WO1997039120A2 - Antisense inhibitors of vascular endothelial growth factor (vefg/vpf) expression - Google Patents

Abstract

The present invention relates to the inhibition of vascular endothelial growth factor expression with oligonucleotides. The oligonucleotides of the present invention are thought to bind to target nRNA in a sequence specific manner and prevent expression of the encoded gene. Chemical modifications of the oligonucleotides for increasing their stability and binding efficiency are disclosed. These modifications increase the stability and the efficiency of the oligonucleotides contemplated in this invention. Oligonucleotides compositions can be used in ex vivo therapies for the treatment of macrophages or in vivo therapies by injection, inhalation, topical treatment or other routes of administration.

Description

ANTISENSE INHIBITORS OF VASCULAR ENDOTHELIAL GROWTH FACTOR (VEgF/VPF) EX_¬ PRESSION

CROSS-REFERENCE TO RELATED APPLICATIONS This application depends for pπoπty upon a copending provisional application having Serial Number 60/015,752, filed Apπl 17, 1996.

STATEMENT REGARDING FEDERALLY SPONSORED

RESEARCH OR DEVELOPMENT

Not applicable.

BACKGROUND OF THE INVENTION Field ofthe Invention

The present invention relates to the cellular inhibition of vascular endothelial growth factor expression with oligonucleotides. The oligonucleotides of the present invention are thought to bind to target mRNA m a sequence specific manner and prevent expression of the encoded VEGF gene. Chemical modifications to the oligonucleotides are disclosed for increasing the stability and binding efficiency of the oligonucleotides. The present oligonucleotide compositions can be used m ex vivo therapies for the treatment of macrophages or in vivo therapies by injection, inhalation, topical treatment or other routes of administration. Description ofthe Related Art

Vascular endothelial growth factor (VEGF), also known as vascular permeability factor, compπses a family of homodimeπc secretory glycoproteins ranging in size from 34 to 46 kilodaltons.

It is secreted by a vanety of cell types in response to hypoxia and certain regulatory factors. Four isotypes of VEGF are known. They anse by alternative splicing of mRNA from a single gene. (Keck et al., 1989; Leung et al., 1989; Connolly and Plander, 1989; Tischer et al., 1991).

VEGF is necessary for the formation of blood vessels (angiogenesis) dunng growth and developmental processes, and for tissue repair. (Ferrera, et al., 1996; Carmeliet et al., 1996; Thomas, 1996; Dvorak et al., 1995a,b; Folkman, 1995; Ferrera, et al., 1992). This growth factor induces vascular permeability, is a chemotactic for monocytes and osteoblasts, and is a selective mitogen for endothelial cells. Receptor proteins for VEGF (KDR and Flt-1 in humans) belong to the transmembrane tyrosine kinase family. Overman et al., 1992; de Vnes et al., 1992). Activation ofthe receptor initiates a cascade of events leading to markedly enhanced rates of vascular endothelial cell proliferation and eventual neovascularization. VEGF is more selective at inducing endothelial cell proliferation than any other protein factor involved in angiogemsis. Unfortunately, under certain conditions, the presence of VEGF may have deleterious health effects.

Abnormally high concentrations of VEGF are associated with diseases charactenzed by a high degree of vascularization or vascular permeability. Examples of such afflictions include diabetic retinopathy, aggressive cancers, psonasis, rheumatoid arthritis, and other inflammatory conditions. (D'Amore, 1994; Dvorak et al., 1995 a,b; Folkman, 1995). Compositions and methods are needed for selectively decreasing abnormally high VEGF concentrations m order to reduce VEGF-mediated neovascularization. These methods and compositions can be used to slow the progression of diseases characterized by vascularization and vascular permeability. One method for reducing VEGF concentrations involves the use of antisense oligonucleotides. (Wagner, 1994). The central advantage of this technique is the specificity with which inhibition can be achieved. Useful oligonucleotides are thought to bind specific sequences of mRNA and interfere with the expression of encoded genes. Reduced protein expression may result from the inhibition of ribosome function, reduced concentrations of translatable substrate mRNA. or other mechanisms. In addition, oligonucleotides can reduce mRNA concentrations by an ohgonucleotide-mediated increase in the rate of degradation of mRNA molecules. Generally, oligonucleotides of approximately 15 bases are sufficient to provide sequence-specific binding to intended RNA targets, although shorter oligonucleotides do sometimes bind. (Uhlmann and Peyman 1990). However, antisense oligonucleotides having between 11-30 bases have been used to reduce protein expression in in vitro expenments. (Reviewed in Uhlman and Peyman, 1990).

A number of obstacles must be overcome before the potential advantages of an antisense treatment strategy can be realized in treating disease. For example, antisense oligonucleotides are large (-3,000 to 10.000 D) hydrophilic compounds and must cross hydrophobic cellular membranes before binding their targets in the cytosol or nucleus. (Uhlmann and Peyman. 1990: Milhgan et al., 1993). Thus, methods are needed to facilitate transport of VEGF antisense oligonucleotides across cell membranes. Therapeutic oligonucleotides must also be nontoxic and should not interfere with normal cellular metabolism. To minimize these nonspecific effects, they must bind their cognate sequences with high specificity and affinity.

Oligonucleotides with a natural phosphodiester backbone are highly susceptible to serum and cellular nucleases. Random 17 base-long oligonucleotide sequences have a half-life of less than 3 minutes in serum (Bishop et al., 1996). Oligonucleotides with increased stabilities are needed before they can be used as therapeutics in the treatment of neovascular disease. Substitution of the phosphodiester groups with phosphorothiotates to increase oligonucleotide half-lifes. They should be chemically inert and nuclease resistant m a vanety of chemical environments. However, such oligonucleotides have not previously been shown to inhibit VEGF expression in a selective manner. One disadvantage of previously known phosphorothiotate oligonucleotides is that they require concentrations over 1 micromolar (μM) to reduce VEGF expression (Nomura et al., 1995, Robinson et al., 1996) At these concentrations, those oligonucleotides are toxic (Woolf et al., 1992; Stem and Cheng, 1993; Stein and Kreig, 1994, Wagner, 1994; Fennewald et al., 1996) and the observed effects probably are the result of this nonspecific toxicity (Fennewald et al., 1995). Novel oligonucleotide inhibitors are needed that demonstrate a true antisense effect by inhibiting VEGF expression at nontoxic concentrations. These oligonucleotides will likely have higher association constants and/or an increased specificity for their target mRNA sequences than pnor VEGF antisense oligonucleotides. There are several possible explanations for the limited effectiveness of pnor VEGF antisense oligonucleotides. One possibility is that target RNA sequences may be confined in macromolecular structures that stencally block oligonucleotide binding For example, RNA binding proteins and protein translation complexes may block oligonucleotide binding Alternatively, oligonucleotides may not be able to bind unfavorable conformations of the mRNA In addition, the location of effective target sequences is vanable. Effective target sequences may be located anywhere on target mRNA transcπpts and oligonucleotides targeted to translation initiation codons or to the 5' untranslated regions are not always effective. (Wagner et al., 1993, Fenster et al., 1994). Nonspecific interactions between oligonucleotides and other molecules, such as proteins, can also lead to vaπable biological activity. (Woolf et al., 1992; Stein and Cheng, 1993). Furthermore, the oligonucleotides themselves may adopt unexpected tertiary and quaternary structures that bind DNA at unexpected locations. Such aberrant binding has the potential to produce undesired biological effects (Chaudhary et al , 1995).

Other difficulties have also been encountered in the search for efficacious antisense oligonucleotides. The affinity of oligonucleotides for their RNA targets increases with length and with increased G-C content Yet, longer oligonucleotides tend to bind RNA sequences nonspecifically and oligonucleotides. Moreover oligonucleotides with a high G-content tend to form G-quartets reducing the amount of the free-coil form of oligonucleotide thought to be required for antisense binding (Bishop et al, 1996). Thus, oligonucleotides are needed that are short and have a high affinity for their target sequences and that do not form G quartets despite having a high G content.

As previously noted, oligonucleotides are large hydrophilic compounds that must cross hydrophobic cellular membranes before they can bind their targets in the cytosol or nucleus (Uhlmann and Peyman, 1990; Milhgan et al., 1993). However, because of their large size, their hydrophilic nature and negative charge oligonucleotides do not efficiently cross cell membranes. In the absence of cellular uptake enhancers, oligonucleotides tend to accumulate in pennuclear endosomal compartments of treated cells. (Fisher et al., 1993, Guy-Caffey et al, 1995). In cases, transport of oligonucleotides across the plasma membrane or the membranes of the endosomal compartments limits their internalization rate and their activity. Therefore, new compositions and methods are needed to enhance the rate oligonucleotides cross lipid bilayers.

One class of lipid uptake enhancers includes a positively-charged head group that binds nucleic acids, and a membrane interactive tail that is thought to interact with membrane components. These compositions may facilitate oligonucleotide penetration of the cell presumably by transiently disrupting cell membranes. Unfortunately, the activity of many cationic lipid preparations, such as Lφofectin®, a 1:1 (mass) liposomal mix of the cationic lipid DOTMA and the fusogenic lipid dioleoyl phosphotidylethanolamine (DOPE) (Life Technologies, Inc., Gaithersburg. MD), are highly sensitive to factors such as the composition and quantity of nucleic acid, the target cell type, and the concentration of serum m the cell growth medium. In addition, some preparations are themselves cytotoxic These constraints severely limit the utility of many of these compounds as oligonucleotide delivery agents for therapeutic use in animal systems Improved delivery systems that are compatible with oligonucleotides must be identified. In summary, the progression of many diseases is associated with increased angiogenesis and vascular permeability caused by the over expression of VEGF. New compositions and methods for specifically reducing VEGF expression would be useful in the treatment of these diseases. Antisense oligonucleotide treatment is an attractive approach because of its potential selectivity

Unfortunately, many of the known VEGF antisense oligonucleotides only work at concentrations that are toxic to cells and exhibit only nonspecific effects. Furthermore, previous antisense oligonucleotides are chemically and biologically labile and those that are more stable tend to have unacceptably low affinities for their target sequences and they do not readily penetrate cell membranes and therefore have difficulty reaching their biological targets. Lastly, oligonucleotides with a high G content tend to form G quartets. New antisense oligonucleotide compositions are required that are nontoxic and have increased affinity for their mRNA target sequences These compositions should have improved biological stability including increased resistance to degradation by nucleases. In addition, useful oligonucleotides should not aggregate regardless of their sequence. New compositions are also required that facilitate the transport of oligonucleotides across cell membranes.

SUMMARY OF THE INVENTION The present invention provides compositions and methods for slowing the progression of diseases associated with increased angiogenesis and vascular permeability. The present antisense oligonucleotide compositions are markedly supenor to pnor oligonucleotides at selectively inhibiting the expression of VEGF by producer cells and they are intended for use in the treatment of such diseases. The selectivity of the present invention is provided by antisense oligonucleotides that specifically bind VEGF mRNA molecules and block expression of VEGF.

The present invention provides oligonucleotides and methods for making and using them, with chemical modifications to increase their affinity and specificity for target mRNA sequences. The present oligonucleotides have improved biological stability and high affinities for their target sequences. The oligonucleotides are relatively inert to chemical and biological challenges in both hydrophobic and hydrophilic environments and they resist aggregation regardless of their sequence. The invention provides VEGF antisense oligonucleotides that are both effective and nontoxic. Specifically, this invention is for new oligonucleotide compositions that, when used to treat cells at concentrations below 1 micromolar, cause a decrease in the cellular production of

VEGF. At these concentrations the present antisense oligonucleotides are nontoxic and do not interfere with cellular metabolism.

The invention also provides compositions and methods that allow oligonucleotides to readily penetrate cell membranes to reach their biological targets. This is accomplished by providing methods of making and using antisense oligonucleotides with cellular uptake enhancers. The cellular uptake enhancers are nontoxic, are compatible with VEGF antisense oligonucleotides and facilitate the efficient penetration of oligonucleotides through cell membranes.

For the purposes of this invention, the term "oligonucleotide" includes nucleic acid polymers and chemical structures resembling nucleic acid polymers. Equivalents of ribose or deoxyribose may be substituted into the structures so long as the base moieties attached to the structure can maintain the hydrogen bonds required for specific binding to their target sequences. Similarly, oligonucleotides may contain chemical equivalents of the phosphodiester backbone such as phosphothioester linkages. In addition, oligonucleotides may include base moieties that are chemically modified. Specifically, oligonucleotides may include but are not limited to C5-(propynyl or hexynyl) uridine or cytidine residues, 6-aza-uridine or cytidine residues and pyrimidines with both

C5 and 6 aza modifications.

The term "VEGF" is meant to include all proteins in the class known as vascular endothelial growth factors. The term VEGF includes at least the four known human isotypes that are thought to arise by alternative splicing of mRNA and any homologous protein that has a similar biological function. The known proteins include those that are encoded from mRNA species known in the art as

VEGF206,VEGF 185,VEGF 165,andVEGF 121.

Antisense oligonucleotides of the present invention are prepared as follows. A sequence of approximately 15-30 nucleotides and preferably about 19 nucleotides is identified on an mRNA that encodes VEGF. The sequences of VEGF mRNA molecules are known in the art. The RNA sequence can be anywhere on any mRNA that encodes any protein in the VEGF family of proteins.

More prefeσed are antisense oligonucleotides that are complementary to rnRNA's encoding human VEGF 206, VEGF 185, VEGF 165 and VEGF 121. Most preferred are oligonucleotides that bind sequences found on all ofthe VEGF mRNAs. (See Table 1).

