US20070010474A1 - NF-kappaB oligonucleotide decoy molecules

Info

Abstract

Description

Claims

US20070010474A1

Publication number: US20070010474A1
Application number: US11/521,714
Authority: US
Inventors: Leslie McEvoy; Christi Parham; Jie Zhang; Rolf Ehrhardt
Original assignee: Individual
Current assignee: Individual
Priority date: 2003-12-02
Filing date: 2006-09-14
Publication date: 2007-01-11
Also published as: WO2005056020A8; US20070078102A1; EP1691817A2; WO2005056020A3; JP2007512845A; US7378509B2; WO2005056020B1; US20050182012A1; WO2005056020A2

The present invention concerns double-stranded NF-κB decoy oligodeoxynucleotide (NF-κB dsODN) molecules that contain a core sequence capable of specific binding to an NF-κB transcription factor. In a particular aspect, the invention concerns NF-κB decoy molecules that preferentially bind p50/p65 and/or cRel/p50 heterodimers over p50/p50 homodimers. In another aspect, the invention concerns NF-κB decoy molecules with improved binding affinity to p65.

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation application of copending U.S. application Ser. No. 11/368,060 filed on Mar. 3, 2006, which is a continuation-in-part of U.S. application Ser. No. 11/004,512 filed on Dec. 2, 2004, which claims priority under 35 U.S.C. § 119(e)(1) of U.S. provisional patent application Ser. No. 60/526,623, filed on Dec. 2, 2003, and U.S. Provisional patent application Ser. No. 60/612,029, filed on Sep. 21, 2004, the entire disclosures of which are hereby expressly incorporated by reference.

BACKGROUND OF THE INVENTION

1. Field of the Invention
The present invention concerns NF-κB oligonucleotide decoy molecules, and their use in the treatment of various NF-κB related diseases and pathological conditions.
2. Description of the Related Art
NF-κB is a family of inducible dimeric transcription factors composed of members of the Rel family of DNA-binding proteins that recognize a common sequence motif. In its active DNA-binding form, NF-κB is a heterogeneous collection of dimers, composed of various combinations of members of the NF-κB/Rel family. At present, this family is composed of 5 members, termed p52, p50, p65, cRel and Rel B. The homology between the members of the Rel family is through the Rel homology domain, which is about 300 amino acids in size and constitutes the DNA-binding domain of these proteins.
Different NF-κB dimers exhibit different binding affinities for NF-κB sites bearing the consensus sequence GGGRNNYYCC (SEQ ID NO: 1) where R is purine, Y is pyrimidine and N is any base. The Rel proteins differ in their abilities to activate transcription, such that only p65/RelA and c-Rel were found to contain potent transcriptional-activation domains among the mammalian family members. NF-κB is found in its inactive form in the cytoplasm, where it is bound to the 43-kDa protein IKB that covers the nuclear localization signal region of the p65/p50 dimer. Activation of NF-κB starts with the proteolytic destruction of IKB followed by the transport of the RelA/p50 complex into the nucleus, where it binds to its recognition site on the DNA and activates transcription of target genes. For further review of the NFκB family see, for example, Gomez et al., Frontiers in Bioscience 2:49-60 (1997).
p52 and p50 do not contain transactivation domains. Dimers composed solely of p52 and/or p50 proteins that lack transcriptional activation domains are generally not activators of transcription and can mediate transcriptional repression.
The transcription factors of the Rel/NF-κB family are key regulators of immune and inflammatory responses, and contribute to lymphocyte proliferation, survival and oncogenesis. Thus, NF-κB plays a key role in the expression of several genes involved in the inflammation, cell proliferation and immune responses. (D'Acquisto et al., Gene Therapy 7: 1731-1737 (2000); Griesenbach et al., Gene Therapy 7, 306-313 (2000); Morishita et al., Gene Therapy 7: 1847-1852 (2000)). Among the genes regulated by NF-κB are many which play critical roles in various diseases and conditions, such as rheumatoid arthritis, systemic lupus erythematosus, restenosis, myocardial infarction, ischemia reperfusion injury, glomerulonephritis, atopic dermatitis, saphenous vein graft, Alzheimer's disease, to name a few. See, e.g. Khaled et al. Clinical Immunology and Immunopathology 86(2): 170-179 (1998); Morishita et al., Nature Medicine 3(8): 894-899 (1997); Cho-Chung et al., Current Opinion in Molecular Therapeutics 1(3): 386-392 (1999); Nakamura et al., Gene Therapy 9: 1221-1229 (2002); Shintani et al., Ann. Thorac. Surg. 74: 1132-1138 (2002); and Li et al., J. Neurochem. 74(1): 143-150 (2000).
NF-κB decoys have been proposed for the inhibition of neointimal hyperplasia after angioplasty, restenosis and myocardial infarction (Yoshimura et al., Gene Therapy 8: 1635-1642 (2001); Morishita et al., Nature Medicine 3(8): 894-899 (1997)). The greater inhibition of reperfusion injury, acute, and chronic rejection after transplantation results in a prolongation of allograft survival and decrease in graft coronary artery disease. (Feeley et al., Transplantation 70(11): 1560-1568 (2000)). In vivo transfection of an NF-κB decoy provides a novel strategy for treatment of acute myocarditis. (Yokoseki et al., supra). Ueno et al., supra reported that blocking NFκB by NFκB decoy prevented ischemia reperfusion injury in the heart.
It has been shown (Ziegler-Heitbrock et al, J. Leukoc. Biol. 55(10:73-80 (1994); Kastenbauer and Ziegler-Heitbrock, Infect. Immunol. 67(4):1553-9 (1999)) that when a human monocyte cell line, Mono Mac 6, was pre-treated for two days with low doses of lipopolysaccharide (LPS), the response to subsequent LPS stimulation was strongly reduced. Upon stimulation of these LPS-tolerant cells with LPS, these cells exhibited a predominance of the p50 homodimer as shown by the gel shift assay. The authors then tested the effect of the altered NF-κB complexes on gene expression via reporter gene analysis. NF-κB-dependent HIV-1 LTR reporter gene constructs were transfected into Mono Mac 6 cells, followed by pre-culture with and without LPS, and luciferase activity was measured. When LPS-tolerant cells were tested, LPS stimulation did not increase transactivation of the NF-κB-dependent HIV-1 LTR reporter gene. This indicates that the NF-κB complexes present in LPS-tolerant cells are functionally inactive. This also was applicable to the transcription of the NF-κB-controlled TNF gene. Using a TNF promoter-controlled luciferase reporter construct, LPS-tolerant cells showed only a minimal response to LPS stimulation. Therefore, it was concluded that the p50 homodimers induced by LPS tolerance lack transactivation activity. These p50 homodimers instead occupy the cognate NF-κB-binding sites and prevent transactivation and therefore transcription by the p50/p65 complex.

SUMMARY OF THE INVENTION

The present invention concerns double-stranded NF-κB decoy oligodeoxynucleotide (NF-κB dsODN) molecules that contain a core sequence capable of specific binding to an NF-κB transcription factor. In a particular aspect, the invention concerns NF-κB decoy molecules that preferentially bind p50/p65 and/or cRel/p50 heterodimers over p50/p50 homodimers when p50/p50 homodimers are present. The selective decoy molecules of the invention, by not blocking p50/p50 homodimers, allow these homodimers to block the promoters of NF-κB regulated genes, which provides an additional level of negative regulation of gene transcription. As a result, the selective NF-κB decoy molecules are particularly potent inhibitors of NF-κB activity both in vitro and in vivo.
If p50/p50 homodimers are not present, are present only in small amounts, or do not play a significant role in the disease state, the invention provides NF-κB decoy molecules with high binding affinity for p65.
In one aspect, the invention concerns double-stranded NF-κB decoy oligodeoxynucleotide (NF-κB dsODN) molecules that preferentially bind p50/p65 and/or cRel/p50 heterodimers relative to p50/p50 homodimers.
In another aspect, the invention concerns an NF-κB double-stranded decoy oligodeoxynucleotide (dsODN) molecule, comprising a sense and an antisense strand, which preferentially binds p50/p65 and/or cRel/p50 heterodimers over p50/p50 homodimers when p50/p50 homodimers are present and/or exhibits a p65 binding affinity of 45 or less, as determined by measuring the molar excess required to compete at least 50% of binding of p65/p50 in an electromobility shift assay to the non-mammalian NF-κB promoter from HIV (sequence 113/114).
In yet another aspect, the invention concerns The dsODN molecule of claim 1 characterized by a specificity/affinity factor of at least about 20, where the specificity/affinity factor is determined in a competitive binding assay, and is defined as follows:
Specificity/affinity factor=(S _p50/p50 −S _p65/p50)×S _p50/p50 /S _p65/p50
where S_p50/p50equals the molar excess of said dsODN molecule required to compete 50% of the binding of p50/p50 to the non-mammalian NF-κB promoter from HIV (sequence 113/114) and S_p65/p50equals the molar excess of said dsODN molecule required to compete 50% of the binding of p65/p50 to the non-mammalian NF-κB promoter from HIV (sequence 113/114), and wherein the score (S) is assigned as 100 if the decoy is unable to compete at least 50% of the binding at any molar ratio tested.
In another aspect, the present invention concerns NF-κB dsODN molecules that have a specificity/affinity factor of at least about 25, or at least about 30, or at least about 35, or at least about 40, or at least about 50 or at least about 60, or at least about 70, or at least about 80. In a preferred embodiment, the decoy molecules of the invention additionally show increased binding affinity to the p50/p65 heterodimers and/or have improved stability in vivo.
In a further aspect, the invention concerns a dsODN molecule comprising in its first strand, in 5′ to 3′ direction, a sequence of the formula FLANK1-CORE-FLANK2, wherein
CORE is selected from the group consisting of GGGATTTCC (SEQ ID NO: 11); GGACTTTCC (SEQ ID NO: 13); GGATTTCC (SEQ ID NO: 19); GGATTTCCC (SEQ ID NO: 21); and GGACTTTCCC (SEQ ID NO: 25);
FLANK1 is selected from the group consisting of AT; TC; CTC; AGTTGA (SEQ ID NO: 79), and TTGA (SEQ ID NO: 80);
FLANK2 is selected from the group consisting of GT; TC; TGT; AGGC (SEQ ID NO: 88); and AG.
In a specific embodiment, CORE is selected from the group consisting of GGGATTTCC (SEQ ID NO: 11); GGACTTTCC (SEQ ID NO: 13); and GGATTTCC (SEQ ID NO: 19); FLANK1 is AT and FLANK2 is GT; or FLANK1 is TC and FLANK2 is TC; or FLANK1 is CTC and FLANK2 is TGT; or FLANK1 is AGTTGA (SEQ ID NO: 79) and FLANK 2 is AGGC (SEQ ID NO: 88); or FLANK1 is TTGA and FLANK2 is AG.
In another specific embodiment, CORE is GGGATTTCC (SEQ ID NO: 11); or GGACTTTCC (SEQ ID NO: 13), FLANK1 is AGTTGA (SEQ ID NO: 79) and FLANK 2 is AGGC (SEQ ID NO: 88).
In yet another specific embodiment, The dsODN molecule of claim 14 wherein CORE is GGACTTTCC (SEQ ID NO: 13), FLANK1 is AGTTGA (SEQ ID NO: 79) and FLANK 2 is AGGC (SEQ ID NO: 88).
The NF-κB dsODN molecules include a second strand that is at least partially complementary to said first strand, and may have a phosphodiesterate, phosphorothioate, mixed phosphodiesterate-phosphorothioate, or any other modified backbone.
The two strands may be connected to each other solely by Watson-Crick base pairing and/or by covalent bonds.
In a further aspect, the invention concerns an NF-κB dsODN molecule comprising a sequence, in 5′ to 3′ direction, selected from the group consisting of SEQ ID NOs 26 through 77 and 10.
In a still further aspect, the invention concerns an NF-κB dsODN molecule comprising a sequence, in 5′ to 3′ direction, selected from the group consisting of SEQ ID NOs: 26 through 34.
In another aspect, the invention concerns an NF-κB dsODN molecule comprising a sequence, in 5′ to 3′ direction, selected from the group consisting of SEQ ID NOs: 26 through 31.
In yet another aspect, the invention concerns an NF-κB dsODN molecule comprising the sequence of SEQ ID NO: 30.
In a particular embodiment, the NF-κB dsODN molecule is 12 to 28, or 14 to 24, or 14 to 22 base pairs long, and may comprise modified or unusual nucleotides.
In a further aspect, the invention concerns an NF-κB dsODN molecule which exhibits a p65 competitive binding affinity of 45 or less.
In a particular embodiment, the dsODN with good p65 binding affinity comprises in its sense strand, in 5′ to 3′ direction, a sequence of the formula FLANK1-CORE-FLANK2, wherein
CORE is selected from the group consisting of GGGGACTTTCCC (SEQ ID NO: 9); GGGACTTTCC (SEQ ID NO: 5); GGACTTTCCC (SEQ ID NO: 25); GGGATTTCC (SEQ ID NO: 11); and GGACTTTCC (SEQ ID NO: 13);
FLANK1 is selected from the group consisting of AGTTGA (SEQ ID NO: 79); CTC; TC; CT; CCTTGAA (SEQ ID NO: 6); and CT; and
FLANK 2 is selected from the group consisting of AGGC (SEQ ID NO: 88); TGT; TC; AGG; TCC; and TCA.
Just as in other aspects of the invention, the latter dsODN molecules may have hybrid or otherwise modified backbones, strands that are partially or fully complementary, and connected to each other, completely or partially, by Watson-Crick base pairing and/or by other covalent or non-covalent means.
In another aspect, the invention concerns a composition comprising an NF-κB double-stranded decoy oligodeoxynucleotide (dsODN) as described above. The composition may, for example, be a pharmaceutical composition.
In yet another aspect, the invention concerns a method for the treatment of an inflammatory, immune or autoimmune disease, comprising administering to a mammalian subject in need an effective amount of an NF-κB double-stranded decoy oligodeoxynucleotide (dsODN) molecule described above.
The invention further concerns a method for the treatment of cancer, comprising administering to a mammalian subject in need an effective amount of an NF-κB double-stranded decoy oligodeoxynucleotide (dsODN) molecule herein.
In a further aspect, the invention concerns a method for the treatment of reperfusion injury or restenosis, comprising administering to a mammalian subject in need an effective amount of an NF-κB double-stranded decoy oligodeoxynucleotide (dsODN) molecule herein.
In all aspects, the mammalian subject is preferably human.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a graph showing the p65/p50 binding of certain NF-κB decoy molecules.
FIG. 2 is a graph showing the p50/p50 binding of certain NF-κB decoy molecules.
FIG. 3 shows an electrophoretic mobility-shift assay (EMSA) analysis that identifies the NF-κB proteins contained in complexes bound to a radiolabeled oligonucleotide probe with the NF-κB transcription factor binding site.
FIG. 4 shows that topical application of NF-κB decoy suppresses inflammation in a dose dependent manner in a mouse model of atopic dermatitis.
FIG. 5 shows that cessation of the betamethasone, but not NF-κB decoy treatment results in a rebound of ear swelling and inflammation in the Dp injected ear.
FIG. 6 shows that topical NF-κB decoy treatment decreases the expression of pro-inflammatory cytokines (IL-1B and IL-6) in Dp induced murine atopic dermatitis.
FIG. 7 shows that topical application of NF-κB decoy suppresses inflammation and the infiltration of inflammatory cells responsible for atopic dermatitis, and decreases epidermal hyperproliferation, cellular infiltration and degranulation of mast cells.
FIG. 8 shows that administration of NF-κB decoy into the arthritic joints of rats with collagen induced arthritis (CIA) leads to amelioration of arthritis.
FIG. 9 shows that administration of NF-κB decoy into the arthritic joints of rats with adjuvant induced arthritis (AIA) leads to amelioration of arthritis.
FIG. 10 shows that NF-κB decoy reverses weight loss in a murine model of TNBS-induced colitis.
FIG. 11 shows that NF-κB decoy reduces inflammation in a murine model of TNBS-induced colitis.
FIG. 12 shows that NF-κB decoy reverses weight loss in oxazolone-induced colitis.
FIG. 13 shows that NF-κB decoy increases survival in oxazolone-induced colitis.
FIG. 14 shows the relationship of NF-κB affinity of various decoy molecules as predicted by bioinformatics versus the specificity/affinity factor experimentally generated.
FIG. 15 is a scatter plot of bioinformatics matrix binding score versus competition score when binding to p65/p50. Lower competition scores imply higher binding affinity.
FIG. 16 is a scatter plot of the difference of competition scores between binding with p50/p50 and binding with p65/p50, versus the ratio of the competition scores (p50/p50 versus p65/p50). The plots within the smaller square represent the decoys with the best specificity for p65/p50 vs. p50/p50, whereas the plots outside the smaller and larger square have poor specificity, and the plots between (outside the smaller square and inside the larger square) have intermediate specificity.
FIG. 17 shows that a phosphorothioated NF-κB decoy is stable to nucleases contained in mouse serum as compared to a corresponding decoy having a phosphodiester backbone.
FIG. 18 shows quantitated results from a competitive Electrophoretic Mobility-Shift Assay (EMSA) which indicates the ability of the NF-κB decoy to specifically bind to NF-κB subunits as compared to a scramble ODN. Identification of the complexes is shown by antibody supershift.
FIG. 19 shows NF-κB decoy efficacy in a chronic murine inflammation model. NC/Nga mice injected intradermally with Dp-extract showed ear swelling. Treatments initiated on day 11 (indicated by the arrow) include 1% scramble ODN, 0.25% and 1% NF-κB decoy and 0.1% betamethasone valerate (BMV). The error is expressed as standard error of the mean (SEM) (n=8, P<0.001). These results are representative of 10 independent experiments.
FIG. 20 shows NF-κB decoy efficacy in an acute murine inflammation model. A single application of 2 μg/ml phorbol 12-tetradecanoate 13-acetate (TPA) induced an acute inflammatory response as noted by a dramatic increasee in ear thickness. Treatments include 1% scramble ODN, 0.1% 0.25% and 0.5% NF-κB decoy and 0.1% betamethasone valerate (BMV). The error is expressed as SEM (n=8, *P<0.002). These results are representative of 5 independent experiments.
FIG. 21 shows EMSA analysis that was performed using ³²P-labeled probe and nuclear extracts prepared from control, inflamed, 1% scramble ODN, 1% NF-κB decoy or BMV treated NC/Nga ears. The identities of the NF-κB proteins contained in complexes bound to the radiolabeled oligonucleotide probe were determined by pre-incubating extracts with antibodies specific for each family member of the NF-κB proteins. The complexes indicated in the brackets contain p65 and p50 homo and heterodimers. Antibody/NF-κB protein complexes of slower mobility are indicated by the asterisk (right panel; *). These results are representative of 3 independent experiments.
FIG. 22 shows the results of a quantitation gel shift analysis that was performed by normalizing the NF-κB specific band intensities to the values obtained from a TransAM assay performed with the same nuclear extracts for the ubiquitously expressed transcription factor, NF-YA. The y-axis represents normalized values to inflamed tissue, following which ratios were obtained with values for NF-κB band intensities to NF-YA absorbance readings. Each data point is representative of 4 independent ear samples.
FIG. 23 shows that the tested NF-κB decoy suppresses inflammation. (A) Hematoxylin and Eosin (H&E)-stained sections showed Dp-induced hyperproliferated epidermis and edematous dermis, with massive increase in lymphocyte infiltrations. Sections were observed at a magnification of 10×, scale bar is 100 μm. (B) Toludine blue stained skin sections showing the number of mast cells (indicated by arrows). (C) Sections were stained with congo red to quantitate eosinophils (indicated by arrows). (D) CD4⁺T cells were detected by immunoflouresence microscopy (indicated by arrows). Sections were observed at a magnification of 40×, scale bar in B, C, and D is 25 μm.
FIG. 24 shows quantitative representation of the decrease in mast cells (*P<0.02 for 0.25% NF-κB decoy and P<0.04 for BMV), eosinophils (*P<0.01 for 0.25% NF-κB decoy and BMV) and CD4⁺T cells upon topical NF-κB decoy and BMV treatment.
FIG. 25 shows NF-κB decoy localization in inflamed tissue using 1% biotinylated NF-κB decoy. 1% NF-κB decoy-treated ears were stained for CD207⁺LC, CD4⁺cells, CD117⁺mast cells and MCP-1⁺macrophages. Laser scanning confocal microscopy images show that the NF-κB decoy is co-localized with cell markers. Sections were observed at a magnification of 60×, scale bar is 200 μm. Lower panel is a higher magnification of images taken at 60×.
FIG. 26 shows NF-κB decoy delivery in normal porcine skin. Co-localization of 0.5% NF-κB decoy and nuclei is shown in the left panel. Non-treated skin as a negative control for avidin staining shows minimum background staining on the right panel. Sections were observed at a magnification of 40×, scale bar is 25 μm.
FIG. 27 shows the results of pharmacokinetic-PCR assay that was performed on a pig skin to detect NF-κB decoy in ng levels. Increased uptake of the tested NF-κB decoy is shown in tape stripped samples. These results are representative of 2 independent experiments.
FIG. 28 shows that NF-κB decoy inhibits proinflammatory regulators. 1% NF-κB decoy and BMV suppressed cytokine production. Inflamed ears after 14 days of treatment twice-a-day (BID) were stained with anti-IL-1β, anti-TNF-α, anti-MIP2α and anti-ICAM antibodies. Sections were observed at a magnification of 40×, scale bar is 25 μm.
FIG. 29 shows that an NF-κB decoy suppresses key proinflammatory mRNA levels. Ubiquitin normalized mRNA levels of key proinflammatory genes from inflamed ears treated with 1% Scramble ODN, 1% topical NF-κB decoy and betamethasone valerate (BMV) for 7 days BID is represented on y-axis. The error is expressed as SEM (n=4). These results are representative of 3 independent experiments.
FIG. 30 shows that NF-κB decoy induces apoptosis. Dustmite induced inflammed ears were analyzed for apoptosis by tunnel assay (upper panel) and for proliferation using Ki67staining (lower panel) after 2 days of treatment with BMV and 1% NF-κB decoy. Sections were observed at a magnification of 20×, scale bar is 100 μm.
FIG. 31 shows the apoptotic and mitotic index. Quantitative representation as mitotic and apoptotic index as percent positive cells in epidermis and dermis is shown. Cells were quantitated from 6 independent fields from 3 independent samples. For apoptotic index: compare to inflamed sample P<0.0001 for NF-κB decoy and P<0.000 1 for BMV. For mitotic index: compare to inflamed sample P<0.006 for NF-κB decoy and P<0.0003 for BMV. The error is expressed as SEM. These results are representative of 3 independent experiments.
FIG. 32 shows that NF-κB decoy does not cause skin atrophy. Normal mouse ears were treated with BMV and 1% NF-κB decoy for 14 days twice-a-day (BID). The error is expressed as SEM (n=8, P<0.0001). These results are representative of 5 independent experiments.
FIG. 33 shows picro sirus red stained sections. Dermal thickness of normal treated ears was 93±12 μm, 1% NF-κB decoy treated ear was 89±10 μm and BMV treated ear was 66±7 μm. Sections were observed at a magnification of 40×, scale bar is 25 μm.
FIG. 34 shows ubiquitin normalized mRNA levels of extracellular matrix genes. Collagen I (Col I), Collagen III (Col III), Tenacin C and elastin are represented on the y-axis, treatments include 1% NF-κB decoy and BMV. The error is expressed as SEM (n=4, BMV samples were compared to normal samples P<0.001 for Col I and Col III, P<0.03 for tenacin C and P<0.002 for elastin. These results are representative of 3 independent experiments
FIG. 35 shows NF-κB decoy withdrawal does not cause rebound of swelling and inflammation. Pharmacokinetic-PCR analysis shows decreased 1% NF-κB decoy after 3 days. Ears were analyzed by real-time quantitative PCR. The error is expressed as SEM (n=8). These results are representative of 2 independent experiments.
FIG. 36 shows Dp-Ag recall response after withdrawl of NF-κB decoy treatment. NC/Nga mice injected intradermally with Dp-extract showed ear swelling. BMV and 1% topical NF-κB decoy treatment was initiated on day 13 (indicated by the arrow). Ears were treated for 14 days twice-a-day (BID). Topical treatment was then stopped on day 28 (*). Following which a single intradermal injection of Dp-extract was performed in all groups. Ear swelling was measured for the next 2 weeks. Data is plotted as SEM (n=8).
FIG. 37 shows that NF-κB decoy restores normal levels of transepidermal water loss (TEWL). TEWL measures of inflamed ears treated with BMV or 1% topical NF-κB decoy for 14 days BID show lack of normal restoration in BMV treated samples (marked with *), while 1% NF-κB decoy treated ears indicated normal TEWL. The error is expressed as SEM (n=8, P<0.005). These results are representative of 3 independent experiments.
FIG. 38 also shows that NF-κB decoy restores normal levels of TEWL. Topical BMV and 1% topical NF-κB decoy treatment was performed on normal BALB/c ears for 14 days BID, followed by acetone treatment. Barrier recovery was determined at 3, 6, 24 hrs. The 100% value is from the untreated normal control animals. Results are plotted as % recovery and the error is expressed as SEM (n=8, P<0.005). These results are representative of 3 independent experiments.
FIG. 39 shows that NF-κB decoy treatment does not increase expression of inflammation-associated stratum comeum genes. Ubiquitin normalized mRNA levels of inflammation associated stratum comeum genes from inflamed ears treated with BMV and 1% topical NF-κB decoy for 14 days BID are represented on the y-axis. Error is expressed as SEM (n=5, P<0.005). These results are representative of 2 independent experiments.
FIG. 40 shows that NF-κB decoy treatment restores goblet cell function. Representative sections of trefoil factor family (TFF) and mucopolysaccharide staining in the colon assessing goblet cell function from vehicle (A, C, E) and 1.5 mg/kg NF-κB decoy (B, D, F) treatment groups seven days after induction of TNBS colitis are shown.
FIG. 41 shows that NF-κB decoy reverses trinitrobenzesulfonic acid (TNBS)-induced colonic inflammation. Mice were treated intrarectally with saline vehicle (A), 1.5 mg/kg NF-κB decoy (B), 0.3 mg/kg budesonide (C), or 1.5 mg/kg scramble ODN (D) on days 2 and 4 post TNBS induction. Mice were sacrificed on day 7, fixed in formalin, and stained with hematoxylin and eosin. Shown are representative sections from each treatment group.
FIG. 42 shows that NF-κB decoy treatment restores colonic tissue homeostasis. Ki67 staining of colonic sections from inflamed vehicle treated mice and mice receiving 2 doses of 1.5 mg/kg Decoy at day 2 and 4 post TNBS induction. Colons were harvested at day 7 and processed for analysis. Proliferation in the mucosal epithelium is clearly reduced and restricted to normal zones of the intestinal crypts in the NF-κB decoy-treated mice (B) as compared to inflamed vehicle controls (A). Elevated proliferation in the inflamed colons was also evident in the lamina propria and muscularis layers (C). In NF-κB decoy treated mice, normal levels of proliferation were also observed in the lymphoid follicles (D). Smooth muscle actin staining revealed a very disrupted muscularis in the inflamed vehicle treated mice that was returned to normal after NF-κB decoy treatment (E versus F, respectively).
FIG. 43 shows that NF-κB decoy is effectively delivered to inflamed regions of the colon. NF-κB decoy (2.5 mg/kg) labeled with a HEX flourophore was intrarectally applied to the colon at the day 4 post TNBS-induction time point and harvested 6 hr after administration. Sytox green nuclear counterstaining overlayed with the red Hex-labeled NF-κB decoy indicates the absence of NF-κB decoy in relatively normal colonic regions (A) and a very concentrated localization within inflamed lesions (B) throughout the colon. At higher magnification it is apparent that the NF-κB decoy co-localizes in the cell nucleus.
FIG. 44 shows the amount of NF-κB decoy present in the tissue that was quantitated based on the percent of p65/p50 heterodimer remaining and normalized to vehicle control values. Data represent mean±S.E.M. values (n=3-5mice/group).
FIG. 45 shows that NF-κB decoy inhibits p65/p50 complex in the colon. Colons from TNBS-induced mice were analyzed for tissue NF-κB levels by electrophoretic mobility shift assay using ³²P-labeled probe and nuclear extracts prepared from inflamed vehicle controls and NF-κB decoy treated mice.
FIG. 46 shows quantitation of transcription factor levels over an eight day timecourse that was graphed with trend line for regression analysis.
FIG. 47 shows that NF-κB decoy reduces inflammatory cell infiltration and neutrophil activity. Immunohistochemistry for CD45+ cells indicates a reduction in the number of leukocytes after NF-κB decoy treatment (1.5 mg/kg) as compared to vehicle controls (A versus B). Neutrophil infiltration was specifically determined using a cell type specific antibody and by measuring myeloperoxidase levels in the TNBS-induced colons. The number of neutrophils is clearly reduced in histological sections from the distal colon after NF-κB decoy treatment (C versus D).
FIG. 48 shows a significant decrease in MPO activity that was observed as a result of NF-κB decoy treatment at the day 7 time point. Results are expressed as units of MPO activity per mg protein in colons from 5-8 animals per group. Bars represent the mean±S.E.M from two independent experiments. Asterisk indicates significant difference compared with saline-treated group (p<0.05).
FIG. 49 shows that NF-κB decoy reduces tissue cytokine mRNA levels. Ubiquitin normalized mRNA levels of TNFα (A), IL-6 (B), IL-1β (C) and MCP-1 (D) from inflamed colons after two intrarectal treatments with vehicle, 1.5 mg/kg NF-κB decoy, 0.3 mg/kg budesonide or 1.5 mg/kg scramble ODN. Data is plotted as mean±S.E.M. (n=5-8 mice/group) from two independent experiments. Asterisk indicates significant difference compared with saline-treated group (p<0.05).
FIG. 50 shows that NF-κB decoy reverses disease activity in the oxazolone colitis model. Mean percent body weight of 3 independent experiments (n=25-35 mice per group). Asterisk indicates a significant difference between NF-κB decoy and vehicle treated groups at the day 2 and 3 time points (p<0.01).
FIG. 51 shows representative H&E stained sections from various treatment groups.
FIG. 52 shows colonic sections from a representative experiment that were assessed by a blinded pathologist and quantitated in the histological colitis score. Bars represent the mean±S.E.M (n=5/group). Asterisk indicates significant difference compared with saline-treated group (p<0.05).
FIG. 53 shows the mean disease activity index (DAI) at day 10. Colitis was induced with 5% DSS in the drinking water for 10 days and mice were intrarectally treated with saline vehicle, 2.5 mg/kg NF-κB decoy, budesonide, or scramble ODN every other day for a total of 5 treatments. Data represent mean±S.E.M. (n=15-25 mice/group) from 3 independent experiments. Asterisk indicates significant difference compared with saline-treated group (p<0.05).
FIG. 54 shows representative H&E stained sections from treatment groups in FIG. 53.
FIG. 55 shows the histological colitis score. Bars represent the mean±S.E.M (n=6-14/group).
FIG. 56 shows the mean disease activity index (DAI) 10 days after the start of the second DSS cycle. Data represent mean±S.E.M. (n=15-25 mice/group) from 3 independent experiments. Asterisk indicates significant difference compared with saline-treated group (p<0.05).

