CROSS-REFERENCE TO RELATED APPLICATIONS

Priority is hereby claimed to US provisional application Serial No. 60/624,906, filed 04 November 2004, incorporated herein by reference.

INCORPORATION BY REFERENCE

All of the papers and prior patents cited below are incorporated herein by reference.

INTRODUCTION

The present invention fills a gap in crafting conjugates that contain ribonucleic acids (RNA) in general, and short interfering RNA in particular, and using the conjugates as therapeutic agents. Short interfering RNAs (siRNAs) have been dubbed in the popular scientific literature as "one of the hottest things in target validation in the last five years." This enthusiasm is due to the fact that siRNAs have been shown to be efficacious in specifically suppressing the expression of targeted genes.

This sequence-specific post-transcriptional silencing of gene expression via the action of siRNAs is a naturally occurring process first discovered in fungi and plants, and later shown to occur in bacterial and animal cells. Short interfering RNAs are short (ca. 21-25 nucleotides) RNA fragments, obtained either enzymatically or by chemical synthesis. Short interfering RNAs function by inducing sequence-specific degradation of targeted messenger RNAs (mRNAs). The potential of siRNAs as potent antiviral and anticancer drugs is widely accepted in the scientific community. However, before this potential becomes a reality in the form of pharmaceutical compositions, fundamental problems of delivering the siRNAs to their intended targets must be solved.

At the root of the delivery problem is that siRNAs are polyanions. Thus, unassisted permeation of siRNAs across lipid bilayers is negligible. siRNAs are conventionally delivered to cells using cationic liposomes, or polyplexes with

polyethyleneimines. Although the use of liposomes to deliver siRNAs has shown some success, the major disadvantage of a liposome delivery vehicle is that a number of cell types cannot be transfected using liposomes. Also, several cell types cannot be liposome-transfected with an efficiency that produces significant biological effects. Moreover, in experiments using liposome delivery vehicles several different manipulations of the cells are required. In short, the process is cumbersome.

Similarly, polyethyleneimines are difficult to deliver into cells because they are high molecular- weight polymers. Thus, in vivo delivery of polyethyleneimine- siRNA polyplexes is plagued by all the obstacles inherent in the systemic delivery of a high molecular-weight cationic complex: the complexes must make their way to the intended site, extravasate into the targeted tissue, etc. Therefore a method of moving siRNA molecules from the extracellular environment into the cytoplasm of target cells would be a significant breakthrough in the therapeutic use of siRNAs to silence genes in vitro {e.g., in cultured and somatic cells) and in vivo {e.g., systemic delivery of siRNAs as drugs).

SUMMARY OF THE INVENTION

The invention is thus a method to create an efficacious cell delivery system for siRNAs that mimics a naturally occurring process, and the resulting siRNA delivery system. Many antibiotics and low molecular-weight enzyme inhibitors have been conjugated to amines. The amine moiety is often crucial for increasing biological activity. In many of these compounds, it is the amine moiety that provides the structural elements required to ferry the conjugate into cells. In particular, the antibacterial activities of many antibiotics are due (in part) to amines or polyamines conjugated to glycosidic, aromatic, or polyketide moieties. The cationic polyamine residues function to facilitate transport of the antibiotics into the cells. A fitting example is streptomycin (see Fig. 4), which is an aminoglycoside wherein the transporting residues are aminoguanidino groups. Similarly, in the bleomycin-phleomycin group of antibiotics (see Fig. 2), modified spermine and spermidine groups are the transporting residues. In spermidine-conjugated antibiotics and antitumorals (such as spergualin, laterosporamine, the edeins, glisperins A, B,and C, and glycocinamoylspermidines) (see Fig. 3), a conjugated spermidine moiety is the transporter of the biologically active agent into the cells.

Mention should also be made of the aminoglycoside antibiotics called kanamycins. In kanamycins, an amino moiety ferries the glycosides into the cells. This is also true of squalamine (see Fig. 7), a broad-spectrum, steroidal antibiotic isolated from the tissues of the dogfish shark. In these compounds, a sulfated bile acid is fused to the aminopropyl primary amine of spermidine. It is the spermidine portion of the molecule that acts to carry the remainder of the molecule across cell membranes.

In the present invention, natural and/or synthetic polyamine groups are covalently bonded to siRNA molecules via bonds that either: 1) are broken in the cytoplasm and set the siRNA moiety free; or 2) remain intact, while still enabling the siRNA portion of the conjugate to bind to the RNA-inducing silencing complex (RISC) molecule and thereby to cleave the specific mRNA targeted by the siRNA. (See Dykxhoorn et al., Nature Reviews, RNAi Collection, p.7, December 2003).

The preferred polyamines for use in the present invention are preferably spermidine and derivatives thereof, spermine and derivatives thereof, and hirudonine and derivatives thereof. See Figs. 1 and 6. Most cells take up polyamines by carrier-mediated, energy-dependent mechanisms. Many cells (for instance, human fibroblasts, mouse leukemia cells, rat Morris hepatoma cells, etc.) appear to have a single transporter for all polyamines. Thus, as a general proposition, the specificity of the transporter is not terribly stringent. For example, it is known from previous work that derivatives of polyamines substituted with alkyl substituents are also efficiently transported into cells by the same transporter system that mobilizes the natural polyamines.