TABLE 1

Anti-VEGF Oligonucleotides

I.D. Description Sequence Modification

T30615: antisense to mRNA 185- 203+ 5' -g*c*g*c*t*g*a*t*a«g*a*c*a*t*c*c*a*t*g -3' total PT (phosphorothioate) DNA

T30639: var. of T3061S 5'-g*C*g*C*U*g*a*U*a^«g*a*C*a*U*C*C*a*U*g-3 ' total PT, C5-propynyl pyrimidines DNA

T30640: mRNA seq. 204-222 5' -C*g*a*U*U^»g*g*a*U*g*g*C*a*g*U*a*g*C*t-3 ' total PT, CS-propynyl pyrimidines

T30641: mRNA seq. 232-2S0 5' -U*a*C*U*C*C*U*g*g*a*a*g*a*U*g*U*C*C*a-3 ' total PT, C5-propynyl pyrimidines

T30847: var. of T30639 5' -g*C*g*C*U*g-a*U*a-g-a^«C*a-U*C*C*a*U*g-3 ' 4 PD linkages

T30848: var. of T30639 S^t-g*C*g-C*U-g-a*U*a-g-a*C*a-U*C*C*a*U*g-3' 6 PD linkages

T30849: var. of T30639 5' -g*C*g*C*U*g-a*U*a*g-a*C*a*U*C*C^«a*U*g-3' 2 PD linkages

T30876: mRNA seq. 224-242 5 ' -g*a*a*g*a*U*g*U*C*C*a*C*C*a*g^«g*g*U*C-3 ' total PT, C5-propynyl pyrimidines

T30877: mRNA seq. 406-424 5' -a*g*g*a*a*g*C*U*C*a*u*c*U*C*U*C*C*U*a-3' total PT, C5-propynyl pyrimidines

T30878: mRNA seq. 522-540 5'-U*a*C*a*C*g*U*C*U*g*C*g*g*a*U*C*U*U*g*-3' total PT, C5-propynyl pyrimidines

T30879: mRNA seq. 575-593 5^,-U*a*a*C*U*C-a-a*g*C-U*g*C*C*U*C*g*C-C--3' total PT, C5-propynyl pyrimidinea

T30886: mRNA seq. 171-189 S• -C*C*a*U*g*a*a*C*U*U*C*a*C*C*a*C*U*U*c-3' total PT, C5-propynyl pyrimidines

T30887: mRNA seq. 176-194 5' -g*a*C^«a*U*C*C*a*U*g*a*a*c*t*t*c*a*c*c-3' total PT, C5-propynyl pyrimidines

T30888: mRNA seq. 199-217 5' -g*g*a*U*g*g*C*a*g*U*a*g^»C*U*g*C*g*C*U-3' total PT, C5-propynyl pyrimidines

T30889: mRNA seq. 195-213 5 > -g*g*C*a*g*U*a*g*C*U*g*C*g*C*U*g*a*U*a-3 ' total PT, CS-propynyl pyrimidines

T30890: var. of T30639 5' -g*C*g*C*t*g*a*t*a*g*a*C*a*t*C*C*a*t*g-3 ' total PT, C5-propynyl C only

T30891: var. of T30639 5' -g*c*g*c*U*g*a*U*a*g*a»c*a*U*c*c*a*U*g-3 * total PT, C5-propynyl U only

T30892: var. of T30639 5" -g*c*g*c*t*g*a*U*a*g*a*C*a*U*C*c*a*t*g-3' total PT, 4 CS-propynyl pyrimidines

T30893: var. of T30639 5" -g*C*g*c*U*g*a*U*a^«g*a*C^«a*t*C*c*a*U*g-3 ' total PT, 6 CS-propynyl pyrimidines

S96-5296: var. of T30639 5' -g*C*g*C*U-g-a-U^«a-g-a-C^,a-U*C-C*a*U«g-3' [AL] -Lip-1 8 PD; C5-propynyls, lipid tether

S96-5297: var. of T30639 S' -g*C*g*C*U-g-a-U"a-g-a-C*a-U*C-C*a*U*g-3' (AL) -pyrene 8 PD; C5-proρynyls, pyrene tether

T30688: var. of T30615 5'-g*C*g*C*U*a*U*g*a*C*a*U*C*C*a*U*g-3' total PT, C5-hexynyl pyrimidines DNA

T30692: 2-base mismatch 5' -g*C*g*C*U^»a*C*a*g*a*C*a*U*U*C*a*U*g-3' total PT, C5-propynyl pyrimidinea, DNA version of T30639

T30807: 'sense' DNA of T30615 5' -c*a*t*g*g*a*t*g*t*c*t*a*t*c*a*g*c*g*c-3' total phosphodiester, DNA ^"■

T30807: 'sense' RNA of T3061S 5' -c*a*t*g*g*a*t*g*c*c*t*a*t*c*a*g*c*g*c-3 ' total phosphodiester, RNA

+ human VEGF mRNA sequence from Leung et al.. Science, 246:1306, 1989. Initiator codon at base 57.

* phosporothloate linkage. O

4

- phosphodiester linkage.

1 C,U represent modified bases

A senes of complementary or "antisense" oligonucleotides are prepared For the purposes of this invention, "antisense" means that the oligonucleotides have sequences complementary to mRNA sequences such that they will bind those sequences through specific hydrogen bonding patterns However, an antisense oligonucleotide can have mismatches or imperfect hydrogen bonding patterns as long as the oligonucleotide has anti-VEGF activity at concentrations below 1 micromolar.

Antisense oligonucleotides contemplated in this invention mclude modifications that improve their biological stability. Biological stability is improved by incorporating nuclease resistant linkages, such as phosphorothioate linkages, between vanous or all nucleotide residues. The present oligonucleotides also include chemically modified bases at vanous or all pyπmidine locations. These modified bases include CS-propynyl pynmidines, C5-hexynyl pynmidmes or 6-aza- pynmidmes or combmed C5 and 6-aza pynmidme denvatives and may further stabilize the oligonucleotides of the present invention.

Antisense oligonucleotides contemplated in this invention mclude modifications that improve their binding affinity for their target sequences Binding affinity is improved by incorporating vanous chemical moieties into pynmidme bases The present oligonucleotides include chemically modified bases at vanous or all pyπmidine locations These modified bases mclude CS- propynyl pynmidines, C5-hexynyl pynmidmes or combined C5 and 6-aza pynmidme denvatives.

Antisense oligonucleotide binding can be to actual mRNA or to chemically synthesized RNA sequences which are identical to sequences found on VEGF mRNAs This binding can be demonstrated in a vanety of ways. One method for observing binding is descπbed in Example III. This method involves mixing antisense oligonucleotides with chemically synthesized RNA sequences of the same length, allowing the antisense oligonucleotide to anneal m an initial heating and cooling step, and observing the absorbance change of the mixture at 260 nm on heating Binding can also be measured by other methods such as, nuclease protection expenments. oligonucleotide extension expenments, NMR, gel electrophoresis or other techniques well known to those of skill in the art.

For the purposes of this invention "improved binding affinity" or "stability" means that the oligonucleotide has a higher melting temperature (T„-) when assayed with its target RNA sequence than an oligonucleotide without the modification. Melting point assays, as descnbed in Example HI, are used for this determination Chemical modifications that increase the binding affinity of antisense ohgonucleotide/mRNA target sequence duplexes are contemplated for use by the present invention. In general, antisense oligonucleotides having a T_m above 45^CC in the descπbed assay are contemplated More preferred are oligonucleotides having a T_m above 50°C

Certain oligonucleotides of the present invention include chemical modifications that improve their activity over previously known VEGF antisense oligonucleotides. Improved activity means that lower concentrations of oligonucleotide are required to inhibit VEGF expression in vivo Although the invention is not intended to be limited by the mode of action of these modifications, increased binding affinity and biological stability are thought to be at least partially responsible for the increased activity of the presently contemplated oligonucleotides Specific chemical modifications, as set forth above, are used to increase the activity ofthe present oligonucleotides. Antisense oligonucleotides contemplated by the present invention are also nontoxic at concentrations below approximately 1 μM. Toxicity is measured according to the method set forth in Example V.

Antisense oligonucleotides of the present invention reduce VEGF production in treated cells. In one method cells are treated by placing them in direct contact with the oligonucleotide compositions so that the oligonucleotide can be internalized in the cell and reach its target mRNA sequence Pnor to treatment, the oligonucleotide is dissolved or suspended m a liquid or incoφorated mto a solid. Suitable liquid and solid formulations are known in the art and can be chosen by well known methods Formulated oligonucleotides are placed in direct contact with cells. In other methods, the formulated oligonucleotides are positioned such that oligonucleotides can reach their target cells through diffusion, dispersion or like means. The present mvention does not require the oligonucleotide formulation to directly contact target cells. The invention only requires that the oligonucleotide reach target cells. For example, an oligonucleotide could be introduced into the blood stream but diffuse out ofthe blood before reaching target tumor cells, arthntic cells, or the like. Alternatively, the oligonucleotide could be mixed into a powder which is applied directly to the skin and diffuse to underlying cells.

Cells treated with the present antisense oligonucleotides produce, at most, approximately 90% of the VEGF that is produced by untreated cells under the same conditions. This affect is observed when oligonucleotide solution concentrations are below approximately 1 micromolar (μM) One method for measuπng reduced cellular VEGF production is descπbed in Example VI However, other methods can be used to detect the reduction m VEGF production, if they are as sensitive as the method descπbed m Example V. The percent of VEGF produced by treated cells is determined by measuπng the amount of VEGF produced by untreated cells and treated cells The percentage equals the amount produced by treated cells divided by the amount produced by untreated cells multiplied by 100. The untreated and treated cells are intended to be approximately identical in all respects except with regard to the presence or absence of oligonucleotide formulations. Thus, the cells used m the assay are of the same type, passage number, phenotype and are m the same stage of growth The cells are grown under the identical conditions including identical media (except for changes due to the presence or absence of the oligonucleotide formulation itself), temperature and atmosphere. Under these conditions, cells treated with the antisense oligonucleotides contemplated by this invention produce, at most, approximately 90% of the VEGF as produced by identical untreated cells when antisense oligonucleotides are used at concentrations of up to 1 μM in solutions or a similar mole percent if used in solid formulations.

Preferred oligonucleotides incorporate certain chemical modifications that increase their resistance to nucleolytic degradation. Chemical modifications contemplated m these embodiments are modifications of the common naturally occumng chemistnes found m oligonucleotides. Certain chemical moieties contemplated in this invention include phosphorothioate linkages. These may be positioned between some or all of the nucleoside residues. The most preferred oligonucleotide contains 10 phosphorothioate and 8 phosphodiester bonds. In addition to nucleotide linkages, chemical modifications to the base moieties may increase resistance to nuclease degradation. More specifically, modifications to pyndine πngs including C5-propynyl or hexynyl groups and/or 6-aza- pyπdme modifications are contemplated

One method for measuπng nuclease resistance is by determining the half-life of oligonucleotides m blood serum. This is accomplished by standard methods well known m the art. For the purposes of the present invention, a chemical moiety decreases the rate of degradation of antisense oligonucleotides by nucleases if the oligonucleotide has a longer serum half-life with the moiety than it would have without the moiety Phosphorothioate containing oligonucleotides have half-lives of well over 24 hours while their counterparts which contain only phosphodiester bonds have serum half lives of under 3 hours.

Oligonucleotides are contemplated that contain chemical modifications in their pyπmidine πngs. Preferred oligonucleotides contain either C5-propynyl pynmidines, C5-hexynyl pynmidmes and/or 6-aza pynmidmes. These modifications increase their T_ms, biological stability, and their activity. The synthesis of nucleotide precursors containing these modifications is descπbed in Example 1 The synthesis of oligonucleotides from these and other protected nucleotides is by standard phosphoramidite chemistry well known in the art. Certain embodiments of the present invention are directed to cellular uptake enhancement compositions that improve the activity of oligonucleotides. Generally, these compositions enhance the transport of a liquid across the liquid bilayer. In some embodiments, the oligomer is covalently conjugated to a hpophihc molecule. This improves the oligonucleotide membrane association and permeability properties, such as cholesterol, fatty acids or other hpophihc tether. These molecules can be chemically linked to oligonucleotides by standard methods well known m the art.

In other embodiments, uptake enhancers such as cationic lipids or liposomal preparations may be used. These agents are attractive because of their versatility. These embodiments have the advantage that the same delivery vehicle may be used to administer a mixture of oligonucleotides. One embodiment specifically contemplates the use of the liposomal preparation Cellfectm®. Other embodiments include a class of polyammolipid uptake enhancers, including spermidine-cholesterol (SpdC). This latter compound has the advantage of functioning particularly well even m the presence of serum Compositions and methods of prepanng antisense oligonucleotides with cellular uptake enhancers are descπbed in Examples IV, VI, and VII.

Certain oligonucleotides contemplated in the present invention are in the salt form. A salt form is a form m which the oligonucleotide is associated with a positively charged (cationic) atoms or molecules. Suitable cations include but are not limited to sodium, potassium, ammonia, spermidine or polyamino lipids such as spermidine-cholesterol and the like.

Certain embodiments contemplated in the present invention compnse a liposome. Suitable liposomes are well known in the art. Certain liposome compositions specifically contemplated by the present invention include Cellfectm®. Other compositions include spermidine-cholesterol mixed with DOPE. Liposomal preparations are prepared by methods well known in the art.

Certain embodiments of the present invention contemplate delivery of its oligonucleotide compositions through sustained delivery systems, including but not limited to polymenc release devices, for example polycaprolactone or blends of polycaprolactone with methoxypolyethylene glycol Methods and compositions for incoφorating the present antisense oligonucleotides into sustained delivery systems are well known m the art.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1. Representative modified bases of the invention that are used to replace the natural bases m the synthesis of antisense oligonucleotides. Figure 2. Synthetic scheme for the preparation of 5-(l -hexynyl or propynyl)-6-aza-2'- deoxyuπdme phosphoramidite (see also Figure 1).

Figure 3. Effect of antisense oligonucleotides on VEGF production by normal human keratinocyte m culture

Figure 4. Effect of oligonucleotide (Sequence ID No 2) administered with or without Cellfectm® on VEGF expression by keratinocytes

Figure 5. Effect of oligonucleotide (Sequence ID No. 27) administered with or without

Cellfectm® on VEGF expression by keratinocytes.

Figure 6. Effect of different cellular uptake enhancers on the activity of T30639

(Sequence ID No. 2) with keratinocytes. Figure 7. Short term cellular exposure to oligonucleotide formulations of the invention and long term inhibition of VEGF expression.

Figure 8. Short term cellular exposure to oligonucleotide formulations of the invention and long term inhibition of VEGF expression.

Figure 9. Short term cellular exposure to oligonucleotide formulations of the invention and long term inhibition of VEGF expression. Figure 10. Effect of end-modified, chimeric VEGF antisense oligonucleotides on VEGF expression, in the presence or absence of uptake enhancer.

DESCRIPΗON OF PREFERRED EMBODIMENTS The preferred embodiment includes an antisense oligonucleotide that binds to a sequence common to multiple VEGF encoding mRNA molecules and prevents the expression of VEGF in vivo. Preferred oligonucleotides contain phosphorothioate linkages in place of several of the phosphodiester linkages and other chemical modifications that increase the affinity of the oligonucleotide for its target mRNA sequence. In preferred compositions oligonucleotides are formulated with cell uptake enhancers that improve their ability to cross the cell membrane.

