HK1191566B - Peptide-based in vivo sirna delivery system - Google Patents

Description

Peptide-based in vivo siRNA delivery system

Background

Delivery of polynucleotides and other substantially cell membrane impermeable compounds into living cells is highly limited by the complex membrane system of the cell. Drugs used for antisense, RNAi and gene therapy are relatively large hydrophilic polymers and often have a high negative charge. Both of these physical characteristics severely limit the direct diffusion of these drugs across cell membranes. Thus, the major obstacle to polynucleotide delivery is the delivery of polynucleotides across cell membranes to the cytoplasm or nucleus.

One means that has been used to deliver small nucleic acids in vivo is to attach the nucleic acid to a small targeting molecule or a lipid or sterol. Although some delivery and activity has been observed with these conjugates, the extremely large nucleic acid doses required to utilize these methods are impractical.

Numerous transfection reagents have also been developed that achieve moderately effective delivery of polynucleotides into cells in vitro. However, the use of these same transfection reagents to deliver polynucleotides in vivo is complex and ineffective due to toxicity in vivo, adverse serum interactions, or poor targeting. Transfection reagents, cationic polymers and lipids that work well in vitro often form larger cationic electrostatic particles and destabilize cell membranes. The positive charge of the in vitro transfection reagent facilitates association with the nucleic acid via charge-charge (electrostatic) interactions, thereby forming a nucleic acid/transfection reagent complex. The positive charge is also beneficial for non-specific binding of the carrier to the cell and for membrane fusion, destabilization or disruption. Destabilization of the membrane facilitates delivery of the substantially cell membrane impermeable polynucleotide across the cell membrane. While these properties aid in nucleic acid transfer in vitro, they can cause toxicity and ineffective targeting in vivo. Cationic charge results in interactions with serum components, causing destabilization of polynucleotide-transfection reagent interactions, poor bioavailability, and poor targeting. The membrane activity of transfection reagents that are effective in vitro often results in toxicity in vivo.

For in vivo delivery, the vector (nucleic acid and associated delivery agent) should be small, less than 100nm in diameter and preferably less than 50 nm. Smaller complexes of less than 20nm or less than 10nm would be more useful. Delivery vehicles larger than 100nm rarely reach cells other than vascular cells in vivo. Complexes formed by electrostatic interactions tend to aggregate or separate when exposed to physiological salt concentrations or serum components. Furthermore, the cationic charge on the in vivo delivery vehicle leads to adverse serum interactions and thus poor bioavailability. Interestingly, the high negative charge can also inhibit targeted in vivo delivery by interfering with the interaction necessary for targeting, i.e., binding of the targeting ligand to a cellular receptor. Thus, near neutral vectors are required for in vivo distribution and targeting. Without careful regulation, membrane disruption or destabilizing activity is toxic when used in vivo. It is easier to balance vector toxicity with nucleic acid delivery in vitro than in vivo.

One means of reversibly modulating the membrane-disrupting activity of membrane-active polyamines is described by Rozema et al in U.S. patent publication 20040162260. Membrane active polyamines provide a means of disrupting cell membranes. The pH-dependent reversible regulation provides a means to limit the activity of entry into the endosome of the target cell, thereby limiting toxicity. The method relies on amine modification of polyamines with 2-propionic acid-3-methyl maleic anhydride.

This modification converts the polycation to a polyanion by converting the primary amine to paired carboxyl groups (β carboxyl and γ carboxyl groups) and reversibly inhibits the membrane activity of the polyamine. Rozema et al (Bioconjugate chem.2003,14,51-57) reported that the beta carboxyl group did not exhibit a complete apparent negative charge and could not inhibit membrane activity alone. It is reported that the addition of gamma carboxyl group is necessary for effective inhibition of membrane activity. To enable co-delivery of the nucleic acid with the delivery vehicle, the nucleic acid is covalently linked to a delivery polymer. Using its biologically labile conjugate delivery system, it can be shown to deliver polynucleotides into cells in vitro. However, because the carrier is highly negatively charged with high negative charge density in the case of both nucleic acids and modified polymers, this system is not efficient for in vivo delivery. The negative charge may inhibit cell-specific targeting and enhance non-specific uptake by the reticuloendothelial system (RES).

Rozema et al in U.S. patent publication 20080152661 improved the process of U.S. patent publication 20040162260 by eliminating the high negative charge density of the modified membrane active polymer. By substituting the gamma carboxyl group of 2-propionic acid-3-methylcaleic anhydride with neutral hydrophilic targeting (galactose) and steric stabilization (PEG) groups, Rozema et al are able to retain overall water solubility and reversible inhibition of membrane activity while incorporating effective in vivo hepatocyte targeting. As before, the polynucleotide is covalently linked to the transfection polymer. Covalent attachment of the polynucleotide to the transfection polymer is maintained by preventing dissociation of the polynucleotide from the transfection polymer to ensure co-delivery of the polynucleotide and transfection polymer to the target cells during in vivo administration. Co-delivery of the polynucleotide with the transfection polymer is desirable because the transfection polymer provides trans-cellular membrane transport of the polynucleotide from outside the cell or from inside the endocytic compartment to the cytoplasm. U.S. patent publication 20080152661 demonstrates that polynucleotides, particularly RNAi oligonucleotides, can be efficiently delivered to liver cells in vivo using the novel improved physiologically responsive polyconjugates.

However, covalent attachment of nucleic acids to polyamines has inherent limitations. Modifying the transfection polymer to link both the nucleic acid and the masking agent is complicated by charge interactions. The attachment of negatively charged nucleic acids to positively charged polymers tends to aggregate, thereby limiting the concentration of the mixture. Aggregation can be overcome by the presence of excess polycation or polyanion. However, this solution limits the nucleic acid to polymer ratio that can be formulated. Furthermore, the attachment of negatively charged nucleic acids to unmodified cationic polymers results in condensation and aggregation of the complex and inhibits polymer modification. Formation of polymer modifications with negative electropolymers weakens the attachment of nucleic acids.

The technology described in U.S. patent publication 20080152661 was further improved by Rozema et al in U.S. provisional application 61/307,490. In U.S. provisional application 61/307,490, Rozema et al demonstrated that by careful selection of targeting molecules and independent attachment of appropriate targeting molecules to both siRNA and delivery polymers, the siRNA and delivery polymers can be uncoupled but still retain both components for effective targeting of cells in vivo and achieve highly functional targeted delivery of the siRNA. The delivery polymers used in both U.S. patent publication 20080152661 and U.S. provisional application 61/307,490 are relatively large synthetic polymers poly (vinyl ethers) and poly (acrylates). Larger polymers can be modified with both targeting ligands for cell specific binding and PEG for increased shielding. Larger polymers are necessary for efficient delivery, possibly through increased membrane activity and improved protection of nucleic acids in vivo within cells. Larger polycations interact more strongly with the membrane and with anionic RNA.

We have developed here an improved siRNA delivery system using a much smaller delivery peptide. The improved system provides effective siRNA delivery with reduced toxicity and thus a wider therapeutic window.

Disclosure of Invention

In a preferred embodiment, the invention features a composition for delivering an RNA interference polynucleotide into a liver cell in vivo, comprising: a) asialoglycoprotein receptor (ASGPr) targeted reversibly masked melittin peptides (delivery peptides) and b) RNA interference polynucleotides conjugated with hydrophobic groups containing at least 20 carbon atoms (RNA conjugates). The delivery peptide and siRNA conjugate are synthesized separately and may be supplied in separate containers or in a single container. The RNA interference polynucleotide is not conjugated to a delivery peptide.

In another preferred embodiment, the invention features a composition for delivering an RNA interference polynucleotide into a liver cell in vivo, comprising: a) ASGPr-targeted reversibly masked melittin peptides (delivery peptides) and b) RNA interference polynucleotides conjugated with galactose clusters (RNA conjugates). The delivery peptide and siRNA conjugate are synthesized separately and may be supplied in separate containers or in a single container. The RNA interference polynucleotide is not conjugated to a polymer.