Polyamines are also ideal carriers for siRNAs because of the large binding affinity of polyamines to ribonucleic acids. Polyamines, especially spermine, strongly bind to ribosomes, and are constitutive parts of transfer RNAs (t-RNAs). The strongly basic polyamines bind to t-RNAs by hydrogen bonds as well as by electrostatic charges (Frydman et al., Proc. Natl. Acad. Sd (USA) (1992) 89:9186; Fernandez et al., (1994) Cell. MoI. Biol. 40:93). Transfer RNAs are ribonucleotides roughly twice the size of siRNAs, but the ribonucleotide chains have similar structures. Without being bound to any particular biological mechanism, in t-RNA complexes, spermine or spermidine bind to a loop of ribonucleotides; in siRNAs the polyamine could bind to and stabilize the annealed double-strand chain of the siRNA until the siRNA binds to the RNA-inducing silencing complex.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 depicts the chemical structures of putrescine, spermidine, and spermine.

Fig. 2 depicts the chemical structures of the bleomycin class of compounds.

Fig. 3 depicts the chemical structures of spergualin and edeines A and B.

Fig. 4 depicts the chemical structure of streptomycin.

Fig. 5 depicts the chemical structures of 4-coumaroylagmatine and hordatine A, B, and M.

Fig. 6 depicts the chemical structure of hirudonine.

Fig. 7 depicts the chemical structure of squalamine.

Fig. 8 is a schematic illustrating the siRNA-mediated, sequence-specific cleavage of mRNA.

Fig. 9 is a schematic illustrating the siRNA-mediated, sequence-specific cleavage of mRNA, and illustrating the RNA-inducing silencing complex (RISC).

Figs. 1OA and 1OB depict the chemical structures of protected ribonucleosides that can be used to fabricate oligoribonucleotide siRNAs.

Fig. 11 is a reaction scheme illustrating phophoramadite-based oligoribonucleotide synthesis.

Fig. 12 is a reaction scheme illustrating deprotection of the ribonucleoside shown in Fig. 1OB

Fig. 13 depicts the chemical structure of three different protected ribonucleosides for use in the present invention.

Fig. 14 depicts a series of polyamine-siRNA conjugates according to the present invention.

Fig. 15 depicts another series of polyamine-siRNA conjugates according to the present invention.

Fig. 16 depicts yet another series of polyamine-siRNA conjugates according to the present invention.

Fig. 17 depicts a general reaction scheme for fabricating polyamine-siRNA conjugates according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION A first embodiment of the invention is directed to a conjugate comprising a polyamine covalently bonded to a ribonucleic acid (RNA). It is preferred that the

RKA is an oligo-RNA, and most preferably that the RNA is a short interfering RNA (siRNA).

The polyamine portion of the conjugate can any polyamine, without limitation. The preferred polyamines are those selected from the group consisting of putrescine, spermine, spermidine, hirudonine, and derivatives thereof. Explicitly included within those derivatives are the conformationally restricted polyamine compounds disclosed in U.S. Patent Nos. 6,392,098 and 6,794,545. More specifically, for conformationally-restricted polyamines, the polyamine portion of the conjugate is preferably selected from compounds of Formula I:

E— NH- D— NH- B— A— B— NH- D— NH- E (I)

wherein A is selected from the group consisting OfC ₂ - to C ₆ -alkene and C ₃ - to C ₆ -cycloalkyl, cycloalkenyl, and cycloaryl;

B is independently selected from the group consisting of a single bond and Ci- to C ₆ -alkyl and alkenyl;

D is independently selected from the group consisting OfC ₁ - to C ₆ -alkyl and alkenyl, and C ₃ - to C ₆ -cycloalkyl, cycloalkenyl, and cycloaryl; and

E is methyl; and pharmaceutically-suitable salts thereof.

As detailed in greater detail hereinbelow, these conformationally-constrained polyamines can be made according to the following general reaction scheme: A compound of Formula II:

HO— B— A— B-OH (II)

is reacted with a protecting reagent, preferrably mesitylenesulfonyl chloride, to yield a compound of Formula III:

PROT— 0— B— A— B— O— PROT (III)

wherein PROT is the protecting group.

Then, the Formula III compound is reacted with a compound of Formula IV:

E-N(PROT) -D-NH-(PROT) (IV)

to yield a compound of Formula V:

E-N(PROT) -D-N(PROT) -B-A-B-N(PROT) -D-N(PROT)-E (V)

It is much preferred that the protecting group, PROT, in both the Formula III intermediate and the Formula IV intermediate be a mesitylenesulfonyl moiety.

The Formula V compound is then deprotected to yield a compound of Formula I (E- NH- D— NH- B— A— B— NH- D— NH- E).

A second embodiment of the invention is directed to a method of mobilizing RNA into a living cell. The method comprises conjugating the RNA to a polyamine to yield a conjugate; and then contacting the conjugate to the living cell. It is preferred that the RNA is conjugated to a polyamine selected from the group consisting of putrescine, spermine, spermidine, hirudonine, and derivatives thereof. It is also preferred that a siRNA is conjugated to the polyamine.

A third embodiment of the invention is directed to a composition of matter for mobilizing RNA into a living cell, the composition comprising a polyamine covalently bonded to a ribonucleic acid (RNA), in combination with a pharmaceutically suitable carrier.

In the present invention, polyamine transporters are bound to the siRNAs by collapsible bonds that will release the ribonucleotides as they enter the cells. Crafting discrete conjugates of polyamines with siRNAs that are devoid of systemic toxicity when administered in vivo, opens the way to study gene function by modulating naturally occurring processes in mammalian cells, ultimately leading to new gene-specific therapeutic agents for treating disease. Thus, the conjugates disclosed herein are highly useful for delivering siRNAs from outside a cell and into the cytoplasm and/or nucleus.