Oligonucleotides of the present invention may range in length from approximately 17 residues to 30 residues in length. Preferred oligonucleotides are 19 nucleotides long. Their sequences are selected based on their complementarity to the mRNA molecules that encode the VEGF genes. The region of the mRNA molecule that is complimentary to the oligonucleotide is called the target sequence. Preferred antisense oligonucleotides are complementary to target sequences that are found in each of four known VEGF mRNA molecules including VEGF 206, VEGF 185, VEGF 165, and VEGF 121.

Oligonucleotides are contemplated that contain chemical modifications that improve their binding affinity for target mRNA. Preferred oligonucleotides contain either C5-propynyl pyrimidines, C5 -hexynyl pyrimidines and/or 6-aza pyrimidines. Preferred modifications increase the temperature at which the oligonucleotide dissociates from its target sequence. The synthesis of nucleotide precursors containing these modifications is described in Example I. The synthesis of oligonucleotides from protected nucleotides is by standard phosphoramidite chemistry and is well known in the art. Preferred oligonucleotides incoφorate certain chemical modifications that increase their resistance to nucleolytic degradation. Although the invention is not limited by the mechanism for this resistance, the chemically modified nucleotides are thought to resist nuclease digestion by interfering with oligonucleotide binding in the substrate binding pocket of nucleases. The preferred nuclease resistant oligonucleotides contain phosphorothioate linkages between at least some of the nucleotide residues. The most preferred oligonucleotide contains 10 phosphorothioate and 8 phosphodiester linkages.

In preferred compositions, the oligonucleotides of the present invention are formulated or mixed with cell uptake enhancers that increase their ability to penetrate cell membranes. Cell uptake enhancers contemplated for use in this invention include dioleoyl phosphotidylethanolamine, Cellfectin®, spermidine-cholesterol and the like. Most preferred is a 1:1 mixture by mass of spermidine-cholesterol and dioleoyl phosphotidylethanolamine. This formulation is mixed with 10 nanomolar to 1 micromolar concentrations of oligonucleotide according to standard methods well known in the art.

Oligonucleotide compositions contemplated by the present invention are selected based on their in vivo activity. Preferred compositions are not substantially cytotoxic to cells with oligonucleotide concentrations up to 1 micromolar. Standard cytotoxicity assays as descπbed in

Example I are used m making this determination. The present compositions must also demonstrate an ability to reduce cellular VEGF production at concentrations below 1 micromolar.

The present invention has been descπbed m terms of particular embodiments found or proposed to compπse preferred modes for the practice of the invention. It will be appreciated by those of skill in the art that, m light of the present disclosure, numerous modifications and changes can be made in the particular embodiments exemplified without departing from the intended scope of the invention All such modifications are intended to be included withm the scope of the appended claims

EXAMPLES Example I Method for the preparation of modified bases for incorporation into synthetic oligonucleotides:

The modified bases that increase the binding affinity and/or specificity of the synthetic oligonucleotides are shown in Fig. 1 The synthetic scheme for the prepanng 5-(l -hexynyl or propynyl)-6-aza-2'-deoxyundιne phosphoramidite is shown m Fig. 2 This synthesis provided the building block for prepanng the antisense oligonucleotides containing 6-aza-U Similar schemes have been used to synthesize 6-aza-C. The detailed synthetic methodology for the preparation of 5- (l-hexynyl)-6-aza-2'-deoxyuπdιne phosphoramidite is descnbed below. In a similar manner, 5- propynyl denvative was prepared starting from 5-ιodo denvative 7

3'.5'-Dι-0-ρ-toluoyl-5-ιodo-6-aza-2'-deoxyuπdme r7aV

Chlorotnmethylsilane (0.5 ml) was added to a suspension of 5-ιodo-6-azauracιl (5, 8g, 33.47 mmol) m 1,1, 1,3,3 ,3-hexamethyldιsιlazane (HMDS, 80 ml) and the mixture was heated under reflux for 6 h The reaction mixture was cooled to room temperature and HMDS evaporated in vaccuo. The residue was dπed under high vacuum for 4 h The dπed silyl denvative was dissolved m dichloromethane (60 ml). l-chloro-2-deoxy-3,5-dι-0-/Moluoyl-b-D-er>-/rtro-pentofuranose (6, 16.3 g, 42 mmol) and zmc chloπde (0.46 g, 3.35 mmol) were added to this solution and the mixture was stirred under an argon atmosphere for 24 h. The reaction mixture was diluted with dichloromethane (250 ml) and the dichloromethane solution was washed with saturated aqueous NaHC03 solution (100 ml). Aqueous layer was extracted with dichloromethane (4 x 100 ml) and the combined organic layer was dπed (Na2SO_ι) and evaporated. The residue was puπfied by silica gel column ( 4 x 15 cm) chromatography and product elutes in dichloromethane containing 0-5% methanol. The dπed anomenc product weights 15g. Pure b anomer was obtained by tπturatmg with a mixture of dichloromethane and methanol (4: 1 , 200 ml). The solid was collected by filtration and evaporation. 10.5 g of pure b anomer was obtained after repeating this tnturation process, mp 204-205 °C. *H NMR (DMSO--Y6): d 2.35, 2.37 (2s, 6 H, 2 CH3), 2.80 ( m, 2 H, C2*H and C2^»H), 4.43 (s, 3 Η, C-VH, C5-H2), 5.55 (br s, 1 Η, CyH), 6.39 (t, J = 6.0 Ηz, 1 Η, C\'H), 7.28 (t, 4 Η, Tot), 7.85 (t, 4 Η, Tol), 12.42 (br s, 1 Η, NH). Anal. Calcd. for C24Η24IN3O7: C, 48.58; H, 4.08; N, 7.08. Found: C, 48.85; H, 3.80; N, 6.92.

3^,.5^,-Dι-0-/--toluovl-5-fl-hexvnvlV6-a7a-2'-deoxvuπdιne r8a1:

3',5'-Dιtoluoyl-5-ιodo-6-aza-2'-deoxyuπdιne (7, 3.84 g, 6.5 mmol) was dπed by coevaporation with dry DMF (25 ml) and dissolved m DMF (30 ml) to which Cul (0.25 g, 1.3 mmol), tnethylamme 1.82 ml, 13 mmol), 1-hexyne (2.23 ml, 19.5 mmol) and tetrakιs(tπphenylphosphme)palladιum (0.75 g, 0.65 mmol) were added under an argon atmosphere. The reaction mixture was stirred at room temperature for 18 h and an additional 0.5 g of tetrakιs(tπphenyl-phosphme) palladium was added. After 48 h, the solvent was evaporated and the residue coevaporated with toluene. The residue was puπfied by silica gel column chromatography and the product elutes in dichloromethane contaimng 0-5% ethyl acetate to yield 0.9 g of the title compound, mp 198-200 °C. *H NMR (DMSO-</6): d 0.87 (t, 3 H, CH3), 1 45 (m, 4 Η, 2, CH2), 2.37, 2.39 (2s, 6 Η, 2 CH3), 2.45 ( m, 3 Η, CH2 and C2"H), 2,84 (m, 1Η, CγH), 4.45 (s, 3 Η, C-vH, C5H2), 5.59 (br s, 1 Η, CyH), 6.49 (t, J = 6.3 Ηz, 1 Η, CyH), 7.31 (dd, 4 Η, Tol), 7.88 (dd, 4 Η, Tol), 12.40 (br s, 1 Η, NH). Anal. Calcd. for C3θΗ33N3θ7' C, 65.80; H, 6.07, N, 7.68. Found: C, 65.61; H, 5.73; N, 7.29.

6-Aza-5-d -hexynyl1-2'-deoxyundme (9a, 1:

A mixture of 3',5'-dι-0-p-toluoyl-5-(l-hexynyl)-2'-deoxyundιne (8a, 0.8 g, 1.47 mmol) methanol (55 ml) and sodium methoxide (25% solution m methanol, 1.28 ml) was stirred at room temperature for 2 h. The reaction was quenched by the addition of Dowex 50X8 (H⁺) resm. The resm was removed by filtration and the filtrate was evaporated. The residue was puπfied by silica gel column chromatography using dichloromethane containmg 0-4% methanol as the eluent to yield

0.38 g (84%) of the title compound as a very hygroscopic sohd^H NMR (DMSO-ctø). d 0.88 (t, 3 H, CH3), 1.47 (m, 4 Η, 2, CH2), 2.02 ( m, 1 Η, C2^«H), 2.33 (m, 1Η, C2'H), 2.46 (m, 2 Η, CH2), 3.40 (m, 2 H, C5-H2), 3.68 (m, 1 Η, C-vH, ), 4.21 (br s, 1 Η, CyH), 4.61 (t, 1 Η, CyOH), 5.17 (d, 1 Η, CyOH), 649 (dd, 1 Η, CyH), 12.27 (br s, 1 Η, NH).

V.O.f4.4'-DιmethoxvtnMV6-aza-5-π-hexvnvn-2 deoxvuπdιne πθal:

4,4'-dιmethyoxytπtyI chlonde (0.51 g, 1.5 mmol) was added to a solution of 5-(l-hexynyl)-6- aza-2'-deoxyuπdιne (0.38 g, 1.24 mmol) in dry pyndine (10 ml). After stirπng for 6 h, an additional 0.5 g of DMT-C1 was added and the reaction mixture was stirred overnight. The reaction mixture was diluted with dichloromethane (100 ml) and the organic solution washed with water (20 ml). The aqueous layer was extracted with dichloromethane and the combined organic layer was dried (Na2Sθ4) and evaporated. The residue was coevaporated with toluene (10 ml) and puπfied by chromatography over a silica gel column ( 2 x 10 cm). The product was eluted with dichloromethane containing 0-1.5% methanol. Yield, 0.45 g ^!Η NMR (DMSO-cfe)^* d 0 85 (t, 3 H, CH3), 1.37 (m, 4 Η, 2, CH2), 2.08 ( m, 1 Η, Cγ'H), 2.37 (m, 3Η, CH2 and CγH), 3 06 (m, 2 Η, Cs'H2), 3.72 (s, 6 Η, 2 OMe), 3 87 (m, 1 Η, CyH, ), 4.20 (m, 1 Η, CyH), 5.23 (d, 1 Η, CyOH), 6.35 (dd, 1 Η, CyH), 6.83 (m, 4 Η, DMT), 7.16-7.25 (m, 9 Η, DMT), 12.30 (br s, 1 Η, NH).

5'.O.f4.4'-Dιmethoxvtπtvn-6-aza-5-n-hexvnvn-2^,-deoxvundιne-3^,-0-r2-cvanoethvlV^.N- dnsopropvlphosphoramidite Cl lal:

5^,-0-(4,4'-Dιmethoxytntyl)-6-aza-5-(l-hexynyl)-2'-deoxyuπdιne (10a) on reaction with 2- cyanoethyl-N,N-dnsopropylchlorophosphoramιdιte in dichloromethane in the presence of N,N- dnsopropylethylamine provided the phosphoramidite lla by well known methods

Example II Oligonucleotide design and synthesis:

Antisense oligonucleotide sequences were selected that can bind complementary mRNA target sequences shared by all splice vanants of VEGF mRNAs The sequence of exemplary synthetic oligonucleotides are shown m Table 1. To improve their binding affinity for mRNA targets, oligonucleotides were synthesized with pynmidines having C5-propynyl or C5-hexynyl groups as shown in Figure 1. (Wagner et al., 1993) Other modified bases, including 6-aza-dU and 6-aza-dC were also contemplated. (Figures 2). Combinations of these modifications were also contemplated. Oligonucleotide T30691 (Sequence ID No. 27) which was complementary to the antisense oligonucleotide T30639 (Sequence ID No. 2) was used as a control in the following expenments. It was the same size and contains the same backbone and base modifications as T30639 (Sequence ID No. 2). Example m T_m analysis of antisense oligonucleotide-RNA heteroduplex interaction:

The temperature (T^ of antisense oligonucleotide RNA duplexes was used to estimate binding affinity. The T_m was measured in a diode array spectrophotometer equipped with a temperature controlled cell holder (Hewlett Packard Model 8452). Antisense oligonucleotide was mixed with a synthetic RNA target of the same size (each at 1 μM), in a buffer consisting of 2 mM sodium phosphate, pH 7.0, 18 mM NaCl, and 1 mM EDTA. The solution, prepared in a spectrophotometer cell, was heated to 90°C for 10 mm, cooled to 20°C over 10 min, and equilibrated for 10 min to allow duplex formation. To measure the melting temperature (T.J of the duplex, the cell was slowly heated from 20°C to 80°C at a rate of l°C/mιn, and the absorbance at 260 nm was measured as a function of temperature. A nse m absorbance signals the melting or separation of the duplex into single stranded oligomers. The T_m of duplex formation was obtained from the melting curve data using equations descπbed by standard methods (Pughsi and Tinoco, 1989). The T_m data are shown in Table 2. Incoφoration of C5-propyne modified bases or C5-hexynyl-modιfied bases mto phosphorothioate oligonucleotides leds to marked increases in T_m values over unmodified oligonucleotides. This was indicative of a significant improvement in the affinity of the antisense oligonucleotide for its target sequence.

