the groups. Two LODERs were stitched to the pancreatic tumor via a laparotomy.

To provide evidence that tumor growth inhibition occurred due to K-ras silencing, we assessed the effect of the siG12D LODER on K-ras protein expression in tumor tissues. K-ras protein was detected by immunohistochemistry (IH) staining in pancreatic tumor tissues 24 days after treatment with siG12D LODERs, empty LODERs, or untreated. Image analysis software was used to calculate the percentage of K-ras positively-stained cells vs. non-stained cells, in 0.1 x 0.1 mm ² squares at progressive distances from the LODER placement sites (Figure 5A-D). Overall, K-ras staining in siG12D LODER-treated tumors was weaker compared to empty LODER or untreated tumors (representative tissue section shown in Figure 5A). The ratio of K-ras positively-stained vs. non-stained cells (positive/negative ratio) in siG12D LODER- treated tumors was approximately 10-fold lower compared to the two other groups. This finding indicates that at approximately one month after LODER implantation, the effect can be observed at a radius of at least 1 mm away from the LODER location. Thus, siG12D molecules are released from LODERs, spread in tumor tissue, enter tumor cells, and lead to silencing of K-ras expression.

To corroborate these results, we assessed the effect of the siG12D LODER on the K-ras pathway. Functional K-ras induces the phosphorylation of Akt and ERK. Here, tumor tissues were tested from mice bearing pancreatic Pancl-LUC-derived tumors from 4 treatment groups: untreated, injected with siG12D or implanted with empty LODERs or siG12D LODERs. Thirty nine days post- treatment, we performed IH analysis of tumor tissues for Akt and ERK. Akt phosphorylation was induced in all treatment groups, except for the untreated group. We calculated the ratio of positively P-ERK-stained vs. non-stained (positive/negative ratio) nuclei in 0.1 x 0.1 mm squares at progressive distances from LODER placement site. The positive/negative ratio was low in areas adjacent to siG12D LODERs, compared to distant areas and the control group. These results are in line with the K-ras staining results, showing that K-ras silencing by the siG12D LODER leads to the suppression of its downstream signaling pathway as well.

In order to confirm that silencing K-ras in the pancreatic tumor inhibits tumor cell growth, we performed CDC47 IH staining. We estimated the percentage of positively stained cells using image analysis software. The results revealed that the siG12D LODER inhibited the growth of tumor cells around the LODER, and this effect extended up to 2 mm far from the LODER. These results confirm that the observed tumor growth inhibition was due to the silencing of K- ras by the siG12D LODER.

Thus, LODER-driven siG12D significantly inhibited in vivo growth of pancreatic tumors in both subcutaneous and orthotopic mouse models.

Example 6: K-ras expression in human pancreatic tissue.

Over 90% of human pancreatic ductal adenocarcinoma (PDA) tumors involve mutations in the K-ras oncogene, and PDA is believed to be addicted to the K-ras mutation. Next we investigated K-ras expression and localization in normal pancreatic tissues and pancreatic tumors. K-ras IH staining revealed that while in normal tissue the K-ras protein was found in the cytoplasm, in the tumor tissue, K-ras was localized on the membrane of PanIN (pancreatic intraepithelial neoplasia) lesions (Figure 6). K-ras staining could clearly depict tumor cell invasions into tissue determined pathologically as the "normal tumor border".

Example 7: The effect of the siG12D LQDER on epithelial-to-mesenchymal transition

In the course of tumor development, cancer cells may undergo epithelial-to-mesenchymal transition (EMT). In this process, the cells acquire mesenchymal characteristics and lose epithelial ones. It was previously reported that K-ras inhibition reduces EMT, as manifested by a decrease in cell migration and reduction in contact inhibition Singh et al 2009.

In order to study the effects of siG12D on the EMT process and its consequent increased cellular migration and motility, we conducted a series of experiments in which we examined different aspects of this process. We analyzed the migration characteristics of Panel cells following transfection with siG12D or scrambled siRNA, using the scratch and transwell- migration assays. TGF-β was used as a positive control. Migration ability was assessed using a modified Boyden chamber assay, which revealed that siG12D inhibited Panel migration by more than 30% (Figure 7a). To test migration through a scratch assay. Panc-1 cells treated with K-RAS (G12D) siRNA or scrambled siRNA were seeded in 24W plates and scratched with plastic tips, migration was observed using light microscopy. The scratch assay revealed that after 24 hours, the non-treated cells and cells transfected with scrambled siRNA migrated and narrowed the gap created by the scratch, whereas the siG12D transfected cells barely migrated (Figure 7B and Figure 8). Treatment of siG12D in vitro was followed by measurement of genes whose expression (high or low) is characteristic of EMT. Treatment resulted in reduced expression of mesenchymal genes (Zebl and Slug) and augmented expression of epithelial genes (E-Cadherin), whereas TGFbeta treated cells (EMT positive control) exhibited gene expression characteristic of EMT, and untreated cells and cells treated with scrambled siRNA, HI 9 siRNA, exhibited more neutral gene expression (Figure 9).