In a preferred embodiment, the ASGPr-targeted reversibly masked melittin peptide comprises a melittin peptide reversibly modified by reaction with a masking agent containing an ASGPr ligand via a primary amine on the peptide. An amine is reversibly modified if cleavage of the modifying group restores the amine. Reversible modification of melittin peptides by the masking agents described herein reversibly inhibits the membrane activity of melittin peptides. In the masked state, reversibly masked melittin peptides do not exhibit membrane disruptive activity. It is desirable to reversibly modify more than 80% or more than 90% of the amines on melittin peptides to inhibit membrane activity and provide cell targeting functions, i.e. the formation of reversibly masked melittin peptides.

A preferred masking agent containing an ASGPr ligand has a neutral charge and comprises a galactosamine or galactosamine derivative having a disubstituted maleic anhydride amine-reactive group. Another preferred masking agent containing an ASGPr ligand comprises a galactosamine or galactosamine derivative having a peptidase-cleavable dipeptide-p-amidobenzylamine reactive carbonate derivative. Reaction of the amine-reactive carbonate with the amine reversibly modifies the amine to form an amidobenzyl carbamate linkage.

In a preferred embodiment, the bee toxin peptide comprises a Apis mellifera (Apis florea) (Apis cerana or Apis cerana), Apis mellifera (Apis mellifera) (western or european or Apis mellifera), Apis mellifera (Apis dorsiata), Apis cerana (Apis cerana) (eastern honey bee) or a derivative thereof. A preferred melittin peptide comprises the sequence: xaa₁-Xaa₂-Xaa₃-Ala-Xaa₅-Leu-Xaa₇-Val-Leu-Xaa₁₀-Xaa₁₁-Xaa₁₂-Leu-Pro-Xaa₁₅-Leu-Xaa₁₇-Xaa₁₈-Trp-Xaa₂₀-Xaa₂₁-Xaa₂₂-Xaa₂₃-Xaa₂₄-Xaa₂₅-Xaa₂₆Wherein:

Xaa₁is leucine, D-leucine, isoleucine, norleucine, tyrosine, tryptophan, valine, alanine, dimethylglycine, glycine, histidine, phenylalanine or cysteine,

Xaa₂is isoleucine, leucine, norleucine or valine,

Xaa₃is glycine, leucine or valine,

Xaa₅is isoleucine, leucine, norleucine or valine,

Xaa₇is lysine, serine, asparagine, alanine, arginine or histidine,

Xaa₁₀is alanine, threonine or leucine,

Xaa₁₁is a compound of the amino acid sequence of threonine or cysteine,

Xaa₁₂is glycine, leucine or tryptophan,

Xaa₁₅is a compound of threonine or alanine,

Xaa₁₇is isoleucine, leucine, norleucine or valine,

Xaa₁₈is a serine or a cysteine, and the amino acid is,

Xaa₂₀is isoleucine, leucine, norleucine or valine,

Xaa₂₁is lysine or alanine, and can be used as the active ingredient,

Xaa₂₂is an asparagine or an arginine,

Xaa₂₃is lysine or alanine, and can be used as the active ingredient,

Xaa₂₄is arginine or lysine, and can be used as the active ingredient,

Xaa₂₅is lysine, alanine or glutamine,

Xaa₂₆optionally and if present, glutamine, cysteine, glutamine-NH₂Or cysteine-NH₂(ii) a And the number of the first and second electrodes,

and Xaa₂₁、Xaa₂₃And Xaa₂₅At least two of them are lysine。

A more preferred melittin comprises the sequence:

Xaa₁-Xaa₂-Xaa₃-Ala-Xaa₅-Leu-Xaa₇-Val-Leu-Xaa₁₀-Xaa₁₁-Xaa₁₂-Leu-Pro-Xaa₁₅-Leu-Xaa₁₇-Ser-Trp-Xaa₂₀-Lys-Xaa₂₂-Lys-Arg-Lys-Xaa₂₆wherein:

Xaa₁is leucine, D-leucine, norleucine or tyrosine,

Xaa₂is isoleucine, leucine, norleucine or valine,

Xaa₃is glycine, leucine or valine,

Xaa₅is isoleucine, valine, leucine or norleucine,

Xaa₇is lysine, serine, asparagine, alanine, arginine or histidine,

Xaa₁₀is alanine, threonine or leucine,

Xaa₁₁is a compound of the amino acid sequence of threonine or cysteine,

Xaa₁₂is glycine, leucine or tryptophan,

Xaa₁₅is a compound of threonine or alanine,

Xaa₁₇is isoleucine, leucine or norleucine,

Xaa₂₀is isoleucine, leucine or norleucine,

Xaa₂₂is asparagine or arginine, and

Xaa₂₆is glutamine or cysteine.

A most preferred melittin comprises the sequence:

Xaa₁-Xaa₂-Gly-Ala-Xaa₅-Leu-Lys-Val-Leu-Ala-Xaa₁₁-Gly-Leu-Pro-Thr-Leu-Xaa₁₇-Ser-Trp-Xaa₂₀-Lys-Xaa₂₂-Lys-Arg-Lys-Xaa₂₆wherein:

Xaa₁、Xaa₂、Xaa₅、Xaa₁₇and Xaa₂₀Independently isoleucine, leucine or norleucine,

Xaa₁₁is a compound of the amino acid sequence of threonine or cysteine,

Xaa₂₂is asparagine or arginine, and

Xaa₂₆is glutamine or cysteine.

One preferred masking agent comprises a neutral hydrophilic disubstituted alkyl maleic anhydride:

wherein R1 comprises a cell targeting group. One preferred alkyl group is methyl or ethyl. One preferred targeting group comprises an asialoglycoprotein receptor ligand. One example of a substituted alkyl maleic anhydride consists of a 2-propionic acid-3-alkyl maleic anhydride derivative. A neutral hydrophilic 2-propionic acid-3-alkylcarboxylic acid anhydride derivative is formed by linking a neutral hydrophilic group to 2-propionic acid-3-alkylcarboxylic acid anhydride via a 2-propionic acid-3-alkylcarboxylic acid anhydride γ -carboxyl group:

wherein R1 comprises a neutral ASGPr ligand and n =0 or 1. In one embodiment, the ASGPr ligand is attached to the anhydride via a short PEG linker.

One preferred masking agent comprises a hydrophilic peptidase (protease) cleavable dipeptide-p-amidobenzylamine reactive carbonate derivative. The enzymatically cleavable linker of the present invention employs a dipeptide linked to an amidobenzyl-activated carbonate moiety. An ASGPr ligand is attached to the amino terminus of the dipeptide. The amidobenzyl activated carbonate moiety is at the carboxyl terminus of the dipeptide. Peptidase cleavable linkers suitable for use with the present invention have the structure:

wherein R4 comprises an ASGPr ligand and R3 comprises an amine-reactive carbonate moiety, and R1 and R2 are amino acid R groups. A preferred activated carbonate is p-nitrophenol. However, other amine reactive carbonates known in the art are readily substituted for the p-nitrophenol. The reaction of activated carbonate with melittin amine allows the targeting compound asialoglycoprotein receptor ligand to be linked to melittin peptide via a peptidase-cleavable dipeptide-amidobenzyl carbamate linkage. Enzymatic cleavage of the dipeptide removes the targeting ligand from the peptide and triggers an elimination reaction that causes regeneration of the peptide amine.

The dipeptides Glu-Gly, Ala-Cit, Phe-Cit ("Cit" is the amino acid citrulline) are shown in example 3. Neutral amino acids are preferred, although charged amino acids are also permissible.

Preferably, the masking agent provides a targeting function by affinity for a cell surface receptor, i.e., the masking agent contains a ligand for a cell surface receptor. Preferably the masking agent comprises a saccharide having affinity for ASGPr, including (but not limited to): galactose, N-acetyl-galactosamine and galactose derivatives. Galactose derivatives having affinity for ASGPr are well known in the art. An essential feature of reversibly modified melittin is that more than 80% of the melittin amines (in the peptide population) are modified with a physiologically labile reversible covalent bond linked ASGPr ligand.