TABLE 2

T30807 (Antisense DNA)

T30615 T30639 T30688 T306s2

(unmodified) (PrOPynyl) (hexynyl) (propynyl m~smatch)

T_m (°C) 43 53 49 345

AG at 37°C (kcal/mol) 1 1.5 -13.4 -1 1.9 -8.1

Ah (kcallmol) 136 -91.6 -79.8 -103

AS (eu) 402 -252 -219 -306

T30808 (antisense RNA)

T30615 T30639 T30688 T30692

T_m (°C) 42 57 53.5 43.5

AG at 37°C (kcal/mol) -1 1.0 -14.3 -13.5 -1 1.3

AH (kcal/mol) -128 -88.3 -90.7 -1 19

AS (eu) -378 -239 -249 -347 z

Example TV Preparation of uptake enhancers:

The unassisted uptake of antisense oligonucleotides by cells was low (Fisher et al., 1993; Guy-Caffey et al., 1995). To improve penetration into cells, a number of commercially available uptake enhancers as well as formulations of novel polyammohpids synthesized by the inventors are used (Gao et al., 1989; Guy-Caffey et al., 1995). Example V Cytotoxicity assays:

Cells were seeded at a density of 500 cell/well in a 96 well plate. One day after plating, the cells were exposed to senally diluted oligonucleotide formulations (4 wells per dilution). After one day or four days of exposure, the effect on cellular viability was determined with a nonradioactive assay system (Cell Titer 96 Aqueous cell proliferation assay, Promega Coφ.). No toxicity was observed when the present oligonucleotides were at concentrations below 1 μM. Example VI Cellular testing of oligonucleotides:

The activity of antisense oligonucleotides, their modified counteφarts, and vanous formulations were evaluated using cultured human keratinocytes, a pnmary cell line that secretes VEGF under normal culture conditions (Ballaun et al., 1995, Frank et al., 1995) Cells were plated in 48-well plates at a density of 50-100,000 cells/well/0.5 ml KGM medium (Clonetics). A sensitive ELISA-based protein assay system (R&D Systems) was used to measure VEGF protein levels m the cell supernatant Preliminary measurements showed that when NHEK cells were grown in the recommended medium, 50,000 cells plated m 0.5 ml medium produce about -150-200 pg of VEGF in 15 hours (i.e., -300-400 pg/ml m the supernatant of untreated control wells). Cells were also incubated for 15 hours with the oligonucleotide formulation Three of four anti-VEGF oligonucleotides demonstrate activity m the 0.2 μM range, in the presence of 10 μg/ml Cellfectm® The control sense oligonucleotide had no effect (not shown) Results are shown m Figure 3

For the evaluation of the antisense effect, the oligonucleotides were administered to cells in the presence or absence of uptake enhancers In preliminary expenments phosphorothioate oligonucleotides, without base modifications were in effective at concentrations below 1 μM and there was no significant effect observed in the absence of earners (data not shown). At concentrations above 1 μM oligonucleotides tended to inhibit VEGF expression nonspecifically (data not shown). These nonspecific effects were known in the art. (Stein et al., 1993, Wagner, 1994). To avoid these nonspecific effects, oligonucleotides were mixed with uptake enhancers. Cellfectin®, a liposomal preparation of a tetrapalmitylspermine (T_M-TPS) with dioleoyl phosphotidyethanolamine (DOPE) (T_M-TPS/DOPE in 1:1.5 mass ratio, from Life Technologies, Inc.) was more effective and less toxic than other commercially available delivery agents tested.. Oligonucleotides formulated with liposomal preparations of the polyammolipid SpdC (Guy-Caffey et al., 1995; SpdC/DOPE, 1 1 by mass) were even more effective. In typical cell culture expenments, oligonucleotides (10 nM to 1 μM) were dissolved m water -20-40 μl of an aqueous solution of uptake enhancer at room temperature, and incubated for ~10 mm. That solution was mixed with 0.5 ml of warm growth medium and added to cells. Cells were incubated for 15 hours in the presence of the oligonucleotide. After the incubation, the supernatant was collected and either used immediately for ELISA or saved at -80°C for future analysis (no significant difference m VEGF levels was observed between never frozen or frozen and thawed supernatant samples). As shown m Figure 4, the antisense oligonucleotide T30639 (Sequence ID No. 2) was more active in the presence of Cellfectm®, whereas the control 'sense' oligonucleotide T30691 (Sequence ID No. 27) had little effect except at the highest concentration used, as shown m Figure 5.

Figure 6 shows the effect of administeπng 0.1 μM or 0.2 μM oligonucleotide (Sequence ID No. 2) with vanous cationic lipid formulations SpdC, spermidine-cholesterol (Guy-Caffey et al., 1995); DC-Chol (Gao and Huang, 1991); CS, cholate-spermidme; DCS, deoxycholate-spermidme; cF, Cellfectin® (Life Technologies, Inc.). Liposomal preparations of each cationic lipid were prepared by mixing with the fusogenic lipid DOPE (1.1 mass ratio) and were stored after lyophilization until use. The liposomes were resuspended in 5% dextrose (to 1 mg/ml) pnor to use, and stored at 4°C for use withm two weeks. Oligonucleotides were mixed with the cationic liposomal preparations just before cellular treatment, as descnbed above.

Figures 7-9 show the results from cell incubations with varying concentrations of the antisense oligonucleotides T30639 (Sequence ID No. 2), or its chimeπc phosphodiester- phosphorothioate version T30848 (Sequence ID No. 6). (See Table 1). Figure 7 shows the effect of 0.1 μM oligonucleotide, Figure 8 shows the effect of for 0.2 μM oligonucleotide and Figure 9 was for 0.4 μM oligonucleotide. In each expenment cells were treated for 4 hours m medium supplemented with the antisense oligonucleotide premixed with SpdC/DOPE. Then the medium was replaced with fresh unsupplemented medium. Graph 1 shows the percent inhibition in VEGF production 16 hours after the oligonucleotide composition was washed out of the culture, Graph 2 is 40 hours after oligonucleotide wash out, and Graph 3 is 64 hours after oligonucleotide wash out. The amount of VEGF level in the harvested medium was then determined. The moφhology of cells at the end ofthe ~3 day incubation peπod was normal. The long term effects of the oligonucleotide on VEGF production are set out m Figures 7-9. In the graphs the symbol (Δ) is for 0.1 μM T30848 (Sequence ID No. 6). The symbol ( ) is for 0.1 μM T30639 (Sequence ID No. 2). Figure 10 shows the results m similar expenments with oligonucleotides denvatized with hpophilic groups. S96-5296 (Sequence ID No. 20) is modified at the 3'-end with a C-16 lipid group and contains 8 phosphodiester and 11 phosphorothioate linkages. S96-5297 (Sequence ID No. 21) has the same backbone and is end-modified with a 3 '-pyrene moiety. These hydrophobic moieties aid in the uptake and activity of the oligonucleotides when mixed with Cellfectin® and the phosphodiester linkages increase the activity of the oligonucleotides. The symbol ( ) is for S96-5296 (Sequence ID No. 20), the symbol ( ) is for S96-5296 (Sequence ID No. 20) with 10 ug/ml Cellfectin®, the symbol (o) is for S96-5297 (Sequence ID No. 21), the symbol (•) is for S96-5297

(Sequence ID No. 21) with 10 ug/ml Cellfectin®, the symbol (D) is for 0.2 μM T30639 (Sequence ID No. 2) with 10 ug/ml Cellfectin®.

Example VII Anti-VEGF activity of antisense oligonucleotides:

Phosphorothioate containing antisense oligonucleotides without base modifications appeared to have no significant effect on the cellular production of VEGF, except for some sequence- independent nonspecific inhibition at concentrations exceeding 1 μM (data not shown) The results were consistent with other studies showing that low, submicromolar doses of simple phosphorothioate oligonucleotides were ineffective inhibitors, and at high levels, the same oligonucleotides may exert nonspecific effects on cellular metabolism (reviewed in Stein and Cheng, 1993; Wagner, 1994). However, phosphorothioate containing oligonucleotides containing C5- propyne-contaming pynmidmes (Wagner et al., 1993) specifically inhibit the cellular production of VEGF. See Figure 3. These modified oligonucleotides have melting temperatures that were about 15°C higher than their unmodified counteφarts. See Table 2. This suggests that modified oligonucleotides bind their targets with greater affinity than unmodified forms

Optimal oligonucleotide to Cellfectin® ratio: Cellular uptake of the ohgonucleotide- cationic lipid mix was determined partly by the chemical nature of each component in the formulation, partly by their concentration and relative mass ratios, and partly by the endocytic properties of the target cell. With oligonucleotide T30639 (Sequence ID No. 2) coadministered with Cellfectin® to NHEK cells, the ratio of oligonucleotide to TMTPS of 1 3 (by mass), resulted m optimal activity In a related experiment, the concentration of the oligonucleotide was altered while maintaining the ratio of the oligonucleotide and cationic lipid, and the effect on VEGF expression was measured relative to the 'sense' control T30691 (Sequence ID No. 27) Oligonucleotide T30639 (Sequence ID No. 2) showed specific anti-VEGF activity, while the control oligonucleotide had no effect. (See Figures 4 and 5).

Effect of formulations of spermidine-cholesterol+DOPE or DC-chol+DOPE (liposomal preparations; 1:1 by wt) on oligonucleotide efficacy: A number of uptake enhancers are used with nucleic acid therapeutics (Behr, 1994; Guy-Caffey et al., 1995, Lewis et al., 1996). One of these compounds was spermidine-cholesterol conjugate (SpdC) (Guy-Caffey et al., 1995). This compound was non toxic to cells at concentrations far greater than required by this invention and was not toxic to rodents when treated for up to 1 week. Another cationic lipid, DC-Chol (Gao and Huang, 1991), was approved for therapeutic use in gene therapy and was relatively non toxic in cellular systems. Testing liposomal preparations of SpdC/DOPE and DC-Chol/DOPE (SpdC or DC-Chol with dioleoyl phosphotidylcholme) at mass ratios ranging from 1.0 5, 1.1, 1.1 5, and 1.2 a 1 1 ratio appears most effective in anti-VEGF assays with NHEK cells. See Figure 5 Compositions with cationic reagents appear 20-40% more active than those with Cellfectin® (Figure 6).

Short duration of exposure to oligonucleotide formulation is sufficient to observe long term inhibitory effect on VEGF production: VEGF expression was reduced after relatively bπef exposures to the compositions disclosed m this invention. For example, incubations of 4 hours demonstrated more anti-VEGF activity than was observed with overnight oligonucleotide exposures.

See Figure 7-9. Suφnsingly, the effect lasted for at least 3 days, the entire duration of the expeπment. Figure 7-9.

Other expenments showed that antisense oligonucleotides with mixed phosphorothioate- phosphodiester chimeπc backbones were potent inhibitors of VEGF expression. In particular, the chimeπc vanants of T30639 (Sequence ID No. 2) contaimng 10 phosphorothioate and 8 phosphodiester linkages, and lipid end-modifications S96-5296 (Sequence ID No. 20) and S96-5297

(Sequence ID No 21) demonstrated excellent activity in the presence of SpdC/DOPE. (Figure 7)

Inhibition of VEGF lasted for over 3 days after only a 4 hour incubation. (Figure 7) In the absence of SpdC, the chimeπc oligonucleotide backbones do not affect VEGF expression. Thus, oligonucleotides with fewer phosphorothioate linkages may have improved efficacy and reduced nonspecific effects. Uptake appear crucial to oligonucleotide efficacy.

Example Vm In Vivo VEGF Inhibition:

A. Specific Aims

Increased expression of Vascular Endothelial Growth Factor (VEGF) has been implicated m the progression of ocular neovasculaπzation associated with proliferative diabetic retinopathy, neovascular glaucomas, and many other blinding conditions. Retinal ischemia leads to increased synthesis of the angiogenic protein VEGF, which tnggers the proliferation of vascular endothelial cells, resulting in the formation of an abnormally large number of blood vessels in the retma, optic nerve, and ins. As yet, there is no accepted therapeutic treatment for this condition. Our overall objective is to apply rational design and testing procedures to identify novel, potentially therapeutic antisense oligonucleotide inhibitors of VEGF expression, with the aim of treating retinal lschemia- associated neovascularization m humans. Our recent in vitro data m human cell culture systems indicate that we can prepare specific oligonucleotide formulations that inhibit the cellular expression of VEGF by more than 50% in the submicromolar concentration range. Our goal for this proposal is to extend our in vitro findings into a rat model of VEGF-associated neovasculaπzation. Our specific

1. Synthesize a 'library' of antisense oligonucleotides directed agamst rat VEGF mRNA. There are 3 major and 1 minor splice vanants of VEGF Ten oligonucleotides will be targeted to sequences in the common region of the RNAs. They will also contain nuclease resistant backbones, and modified bases to improve binding affinity to target mRNAs 2. Establish rat cell monolayer and spheroid models to evaluate the activity and toxicity of the antisense oligonucleotides and their formulations. The C6 ghoma cell line will be used for this because it has been widely used for studying VEGF function. The spheroid (mass of cells) will be useful for assessing whether our oligonucleotides are able to penetrate cell layers 3. Evaluate the efficacy or oligonucleotides using vanous cellular uptake enhancers.

Compounds developed at Aronex will compared with commercially available agents.

4. Develop a proof-of-concept assay for obtaining data to support the antisense mechanism. This in vitro expeπment is designed to test whether we can specifically target just one isotype of VEGF, to answer the question whether our oligonucleotides are really working by the predicted mode. This will aid in the future design of antisense oligonucleotides.

5. Use an in vivo rat model of ms neovasculaπzation for testing of the most promising antisense compounds. These studies will be done in collaboration with Dr Anthony Adamis at Harvard.

At the completion of the proposed studies, we expect to have quantitative information about the in vivo efficacy of 1-2 oligonucleotides. Some information about the dosing, potential mechanism of action, cellular availability, and potential toxicity will also be known If any of the oligonucleotides reduce vasculaπzation and/or VEGF expression by 20% in vivo, we would consider that to be a positive development and proceed to more detailed studies in animal models dunng Phase 11 ofthe research oligonucleotide is selected, the nsk of nonspecific binding to other RNAs would be unacceptably high, and choosing a sequence with high G-content may lead to undesirable G-quartet formation, which reduces the availability of free-coil form available for binding (Bishop et al, 1996) An alternative approach, which we propose to take, is to use selectively modified oligonucleotides containing C5-propynyl pynmidmes, a modification that leads to very efficient binding without significant toxicity (Wagner et al., 1993, Fenster et al., 1994). We have recently shown that the propynyl modification seems to work especially well (see Preliminary Results).