1. Polyamine Transport of Natural Products into Cells:

The recognition of the ubiquity and essentiality of the polyamines in animal tissues, as well as the ability to culture such tissues, has resulted in numerous studies of polyamine transport in mammalian cells. The inhibition of growth of tumor cells by specific inhibitors of polyamine synthesis and the restoration of growth by exogenous polyamines has been a major stimulus to the study of polyamine uptake. For instance, several modes of uptake of [ ¹⁴ C]- benzylamine are evident: adsorption

to cell surfaces (the adsorbed molecules are removable in large part by washing); a saturable, intracellular transport indicative of a specific process; and a slower, nonsaturable uptake, suggestive of a mix of diffusion and internal binding (Cohen, A Guide to the Polyamines, Oxford, 1998, pp. 467).

There are three natural polyamines present in mammalian cells; putrescine, spermidine, and spermine (see Fig. 1). As summarized by Seiler and Dezeure (1990) Int. J. Biochem, 22:211, polyamine uptake in mammalian cells is generally specific, saturable, requires energy, and is carrier-mediated. Polyamine transport can also be regulated in a process requiring RNA and protein synthesis. A review of the data on three membrane receptors of mammals has described marked similarities or homologies of these proteins to the numerous transporter proteins for amino acids and polyamines isolated from eubacteria and fungi (Reizer et al., Protein Science (1993) 2:20). These data suggest that the receptor proteins from mammal cells may be described as modified transport proteins sharing a common origin with the transport proteins of these microbes.

Many cells have a single transporter for putrescine, spermidine, and spermine, with K _m values in the micromolar range. Both sodium-dependent and sodium-independent systems have been detected. Stressing the therapeutic importance of polyamine transport in cancer, reviewers pointed to the decreased virulence of a leukemia cell line lacking a transport system. Several new inhibitors of putrescine uptake inhibit the uptake of this diamine more actively than that of spermidine, see Minchin et al. (1991), Eur. J. Biochem, 200:457. This also indicates that the transport of these different polyamines in melanoma cells are different.

The uptake of putrescine in human platelets is saturable and energy- dependent but appears complex. Studies with human erythrocytes, which are unable to synthesize the polyamines, demonstrated polyamine receptor sites at the cell surface and the uptake of putrescine, spermidine, and spermine largely ( >95%) into an internal soluble compartment. Moulinoux et al. (1984), Biochimie, 66:385. The uptake was minimal at 4 ⁰ C, and this related mainly to binding to cell stroma. At 37 ⁰ C, uptake of putrescine and spermidine was relatively rapid from serum whereas spermine entered slowly. The K _m values for the saturable putrescine and spermidine uptakes from plasma were 125 and 3.6 μM, respectively.

In one strain of mouse cells, both Na+ and spermidine were shown to enter cells in a 1:1 relationship. Khan et al. (1990) Cell. MoI. Biol, 36:345. The transport appeared to be ATP-independent because it was unaffected by 2-deoxyglucose, which depletes ATP. An examination of many more strains of mammalian cells showed that spermidine uptake was affected by Na+ concentration, although somewhat differently in each case. Khan et al. (1990), Pathobiology, 58:172. The uptake was generally inhibited by ionophores and some polyamine analogs. Preloading the cells with asparagine accelerated putrescine uptake in two strains, as well as of putrescine uptake in neuroblastoma cells. Rinehart and Chen (1984) J. Biol. Chem. 259:4750. The uptake of putrescine by human colon cancer cells is stimulated > 300-fold by asparagine. McCormack and Johnson (1991) Am. J. Physiol. 256:G868. The nature of the asparagine effect on putrescine uptake appears to be greatly exaggerated in some human cancer cells. It is known that asparagine activates a membrane Na+/H+ antiport, provokes an extrusion of H+, and causes Na+- dependent intracellular alkalinization. Fong and Law (1988) Biochem. Biophys. Res. Commun.155:937.

Studies with a wide variety of animal cells indicated the common responses of polyamine transport mechanisms. Nevertheless, some tissue cells and cancer cells appear to possess a single common mechanism. In contrast, others, like the pig renal LLC-PK cell line, revealed several more discriminating systems. The latter cells contained both sodium-dependent and sodium-independent transporters localized in different cell areas. Van der Bosch et al. (1990), Biochem. J. 265:609.

By taking advantage of the aforementioned transport mechanism of polyamines into animal cells, different drugs and chemical agents were covalently attached to the natural polyamines and were thus ferried into the cells. A relatively non-toxic compound, a nitroimidazole-spermidine conjugate entered cells and inhibited uptake of the free spermidine, see Holley et al. (1992), Biochem. Pharmacol. 43:763. It can also be mentioned that a spermine conjugate of an iron chelator is efficiently transported into tumor cells, even if the terminal substituents are quite large. See Bergeron et al. (2003), J. Med. Chem. 46:5478, 2003. Polyamine conjugates with naphtalimides are also efficiently delivered into the cells, Lin et al. (2003), Biochem. Soc. Trans. 31(2):407; indenoisoquinolines (topoisomerase I inhibitors) conjugated with polyamines are delivered into cancer cells, Nagarayan et al. (2003), J. Med. Chem. 46:5712; the spermine conjugate of

the protease inhibitor known as the Bowman-Birk inhibitor ("BBI," an 8000 Da polypeptide) is very effective in localizing BBI in lung and liver, and has none of the unacceptable toxicity of the polylysine conjugate, Kennedy et al. (2003), Pharm. Research, 20(12): 1908.