In another embodiment, the melittin peptides of the invention are further modified at the amino or carboxy terminus by covalent attachment of a steric stabilizer or ASGPr ligand-steric stabilizer conjugate. Amino or carboxyl terminal modifications can be attached to the peptide during synthesis using standard methods in the art. Alternatively, amino or carboxyl terminal modifications may be made by modifying cysteine residues on melittin peptides having amino or carboxyl terminal cysteine residues. One preferred steric stabilizer is polyethylene glycol. Preferably the polyethylene glycol has from 1 to 120 ethylene units. In another embodiment, it is preferred that the polyethylene glycol is less than 5kDa in size. For ASGPr ligand-steric stabilizer conjugates, the preferred steric stabilizer is polyethylene glycol having from 1 to 24 ethylene units.

The RNAi polynucleotide conjugate and the delivery peptide are administered to a mammal in a pharmaceutically acceptable carrier or diluent. In one embodiment, the delivery peptide and RNAi polynucleotide conjugate can be mixed in solution prior to administration to a mammal. In another embodiment, the delivery peptide and the RNAi polynucleotide conjugate can be co-administered to the mammal in separate solutions. In yet another embodiment, the delivery peptide and the RNAi polynucleotide conjugate can be administered to the mammal sequentially. For sequential administration, the delivery peptide can be administered prior to administration of the RNAi polynucleotide conjugate. Alternatively, for sequential administration, the RNAi polynucleotide conjugate may be administered prior to administration of the delivery peptide.

Other objects, features and advantages of the present invention will become apparent from the following detailed description when considered in conjunction with the accompanying drawings.

Brief description of the drawings

FIG. 1 lists tables of melittin peptides suitable for use in the present invention.

FIG. 2 is a graph illustrating the linkage of GalNAc cluster to RNA.

FIG. 3 illustrates a graph of (A) Blood Urea Nitrogen (BUN) levels and (B) creatinine levels in primates treated with reversibly modified melittin siRNA delivery peptides and siRNA-cholesterol conjugates.

FIG. 4 illustrates a graph of (A) aspartate Aminotransferase (AST) levels and (B) alanine Aminotransferase (ALT) levels in primates treated with reversibly modified melittin siRNA delivery peptides and siRNA-cholesterol conjugates.

Figure 4 illustrates a graph of the knockdown of endogenous factor VII levels in primates treated with reversibly modified melittin siRNA delivery peptides and siRNA-cholesterol conjugates.

Detailed Description

Described herein is an improved method for delivering RNA interference (RNAi) polynucleotides into mammalian liver cells in vivo. We describe an in vivo RNAi polynucleotide delivery system employing the small delivery peptide melittin derived from melittin, as well as individual targeting RNAi polynucleotides. Efficient delivery of RNAi polynucleotides to the liver was observed by using RNAi polynucleotide conjugate molecules targeted to the liver and reversible inhibitory melittin peptides targeted to the asialoglycoprotein receptor.

Because melittin and RNAi polynucleotides are independently targeted to hepatocytes, the concentrations of melittin and polynucleotides and the ratio therebetween are limited only by the solubility of the components, and not by the solubility of the associated complexes or the ability to make complexes. Furthermore, the polynucleotide and melittin may be mixed at any time prior to administration, or even administered separately, thus allowing the individual components to be stored separately in solution or dry form.

The present invention includes a conjugate delivery system having the following composition:

y-bee toxin- (L-M)_xAdding N-T into the mixture, adding N-T,

wherein N is an RNAi polynucleotide, T is a polynucleotide targeting moiety (a hydrophobic group having 20 or more than 20 carbon atoms or a galactose cluster), melittin is a melittin peptide or derivative derived from melittin as described herein, and masking agent M contains a generic peptideThe physiologically labile reversible linkage L is covalently linked to the ASGPr ligand as described herein of melittin. Cleavage of L restores the unmodified amines on melittin. Y is an optional polyethylene glycol (PEG) or ASGPr ligand-PEG conjugate, and if present, comprises a cysteine attached to the amino terminus, the carboxyl terminus, or both the amino and carboxyl termini of melittin. Preferably Y is attached to the amino terminus or the amino-terminal cysteine. x is an integer greater than 1. In the unmodified state, melittin has membrane activity. However, the delivery peptide melittin- (L-M)_xHas no membrane activity. Reversible modification of the primary amine of melittin by linking M reversibly inhibits or inactivates the membrane activity of melittin. A sufficient percentage of primary amines of melittin are modified to inhibit membrane activity of the polymer and provide hepatic cell targeting. Preferably, the value of x is greater than 80% and more preferably greater than 90% of the primary amine on melittin, as determined by the amount of amine on melittin in the absence of any masking agent. More specifically, the value of x is greater than 80% and up to 100% of the primary amines on melittin. It should be noted that melittin typically contains 3 to 5 primary amines (amino-terminal (if unmodified) and typically 2 to 4 lysine residues). Thus, a percentage of amine modifications are intended to reflect a percentage of modifications on amines in the melittin peptide population. After cleavage of the reversible linkage L, the unmodified amine is restored, thereby restoring melittin to its unmodified membrane active state. The reversible linkage is preferably a pH labile linkage. Another preferred reversible linkage is a protease cleavable linkage. Reversible masked membrane active polymer melittin- (L-M) targeting ASGPR_x(delivery peptides) and the polynucleotide conjugates T-N are synthesized or manufactured separately. Neither T nor N is directly or indirectly covalently linked to melittin, L or M. In vivo hepatic delivery of polynucleotides does not require electrostatic or hydrophobic association of the polynucleotide or polynucleotide conjugate with a masked or unmasked polymer. The masked polymer and the polynucleotide conjugate may be supplied in the same container or in separate containers. They may be administered in combination, co-administration or sequentially prior to administration.

Hydrophilic groups indicate in qualitative terms that the chemical moiety is water-loving. Typically, these chemical groups are soluble in water and are hydrogen bond donors or acceptors in the presence of water. The hydrophilic group may be charged or uncharged. The charged groups can be positively charged (anionic) or negatively charged (cationic) or both (zwitterionic). Examples of hydrophilic groups include the following chemical moieties: carbohydrates, polyoxyethylene, certain peptides, oligonucleotides, amines, amides, alkoxyamides, carboxylic acids, sulfur and hydroxyl groups.

Hydrophobic groups indicate in qualitative terms that the chemical moiety is water-repellent. Typically, these chemical groups are not soluble in water and do not tend to form hydrogen bonds. Lipophilic groups are soluble in fats, oils, lipids and non-polar solvents and are rarely or even not capable of forming hydrogen bonds. Hydrocarbons containing two (2) or more carbon atoms, certain substituted hydrocarbons, cholesterol, and cholesterol derivatives are examples of hydrophobic groups and compounds.

The hydrophobic group is preferably a hydrocarbon containing only carbon and hydrogen atoms. However, non-polar substituted or non-polar heteroatoms that maintain hydrophobicity and include, for example, fluorine may be permitted. The term includes aliphatic groups, aromatic groups, acyl groups, alkyl groups, alkenyl groups, alkynyl groups, aryl groups, aralkyl groups, aralkenyl groups, and aralkynyl groups, each of which can be straight, branched, or cyclic. The term hydrophobic group also includes: sterols, steroids, cholesterol, and steroids and cholesterol derivatives.

As used herein, a membrane-active peptide is a surface-active, amphiphilic peptide capable of inducing one or more of the following effects on biological membranes: altering or disrupting the membrane to allow non-membrane permeable molecules to enter the cell or cross the membrane, forming pores in the membrane, disrupting the membrane, or otherwise disrupting or dissolving the membrane. As used herein, a membrane or cell membrane comprises a lipid bilayer. The alteration or disruption of the membrane can be functionally determined by the activity of the peptide in at least one of the following assays: haemolysis (haemolysis), liposome leakage, liposome fusion, cell lysis and endosomal release. Membrane active peptides that cause lysis of cell membranes are also known as membrane lytic peptides. Peptides that preferentially cause endosomal or lysosomal destruction compared to the plasma membrane are considered to have endosomal solubility. The effect of membrane active peptides on the cell membrane may be transient. Membrane-active peptides have an affinity for the membrane and denature or deform the bilayer structure.