Approaches to improve the nuclease resistance of oligonucleotides: Oligonucleotides with a natural phosphodiester backbone are highly susceptible to serum and cellular nucleases We have determmed that a random sequence 17-base oligonucleotide has a half-life of less than 3 mmutes in serum (Bishop et al., 1996). One alternative is to use oligomers with phosphorothioate backbone (Stem et al., 1991), a modification that markedly improves the serum half-life of oligonucleotides to a day or more. The use of phosphorothioate linkages is believed to lead to some nonspecific effects at high levels, but as discussed below, we are proposing to synthesize specially modified oligonucleotides that work at very low concentrations, thus reducing the nsk of nonspecific interaction. Oligonucleotides with alternative backbones have been tested but all have far more nonspecific effects than phosphorothioate oligonucleotides. Approaches to improve the cellular uptake of oligonucleotides: Subcellular distribution studies show that cells treated with fluorescent oligonucleotides accumulate in peπnuclear endosomal compartments (Guy-Caffey et al, 1995) The rate-limiting step in the internalization process appears to be transport of oligonucleotides across the plasma membrane or the membranes of the endosomal compartments. There are two potential ways to enhance die transport of oligonucleotides across the lipid bilayer of membranes. In the first approach, the oligomer is covalently conjugated to a compound that improves its membrane association and permeability properties, e.g., by conjugating to cholesterol (Letsinger et al., 1989) We have recently identified a novel propπetary modification, a hpophihc ferrocene tether (see Expenmental Design), that seems to markedly improve the efficacy of antisense oligonucleotides. Alternatively uptake enhancers such as cationic lipids or liposomal preparations may be used. These agents are attractive because of their versatility- the same delivery vehicle may be coadministered with a vanety of oligonucleotides. The design of these cationic lipids incoφorates a positively-charged head group that binds to the nucleic acid, and a membrane interactive tail that is proposed to interact with fusogenic hpids and/or destabilize cellular membranes The activity of many cationic hpid preparations (such as Lipofectin) is influenced by factors such as composition and quantity of nucleic acid, cell type, and the concentration of serum in the cell growth medium. In addition, some preparations are cytotoxic. These constraints severely limit the utility of many of these compounds as delivery agents for therapeutic oligonucleotides m animal systems, and there continues to be a tremendous demand for effective uptake enhancers. We have synthesized a novel delivery vehicle, spermidine-cholesterol (SpdC) that improves the cellular uptake and membrane permeability of oligonucleotides, even in the presence of serum (Guy-Caffey et al., 1995) Formulations using this compound will be evaluated as part of the proposed studies.

Our aim is to identify antisense formulations that inhibit VEGF expression by cells m the eye, with concomitant reduction in disease-associated angiogenesis. These studies were prompted by our recent discovery that chemically modified antisense oligonucleotides can have potent inhibitory activity at submicromolar doses. In addition the development of an animal model of VEGF- associated neovasculaπzation by Dr. Adamis will allow us to test the most efficacious compounds in vivo. Preliminary results Summary- The objective of our preliminary, ongoing seπes of expenments has been to discover and/or test general pπnciples and approaches to improve the activity of antisenses oligonucleotides in cell-based models of VEGF expression. Pπmaπly, our aim has been to: • develop quantitative cell-based screening assays to measure the effect of antisense oligonucleotides on VEGF protein levels. • synthesize oligonucleotides containing structural modifications that lead to improved binding affinity for target mRNA, greater nuclease resistance, and greater specificity. • test formulations using novel uptake enhancers (some developed at Aronex) to increase the cellular internalization and membrane penetration ofthe administered oligonucleotide Oligonucleotide design: Because we expect one of our compounds to be eventually tested m humans, this work was initiated using antisense oligonucleotides directed against human VEGF mRNAs. For achieving maximal inhibition of expression, we selected antisense oligonucleotides that are complementary to sequences shared by all four VEGF mRNAs (Table 3)

Antisense oligonucleotides^*

(initiator AUG codon at mRNA seq. 57)

T30638. mRNA seq. 87-105 5'-a*g*a*g*C*a*g*C*a*a*g*g*C*g*a*g*g*C*t-3 T30639: mRNA seq. 185-023 5'-g*C*g*C*U*g*a*U*a*g*a*C*a*U*C*C*a^*'^*U*g -3 T30640- mRNA seq.204-222 5'-C*g*a*U*U*g*g*a*U*g*g*C*a*g*U*a*g*C*t-3 T30641: mRNA seq. 232-250 5'-U*a*C*U*C*C*U*g*g*a*a*g*a*U*g*U*C*C*a-3

Total phosphorothioate (*) backbone to confer nuclease resistance C-5 propynl pynmidines to improve binding affinity to RNA target

Table 3 Antisense oligonucleotides directed against human VEGF

In agreement with initial reports (Wagner et al, 1993), we found that the C5-propynyl pyπmidine base substitutions increased the Tm of duplex formation, an indicator of the affinity ofthe strands for each other, from -60° C for an unmodified oligonucleotide to over 80° C. For use as controls, we synthesized a 'sense' oligonucleotide ofthe same size and modifications T30691. Cellular testing of oligonucleotides: The activity of antisense oligonucleotides, their modified counteφarts, and vanous formulations was evaluated using cultured human keratinocytes, a pnmary cell line that secretes VEGF. Cells were plated at a density of 50,000 cells/well of a 46 well plate m 0.5 ml KGM medium (Clonetics) The level of VEGF secreted into the growth medium was measured by an enzyme linked immunosorbent assay (ELISA) (R&D Systems). The assay is linear over the 5- 1000 pg/ml range. Our measurements show that when NHEK cells are grown in the recommended medium, 100,000 cells plated in 0.5 ml medium produce "150 pg of VEGF 15 hours (^"300 pg/ml in the supernatant untreated control)

Most cell-based assays were done usmg a commercially available cationic liposome formulation, marketed as transfection agent for gene delivery into cells (Cellfectin, from Life Technologies). Other commercial preparations of transfection agents were found to be either toxic or relatively ineffective (7 tested). One cuπous effect of Cellfectin is that when administered to cells alone, as a control, it actually enhances VEGF production. The reason for this is not known. More recently, we have begun to use formulations of spermidine-cholesterol (Guy-Caffey et al., 1995). In very early experiments, using phosphorothioate antisense oligonucleotides without C5- propynyl modifications, we observed no effect on VEGF levels (up to 5 μM extracellular concentration, in the presence or absence of uptake enhancers). At higher levels, there was a small amount of what appeared to be nonspecific inhibition (data not shown). When C5-propyne containing pyrimidines (Wagner et al., 1993) were substituted for cytosines and thymidines, there was a 20° C improvement in the Tm measured in vivo, indicating that the oligonucleotide can bind to its synthetic RNA target with much greater affinity than the unmodified variant (data not shown). We tested the effect of these oligonucleotides in the presence or absence of uptake enhancers, and also varied the ratio of uptake enhancer to oligonucleotide, in an effort to identify an optimal formulation. We observed that different oligonucleotides have different effects on the VEGF levels.

02|ιMolgo4-10μøMeF

Figure 3. Three of four oligonucleotides had activity in the 0.2 μM range, in the presence of 10 μg/ml Cellfectin. The control sense oligonucleotide had no effect (not shown). The most potent of the test oligonucleotides, T30639, has since been used for subsequent optimization studies.

Optimal oligonucleotide to uptake enhancer ratio: In a follow-up experiment, we maintained the ratio of oligonucleotide (T30639 antisense and T30691 sense control) to the cationic lipid component of cF at 1:3 mass ratio and measured the effect on VEGF production. Again T30639+cF showed specific anti-VEGF activity, while the control oligonucleotide had no effect Figure 11.

VtMit |ι-ιin..«>

cβpe-wdM-id* eonawtm-lon (μMl

Figure 11. Effect of oligonucleotide Cellfectin (1:3) formulation on VEGF expression. At high concentrations, there appeared to be nonspecific effects (not shown). It is important to note that we have not made an attempt to separate free uptake enhancer from bound matenal. This probably happens de facto when we change the relative ratios of one to the other.

Effect of oligonucleotide size: In the next expenment, we asked whether it would be feasible to reduce the oligonucleotide size while maintaining the specific anti-VEGF activity Shorter oligomers are also cheaper to synthesize. However, using the same NHEK assay, we found that the 19-base oligonucleotide was more efficacious than the 16 or 14 base derivatives, all oligonucleotides administered with 10 μg/ml Cellfectin. Changing the ratio of Cellfectin to oligonucleotide did not alter the relative activity (not shown). Effect of formulations of spermidine-cholesterol+DOPE or DC-chol+DOPE (liposomal preparations; 1 : 1 by wt) on oligonucleotide efficacy: recently, we have begun to explore alternatives to Cellfectin, which may be somewhat toxic to cells after a long term exposure. One uptake enhancer, spermidine-cholesterol conjugate, (SpdC) (Guy-Caffey et al. 1995,) has been found to be not toxic to cells at levels used, and there is no detectable toxicity m rodents treated for up to 1 week. The cationic lipid DC-Chol (Gao et al., 1991) has been approved for clinical tπals of gene therapy,and it has very low level of toxicity in cellular systems. The preliminary data indicate that formulations of these novel lipids were 20-40% more potent than Cellfectin in parallel expenments.

Short duration of exposure to oligonucleotide formulation is sufficient to observe long term inhibitory effect on VEGF production: In all of the above expenments, we had been incubating the cells with the oligonucleotide+uptake enhancer overnight. We then asked whether a shorter duration of exposure of cells to the oligonucleotide might achieve the same level of anti-VEGF activity. We discovered that indeed, a wash after 4 hours and return to fresh unsupplemented medium did not diminish the anti-VEGF activity (Figure 12). Moreover the effect lasted for up to 3 days (length of experiment).

Tune (hours) after addition of 0.2 μM T30B39 + 10 μg/ml cF

Figure 12. Four hour oligonucleotide+ cF treatment, then replacement with plain medium. The inhibition continues for almost 3 days. There was no significant inhibition by the control sense oligonucleotide (not shown).

In fact, the level of inhibition in relation to control (no oligonucleotide) was actually much better than observed previously. One reason for this may be that m the single incubation expenment, the VEGF protein continued to be synthesized from preexisting mRNA (not yet blocked by the antisense oligomer) and accumulated in the medium. By replacing the medium at 4 h, this source of 'background' VEGF was eliminated. These data have an important implication in thei/i vivo testing system because the long-term antisense effect suggests that the drug will not have to be administered frequently. This aspect will have to be checked in the proposed in vitro and in vivo assays.

Ferrocene-conjugated oligonucleotide: We have recently discovered that a metallocene- modified oligonucleotide formulated with an uptake enhancer is the most effective VEGF inhibitor in our in vitro assays, with very little toxicity in the concentration range used (Figure 13) The oligonucleotide formulated with Cellfectin has specific anti-VEGF activity 20 μM concentration. The ferrocene tether has been designed to improve the membrane association of the oligonucleotide (D. Mulvey, Aronex, personal comm.). Furthermore, we postulate that the lipophilic iron moiety may aid in cellular targeting and transmembrane movement of the oligonucleotide, perhaps by exploiting the active transport systems ofthe cell. Further work on the mechanism by which modification is beyond the scope of this grant and is the subject of a separate study. However, the fact that we have observed high activity with ferrocene-modified oligonucleotides suggest that this avenue should be explored as we test oligonucleotides for testing in the in vivo model.

Figure 13. The potent antisense effect ofthe ferrocene-conjugated variant of T0639 antisense

As described in detail in the Experimental Design Section, the adult rat model of iris neovascularization provides a means to test the activity of the antisense oligonucleotides in quantitative manner. In this assay, rats are placed in a hypoxic chamber for 1-21 days, and the increase in the vascularization ofthe iris is quantified by digital imaging. As the data show (Figure 14 there is a clear progression in the degree of vasculature with increasing length of incubation. The retinal RNA level also rises but not to the same extent (Figure 15). Encouraged by our preliminary data, and the availability of the rat model, data, we are now proposing to obtain similar proof of concept in an in vivo model of angiogenesis. -j_rfc V-^*-^*-*! D-f-1-.ifv Diff^prπnrc VEGF mRNA Ltvtli of Rit Rcllna tn Hypo»l»

Figure 15

C. Relevant Experience

The Principal Investigator, Nilabh Chaudhary, Ph.D., has broad-based experience in cellular an molecular biology. He received his doctorate in Biochemistry from the University of Western Ontario, London, Canada, in 1984. Subsequently he was awarded a Damon Runyon- Walter Winchell Cancer Fund Fellowship to continue his postdoctoral studies in Cell Biology in Dr. Gunter Blobel's laboratory at The Rockefeller University, New York. In 1986, he was appointed an associate of The Howard Hughes Medical Institute in the same laboratory. Dr Chaudhary joined Triplex (recently renamed Aronex) in 1992 as a Research Scientist to initiate a program in cell biology directed toward the development of techniques for improving the cellular and nuclear uptake of nucleic acid therapeutics. He has closely collaborated with a team of organic chemists to design, synthesize, and test novel oligonucleotide modifications and uptake enhancement reagents. Dr. Chaudhary has co- authored scientific papers on the structure-function relationship of potentially therapeutic oligonucleotides, and devised approaches to enhance their cellular internalization and efficacy. He has experience in the design of cell-based assay systems, immunochemical techniques, microquantitation of proteins, nucleic acid purification and molecular cloning techniques, subcellular fractionation, membrane protein and lipid isolation, and fluorescence microscopy.

Anthony P. Adamis, M.D., is a collaborator and consultant on this project. He is an Assistant Professor in the Department of Ophthalmology at Harvard Medical School and a Research Associate in the Department of Surgery at Children's Hospital, Boston. In 1994, Dr. Adamis and his colleagues demonstrated, for the first time, a causative link between increased ocular VEGF levels, angiogenesis. and progression of proliferative diabetic retinopathy, a primary cause of blindness. In studies conducted since then, be has confirmed the physiological role of hypoxia in stimulating VEGF expression, leading to neovascularization in the eye. His breadth of research experience encompasses clinical studies in patients, development of rodent models of disease design of cell-based proof-of- concept assays, and utilization of molecular biology and cloning techniques. He has published over 20 papers, with 10 in the field of angiogenesis. D. Experimental Design and Methods Summary of approach: Our objective in this proposal is to demonstrate the efficacy of anti- VEGF antisense compounds m an animal model of angiogenesis In preparation for this, we are proposing to carry out a seπes of m vitro expenments that will guide us toward the most promising antisense formulation for in vivo testing. We will begin by screening a 'library' of ten candidate antisense oligonucleotides (19 bases) targeted to the rat VEGF mRNA The oligonucleotides will contain C5-propynyl pynmidmes to improve binding affinity for target mRNA, and phosphorothioate internucleotide linkages to confer nuclease resistance.

For cellular testing, we plan to use the rat C6 (ghoma) cells, which respond to hypoxia by producing copious amounts of VEGF mRNA and protem Assays earned out in 96-well format will be used screen the activity of the vanous antisense or control oligonucleotide preparations The time course of their effect on the level of secreted VEGF m the extracellular medium will be monitored by ELISA. To improve cellular uptake, oligonucleotides will be coadmintstered with novel uptake enhancers. Different ratios of nucleic acids and lipids will be tested. In addition, the two 'best' antisense sequences will be selected for conjugation to a 3'-hpophιhc ferrocene tether, a modification that may contribute to the cellular entry of the antisense oligonucleotide. Also, the effect of the two best oligonucleotides (or their formulations) on VEGF mRNA levels will be determined by Northern blotting (and compared to the effect of appropnate controls).