Natural products are a good sampler of polyamines as vectors for intracellular delivery. Mention should be made of the bleomycins (see Fig. 2 for the various chemical structures), with their wide array of polyamine vectors, as well as the spermidine conjugates spergualin and the edeins (see Fig. 3), and streptomycin, a diguanidino derivative (see Fig. 4). The aminoguanidines found in plants, such as cumaroylagmatine, and the hordatines (see Fig. 5) illustrate the importance of the strongly basic guanidine residue for transport, a residue also found in hirudonine (see Fig.6), an important bacterial and plant polyamine. Squalamine (see Fig. 7) is a spermidine-steroid, where the spermidine moiety is the delivery vector into the cells. All of these natural products include a polyamine moiety.

The present inventor has found that the specificity of the polyamine transport mechanism is surprisingly permissive. After surveying 24 polyamine-like compounds, including spermine and homospermine analogues, pentamines and different oligoamines, it was found that they are efficiently transported into human cells. The subject conjugate are thus expected to be efficiently transported into mammalian cells, including human cells.

2. Syntheses of Small Interfering RNAs (siRNAs):

Small interfering RNAs (siRNAs) are double-stranded fragments of about 21-23 ribonucleotides. It has been shown that siRNA molecules are the mediators of mRNA degradation, and that chemically synthesized duplexes with the fragment pattern mentioned above are capable of guiding mRNA cleavage. Elbashir et al. (2001), Genes and Development, 15:188. The currently accepted chain of events in the siRNA-mediated cleavage of mRNA is presented schematically in Fig. 8. As shown in Fig. 8, siRNAs include a paired sense strand (red, shown in 5' to 3' at the top of Fig. 8) and an antisense strand (blue, shown 3' to 5' at the top of Fig. 8), with a 3' overhang, usually (TT) or (UU), see reference "a" in Fig. 8. The siRNA pathway starts when a long double-stranded (ds) RNA is cleaved by the RNase m enzyme having the trivial name "Dicer," into siRNAs in an ATP-dependent reaction. These siRNAs are then incorporated into the RNA-inducing silencing complex (

RISC). Once uncoupled, the single-stranded antisense strand of the siRNA guides the RISC complex to messenger RNA (mRNA, which is single-stranded) and targets a complementary sequence of the mRNA. This results in the endonucleolytic cleavage of the targeted mRNA (as shown in reference "b" of Fig. 8). There are also, however, data suggesting that in chemically synthesized siRNA duplexes, both the sense as well as the antisense strand target the mRNA. This process is illustrated schematically in Fig. 9. As shown in Fig. 9, a transfected siRNA is incorporated into the RISC, and either the sense or the antisense strand (they are not delineated in Fig. 9) can serve to recognize the complementary sequence in the targeted mRNA. See Duxbury et al. (2004), J. Surgical Res., 117:339.

Synthesis, purification and annealing of siRNAs by chemical processes is becoming increasingly popular. See, for example, Micura (2002), Angew. Chem. Int. Ed, 41(13):2265; and Hobartner et al. (2003), Monatshefte fur Chemie, 134:851, 2003. Chemically synthesized RNA oligonucleotides are by key components of siRNA technology. The coupling of the nucleosides is achieved by conventional and well-known phosphoramidite chemistry as illustrated in Fig. 10. Because the process is well-known to those skilled in the art, it will not be described in great detail. See the citations in the following paragraphs for a full treatment. There is also a thriving commercial market in custom RNA synthesis. For example, siRNAs of any specified sequence can be purchased from, for example, SynGen, Inc. (San Carlos, California), Midland Certified Reagent Company, Inc. (Midland, Texas), and Dharmacon, Inc. (Boulder, Colorado), among many other companies. Additionally, many university-based labs also sell custom RNA synthesis services to the public (e.g., The University of Wisconsin Biotechnology Center [Madison, Wisconsin] and Kansas State University [Manhattan, Kansas], among many others).

Improvements in the structure of suitable protecting groups has taken routine RNA synthesis to the level of product quality and accessible oligonucleotide length as is the case for DNA synthesis. The need for robust RNA synthesis strategies resulted in the crafting of new and sophisticated protecting groups. These procedures were introduced in 1998 and are covered by various patents (see, for example, Pitsch et al., US Patent 5,986,084), and described in the scientific literature (Pitsch et al., (2001) He/v. CMm Acta, 84:3773. The most common procedure (used in commercial automated DNA synthesizers after some technical adjustments) makes use of the "TOM" protecting group ( 2'-O-triisopropylsilyloxymethyl) (see

Fig. 10A, which is a protected ribonucleoside with the TOM protecting group at the 2'-0 position). This protecting group is superior to the classical 2'~0-tert- butyldimethylsilyl (TBDMS) group used in DNA synthesis.

The construction of a protected nucleoside such as shown in Fig. 1OA starts with the N-acetylation at the exocyclic amino groups of the nucleobases of the ribonucleosides, followed by tritylation at the 5' oxygen to give 5'-0-DMT. This is then followed by "TOMylation" at 2' oxygen to give the 2'-0-TOM derivative. Lastly, a "phosphitylation" step at the 3' oxygen using 2-cyanoethyl diisopropylphosphoramido chloridite yields the 3'-0-CEPA derivative shown in Fig. 1OA. The incorporation of the phosphoramidites into oligoribonucleotides is well documented. See Micura et al. (2001), Nucleic Acid Res. 29:3997; Hobartner et al. (2002), Angew. Chem. Int. Ed. 41 :605; Micura et al. (2000), Angew. Chem. Int. Ed. 39:922; Micura et al. (2001), Nucleosides, Nucleotides, Nucleic Acids, 20:1287; and Ebert et al. (2000), HeIv. Chim. Acta, 83:2238).