Delivery of polynucleotides to cells is mediated by melittin peptides that disrupt or destabilize the plasma membrane or internal vesicle membranes (such as endosomes or lysosomes), including forming pores in the membrane or disrupting endosomes or lysosomal vesicles, thereby allowing release of the contents of the vesicles into the cytoplasm.

Endosomolytic peptides are peptides that are released from a closed vesicle (e.g., an endosome or lysosome) within a cell's inner membrane by a compound (e.g., a polynucleotide or protein) that is capable of causing endosomal disruption or lysis or providing a generally impermeable cell membrane in response to an endosomal specific environmental factor (e.g., a decrease in pH or the presence of a lytic enzyme (protease)). The endosomolytic polymer undergoes a change in its physicochemical properties within the endosome. This change can be a change in the solubility of the polymer or the ability to interact with other compounds or membranes due to a change in charge, hydrophobicity, or hydrophilicity. Exemplary endosomolytic peptides have a pH labile or enzyme sensitive group or bond. Thus, a reversibly masked membrane active peptide with a masking agent attached to the polymer through a pH labile linkage can be considered an endosomolytic polymer.

Melittin as used herein is a small amphiphilic membrane-active peptide containing from about 23 to about 32 amino acids derived from the naturally occurring melittin. Naturally occurring melittin contains 26 amino acids and is predominantly hydrophobic at the amino terminus and hydrophilic (cationic) at the carboxyl terminus. The melittin of the present invention may be isolated from biological sources or may be synthetic. Synthetic polymers are formulated or manufactured by "artificial" chemical processes and are not produced by naturally occurring biological processes. As used herein, melittin encompasses the naturally occurring melittin peptides of the melittin family, which can be found, for example, in the venom of the following species: small bees, Italian bees, Chinese bees, big bees, frontal spotted wasps (Vespula maculi), big wasps (Vespa magnica), black-chest wasps (Vespa velutina), hornet (Polistes sp.) HQL-2001 and Asian non-hornet (Polistes hebraeus). As used herein, melittin also encompasses synthetic peptides having an amino acid sequence identical or similar to the naturally occurring melittin peptide. In particular, the melittin amino acid sequence encompasses that shown in figure 1. In addition to amino acids that retain the high membrane activity inherent to melittin, 1 to 8 amino acids may be added to the amino or carboxyl terminus of the peptide. Specifically, cysteine residues may be added to the amino or carboxyl terminus. The list in fig. 1 is not intended to be exhaustive, as other conservative amino acid substitutions are readily envisaged. Synthetic melittin peptides may contain naturally occurring L-amino acids or enantiomeric D-amino acids (inversions). However, melittin peptides should contain essentially all L-or all D-amino acids, but amino acids with an inverted stereocenter may be attached at the amino-or carboxy-terminus. The melittin amino acid sequence may also be inverted (retro). The reverse melittin may have either L-type amino acids or D-type amino acids (reverse inversion). Two melittin peptides may also be covalently linked to form a melittin dimer. Melittin may have, in addition to a masking agent, a modifying group attached to the amino terminus or carboxy terminus of the peptide that enhances tissue targeting or promotes circulation in the body. However, as used herein, melittin does not include chains or polymers containing more than two melittin peptides covalently linked to each other or covalently linked to another polymer or backbone.

Masking

Melittin peptides of the present invention comprise reversibly modified melittin peptides, wherein the reversible modification inhibits membrane activity, neutralizes melittin to reduce positive charges and form polymers with near neutral charges, and provides cell-type specific targeting. The melittin is reversibly modified by reversibly modifying a primary amine on the peptide.

Melittin peptides of the invention are capable of disrupting the plasma membrane or the lysosomal/endocytic membrane. However, when peptides are administered in vivo, membrane activity can cause toxicity. Therefore, reversible masking of the membrane activity of melittin is essential for in vivo use. This masking is achieved by reversibly linking a masking agent to melittin to form reversibly masked melittin, i.e. a delivery peptide. In addition to inhibiting membrane activity, masking agents also provide cell-specific interactions, i.e., targeting.

An essential feature of a masking agent is that it inhibits the membrane activity of the polymer and provides hepatic cell targeting in vivo in the aggregated state. Melittin has membrane activity in the unmodified (unmasked) state and does not have membrane activity (inactivated) in the modified (masked) state. A sufficient number of masking agents are attached to the peptide to achieve the desired degree of inactivation. The degree of melittin modification required by attachment of the masking agent is readily determined using appropriate peptide activity assays. For example, if melittin has membrane activity in a given assay, a sufficient amount of a masking agent is attached to the peptide to achieve the desired degree of inhibition of membrane activity in that assay. Preferably, greater than or equal to 80% or greater than or equal to 90% of the primary amine groups on the population of melittin peptides are modified, as determined by the amount of primary amines on the peptides in the absence of any masking agent. The masking agent is also preferably characterized in that its attachment to the peptide reduces the positive charge of the polymer, thereby forming a more neutral delivery peptide. It is desirable that the masking peptide remain water soluble.

As used herein, the peptide is masked if the modified melittin does not exhibit membrane activity and exhibits in vivo cell-specific (i.e., hepatocyte) targeting. Melittin is reversibly masked if the bond linking the masking agent to the peptide is cleaved such that the amine on the peptide is restored, thereby restoring membrane activity.

Another essential feature is that the masking agent is covalently bound to melittin via a physiologically labile reversible bond. By using a physiologically labile reversible linkage, the masking agent can be cleaved from the peptide in vivo, thereby unmasking the peptide and restoring the activity of the unmasked peptide. By choosing the appropriate reversible linkage, it is possible to form a conjugate that restores melittin activity after having been delivered to or targeted to the desired cell type or cell location. The reversibility of the linkage provides for selective activation of melittin. The reversible covalent linkage contains a reversible or labile bond that may be selected from the group consisting of: physiologically labile bonds, cell physiologically labile bonds, pH labile bonds, very pH labile bonds, extremely pH labile bonds, and protease cleavable bonds.

As used herein, a masking agent comprises a preferably neutral (uncharged) compound having an ASGPr ligand and an amine-reactive group, wherein reaction of the amine-reactive group with an amine on a peptide causes the ASGPr ligand to be bound to the peptide by a physiologically labile reversible covalent bond. The amine reactive group is selected such that cleavage in response to appropriate physiological conditions (e.g., a decrease in pH, such as in endosomes/lysosomes, or enzymatic cleavage, such as in endosomes/lysosomes) causes regeneration of the melittin amine. ASGPr ligands are groups, usually sugars, with affinity for asialoglycoprotein receptors. Preferred masking agents of the present invention are capable of modifying (forming reversible bonds with) a polymer in aqueous solution.

Preferably, the amine reactive group comprises disubstituted maleic anhydride. One preferred masking agent is represented by the following structure:

wherein R1 comprises an asialoglycoprotein receptor (ASGPR) ligand and R2 is alkyl, such as methyl (-CH)₃) Ethyl (-CH)₂CH₃) Or propyl (-CH)₂CH₂CH₃)。

In some embodiments, the galactose ligand is linked to the amine reactive group through a PEG linker, as illustrated by the following structure:

wherein n is an integer between 1 and 19.

Another preferred amine-reactive group comprises a dipeptide-amidobenzylamine-reactive carbonate derivative represented by the following structure:

wherein:

r1 is the R group of amino acid 1,

r2 is the R group of amino acid 2,

r3 is-CH₂-O-C (O) -O-Z, wherein Z is

A halide compound, a halogen atom,

and R4 comprises an ASGPr ligand.

The ASGPr ligand is linked to melittin peptide via a peptidase-cleavable dipeptide-amidobenzyl carbamate linkage by reaction of activated carbonate with melittin amine.

Enzymatic cleavage of the dipeptide removes the targeting ligand from the peptide and triggers an elimination reaction that causes regeneration of the peptide amine. Although the above structure shows a single masking agent attached to melittin peptide, in practice several masking agents are attached to melittin peptide; preferably more than 80% of the amines on the melittin peptide population are modified.