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

A. Definitions
Unless defined otherwise, technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Singleton et al., Dictionary of Microbiology and Molecular Biology 2nd ed., J. Wiley & Sons (New York, N.Y. 1994), and March, Advanced Organic Chemistry Reactions, Mechanisms and Structure 4th ed., John Wiley & Sons (New York, N.Y. 1992), provide one skilled in the art with a general guide to many of the terms used in the present application.
The term “double-stranded” is used to refer to a nucleic acid molecule comprising two complementary nucleotide strands connected to each other solely by Watson-Crick base pairing. The term specifically includes molecules which, in addition to the double-stranded region formed by the two complementary strands, comprise single-stranded overhang(s).
The terms “oligonucleotide decoy,” “double-stranded oligonucleotide decoy,” “oligodeoxynucleotide decoy,” and “double-stranded oligodeoxynucleotide decoy” are used interchangeably, and refer to short nucleic acid molecules comprising a double-stranded region, which bind to and interfere with a biological function of a targeted transcription factor. Accordingly, the terms “NF-κB oligonucleotide decoy,” “double-stranded NF-κB oligonucleotide decoy,” “NF-κB oligodeoxynucleotide decoy,” and “double-stranded NF-κB oligodeoxynucleotide decoy” are used interchangeably, and refer to short nucleic acid molecules comprising a double-stranded region, which bind to and interfere with a biological function of an NF-κB transcription factor. The term “double-stranded” is used to refer to a nucleic acid molecule comprising two complementary nucleotide strands connected to each other by Watson-Crick base pairing. The term specifically includes NF-κB oligodeoxynucleotide decoy molecules which, in addition to the double-stranded region formed by the two complementary strands, comprise single-stranded overhang(s). In addition, the term specifically includes NF-κB oligodeoxynucleotide decoy molecules in which, in addition to the double-stranded region, the two strands are covalently linked to each other at their 3′ and/or 5′ end.
The term “NF-κB” is used herein in the broadest sense and includes all naturally occurring NF-κB molecules of any animal species, including all combinations of members of the NF-κB/Rel family, e.g. p52, p50, p65, cRel and Rel B.
The term “transcription factor binding sequence” is a short nucleotide sequence to which a transcription factor binds. The term specifically includes naturally occurring binding sequences typically found in the regulatory regions of genes the transcription of which is regulated by one or more transcription factors. The term further includes artificial (synthetic) sequences, which do not occur in nature but are capable of competitively inhibiting the binding of the transcription factor to a binding site in an endogenous gene.
As used herein, the phrase “modified nucleotide” refers to nucleotides or nucleotide triphosphates that differ in composition and/or structure from natural nucleotides and nucleotide triphosphates.
As used herein, the terms “five prime” or “5′” and “three-prime” or “3′” refer to a specific orientation as related to a nucleic acid. Nucleic acids have a distinct chemical orientation such that their two ends are distinguished as either five-prime (5′) or three-prime (3′). The 3′ end of a nucleic acid contains a free hydroxyl group attached to the 3′ carbon of the terminal pentose sugar. The 5′ end of a nucleic acid contains a free hydroxyl or phosphate group attached to the 5′ carbon of the terminal pentose sugar.
As used herein, the term “overhang” refers to a double-stranded nucleic acid molecule, which does not have blunt ends, such that the ends of the two strands are not coextensive, and such that the 5′ end of one strand extends beyond the 3′ end of the opposing complementary strand. It is possible for a linear nucleic acid molecule to have zero, one, or two, 5′ overhangs.
As used herein, the terms “preferential binding,” “preferentially bind” and their grammatical equivalents are used to mean that the specificity/affinity factor is at least about 20, or at least about 25, or at least about 30, or at least about 35, or at least about 40, where the specificity/affinity ratio is defined as follows:
Specificity/affinity factor=(S _p50/p50 −S _p65/p50)×S _p50/p50 /S _p65/p50
where S_p50/p50equals the molar excess of decoy required to compete 50% of the binding of p50/p50 to the non-mammalian NF-κB promoter from HIV (sequence 113/114, SEQ ID NO: 48, and its complement) and S_p65/p50equals the molar excess of decoy required to compete 50% of the binding of p65/p50 to the non-mammalian NF-κB promoter from HIV (sequence 113/114). The score (S) is assigned as 100 if the decoy is unable to compete at least 50% of the binding at any molar ratio tested.
The term “binding affinity” refers to how tightly a given transcription factor will bind to a corresponding oligonucleotide decoy, which can be measured by various experimental approaches, including electromobility shift assays (EMSA) or TransAM assays.
The term “competition ratio” describes the ability of a test decoy sequence to compete with a defined sequence for binding and retention of the transcription factor when compared to the defined sequence competing with itself in the TransAm assay. A smaller ratio refer to a higher competition ability to bind the transcription factor.
As used herein, the term “inflammatory disease” or “inflammatory disorder” refers to pathological states resulting in inflammation, typically caused by neutrophil chemotaxis. Examples of such disorders include inflammatory skin diseases including psoriasis, eczema and atopic dermatitis; systemic scleroderma and sclerosis; responses associated with inflammatory bowel disease (IBD) (such as Crohn's disease and ulcerative colitis); ischemic reperfusion disorders including surgical tissue reperfusion injury, myocardial ischemic conditions such as myocardial infarction, cardiac arrest, reperfusion after cardiac surgery and constriction after percutaneous transluminal coronary angioplasty, stroke, and abdominal aortic aneurysms; cerebral edema secondary to stroke; cranial trauma, hypovolemic shock; asphyxia; adult respiratory distress syndrome; acute-lung injury; Behcet's Disease; dermatomyositis; polymyositis; multiple sclerosis (MS); meningitis; encephalitis; uveitis; osteoarthritis; lupus nephritis; autoimmune diseases such as rheumatoid arthritis (RA), Sjorgen's syndrome, vasculitis; diseases involving leukocyte diapedesis; central nervous system (CNS) inflammatory disorder, multiple organ injury syndrome secondary to septicaemia or trauma; alcoholic hepatitis; bacterial pneumonia; antigen-antibody complex mediated diseases including glomerulonephritis; sepsis; sarcoidosis; immunopathologic responses to tissue/organ transplantation; inflammations of the lung, including pleurisy, alveolitis, vasculitis, pneumonia, chronic bronchitis, bronchiectasis, diffuse panbronchiolitis, hypersensitivity pneumonitis, idiopathic pulmonary fibrosis (IPF), and cystic fibrosis; etc. The preferred indications include, without limitation, inflammatory skin conditions, such as dermatitis, eczema, rheumatoid arthritis (RA), rheumatoid spondylitis, gouty arthritis and other arthritic conditions, chronic inflammation, autoimmune diabetes, multiple sclerosis (MS), asthma, systhemic lupus erythrematosus, adult respiratory distress syndrome, Behcet's disease, psoriasis, chronic pulmonary inflammatory disease, graft versus host reaction, Crohn's Disease, ulcerative colitis, inflammatory bowel disease (IBD), Alzheimer's disease, and pyresis, along with any disease or disorder that relates to inflammation and related disorders.
The terms “apoptosis” and “apoptotic activity” are used in a broad sense and refer to the orderly or controlled form of cell death in mammals that is typically accompanied by one or more characteristic cell changes, including condensation of cytoplasm, loss of plasma membrane microvilli, segmentation of the nucleus, degradation of chromosomal DNA or loss of mitochondrial function. This activity can be determined and measured, for instance, by cell viability assays, FACS analysis or DNA electrophoresis.
The terms “cancer” and “cancerous” refer to or describe the physiological condition in mammals that is typically characterized by unregulated cell growth. Examples of cancer include, without limitation, carcinoma, lymphoma, leukemia, blastoma, and sarcoma. Specific examples of such cancers include squamous cell carcinoma, small-cell lung cancer, non-small cell lung cancer, breast cancer, pancreatic cancer, glioblastoma multiforme, cervical cancer, stomach cancer, bladder cancer, hepatoma, colon carcinoma, and head and neck cancer. In a preferred embodiment, the cancer includes breast cancer, ovarian cancer, prostate cancer, and lung cancer.
The term “treatment” refers to both therapeutic treatment and prophylactic or preventative measures, wherein the object is to prevent or slow down (lessen) the targeted pathologic condition or disorder. For purposes of this invention, beneficial or desired clinical results include, but are not limited to, alleviation of symptoms, diminishment of extent of disease, stabilized (i.e., not worsening) state of disease, delay or slowing of disease progression, amelioration or palliation of the disease state, and remission (whether partial or total), whether detectable or undetectable.. Those in need of treatment include those already with the disorder as well as those prone to have the disorder or those in whom the disorder is to be prevented. In tumor (e.g., cancer) treatment, a therapeutic agent may directly decrease the pathology of tumor cells, or render the tumor cells more susceptible to treatment by other therapeutic agents, e.g., radiation and/or chemotherapy.
A “subject” is a vertebrate, preferably a mammal, more preferably a human.
The term “mammal” is used herein to refer to any animal classified as a mammal, including, without limitation, humans, higher primates, rodents, domestic and farm animals, and zoo, sports, or pet animals, such as sheep, dogs, horses, cats, cows, etc. Preferably, the mammal herein is human.
B. Detailed Description
It is known that the p50/p50 homodimer blocks activation of pro-inflammatory genes, while the p65/p50 heterodimer acts by turning on pro-inflammatory genes, and as a result, is a major component in the pathogenesis of inflammation. One idea underlying the present invention is that by designing decoy molecules which could bind p65/p50 and/or cRel/p50 heterodimers and not p50/p50 homodimers, or which would preferentially bind p65/p50 and/or cRel/p50 heterodimers, one could provide extra blockade of NF-κB driven promoters by leaving p50/p50 homodimers behind to occupy these sites. As a result, such selective decoy molecules have the potential to block NF-κB activity, such a pro-inflammatory activity associated with the NF-κB pathway, more efficiently than NF-κB decoys known in the art.
Another objective of the present invention is to provide NF-κB decoy molecules with improved binding affinity for p65. This class of NF-κB dsODN molecules is particularly useful in situations where p50/p50 homodimers are present in small concentrations only, like in the case of dermatitis, and therefore p65/p50 specificity is not critical.
Design of NF-κB Decoys with Improved Properties
1. Design of NF-κB dsODN Molecules
The oligonucleotide decoys of the present invention have been designed taking advantage of the crystal structure of the p50/p65 heterodimer bound to the immunoglobulin light-chain gene (Chen et al, Nature 391(6665):410-3 (1998)) which contains the consensus sequence of 5′-GGGACTTTCC-3′ (SEQ ID NO: 2). The authors showed that p50 contacts the 5-base-pair subsite 5′-GGGAC-3′ (SEQ ID NO: 3) and that p65 contacts the 4-base-pair subsite 5′TTCC-3′ (SEQ ID NO: 4). The DNA contacts by the p50/p65 heterodimer are similar to those in the homodimer structures (Ghosh et al, Nature 373(6512):303-10 (1995); Muller et al, FEBS Lett. 369(1):113-7 (1995)).
In addition, each dsODN sequence was analyzed, using bioinformatics methods which give a score of how well a decoy is predicted to bind to NF-κB. Subsequently, the specificity/affinity factor was experimentally generated using traditional binding assays (e.g. competitive binding assay), but could easily be derived from alternative binding assays, including the TransAM™ method (Active Motif, Carlsbad, Calif.), which is an ELISA-based method for detecting and quantifying transcription factor activation. The predicted binding affinity was derived using a conventional approach to assess the TF binding affinity, using the TF binding sites matrix system that statistically summarized the experimental TF-DNA binding data. The analysis was conducted using the most updated version (8.2, June 2004) of the TRANSFAC database (Wingender et al., Nucleic Acids Res 24:238-41 (1996); Wingender et al., Pac Symp Biocomput 477-85 (1997); Wingender et al., Nucleic Acids Res. 25:265-8 (1997)). TRANSFAC collects position-weight-matrices for DNA-TF binding. The tool Match (Kel et al., Nucleic Acids Res. 31:3576-9 (2003)) was used to assess the binding affinity of decoys using the matrices of TRANSFAC, and actual specificity/affinity factors were compared. The data are illustrated in FIGS. 14 and 15, and discussed below. The exact scores will be assay dependent, however, the relative affinity would remain valid, regardless of the specific assay used.
In one embodiment, the NF-κB dsODN molecules of the present invention consist of two oligonucleotide strands which are attached to each other by Watson-Crick base pairing. While typically all nucleotides in the two strands participate in the base pairing, this is not a requirement. Oligonucleotide decoy molecules, where one or more, such as 1-3 or 1 or 2 nucleotides are not involved in base pairing are also included. In addition, the double stranded decoys may contain 3′ and/or 5′ single stranded overhangs.
In another embodiment, the NF-κB dsODN molecules of the present invention comprise two oligonucleotide strands which are attached to each other by Watson-Crick base pairing, and are additionally covalently attached to each other at either the 3′ or the 5′ end, or both, resulting in a dumbell structure, or a circular molecule. The covalent linkage may be provided, for example, by phosphodiester linkages or other linking groups, such as, for example, phosphothioate, phosphodithioate, or phosphoamidate linkages.
Generally, the dsODN molecules of the invention comprise a core sequence that is capable of specific binding to an NF-κB transcription factor, flanked by 5′ and/or 3′ sequences, wherein the core sequence typically consists of about 5 to 14, or about 6 to 12, or about 7 to about 10 base pairs; and the flanking sequences are about 2 to 8, or about 2 to 6, or about 2 to 4, or about 4 to 8, or about 4 to 6 base pairs long. The molecule typically comprises an about 12 to 28, preferably about 14 to 24 base-pair long double-stranded region composed of two fully or partially complementary strands (including the core and flaking sequences).
Changing the core sequence (including its length, sequence, base modifications and backbone structure) it is possible to change the binding affinity, the stability and the specificity of the NF-κB decoy molecule. Indeed, the NF-κB dsODN molecules of the present invention, which bind the p65/p50 and/or cRel/p50 heterodimers with high affinity and exhibit no or only low affinity binding for the p50/p50 homodimers, were designed by deleting or changing targeted residues in the binding site (core) of a consensus oligonucleotide decoy, based on the crystal structure of the p65/p50 heterodimer binding to DNA, and follow up testing.
In addition, changes in the flanking sequence have a genuine impact on and can significantly increase the in vivo stability of the NF-κB decoy molecule, and may affect binding affinity and/or specificity. In particular, the shape/structure of the NF-κB decoy molecule can be changed by changing the sequences flaking the core binding sequence, which can result in improved stability and/or binding affinity. The shape and structure of the DNA are influenced by the base pair sequence, length of the DNA, backbone and nature of the nucleotide (i.e. native DNA vs. modified sugars or bases). Thus, the shape and/or structure of the molecule can also be changed by other approaches, such as, for example, by changing the total length, the length of the fully complementary, double-stranded region within the molecule, by alterations within the core and flanking sequences, by changing the backbone structure and by base modifications.
The nucleotide sequences present in the decoy molecules of the present invention may comprise modified or unusual nucleotides, and may have alternative backbone chemistries. Synthetic nucleotides may be modified in a variety of ways, see, e.g. Bielinska et al. Science 250:997-1000 (1990). Thus, oxygens may be substituted with nitrogen, sulfur or carbon; phosphorus substituted with carbon; deoxyribose substituted with other sugars, or individual bases substituted with an unnatural base. Thus replacement of non-bridging oxygen atoms of the internucleotide linkage with a sulfur group (to yield a phosphorothioate linkage) has been useful in increasing the nuclease resistance of the dsODN molecule. Experiments determining the relationship between the number of sulfur modifications and stability and specificity of the NF-κB dsODN molecules herein are set forth in the Example below.
In each case, any change will be evaluated as to the effect of the modification on the binding ability and affinity of the oligonucleotide decoy to the NF-κB transcription factor, effect on melting temperature and in vivo stability, as well as any deleterious physiological effects. Such modifications are well known in the art and have found wide application for anti-sense oligonucleotide, therefore, their safety and retention of binding affinity are well established (see, e.g. Wagner et al. Science 260:1510-1513 (1993)).
Examples of modified nucleotides, without limitation, are: 4-acetylcytidin, 5-(carboxyhydroxymethyl)uridine, 2′-O-methylcytidine, 5-carboxymethylaminomethyl-2-thiouridine, 5-carboxymethylaminomethyluridine, dihydrouridine, 2′-O-methylpseudouridine, β,D-galactosylqueuosine, 2′-O-methylguanosine, inosine, N6-isopentenyladenosine 1-metyladenosine, 1-methylpseudouridine, 1-methylguanosine, 1-methylinosine, 2,2-dimethylguanosine, 2-methyladenosine, 2-methylguanosine 3-methylcytidine 5-methylcytidine, N6-methyladenosine, 7-methylguanosine, 5-methylaminomethyl-2-thiouridine, β,D-mannosylqueosine, 5-methoxycarbonylmethyl-2-thiouridine, 5-metoxycarbonalmethyluridine, 5-methoxyuridine, 2-methylthio-N6-isopentenyladenosine, N-((9-beta-D-ribofuransyl-2-methylthiopurine-6-yl)carbamoyl)threonine, N-((9-beta-D-ribofuranosyl)purine-6-yl)N-methylcarbamoyl)threonine, uridine-5-oxyacetic acid-methylester uridine-5-oxyacetic acid, wybutoxosine, pseudouridine queuosine, 2-thiocytidine, 5-methyl-2-thiouridine, 2-thiouridine, 4-thiouridine, 5-methyluridine, N-((9-beta-D-ribofuranosylpurine-6-yl)-carbamoylthreonine, 2′-O-methyl-5-methyluridine, 2′-O-methyluridine, 3-(3-3-amino-3-carboxy-propyl)uridine(acp3)u, and wybutosine.
In addition, the nucleotides can be linked to each other, for example, by a phosphoramidate linkage. This linkage is an analog of the natural phosphodiester linkage such that a bridging oxygen (—O—) is replaced with an amino group (—NR—), wherein R typically is hydrogen or a lower alkyl group, such as, for example, methyl or ethyl. Other likages, such as phosphothioate, phosphodithioate, etc. are also possible.
The decoys of the present invention can also contain modified or analogous forms of the ribose or deoxyribose sugars generally present in polynucleotide structures. Such modifications include, without limitation, 2′-substituted sugars, such as 2′-O-methyl-, 2′-O-allyl, 2′-fluoro- and 2′azido-ribose, carboxylic sugar analogs, α-anomeric sugars, epimeric sugars, such as arabinose, xyloses, lyxoses, pyranose sugars, furanose sugars, sedoheptuloses, acyclic analogs and abasic nucleoside analogs, such as methyl riboside.
In general, the oligonucleotide decoys of the present invention are preferably comprised of greater than about 50%, more preferably greater than about 80%, most preferably greater than about 90% conventional deoxyribose nucleotides.
The NF-κB dsODN decoys of the present invention can be further modified to facilitate their localization, purification, or improve certain properties thereof. For example, a nuclear localization signal (NLS) can be attached to the decoy molecules, in order to improve their delivery to the cell nucleus. The NF-κB/Rel proteins include a common Rel homology domain, which encompasses the NLS. In a preferred embodiment such naturally occurring NLS, or a variant thereof, is used in the decoy molecules of the present invention.
In addition, the NF-κB decoy molecules of the invention may be conjugated with carrier molecules, such as peptides, proteins or other types of molecules, as described, for example, in the following references: Avrameas et al., J Autoimmun 16, 383-391 (2001); Avrameas et al., Bioconjug. Chem. 10: 87-93 (1999); Gallazzi et al., Bioconjug. Chem. 14, 1083-1095 (2003); Ritter, W. et al., J. Mol. Med. 81, 708-717 (2003).
The NF-κB decoy molecules of the invention may further be derivatized to include delivery vehicles which improve delivery, distribution, target specific cell types or facilitate transit through cellular barriers. Such delivery vehicles include, without limitation, cell penetration enhancers, liposomes, lipofectin, dendrimers, DNA intercalators, and nanoparticles.
2. Synthesis of NF-κB dsODN Molecules
The NF-κB sdODN decoy molecules of the present invention can be synthesized by standard phosphodiester or phosphoramidate chemistry, using commercially available automatic synthesizers. The specific dsODN molecules described in the Examples have been synthesized using an automated DNA synthesizer (Model 380B; Applied Biosystems, Inc., Foster City, Calif.). The decoys were purified by column chromatography, lyophilized, and dissolved in culture medium. Concentrations of each decoy were determined spectrophotometrically.