Example 8: A model for metastatic pancreatic cancer

Procedure: Female C57/B6 mice (age ~6w) were intrapancreatically injected with 100,000 Panc02 cells. Two months post-injection, mice were sacrificed and liver, spleen and lung tissues were macroscopically examined for the presence of metastases (Figure 10). Example 9: siRNA sequence of genes shown to induce EMT, namely SLUG, Twist, Zebl, and

Zeb2.

Down-regulation of a gene from a group including these genes is accomplished by siRNA encompassed in LODER and released within the tumor along an extended period. The effects of these LODER encapsulated siRNA will be tested using in in vivo mouse models of pancreatic cancer, such as Panel and Panc02 orthotopic and subcutaneous cell grafts.

Example 10: Clinical study

We have translated our preclinical studies including in vitro and in vivo results and initiated a Phase I clinical study (NCT01188785). The aim of this phase I study is to assess safety, tolerability and to determine preliminary therapeutic benefits. Here we present a case of a 66- year old female patient with locally advanced, non-operable PDA). One 25μ§ siG12D LODER was implanted into the tumor under endoscopic-ultrasound (EUS) guidance. The patient initiated Gemcitabine chemotherapy 16 days later. The treatment was very well tolerated, with no significant side-effects. The serum-based tumor-marker CA 19-9 decreased following the siG12D LODER implantation even before the administration of the first line chemotherapy treatment, and eventually reached normal values. The patient then received local radiation treatment. As shown in Figure 11, a CT scan performed of siG12D showed a significant reduction in tumor size nine months post-insertion (bottom) compared to the time of insertion (top). This patient has survived more than 18 months after the siG12D LODER implantation. These findings show the safety, tolerability, and clinical benefit of the approach for patients with locally advanced non-operable PDA.

Example 11: Comparison of overall survival of patients with Locally Advanced Pancreatic Cancer (LAPC), receiving standard-of-care (SOC) with or without TNFerade (Herman et al 2013) vs. siG12D-LODER in combination with Gemcitabine or FOLFIRINOX. As shown in Figure 12, the described DDD's exhibited superior prolongation of life than the SOC.

Example 12: Measurement of ability of siG12D-LODER to reduce metastasis formation in mice with pancreatic cancer

Experimental design: Mice bearing injected pancreatic Panc02 tumors, stably expressing luciferase, are stratified into groups according to luciferase measurements and treated in a time window before the appearance of metastases as follows: • Group 1: 2 EMPTY LODERS are implanted into/adjacent to orthotopic tumors.

• Group 2: 2 siG12D LODERS are implanted into/adjacent to orthotopic tumors.

Tumor growth follow-up is performed by luciferase measurements. The rate of occurrence and size of metastases is measured macroscopically post mortem. Panc02-luc is a sub-clone of the murine pancreatic carcinoma cell line Panc02 (produced from tumors from K-ras mutated transgenic mice) which stably expresses the luciferase reporter gene. Panc02-luc cells can be grown in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (FCS) and 100 units/ml penicillin/streptomycin, according to the ATCC protocol. Mycoplasma tests can be performed after cell defrost and prior to cell injection to assure that the cell line is mycoplasma free.

Mice: Female C57/B6 mice, 6-7 week old, may be purchased from Harlan, Israel.

Tumor implantation

Mice are anesthetized, and the target skin is shaved and sterilized. An incision is made in the median upper abdominal region, the pancreas is carefully pulled out, and an inoculum (e.g. 100,000) Panc02-luc cells are injected into the tail of pancreas. The pancreas is returned into the abdominal space, and the incision is stitched.

Tumor growth follow-up

Tumor growth is assessed by Luciferase measurements using the IVIS camera. When tumor size reaches ~0.5cm in diameter, mice are divided into several groups that are stratified according to Luciferase intensity levels.