Example 3 shows the dipeptides Glu-Gly, Ala-Cit, Phe-Cit ("Cit" is the amino acid citrulline). For the aforementioned structures, Glu-Gly, Ala-Cit, Phe-Cit represent R2-R1. Neutral amino acids are preferred, although charged amino acids are permissible. Other amino acid combinations are possible as long as they are cleaved by endogenous proteases. In addition, 3 to 5 amino acids may be used as linkers between the amidobenzyl and the targeting ligand.

As with maleic anhydride-based masking agents, the ASGPr ligand may be attached to the peptidase-cleavable dipeptide-amidobenzyl carbonate via a PEG linker.

The film-active polyamine may be conjugated to the masking agent in the presence of an excess of masking agent. Excess masking agent may be removed from the conjugated delivery peptide prior to administration of the delivery peptide.

In another embodiment, the melittin peptides of the invention are further modified at the amino or carboxy terminus by covalent attachment of a steric stabilizer or ASGPr ligand-steric stabilizer conjugate. Preferably the hydrophobic end is modified; i.e. the amino-terminal end of melittin with the "normal sequence" and the carboxy-terminal end of inverted melittin. One preferred steric stabilizer is polyethylene glycol. Amino or carboxyl terminal modifications can be attached to the peptide during synthesis using standard methods in the art. Alternatively, amino or carboxyl terminal modifications may be made by modifying cysteine residues on melittin peptides having amino or carboxyl terminal cysteine residues. Preferably the polyethylene glycol has from 1 to 120 ethylene units. In another embodiment, it is preferred that the polyethylene glycol is less than 5kDa in size. For ASGPr ligand-steric stabilizer conjugates (NAG-PEG modification), the preferred steric stabilizer is polyethylene glycol having 1 to 24 ethylene units. When the terminal PEG modification is combined with reversible masking, it further reduces the toxicity of melittin delivery peptides. Terminal NAG-PEG modification will enhance efficacy.

Steric stabilizer

As used herein, a steric stabilizer is a non-ionic hydrophilic polymer (natural, synthetic or non-natural) that prevents or inhibits intramolecular or intermolecular interactions of the molecules to which it is attached relative to molecules that do not contain the steric stabilizer. Steric stabilizers prevent the molecules to which they are attached from participating in electrostatic interactions. Electrostatic interactions are non-covalent associations of two or more substances due to the attractive forces between positive and negative charges. Steric stabilizers can inhibit interaction with blood components and thus inhibit opsonization, phagocytosis, and uptake by the reticuloendothelial system. Thus, steric stabilizers can increase the cycle time of the molecule to which they are attached. Steric stabilizers may also inhibit molecular aggregation. One preferred steric stabilizer is polyethylene glycol (PEG) or a PEG derivative. PEG molecules suitable for the present invention have about 1 to 120 ethylene glycol monomers.

ASGPR ligands

The targeting moiety or group enhances the pharmacokinetic or biodistribution properties of the conjugate to which it is attached, thereby improving the cell-specific distribution and cell-specific uptake of the conjugate. Galactose and galactose derivatives have been used to target molecules to hepatocytes in vivo through their binding to asialoglycoprotein receptors (ASGPr) expressed on the surface of hepatocytes. As used herein, an ASGPr ligand (or ASGPr ligand) comprises galactose and a galactose derivative having an affinity for ASGPr equal to or greater than the affinity of galactose for ASGPr. The binding of the galactose targeting moiety to ASGPr helps the delivery peptide to target the liver cells in a cell-specific manner and engulf the delivery peptide cells into the liver cells.

The ASGPr ligand may be selected from the group comprising: lactose, galactose, N-acetyl galactosamine (GalNAc), galactosamine, N-formyl galactosamine, N-acetyl-galactosamine, N-propionyl galactosamine, N-butyryl galactosamine, and N-isobutyryl-galactosamine (Iobst, s.t. and Drickamer, k.j.b.c.1996,271, 6686). The ASGPr ligand can be monomeric (e.g., having a single galactosamine) or multimeric (e.g., having multiple galactosamines).

In one embodiment, melittin peptides are reversibly masked by linking ASGPR ligand masking agents to > 80% or > 90% of primary amines on the peptide.

Labile linkages

A bond or linker is a connection between two atoms that connects one chemical group or segment of interest to another chemical group or segment of interest through one or more covalent bonds. For example, the linkage may attach the masking agent to the peptide. The formation of a linkage may join two separate molecules to form a single molecule, or it may join two atoms in the same molecule. The linkage may be electrically neutral or may carry a positive or negative charge. The reversible or labile linkage contains a reversible or labile bond. The linkage may optionally include a spacer that increases the distance between the two joining atoms. The spacers may further add flexibility and/or length to the linkage. Spacers may include, but are not limited to, alkyl, alkenyl, alkynyl, aryl, aralkyl, aralkenyl, aralkynyl; each of which may contain one or more heteroatoms, heterocycles, amino acids, nucleotides and sugars. Spacer groups are well known in the art and the foregoing list is not intended to limit the scope of the invention.

A labile bond is a covalent bond other than a covalent bond with a hydrogen atom that can be selectively broken or cleaved without breaking or cleaving other covalent bonds in the same molecule. More specifically, a labile bond is a covalent bond that is less stable (thermodynamically) or breaks more rapidly (kinetically) under the appropriate conditions as compared to other non-labile covalent bonds in the same molecule. Cleavage of an intramolecular labile bond can result in the formation of two molecules. For those skilled in the art, bond cleavage or instability is generally measured by the half-life of bond cleavage (t)_1/2) (time required for half bond cleavage). Thus, labile bonds encompass bonds that can be selectively cleaved more rapidly than other bonds of the molecule.

Suitable conditions are determined by the type of labile bond and are well known in organic chemistry. Labile bonds may be sensitive to pH, oxidizing or reducing conditions or reagents, temperature, salt concentration, the presence of enzymes (such as esterases, including nucleases and proteases), or the presence of added reagents. For example, an increase or decrease in pH is a suitable condition for pH labile bonds.

The rate at which the labile group will undergo conversion can be controlled by altering the chemical composition of the molecule containing the labile group. For example, the addition of a particular chemical moiety (e.g., an electron acceptor or donor) near the labile group can affect the particular conditions (e.g., pH) under which the chemical conversion will occur.

As used herein, a physiologically labile bond is a labile bond that is cleavable under conditions typically encountered in a mammalian body or under conditions similar to those encountered in a mammalian body. The physiologically labile linking group is selected to undergo a chemical transformation (e.g., cleavage) when present under certain physiological conditions.

As used herein, a physiologically labile bond of a cell is a labile bond that is cleavable under intracellular conditions in a mammal. Mammalian intracellular conditions include chemical conditions found in or similar to those encountered in mammalian cells, such as pH, temperature, oxidizing or reducing conditions or agents, and salt concentrations. Mammalian intracellular conditions also include the presence of enzymatic activities that are normally present in mammalian cells, such as enzymatic activities from proteolytic or hydrolytic enzymes. Physiologically labile bonds of a cell may also be cleaved in response to administration of a pharmaceutically acceptable exogenous agent. Physiologically labile bonds with a half-life of less than 45 minutes that are cleaved under appropriate conditions are considered to be very labile. Physiologically labile bonds with a half-life of less than 15 minutes that are cleaved under appropriate conditions are considered to be extremely labile.

Chemical transformation (cleavage of labile bonds) can be initiated by the addition of pharmaceutically acceptable agents to cells, or can occur spontaneously when molecules containing labile bonds reach the appropriate intracellular and/or extracellular environment. For example, pH labile bonds are cleaved when a molecule enters an acidic endosome. Thus, a pH labile bond can be considered an endosomal cleavable bond. The enzymatically cleavable linkage will be cleaved upon exposure to an enzyme, such as an enzyme present in the endosome or lysosome or cytoplasm. When the molecule enters the more reductive environment of the cytoplasm, the disulfide bonds are cleaved. Thus, disulfide bonds may be considered cytoplasmic cleavable bonds.