3. Characterization of NF-κB dsODN Molecules
The NF-κB decoy molecules of the present invention can be conveniently tested and characterized in a gel shift, or electrophoretic mobility shift (EMSA) assay. This assay provides a rapid and sensitive method for detecting the binding of transcription factors to DNA. The assay is based on the observation that complexes of protein and DNA migrate through a non-denaturing polyacryamide gel more slowly than free double-stranded oligonucleotides. The gel shift assay is performed by incubating a purified protein, or a complex mixture of proteins (such as nuclear extracts), with a ³²P end-labeled DNA fragment containing a transcription factor-binding site. The reaction products are then analyzed on a non-denaturing polyacrylamide gel. The specificity of the transcription factor for the binding site is established by competition experiments, using excess amounts of oligonucleotides either containing a binding site for the protein of interest or a scrambled DNA sequence. The identity of proteins contained within a complex is established by using an antibody which recognizes the protein and then looking for either reduced mobility of the DNA-protein-antibody complex or disruption of the binding of this complex to the radiolabeled oligonucleotide probe.
The ability of a NF-κB decoy to bind to and block the activity of an NF-κB transcription factor can be determined in traditional binding assays (e.g. competitive binding assay), including the TransAM™ method (Active Motif, Carlsbad, Calif.), which is an ELISA-based method for detecting and quantifying transcription factor activation. Briefly, a target sequence, in this case a natural NF-κB binding site, is immobilized on the plate, and a nuclear extract containing NF-κB is incubated in the wells, in the presence or absence of decoy at various concentrations calculated as the molar ration of decoy:plate bound sequence. Positive control wells include decoy with the same sequence as the target DNA on the plate. The data obtained are presented as the ratio of the absorbance of the test decoys and the absorbance of the positive control decoy. Accordingly, lower ratios represent better binding.
In designing the selective NF-κB decoys herein, based on the crystal structure of p65/p50 heterodimer binding to DNA, targeted residues in the binding site (core) of the consensus oligonucleotide decoy were deleted. The ultimate goal was to design a double-stranded oligonucleotide which was able to bind p65/p50 and/or cRel/p50 heterodimers, preferably both the p65/p50 and cRel/p50 heterodimers, with high affinity and exhibited low affinity for p50/p50 homodimers. To achieve this aim, a variety of NF-κB decoys were tested for their ability to bind the different NF-κB proteins in a gel shift assay as described in the following Example 1.
4. Use of NF-κB dsODN Molecules, and Treatment Methods
NF-κB is involved in the regulation of the transcription of numerous genes. A representative grouping and listing of genes transcriptionally activated by NF-κB is provided below.
Cytokines/chemokines and their modulators, such as, for example, interferon-γ (IFN-γ), interferon-β (IFN-β), interleukins, such as, IL-1, Il-2, IL-6, IL-8, IL-9, IL-10, IL-11, IL-12, IL-13, IL-15, lymphotoxin-α, lymphotoxin-β, TNF-α, MIP-1, MIP-2, MIP-3, RANTES, TNF-α, TRAIL.
Immunoregulators, such as, for example, BRL-1, CCR5, CCR7, CD137, CD154, CD40 and CD40 ligand, CD48, CD83, CD23, IL-2 receptor α chain, certain immunoglobulin heavy and light chains, MHC Class I antigen, T cell receptor subunits, TNF-receptor (p75/80).
Proteins involved in antigen presentation, such as, for example, Complement B, Complement component 3, TAP1, and tapasin.
Cell adhesion molecules, such as, for example, E- and P-selectin, ICAM-1, MadCAM-1, VCAM-1, and Tenascin-C.
Acute phase proteins, such as, for example, angiotensinogen, β-defensin-2, complement factors, tissue factor-1 (TF-1), urokinase-type plasminogen activator.
Stress response genes, such as, for example, angiotensin-2, COX-2, MAP4K1, Phospholipase A2.
Cell surface receptors, such as, for example, CD23, CD69, EGF-R, Lox-1, Mdr1.
Regulators of apoptosis, such as, for example, Bf11, Bcl-xL, Caspase-11, CD95 (Fas), TRAF-1, TRAF-2.
Growth factors and their modulators, such as, for example, G-CSF, GM-CSF, EPO, IGFBP-1, IGFBP-2, M-CSF, VEGF-C.
Early response genes, such as, for example, TIEG, B94, Egr-1. In addition, NF-κB regulates the transcription of other transcription factors, such as c-myc-, c-myb, A20, junB, p53, WT1, and viruses.
Thus, inhibition of NF-κB induced gene expression, including expression of pro-inflammatory cytokines, such as IL-1 and TNF-α, and immune modulators, by the NF-κB decoy molecules herein is useful in the prevention and treatment of inflammatory, immune and autoimmune diseases, such as rheumatoid arthritis (RA) (Roshak et al., Current Opinion in Pharmacology 2:316-321 (2002)); Crohn's disease and inflammatory bowel disease (IBD) (Dijkstra et al., Scandinavian J. of Gastroenterology Suppl. 236:3741 (2002)); colitis; pancreatitis (Eeber and Adler, Pancreatology 1:356-362 (2001)), periodonitis (Nichols et al., Annals of Periodontology 6:20-29 (2001)); lupus (Kammer and Tsokos, Current Directions in Autoimmunity 5:131-150 (2002)); asthma (Pahl and Szelenyi, Inflammation Research 51:273-282 (2002)); and ocular allergy (Bielory et al., Opinion in Allergy and Clinical Immunology 2:435-445 (2003)), inflammatory skin diseases, such as atopic dermatitis/eczema, psoriasis, and the like.
A more detailed list of diseases and pathogenic conditions, which are targeted for prevention and/or treatment by the NF-κB decoys of the present invention includes psoriasis, eczema and atopic dermatitis; systemic scleroderma and sclerosis; responses associated with inflammatory bowel disease (IBD) (such as Crohn's disease and ulcerative colitis); ischemic reperfusion disorders including surgical tissue reperfusion injury, myocardial ischemic conditions such as myocardial infarction, cardiac arrest, reperfusion after cardiac surgery and constriction after percutaneous transluminal coronary angioplasty, stroke, and abdominal aortic aneurysms; cerebral edema secondary to stroke; cranial trauma, hypovolemic shock; asphyxia; adult respiratory distress syndrome; acute-lung injury; Behcet's Disease; dermatomyositis; polymyositis; multiple sclerosis (MS); meningitis; encephalitis; uveitis; osteoarthritis; lupus nephritis; autoimmune diseases such as rheumatoid arthritis (RA), Sjorgen's syndrome, vasculitis; diseases involving leukocyte diapedesis; central nervous system (CNS) inflammatory disorder, multiple organ injury syndrome secondary to septicaemia or trauma; alcoholic hepatitis; bacterial pneumonia; antigen-antibody complex mediated diseases including glomerulonephritis; sepsis; sarcoidosis; immunopathologic responses to tissue/organ transplantation; inflammations of the lung, including pleurisy, alveolitis, vasculitis, pneumonia, chronic bronchitis, bronchiectasis, diffuse panbronchiolitis, hypersensitivity pneumonitis, idiopathic pulmonary fibrosis (IPF), and cystic fibrosis, and the like.
Rheumatoid Arthritis
Rheumatoid arthritis (RA) is a chronic systemic autoimmune inflammatory disease that mainly involves the synovial membrane of multiple joints with resultant injury to the articular cartilage. Efficacy of the NF-κB decoy molecules of the present invention in the prevention and/or treatment of arthritis can be evaluated in a collagen-induced arthritis (CIA) model (Terato et al. Brit. J. Rheum. 35:828-838 (1966)), in the adjuvant-induced arthritis model (Taurog et al., Cell Immunol 75:271-82 (1983); Taurog et al., Cell Immunol 80:198-204 (1983)), or a model of antibody-mediated arthritis induced by the intravenous injection of a cocktail of four monoclonal antibodies, as described by Terato et al., J. Immunol. 148:2103-8 (1992); Terato et al., Autoimmunity 22:137-47 (1995). Candidates for the prevention and/or treatment of arthritis can also be studied in transgenic animal models, such as, for example, TNF-α transgenic mice (Taconic). These animals express human tumor necrosis factor (TNF-α), a cytokine which has been implicated in the pathogenesis of human rheumatoid arthritis. The expression of TNF-α in these mice results in severe chronic arthritis of the forepaws and hind paws, and provides a good mouse model for study of inflammatory arthritis. As shown in Examples 7 and 8, a representative NF-κB decoy molecule of the present invention showed significant efficacy in the collagen-induced and adjuvant-induced arthritis models.
Skin Diseases and Related Conditions
The efficacy of the NF-κB decoy molecules of the present invention can be tested in various animal models of inflammatory skin conditions. Such models are well known in the art and include Dustmite Ag (Dp) induced contact dermatitis in NC/Nga mice (Sasakawa, T et al., Int Arch Allergy Immunol 126:23947 (2001); Sasakawa et al., Int Arch Allergy Immunol 133:55-63 (2004)). Results of testing a representative NF-κB decoy of the invention in this model are described in Examples 3, 4, and 6.
Since porcine skin is more similar to human skin than rodent skin, the NF-κB decoy molecules of the present invention can also be tested for efficacy in the treatment of inflammatory skin conditions, such as, for example, dermatitis, and eczema in pig skin inflammatory models. The NF-κB decoy molecules of the present invention effectively block NF-κB activity in this model, and block expression of key inflammatory genes.
Existing therapies for the treatment of these conditions are inadequate, raising serious issues of side effects during long term application, especially in children. The NF-κB decoys are devoid of the known side effects of steroid therapy and involve limited systemic exposure.
Skin is known for its functional role as a protective barrier. It is now clear that skin is also a very dynamic organ that can elicit immunological responses. However, not all immunological reactions in the skin are beneficial, some harmful reaction (inflammation) might result from reacting to invaders, such as allergens (contact dermatitis, atopic dermatitis an other kin inflammatory diseases) that require down-regulation/medication. Other reactions might occur from autoreactive lymphocytes being exposed to skin antigens, such as epithelia base membrane component type VII collage or cell surface components or keratinocytes, or still unknown cell components, which can lead to pemphigus vulgaris (Anhalt et al., N. Engl. J. Med. 323(25):1729-35 (1990)), or epidermyosis bullosa (Woodley et al., N. Engl. J. Med. 310(16):107-13 (1984)), orpsoriasis (Stern, Lancet 350(9074):349-53 (1997)). Most of these inflammatory skin responses involve mononuclear cells, in particular T cells. Observations of peri- and intrafollicular inflammation and infiltration of these follicles by T cells and other immune cells, such as macrophages and dendritic cells, point to a T cell mediated autoimmune syndrome in alopecia areata (AA) as well (McElwee et al., J. Invest. Dermatol. 119(6):1426-33 (2002)).
A number of studies have demonstrated the crucial role of cytokines released from immune cells infiltrating the affected sites for initiating and maintaining fibrosis in the skin of scleroderma patients (Fagundus et al., Clin. Dermatol. 12(3):407-17 (1994); Postlethwaite, Current Opin. Rheumatol. 7(6):535-40 (1995)). Thus, drugs that target immune competent cells, such as T cells, should be very successful in treating a broad spectrum of inflammatory conditions of the skin.
Inflammatory Bowel Disease (IBD)
The term “inflammatory bowel disease” or “IBD” is commonly used to ulcerative colitis and Crohn's disease, which are chronic inflammatory diseases of the gastrointestinal tract of unknown etiology. Traditional treatment is based on corticosteroid therapy and the administration of salazopyrine. More recently, other therapeutics, such as an anti-TNF-α antibody, have been developed. However, there is a great need for new therapies, partly due to the known side effects of steroid therapy, and since in a significant subgroup of patients current therapies fail to induce remission. See, e.g. Mee et al., Gut 2:1-5 (1979). Mouse models are described by Matsumoto, et al., Gut 43:71-78 (1998); and Strober et al., J Clin Invest 107:667-670 (2001).
Similarly to certain skin diseases, inflammatory disease of the gastrointestinal tract involves immune competent cells. For example, in inflammatory bowel disease (ulcerative colitis and Crohn's disease), T cells are the main culprit of the inflammation. The balance between pro- and anti-inflammatory cytokines secreted by T cells regulates both the initiation and perpetuation of inflammatory bowel disease. Most of these secreted factors are regulated directly or indirectly by NF-κB (Neurath et al., Natl. Med. 8(6):567-73 (2002); Strober et al., J. Gastroenterol. 38 Suppl 15:55-8 (2003)). Other diseases of the gastrointestinal tract, such as irritable bowel disease (IBS), gastritis, Barrett syndrome, peptide lulcer, reflux disease, acute and chronic pancreatitis, cholecystitis, result from similar, albeit more chronic dysregulated mucosal immune systems and thus are expected to be responsive to treated with NF-κB inhibitors.
Other Therapeutic Utilities
Since NF-κB plays a pivotal role in the coordinated transactivation of cytokine and adhesion molecule genes involved in atherosclerosis and lesion formation after vascular injury (Yoshimura et al., Gene Therapy 8: 1635-1642 (2001)); neuronal damage after cerebral ischemia (Ueno et al., J. Thoracic and Cardiovascular Surgery 122(4): 720-727 (2001)); chronic airway inflammation (Griesenbach et al., Gene Therapy 7, 306-313 (2000)); progression of autoimmune myocarditis (Yokoseki et al., Circ. Res. 89: 899-906 (2001)); acute rejection and graft arteriopathy in cardiac transplantation (Suzuki et al., Gene Therapy 7: 1847-1852 (2000)); and myocardial infarction (Morishita et al., Nature Medicine 3(8): 894-899 (1997)), NF-κB decoy molecules also find utility in the treatment of such diseases and conditions.
Recent evidence indicates that NF-κB and the signaling pathways that are involved in its activation are also important for tumor development. See, e.g. Karin et al., Nat. Rev. Cancer 2(4):301-10 (2002). Therefore, blocking NF-κB by the decoy molecules of the present invention finds utility in the prevention and treatment of cancer, offering a new anti-cancer strategy, either alone or in combination with other treatment options.
Delivery of the NF-κB dsODN Molecules
The route of delivery of the NF-κB decoys of the present invention depends on the disease or pathological condition the prevention and/or treatment is targeted. For certain indications, a preferred mode of delivering the NF-κB decoys of the present invention is pressure-mediated transfection, as described, for example, in U.S. Pat. Nos. 5,922,687 and 6,395,550, the entire disclosures of which are hereby expressly incorporated by reference. In brief, the NF-κB decoy molecules are delivered to cells in a tissue by placing the decoy nucleic acid in an extracellular environment of the cells, and establishing an incubation pressure around the cells and the extracellular environment. The establishment of the incubation pressure facilitates the uptake of the nucleic acid by the cells, and enhances localization to the cell nuclei.
More specifically, a sealed enclosure containing the tissue and the extracellular environment is defined, and the incubation pressure is established within the sealed enclosure. In a preferred embodiment, the boundary of the enclosure is defined substantially by an enclosing means, so that target tissue (tissue comprising the target cell) is subjected to isotropic pressure, and does not distend or experience trauma. In another embodiment, part of the enclosure boundary is defined by a tissue. A protective means such as an inelastic sheath is then placed around the tissue to prevent distension and trauma in the tissue. While the incubation pressure depends on the application, incubation pressures about 300 mmHg-1500 mmHg above atmospheric pressure, or at least about 100 mmHg above atmospheric pressure are generally suitable for many applications.
The incubation period necessary for achieving maximal transfection efficiency depends on parameters such as the incubation pressure and the target tissue type. For some tissue, such as human vein tissue, an incubation period on the order of minutes (>10 minute) at low pressure (about 0.5 atm) is sufficient for achieving a transfection efficiency of 80-90%. For other tissue, such as rat aorta tissue, an incubation period on the order of hours (>1 hour) at high pressure (about 2 atm) is necessary for achieving a transfection efficiency of 80-90%.
Suitable mammalian target tissue for this type of delivery includes blood vessel tissue (in particular veins used as grafts in arteries), heart, bone marrow, and normal and tumor connective tissue, liver, genital-urinary system, bones, muscles, gastrointestinal organs, endocrine and exocrine organs, synovial tissue and skin. A method of the present invention can be applied to parts of an organ, to a whole organ, or to a whole organism. In one embodiment a nucleic acid solution can be perfused into a target region (e.g. a kidney) of a patient, and the patient is subject to pressure in a pressurization chamber.
For other applications, the NF-κB decoys of the present invention can be administered by other conventional techniques. For example, retrovial transfection, transfection in the form of liposomes are among the known methods suitable for transfection. For details see also Dzau et al., Trends in Biotech 11:205-210 (1993); or Morishita et al., Proc. Natl. Acad. Sci. USA 90:8474-8478 (1993). When administered in liposomes, the decoy concentration in the lumen will generally be in the range of about 0.1 μM to about 50 μM per decoy, more usually about 1 μM to about 10 μM, most usually about 3 μM.
Dosing is dependent on severity and responsiveness of the disease state to be treated, with the course of treatment lasting from several days to several months, or until a cure is effected or a diminution of the disease state is achieved. Optimal dosing schedules can be calculated from measurements of drug accumulation in the body of the patient. Persons of ordinary skill can easily determine optimum dosages, dosing methodologies and repetition rates. Optimum dosages may vary depending on the relative potency of individual oligonucleotides. In general, an effective dose is from 0.01 μg to 100 g per kg of body weight. Persons of ordinary skill in the art can easily estimate repetition rates for dosing based on measured residence times and concentrations of the drug in bodily fluids or tissues. In addition to the potency of the specific decoy molecule delivered, the effective dose will depend on the target disease, the route of delivery, the formulation used, the severity of the disease, the age, sex, and overall condition of the patient to be treated, and other similar considerations.
For treatment of rheumatoid arthritis and inflammatory bowel disease (IBD), murine experiments typically use about 25-100 μg decoy in 50 to 100 μl injectable formulation. Based on these dose regimens, determination of the proper dosage for other species, including humans, can be assisted by mathematical calculations that have been developed for inter-species scaling. See, for example, Mordenti et al., Pharm Res 8:1351-9 (1991).
For topical application, the decoys of the present invention are administered in the form of conventional topical formulations, such as, for example, creams, ointments, gels, suspensions, and the like. Preferably, such topical formulations will contain one or more penetration enhancers and/or surfactants to assure efficient delivery through the skin. Topical formulations for delivery of dsODN molecules are described in co-pending application U.S. Provisional Application Ser. No. 60/612,046, filed on Sep. 21, 2004, the entire disclosure of which is hereby expressly incorporated by reference. Topical formulations specifically include aqueous, emulsion-based and liposome formulations. For the treatment of dermatitis/eczema, a typical formulation may be a topical cream containing about 0.1 to 5% by weight, or about 0.2 to 3% by weight, or about 0.25 to 1% by weight of active ingredient. In order to increase half-life, for dermatological applications, at least one and preferably both strands of the NF-κB decoy molecules of the present invention are fully phosphorothioated.
For other delivery routes, the most suitable concentration can be determined empirically. The determination of the appropriate concentrations and doses is well within the competence of one skilled in the art. Optimal treatment parameters will vary depending on the indication, decoy, clinical status of the patient, etc., and can be determined empirically based on the instructions provided herein and general knowledge in the art.
The decoys may be administered as compositions comprising individual decoys or mixtures of decoys. Usually, a mixture contains up to 6, more usually up to 4, more usually up to 2 decoy molecules.
Many anti-inflammatory and anti-rheumatic drugs, including glucocorticoids, aspirin, sodium salicylate, and sulfosalazine, are inhibitors of NF-κB activation. For the treatment of inflammatory and autoimmune diseases and conditions, the NF-κB decoy molecules of the present invention can optionally be administered in combination with such drug treatments. Combination treatment includes simultaneous administration as well as consecutive administration of two or more drugs in any order. Thus, for example, the topical anti-inflammatory applications, the NF-κB decoys of the invention can be administered in combination with betamethasone or similar therapeutic agents.
In cancer therapy, the administration of the NF-κB decoy molecules can be combined with other treatment options, including treatment with chemotherapeutic anticancer agents and/or radiation therapy.
Further details of the invention will be apparent from the following non-limiting Examples.