Another building block for oligoribonucleotides is depicted in Fig. 1OB. This protected ribonucleoside was first reported in 1998 in U.S. Patent No. 6,111,086, see also Scaringe (2000), Methods in Enzymology, 317:3. The rationale behind the protected nucleoside shown in Fig. 1OB was the desire for a mildly acidic aqueous condition for the final deprotection at the 2'-0 group of the synthetic ribonucleotide. This cannot be achieved if the protecting group at 5'-0 is DMT, which is itself labile to mild acidic conditions. DMT was therefore replaced with the 5'-O-bis(trimethylsiloxy)cyclododecyloxysilyl ether (DOD), together with the 2'-O- bis(2-acetoxyethyloxy)methyl (ACE) orthoester. Herein the 3'-OH group is derivatized as the methyl-N,N-diisopropylphosphoramidite because the cyanoethyl group (used in the TOM-protected nucleosides of Fig. 10A) proved to be unstable to the fluoride reagents needed to cleave DOD. The coupling yields with this protected ribonucleoside are higher than 99%.

After the oligonucleotide assembly, the phosphate methyl protecting groups are removed with disodium 2-carbamoyl-2- cyanoethylene-l,l-dithiolate trihydrate (S ₂ Na ₂ ) in DMF (see Fig.12). Then basic conditions (40% aqueous methylamine) cause oligonucleotide cleavage from the solid support, along with the removal of the acyl protecting groups on the exocyclic amino groups and the acetyl groups on the 2'-orthoesters. The resulting 2'-O-bis(2-hydroxyethyloxy) methyl orthoesters are ten times more acid labile than before the removal of the acetyl groups. As a

consequence, very mild conditions (pH 3.8, 30 min, 60°C) are all that is required for the final deprotection step. The 5'-O- DOD group is cleaved with fluoride reagents. The published HPLC chromatograms of the crude oligoribonucleotides obtained by this procedure are impressive, exhibiting almost no by-products.

The general scheme of the automated synthesis of oligoribonucleotides is depicted in Fig.l 1 (PG= protecting group). The synthesis starts from the 3 '-end by attachment to the resin. Deprotection is achieved in two steps, without degradation of the RNA products, first with CH ₃ NH ₂ in ethanol/water, followed by Bu ₄ NF in tetrahydrofuran. The process can be repeated up to about a 150-mer product before significant product degradation results.

3. Synthesis of Conformationally-Constrained Polyamines:

A host of naturally occurring polyamines, including putrescine, spermine, spermidine, and hirudonine can be purchased commercially from a number of worldwide suppliers such as Aldrich Chemical Co, Milwaukee, Wisconsin and Fisher Scientific, Hampton, New Hampshire. The Aldrich catalog numbers are: putrescine (Dl,320-8), spermine (S383-6), and spermidine (S382-8). The Fisher catalog number for hirudonine (diguanylspermidine, CAS No. 2465-97-6) is ICN222595.

Conformationally-constrained polyamines suitable for use in the present invention are preferably synthesized as disclosed in U.S. Patent Nos. 6,392,098 and 6,794,545. Briefly, various rigid moieties, either cyclic moieties or double- or triple- bonded moieties, are introduced into the backbone of a polyamine.

The first targeted location was the central 1,4-diaminobutane segment of a polyamine. In its staggered conformation, four semi-eclipsed conformational rotamers are possible around the diaminobutane segment. The four have enantiomeric relationships. Introduction of a bond between the C-I and C-3 positions or the C-2 and C-4 positions of the central diaminobutane segment generates a cyclopropane ring. Introduction of an additional bond between the C-2 and C-3 positions generates a conformationally restricted alkene derivative. Cyclobutyl, cyclopentyl, and cyclohexyl moieties can be introduced into the structure following the same strategy.

Using this approach four conformationally semi-rigid structures were obtained which mimic the four semi-eclipsed conformational structures of spermine.

Two of the semi-rigid structures are epimers of the other two.

For purposes of the present invention, it is important to note that the cis and trans isomers of the subject compounds assume very distinct three-dimensional conformations due to the restricted bond rotation afforded by the centrally-located ring structure or unsaturation. All geometric isomers (optically active or otherwise), including pure isolated cis forms and pure isolated trans forms of the polyamines, and mixtures thereof, are explicitly within the scope of the present invention. Additionally, all positional isomers of the subject compounds are explicitly within the scope of the present invention. When A or D of formula I is a cyclical moiety, the two B substituents or the amino moieties, respectively, may be oriented in the 1,2 or 1,3 or 1,4 position with respect to each other.

(a) Spermine Analogs Containing a Cyclopropyl Ring:

Cis and trans cyclopropyl analogs of spermine were prepared via the reactions illustrated in Schemes 1, 2, 3, 4, 4A, 5, and 5A, hereinbelow.

With reference to Schemes 1 and 2, the cyclopropyl diesters 1 and 2 were first converted into their hydrazides 103 and 4, and the hydrazides converted into the diamines 5 and 6, respectively. The diamines 5 and 6 were then mesitylated to give the amides 7 and 8, and the amides were then alkylated with 9 to give 10 and 11, respectively. Hydrolysis of the protective groups yielded the trans analog 12 and the cis analog 13.