Calibration of appearance of metastasis

1. Mice are implanted with tumors as described above.

2. Groups of mice are sacrificed at various times past tumor implantation.

3. Macroscopic and histological analyses (e.g. H&E sections) of the liver, spleen and lungs are performed to determine metastasis formation.

The time chosen for LODER treatment will be decided based on the time after tumor implantation when metastases are first detected in at least 20% of the mice. Treatment groups

Mice with tumors are divided into treatment groups at a time point before metastases first appear (according to the calibration test).

Group 1 : Two EMPTY LODERS implanted into/adjacent to orthotopic tumors.

Group 2: Two siG12D LODERS (5\ig siRNA) implanted into/adjacent to orthotopic tumors.

LODER implantation

1. Mice will be anesthetized and incised and the pancreas pulled out as described above.

2. For tumors bigger than 0.5cm in diameter, the surgeon implants the appropriate

LODERs into the tumor and then stitches the insertion site.

3. For tumors smaller than 0.5cm in diameter, the surgeon affixes the appropriate

LODERs to the tumor using a needle.

4. The pancreas is returned into the abdominal space

5. The incision is stitched.

Post-treatment follow-up and analysis

1. Primary tumor growth is assessed, e.g. by luciferase measurement. Luciferase can be visualized by injection of Luciferin and measured using an IVIS camera. Visualization is performed before treatment (in order to divide mice into groups) and at various time points after treatment.

2. Macroscopic analysis: At the time when mice are expected to bear metastases, mice are sacrificed, and the liver, spleen and lung tissues are examined. The diameter of visible metastases is determined, e.g. using a caliper.

3. Histology analysis: Liver, spleen and lung tissues are fixed (e.g. with formalin), embedded in paraffin, and tissue slides are prepared and stained with Hematoxylin and Eosin (H&E). Tissue scan for microscopic metastases is performed by a pathologist.

We expect to see a lower count of macroscopic and microscopic metastases in the treatment group compared to the control. Metastases forming in the treatment group may also be smaller than those in the control group. These results will indicate a metastasis formation inhibitory effect of siG12D LODER treatment. Example 13: Assessment of ability of siG12D LQDER to prolong the survival of mice with

Panc02 tumors and metastases

General experimental design: Similar to the previous Example. Past treatment follow-up and analyses 1. Primary tumor growth: similar to the previous Example.

2. Overall survival: Following treatment, mice are monitored daily for survival. Severely morbid mice may be euthanized. A survival curve will be produced.

3. Histology analysis: similar to the previous Example. May also include analysis of proliferation markers, e.g. Ki67 and pHH3. Tissue scan for microscopic metastases may be performed by a pathologist.

We expect to see an increased survival in the treatment group compared to the control. Determining the metastatic load by post-mortem examination of liver, spleen and lung tissue will show whether death and morbidity were due to metastases or to the primary tumor. A decrease in proliferation markers in metastases in the treatment group will indicate that tumor local treatment exerts a systemic effect on disseminated metastases.

Example 14: Measuring levels of pancreatic cancer circulating tumor cells (CTCs) in subjects' blood following treatment with siG12D LQDER

Protocol: Sample collection is typically performed within 24h of patients' blood extraction. Whole blood sample is washed with ECL (erythrocyte lysis) buffer until red cells are removed. White blood depletion step: CD47 depletion may be performed using a kit. RNA preparation from CTCs: RNA may be prepared using the Tri-Zol (Invitrogen) reagent.

Relative quantitation of CTC markers: CTC markers relative level is assessed, e.g. by Real- Time PCR, using markers such as EGFR and CK19. Results may be compared to previously collected data from the same patient. Follow-up radiology: Patients receiving siG12D LODER treatment or chemotherapy without LODER (control group) are subjected to imaging (e.g. by CT scan) of the original tumor and metastases during a follow-up period. The data are analyzed by a radiologist to monitor disease progression. Example 15: 2'-QMe Modification of siG12D

Materials and Experimental Methods

Stability studies l-μΐ of modified siRNA was incubated at with 9-μ1 50% human serum, at 37°C. The incubation was stopped at various time points (0 min, 10 min, 30 min, 4h, 24h, 1 week, 2 weeks and 4 weeks) by the addition of ΙΟμΙ sample buffer and freezing at -70°C. siRNA degradation was evaluated using urea-acrylamide gel-electrophoresis. Gels were stained with ethidium bromide, and band intensity was measured using Image Gauge™ software. Stability was evaluated by calculating the ratio between the intensity at each time point vs. the 0-min time point. Values were classified into five stability groups: Level 0: very unstable; 95-100% degraded; Level 1: unstable; 75-95% degraded; Level 2: medium stability; 50-75% degraded; Level 3: stable; 25-50% degraded; Level 4: very stable. 0-25% degraded.