EXAMPLE 1

Design and Testing of NF-κB Decoy Molecules
Design
NF-κB dsODN decoy molecules were designed and tested for their ability to bind and/or compete for binding of NF-κB. In a particular aspect, the goal of the invention was to design NF-κB decoy molecules that preferentially bind p65/p50 and/or cRel/p50 heterodimers over p50/p50 homodimers. As a result of not blocking p50/p50 homodimers, the selective decoy molecules of the invention allow these homodimers to block the promoters of NF-κB regulated genes, which provides an additional level of negative regulation of gene transcription.
In designing the oligonucleotide decoys, information available from crystal structure studies and computational analysis of the known NF-κB binding sites were utilized.
As discussed above, based on study of the crystal structure of the p50/p65 heterodimer bound to the immunoglobulin light-chain gene, which contains the consensus sequence of 5′-GGGACTTTCC-3′ (SEQ ID NO: 2), it has been shown that p50 contacts the 5-base-pair subsite 5′-GGGAC-3′ (SEQ ID NO: 3) and that p65 contacts the 4-base-pair subsite 5′TTCC-3′ (SEQ ID NO: 4). A series of NF-κB oligonucleotide decoys were designed, which contained fewer numbers of G's at the 5′ end of the consensus binding site with the aim to prepare decoy molecules that would have lower affinity for the p50/p50 homodimer but still bind the p65/p50 heterodimer. These oligonucleotide decoys were assigned “core” and “flank” letter codes for ease of identification and presentation. The cores were assigned letter codes “A” through “L” and the flanks “T” through “Z′”. The decoys were tested in the gel shift assay to determine their ability to compete with a high affinity radiolabeled oligonucleotide for NF-κB binding The NF-κB-binding DNA consensus sequences were selected from publications of NF-κB related DNA-protein interactions, including: Blank et al., EMBO J. 10:4159-4167 (1991); Bours et al. Mol. Cell. Biol. 12:685-695 (1992); Bours et al. U. Cell 72:729-739 (1993); Duckett et al. Mol. Cell. Biol. 13:1315-1322 (1993); Fan C. -M., Maniatis T., Nature 354:395-398 (1991); Fujita et al., Genes Dev. 6:775-787 (1992); Fujita et al. Genes Dev. 7:1354-1363 (1993); Ghosh et al., Nature 373:303-310 (1995); Ghosh et al., Cell 62:1019-1029 (1990); Grumont et al., Mol. Cell. Biol. 14:8460-8470 (1994); Henkel et al., Cell 68:1121-1133 (1992); Ikeda et al. Gene 138:193-196 (1994); Kunsch et al., Mol. Cell. Biol. 12:4412-4421 (1992); LeClair et al., Proc. Natl. Acad. Sci. USA 89:8145-8149 (1992); Li C. -C. et al., J. Biol. Chem. 269:30089-30092 (1994);
Matthews et al., Nucleic Acids Res. 21:1727-1734 (1993); Mueller et al., Nature 373:311-317 (1995); Neri et al., Cell 67:1075-1087 (1991); Nolan et al., Cell 64:961-969 (1991); Paya et al., Proc. Natl. Acad. Sci. USA 89:7826-7830 (1992); Plaksin et al., J. Exp. Med. 177:1651-1662 (1993); Schmid et al., Nature 352:733-736 (1991); Schmitz M. L., Baeuerle P. A. EMBO J. 10:3805-3817 (1991); Ten et al., EMBO J. 11:195-203 (1992); Toledano et al., J. Mol. Cell. Biol. 13:852-860 (1993); Urban et al., EMBO J. 10:1817-1825 (1991).
Based on this available information, a set of decoys was generated for initial screening. These decoys include a “mutation decoy”, the scrambled decoys, decoys with different length at their 5′ or 3′ end, and decoys with alternative base composition within the core region and/or in the flanking sequences.
To better understand the base-composition near the core binding sites of NF-κB, the core binding sites were computationally aligned (forward strand only) with known binding sequences. Based on this alignment, several major groups of decoys with slightly different core binding sites were created.

The major core and flanking sequences are listed in Table 1.

TABLE 1


LETTER	SEQ	LETTER

CODE	CORE	ID	CODE	FLANK

A	GGGACTTTCC	5	T	CCTTGAA . . . TCC
				(SEQ ID NO:6 and
				SEQ ID NO:8)

B	GGGGACTTTCC	7	U	AT . . . GT
				(SEQ ID NO:12 and
				SEQ ID NO:14)

C	GGGGACTTTCCC	9	V	TC . . . TC
				(SEQ ID NO:16 and
				SEQ ID NO:18)

D	GGGATTTCC	11	W	CTC . . . TGT
				(SEQ ID NO:20 and
				SEQ ID NO:22)

E	GGACTTTCC	13	W′	CTC . . . TCA
				(SEQ ID NO:82 and
				SEQ ID NO:83)

F	GACTTTCC	15	X	CT . . . TC
				(SEQ ID NO:84 and
				SEQ ID NO:85)

G	GACTTTCCC	17	Y	TC . . . CA
				(SEQ ID NO:86 and
				SEQ ID NO:87)

H	GGATTTCC	19	Z	AGTTGA . . . AGGC
				(SEQ ID NO:79 and
				SEQ ID NO:88)

I	GGATTTCCC	21	Z′	TTGA . . . AGGC
				(SEQ ID NO:80 and
				SEQ ID NO:89)

J	GATTTCC	23	Z-4	TTGA . . . AG
				(SEQ ID NO:90 and
				SEQ ID NO:91)

K	GATTTCCC	24	Z-2	GTTGA . . . AGG
				(SEQ ID NO:81 and
				SEQ ID NO:92)

L	GGACTTTCCC		25

Oligonucleotide Synthesis
The sequences of the NF-κB decoys used in the Examples herein include 5′AGTTGAGGACTTTCCAGGC-3′ (SEQ ID NO: 30) and 5′-GCCTGGAAAGTCCTCAACT-3′ (SEQ ID NO: 95). The scrambled ODN sequence was 5′-CAGTAGTATGTGAGCCTGC-3′ (SEQ ID NO: 46) and 5′-GCAGGCTCACATACTACTG-3′ (SEQ ID NO: 96). After synthesis, the ODNs were purified using ion exchange chromatography (Avecia Biotechnology). All ODNs used were phosphorothioates and were annealed by combining equal molarities of each strand at room temperature. Concentrations of ODNs were determined by spectrophotometry.
Electrophoretic Mobility Shift Assay (EMSA)
The EMSA assay was employed to characterize oligonucleotide decoys for the NF-κB transcription factor. Using a radiolabeled oligonucleotide probe (non-mammalian, based on NF-κB promoter from HIV, sequence 113/114, SEQ ID NO: 48, and its complement), which exhibits high affinity for relevant members of the NF-κB family, binding of p65/p50, cRel/p50 and p50/p50 was tested using a nuclear extract from an activated monocyte cell line. Using the above modified oligonucleotides to compete for binding for the above-mentioned NF-κB family members, it was possible to compare the binding affinity of these oligonucleotides against the high affinity radiolabeled probe and each other. This assay has also enabled the design of a decoy which selectively binds particular members of the NF-κB family. By using increasing concentrations of various oligonucleotides, it was observed that, by deleting or changing targeted residues in the binding site, it was possible to specifically decrease the binding of the decoy molecule to p50/p50 homodimers, while retaining the affinity for p65/p50 and cRel/p50 heterodimers.
The NF-κB gel shift assays (EMSA) were performed as follows. A double-stranded oligonucleotide containing a consensus NF-κB binding site (5′ AGTTGAGGGGACTTTCCCAGGC 3′) (SEQ ID NO: 78) was end-labeled with γ³²P-ATP using T4 Polynucleotide Kinase (Promega). Alternatively, a double-stranded ODN containing a consensus NF-κB binding site from the IL-6 promoter (5′ TGTGGGATTTTCCCATGACTC 3′) (SEQ ID NO: 94) was end-labeled with γ³²P-ATP using T4 Polynucleotide Kinase (Promega). One microgram of a nuclear extract prepared from LPS stimulated THP-1 cells (human monocyte cell line) was incubated with 35 fmol of radiolabeled probe in the presence or absence of competing unlabeled NF-κB double-stranded oligonucleotides (dsODN) or scrambled dsODN. The incubations were carried out at room temperature for 30 minutes in a 20 μl reaction volume composed of 10 mM Tris- HCl pH 8, 100 mM KCL, 5 mM MgCl2, 2 mM DTT, 10% Glycerol, 0.1% NP-40, 0.025% BSA and 1 μg Poly-dIdC. The reactions were loaded onto a 6% polyacrylamide gel, subjected to electrophoresis and dried. The dried gels were imaged and quantitated using a Typhoon 8600 PhosphorImager (Amersham) and ImageQuant software. The identity of the NF-κB proteins contained in complexes bound to the radiolabeled oligonucleotide probe were identified by pre-incubating the reactions for 5 to 10 minutes with individual antibodies specific for each member of the NF-κB family prior to the addition of the radiolabeled probe. All antibodies were purchased from Active Motif (p65 and p50) or Santa Cruz Biotechnology (p52, RelB, cRel).
In Vitro Serum Stability Assay
To increase ODN stability, the normal phosphodiester backbone was modified to contain phosphorothioate linkages. Stability enhancement from the phosphorothioate backbone was verified by incubating NF-κB decoy containing either a phosphodiester or phosphorothioate backbone in serum for different time intervals and assessing degradation by gel electrophoresis.
Mouse serum (United States Biologicals) was diluted to a final working concentration of 70% using Phosphate Buffered Saline (PBS). NF-κB decoy containing either a phoshodiester (PO) or phosphorothioate (PS) backbone was diluted into serum for a final concentration of 4 μM. Gentamycin was included to prevent bacterial growth. The reactions were allowed to incubate at 37° C. and 10 μl aliquots were removed at 0, 30, 75 and 120 minutes. The reactions were stopped by the addition of 1 μl 0.5M EDTA and quickly frozen on dry ice. The samples were loaded onto a 20% polyacrylamide gel and subjected to electrophoresis. The gels were imaged and quantitated using a Typhoon 8600 PhosphorImager (Amersham) and ImageQuant software.
The results show that as a phosphorothioate, the NF-κB decoy was very stable in mouse serum (FIG. 17). Although some degradation was observed by 30 minutes incubation, only minimal degradation was observed in subsequent time points. In contrast, the phosphodiester NF-κB decoy was degraded relatively quickly and completely degraded at 120 min.
FIG. 18 shows that nuclear extracts from a human monocytic cell line has 2 strong DNA-protein complexes when incubated with an NF-κB-containing consensus sequence from the IL-6 promoter. A competitive reduction in the density of these bands was observed with NF-κB decoy addition that was concentration dependent. The band densities remained unchanged in samples treated with a negative control ODN containing a scrambled sequence (scramble ODN) at up to 10-fold excess over the radiolabeled probe. The NF-κB complexes, identified by antibody supershifts, were shown to either contain a p65/p50 complex (top band) or a combination of p50/cRel/p52 complexes (lower band). This verifies that the NF-κB decoy binds the proinflammatory NF-κB heterodimeric complexes in a dose- and sequence-specific manner and has improved stability due to the chemically-modified backbone.
Nuclease Degradation and Chemistry Modifications
Native DNA is subject to rapid degradation inside of a cell, primarily through the action of 3′ exonucleases, but also as a result of endonuclease attack. Therefore, oligonucleotide decoys are designed, they are modified to enhance their stability. Replacing one of the non-bridging oxygen atoms of the internucleotide linkage with a sulfur group, creating what is referred to as a phosphorothioate (PS) oligodeoxynucleotide, has been highly successful. The molecules are relatively nuclease resistant; however, they have been shown to exhibit nonspecific protein binding relative to 3′-terminally modified and unmodified oligonucleotide decoys (Brown et al, J. Biol. Chem. 269(43):26801-5 (1994)). Therefore, a set of experiments was performed to determine how many sulfurs were required at the 3′, 5′ or an internal site to provide nuclease resistance to our oligonucleotide decoys while maintaining the achieved specificity.
The Analysis of the EMSA Results
As discussed earlier, one goal of the work disclosed herein has been to develop NF-κB oligonucleotide decoy molecules that preferentially bind to the NF-κB p65/p50 and/or cRel/p50 heterodimer relative to the p50/p50 homodimer. The experimental results showed that the binding to p65/p50 and cRel/p50 were generally equivalent, therefore, only the p65/p50 bands were quantitated in our analysis.
FIG. 1 shows the p65/p50 binding of certain NF-κB decoy molecules. FIG. 2 shows the p50/p50 binding of certain NF-κB decoy molecules.
Preferential binding was quantified using the specificity/affinity factor, calculated as follows:
Specificity/affinity factor=(S _p50/p50 −S _p65/p50)×S _p50/p50 /S _p65/p50
where S_p50/p50equals the molar excess of decoy required to compete 50% of the binding of p50/p50 to the non-mammalian NF-κB promoter from HIV (sequence 113/114) and S_p65/p50equals the molar excess of decoy required to compete 50% of the binding of p65/p50 to the non-mammalian NF-κB promoter from HIV (sequence 113/114, where the reverse strand corresponding to sequence 113 is designated as “114”). The score (S) is assigned as 100 if the decoy is unable to compete at least 50% of the binding at any molar ratio tested.