As used herein, a pH labile bond is a labile bond that is selectively broken under acidic conditions (pH < 7). These bonds may also be referred to as endosomal labile bonds because the pH of cellular endosomes and lysosomes is less than 7. The term pH labile includes bonds that are pH labile, very pH labile, and extremely pH labile.

However, if the anhydride is a cyclic anhydride, the reaction with the amine produces an amic acid, i.e., a molecule in which the amide and acid are in the same molecule, the presence of two reactive groups (amide and carboxylic acid) in the same molecule accelerates the reverse reaction, in particular, the product maleic acid of a primary amine with maleic anhydride and maleic anhydride derivatives is 1 × 10 faster than its acyclic analogs⁹To 1 × 10¹³The amine and anhydride were recovered back (Kirby 1980).

The reaction of an amine with an anhydride to form an amide and an acid.

Reaction of an amine with a cyclic anhydride to form an amic acid.

The cleavage of amic acid to form amines and anhydrides is pH dependent and greatly accelerated at acidic pH. This pH-dependent reactivity can be exploited to form reversible pH-labile bonds and linkers. Cis-aconitic acid has been used as such a pH sensitive linker molecule. The gamma-carboxylate is first coupled to the molecule. In a second step, an alpha or beta carboxylate ester is coupled to a second molecule to form a pH sensitive coupling of the two molecules. The half-life of this linker at pH5 for cleavage was between 8 and 24 hours.

Structure of maleic anhydride and maleic anhydride.

Aconitic acid

Maleic anhydride

The pH at which lysis occurs is controlled by adding chemical components to the labile moiety. The rate of conversion of maleic acid to amine and maleic anhydride strongly depends on the substitution of the maleic anhydride system (R2 and R3). When R2 is methyl, the conversion rate is 50 times higher than when R2 and R3 are hydrogen. This rate increase is significant when alkyl substitution (e.g., 2, 3-dimethylmaleic anhydride) is present at both R2 and R3: which is 10,000 times faster than unsubstituted maleic anhydride. Cleavage of the maleamic acid bond formed by modifying the amine with 2, 3-dimethylmaleic anhydride at pH5 restores half-lives of anhydride and amine of between 4 and 10 minutes. It is expected that if R2 and R3 are groups greater than hydrogen, the rate of conversion of the amide-acid to the amine and anhydride will be faster than if R2 and/or R3 are hydrogen.

Very unstable bond of pH: the cleavage half-life of the pH-labile bond at pH5 is less than 45 minutes. The construction of pH-labile bonds is well known in the chemical art.

Extremely unstable bond of pH: the cleavage half-life of the extremely labile bond at pH5 is less than 15 minutes. The construction of extremely pH labile bonds is well known in the chemical art.

Disubstituted cyclic anhydrides are particularly useful for attaching masking agents to melittin peptides of the invention. It provides a physiological pH labile linkage, readily modifies amines, and restores those amines after lysis at the reduced pH seen in endosomes and lysosomes in cells. Second, the α or β carboxylic acid groups produced upon reaction with the amine appear to contribute only about the expected negative charge of 1/20 to the polymer (Rozema et al Bioconjugate Chemistry 2003). Thus, modifying a peptide with disubstituted maleic anhydride effectively neutralizes the positive charge of the peptide, rather than producing a peptide with a high negative charge. For in vivo delivery, a near neutral delivery peptide is preferred.

RNAi polynucleotide conjugates

We have found that conjugation of RNAi polynucleotides to polynucleotide targeting moieties (hydrophobic groups or galactose clusters) and co-administration of RNAi polynucleotide conjugates with the above-described delivery peptides provides efficient, functional delivery of RNAi polynucleotides to liver cells, particularly hepatocytes, in vivo. Functional delivery means delivery of the RNAi polynucleotide into a cell and with the desired biological activity, sequence-specific inhibition of gene expression. Many molecules, including polynucleotides, administered to the vascular structures of mammals are normally cleared by the liver itself. The following liver-achieved polynucleotide clearance was not considered functional delivery: in which the polynucleotide is degraded or otherwise processed for removal by itself and in which the polynucleotide does not cause sequence-specific inhibition of gene expression.

RNAi polynucleotide conjugates are formed by covalently linking an RNAi polynucleotide to a polynucleotide targeting moiety. Polynucleotides are synthesized or modified to contain reactive groups a. The targeting moiety is also synthesized or modified to contain a reactive group B. The reactive groups a and B are selected so that they can be linked by covalent bonding using methods known in the art.

The targeting moiety can be attached to the 3 'or 5' end of the RNAi polynucleotide. For siRNA polynucleotides, the targeting moiety may be attached to the sense strand or the antisense strand, but the sense strand is preferred.

In one embodiment, the polynucleotide targeting moiety consists of a hydrophobic group. More specifically, the polynucleotide targeting moiety consists of a hydrophobic group having at least 20 carbon atoms. Hydrophobic groups used as polynucleotide targeting moieties are referred to herein as hydrophobic targeting moieties. Exemplary suitable hydrophobic groups may be selected from the group consisting of: cholesterol, di-cholesterol, tocopherol (tocophenol), di-tocopherol, didecyl, didodecyl, dioctadecyl, isoprenoid and cholenamide (choleamide). Hydrophobic groups having 6 or less than 6 carbon atoms are not effective as polynucleotide targeting moieties, while hydrophobic groups having 8 to 18 carbon atoms provide incremental polynucleotide delivery as the size of the hydrophobic group is increased (i.e., the number of carbon atoms is increased). Attachment of the hydrophobic targeting moiety to the RNAi polynucleotide does not provide efficient functional in vivo delivery of the RNAi polynucleotide in the absence of co-administration of the delivery peptide. Although others have reported that siRNA-cholesterol conjugates can deliver siRNA (siRNA-cholesterol) into liver cells in vivo in the absence of any additional delivery vehicle, high siRNA concentrations are required and delivery efficacy is poor. Delivery of polynucleotides is greatly improved when combined with the delivery peptides described herein. By providing siRNA-cholesterol along with the delivery peptide of the present invention, the efficacy of siRNA-cholesterol is increased by about 100-fold.

Hydrophobic groups suitable for use as targeting moieties for polynucleotides may be selected from the group consisting of: alkyl, alkenyl, alkynyl, aryl, aralkyl, aralkenyl and aralkynyl (each of which may be linear, branched or cyclic), cholesterol derivatives, sterols, steroids and steroid derivatives. The hydrophobic targeting moiety is preferably a hydrocarbon containing only carbon and hydrogen atoms. However, substitutions or heteroatoms, such as fluorine, which maintain hydrophobicity may be permissible. The hydrophobic targeting moiety can be attached to the 3 'or 5' end of the RNAi polynucleotide using methods known in the art. For RNAi polynucleotides having 2 strands (such as siRNA), a hydrophobic group can be attached to either strand.

In another embodiment, the polynucleotide targeting moiety comprises a galactose cluster (galactose cluster targeting moiety). As used herein, a galactose cluster comprises molecules having 2 to 4 terminal galactose derivatives. As used herein, the term galactose derivative includes galactose and galactose derivatives having an affinity for an asialoglycoprotein receptor equal to or greater than the affinity of galactose for an asialoglycoprotein receptor. The terminal galactose derivative is attached to the molecule through its C-1 carbon. Asialoglycoprotein receptor (ASGPr) is unique to hepatocytes and binds to branched galactose terminal glycoproteins. A preferred galactose cluster has 3 terminal galactosamines or galactosamine derivatives, each with affinity for the asialoglycoprotein receptor. A more preferred galactose cluster has 3 terminal N-acetyl-galactosamine. Other terms commonly used in the art include triantennary galactose, trivalent galactose and galactose trimer. The binding affinity of the triantennary galactose derivative cluster to ASGPR is known to be greater than that of the biantennary or monoantennary galactose derivative structures (Baenziger and Fiete,1980, Cell,22, 611-945; Connolly et al, 1982, J.biol.chem.,257, 939-945). Multivalence is required to achieve nM affinity. Linking a single galactose derivative with affinity for the asialoglycoprotein receptor does not enable functional delivery of the RNAi polynucleotide into hepatocytes in vivo when co-administered with a delivery peptide.