Referring now to Scheme 3, in a separate reaction, the trans cyclopropyl diester 1 was converted into the amide 14 by reaction with benzylamine (BnNH ₂ ), the amide reduced to the amine 15, and the amine alkylated to 16. The phthalyl residues were then cleaved with hydrazine to give 17. Compound 17 was then either deprotected by hydrogenolysis to give 18; or fully alklyated to 19, and the benzyl residues cleaved by hydrogenolysis to give 20.

With reference to Scheme 4, the amine 15 was also alkylated with 21 to give 22. Compound 22 was then deprotected to yield the trans cyclopropyl analog 23.

An alternative (and preferred) route to 23 is given in Scheme 4A. Here, 3- ethylamino propionitrile 101 was converted into the corresponding amine 102, which was then mesitylated to yield 3. In a parallel synthesis, the cis diester 1 was reduced to the dialcohol 15', which was then mesitylated to yield the dimesityl derivative 16'. Reacting 3 and 16' in the presence of sodium hydride yields 22'.

In the same fashion as Scheme 4, 22' was then deprotected to yield the trans cyclopropyl analog 23.

Referring now to Scheme 5, in a separate reaction, the cis cyclopropyl diester 2 was reduced to the dialcohol 24. The dialcohol was then converted into the amine 25, and the amine protected by mesitylation to 26. Compound 26 was then alkylated with 9 to yield 27, and then deprotected to yield the cis cyclopropyl tetramine 28.

An alternative (and preferred) route to 28 is given in Scheme 5A. Here, the cis cyclopropyl diester 2 is reduced to the dialcohol 24 in the same fashion as in Scheme 5. Compound 24 was then protected by mesitylation to yield 25'. Compound 25' was then reacted with 3 to yield 27. Deprotecting yields the tetramine 28.

(b) Spermine Analogs Containing a Cyclobutyl Ring: Cis and trans cyclobutyl analogs of spermine were prepared via the reactions illustrated in Schemes 6, 7, 8, 8A, 9, and 9 A, hereinbelow.

Referring now to Schemes 6 and 7, the synthesis of the cyclobutyl derivatives started with the trans and cis 1,2-diaminobutanes 29 and 30, respectively. These compounds were first converted to the amides 31 and 32, and then alkylated to 33 and 34, respectively. Compounds 33 and 34 were then deprotected to yield the trans tetramine 35 (Scheme 6) and the cis tetramine 36 (Scheme 7).

With reference to Schemes 8 and 9, in separate reactions, the trans cyclobutyl diester 37 and the cis cyclobutyl diester 38 were reduced to the respective dialcohols 39 and 40, the dialcohols converted into the diamines 41 and 42. The diamines 41 and 42 were then protected by mesitylation to yield 43 and 44, respectively. These compounds were then alkylated to give 45 and 46. The protecting groups were then removed to yield the trans cyclobutyl tetramine 47 (Scheme 8) and the cis tetramine 48 (Scheme 8).

Alternative (and preferred) routes to 47 and 48 are given in Schemes 8A and 9 A, respectively. The cis and trans di esters 37 and 38 were reduced to the respective dialcohols 39 and 40 in the same fashion as in Schemes 8 and 9. Compounds 39 and 40 were then mesitylated to yield 41' and 42', respectively. Reaction of 41' and 42' with 3 yields 45 (Scheme 8A) and 46 (Scheme 9A). Deprotecting yields the desired products 47 and 48.

(c) Spermine Analogs Containing an Unsaturation:

Cis and trans unsaturated analogs of spermine were prepared via the reactions illustrated in Schemes 10, 10A, 11, and 1 IA.

Referring to Scheme 10, the trans diester 49 was reduced to the dialcohol 50, which was then converted into the trans diamine 51. Referring to Scheme 11, the cis diamine 52 was obtained from the commercially available cis dialcohol 43'. With reference to both Scheme 10 and Scheme 11, compounds 51 and 52 were protected by mesitylation to give 53 and 54, respectively. Compounds 53 and 54 were alkylated to 55 and 56, and lastly deptrotected to yield the trans tetramine 57 (Scheme 10) and the cis tetramine 58 (Scheme 11).

Alternative (and preferred) routes to 57 and 58 are given in Schemes 1OA and HA, respectively. The cis and trans dialcohols 50' and 50 were obtained in the same fashion as in Schemes 10 and 11. Compounds 50' and 50 were then mesitylated to yield 51' and 52', respectively. Reaction of 51' and 52' with 3 yields 55 (Scheme 10A) and 56 (Scheme 1 IA). Deprotecting yields the desired products 57 and 58.

Following the above general protocols, and using suitable and well known starting reagents, all of the compounds of Formula I, including those where A and D are independently C ₆ - or C ₆ -cycloalkyl, cycloalkenyl, or cycloaryl, can be readily obtained. An illustrative listing of compounds of Formula I are presented in Table 1.

32

The pure conjugates, as well as pharmaceutically-suitable salts thereof, are explicitly within the scope of the present invention. By the term "pharmaceutically- suitable salts" is meant any salt form of the subject conjugates which renders them more amenable to administration by a chosen route. A wide range of such salts are well known to those of skill in the pharmaceutical art. The preferred pharmaceutically-suitable salts are acid addition salts such as chlorides, bromides, iodides and the like.

4. Syntheses of Polyamine Conjugates of siRNAs:

Based on the above mentioned data, the three nucleosides shown in Fig. 13 can be prepared. Nucleoside A is a 2'-O- TOM nucleoside, nucleoside B is a 2'-O- ACE nucleoside, and nucleoside C is a 5'- thiol nucleoside. Nucleoside C is prepared from 5'- thioriboside that is acetylated in its nucleobase, converted into its 2'-0-ACE derivative and finally phosphitylated at 3' position.