A preferred decoy molecule will have a lower score for the p65/p50 heterodimer and a higher score for the p50/p50 homodimer. The specificity of the decoy to p65/p50 heterodimer versus p50/p50 homodimer is proportional to their difference of score (score p50/p50-score p65/50). The results of the EMSA competition binding experiments, performed as described above, are summarized in Table 2A, where the decoy molecules are listed starting with the most specific decoys (highest specificity/affinity factor).

Flank

(p50-

affinity

ID

ID	Sequences	Alias	p65	p50	p65)*p50	p65)/p65	factor	NO

173	TTGAGGACTTTCCAG	E-Z-4	55	100	4500	0.82	81.82	26

177	CTCGGACTTTCCTGT	E-W	57.5	100	4250	0.74	73.9	27

151	AGTTGAGGGATTTCCAGGC	D-Z	36	69	2277	0.92	63.25	28

207	AGTTGAGGACTTTCCCAGGC	L-Z	20	45	1125	1.25	56.25	29

153	AGTTGAGGACTTTCCAGGC	E-Z	62.4	96	3225.6	0.54	51.69	30

265	AGTTGAGGATTTCCCAGGC	I-Z	54	84	2520	0.56	46.67	31

235	TCGGACTTTCCCTC	L-V	37	62.5	1593.75	0.69	43.07	32

117	ATGGACTTTCCGT	E-U	72.5	100	2750	0.38	37.93	33

227	TCGGATTTCCTC	H-V	74	100	2600	0.35	35.14	34

155	TCGGACTTTCCTC	E-V	81	97	1552	0.20	19.16	35

281	TTGAGGACTTTCCAGGC	E-Z	87	100	1300	0.15	14.94	36
		EVEN

121	TCGGGACTTTCCTC	A-V	25	34	306	0.36	12.24	37

263	AGTTGAGGATTTCCAGGC	H-Z	73	82	738	0.12	10.11	38

295	TGAGGACTTTCCAGGCTC		92	100	800	0.09	8.70	39

289	TGAGGACTTTCCAGGC		93	100	700	0.08	7.53	40

309	CCTTGAAGGGATTTCCCTCC	M-T	45	50	250	11	5.56	10

279	TTGCGGACTTTCCAGGC	E-Z A->C	48	52	208	0.08	4.33	41
		EVEN

191	CTGGGACTTTCCTC	A-X	19	22.5	78.75	0.18	4.14	42

141	GTTGAGGGACTTTCCAGG	A-Z-2	7.5	10	25	0.33	3.33	43

123	CTCGGGACTTTCCTGT	A-W	8	10	20	0.25	2.50	44

195	TCGGGGACTTTCCCTC	C-V	8.5	9	4.5	0.06	0.53	45

103	CAGTAGTATGTGAGCCTGC		100	100	0	0.00	0.00	46

107	TTGCCGTACCTGACTTAGCC	SCRAMBLED	100	100	0	0.00	0.00	47

113	AGTTGAGGGGACTTTCCCAGGC	C-Z	5	5	0	0.00	0.00	48

129	TCGGGATTTCCTC	D-V	37.5	37.5	0	0.00	0.00	49

145	AGTTGAGGGACTTTCCAGGC	A-Z	8	8	0	0.00	0.00	50

157	AGTTGAGACTTTCCAGGC	F-Z	100	100	0	0.00	0.00	51

159	AGTTGAGACTTTCCCAGGC	G-Z	100	100	0	0.00	0.00	52

169	GGACTTTCC	E	100	100	0	0.00	0.00	53

171	AGGACTTTCCA	E-A	100	100	0	0.00	0.00	54
		FLANK

179	CTGGACTTTCCTC	E-X	100	100	0	0.00	0.00	55

183	AAGAGGACTTTCCAGAG	E-AG	100	100	0	0.00	0.00	56
		FLANK

185	ATATGGACTTTCCTTAA	E-AT	100	100	0	0.00	0.00	57
		FLANK

187	CAACGGACTTTCCACAC	E-CA	100	100	0	0.00	0.00	58

189	CAGTGGACTTTCCACTG	E-CAGT	100	100	0	0.00	0.00	59
		FLANK

213	TCGACTTTCCCTC	G-V	100	100	0	0.00	0.00	60

221	CTGGGGACTTTCCCTC	C-X	25	25	0	0.00	0.00	61

229	TCGGATTTCCCTC	I-V	100	100	0	0.00	0.00	62

231	TCGATTTCCTC	J-V	100	100	0	0.00	0.00	63

233	TCGATTTCCCTC	K-V	100	100	0	0.00	0.00	64

239	CTCGGGGACTTTCCCTCA	C-W′	9	9	0	0.00	0.00	65

241	CTCGGACTTTCCTCA	E-W′	100	100	0	0.00	0.00	66

273	TTGAGGATTTCCAGGC	H-Z′	100	100	0	0.00	0.00	67
		(−3′2BP)

275	TTGAGGATTTCCAGGCT	H-Z′	100	100	0	0.00	0.00	68
		(−3′1BP)

277	TTGAGGATTTCCAGGCTC	H-Z′	100	100	0	0.00	0.00	69

283	TGAGGACTTTCCAGG	E-Z-3	100	100	0	0.00	0.00	70

285	GAGGACTTTCCAG	E-Z-6	100	100	0	0.00	0.00	71

287	GTTGAGGACTTTCCAGGC		100	100	0	0.00	0.00	72

291	GAGGACTTTCCAGGC		100	100	0	0.00	0.00	73

293	AGGACTTTCCAGGC		100	100	0	0.00	0.00	74

297	AGGACTTTCCAGGCTC		100	100	0	0.00	0.00	75

259	TTGAGGACTTTCCAGGCTC	E-Z′	87	85	−170	−0.02	−1.95	76

197	CTCGGGGACTTTCCCTGT	C-W	26	17.5	−148.75	−0.33	−5.72	77

E-Z minus 4 is E-Z with the two 5′ and 3′ bases deleted
E-Z even is E-Z minus the two 5′ bases
E-Z a to C Even is E-Z with the A in position 6 changed to a C
A-Z-2 is A-Z with the 5′ and 3′ bases deleted
E-A is the E core with an A added to the 5′ and 3′ ends
E-AG Flank is E core with AAGA as the 5′ flank and AGAG as the 3′ flank
E-AT Flank is E core with ATAT as the 5′ flank and TTAA as the 3′ flank
E-CA Flank is E core with CAAC as the 5′ flank and ACAC as the 3′ flank
E-CAGT Flank is E core with CAGT as the 5′ flank and ACTG as the 3′ flank
H-Z′ (−3′2BP) is H-Z with the two 5′ bases deleted
H-Z′ (−3′1BP) is H-Z with two 5′ bases deleted and a T added at the 3′ end
E-Z-3 is E-Z with AGT deleted from the 5′ end with C deleted from the 3′ end
E-Z-6 is E-Z with AGTT deleted from the 5′ end and GC deleted from the 3′ end

The data set forth in Table 2A suggest that the decoys with better p65/p50 specificity most likely share the “E” or “D” or “H” or “L” or “I” core sequence and “Z” or “W” or “V” or “U” or “Z-4” flanking sequences. In a more preferred group, the core sequence is “E” or “D” and the flanking sequence is “Z.” The decoy designated 153/154 was chosen as best from the top few candidates with the consideration of other parameters (see below).

As discussed earlier, for certain uses of the NF-κB decoys herein specificity is not a requirement. For such applications, it is advantageous to select NF-κB decoy molecules with high binding affinity for p65. The following Table 2B shows the p65 binding affinity for some NF-κB decoy molecules tested.

	TABLE 2B


	Core-Flank		SEQ ID

ID	Sequences	Alias	p65	NO

113	AGTTGAGGGGACTTTCCCAGGC	C-Z	5	48

141	GTTGAGGGACTTTCCAGG	A-Z-2	7.5	43

123	CTCGGGACTTTCCTGT	A-W	8	44

145	AGTTGAGGGACTTTCCAGGC	A-Z	8	50

195	TCGGGGACTTTCCCTC	C-V	8.5	45

239	CTCGGGGACTTTCCCTCA	C-W′	9	65

191	CTGGGACTTTCCTG	A-X	19	42

207	AGTTGAGGACTTTCCCAGGC	L-Z	20	29

121	TCGGGACTTTCCTC	A-V	25	37

221	CTGGGGACTTTCCCTC	C-X	25	61

197	CTCGGGGACTTTCCCTGT	C-W	26	77

151	AGTTGAGGGATTTCCAGGC	D-Z	36	28

235	TCGGACTTTCCCTC	L-V	37	32

129	TCGGGATTTCCTC	D-V	37.5	49

309	CCTTGAAGGGATTTCCCTCC	M-T	45

279	TTGCGGACTTTCCAGGC	E-Z A->C	48	41
		EVEN

265	AGTTGAGGATTTCCCAGGC	I-Z	54	31

173	TTGAGGACTTTCCAG	E-Z-4	55	26

177	CTCGGACTTTCCTGT	E-W	57.5	27

153	AGTTGAGGACTTTCCAGGC	E-Z	62.4	30

117	ATGGACTTTCCGT	E-U	72.5	33

263	AGTTGAGGATTTCCAGGC	H-Z	73	38

227	TCGGATTTCCTC	H-V	74	34

155	TCGGACTTTCCTC	E-V	81	35

281	TTGAGGACTTTCCAGGG	E-Z EVEN	87	36

259	TTGAGGACTTTCCAGGCTC	E-Z′	87	76

295	TGAGGACTTTCCAGGCTC		92	39

289	TGAGGAGTTTCCAGGC		93	40

103	CAGTAGTATGTGAGCCTGC		100	46

107	TTGCCGTACGTGACTTAGCC	SCRAMBLED	100	47

157	AGTTGAGACTTTCCAGGC	F-Z	100	51

159	AGTTGAGACTTTCCCAGGC	G-Z	100	52

169	GGACTTTCC	E	100	53

171	AGGACTTTCCA	E-A FLANK	100	54

179	CTGGACTTTCCTC	E-X	100	55

183	AAGAGGACTTTCCAGAG	E-AG FLANK	100	56

185	ATATGGACTTTCCTTAA	E-AT FLANK	100	57

187	CAACGGACTTTCCACAC	E-CA FLANK	100	58

189	CAGTGGACTTTCCACTG	E-CAGT	100	59
		FLANK

213	TCGACTTTCCCTG	G-V	100	60

229	TCGGATTTCCCTC	I-V	100	62

231	TCGATTTCCTC	J-V	100	63

233	TCGATTTCCCTC	K-V	100	64

241	CTCGGACTTTCCTCA	E-W′	100	66

273	TTGAGGATTTCCAGGC	H-Z′	100	67

275	TTGAGGATTTCCAGGCT	H-Z′	100	68

277	TTGAGGATTTCCAGGCTC	H-Z′	100	69

283	TGAGGACTTTCCAGG	E-Z-3	100	70

285	GAGGACTTTCCAG	E-Z-6	100	71

287	GTTGAGGACTTTCCAGGC		100	72

291	GAGGACTTTCCAGGC		100	73

293	AGGACTTTCCAGGC		100	74

297	AGGACTTTCCAGGCTC		100	75

A similar analysis was applied to evaluate the chemical modification of DNA backbone for tested decoys.

TABLE 3A


S_p50/p50/S_p65/p50and specificity/affinity factor for 153/154 with various DNA backbones

						Specificity/Affinity
backbone	p65	p50	p50-p65	(p50-p65) * p50	(p50 -p65)/p65	Factor

PO/H11	7.5	40	32.5	1300	4.33	173.33
H7/H7	39	100	61	6100	1.56	156.41
H6/H6	40	100	60	6000	1.50	150.00
H3/H3	45	100	55	5500	1.22	122.22
H11/PO	7.5	33	25.5	841.5	3.40	112.20
H5/H5	50	100	50	5000	1.00	100.00
PO	58	92	34	3128	0.59	53.93
PO/H5	65	100	35	3500	0.54	53.85
H8/H8	29	53	24	1272	0.83	43.86
H10/H10	21	42	21	882	1.00	42.00
H5/PO	59	83	24	1992	0.41	33.76
H10/H8	9	21	12	252	1.33	28.00
H9/H9	17	30	13	390	0.76	22.94
H4/H4	85	100	15	1500	0.18	17.65
H8/H10	33	39	6	234	0.18	7.09
H11/H11	24	27	3	81	0.13	3.38
PS/PO	5	5	0	0	0.00	0.00
PO/PS	5	5	0	0	0.00	0.00

In the foregoing Table 3A, where there are two designations for the backbone chemistry, the first one indicates the chemistry of strand 153 and the second the chemistry of strand 154. Fully phosphodiester bonds are designated “PO,” fully phosphorothioate backbones are designated “PS.” Hybrid backbones are designated with an “H” followed by the number of phosphorothioate backbone linkages, starting from the 3′ end. Thus, H3 means that the three most 3′ linkages are phosphorothioate and the rest of the backbone linkages are phosphodiester. If only one designation is shown (such as just PO), both strands have the same backbone chemistry.
The data set forth in Table 3A indicate that if either strand or both strands are fully phosphothioated (e.g. PS/PO or PO.PS) then the decoy has a high affinity for both p65/p50 and p50/p50, and therefore lacks the specificity desired. Generally, although not always, a higher number of phosphorothioate linkages resulted in reduced specificity. Generally, hybrid strands with more than 8 phosphorothioate linkages lacked specificity, whereas those with fewer than 7 retained acceptable affinity and specificity. However, H4/H4 has extremely low affinity, while H3/H3 and H5/H5 were both in the acceptable range. H11/PO and PO/H11 has good affinity and specificity. Based on half-life, specificity and affinity, H3/H3, H5/H5, H6/H6, and H7/H7 were identified as the optimal backbone for the 153/154 decoy. Optimal backbone chemistries for other decoys can be tested and determined in an analogous manner.

The effect of backbone chemistry on the p65 binding affinity is illustrated by the data set forth in Table 3B below.

	TABLE 3B


	backbone	p65

	PS/PO	5
	PO/PS	5
	PS/PS	5
	PO/H11	7.5
	H11/PO	7.5
	H10/H8	9
	H9/H9	17
	H10/H10	21
	H11/H11	24
	H8/H8	29
	H8/H10	33
	H7/H7	39
	H6/H6	40
	H3/H3	45
	H5/H5	50
	PO	58
	H5/PO	59
	PO/H5	65
	H4/H4	85

Just as before, if increased p65 binding affinity is the goal, preferably decoys with an affinity score below 40, or below 35, or below 30, or below 25, or below 20 are selected. Since this data was generated in a competitive binding assay, smaller scores indicate better binding affinity, as they represent the amount of competing material left binding after the competition. The results set forth in Table 3B show that, with a few exceptions, such as H4 on both strands or H5 on one strand, which decreases p65 binding affinity, increasing the level of phosphorothioate substitutions generally increases p65 binding affinity.
Specificity Relative to Other Transcription Factors
Decoy molecules must also specifically block only the target transcription factor and not non-specifically bind and block unrelated transcription factors. It has also been established that it is possible to design NF-κB decoy molecules which do not exhibit any non-specific effects on unrelated promoters using EMSA. Specifically, using radiolabeled oligonucleotide probes corresponding to the promoter sequences for the ubiquitous transcription factor Oct-1, it demonstrated that 153/154 (wherein “154” designates the reverse sequence corresponding to the sequence “153”) PO and H3 NF-κB decoy did not show any binding affinity for the promoter (data not provided). This is important because any non-specific effects of an oligonucleotide to other important proteins in the cell could result in unwanted toxicity of the decoy for the treatment individual.
Half-Life
Native DNA is subject to rapid degradation inside a cell, primarily through the action of 3′ exonucleases, but also as a result of endonuclease attack. Therefore, when oligonucleotide decoys are designed, they are modified to enhance their stability. Replacing one of the non-bridging oxygen atoms of the internucleotide linkage with a sulfur group, creating what is referred to as a phosphorothioate oligodeoxynucleotide, has been highly successful. The molecules are relatively nuclease resistant; however, they have been shown to exhibit non-specific protein binding relative to 3′-terminally modified and unmodified oligonucleotide decoys (Brown et al., J. Biol. Chem. 269:26801-5 (1994)). Therefore, a set of experiments were performed to determine how many sulfurs were required at the 3′- or 5′-end, or at an internal site to provide nuclease resistance to the oligonucleotide decoys herein, while maintaining specificity.
Binding specificity was assessed by the gel shift assay described above. 3′-exonuclease resistance was assessed using a standard snake venom assay (Cummins et al., Nucleic Acids Res. 23:2019-24 (1995)). To assess the resistance of the decoys to more relevant mammalian nuclease activity, as assay was adapted in which cytoplasmic and nuclear extracts were prepared from activated macrophages. (Hoke et al., Nucl. Acids Res. 19(20):5743-8 (1991)). The activity of the extracts was confirmed with positive controls in each assay. It was determined that capping the 3′-ends of each strand of the decoy with a few sulfur groups was sufficient to protect it from nuclease degradation.
Together these data indicate that for a p50/p65-selective NF-κB decoy 3-5 sulfurs at the 3′ ends of a 19-mer oligonucleotide duplex are sufficient to protect the decoy from nuclease degradation. Additionally, it was able to maintain specific subunit binding within the transcription factor family as well as lack of binding to irrelevant transcription factors. These data demonstrate that the present invention provides methods and means for designing specific and long-lasting oligonucleotide decoys targeting transcription factors, in particular NF-κB.

EXAMPLE 2

NF-κB Decoy Molecules Comprising a Nuclear Localization Signal
In order to determine the ability of a nuclear localization signal (NLS) containing peptide to improve the entry of an oligonucleotide decoy into the nucleus, a peptide with the NLS sequence based on the simian virus 40 large tumor antigen (PKKKRKVEDPYC) (SEQ ID NO: 93) was synthesized by Sigma Genosys and conjugated to the NF-κB 153 H3 oligonucleotide as follows. Briefly, 6.5 nmols of oligonucleotide was first incubated with 40-fold molar excess of the linker Sulfo-SMCC (Pierce) at room temperature for 2 hours. After removal of excess linker from the reaction by a NAP-10 column (Pharmacia Biotech), the activated oligonucleotide was incubated with 5-fold molar excess of the NLS peptide at room temperature overnight. To assess the percentage of oligonucleotide successfully conjugated to the NLS peptide, the reaction was analyzed by loading 1 μl onto a 20% PAGE gel (non-denaturing). The gel was stained with SYBR Gold (Molecular Probes) and visualized on a Typhoon Phosphorimager (Amersham). The concentration of the NLS-peptide conjugated single strand 153 H3 was determined by OD absorbance. The conjugate was then annealed to its complemetary strand 154 H3 (in equal molar amounts) containing a biotin molecule at its 5′ end. The presence of the biotin molecule on the now double stranded NLS decoy was to enable visualization (via streptavidin) of the localization through the use of microscopy.
The following examples describe results obtained by the administration of the NF-kB decoy “153/154.” In the dermatological models, both strands of the NF-kB decoy 153/154 are fully phosphorothioated, i.e. the 153/154 molecule is “PS/PS.”

EXAMPLE 3

NF-κB Decoy Reduces Ear Swelling in Murine Atopic Dermatitis
To determine the efficacy and effective dose range of NF-kB decoy in Dustmite Ag (Dp) induced contact dermatitis in NC/Nga mice.
Method
Dp Ag induced dermatitis was induced as previously described (Sasakawa, T et al., Int Arch Allergy Immunol 126:239-47 (2001); Sasakawa et al., Int Arch Allergy Immunol 133:55-63 (2004)). Briefly, Six week old male NC/Nga mice were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 11, the DP injected ear was topically treated 2 times a day for 10-12 days with 20 μl of vehicle, vehicle containing 0.25% or 0.1 % NF-κB decoy or betamethasone as a control. The ear thickness was measured with an ear thickness gauge (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
Results
To examine the effect of NF-κB decoy treatment on established atopic dermatitis, the ear swelling/inflammation, as measured by ear thickness of Dp-injected NC/Nga mice with or without NF-κB decoy treatment, was examined. Dp extract was injected on the ventral side of the right ear on days 0, 2, 4, 7, 9, and 11. Topical NF-κB decoy was initiated on day 11 after the final Dp injection. Thickening of the ear injected with DP Ag was observed as early as 24 hours later and rapidly increased until day 11. Ear thickness was maintained for up to two weeks after the final Dp injection in both the untreated and vehicle treated ears. As shown in FIG. 4, topical 0.25% NF-κB decoy or betamethasone (BID) treatment resulted in a rapid decrease in ear thickness while the 0.1% NF-κB decoy treatment group resulted in little to no decrease in ear thickness similar to the vehicle control.
Conclusion
Topical application of NF-κB decoy suppresses inflammation in a dose dependent manner in this mouse model of atopic dermatitis.

EXAMPLE 4

Cessation of NF-κB Decoy Treatment Does Not Result in a Rebound of Ear Swelling and Inflammation in the Dp Injected Ear
The purpose of this experiment was to determine whether or not NF-κB decoy efficacy is maintained when treatment is stopped.
Method
Dp Ag-induced dermatitis was induced as previously described (Sasakawa et al, supra). Briefly, six-week old male NC/Nga mice were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 11, the DP injected ear was topically treated 2 times a day for 10-12 days with 20 μl of vehicle, vehicle containing 0.25% NF-kB decoy or betamethasone as a control. Topical treatment was then stopped for 2 weeks. The ear thickness was measured with an ear thickness gauge (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
Results
The ability of NF-κB decoy treatment to maintain the decrease in ear thickness after discontinuation of treatment was examined. As shown in FIG. 5, discontinuation of the NF-κB decoy treatment resulted in no increase in ear swelling, while the cessation of betamethasone in the Dp injected ears resulted in a significant increase of ear swelling.
Conclusion
Cessation of the betamethasone, but not NF-κB decoy treatment results in a rebound of ear swelling and inflammation in the Dp injected ear.