Galactose

The galactose cluster contains 3 galactose derivatives, each linked to a central branch point. The galactose derivative is linked to the central branch point through the C-1 carbon of the sugar. The galactose derivative is preferably linked to the branch point via a linker or spacer. A preferred spacer is a flexible hydrophilic spacer (U.S. Pat. No. 5885968; Biessen et al J.Med.chem.1995 Vol.39, page 1538-1546). A preferred flexible hydrophilic spacer is a PEG spacer. One preferred PEG spacer is PEG₃A spacer. The branch point may be any small molecule that allows attachment of 3 galactose derivatives and further allows attachment of the branch point to the RNAi polynucleotide. An exemplary branch point group is dilysine. The dilysine molecule contains 3 amino groups, 3 galactose derivatives can be linked; also, a carboxyl reactive group, a dilysine, may be attached to the RNAi polynucleotide. Attachment of the branch point to the RNAi polynucleotide may occur through a linker or spacer. A preferred spacer is a flexible hydrophilic spacer. A preferred flexible hydrophilic spacer is a PEG spacer. One preferred PEG spacer is PEG₃Spacer (3 ethylene units). The galactose cluster may be linked to the 3 'or 5' end of the RNAi polynucleotide using methods known in the art. For RNAi polynucleotides having 2 strands, such as siRNA, galactose clusters may be linkedEither chain.

One preferred galactose derivative is N-acetyl-galactosamine (GalNAc). Other saccharides having affinity for asialoglycoprotein receptors may be selected from the group comprising: galactose, galactosamine, N-formyl galactosamine, N-acetyl galactosamine, N-propionyl-galactosamine, N-N-butyryl galactosamine and N-isobutyryl galactosamine. The affinity of numerous galactose derivatives for asialoglycoprotein receptors has been studied (see, e.g., Iobst, S.T. and Drickamer, K.J.B.C.1996,271,6686) or readily determined using methods typical in the art.

One embodiment of galactose Cluster

Galactose cluster with PEG spacer between branch point and nucleic acid

The term polynucleotide or nucleic acid or polynucleic acid is a term in the art, meaning a polymer containing at least 2 nucleotides. Nucleotides are monomeric units of polynucleotide polymers. Polynucleotides having less than 120 monomer units are often referred to as oligonucleotides. Natural nucleic acids have a deoxyribose phosphate or ribose phosphate backbone. A non-natural or synthetic polynucleotide is a polynucleotide that is polymerized in vitro or in a cell-free system and contains the same or similar bases but may contain a backbone of a type other than a ribose phosphate or deoxyribose phosphate backbone. Polynucleotides can be synthesized using any technique known in the art. Polynucleotide backbones known in the art include: PNA (peptide nucleic acid), phosphorothioate, phosphorodiamidate, morpholino (morpholino), and other variants of the phosphate backbone of natural nucleic acids. Bases include purines and pyrimidines, which further include the natural compounds adenine, thymine, guanine, cytosine, uracil, inosine, and natural analogs. Synthetic derivatives of purines and pyrimidines include, but are not limited to, modifications that place new reactive groups such as, but not limited to, amines, alcohols, thiols, carboxylates, and alkyl halides on the nucleotides. The term base encompasses any known base analog of DNA and RNA. The polynucleotide may contain ribonucleotides, deoxyribonucleotides, synthetic nucleotides, or any suitable combination. The polynucleotides may be polymerized in vitro, they may be recombinant, contain chimeric sequences or derivatives of these groups. The polynucleotide may include a terminal cap moiety at the 5 'end, the 3' end, or both the 5 'and 3' ends. The cap moiety may be, but is not limited to, an inverted deoxyabasic moiety, an inverted deoxythymidine moiety, a thymidine moiety, or a3' glyceryl modification.

An RNA interference (RNAi) polynucleotide is a molecule capable of degrading or inhibiting translation of a transgenic messenger RNA (mrna) transcript in a sequence-specific manner by interacting with the RNA interference pathway machinery of mammalian cells to induce RNA interference. The two major RNAi polynucleotides are small (or short) interfering RNAs (sirnas) and micrornas (mirnas). The RNAi polynucleotide may be selected from the group consisting of: siRNA, microrna, double-stranded RNA (dsrna), short hairpin RNA (shrna), and expression cassettes encoding RNA capable of inducing RNA interference. siRNA comprises a double-stranded structure, which typically contains 15 to 50 base pairs and preferably 21 to 25 base pairs, and the nucleotide sequence is identical (fully complementary) or nearly identical (partially complementary) to a coding sequence in a target gene or RNA expressed in a cell. The siRNA may have dinucleotide 3' overhangs. The siRNA may be composed of two annealed polynucleotides or a single polynucleotide forming a hairpin structure. The siRNA molecules of the invention comprise a sense region and an antisense region. In one embodiment, the siRNA of the conjugate is assembled from two oligonucleotide fragments, wherein one fragment comprises the nucleotide sequence of the antisense strand of the siRNA molecule and the second fragment comprises the nucleotide sequence of the sense region of the siRNA molecule. In another embodiment, the sense strand is linked to the antisense strand by a linker molecule such as a polynucleotide linker or a non-nucleotide linker. Micrornas (mirnas) are small, non-coding RNA gene products of about 22 nucleotides in length that direct disruption or translational repression of their mRNA targets. If there is partial complementarity between the miRNA and the target mRNA, translation of the target mRNA is repressed. If the complementarity is extensive, the target mRNA is cleaved. For mirnas, the complex binds to a target site that is usually located in the 3' UTR of mrnas that usually share only partial homology with mirnas. The "seed region," a stretch of about seven (7) contiguous nucleotides at the 5' end of the miRNA that forms complete base pairing with its target, plays a key role in miRNA specificity. Binding of the RISC/miRNA complex to mRNA can lead to repression of protein translation or cleavage and degradation of mRNA. Recent data indicate that mRNA cleavage occurs preferentially if there is complete homology along the entire length of the miRNA and its target rather than showing complete base pairing only in the seed region (pilai et al 2007).

The RNAi polynucleotide expression cassette can be transcribed in a cell to produce a small hairpin RNA that can act as an siRNA, independent sense and antisense strand linear sirnas, or an miRNA. The DNA transcribed by RNA polymerase III contains a promoter selected from the group consisting of: the U6 promoter, the H1 promoter, and the tRNA promoter. RNA polymerase II promoters include U1, U2, U4 and U5 promoters, snRNA promoter, microRNA promoter and mRNA promoter.

A list of known miRNA sequences can be found in databases maintained by research organizations such as: the Wellcome Trust Foundation Sanger Institute (Wellcome true Sanger Institute), the Pennsylvania Bioinformatics Center (Penn Center for Bioinformatics), the Meremon Kaltren Cancer Center (medical Sloan Kettering Cancer Center), and the European molecular biology laboratory (European molecular biology laboratory). Known effective siRNA sequences and homologous binding sites are also well described in the relevant literature. RNAi molecules are readily designed and produced by techniques known in the art. In addition, computational tools exist to increase the chances of finding potent and specific sequence motifs (Pei et al 2006, Reynolds et al 2004, Khvorova et al 2003, Schwarz et al 2003, Ui-Tei et al 2004, heal et al 2005, talk et al 2004, Amarzguioui et al 2004).

The polynucleotides of the invention may be chemically modified. Non-limiting examples of such chemical modifications include: phosphorothioate internucleotide linkages, 2 '-O-methyl ribonucleotides, 2' -deoxy-2 '-fluoro ribonucleotides, 2' -deoxyribonucleotides, "universal base" nucleotides, 5-C-methyl nucleotides, and inverted deoxy abasic residues. These chemical modifications, when used in various polynucleotide constructs, have been shown to retain polynucleotide activity in cells while at the same time increasing the serum stability of these compounds. Chemically modified sirnas may also minimize the possibility of activating interferon activity in humans.