In an effort to provide evidence for the antisense mechanism, a separate seπes of in vitro expenments is planned. C6 cells will be treated with antisense oligonucleotides specially designed to be VEGF isotype-specific, i.e , to target only one or two species of VEGF mRNA (3 major, one minor in the rat). RNAse protection assay will be used to measure the relative levels of each species of VEGF mRNA. In pnnciple, if the antisense effect is truly sequence-specific, only the expression of the targeted isotype should be down regulated Oligonucleotides of different sequence should be ineffective. The cellular toxicity of the most effective antisense compounds will be assayed in two different cell lines, and the two least toxic formulations will be tested in C6 cell spheroid models, designed to determine whether oligonucleotides can penetrate across cell layers. VEGF rnRNA levels in successive layers of cells in the spheroid will be determmed by m situ hybπdization The utility of uptake enhancers and tethers will also be checked m this model. The anti-angiogemc activity of the most effective anti-VEGF oligonucleotide will then be evaluated m animals, using a rat eye model of ins neovasculanzation. Albmo rats will be placed in low oxygen chamber (up to 2 weeks) and the vasculaπzation in the ins monitored by a noninvasive, quantitative digital imaging procedure. In this model, increased vascularization is noticeable after only 1-2 days of hypoxia The test oligonucleotide (or formulations) will be introduced directly into one eye ofthe rat, with the other eye seeing as an untreated as control. After up to 1 week of exposure, any effect on vascular growth will be quantified. Changes m the levels of VEGF protem (m the vitreous, if possible), and mRNA levels in the retina will be checked by ELISA and Northern blotting respectively Any side effect will be noted. Depending on the initial results, a multidose experiment will be attempted. Effect of control oligonucleotides will also be checked. The least toxic, most effective formulations (>20% inhibition of neovascularization) will then be tested in an extended series of animal trials, as part of Phase II of theses studies. Selection of antisense and control oligonucleotides: To achieve maximal blocking of VEGF expression, we will synthesize antisense oligonucleotides that bind 10 the common region of all VEGF isotypes in the rat (Conn et al., 1990). Ten essentially randomly selected oligonucleotides, with no obvious haiφin motifs or G-rich stretches, will be synthesized for testing in the first round of screening. 5 will be rat-specific and 5 sequences will be chosen to bind to human mRNA as well. It is not clear whether evolutionarily conserved sites are 'good' or 'bad' antisense targets, although most recent evidence suggests there is no predictable, preferred location for antisense targeting. All synthetic oligonucleotides will contain C5-propynyl pyrimidines (to improve binding affinity for target) and phosphorothioate linkages (nuclease resistance) (Wagner et al., 1993, Fenster et al., 1994). We already have several 'irrelevant' oligonucleotides that we use as controls, but if necessary, we will synthesize a control oligonucleotide of the same size and base composition as the antisense sequence. The oligonucleotides will be synthesized, purified (>95%, by HPLC), and characterized by The Oligonucleotide Synthesis Group at Aronex.

Oligonucleotides for mechanism of action studies: For obtaining data that supports the antisense mechanism of action, several ("4; depending or efficacy) of 20-mer isotype-specific oligonucleotides will be prepared. An oligonucleotide directed against a sequence found only on VEGF- 165 mRNA should not bind to VEGF- 120 mRNA. Similarly, a 20-base oligonucleotide complementary to the splice junction of VEGF-120 (i.e., 10 bases per exon) should not be able to bind well to VEGF-165. For use as control, oligonucleotides with reversed sequences will be synthesized (two halves will be reversed). The effect of these oligonucleotides on VEGF expression will be determined by comparing the relative levels of the various mRNAs. We will use the RNAse protection assays to quantify the relative levels of each mRNA (Ambion, Austin, TX) . The probes for doing this (ranging from "150 to 250 bases long) have already been prepared using rat mRNA sequence-specific primers and RT-PCR technology (Perkin Elmer). Cell culture: The biological screening will be conducted in C6 glial cells derived from rat glioma. The predominant isoforms of VEGF in this cell line are VEGF-165 (amino acids) and VEGF-120 (46% each), while VEGF-188 accounts for only about "8% (Bacic et al., 1995). This cell line has been widely used to investigate VEGF structure and function. To induce VEGF synthesis by stimulating with hypoxia, cells will be placed in a low oxygen chamber (GasPak Plus anaerobic culture chamber (BBL Microbiology Systems) with hydrogen and palladium catalyst to remove all oxygen (Stein et al., 1995). Typical incubations times will range from 6-18 h. Alternatively the cultures will exposed to 100-300 μM cobalt chloride, which interferes with the heme-dependent hypoxia response system and activates a hypoxia response factor that induces the transcnption of VEGF mRNA.

Evaluation of antisense oligonucleotide activity m vitro: C6 cells, grown in monolayers, will be maintained in Dulbecco's medium with 5% fetal bovine serum and antibiotics. For the preliminary testing of anti-VEGF oligonucleotides, cells will be plated at a density of 10,000 or 20,000 cells/well, m a 96 well dish. After 1 day of recovery, the cells will be treated with oligonucleotide (in .25 ml medium). Two types of medium will be tested, the regular serum-contammg C6 medium, or Optimem (Life Technologies), the reduced-serum medium that is often used to improve transfection efficiency by reducing interference by serum components. After varying peπods of incubation in the hypoxia chamber, the supernatant will be transferred to a fresh plate for further analysis by ELISA. As a rule, when new formulations are tested we examine the cellular moφhology through a microscope to look for unusual changes or any obvious signs of toxicity.

For RNA analysis, a larger number of cells (>2xl0⁶ to 10⁷ cells in T75 flask) will be treated with a select number of formulations. After oligonucleotide treatment (and exposure to hypoxia, etc.) the supernatant will again be saved for ELISA, and RNA will be isolated and analyzed using methods descπbed below.

VEGF ELISA Assay: There is no commercial kit available yet for rodent systems so we are devising one using antibodies known to react well with rat VEGF (RDI-1020 or RDI-4060 from Research Diagnostics, Inc., and another from R&D Systems). Other antisera to VEGF are also available so we will choose the best combination. ELISA reagents (enzyme-linked second antibody, substrate) have been purchased from Pierce.

RNA extraction, Northern blots, RNAse protection assays, and hybπdization probes: VEGF mRNA size is in the range of 3.8 to 4 kilobases, mainly because of the long untranslated region. For RNA analysis, total RNA will be isolated from treated or untreated cells by the RNAzol method (Tel-Test, Inc., Fπendswood, TX). For use as probes, VEGF-specific segments corresponding to the common region and isotype-specific probes have already been generated by a combination of reverse transcπptase-polymerase chain reaction (RT-PCR kit, Perkin Elmer) usmg C6 RNA and VEGF- specific pnmers followed by size selection of cDNAs onginating from different mRNAs, and selective amplification usmg isotype-specific pπmers These probes will be cloned into PCRII plasmid cloning vector (Invitrogen) and the sequence will be confirmed by sequencing (Sequenase, USB). The PCRII vector allows the RNA polymerase dependent production of radiolabeled RNA probes for use in RNAse protection assays (kit from Ambion, Austin, TX). A β-actin probe will be used to normalize the RNA levels. For Northern blots, RNA (up to 20 μg) will be fractionated on formaldehyde gels, transferred to nylon, and probed with radiolabeled probes according to standard procedures, with β-actin to normalize RNA levels. In all RNA assays, phosphσrimaging (Fuji Phosphoπmager) will be used to quantify the relative levels of radioactivity. Expenments to support the antisense mechanism of action: The anttsense-ohgonucleotides in this study are complementary to the mRNA sequence encoding VEGF mRNA. However, their inhibitory effect m biological system does not necessaπly prove an antisense mechanism of action. In fact, recent analyses indicate that many oligonucleotides may interfere nonspecifically with cellular metabolism, especially at concentrations above 1 μM (reviewed in Stein and Cheng, 1993). Proof of antisense mechanism is deceptively difficult, and has not really been shown except by circumstantial evidence. Our current expenment, though indirect, has been designed to obtain evidence for probable antisense mechanism. In bπef, we have synthesized "isotype-specific" antisense oligonucleotides which will be used in a rat cell-based assay to selectively inhibit the expression of a smgle VEGF vanant. Rat C6 cell (glioma ongm) produce three mam types of VEGF isotypes. RNAse protection assays have shown that about 45% of VEGF m C6 cell Ime is 120 amino acid vanant, 45% is 165aa vanant , and the remaining is 188 aa vanant (Bacic et al., 1995). Cells will be treated with antisense oligonucleotide specific for one or more isotypes, and then mRNA transcnption will be stimulated either by hypoxia, or by the addition of 200 μM cobalt chloride, which mimics hypoxia. After 9 h exposure, cellular levels of VEGF mRNA will be monitored by RNAse protection assay (Ambion , Austin, TX).). Very important, isotype-specific probes ("100-200 bases long) that we have already generated by RT-PCR techniques will be used to quantify the level of each VEGF mRNA species, and levels will be normalized relative to β-actin. If antisense mechanism is operative, and a smgle isotype-specific oligonucleotide is used, only the expression of one species of VEGF should be reduced relative to others. On the other hand, an oligonucleotide complementary to a common region of all VEGFs should reduce the expression of every VEGF.

Synthesis of ferrocene-conjugated oligonucleotides: After the first few round of screening, the most active oligonucleotide with fewest side effects will be attached through the 3 end to a ferrocene tether with the hope of further increasing the uptake and specific activity of the oligonucleotide (see preliminary results; T30781 (+ferocene) vs. T30639 (unmodified)). If effective, the activity of this oligonucleotide will he compared to that of a random sequence control containing the same tether. It is hypothesized that the ferrocene moiety may allow the oligonucleotide to exploit the active transport or permeation system (iron?) of the cell, but the mechanism has not yet been studied. Use of uptake enhancers: In most instances, to facilitate cellular entry, the oligonucleotides will be administered to cells m the presence of cationic lipid reagents. Developed as transfection agents for gene delivery, many cationic lipids are now available commercially, but only Cellfectin (Life Technologies) was found to be consistently effective m our assay (of 7 major lipids tested). Cellfectin is a 1:1.5 (wt/wt) liposomal mix of the polyammolipid tetrapalmityl-spermme and the phospholipid dioleoyl phosphotidyl ethanolamine (DOPE). Another lipid we are working with is DC- chol, developed by Leaf Huang (Gao and Huang, 1991), and approved for clinical tnals for gene therapy (Rgene Therapeutics, The Woodlands, TX). In addition, we have designed and synthesized (Guy-Caffey et al, 1995) a series of novel polyammolipid uptake enhancers that markedly increase the cellular uptake of oligonucleotides, even in the presence of serum and without significant associated toxicity. The most effective of these is SpdC (spermidine-cholesterol). When injected i.v. in mice, SpdC enhanced oligonucleotide delivery to many tissues, and no toxicity was observed when SpdC was injected into mice at concentrations as high as 100/mg/kg/day for 1 week (data not shown).

We have prepared liposomal preparations of SpdC/DOPE and DC-chol/DOPE (SpdC or DC-chol with dioleoyl phosphotidylcholine) at mass ratios ranging from 1:0.5, 1:1, 1:1.5, and 1:2. The ratio that worked most effectively in our earlier anti-VEGF assays in NHEK cells was 1:1, for both cationic lipids. We are proposing to investigate the utility of these preparations in enhancing the activity of anti-VEGF oligonucleotides in the C6 cell line, in C6 spheroids, and eventually in the eye model.

The exact mechanism by which these uptake methods work is still controversial. Mixing the cationic lipid with the nucleic acid almost always yields microprecipitates that enter cells efficiently. To maintain consistency, and perhaps to discover an underlying principle, we will monitor the size of the particles by using a Coulter particle sizing apparatus. Formulations for injection in vivo: We have used the term formulation loosely in this proposal to convey the message of using a multicomponent mixture to deliver drugs. However, as we approach the in vivo testing stage, it would be prudent to devise the optimal formulation for injection into the eye, especially if uptake enhancers are involved. Other parameters to consider would be the amount of antisense and concentration, salinity, particle size, pH, and vehicle. Dr. Joe Wyse will assist us choosing and preparing the optimal formulation.

Cytotoxicity assays: Cells will be seeded at a density of 500 cell/well in a 96 well plate. One day after plating, the cells will be exposed to serially diluted oligonucleotide formulations (4 wells per dilution). After one day or four days of exposure, the effect on cellular viability will be monitored using a nonradioactive assay system (Cell Titer 96 Aqueous cell proliferation assay, Promega Coφ.). For the most potent oligonucleotide, this assay will be done in three separate cell lines (including C6, NHEK, and a fibroblast cell line).

Evaluation of antisense activity in spheroids: C6 cells, normally grown in monolayers (4.5 g glucose/1, DMEM, 5% FCS plus antibiotics) can be induced to grow in spheroids or aggregates of cells about 0.4 to 0.8 mm. It would be informative to know whether our antisense oligonucleotides, formulated with lipids or otherwise can go across the layers of cells of a spheroid and still have biological activity. To prepare spheroids, the method described by Stein et al, (1995), will be used. C6 cells will be transferred from confluent cultures to nonadherent bacteriological dishes, and grown for 48 hours. The emerging spheroblasts will be transferred to spinner flasks, grown for an additional 10 days (80 φm), and the spheroids will be sorted into uniform size by sedimentation through a 10 ml pipet. Growth will be continued for an additional 6 weeks, with a medium change every other day. The flask will be flushed each day with 95% air + 5% CO₂ to insure adequate oxygenation and pH. The spheroids will be treated with antisense formulations or appropriate controls, exposed to hypoxia to induce VEGF synthesis for up to 1 day, and then the level of VEGF mRNA in spheroid sections will be examined by in situ hybridization. For this, the spheroids will be fixed with 4% paraformaldehyde, frozen, sectioned into 10 μm thick pieces, and processed for in Situ hybridization with ³⁵S-labeled DNA or RNA probes for VEGF generated as described earlier.