EXAMPLE 5

NF-κB Decoy Treatment Reduces Expression of Key Pro-Inflammatory Genes in Dp Injected NC/Nga Mice
The goal of this study was to evaluate the effect of NF-κB decoy treatment on elevated pro-inflammatory cytokine expression in Dp injected NC/Nga mice.
Method
Right ears of Dp-injected mice were removed 1 day after the final decoy treatment. Part of the ear was flash frozen in liquid nitrogen and store at −80° C. Total RNA was isolated from the ears using Qiazol (Qiagen) according to manufacturer's instruction. The expression of the mouse genes were assayed by real-time quantitative PCR with an ABI PRISM 7900 Sequence Detector System (Applied Biosystems, Foster City, Calif.). All procedures were carried out as previously described (Hurst et al., Immunol 169:443-53 (2002)). Briefly, Dnase-treated total RNAs were mixed with random hexamers (Gibco-BRL), Oligo dt (Boehringer), and the first strand cDNAs were synthesized with SurperScript II reverse transcriptase. Primers for the respective genes were designed using the primer design software Primer Express (Applied Biosystems). Primers were synthesized by Sigma Genosys (Woodlands, Tex.). The quantitative PCR was performed using TaqMan PCR reagent kits according to the manufacturer's protool (Applied Biosystems). Sample cDNAs equivalent to 25 ng of RNA were examined in each reaction in a 384-well PCR plate. Levels of ubiquitin were measured for each sample, and used as internal standard. Cytokine levels are expressed as relative expression to ubiquitin levels.
Results
The effects of NF-κB decoy treatment on pro-inflammatory gene expression in the skin was analyzes by real-time quantitative PCR. As shown in FIG. 6, vehicle or non treated Dp injected mice showed a marked increase in expression levels of both IL-1β and IL-6, while NF-κB decoy treatment decreased the expression levels of both cytokines similar to that of protopic and betamethasone.
Conclusion
Topical NF-κB decoy treatment decreases the expression of pro-inflammatory cytokines (IL-1β and IL-6) in Dp induced murine atopic dermatitis.

EXAMPLE 6

The Efficacy of NF-κB Decoy Therapy in Dustmite Ag (Dp) Induced Contact Dermatitis in NC/Nga Mice
This experiment investigates the efficacy of NF-κB decoy in Dustmite Ag (Dp) induced contact dermatitis in NC/Nga mice.
Method
Right ears of Dp-injected mice were removed 1 day after the final decoy treatment. Part of the ear was fixed in 10% phosphate buffered formalin (pH 7.2) and embedded in paraffin, and 3 micron sections were cut. Then the samples were stained with hematoxylin and eosin for histology and toluidine blue for detection of degranulated mast cells.
Results
Histological examination of the skin lesions was performed on day 26. H&E staining showed severe epidermal hyperplasia and cellular infiltration into the dermis of vehicle treated ears injected with Dp, and treatment with NF-κB decoy or betamethasone produced both a decrease in epidermal hyperplasia and cellular infiltrate. As shown in FIG. 7, most of the mast cells in the Dp-injected, vehicle treated mice were degranulated while treatment with NF-κB decoy or betamethasone demonstrated a decrease in degranulated mast cells.
Conclusion
Topical application of NF-κB decoy suppresses inflammation and the infiltration of inflammatory cells (i.e mast cells) responsible for atopic dermatitis. Also, NF-κB decoy treatment decreased epidermal hyperproliferation, cellular infiltration and degranulation of mast cells.

EXAMPLE 7

NF-κB Decoy Molecules Reduces Ear Swelling in Acute and Chronic Inflammation
The following studies further examine the use of topical NF-κB decoy for targeting Th1/Th2 driven skin inflammation in experimental atopic dermatitis.
Method
Oligonucleotide Synthesis: The NF-κB decoy is a 19 base pair (molecular weight of 13 kDa) double-stranded deoxyribonucleic acid comprised of a random mixture of R- and S-chiral forms of oligonucleotides. The sequence of the NF-κB decoy used in this study was: 5′AGTTGAGGACTTTCCAGGC-3′ (SEQ ID NO: 30) and 5′-GCCTGGAAAGTCCTCAACT-3′ (SEQ ID NO: 95). The scrambled oligonucleotides (ODN) sequence was 5′-CAGTAGTATGTGAGCCTGC-3′ (SEQ ID NO: 46) and 5′-GCAGGCTCACATACTACTG-3′ (SEQ ID NO: 96). All oligonucleotides used were phosphorothioates and were annealed by combining equal molarities of each strand at room temperature.
The oligonucleotides described above are also used in Examples 8-14 described below.
Indicated concentrations of ODN were added to Corgentech's gel based formulation, for topical application (U.S. patent application Ser. No. 11/233,511).
Animal Models:
In order to demonstrate that topical NF-κB decoy is effective in reducing skin inflammation in a mouse model of chronic inflammation, dustmite Ag (Dp)-induced dermatitis was induced in six week old male NC/Nga mice (Charles River Lab). Animals were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 12, the Dp injected ear was topically treated 2 times a day for 11-14 days with 20 μl of vehicle, vehicle containing 0.25% or 1% NF-κB decoy, non-specific scramble ODN, or topical 0.1% betamethasone valerate (BMV) as the positive control. The ear thickness was measured with snap caliper (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
2. In order to demonstrate that topical NF-κB decoy is effective in reducing skin inflammation in an acute mouse treatment model, 12-O-tetradecanoylphorbol 13-acetate (TPA, phorbol 12-tetradecanoate 13-acetate) (TPA)-induced dermatitis was initiated by topical application of 2 μg of TPA on 7-8 week old female Balb/C mice ears. Six hours after induction, mice were treated topically with 20 μl of vehicle, vehicle containing 1% of scramble ODN, 0.1%, 0.25% or 0.5% NF-κB, or topical 0.1% BMV as the positive control. The ear thickness was measured with ear thickness gauge (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after induction.
Results
For the chronic atopic dermatitis model, a dose-dependent inhibition of inflammation following NF-κB Decoy treatment was observed (FIG. 19). The highest topical dose of 1% NF-κB decoy inhibited inflammation by 74%, similar to topical BMV. Ears treated with scrambled ODN, vehicle only, or ears left untreated maintained severe ear thickness (FIG. 19).
For the acute atopic dermatitis model, TPA-induced ear swelling was reduced following application of topical NF-κB decoy; the highest dose (1%) showed similar efficacy to BMV treatment, with over 70% inhibition of ear swelling (FIG. 20)
Conclusion
This data demonstrate that topical NF-κB decoy is effective in reducing skin inflammation in both the chronic and the acute mouse treatment models.

EXAMPLE 8

NF-κB Decoy Molecules Inhibits p65-DNA Binding Activity
The purpose of this experiment was to determine whether NF-κB decoy competitively inhibits NF-κB protein.
Method
Animal Model: Dustmite Ag (Dp)-induced dermatitis was induced in six week old male NC/Nga mice (Charles River Lab). Animals were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 12, the Dp injected ear was topically treated 2 times a day for 11-14 days with 20 μl of vehicle, vehicle containing 0.25% or 1% NF-κB decoy, non-specific scramble ODN, or topical 0.1% betamethasone valerate (BMV) as the positive control. The ear thickness was measured with snap caliper (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
EMSA: The NF-κB gel shift assays were performed as follows. A double-stranded ODN containing a consensus NF-κB binding site from the IL-6 promoter (5′TGTGGGATTTTCCCATGACTC 3′ (SEQ ID NO: 97)) was end-labeled with γ32P-ATP using T4 Polynucleotide Kinase (Promega). 5 μg of a nuclear extract prepared from Dp-Ag injected ears was incubated with 35 fmol of radiolabeled probe in the presence or absence of indicated antibodies. The identity of NF-κB proteins contained in complexes bound to the radiolabeled ODN probe were identified by pre-incubating the reactions for 10 minutes with antibodies specific for each member of the NF-κB family prior to the addition of the radiolabeled probe (all antibodies were purchased from Active Motif (p65 and p50) (Carlsbad, Calif.) or Santa Cruz Biotechnology (p52, RelB, cRel).
NF-YA TransAM Assay: To assess the relative affinities of NF-κB decoy for a NF-YA/SP1 containing complex, the NF-YA TransAM assay (Active Motif, Carlsbad, Calif.) was utilized. The assay was performed according to manufacturer's instructions. A double-stranded ODN containing the consensus NF-YA binding sequence was immobilized on a 96-well plate. A nuclear extract prepared from NC/Nga ears/activated THP-1 cells were incubated and allowed to bind to the immobilized ODN. The unbound material was washed away and the bound NF-YA/SP 1 detected using an antibody that specifically recognizes NF-YA/SP1. The NF-YA/SP1 antibody was detected by a secondary antibody labeled with horseradish peroxidase (HRP), and the amount of HRP in each well was measured using a colorimetric substrate reaction and read using a microplate spectrophotometer.
Results
Gel shift assay was performed with labeled probe and nuclear extracts of inflamed Dp-injected ears to determine NF-κB decoy-dependent competitive inhibition to NF-κB protein. Nuclear extracts from inflamed ears of NC/Nga mice showed 3 strong DNA-protein complexes with low mobility (FIG. 21, indicated by the arrows). A reduction in the density of the three bands was observed in samples treated with topical BMV or 1% NF-κB decoy, while band densities remaining unchanged in samples treated with scrambled ODN. The identity of the p50 and p65 NF-κB sub-units was confirmed by a reduction in mobility following incubation of the sample extracts with specific antibodies (FIG. 21; right panel denoted by *). Nuclear extract quality and equal gel loading was confirmed by normalizing the band intensity from the gel shift analysis to the values obtained from a TransAM assay for the ubiquitously expressed transcription factor, nuclear factor-YA (NF-YA) (FIG. 22). The DNA-binding specificity of the NF-κB decoy was tested through competitive inhibition of NF-κB protein.
Conclusion
NF-κB decoy demonstrates a dose-dependent inhibition of p65-DNA binding activity. Further, NF-κB decoy failed to block the binding activity of other known ubiquitously expressed transcription factors including SP1 or NF-YA.

EXAMPLE 9

Topical NF-κB Decoy Suppresses Inflammation
Method
Animal Model: Dustmite Ag (Dp)-induced dermatitis was induced in six week old male NC/Nga mice (Charles River Lab). Animals were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 12, the Dp injected ear was topically treated 2 times a day for 11 -14 days with 20 μl of vehicle, vehicle containing 0.25% or 1% NF-κB decoy, non-specific scramble ODN, or topical 0.1% betamethasone valerate (BMV) as the positive control. The ear thickness was measured with snap caliper (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
Hematoxylin and Eosin Staining: Dp injected mice ears were formalin-fixed, paraffin-embedded and 5-μM sections were then stained with hematoxylin and eosin. Mast Cells Staining: Deparaffinized and rehydrated sections were stained in toluidine blue working solution (1% Toluidine blue stock solution). Eosinophils Staining: Formalin-fixed, paraffin-embedded tissue samples were stained with congo red for detection of eosionophiis.
Results
Histological analyses showed that topical NF-κB Decoy decreased edema and inflammatory cell infiltrate in skin lesions induced by Dp injection into ears of NC/Nga mice. Hematoxylin and eosin (H&E)-stained sections of untreated lesions at day 25 reveal a hyperplastic epidermis and edematous dermis, with an increase in lymphocyte infiltration, clinically resembling human AD (FIG. 23A). In contrast, topical NF-κB decoy or BMV treatment produced a similar and significant decrease in epidermal hyperplasia and cellular infiltration (FIG. 23A). Toludine blue stained ear sections of both untreated and scrambled ODN treated lesions showed an increased number of mast cells (FIG. 23B), while inflamed skin treated with topical NF-κB decoy or BMV showed fewer mast cells (FIG. 23B). In addition, congo red staining revealed a decrease in eosinophil infiltration in skin treated with topical NF-κB decoy or BMV, (FIG. 23C). CD4⁺T cells were also decreased with topical NF-κB decoy and BMV treatment (FIG. 23D). Mice treated with either vehicle control or scrambled-sequence ODN fail to reduce such inflammatory response. A quantitative representation of the decrease in mast cells (60%), eosinophils (70%) and CD4⁺T cells (60%), upon topical NF-κB decoy and BMV treatment was performed (FIG. 24).
Conclusion
Topical application of NF-κB decoy suppresses inflammation and the infiltration of inflammatory cells (i.e. mast cells) responsible for atopic dermatitis.

EXAMPLE 10

Cellular Uptake of Topical NF-κB Decoy
The purpose of this experiment is to examine the specific cellular uptake of topical NF-κB decoy.
Method
Animal Model: Dustmite Ag (Dp)-induced dermatitis was induced in six week old male NC/Nga mice (Charles River Lab). Animals were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 12, the Dp injected ear was topically treated 2 times a day for 11-14 days with 20 μl of vehicle, vehicle containing 0.25% or 1% NF-κB decoy, non-specific scramble ODN, or topical 0.1% betamethasone valerate (BMV) as the positive control. The ear thickness was measured with snap caliper (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
NF-κB Decoy Localization: Cryosections were used for co-localization of 1% biotinylated NF-κB decoy with various cells in inflamed skin. After fixation in 2% paraformaldehyde, sections were blocked with Goat Ig (Sigma, Atlanta, Ga.). Primary antibodies include rat anti-mouse CD4 (R&D Systems, Minneapolis, Minn.), rat anti-mouse CD117 (BD, San Diego, Calif.), rabbit anti-mouse CD207-Langerin (Imgenex, San Diego, Calif.), rat anti-mouse MCP-1 (Abcam, Cambridge, UK), rat IgG2b Iso. (BD), rabbit Ig Iso (Sigma 10 mg/ml). Secondary antibodies include Alexa 546 SAV (Molecular Probes, Eugene, Oreg.), goat anti-rat Alexa 546, goat anti-rabbit Alexa 546 (Molecular Probes). Images were taken at 60× magnification, using laser scanning confocal microscopy.
Results
To examine the specific cellular uptake of topical NF- To examine the specific cellular uptake of topical NF-κB decoy, topical biotinylated-NF-κB decoy was applied to Dp Ag-induced inflamed ears for 2 days (B.I.D). NF-κB decoy distribution in the epidermal cells and throughout the papillary dermis was observed (FIG. 25). Immunohistological analysis revealed an uptake of NF-κB decoy by several key inflammatory cells. The co-localization of NF-κB decoy was detected in langerin-positive, resident antigen-presenting cells (APC)-Langerhans cells (LC), CD117 positive mast cells, CD4⁺T cells and MCP-1 positive macrophages (FIG. 25).
Conclusion
The results show efficient drug delivery to inflamed mouse skin by immunohistological analysis as revealed by the presence of NF-κB decoy in cells pivotal for disease progression and maintenance, including keratinocytes, resident APC-Langerhans cell LCs, mast cells, CD4⁺T cells and macrophages.

EXAMPLE 11

Topical NF-κB Decoy Uptake in Porcine Skin
The purpose of this experiment is to demonstrate effective topical NF-κB decoy delivery into porcine skin tissue.
Method
For pig skin, 0.5% biotinylated NF-κB decoy was applied at 30 μl/cm²area on the back of a light skinned female Yorkshire pig for 24 hrs. Skin was washed thoroughly with PBS to remove any remaining formulation. Duct tape from Intertape Polymer Group (Sarasota, Fla.) was used and the skin was tape-stripped 20 times. The biotinylated-NF-κB decoy was visualized using AlexaFluor₄₈₈-streptavidin (Molecular Probes) and counterstained with Hoechst stain for nuclear co-localization.
Results
The results show an excellent delivery of topical NF-κB decoy into the porcine epidermis and dermis (FIG. 26). Semi-quantitative analyses using Image-Pro software indicates that over 80% cells in treated pig skin are NF-κB decoy positive. This observation was further confirmed using a quantitative real-time PCR-based analytical method to detect the decoy (FIG. 27). Uptake and retention of topical NF-κB decoy by pig skin following application of a 1% decoy topical formulation was 10 ng/mg of tissue processed at 24 hours post-application. Prior tape-stripping of the skin, to model the compromised barrier in atopic dermatitis, increased the uptake of an equivalent application to 20 ng decoy/mg of tissue (FIG. 27). Overall, these observations confirm effective NF-κB decoy delivery into both compromised and uncompromised skin tissue.
Conclusion
The results demonstrate effective topical NF-κB decoy delivery into porcine skin tissue.

EXAMPLE 12

NF-κB Decoy Effects on Inflammation, Apoptosis and Proliferation
The purpose of this experiment is to evaluate the anti-inflammatory effects of NF-κB decoy through the inhibition of essential regulators of inflammation and by induction of apoptosis of key immune cells.
Method
Animal Model: Dustmite Ag (Dp)-induced dermatitis was induced in six week old male NC/Nga mice (Charles River Lab). Animals were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 12, the Dp injected ear was topically treated 2 times a day for 11-14 days with 20 μl of vehicle, vehicle containing 0.25% or 1% NF-κB decoy, non-specific scramble ODN, or topical 0.1% betamethasone valerate (BMV) as the positive control. The ear thickness was measured with snap caliper (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
Apoptosis and Mitotic Index: Formalin-fixed, paraffin-embedded tissue samples were used to identify apoptotic cells by the TUNEL assay. All the reagents were provided in the “In Situ Cell Death Detection Kit” (Roche, Indianapolis, Ind., Cat. No. 1 684 795). Formalin-fixed, paraffin-embedded tissue samples were also used to identify proliferating cells using Ki67 rabbit monoclonal antibody (NeoMarkers Cat. No. RM-9106, Lab Vision Corp., Fremont, Calif.).
Immunoflourescence: Cryosections were used for IL-1β, TNFα, MIP2α and ICAM localization. Primary antibodies include PE-rat anti-mouse IL-1β (R&D Systems, Minneapolis, Minn.), PE-rat anti-mouse TNF (BD), rat anti-mouse MIP2α (R&D Systems, Minneapolis, Minn.), rat anti-mouse ICAM (Serotec, Raleigh, N.C.), rat IgG2b isotype control (BD) (Sigma, Atlanta, Ga.). Secondary antibodies include goat anti-rat Alexa 546 (Molecular Probes, Eugene, Oreg.).
RNA Analysis: Total RNA was isolated from the ears using Qiazol (Qiagen, Valencia, Calif.). The first strand cDNA was synthesized using Reverse Transcriptase II (Invitrogen, Carlsbad, Calif.). The expression of the mouse genes was assayed by real-time quantitative PCR with an ABI PRISM 7900 Sequence Detector System (Applied Biosystems, Foster City, Calif.). Primers were designed to cross at least one intron based on the most current genome sequences using software Primer3 (Whitehead Institute for Biomedical Research, MIT). The quantitative PCR was performed using TaqMan PCR reagent kits according to the manufacturer's protocol (Applied Biosystems, Foster City, Calif.). ECM genes and cytokine mRNA levels were normalized to ubiquitin mRNA levels and expressed relative to normal untreated samples.
Results
Immunohistological analysis of skin lesions showed an increase in inflammatory cytokine expression IL-1β, TNF-α, macrophage inflammatory protein 2-alpha precursor (Mip2α) and a key adhesion molecule, intercellular adhesion molecule 1 (ICAM-1) in the inflamed ears (FIG. 28). Both topical NF-κB decoy and BMV treatment decreased the expression of IL-1β, TNF-α Mip2α and ICAM-1. Tissues treated with the scrambled ODN showed no alteration in the expression of these molecules. These observations indicate that inflamed skin cells readily take up NF-κB decoy and topical NF-κB decoy, similar to BMV, suppresses expression of key regulatory proteins.
For a quantitative measure of the anti-inflammatory effects of topical NF-κB decoy and BMV treatment in the Dp Ag-inflamed ears, real time RTPCR analysis was performed. NF-κB decoy and BMV treatment inhibits mRNA levels of both Th1 and Th2 related proinflammatory genes including chemokine (C-C motif) receptor 3 (CCR3), IL-4, ICAM, thymic stromal lymphopoietin (TSLP), IL-13, IFNγ, IL-1β and TNF-α (FIG. 29). These results confirm the above immunohistochemistry data that NF-κB decoy downregulated important inflammatory related genes and has a similar efficacy profile compared to BMV.
Since blockade of NF-κB function is associated with induction of apoptosis as well as inhibition of proliferation, NF-κB decoy treated Dp Ag-inflamed ears was analyzed by Tunnel assay and by Ki67 staining (FIG. 30). Enhanced apoptosis mediated by BMV and topical NF-κB decoy application on inflammed ears was observed, in both the skin epidermis and dermis (FIG. 30). In contrast, the scramble ODN failed to elicit a similar response. Positive Ki67 staining showed increased proliferation in the epidermal and dermal layers of inflamed ears (FIG. 30; lower panel). Comparative quantitative analyses of both topical NF-κB decoy and BMV treated inflamed tissue revealed that each treatment resulted in a similar outcome: 50-70% apoptotic cells and a 50% decrease in the mitotic index (FIG. 31). The number of Ki67-positive cells with scrambled ODN treatment remained unaffected (FIG. 31). Together, these observations indicate that NF-κB decoy exerts its anti-inflammatory action through the effective inhibition of essential regulators of inflammation and by induction of apoptosis of key immune cells.
Conclusion
Topical NF-κB decoy, similar to BMV, exerts its anti-inflammatory action by blocking expression of major inflammatory regulators and by inducing apoptosis of inflamed skin cells.