Viability studies

Pancl-luc cells were seeded in 96-well plate (0.5 x 10 ⁴ cells/well) to 70% confluence and were transfected with siRNAs. 120 hrs later, a 6-hr viability test was performed using the XTT assay. The indicated 4 levels of staining, namely 1, 2, 3, and 4 as shown in the last column of Figure 14, related to levels of efficacy.

Results siG12D was modified with 2'-oxymethyl ("2'-OMe") at various positions, as set forth in Table 1 below.

Table 1: 2-O-Methyl modifications of siG12D. "o" denotes 2-O-Methyl modification.

5 Sense: 5' GUoUoGGAGCUGAUGGCGoUoAGdTdT 49-50

Anti-sense: 5' CUoACGCCAUCAGCUCCAoAoCdTdT

6 Sense: 5' GUoUoGGAGCUGAUoGGCGoUoAGdTdT 51-52

Anti-sense: 5' CUoACGCoCAUCAGCUCCAoAoCdTdT

7 Sense: 5' GoUoUoGGAGCoUoGAoUoGGCGoUoAGdTdT 53-54

Anti-sense: 5' CUACGCCAUCAGCUCCAACdTdT

8 Sense: 5' oGUoUGoGAoGCoUGoAUoGGoCGoUAoGdTdT 55-56

Anti-sense: 5' CoUAoCGoCCoAUCAGCoUCoCAoACdTdT

9 Sense: 5' oGUoUGoGAoGCoUGoAUoGGoCGoUAoG 57-58

Anti-sense: 5' CoUAoCGoCCoAUCAGCoUCoCAoAC

10 Sense: 5' GoUoUoGGAGCoUoGoAoUoGGCGoUoAG 59-60

Anti-sense: 5' CUACGCCAUCAGCUCCAAC

11 Sense: 5' GoUoUoGGAGCoUoGoAoUoGGCGUAG 61-62

Anti-sense: 5' CUACGCCAUCAGCUCCAAC

12 Sense: 5' oGoUoUoGGAGCoUoGoAoUoGGCGUAoG 63-64

Anti-sense: 5' CUACGCCAUCAGCUCCAAC

13 Sense: 5' oGoUoUGGAGCoUoGoAoUoGGCGoUoAoG 65-66

Anti-sense: 5' CUAoCoGoCCAUCAGCoUoCoCAAC

14 Sense: 5' GUoUGGAGCoUGAoUGGCGoUAG 39-40

Anti-sense: 5' CoUACGCoCAUoCAGCUCoCAAC

15 Sense: 5' GoUoUoGGAGCoUoGAoUoGGCGoUoAG 67-68

Anti-sense: 5' CoUACGCoCAUCAGCUCoCAAC

The modified siRNA were tested for stability in serum, and several of them exhibited improved stability relative to the unmodified siRNA. Next, their effect on the viability of a pancreatic cancer cell line was tested in cell culture. Several of the modified siRNA exhibited improved efficacy relative to the unmodified siRNA (Figure 13). The data is summarized in Figure 14. Example 16: Effect of modified and unmodified siG12D molecules alone and with Gemcitabine

This study compared the effects on cell viability of specific modified and unmodified siG12D (using non-specific siRNA as a control), alone and combined with the chemotherapy drug Gemcitabine (Gemzar®). Experimental procedure: Panel cells were transfected using Lipofectamine™ with unmodified and modified siG12D and a scrambled (non-targeting) siRNA control molecule or were mock transfected (no siRNA). Cell viability was assessed 72 hrs post-transfection by the Methylene Blue (MB) test. Additional groups were transfected as described above, followed by lhr of treatment with ΙΟμΜ Gemzar®. At the exemplary time point of 72h, Gemzar® enhanced the effect of siG12D by nearly 10% (Figure 15). Non-specific siRNA also reduced cell viability (possibly via the TLR3/interferon immune response pathway). Additionally, some modified siRNAs were able to achieve results comparable to un-modified siRNA. Other time points, for example 48h, 72h, 96h and 120h, may be utilized in additional studies. Example 17: Assessment of the effect of siG12D on EMT in pancreatic cells