The processed section will be counterstained with hematoxylin and eosin stain. After several days of autoradiography (Guy-Caffey et al.) the slides would be examined (photographed) by bright field and dark-field illumination.

The distribution of VEGF RNA will indicate the degree of inhibition achieved by the antisense oligonucleotide. Ideally, all layers will show low level of VEGF mRNA. Most likely, the superficial layers will have less VEGF, either because the drug did not penetrate into the layers of cells, or because the cells were more hypoxic at the center and produced more VEGF. If the delivery is only into the superficial layers, we will attempt to devise new delivery approaches.

Evaluation of the anti-angiogenic activity of VEGF antisense oligonucleotides in vivo: The adult rodent model of VEGF-associated iris neovascularization: Adult rats in a hypoxic atmosphere stimulate new blood vessel growth on the iris. The neovascularization is correlated in time with the upregulation of VEGF mRNA levels in the retina. The sequence of ocular events closely reproduce those seen in the monkey model of rubeosis iridis and human iris neovascularization, where ischemic retinal VEGF is known to be causal in the development of iris neovascularization. It is our intention to use this model for testing the activity of the antisense compounds that may reduce angiogenesis. The animal experimentation, to be performed in the Adamis laboratory, involve animal handling and surgery, photography, computer quantification, and Northern analysis of VEGF mRNA.

Adult 350-400 gram Kingston colony albino Sprague-Dawley rats (female) are placed in hypoxic chambers (1% oxygen) for 1-21 days. Biomicroscopic examination and standardized slit camera photography is performed before and after the incubation period. Signs of progressive iris vessel dilation and tortuosity as well as increased vascular density develop in the hypoxic atmosphere. The iris vessels are clearly visible in these albino animals. Standardized photographs of the irises are scanned into a computer program (NIH Image, 1.54D software) and the area of neovascularization quantified in a masked manner through pixel analysis. Iris vascularity shows a progressive increase through day 14. Proliferating cell nuclear antigen (PCNA) and Factor VHI immunostaining has confirmed endothelial cell proliferation beginning day 2; proving the increased vascularity represents angiogenesis. Isolated retinas prepared for Northern blotting demonstrate that the hypoxic animals increase steady-state VEGF mRNA levels in the retina. In summary, adult rats in a hypoxic atmosphere stimulate new vessel growth on the iris. It is our intention to use this model to test the effect of candidate anti-VEGF formulations.

Experiment to determine the effect of interrupted hypoxia on neovascularization: The above rat model has only been used for contmuous exposure to hypoxia. Because the model may prove more useful if the hypoxia can be interrupted, e.g., for repeat administration of drugs we would like to characteπze the degree of ins neovasculanzation for animals removed from hypoxia for short penods on a daily basis. We speculate that removal will synergize the neovasculaπzation smce neonatal rats treated in such a manner have a greater neovascular response compared to animals in constant oxygen (Reynaud and Dorey, 1994). Reoxygenation of hypoxic retina produces reactive oxygen intermediates which are known to increase in retinal VEGF mRNA protem (Kuroki et al. 1995).

Rats will be placed m hypoxic chambers (10% oxygen) for 1, 3, 5, 7, 14, and 21 days (n=3 per time point, 18 total). They will be removed for ore hour per day and exposed to normal room oxygen (21%). At the end ofthe incubation penod, standardized ins photographs will be taken. The animals will be sacnficed and the antenor half of the eye prepared for PCNA staining and light microscopy. The retinas will be isolated and prepared for Northern blotting with a³²P-labeled 558 bp VEGF cDNA random pπme-labeled probe. Retinal VEGF mRNA upregulation will be correlated with the photographic and immunohistochemical documentation of ins neovasculanzation over the 21 day time penod. The area of vasculanty will be quantified from the standardized photographs and compared to animals placed in uninterrupted hypoxia. From this expenment, we will be able to estimate the maximum number of times the animal can be taken out of the hypoxia chamber and dosed, without compromising the hypoxic effect.

Evaluation of antisense effect m vivo: On day 0, baseline photographs will be taken. One eye of an anesthetized animal will be randomized to receive a smgle mtravitreal injection of the VEGF antisense compound. The other eye will be left untreated Doses will be guided by them vitro studies but will be administered in 20 μl or less. The animals will be placed in uninterrupted hypoxia for 7 days. In the first such assay, 6 animals will be treated per formulation (4 formulations: oligo only; ohgo+tether; and each with uptake enhancer). The aim will be to see whether these compounds have any acute effect. If there is noticeable reduction in vasculaπzation, a larger number of animals will be used. The table below summaπzes the statistical rationale for how many might be needed. In total, it is expected that "20 rats per treatment will be used. Table 1. Number of rats needed: Ninety percent of rats m this model develop neovascularization. Assuming a level of significance of 0.05 (α = 0.05) and a power of 0.8 (l-β-0.80) the number of eyes per treatment group can be determined for the varying effectiveness of an angiogenesis blocking agent. Control Group Treatment Group Numberof Eyes Required

90 40 38

90 50 24

90 60 17

90 70 12

Inhibition of neovascularization for these experiments will be defined as a decrease of 20% in the area of vascularization in the treated versus control eyes. If the assumed effectiveness of a particular agent is high, the percentage of eyes developing ins neovasculanzation is m the treatment group will be low, and the number of eyes required for the statistical significance will decrease dramatically.

On day seven, the animals will be photographed and sacπficed. The retinas will be harvested and prepared for Northern blotting (individually). VEGF steady state mRNA will be quantified following normalization to the 285 ribosomal RNA signal, using a Phosphorimager (Molecular Dynamics). Ins vasculaπty will be quantified and compared between treated and control eyes.

F. Vertebrate Animal handling procedures: All animal procedures will be done Children's Hospital, Harvard Medical School, following the guidelines established by the Association for research in Vision and Ophthalmology resolution on the use of animals for research and guidelines established by the Massachusetts Eye and Ear Infirmary Animal Care Committee. In total, 120 Albino female Kingston Colony Sprague-Dawley rats will be used. They will be anesthetized and injected intravitrealy with 20 μl of oligonucleotide or control formulation, using a 30 gauge needle. Gentamycin sulfate will be applied following the injections. Animals will be kept for up to 3 weeks in a 10% oxygen atmosphere. Following the termination of treatment, they will be sacnficed byC0₂ inhalation. All experiments are consistent with OPRR guidelines. I. Literature Cited

Adamis, A.P., et al., 1994, Arch. Ophthalmol. 118:445-450

Adamis, A.P., et al., 1993, Biochem. Biophys. Res. Comm., 193:631-638 Adamis, A.P., et al., 1996, Arch. Ophthal. 114:66-71

Bacic et al., 1995 Pharm. News 2:V, (abst)

Bishop et al., 1996, 3. Biol. Chem.271:5698-5703.

Conn, (3. et al., 1990, Proc. Natl. Acad. Sci. USA 87:2628-2632

D'Amore, P.A., 1994, Investigative Ophthai. Vis. Sci. 35:3974-3979 deYries, C, et al., 1992, Science 255:989-991.

Fenster, S.D., et al., 1994, Biochemistry 33:8391-8398

Ferrera, N., et al., 1992., Endocnne Reviews 13:18-31

Guy-Caffey et al., 1995, J.Biol.Chem. 270:31391-31396

Gao, X., and Huang, L.1991, Biochem. Biophys. Res. Comm. 179:280-285 Miller, J. et al., 1994. Am. J. Pathol. 145: 574-584. Milligan, J.F.. et al., 1993, J.Med.Chem. 36:1923-1937. Plate, K.H., et al., 1992, Nature 359:845-848 Shima, D.T., et al., 1995, Molecular Med. 2:64-75 Shweiki et al., 1992, Nature 362: 841 -844

Stein, CA. et al., 1991, Pharmac. Ther. 52:365-384 Stein, CA and Cheng, Y.C., 1993, Science 261:1004-1012 Terman et al., 1992, Biochem. Biophys. Res. Comm. 187:1579-1586 Uhlmann, F., and Peyman, A., 1990, Chem. Rev.90:543 Wagner, R.W. et al., 1993, 260:1510-1513 Wagner, R.W., 1994, Nature 372:333-335.

REFERENCES CITED

The following references to the extent that they provide procedural details supplementary to those set forth herein, are specifically incoφorated herein by reference.

Ballaun, C, Weninger, W., Uthman, A., Weich, H., Tsehaehler, E. (1995) J. Invest. Ital. Dermatol.

104, 7-10.

Behr, J. (1994) Ital. Bioconjugate Chem. 5, 382-389

Bishop, J. S., Guy-Caffey, J. K., Ojwang, J. O., Smith, S. R, Hogan, M. E., Cossum, P. A.,

Rando, R. F., Chaudhary, N. (1996) J. Biol. Chem. 271, 5698-5703 Carmeliet, P., Ferreriera, V., Breier, G., Pollefeyt, S., Kieckens, L., Gertsenstein, M., Fahrig,

M., Vandenhoeck, A., Haφal, K., Eberhardt, C, Declercq, C, Pawling, J., Moons, L., Collen, D.,

Risau, W., Nagy, A. (1996) Nature 380, 435-439

Chaudhary, N., Bishop, J. S., Jayaraman, K., Guy-Caffey, J. K., in Delivery Strategies For

Antisense Oligonucleotide Therapeutics, S. Akhtar, Ed. (CRC Press, Boca Raton, 1995) pp. 39-60. Connolly, D. T., Plander, J. V. (1989) J. Biol. Chem. 264, 20017-20024

D'Amore, P. A. (1994) Invest. Ophthalmol. Vis. Sci. 35, 3974-3978 de Vries, C, Escobedo, J. A., Ueno, H., Houck, K., Ferrera, N., Williams, L. T. (1992) Science 255,

989-991

Dvorak, H. F., Brown, L. F„ Detmar, M., Dvorak, A. M. (1995a) Am. J. Pathol. 146, 1029-1039 Dvorak, H. F„ Detmar, M., Claffey, K. P., Nagy, J. A., van der Water, L., Senger, D. R. ( 1995b) Int.

Arch. Allergy Immunol. 107, 233-235

Fenster, S. D., Wagner, R. W., Froehler, B. C, Chin, D. J. (1994) Biochemistry 33, 8391-8398

Ferrera, N., Houck, K., Jakeman, L., Leung, D. W. (1992) Endocr. Rev. 13, 18-32

Ferrera, N., Carver-Moore, K., Chen, H., Dowd, M., Lu, L., O'Shea, K. S., Powell-Braxton, L., HiUan, K. J., Moore, M. W. (1996) Nature 380, 439-442

Fisher, T. L., Terhorst, T., Cao, X., Wagner, R. W. (1993) Nuc. Acids Res. 21, 3857

Folkman, J. (1995) Nat. Med. 1, 27-31

Frank, S., Hubner, G., Breier, G., Longaker, M. T., Greenhalgh, D. G., Werner, S. (1995) J. Biol.

Chem. 270, 12607-12613 Gao, X., Huang, L. (1991) Biochem. Biophys. Res. Commun. 179, 280-285

Gao, X., Huang, L. (1991) Biochem. Biophys. Res. Commun. 179, 280-285

Guy-Caffey, J. K., Bodepudi, V., Bishop, J. S., Jayaraman, K., Chaudhary, N. (1995)7. Biol. Chem.

270, 31391-31396

Keck, P. J., Hauser, S. D., Krivi, G., Sanzo, K., Warren, T., Feder, J., Connolly, D. T. (1989) Science 246, 1309-1312

Ledley, F. (1994) Curr. Opin. Biotechoi 5, 626-636 Leung, D. W., Cachianes, G., Kuang, W. J., Goedddel, D. V., Ferrera, N. (1989) Science 246, 1306-

1309

Lewis, J. G., Lin, K.-Y., Kothavale, A., Flanagan, W. M., Matteucci, M. D., DePrince, R. B., Mook,

J., R.A., Hendren, R. W., Wagner, R. W. (1996) Proc. Natl. Acad. Sci. USA 93, 3176-3181 Milhgan, J. F., Matteucci, M. D., Martin, J. C. (1993) J. Med. Chem. 36, 1923-1937

Nomura, M., Yamagishi, S., Harada, S., Hayashi, Y., Yamashima, T., Yamashita, J., Yamamoto, H.

(1995)7. Biol. Chem. 270, 28316-28324

Puglisi, J. D., Tinoco, J., I. (1989) Methods Enzymol. 180, 304-325

Robinson, G. S., Pierce, E. A., Rook, S. L., Foley, E., Webb, R., Smith, L. E. H. (1996) Proc. Natl. Acad. Sci. USA 93, 4851 -4856

Stein, C. A., Cheng, Y.-C. (1993) Science 261, 1004-1012

Stein, C. A., Kreig, A. M. (1994) Antisense Res. Dev. 4, 67-69

Terman, B. I., Dougher-Vermazen, M., Carrion, M. E., Dimitrov, D., Armellino, D. C,

Gospodarowicz, D., Bohlen, P. (1992) Biochem. Biophys. res. Commun. 187, 1579-1586 Thomas, K. A. ( 1996) J. Biol. Chem. 271 , 603-606

Tischer, E., Mitchell, R., Hartman, T., Silva, M., Gospodarowicz, D., Fiddes, J. C, Abraham, J. A.

(1991) 7. Biol. Chem. 266, 11947-11954

Uhlmann, E., Peyman, A. (1990) Chemical Reviews 90, 543-584

Wagner, R. W., Matteucci, M. D., Lewis, J. G., Gutierrez, A. J., Moulds, C, Froehler, B. C. (1993) Science 260, 1510-1513

Wagner, R. W. (1994) Nature 372, 333-335

Winternitz, C. I., Jackson, J. K., Oktaba, A. M., Burt, H. M. (1996) Pharm. Res. 13, 368-375

Woolf, T. M., Melton, D. A., Jennings, C. G. B. (1992) Proc. Natl. Acad. Sci. USA 89, 7305-7309

SEQUENCE LISTING

(1) GENERAL INFORMATION:

(1) APPLICANT: Chaudhary, Nilabh Rao, T. Sudhakar Revankar, Ganapathi R. Cossum, Paul A. Rando, Robert F. Peyman, Anusch Uhlmann, Eugen

(11) TITLE OF INVENTION: Inhibitors of Vascular Endothelial Growth Factor Expression

(ill) NUMBER OF SEQUENCES: 21

(iv) CORRESPONDENCE ADDRESS:

(A) ADDRESSEE: Conley, Rose & Tayon, P.C.