EXAMPLE 13

Lack of Skin Atrophy With NF-κB Decoy Treatment
The purpose of this experiment was to test the local skin-thinning effects of topical application of NF-κB decoy.
Method
Animal Model: Dustmite Ag (Dp)-induced dermatitis was induced in six week old male NC/Nga mice (Charles River Lab). Animals were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 12, the Dp injected ear was topically treated 2 times a day for 11 -14 days with 20 μl of vehicle, vehicle containing 0.25% or 1% NF-κB decoy, non-specific scramble ODN, or topical 0.1% betamethasone valerate (BMV) as the positive control. The ear thickness was measured with snap caliper (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
Collagen Staining: Formalin-fixed, paraffin-embedded tissue samples were used to stain collagen using sirius red F3B dye. In the dermis this stain specifically detects collagen type III. Sections are stained with Alcian blue pH=2.5, to delineate cartilage prior to picro-sirius red staining (0.1%).
Results
Ear thickness measurements of normal ears treated with BMV and 1% NF-κB decoy for 14 days application twice-a-day (BID) were performed. Notably, dramatic decrease in ear thickness was observed in response to BMV treatment (FIG. 32). NF-κB decoy failed to show any such side effects of skin atrophy (FIG. 32). Collagen I and III stain was performed on drug treated ear sections, which further confirmed skin-thinning effects of BMV (FIG. 33).
Further, topical NF-κB decoy did not cause skin atrophy in a 4-week dermal pig and a 21-week mouse carcinogenicity study.
Gene expression of key extra-cellular matrix (ECM) components, including collagen I and III, tenacin C and elastin are decreased over 50% by BMV, without much effect by NF-κB decoy (FIG. 34).
Conclusion
Long-term treatment of topical NF-κB decoy does not cause breakdown of connective tissue genes associated with skin atrophy.

EXAMPLE 14

Withdrawal of Topical NF-κB Decoy Treatment Does Not Cause Disease Rebound and Restores Epidermal Permeability
The purpose of this experiment was to test whether the cessation of NF-κB decoy treatment causes rebound of swelling and inflammation.
Method
Animal Model: Dustmite Ag (Dp)-induced dermatitis was induced in six week old male NC/Nga mice (Charles River Lab). Animals were injected intradermally with 5 μg of Dp extract (Greer Laboratories, Lenoir, N.C.) dissolved in saline on the ventral side of their right ears on days 0, 2, 4, 7, 9, and 11. Starting on day 12, the Dp injected ear was topically treated 2 times a day for 11-14 days with 20 μl of vehicle, vehicle containing 0.25% or 1% NF-κB decoy, non-specific scramble ODN, or topical 0.1% betamethasone valerate (BMV) as the positive control. The ear thickness was measured with snap caliper (Oditest, Dyer, Inc., Lancaster, Pa.) 24 hr after each intradermal injection or treatment.
TEWL Measurements: Transepidermal water loss in Dp-induced inflamed ears was measured using a Tewameter water analyzer (Courage and Khazaka, Cologne, Germany). Inflamed ears were treated with scramble ODN, 1% NF-κB decoy or BMV for 14 days twice-a-day (BID). Ears were analyzed 16-18 hrs after the last application of either scramble ODN, 1% NF-κB decoy or BMV.
PCRPK Analysis: The amount of NF-κB decoy in skin tissue was determined by using in-house developed method. Simply, the method is based on real-time quantitative PCR based assay with specific primer design and reagent modifications. The skin samples undergo alkaline lysis (using P2 buffer of Qiagen Miniprep Kit). Quantitative measure of NF-κB decoy in samples is obtained by comparing with the standard curve that is prepared by spiking known amounts of NF-κB decoy to the same biological matrix (or sample without NF-κB decoy treatment). The final value of NF-κB decoy in tissue was normalized to the net tissue weight.
Results
The abilities of NF-κB decoy and steroid treatment to cause rebound of disease after discontinuation of drug application were compared. Unlike NF-κB decoy, cessation of BMV resulted in an immediate and severe rebound of swelling and inflammation.
Polymerase Chain Reaction Pharmacokinetics (PCRPK) study was performed to determine the presence of NF-κB decoy after treatment withdrawal. Rapid clearance of NF-κB decoy to minimal levels within 3 days of withdrawal was observed (FIG. 35). Furthermore, the absence of therapeutic effective doses of decoy in the skin was confirmed by cutaneous Dp-Ag rechallenge experiments, in which the decoy treated group, showed a similar response to injected dust-mite antigen as the non-treated or scramble group (FIG. 36). These observations indicate that there is no significant accumulation of decoy in the treated tissue and no impaired immune response after 15 days of BID application.
BMV treatment withdrawal causing rebound is known to be associated with barrier disruption; therefore the effects of BMV and topical NF-κB decoy on integrity of mouse stratum corneum (SC) were determined by measuring the rate of Transepidemal Water Loss (TEWL). Increase of TEWL in inflamed ears was restored to normal levels by topical NF-κB Decoy application for 14 days (BID) (FIG. 37). In contrast, BMV treatment failed to achieve complete reduction of TEWL, indicative of the lack of normal restoration of epidermal permeability. Next, the likelihood of topical NF-κB decoy or BMV application (BID for 14 days) on normal ears to exert adverse effects on SC function was assessed. Neither BMV nor NF-κB decoy produced any changes in basal TEWL. However, unlike NF-κB decoy that shows normal kinetics of barrier recovery after acute barrier disruption with acetone, BMV treatment showed a significant delay of barrier recovery at 3 hrs, 6 hrs and 24 hrs (FIG. 38).
Increase in TEWL is associated with dysregulation of several key regulators of SC integrity. NF-κB decoy treatment does not increase expression of inflammation-associated genes like skin-derived antileucoproteinase (SKALP), cytokeratin 6 (CK6), and cytokeratin 16 (CK16) levels, as observed in inflamed skin (FIG. 39). However, BMV treatment further augments this expression (FIG. 39). The inflammation-associated cytokeratin 17 (CK17) mRNA levels were not induced in this inflammation model (data not shown). These data indicate that topical NF-κB decoy treatment restored inflammation-induced compromised barrier function and does not induce inflammation-associated cytokeratin genes.
Conclusion
There is no significant accumulation of decoy in the treated tissue and no impaired immune response after 15 days of BID application.
Further, topical NF-κB decoy treatment restored inflammation-induced compromised barrier function and does not induce inflammation-associated cytokeratin genes.

EXAMPLE 15

Administration of NF-κB Decoy Into Arthritic Joints Led to Amelioration of Collagen-Induced Arthritis
In rheumatoid arthritis (RA) NF-κB plays a pivotal role in the development of arthritis. In the present experiment, it has been investigated whether local administration of an NF-κB decoy could suppress the severity of joint inflammation.
Method
Collagen Induced Arthritis (CIA) was induced using a method previously described by Trentham et al., Arthritis Rheum 25:911-6 (1982). Briefly, 6 week old female DA rats (Charles Rivers) were immunized intradermally with 1 mg of bovine type II collagen (Chondrex) dissolved in 0.5 ml of 0.1M acetic acid at 4° C. and emulsified in 0.5 ml of cold Freund's incomplete adjuvant (Difco, Detroit, Mich.). Onset of arthritis in the ankle joints could be seen between 10 and 12 days. All rats whose onset of arthritis could not be recognized visually by day 13 were excluded from the study. On day 14 after immunization, the rats were anesthetized with isoflurane. Next 50 μl of a 100 μg solution of NF-κB decoy was injected into the articular space of the hind ankle with a 27 gauge needle. As a control, 0.5 mg of prednisolone was injected into the articular space of the hind ankle. Footpad swelling was measured with a caliper every other day for the first 2 weeks then 1× week there after.
Results
Hindpaw swelling in CIA rats injected with NF-κB decoy was measured daily for the first week and then twice a week thereafter. As shown in FIG. 8, one-time injection of NF-κB decoy markedly reduced footpad swelling starting on day 23 and continued to resolve at a faster rate while the inflammation in the scramble control gradually decreased over time. Prednisolone treated animals significantly decreased hindpaw swelling almost immediately after injection and continued to suppress swelling throughout the experiment.
Conclusion
Administration of NF-κB decoy into the arthritic joints of CIA rats led to an amelioration of arthritis.

EXAMPLE 16

NF-κB Decoy Efficacy Studies in an Adjuvant Induced Arthritis (AIA) Model
The efficacy of NF-κB decoy in the prevention and treatment arthritis was further studied in the adjuvant induced arthritis (AIA) model of arthritis.
Method
AIA was induced using a method previously described by Taurog et al. (Cell Immunol 75:271-82 (1983); Cell Immunol 80:198-204 (1983)). Briefly, 7-8 -week old female Lewis rats (Charles Rivers) were immunized intradermally at the base of the tail with 0.2 ml of a 10 mg/ml solution of Freund's adjuvant containing heat-killed mycobacterium tuberculosus H37Ra (Difco, Detroi, Mich.). Onset of arthritis in the ankle joints could be seen between 10 and 12 days. All rats whose unset of arthritis could not be recognized visually by day 12 were excluded from the study. On day 13 after immunization, the rats were anesthetized with isoflurane. Next 50 μl of a 100 μg solution of NF-κB decoy was injected into the articular space of the hind ankle with a 27 gauge needle. As a control, 0.5 mg of prednisolone was injected into the articular space of the hind ankle. Footpad swelling was measured with a caliper daily for the first week and 2 times a week thereafter.
Results
The results shown in FIG. 9 show that local injection of the NF-κB decoy into the inflamed joint in this arthritis model markedly reduced footpad swelling and continued to ameliorate the swelling while in the footpad treated with the scramble decoy, swelling was greatly increased and maintained throughout the study. Prednisolone treatment severely inhibited the footpad swelling compared to that of the scramble decoy control.

EXAMPLE 17

NF-κB Decoy Efficacy Studies in the TNBS-Induced Colitis Model
Trinitrobenzene sulfonic acid (TNBS)-induced colitis in the mouse is one of the most relevant animal models that resembles the etiology of Crohn's disease (CD) in humans. In this colitis model, intestinal inflammation develops as a result of the covalent binding of the haptenizing agent to autologous host proteins with subsequent stimulation of a delayed-type hypersensitivity to TNBS-modified self antigens. Although the relationship of this model to human disease is imperfect, the hapten-induced colitis displays CD-like features, notably transmural mononuclear inflammation and predominant Th-1 activity of the resident mucosal leukocytes. Inflammation and cytokine production in TNBS-treated mice, as well as in CD patients, is associated with activation of transcription factors such as nuclear factor NF-κB.
The procedures used for setting up the TNBS colitis model are very similar to that described in the following publications: Neurath et al., Int Rev Immunol 19:51-62 (2000); Bouma et al., Gastroenterol 123:554-565 (2002), and Bouma and Strober, Nat Rev Immunol 3:521-533 (2003).
In brief, according to this protocol under general anesthesia with isoflurane, colitis is induced in SJL/J mice by intrarectal administration of 2 mg TNBS in a 45% ethanol solution. Intrarectal injection is administered with a 3.5F polyurethane umbilical catheter equipped with a 1 -ml syringe. The catheter is inserted so that the tip is 3 cm proximal to the anal verge and the TNBS was injected with a total volume of 100 μl. To ensure distribution of the TNBS within the entire colon and cecum, mice are held in a vertical position for 30 s after the injection. Changes in body weight are monitored daily, which typically amount to weight losses of 15-30% occurring over a period of approximately 1-4 weeks after TNBS administration. The mice are treated with decoy in the same intrarectal manner with a single administration either prior to TNBS (prophylactic) or at various time points post TNBS induction (therapeutic). In some instances a second decoy treatment may be administered to prolong therapeutic effects if deemed necessary. In addition to body weight measurements, in certain experiments animals are sacrificed at different time points to harvest tissues for mRNA and/or protein expression and histological analysis.
Following this protocol, a single administration of NF-κB decoy given two days post disease induction can substantially reverse the weight loss associated with TNBS colitis (FIG. 10). Weight loss in this model is closely correlated with inflammation and disease severity. By day 2 post induction the mice have generally lost 15-20% body weight indicating disease onset. When vehicle alone was administered at this point, no ameliorative effects were seen and the mice in this group continued to lose weight and suffer from a high mortality rate. In treatment groups receiving either 50 or 100 μg NF-κB decoy on day 2, the disease associated weight loss was reversed very rapidly with almost complete recovery achieved by day 7. There was also a much higher survival rate in the decoy treated groups. The efficacy of NF-κB decoy was corroborated at the histopathological level in colons harvested from mice in both treatment groups at the day 7 post induction time point (FIG. 11). The TNBS-induced animals treated with vehicle alone showed localized inflammation, loss of the mucosal epithelium, an increase in crypt depth (with some sign of crypt branching), dissolution of the muscularis mucosae, and thickening of the muscularis layer (localized to the site of the inflammation). All these features are consistent with human inflammatory bowel disease (IBD) and the animal models presented in the literature. A single treatment with NF-κB decoy on day 2 post induction was able to reverse many of these pathological hallmarks of IBD.

EXAMPLE 18

NF-κB Decoy Efficacy Studies in the Oxazolone Colitis Model
A series of preclinical experiments have been initiated to evaluate the therapeutic potential of an NF-κB decoy of the present invention in inflammatory bowel disease. Oxazolone induced colitis in the mouse is a relevant model for studying disease pathology related to ulcerative colitis (UC) in humans. In this colitis model, intestinal inflammation develops as a result of covalent binding of the haptenizing agent to autologous host proteins with subsequent stimulation of a delayed-type hypersensitivity to oxazolone-mediated self antigens. Using the “classical” haptenazing agent, oxazolone elicits an inflammatory bowel disease involving the distal half of the colon, and has histologic features resembling UC rather than Crohn's disease. In addition, oxazolone colitis is driven by a Th2 as opposed to Th1 response. NF-κB regulates many of the cytokines, chemokines, and cell adhesion molecules contributing to this response.
The procedures used to set up this oxazolone colitis model were similar to those described by Neurath et al., Int Rev Immunol 19:51-62 (2000); Bouma et al., Gastroenterol 123:554-565 (2002), and Bouma and Strober, Nat Rev Immunol 3:521-533 (2003).
In brief, under general anesthesia with isoflurane, mice were presensitized by applying 200 μl of a 3% (w/v) solution of oxazolone (ethoxymethylene-2-phenyl-2-oxazolin-5-one) in 100% ethanol to a 2×2 cm field of shaved abdominal skin. Five days after presensitization, mice were rechallenged intrarectally with 150 μl of 1.5% oxazolone in 50% ethanol or only 50% ethanol, again under general anesthesia with isoflurane. Intrarectal injection was administered with a 3.5F polyurethane umbilical catheter equipped with a 1-ml syringe. The catheter was inserted so that the tip was 3 cm proximal to the anal verge and the oxazolone was injected with a total volume of 150 μl. To ensure distribution of the oxazolone within the entire colon and cecum, mice were held in a vertical position for 30 seconds after the injection. Changes in body weight were monitored daily. Typically 15-20% weight losses were observed over a period of approximately 5-7 days after oxazolone administration. The oxazolone colitis model has an accurate disease progression from which the animals fully recover within 7-10 days.
The mice were treated with the NF-κB decoy in the same intrarectal manner with a single administration either prior to intrarectal oxazolone challenge (prophylactic) or at various time points post oxazolone challenge (therapeutic). In some instances, a second decoy treatment may be administered to prolong therapeutic effects if deemed necessary. In addition to body weight measurements, in certain experiments animals were sacrificed at various time points to harvest tissues for mRNA expression and histological analysis.
It has been found that a single administration of the NF-κB decoy given two days post disease induction could substantially reverse the weight loss associated with oxazolone colitis (FIG. 12). Weight loss in this model is closely associated with inflammation and disease severity. By day 2 post induction, the mice have generally lost 15-20% body weight indicating disease onset. When vehicle alone was administered at this point, no ameliorative effects were see, and the mice in this group continued to lose weight and suffer from a high mortality rate. In treatment groups receiving either 50 or 100 μg NF-κB decoy on day 2, the disease associated weight loss was reversed very rapidly with complete recovery observed in the 100 μg decoy group within 3 days of treatment. There was also a much higher survival rate in the decoy treatment groups (FIG. 13). By day 3, there was 60-70% mortality rate in the vehicle treated group. In contrast, in the treated groups there was only a 20-30% mortality rate.