Additional studies of EMT may utilize various mesenchymal markers (e.g. N-cadherin, Vimentin, S100A4, a- smooth-muscle cell actin [SMA]) and epithelial markers (e.g. E- cadherin, cytokeratin 8 or other cytokeratins, ZO-1, claudins, occludins, snail, slug, ZEB1/2, and Twistl/2) to monitor the EMT process. Example 18: Comparison of Chemotherapy alone vs. Chemotherapy + siG12D-DDD

Patients with metastatic pancreatic cancer are administered 1 or more courses of Chemotherapy alone, for example Gemcitabine and/or FOLFIRINOX, or siG12D-DDD implantation followed by the same Chemotherapy treatment as the first group, beginning 0-120 days after DDD implantation. Disease progression is assessed, for example as described herein. siG12D-DDD is shown to be an effective adjunct to chemotherapy.

Example 19: Preparation of DDD's containing siRNA's against mutated K-ras and an EMT- related gene

DDD's containing siRNA's against mutated K-ras and an EMT-related gene are prepared as described hereinabove, except that the different siRNA's are mixed together prior to mixing with the other solid ingredients.

Example 20: Testing of DDD's containing siRNA's against mutated K-ras and an EMT-related gene

DDD's containing siRNA's against mutated K-ras and an EMT-related gene are tested as described hereinabove. With respect to the jurisdictions allowing it, all patents, patent applications, and publications mentioned herein, both supra and infra, are incorporated herein by reference.

Example 21 : Assessment of ability of siG12D LQDER to decrease the presence of circulating tumor cells (CTC) in mice with Panc02 tumors. General experimental design: C57/B6 mice bearing injected pancreatic and subcutaneous Panc02 tumors, stably expressing luciferase, are stratified into groups according to luciferase measurements and treated post-metastases formation as follows:

• Group 1: 2 EMPTY LODERS will be implanted into/adjacent to orthotopic tumors

• Group 2: 2 siG12D LODERS will be implanted into/adjacent to orthotopic tumors. Tumor growth follow-up is performed by luciferase measurements. The presence of circulating tumor cells in the blood of these mice is determined using quantitative PCR (qPCR) and droplet digital PCR) ddPCR.

When tumor size reaches ~0.5cm in diameter, mice are divided into groups stratified according to Luciferase intensity levels, up to 20 animals in each group (according to the survival following surgery). For subcutaneous tumors, growth may be measured using a caliper. When tumor size reaches ~lcm in diameter, mice are divided into groups stratified according to Luciferase intensity levels.

In other experiments, the treatment groups are divided into treatment groups at the time before metastases first appear (according to the calibration test performed in Example 15). Treatment groups are:

• Group 1: Two EMPTY LODERS implanted adjacent to orthotopic tumors or implanted into subcutaneous tumors.

• Group 2: Two siG12D LODERS (5μg siRNA) implanted into/adjacent to orthotopic tumors. Analysis

1. At certain time points post-treatment, mice are euthanized and blood is extracted. 2. CTC along with leukocytes (the blood nucleated cell fraction) is enriched using Erythrocyte Lysis Buffer (ELB).

3. RNA is isolated, e.g. using Tri-Zol (Invitrogen) reagent. cDNA is prepared, e.g. using the qScript kit (Quanta).

4. CTC markers relative level quantification: relative levels of CTC markers are assessed, e.g. by Real-Time PCR and ddPCR. Marker list may include: EGFR, CK19 and Luciferase (an artificial marker enabled by the genetic labeling of Panc02 cells with the lucif erase gene).

Expected results: By this experiment we plan to see a decrease (in either the orthotopic or the subcutaneous models) in the levels of CTC. If such a decrease will be met, this could imply that the ability of pancreatic tumors in the Panc02 model to disseminate tumor cells into the blood is impaired. Dissemination of tumor cells into the blood is an essential step in the formation of distant organ metastasis. Impairment of this procedure is highly likely to impair metastasis formation.

Example 22: Assessment of ability of siG12D LQDER to decrease the distant organ colonization of disseminated tumor cells.

General experimental design: C57/B6 mice bearing injected subcutaneous Panc02 non-labeled tumors are stratified into groups with equal tumor average size and treated as follows:

• Group 1: 2 EMPTY LODERS will be implanted into/adjacent to orthotopic tumors

• Group 2: 2 siG12D LODERS will be implanted into/adjacent to orthotopic tumors.