Nucleoside B is attached to a polyamine chain at the 5'-O. The linker is preferably a carbamate bond (as shown in Fig. 14). This linker can be affixed either via a chloroformate on the 5'-O, or by adding an the alcohol to an alkyl isocyanate. In either case, the polyamine residue will be attached to the ribose by a carbamate bond. The polyamine chains are protected with alkali-labile protecting groups, such as FMOC, trifluoroacetate, and the like. As the oligonucleotide chain grows from the 3'-end to the 5'-end (as shown in Fig.11), the polyamine conjugated nucleoside will be attached in the last step of the synthesis. Release of the oligonucleotide from the resin under mild alkaline conditions (as shown in Fig. 10) will also cleave the protecting groups on the polyamine chain. The mild acid conditions necessary to free the 2'-0 will not affect the carbamate bond, and an oligonucleotide covalently bound to a polyamine residue will be obtained. The resulting polyamine-RNA conjugates are ferried into living cells in the same fashion as other polyamine conjugates are.

Regarding annealing, the strands of a double-strand siRNAs are constructed independently; the sense strand is constructed with the desired sequence of nucleotides, then the antisense strand is constructed with the corresponding complementary bases. The single strands are incubated together (pH 7.4, 1 min, 9O ⁰ C) to form the duplex. This pairing is known as annealing the siRNAs. In the

preferred embodiment of the present invention, the polyamine moiety is attached to the 5'-O of the sense strand (see Fig. 8).

Nucleoside C is constructed using 5'-thioribose. Polyamine derivatized residues (see Fig.15 for an exemplary list of preferred residues) are attached to short thioalkyl linkers. The N-thioethyl polyamine residues exemplified in Fig. 15 can be obtained starting with S-benzylcysteinamine, and then building up the polyamine chain by successive alkylations following established procedures. Valasinas et al. (2003) and references therein), and finally by deprotecting the thiol group using hydrogenolysis to cleave the benzyl group.

The synthesis of the disulfide-linked polyamine to the 5'-S- oligoribonucleoside is achieved by treatment of the mixture with diamide, a known thiol oxidant. The conjugated nucleoside is then attached to a sense oligoribonucleotide chain as discussed hereinabove. The sense oligoribonucleotide chain is deprotected at the 2'-0 position, the polyamine protecting groups are cleaved, and the strand is thenannealed to the complementary antisense strand. While not being limited to any particular biological mechanism or phenomenon, the rationale behind this approach is as follows: the polyamine will facilitate the transport across the plasma membrane of the siRNA duplex, and the conjugate will be freely translocated into the cytoplasm. The disulfide bond will then be reduced in the cytoplasm by thiols, thereby releasing the siRNA portion of the conjugate. The released siRNA will proceed to cause the sequence-specific mRNA degradation that it was designed to achieve based upon its pre-determined sequence.

Even if the conjugate remains intact after delivery into the cell, the two strands of the siRNA will partially dissociate at the RISC after delivery of the conjugate to the cytoplasm. This will not affect the function of the siRNA duplex, as single-strand antisense siRNAs are able to silence endogenous gene expression in cells..

The third approach to the synthesis of a polyamine conjugate of an oligoribonucleotide uses a linker that successfully mimics a peptidase and sets free amides and esters bound to it by way of an intramolecular-catalyzed cleavage. The polyamine chain and the nucleoside are linked through Kemp's triacid (Kemp and Petrakis (1981), J. Org. Chem. 46:5140). This remarkable triacid has three carboxylates in an all-axial orientation. One carboxyl is bound to an amine (as an amide), and a second carboxyl forms an ester with an alcohol. At a pH of about 6,

the molecule will first release the amine residue via the intramolecular formation of an anhydride. The alcohol will then be released by rearrangement of the intramolecular anhydride (see Fig.16). When entering the cytoplasm, the conjugate will be confronted by a variety of peptidases and lysosomal enzymes that will cleave both the amide bond and the ester bond. This cleavage is assisted by the axial geometry of the carboxylates. Relief of internal compression during anhydride formation thus contributes to the enzyme-driven acceleration of hydrolysis. The van der Waals contact distances in Kemp's acid are very short, and energetically costly desolvation processes can thus be averted.

The synthesis uses Kemp's acid chloride-anhydride, see Fig. 16. By reaction with a polyamine, a polyamide is formed. Ring opening of the anhydride with the 5'-0 alcohol of nucleoside A yields a Kemp acid substituted with an amide and an ester. The Kemp acid thus substituted with an amide and an ester of nucleoside A is then coupled to the growing edge of a sense ribonucleotide (e.g., see Fig. 11) through the 3'-0-CEPA group. The TOM protecting group, the N-acetyl, and the protecting groups on the polyamine are cleaved in mild alkali. The ester amide ribonucleotide is then annealed with the antisense strand and transported into the cytoplasm.

It is expected that the polyamide linkage will be cleaved first by cytoplasmic peptidases that will free the oligoribonucleotide from the polyamine chain. Cleavage of the ester group by the cell's esterases will follow with release of the siRNA duplex. Even if the latter hydrolysis is slow ( hours), it should be pointed out that siRNAs linger in the cells for hours without degradation. Thus there is enough time to achieve hydrolysis of both amide and ester linkages. The disubstituted Kemp acid is stable in plasma at pH 7.4. Nucleoside A is preferred in this synthetic sequence and not nucleoside B because of the sensitivity of the orthoester protecting group at 2'-0 in nucleoside B to low pH conditions (and also to avoid cleavage at the substituted Kemp acid during deprotection of the conjugated oligonucleotide).