(B) STREET: 600 Travis, Suite 1850

(C) CITY: Houston

(D) STATE: Texas

(E) COUNTRY: U.S.A.

(F) ZIP: 77002-2912

( v) COMPUTER READABLE FORM :

(A) MEDIUM TYPE : Floppy dis k

(B) COMPUTER: IBM PC compatible

( C ) OPERATING SYSTEM : Wmdows 95

( D ) SOFTWARE : Microsoft Word 7 . 0a

( vi ) CURRENT APPLICATION DATA:

(A) APPLICATION NUMBER:

(B) FILING DATE:

(C) CLASSIFICATION:

(vn) PRIOR APPLICATION DATA:

(A) APPLICATION NUMBER:

(B) FILING DATE:

(vm) ATTORNEY/AGENT INFORMATION:

(A) NAME: McDamel, C. Steven

(B) REGISTRATION NUMBER: 33,962

(C) REFERENCE/DOCKET NUMBER: 1472-07200

(ix) TELECOMMUNICATION INFORMATION:

(A) TELEPHONE: 713/238-8010

(B) TELEFAX: 713/238-8008

(2) INFORMATION FOR SEQ ID Nθ:l:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: synthetic nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(E) ANTI-SENSE: Y ( iv) ANTI-SENSE : YES

( ix) FEATURE :

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(C) OTHER INFORMATION: /note= "Phosphorothioate linkage between each residue"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:l:

GCGCTGATAG ACATCCATG 19

(2) INFORMATION FOR SEQ ID NO:2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note= "Phosphorothioate linkage between each residue, C5-propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:2:

GCGCUGAUAG ACAUCCAUG 19

(2) INFORMATION FOR SEQ ID NO:3:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note= "Phosphorothioate linkage between each residue, C5-propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:3:

CGAUUGGAUG GCAGUAGCCT 19

(2) INFORMATION FOR SEQ ID NO:4:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid (C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note= "Phosphorothioate linkage between each residue, C5-propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 4:

(2) INFORMATION FOR SEQ ID NO:4:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note= "Phosphorothioate linkage between each residue, C5 propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:4:

UACUCCUGGA AGAUGUCCA 19

(2) INFORMATION FOR SEQ ID NO:5:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 6-7; 9-10; 10-11; 13-14

(D) OTHER INFORMATION: /note= "Phosphorothioate linkage between the indicated residues"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:5:

GCGCUGAUAG ACAUCCAUG 19

(2) INFORMATION FOR SEQ ID NO: 6:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid (C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 3-4; 5-6; 6-7; 9-10; 10-11; 13-14

(D) OTHER INFORMATION: /note= "Phosphorothioate linkages except between indicated residues"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:6:

GCGCUGAUAG ACAUCCAUUG 19

(2) INFORMATION FOR SEQ ID NO:7:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: mιsc_feature

(B) LOCATION: 6-7; 10-11

(C) OTHER INFORMATION: /note= "Phosphorothioate linkage except between indicated residues"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:7:

GCGCUGAUAG ACAUCCAUG 19

(2) INFORMATION FOR SEQ ID NO:8:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(C) OTHER INFORMATION: /note= "Phosphorothioate linkage between all residues, C5-propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:8:

GAAGAUGUCC ACCAGGGUC 19

(2) INFORMATION FOR SEQ ID NO: 9:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages between indicated residues, C5-propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 9:

AGGAAGCUCA UCUCUCCUA 19

(2) INFORMATION FOR SEQ ID NO: 10:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note= "Phosphorothioate linkages, CS- propynyl pyrimidines "

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 10:

UACACGUCUG CGGAUCUUG 19

(2) INFORMATION FOR SEQ ID NO: 11:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note= "Phosphorothioate linkages, CS- propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 11:

UAACUCAAGC UGCCUCGCC 19

(2) INFORMATION FOR SEQ ID NO: 12: (i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note= "Phosphorothioate linkages, CS- propynyl pyrimidines "

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 12:

CCAUGAACUU CACCACUUC 19

(2) INFORMATION FOR SEQ ID NO:13:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl pyrimidines "

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 13:

GACAUCCAUG AACUUCACC 19

(2) INFORMATION FOR SEQ ID NO: 14:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:14:

GGCUGGCAGU AGCUGCGCU 19 (2) INFORMATION FOR SEQ ID NO: 15:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 15:

GGAUGGCAGU AGCUGCGCU 19

(2) INFORMATION FOR SEQ ID NO: 16:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl C residues"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 16:

GCGCTGATAG ACATCCATG 19

(2) INFORMATION FOR SEQ ID NO: 17:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl U residues"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 17: GCGCUGAUAG ACAUCCAUG 19

(2) INFORMATION FOR SEQ ID NO:18:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl pyrimidines at residues 8, 12, 14 and 15"

(xi) SEQUENCE DESCRIPTION: SEQ ID N0:18:

GCGCTGAUAG ACAUCCATG 19

(2) INFORMATION FOR SEQ ID NO:19:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl pyrimidines at residues 2, 5, 8, 12, 15 and 18"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:19:

GCGCUGAUAG ACATCCAUG 19

(2) INFORMATION FOR SEQ ID NO:20:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl pyrimidines, 3' end linked to lipid tether" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:20: GCGCUGAUAG ACAUCCAUG 19

(2) INFORMATION FOR SEQ ID NO:21:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: mιsc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, except between residues 1-5, 8-9, 12-13, 14-15, and 16-19; 3' -terminal pyrene; CS- propynyl pynmidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:21:

GCGCUGAUAG ACAUCCAUG 19

(2) INFORMATION FOR SEQ ID NO:22:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: mιsc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages; C5- hexynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:22:

GCGCUGAUAG ACAUCCAUG 19

(2) INFORMATION FOR SEQ ID NO:23:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: mιsc_feature

(B) LOCATION: 1-19 (D) OTHER INFORMATION: /note- "Phosphorothioate linkages, C5- propynyl pyrimidines"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:23:

GCGCUGACAG ACAUUCAUG 19

(2) INFORMATION FOR SEQ ID NO:24:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: NO

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:24: CATGGATGTC TATCAGCGC 19

(2) INFORMATION FOR SEQ ID NO:25:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: NO

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:25: CATGGATGTC TATCAGCGC 19

(2) INFORMATION FOR SEQ ID NO:26:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl C RESIDUES"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:26:

AGAGCAGCAA GGCGAGGCT 19

(2) INFORMATION FOR SEQ ID NO:27: (i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 19 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(iv) ANTI-SENSE: YES

(ix) FEATURE:

(A) NAME/KEY: misc_feature

(B) LOCATION: 1-19

(D) OTHER INFORMATION: /note- "Phosphorothioate linkages, CS- propynyl C RESIDUES"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:27:

CAUGGAUGUC UAUCAGCGC 19

Claims

WHAT IS CLAIMED IS:

1 An antisense oligonucleotide that reduces cellular VEGF production in cells treated with said antisense oligonucleotide at concentrations of less than about 1 micromolar; said treated cells producing no more than about 90 percent ofthe VEGF that is produced by untreated cells.

2. The antisense oligonucleotide of claim 1 wherein said antisense oligonucleotide bmds to RNA sequences found on mRNA that encodes VEGF.

3 The antisense oligonucleotide of claim 1 wherem said antisense oligonucleotide binds to an RNA sequence found on at least two ofthe mRNAs that encode VEGF.

4. The antisense oligonucleotide of claim 1 wherein said antisense oligonucleotide binds to an RNA sequence found on mRNA VEGF 206.

5 The antisense oligonucleotide of claim 1 wherein said antisense oligonucleotide binds to an RNA sequence found on mRNA VEGF 185

6 The antisense oligonucleotide of claim 1 wherein said antisense oligonucleotide bmds to an RNA sequence found on mRNA VEGF 165.

7. The antisense oligonucleotide of claim 1 wherein said antisense oligonucleotide binds to an RNA sequence found on mRNA VEGF 121.

8. The antisense oligonucleotide of claim 1 which compπses a chemical moiety that decreases the rate of degradation of said antisense oligonucleotide by nucleases

9 The antisense oligonucleotide of claim 1 wherein said oligonucleotide compπses a phosphorothioate group and a phosphodiester group.

10. The antisense oligonucleotide of claim 1 compnsmg a pair of adjacent residues connected through a chemical moiety that resists degradation by nucleases.

11. The antisense oligonucleotide of claim 8 wherem said moiety that decreases the rate of degradation of said antisense oligonucleotide by nucleases is a phosphorothioate group.

12. The antisense oligonucleotide of claim 8 wherem each residue of said oligonucleotide is linked through a phosphorothioate group.

13. The antisense oligonucleotide of claim 1 wherem said oligonucleotide compπses a nucleotide residue selected from the group consisting of C5-propynyl undine, C5-propynyl cytidine, C5- hexynyl undme, C5-hexynyl cytidine, 6-aza-undme, and 6-aza- cytidine.

14. The antisense oligonucleotide of claim 1 wherein said oligonucleotide compπses a phosphorothioate group and a nucleotide residue selected from the group consisting of CS- propynyl undme, C5-propynyl cytidine, C5-hexynyl undme, C5-hexynyl cytidine, 6-aza- undme, and 6-aza- cytidine.

15. The antisense oligonucleotide of claim 1 compnsmg a C5-propynyl undme residue

16. The antisense oligonucleotide of claim 1 compnsmg a C5-propynyl undme residue and a phosphorothioate group.

17. The antisense oligonucleotide of claim 1 compnsmg a C5-propynyl cytidine residue.

18. The antisense oligonucleotide of claim I compnsmg a C5-propynyl cytidine residue and a phosphorothioate group.

19. The antisense oligonucleotide of claim 1 compnsmg a C5-hexynyl undme residue and a phosphorothioate group.

20. The antisense oligonucleotide of claim 1 compnsmg a C5-hexynyl cytidine residue.

21. The antisense oligonucleotide of claim 1 compnsmg a C5-hexynyl cytidine residue and a phosphorothioate group.

22. The antisense oligonucleotide of claim 1 compnsmg a 6-aza-deoxy undine residue.

23. The antisense oligonucleotide of claim 1 compnsmg a 6-aza-deoxy undine residue and a phosphorothioate group.

24 The antisense oligonucleotide of claim 1 compπsing a 6-aza-deoxy cytidine residue.

25. The antisense oligonucleotide of claim 1 compnsmg a 6-aza-deoxy cytidine residue and a phosphorothioate group.

26 A composition compnsmg an antisense oligonucleotide that reduces cellular VEGF production in cells treated with said antisense oligonucleotide at concentrations of less than about 1 micromolar said treated cells produce, at most, about 90 percent of the VEGF that is produced by untreated cells, said composition further compnsmg a cellular uptake enhancer.

27. The composition of claim 26 wherein said cellular uptake enhancer is a hpophihc moiety.

28. The composition of claim 27 wherein said hpophihc moiety compπses cholesterol.

29. The composition of claim 1 wherein said oligonucleotide further compπses an ionic bond to a cation to form a salt.

30. The composition of claim 29 wherein said cation is a cationic lipid.

31. The composition of claim 30 wherem said cationic lipid is a polyammolipid.

32. The composition of claim 31 wherein said polyammolipid is spermidine-cholesterol

33. The composition of claim 26 said cellular uptake enhancer compnsmg a liposome.

34. The composition of claim 33 said liposome compnsmg Cellfectin®.

35. The composition of claim 33 said liposome compnsmg spermidine-cholesterol mixed with DOPE.

36. An antisense oligonucleotide that binds VEGF mRNA and compnses a phosphorothioate group and a nucleotide residue selected from the group consisting of C5 -propynyl undme, C5 -propynyl cytidine, C5 -hexynyl undme, C5 -hexynyl cytidine, 6-aza-deoxy undine, and 6-aza-deoxy cytidine, wherein said antisense oligonucleotide has a duplex melting temperature of at least about 5°C above the melting temperature of an identical antisense oligonucleotide that lacks chemically modified pynmidme residues; said antisense oligonucleotide reduces cellular VEGF production in cells treated with said antisense oligonucleotide at concentrations of less than about 1 micromolar; said treated cells producing no more than about 90 percent of the VEGF that is produced by untreated cells.

37. A composition compπsing an antisense oligonucleotide that binds VEGF mRNA and compπses a phosphorothioate group and a nucleotide residue selected from the group consisting of CS- propynyl undme, C5-propynyl cytidine, C5-hexynyl undme, C5-hexynyl cytidine, 6-aza-deoxy undme, and 6-aza-deoxy cytidine, wherem said antisense oligonucleotide has a duplex melting temperature of at least about 5°C above the melting temperature of an identical antisense oligonucleotide that lacks chemically modified pynmidme residues; said antisense oligonucleotide reduces cellular VEGF production in cells treated with said antisense oligonucleotide at concentrations of less than about 1 micromolar; said treated cells producing no more than about 90 percent of the VEGF that is produced by untreated cells; said composition also compnsmg a polymeπc sustained release compound.

38. In an antisense oligonucleotide that contains phosphorothioate linkages and binds to mRNA encoding VEGF an improvement compnsmg: including a nucleotide residue m the antisense oligonucleotide selected from the group consisting of C5 -propynyl undme, C5-propynyl cytidine, C5-hexynyl undme, C5-hexynyl cytidine, 6-aza-deoxy undine, and 6-aza-deoxy cytidme; wherem said improvement increases the duplex melting temperature of an antisense oligonucleotide by at least about 5°C.

39. A method for reducing cellular VEGF production in a cell compnsmg contacting a cell with antisense oligonucleotide of claim 1, said cell producmg no more than approximately 90 percent of the VEGF that is produced by an uncontacted cell, the concentration of said oligonucleotide bemg less than about 1 micromolar.

40. An antisense oligonucleotide that is selected from the group consisting of sequence ID numbers 2- 22.

41. A composition compnsmg antisense oligonucleotide havmg sequence ID number 2.

42. A composition compnsmg antisense oligonucleotide having sequence ID number 3

43. A composition compnsmg antisense oligonucleotide having sequence ID number 4

44. A composition compnsmg antisense oligonucleotide having sequence ID number 6.

45. A composition compnsmg antisense oligonucleotide having sequence ID number 10.

46. A composition compnsmg antisense oligonucleotide having sequence ID number 20.

47. A composition compnsmg antisense oligonucleotide having sequence ID number 21.

48. A composition compπsing antisense oligonucleotide having sequence ID number 22.