EXAMPLE 19

Additional Studies on TNBS-Induced Colitis Model
To further establish the therapeutic efficacy of locally administered NF-κB decoy in inflammatory bowl disease (IBD), Examples 19, 20 and 21 show additional studies in three different mouse models of acute colitis that display unique features related to human IBD.
The NF-κB decoy used in the experiments described in Examples 19, 20 and 21 was a 19 base pair (molecular weight of 13 kDa) double-stranded deoxyribonucleic acid comprised of a random mixture of R- and S-chiral forms of oligonucleotides. The sequence of the NF-κB decoy used in this study was: 5′AGTTGAGGACTTTCCAGGC-3′ (SEQ ID NO: 30) and 5′-GCCTGGAAAGTCCTCAACT-3′ (SEQ ID NO: 95). The scrambled oligonucleotides (ODN) sequence was 5′-CAGTAGTATGTGAGCCTGC-3′ (SEQ ID NO: 46) and 5′-GCAGGCTCACATACTACTG-3′ (SEQ ID NO: 96). All oligonucleotides used were phosphorothioates and were annealed by combining equal molarities of each strand at room temperature.
Trinitrobenzene sulfonic acid (TNBS) colitis was induced in 5-6 week old male SJL/J mice. 2.5 mg TNBS (Sigma, Atlanta, Ga.) dissolved in 45% ethanol was administered intrarectally on day 0 with a 3.5F polyurethane umbilical catheter (Supplier) equipped with a 1-ml syringe. The catheter is inserted so that the tip is 4 cm proximal to the anal verge and the TNBS is injected with a total volume of 150 μl. To ensure distribution of the TNBS within the entire colon and cecum, mice are held in a vertical position for 30 s after the injection. Changes in body weight are monitored daily. On days 2 and 4 post induction, the animals are treated in the same intrarectal manner with 150 μl of vehicle, vehicle containing specified dose of NF-κB decoy or scramble ODN (negative control ODN), or 0.3 mg/kg budesonide (Sigma, St. Louis, Mo.) as the steroid control. The animals are monitored for 7-10 days then sacrificed to harvest tissues for mRNA and/or protein expression and histological analysis.
Histological Analysis
Harvested colons were formalin-fixed in a Swiss-roll orientation, embedded in paraffin, sectioned longitudinally and stained with hematoxylin-and-eosin, Peroidic acid-Schiff (PAS) or Alcian blue stains. Crypt loss and inflammation were assessed based on severity of lesions and the extent of tissue damage. The colon sections were scored in a blinded fashion by an independent veterinary pathologist using previously described criteria. Fort M M, et al., J Immunol 2005;174:6416-23.
Immunohistochemistry
In addition to assessing the ability of NF-κB decoy treatment to mitigate pro-inflammatory responses in colitis discussed below, the ability of NF-κB decoy treatment to affect mucosal repair processes was evaluated. Mucosal healing in the gastrointestinal tract requires both restitution and regeneration of the epithelium. (See Okamoto and Watanabe, Dig Dis Sci, 50 Suppl 1:S34-8 (2005); Taupin and Podolsky, Nat Rev Mol Cell Biol, 4:721-32 (2003)). Restitution promotes epithelial-cell migration to reseal lesions and re-establish surface-cell continuity. The trefoil factor family of secreted peptides play an essential role in protecting the mucosal epithelium from a range of insults and contribute to restitution and repair. NF-κB activation has been specifically associated with decreased intestinal trefoil factor (ITF) expression in inflamed epithelium. (See Loncar et al., Gut 52:1297-303 (2003); Dossinger et al., Cell Physiol Biochem 12:197-206 (2002)).
Formalin-fixed paraffin-embedded sections were used for immunohistochemstry following heat-induced epitope retrieval using a modified citrate buffer, pH=6.1 (Target retrieval solution, DAKO, Carpinteria, Calif.) in a pressure cooker (Retrever 2100, EM Sciences, Hatfield, Pa.). The immunohistochemical staining was carried out with Autostainer (DAKO, Carpinteria, Calif.) using rabbit monoclonal anti-Ki67 (LabVision, Freemont, Calif.), peroxidase-conjugated mouse anti-smooth muscle actin, EPOS-SMA (DAKO, Carpinteria, Calif.), goat anti ITF (intestinal trefoil factor) (Santa Cruz Biotechnologies, Inc., Santa Cruz Calif.). The primary antibodies were localized by anti-rabbit SuperPicture-HRP reagent (ZYMED, South San Francisco, Calif.), biotinylated anti-goat IgG, biotinylated anti-rat IgG and ABC-HRP (Vector Laboratories, Inc., Burlingame, Calif.). Peroxidase activity was revealed using DAB substrate. The sections were counterstained with hematoxylin before coverslipping.
Immunohistochemical assessment of ITF levels revealed very low levels of staining in the inflamed vehicle-treated control mice (FIG. 40A). After two treatments of 1.5 mg/kg NF-κB decoy, the production of ITF by goblet cells was pronounced in the mucosal epithelium at day 7 post TNBS induction (FIG. 40B). Regeneration, on the other hand, involves epithelial proliferation and differentiation to restore normal function. Reduced goblet cell function is a classic pathological hallmark of disease in murine and human colitis. Re-emergence of secretory granules from goblet cells is evident in the H&E stained sections from mice treated with NF-κB decoy (FIG. 41). The ability of these highly differentiated cells to produce mucopolysaccharides was determined by PAS (FIGS. 40C and 40D) and Alcian blue (FIGS. 40E and 40F) staining of colonic tissues. In both cases, the colons from mice treated with NF-κB decoy demonstrated a dramatic increase in the levels of mucopolysaccharide production in the tissue as compared to saline controls.
To assess changes in cell proliferation due to NF-κB decoy treatment, colons from TNBS-induced mice were stained for Ki67. In the inflamed colons of vehicle treated mice high levels of proliferation were observed not only in the mucosal epithelium lining the crypts, but also throughout the lamina propria and muscularis layers (FIGS. 42A and 42C). In NF-κB decoy treated mice, normal levels of proliferation were observed in the lymphoid follicles (FIG. 42D) suggesting a return to homeostatic conditions. To further evaluate the condition of the colonic muscularis, smooth muscle actin levels were examined. While inflamed colons showed very disrupted actin patterns associated with increased intercellular space, the muscularis of NF-κB decoy treated mice was uniform and well compacted (FIG. 42E versus 42F). These findings suggest the ability of NF-κB decoy to aid the repair of damaged intestinal epithelium and restore colon function.
Laser Scanning Confocal Microscopy
In order to assess NF-κB decoy uptake into the inflamed colon, fluorescently labeled decoy (HEX-NFκB decoy) was intrarectally administered into mice with TNBS-induced colitis as discussed below.
For decoy localization studies, 6-carboxy-2′,4,4′,5′,7,7′-hexachlorofluorescein, succinimidyl ester labeled NF-κB decoy (HEX-NFκB decoy) was applied in an enema form in TNBS-treated mice. Colon harvested after 6 hours was embedded and frozen in Tissue-Tek OTC following formalin fixation. The whole colon of HEX-NFκB decoy treated animals was embedded and frozen in Tissue-Tek OTC following formalin fixation. Frozen sections (10 micrometer thickness) were stained with Sytox Green (Molecular Probes, Inc., Eugene, Oreg.) and mounted in glycergel (DAKO, Carpinteria, Calif.). The frozen sections were photographed using a NIKON laser scanning confocal microscope (NIKON USA, NJ). A NIKON Eclipse E800 light microscope was fitted with the PCM 2000 confocal system. The fluorochromes were sequentially excited with a HeNe laser followed by an Argon laser (using the 488 line for excitation). The confocal images were captured with Simple PCI software, v. 4.0.6 (Compix Inc., Cranberry Township, Pa.).
Throughout the distal colon, the HEX-NFκB decoy concentrated specifically at sites of inflammation, with little to no decoy accumulation observed in non-inflamed areas where the mucosal epithelium is intact. HEX-NFκB decoy uptake and /or accumulation were not seen in the cells of intact mucosal epithelium or in cells in the underlying lamina propria (FIG. 43A). In the inflamed areas with evidence of ulceration, HEX-NFκB decoy accumulation was detected not only in the luminal inflammatory cells but also in cells throughout the lamina propria all the way to the muscularis mucosae (FIG. 43B). Co-localization studies with Sytox green nuclear counterstain revealed that HEX-decoy delivered to cells within the inflamed lesions is in fact primarily localized in the cell nuclei after 6 hours, the site of drug action. Although NFκB decoy was found distributed throughout the colon, from the cecum to the rectum, with this route of administration it was found in a majority but not all lesions after only two applications. The nuclear localization and the restriction to highly inflamed regions indicate a selective delivery of the therapeutic agent to important tissue target sites, i.e. inflamed areas with ulcerative lesions. Co-localization studies with cell-type specific markers revealed uptake by neutrophils, macrophages, and CD4⁺T-cells (data not shown). In order to assess the amount of drug delivered, NFκB decoy levels were measured in the colonic tissue using a quantitative pharmacokinetic-PCR (PCRPK) assay discussed below.
Pharmacokinetic-PCR Assay
The amount of NF-κB decoy in tissue was determined using an in-house developed pharmacokinetic-PCR (PCRPK) method based on a real-time quantitative PCR assay with specifically designed primers and modified reaction mixture.
Six hours after the first intrarectal treatment of 1.5 mg/kg NF-κB decoy on day 2, the mean level of NF-κB decoy detected was 420±90 pg per mg of tissue (FIG. 44). This amounts to approximately 0.27% of the original 1.5 mg/kg dose being retained in the tissue. The detectable level of NF-κB decoy remained constant over 24 hrs (450±130 pg/mg at day 3) and then showed a 45% decline at day 4 (250±80 pg/mg) despite the application of a second dose 6 hrs prior to this harvest time point. By day 6, 48 hours after the last NF-κB decoy treatment, only 40±10 pg/mg were detected. The amount of measurable NF-κB decoy in the tissue was dependent on applied dose. Overall, this data indicates the successful non-viral delivery of our NF-κB decoy into the colonic tissue after topical application, targeting mostly inflamed lesions and infiltrating cells.
NF-κB TransAM Assay
To verify the targeted blockade of NF-κB function upon NF-κB decoy treatment, colons from TNBS-induced mice were analyzed for tissue NF-κB levels at various time points post treatment using an electrophoretic mobility shift assay.
Nuclear extracts prepared from colon tissue were evaluated for transcription factor levels using the NF-κB TransAM assay (Active Motif, Carlsbad, Calif.) according to manufacturer's instructions.
As expected during an inflammatory response, increased binding to a radiolabeled ODN containing NF-κB-binding sites was observed in tissue extracts from vehicle treated mice indicating elevated transcription factor levels (FIG. 45). A reduction in the density of the p65/p50 band was observed in tissues treated with 1.5 mg/kg NF-κB decoy. The identity of the p65, RelB and p50 NF-κB sub-units was confirmed by a reduction in mobility following incubation of the sample extracts with specific antibodies. Kinetic analysis of p65 levels in the TNBS model was carried out using an ELISA-based TransAM assay. p65 binding in the vehicle treated animals increased through day 6 reaching a peak level of 0.82±0.02 A450 units and remaining elevated through day 8 (FIG. 46). In contrast, the amount of p65 available for binding was mitigated upon treatment with 1.5 mg/kg NF-κB decoy beginning on day 2, and after a second dose at day 4 p65 levels remained substantially lower than vehicle controls through day 7 (0.61±0.09 A450 units). Regression analysis of the curve fits for each of these data sets revealed the differences in p65 activation to be statistically significant (p=0.03), thus indicating that a naked phosphorothioated NF-κB decoy is able to significantly reduce levels of this key transcription factor in vivo.
Myeloperoxidase Assay
Infiltration of leukocytes is a hallmark of colitis pathology. Using CD45 as a pan-leukocyte marker, the level of inflammatory cell accumulation by immunohistochemistry was evaluated (FIG. 47).
Myeloperoxidase (MPO) activity can be used as a quantitative measure of the extent of neutrophil infiltration into inflamed tissue. To measure myeloperoxidase (MPO) activity, frozen colonic tissue was pulverized then homogenized in T-PER protein extraction reagent (Pierce, Rockford, Ill.). Homogenates were centrifuged at 16,000×g for 15 min then MPO activity was measured by adding 50 ml of supernatant to 250 ml of an assay reaction mixture containing 0.5% hexadecyltrimethylammonium bromide (in 50 mM potassium phosphate, pH 6.4), 0.17 mg/ml o-dianisidine dihydrochloride, and 0.0015% hydrogen peroxide. After a 30 min incubation period at room temperature, absorbance at 450 nm was measured. Absolute MPO activity was determined based on the generation of a standard curve and normalized with protein concentration as determined by BCA assay (Pierce, Rockford, Ill.).
Seven days after TNBS-induced colitis, the amount of inflammatory cell infiltration was clearly reduced in mice treated intrarectally with 1.5 mg/kg NF-κB decoy as compared to saline vehicle controls (FIG. 47A versus 47B). Using a specific immunohistochemical marker for murine neutrophils the experiment showed that a large proportion of the inflammatory infiltrates are in fact, neutrophils (FIGS. 47C and 47D). Neutrophil accumulation was predominant not only in the ulcerated areas but also in the expanded lamina propria between the colonic crypts. The TNBS-induced acute inflammatory response in the colon is specifically associated with an increase in the infiltration of activated neutrophils into the gut mucosa. (See Buell and Berin, Dig Dis Sci, 39:2575-88 (1994); Yamada et al., Gastroenterology, 102:1524-34 (1992)).
Colitis induction with TNBS followed by vehicle treatment resulted in substantially increased MPO activity levels that peaked at post induction day 2 but remained elevated at the day 7 endpoint. Treatment with 1.5 mg/kg NF-κB decoy on days 2 and 4 post disease induction significantly decreased the colonic MPO activity by 52% at day 7 (p<0.05) (FIG. 48). The scramble ODN had no protective effect, ruling out non-specific inhibitory mechanisms. In summary, this data indicates the ability of NF-κB decoy to reduce infiltration and activity of inflammatory cells in colonic tissue and that the effectiveness of NF-κB decoy was sequence dependent and superior to budesonide.
Cytokine mRNA Quantification
Cytokines and chemokines play a central role in the modulation of the intestinal immune system and there is extensive evidence illustrating their dysregulation in inflammatory bowel disease. Therefore, key inflammatory mediators in the colon were examined to assess the efficacy of NF-κB decoy treatment at the molecular level.
At the time of sacrifice, the colon was removed from the colo-caecal junction to the anal verge. Harvested colons were flushed to remove fecal material and flash-frozen in liquid nitrogen and stored at −80° C. Total RNA was isolated from the colons using Qiazol (Qiagen; Valencia, Calif.). The expression of the mouse genes was assayed by real-time quantitative PCR with an ABI PRISM 7900 Sequence Detector System (Applied Biosystems, Foster City, Calif.). Primers were synthesized at Sigma Genosys. The quantitative PCR was performed using TaqMan PCR reagent kits according to the manufacturer's protocol (Applied Biosystems, Foster City, Calif.). Cytokine mRNA levels are expressed relative to ubiquitin mRNA levels.
As determined by real-time PCR, TNBS-induced inflammation resulted in upregulation of TNFα, IL-1β, IL-6, IL-12 and MCP-1 mRNA levels (FIG. 49). Treatment with 1.5 mg/kg NF-κB decoy resulted in a 5-fold reduction in TNFα levels as compared to vehicle controls at the day 7 endpoint (1.1±0.4 versus 5.1±1.3, respectively; p<0.05). The highly expressed cytokines IL-6 and IL-1β were reduced by NF-κB decoy treatment 2- and 4-fold (P<0.05); respectively. IL-12, a key Th1-inducing cytokine, was down-regulated nearly 4-fold in NF-κB decoy treated mice (data not shown). The monocyte chemoattract protein, MCP-1, decreased approximately 3-fold after NF-κB decoy treatment (p<0.05). In the cases of TNFα and IL-1β, NF-κB decoy treatment demonstrated even greater inhibition than the steroid treated group. The scramble ODN (1.5 mg/kg) had little to no effect on mRNA levels. These important mediators of the inflammatory response are all controlled by NF-κB and affect the differentiation and activation of key immune cells, including macrophages, lymphocytes and polymorphonuclear neutrophils. The ability of NF-κB decoy to down-regulate these important inflammatory mediators indicates one of the mechanisms by which efficacy is achieved in ameloriating colitis.
Statistics
Data are presented as mean values±S.E.M. Statistical significance was determined using the Student's unpaired t-test for two group analysis and by ANOVA with post hoc comparisons by Dunnett's Multiple Comparison Test for multiple group analysis (GraphPad Software, San Diego Calif.).

EXAMPLE 20

Additional Studies on Oxazolone-Induced Colitis Model
The purpose of this experiment was to determine the therapeutic potential of our NF-κB decoy to treat a Th2 cytokine-mediated colitis model.
Oxazolone-induced colitis was initiated by presensitizing C57BL/10 mice with a 3% (w/v) solution of oxazolone (4-ethoxymethylene-2-phenyl-2-oxazolin-5-one; Sigma, Atlanta, Ga.) in 100% ethanol applied topically to a 2×2 cm field of shaved abdominal skin. 5-7 days after presensitization mice are rechallenged intrarectally with 150 ul of 1.5% oxazolone in 50% ethanol or 50% ethanol alone under general anesthesia with isoflurane. The mice are treated with 1.5 or 5 mg/kg NF-κB decoy or 0.3 mg/kg budesonide in the same intrarectal manner with a single administration either prior to intrarectal oxazolone challenge (prophylactic) or at various time points post oxazolone challenge (therapeutic).
Animals receiving saline vehicle demonstrated a 17% weight loss two days after disease induction, which was followed by a steady recovery in body weight that returned to the original starting value by day 5-6 (FIG. 50). While neither budesonide nor NF-κB decoy was able to completely prevent disease onset, presumably as a result of the ethanol damage caused during induction, these two treatment groups recovered much more rapidly than vehicle controls. By day 2, both the NF-κB decoy and budesonide treatment groups began recovering body weight, while the vehicle group remained non-responsive. A significant difference between treated and vehicle groups was observed at day 3, at which point the NF-κB decoy and budesonide treated groups had returned to their approximate starting weights while the vehicle group was still down 12% from its original mean weight (p<0.01). There was a moderate but non-statistically significant increase in survival associated with NF-κB decoy and budesonide (62% and 69%, respectively) when compared with the 35% survival in vehicle controls (data not shown). At the day 6 endpoint, despite what appeared to be a full recovery in body weight, there was a remarkable amount of damage still present throughout the vehicle treated colons (FIG. 5 1). However, in the NF-κB decoy treated group colonic damage was minimal at this time point and limited to the most distal portion of the colon. Quantitative analysis of the histopathological scores indicated over 70% reduction in the level of colitic damage between NF-κB decoy and vehicle treated controls (7.2±2.6 versus 26.7±7.3, respectively) (FIG. 52). Histological scores in the budesonide-treated group were also improved but to a lesser degree than the NF-κB decoy treated mice. Animals receiving NF-κB decoy treatment showed greater improvement than those given budesonide in a majority of the pathological parameters evaluated; including mucoid crypt cysts and abscesses, goblet cell depletion, edema, mononuclear cell infiltration and transmural inflammation. Overall, this data shows that topical delivery of naked NF-κB decoy has a better efficacy profile than budesonide in a Th2-mediated disease model.

EXAMPLE 21

Dextran Sulfate Sodium (DSS)-Induced Colitis Model
The purpose of this experiment was to examine NF-κB decoy efficacy in acute and chronic dextran sodium sulphate (DSS)-induced colitis models which recapitulate epithelial injury conditions and have been characterized by a mixed Th1/Th2 cytokine response.
To induce DSS colitis, female BALB/C mice were fed 5% dextran sodium sulfate (DSS; 35-45 kDa; MP Biomedicals) dissolved in drinking water for 10 days. Animals were administered 2.5 mg/kg NF-κB decoy, scramble ODN (negative control ODN), or vehicle every other day by intrarectal administration from day 0 to day 9. A disease activity index (DAI) was calculated on day 10 based on the summation of scores evaluating weight loss, blood in stool, and stool consistency. Each parameter was graded on a scale of 0 to 4, and the combined scores were divided by 3 to obtain the final disease activity index. The presence of blood in the stool was evaluated with Hemoccult test strips (Supplier). Colon lengths were also measured when mice were sacrificed. Chronic colitis was induced by giving the mice 4% DSS in their drinking water for two cycles. The first cycle of DSS administration was followed by 14 days of plain drinking water. The second cycle of DSS was followed by 3 days of plain drinking water, after which the mice were sacrificed. NF-κB decoy (5 mg/kg) was intrarectally administered every other day for a total of 5 treatments beginning the first day of each DSS cycle.
Animals receiving vehicle only or scramble ODN control treatment achieved mean disease activity index scores of 1.6±0.2 and 1.5±0.3, respectively (FIG. 53). Remarkably, the disease activity index was reduced over 50% in the NF-κB decoy treatment group compared to vehicle and scramble controls, reaching a statistically significant value of 0.78±0.2 (p<0.05). Surprisingly, the steroid treatment was ineffective and in some cases even exacerbating in this model (data not shown). At the day 10 endpoint, in the acute DSS model, crypt elongation, severe lymphocytic infiltration, goblet cell loss, and a high frequency of lymphoid aggregates in the submucosum of vehicle treated animals were observed (FIG. 54). Similar to the quantitative difference in DAI scores in this model, the NF-κB decoy treated colons had significantly reduced histopathology scores, decreased over 55% as compared to vehicle controls (21.5±5.0 versus 47.9±5.6, respectively) (FIG. 55). The recovery of colonic morphology was apparent with resolution of crypt length, return of goblet cells, and reduced number and size of lymphoid aggregates.
A repeated inflammation cycle protocol for the DSS model was used to assess the effectiveness of NF-κB decoy treatment under more clinically-related chronic inflammatory conditions. Treatment in this model was given therapeutically, beginning at the start of the 2^ndcycle. Ten days into the second DSS cycle, the NF-κB decoy treated group had a significantly lower DAI score as compared to the vehicle control (FIG. 56) (p<0.0015). The decreased DAI score in the NF-κB decoy treated mice comprised reductions of over 30% in weight loss (1.9±0.3 versus 2.9±0.2) and stool consistency scores (1.6±0.2 versus 2.4±0.2), and a reduction of over 80% in the amount of fecal occult blood (0.1±0.1 versus 0.6±0.2) observed in this group. No effect was observed with the scramble ODN. Histologically, the NF-κB decoy treated colons still displayed some intermittent areas of inflammation but overall the colitic damage was substantially reduced compared to vehicle treated controls which had very high levels of infiltrates, lamina propria edema, disrupted crypt morphology, loss of goblet cells, and large well-developed lymphoid aggregates (data not shown). The amelioration of disease in the DSS induced colitis model shows the therapeutic potential of NF-κB decoy in a mixed Th1/Th2 mediated disease. These findings in the chronic colitis model further show that therapeutic treatment with NF-κB decoy is also efficacious under chronic inflammatory conditions.
All references cited throughout this disclosure are hereby expressly incorporated by reference.
Although the present invention is illustrated with reference to certain specific embodiments, it is not so limited. Modifications and variations are possible without diverting from the idea of the invention, and will be apparent to those skilled in the art. All such modifications and variations are specifically within the scope herein.

Specificity/

1. A method for the treatment of an inflammatory bowel disease (IBD), comprising administering to a mammalian subject in need an effective amount of an NF-κB double-stranded decoy oligodeoxynucleotide (dsODN) molecule.

2. The method of claim 1 wherein said dsODN comprises in its sense strand, in 5′ to 3′ direction, the sequence of SEQ ID NO:30.

3. The method of claim 2 wherein the antisense strand of said dsODN molecule is at least partially complementary to said sense strand.

4. The method of claim 3 wherein the dsODN molecule further comprises at least one single-stranded overhang.

5. The method of claim 3 wherein the two strands of the dsODN are linked at the 5′ and/or 3′ end by a covalent bond, other than a peptide bond.

6. The method of claim 3 wherein the dsODN molecule comprises modified nucleotides.

7. The method of claim 3 wherein the dsODN molecule has a phosphodiester backbone.

8. The method of claim 3 wherein the dsODN molecule has a phosphorothioate backbone.

9. The method of claim 3 wherein the dsODN molecule has a mixed phosphodiester-phosphorothioate backbone.

10. The method of claim 2 wherein the antisense strand of said dsODN molecule is fully complementary to said sense strand.

11. The method of claim 2 wherein said dsODN molecule has a phosphodiesterate backbone.

12. The method of claim 2 wherein said dsODN molecule has a phosphorothioate backbone.

13. The method of claim 2 wherein said dsODN molecule has a mixed phosphodiesterate-phosphorothioate backbone.

14. The method of claim 2 wherein the sense and antisense strands are connected to each other solely by Watson-Crick base pairing.

15. The method of claim 2 wherein the sense and antisense strands are connected, completely or partially, by cross-links other than Watson-Crick base pairing.

16. The method of claim 2 wherein the sense and antisense strands are covalently linked to each other at their 3′ and/or 5′ end.

17. The method of claim 2 wherein the dsODN molecule comprises modified nucleotides.

18. The method of claim 1 wherein said mammalian subject is human.

19. The method of claim 1 wherein said dsODN molecule is administered by pressure mediated transfection.

20. The method of claim 1 wherein said dsODN molecule is administered by retroviral transfection.

21. The method of claim 1 wherein said dsODN molecule is administered in liposomes.

22. The method of claim 1 wherein said dsODN molecule is administered as a topical formulation.

23. The method of claim 1 wherein said dsODN molecule is administered as an oral formulation.