The fourth approach to conjugate a polyamine with an oligoribonucleotide will be based on the construction of a connector linkage that collapses after a sequence of hydrolytic steps; the first involving an enzymatic cleavage and the second involving a solvolysis that proceeds spontaneously after the first step occurs. The connector linkage is constructed as shown in Fig. 17. A polyamine unit (protected with acid stable groups; e.g., trifluoroacetate) will be bound to the α-

amino group of lysine (PGl could be ε- FMOC) through a carbamate bond; the lysine is then converted into its corresponding amide by treatment with p- aminobenzyl alcohol. Addition of the benzyl alcohol to p-nitrophenyl isocyanate results in the p-nitroanilide Ia. Cleavage of the protecting group at the α-amino residue gives Ib.

Compound Ib will undergo rapid hydrolysis in the presence of trypsin with release of p-nitroaniline. In the present invention, the p-nitroaniline residue is replaced with nucleoside A via a simple displacement reaction (see Fig. 15). After deprotection of PGl at the α-amino group of the lysine residue to give 2a, the amide bond at the lysine residue can be hydrolyzed by trypsin and/or the lysosomal proteases cathepsins B and L, thereby releasing benzyl carbamate 3 (see Fig. 17), that will undergo spontaneous solvolysis to nucleoside A and p-aminobenzyl alcohol.

In short, compound 2a is used as the last step of an oligonucleotide construction to obtain a sense ribonucleotide strand bound through a collapsible linker to the polyamine chain. Cleavage of the TOM protecting groups, as well as of the polyamine protecting groups and PGl at the α-lysine, followed by annealing with the complementary RNA strand, affords a conjugated siRNA complex that is delivered into living cell and then collapses into its component moieties after hydrolysis.

5. Administration and Pharmaceutical Unit Dosage Forms:

Because the above-described conjugates are effective to moblize RNAs into living cells, the conjugates are suitable for therapeutically treating of mammals in vivo, including humans, and for treating mammalian cells in vitro, in any treatment regimen requiring the mobilization of RNA into mammalian cells. In short, the conjugates are useful for moving RNA, including siRNA from an extracellular space into the cytoplasm of a mamallian cell.

Administration of the subject complexes to a human or non-human patient can be accomplished by any means known in the pharmaceutical arts. The preferred administration route is parenteral, including intravenous administration, intraarterial administration, intratumor administration, intramuscular administration, intraperitoneal administration, and subcutaneous administration, either neat or in combination with a pharmaceutical carrier suitable for the chosen administration

route. The treatment method is also amenable to oral administration.

It must be noted, as with all pharmaceuticals, the concentration or amount of the polyamine-RNA conjugate administered will vary depending upon the severity of the ailment being treated, the mode of administration, the condition and age of the subject being treated, and the particular polyamine-RNA conjugate or combination of conjugates being used.

The conjugates described herein are administratable in the form of tablets, pills, powder mixtures, capsules, iηjectables, solutions, suppositories, emulsions, dispersions, food premixes, and in other suitable forms. The pharmaceutical dosage form which contains the conjugates described herein is conveniently admixed with a non-toxic pharmaceutical organic carrier or a non-toxic pharmaceutical inorganic carrier. Typical pharmaceutically-acceptable carriers include, for example, mannitol, urea, dextrans, lactose, potato and maize starches, magnesium stearate, talc, vegetable oils, polyalkylene glycols, ethyl cellulose, poly(vinylpyrrolidone), calcium carbonate, ethyl oleate, isopropyl myristate, benzyl benzoate, sodium carbonate, gelatin, potassium carbonate, silicic acid, and other conventionally employed acceptable carriers. The pharmaceutical dosage form may also contain non-toxic auxiliary substances such as emulsifying, preserving, or wetting agents, and the like.

Solid forms, such as tablets, capsules and powders, can be fabricated using conventional tabletting and capsule-filling machinery, which is well known in the art. Solid dosage forms may contain any number of additional non-active ingredients known to the art, including excipients, lubricants, dessicants, binders, colorants, disintegrating agents, dry-flow modifiers, preservatives, and the like.

Liquid forms for ingestion can be formulated using known liquid carriers, including aqueous and non-aqueous carriers, suspensions, oil-in-water and/or water- in-oil emulsions, and the like. Liquid formulation may also contain any number of additional non-active ingredients, including colorants, fragrance, flavorings, viscosity modifiers, preservatives, stabilizers, and the like.

For parenteral administration, the subject conjugates may be administered as injectable dosages of a solution or suspension of the conjugate in a physiologically- acceptable diluent or sterile liquid carrier such as water or oil, with or without additional surfactants or adjuvants. An illustrative list of carrier oils would include animal and vegetable oils (peanut oil, soy bean oil), petroleum-derived oils (mineral

oil), and synthetic oils. In general, for injectable unit doses, water, saline, aqueous dextrose and related sugar solutions, and ethanol and glycol solutions such as propylene glycol or polyethylene glycol are preferred liquid carriers.

The pharmaceutical unit dosage chosen is preferably fabricated and administered to provide a concentration of conjugate at the point of contact with the target cell of from, for example, about 1 μM to about 10 mM. More preferred is a concentration of from about 1 μMto about 100 μM. This concentration will, of course, depend on the chosen route of administration and the mass of the subject being treated. Concentrations above and below the above-stated ranges are within the scope of the invention.