At 4 days post-treatment, Panc02 cells labeled with luciferase are injected intravenously. Growth of Panc02-luc cells in the lungs is monitored through luciferase measurements.

Treatment groups:

Mice with tumors are divided into treatment groups at the time before metastases first appear.

• Group 1: Two EMPTY LODERS will be implanted into/adjacent to orthotopic tumors.

• Group 2: Two siG12D LODERS (5μg siRNA) will be implanted into/adjacent to orthotopic tumors. Expected results: By this experiment we plan to demonstrate that colonization of distant target organ with disseminated cells may be dependent on factors secreted from the primary tumor. Therefore, disruption of tumor growth with siG12D LODER may inhibit the colonization of distant organs with tumor cells and by this decrease the metastasis formation.

In the claims, the word "comprise", and variations thereof such as "comprises", "comprising", and the like indicate that the components listed are included, but not generally to the exclusion of other components.

References

Barry et al March 2013, Volume 30, Issue 3, pp 251-264)

Brummelkamp et al, Stable suppression of tumorigenicity by virus-mediated RNA Interference. Cancer Cell September · 2:243 (2002) Corbett et al, Induction and chemotherapeutic response of two transplantable ductal adenocarcinomas of the pancreas in C56BL/6, Cancer research 44,717-726, (1984).

Deng et al Cancer Research: April 15, 2010; Volume 70, Issue 8, Supplement 1.

Gupta and Massague, Cell 127, November 17, 2006.

Fleming et al, Molecular Consequences of Silencing Mutant K-ras in Pancreatic Cancer Cells: Justification for K-ras-Directed Therapy. Mol Cancer Res 3(7):413-23 (2005).

Loeher et al, Gemcitabine Alone Versus Gemcitabine Plus Radiotherapy in Patients With Locally Advanced Pancreatic Cancer: An Eastern Cooperative Oncology Group Trial

JOURNAL OF CLINICAL ONCOLOGY VOLUME 29 NUMBER 31 4105 (2011) Makadia and Siegel, Poly Lactic-co-Glycolic Acid (PLGA ) as Biodegradable Controlled Drug Delivery Carrier, Polymers 2011, 3, 1377-1397.

Morioka CY et al, Suppression of invasion of a hamster pancreatic cancer cell line by antisense oligonucleotides mutation-matched to K-ras gene. In Vivo. 2005 May- Jun;19(3):535-8. Normanno et al, Implications for K-ras status and EgFR-targeted therapies in metastatic cRc. Nat. Rev. Clin. Oncol. 6, 519-527 (2009).

Padhani AR, Ollivier L. The RECIST (Response Evaluation Criteria in Solid Tumors) criteria: implications for diagnostic radiologists. Br J Radiol 2001; 74: 983-986.

Park et al, Biodegradable Polymers for Microencapsulation of Drugs. Molecules 2005, 10: 146-161.

Rejiba et al, K-ras oncogene silencing strategy reduces tumor growth and enhances gemcitabine chemotherapy efficacy for pancreatic cancer treatment. Cancer Science 98(7): 1128-1136 (2007).

Singh et al, A gene expression signature associated with "K-Ras addiction" reveals regulators ofEMT and tumor cell survival. Cancer Cell. 2009 Jun 2;15(6):489-500. Wang W et al, Identification of effective siRNA against K-ras in human pancreatic cancer cell line MiaPaCa-2 by siRNA expression cassette. World J Gastroenterol. 2005 Apr 7;11(13):2026-31.

Siegel R, Naishadham D, Jemal A: Cancer, statistics, 2012. CA Cancer J Clin 62:10-29 Maitra A, Hruban RH: Pancreatic cancer. Annu Rev Pathol, 2008, 3:157-88 Valastyan and Weinberg, Cell. 2011, 147(2): 275-292) Kalluri and Weinberg, J Clin Invest. 2009 Jun; 119(6): 1420-8 Brabletz, T., Cancer Cell 22, December 11, 2012 Hotz B, et al, Clin Cancer Res 2007;13:4769-76 Jen- Jung Pan, Muh-Hwa Yang , J. of Gastrointestinal One, Vol 2, No 3 (September 2011) Rachagani, S. et al British Journal of Cancer 2011 104, 1038 - 1048

Andreia de Albuquerque et al., Oncology 2012;82:3-10Herman, J.M. et al, Journal Of Clinical Oncology Volume 31 number 7, (2013)