In one embodiment, the chemically modified RNAi polynucleotides of the invention comprise duplexes having two strands, one or both of which may be chemically modified, wherein each strand has from about 19 to about 29 nucleotides. In one embodiment, the RNAi polynucleotides of the invention comprise one or more modified nucleotides while maintaining the ability to mediate RNAi within a cell or reconstituted in vitro system. The RNAi polynucleotide can be modified, wherein the chemical modification comprises one or more (e.g., about 1, 2,3, 4,5, 6, 7, 8, 9, 10, or more than 10) nucleotides. The RNAi polynucleotides of the invention can comprise modified nucleotides that comprise a certain percentage of the total number of nucleotides present in the RNAi polynucleotide. Thus, an RNAi polynucleotide of the invention can generally comprise about 5% to about 100% of a nucleotide position (e.g., 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or 100% of a nucleotide position) of a modified nucleotide. The actual percentage of modified nucleotides present in a given RNAi polynucleotide depends on the total number of nucleotides present in the RNAi polynucleotide. If the RNAi polynucleotide is single-stranded, the percentage modification can be based on the total number of nucleotides present in the single-stranded RNAi polynucleotide. Likewise, if the RNAi polynucleotide is double-stranded, the percentage modification can be based on the total number of nucleotides present in the sense strand, the antisense strand, or both the sense and antisense strands. In addition, the actual percentage of modified nucleotides present in a given RNAi polynucleotide can also depend on the total number of purine and pyrimidine nucleotides present in the RNAi polynucleotide. For example, wherein all pyrimidine nucleotides and/or all purine nucleotides present in the RNAi polynucleotide are modified.

The RNAi polynucleotide modulates expression of RNA encoded by the gene. Because multiple genes may share some degree of sequence homology with each other, RNAi polynucleotides can be designed to target a class of genes with sufficient sequence homology. Thus, the RNAi polynucleotide can contain sequences that are complementary to a consensus sequence of different gene targets or to a unique sequence of a particular gene target. Thus, RNAi polynucleotides can be designed to target conserved regions of RNA sequences with homology between several genes, thereby targeting several genes in a gene family (e.g., different gene isoforms, splice variants, mutant genes, etc.). In another embodiment, the RNAi polynucleotide can be designed to target sequences unique to a particular RNA sequence of a single gene.

The term complementarity refers to the ability of a polynucleotide to form hydrogen bonds with another polynucleotide sequence through the classical Chinese-Crick (Watson-Crick) type or other unconventional types. In the case of the polynucleotide molecules of the invention, the binding free energy of the polynucleotide molecule to its target (effector binding site) or complementary sequence is sufficient to allow the relevant function of the polynucleotide to proceed, for example enzymatic mRNA cleavage or translational inhibition. The determination of the binding free energy of nucleic acid molecules is well known in the art (Frier et al 1986, Turner et al 1987). Percent complementarity indicates the percentage of bases (e.g., 5, 6, 7, 8, 9, 10 out of 10, 50%, 60%, 70%, 80%, 90%, and 100% complementary) that can form hydrogen bonds (e.g., watson-crick base pairing) with the second polynucleotide sequence in adjacent strands in the first polynucleotide molecule. By fully complementary is meant that all bases in adjacent strands of the polynucleotide sequence will form hydrogen bonds with the same number of adjacent bases in the second polynucleotide sequence.

Inhibiting, down-regulating or knocking-down (knockdown) gene expression means that gene expression, as measured by the level of RNA transcribed from the gene or the level of polypeptide, protein or protein subunit translated from the RNA, is reduced below that observed in the absence of the blocking polynucleotide conjugate of the present invention. The inhibition, down-regulation or knock-down of gene expression with polynucleotides delivered by the compositions of the invention is preferably less than that observed in the presence of control inactive nucleic acids (i.e., nucleic acids with misordering or with inactive mismatches) or in the absence of conjugation of the polynucleotide to the masked polymer.

In vivo administration

In pharmacology and toxicology, the route of administration is the route by which a drug, fluid, poison, or other substance comes into contact with the body. Generally, methods of administration of drugs and nucleic acids for the treatment of mammals are well known in the art and may be applied to the administration of the compositions of the present invention. The compounds of the invention may be administered by any suitable route, most preferably parenterally, in a formulation suitably tailored to the route. Thus, the compounds of the invention may be administered by injection, for example, intravenous, intramuscular, intradermal, subcutaneous or intraperitoneal injection. Accordingly, the present invention also provides a pharmaceutical composition comprising a pharmaceutically acceptable carrier or excipient.

Parenteral routes of administration include intravascular (intravenous, intraarterial), intramuscular, intraparenchymal, intradermal, subdermal, subcutaneous, intratumoral, intraperitoneal, intrathecal, subdural, epidural, and intralymphatic injection using a syringe and needle or catheter. Intravascular in this context means within a tubular structure called a vessel that is connected to a tissue or organ in vivo. Within the lumen of the tubular structure, bodily fluids flow to or from the body part. Examples of body fluids include blood, cerebrospinal fluid (CSF), lymph or bile. Examples of vessels include arteries, arterioles, capillaries, venules, sinusoids, veins, lymphatic vessels, bile ducts, and salivary or other exocrine ducts. Intravascular routes include delivery through a blood vessel such as an artery or vein. The blood circulation system provides for systemic distribution of the drug.

The composition is injected in the form of a pharmaceutically acceptable carrier solution. Pharmaceutically acceptable refers to those properties and/or substances that are acceptable to a mammal from a pharmacological/toxicological standpoint. The phrase pharmaceutically acceptable refers to molecular entities, compositions, and properties that are physiologically tolerable and do not typically produce an allergic or other untoward or toxic reaction when administered to a mammal. Preferably, as used herein, the term pharmaceutically acceptable means approved by a regulatory agency of the federal or a state government or listed in the U.S. pharmacopeia or other generally recognized pharmacopeia for use in animals, and more particularly in humans.

The RNAi polynucleotide-targeting moiety conjugate is co-administered with the delivery peptide. Co-administration means administering the RNAi polynucleotide and the delivery peptide to the mammal such that both are present in the mammal. The RNAi polynucleotide-targeting moiety conjugate and the delivery peptide may be administered simultaneously or they may be delivered sequentially. For simultaneous administration, they may be mixed prior to administration. For sequential administration, the RNAi polynucleotide-targeting moiety conjugate or the delivery peptide may be administered first.

For RNAi polynucleotide-hydrophobic targeting moiety conjugates, the RNAi conjugate can be administered up to 30 minutes prior to administration of the delivery peptide. In addition, for RNAi polynucleotide-hydrophobic targeting moiety conjugates, the delivery peptide can be administered up to two hours prior to administration of the RNAi conjugate.

For RNAi polynucleotide-galactose cluster targeting moiety conjugates, the RNAi conjugate can be administered up to 15 minutes prior to administration of the delivery peptide. In addition, for RNAi polynucleotide-galactose cluster targeting moiety conjugates, the delivery peptide may be administered up to 15 minutes prior to administration of the RNAi conjugate.

Therapeutic effects

The RNAi polynucleotides can be delivered for research purposes or can be delivered to produce a therapeutic change in a cell. Delivery of RNAi polynucleotides in vivo can be used in research reagents and in a variety of therapeutic, diagnostic, target validation, genomic discovery, genetic engineering, and pharmacogenomic applications. We have disclosed RNAi polynucleotide delivery resulting in the inhibition of endogenous gene expression in hepatocytes. Reporter (marker) gene expression levels determined after delivery of the polynucleotide indicate a reasonable expectation of similar levels of gene expression after delivery of other polynucleotides. The level of treatment deemed beneficial by one of ordinary skill in the art varies from disease to disease. For example, hemophilia a and B are caused by deficiencies in the X-linked coagulation factors VIII and IX, respectively. Its clinical course is strongly influenced by the percentage of normal serum levels of factor VIII or IX: <2%, severe; 2-5%, medium; and 5-30%, mild. Thus, an increase from 1% to 2% of the normal level of self-circulation factor may be considered beneficial in severe patients. Levels greater than 6% prevent spontaneous bleeding, but do not prevent secondary bleeding from surgery or injury. Similarly, inhibition of a gene need not be 100% to provide a therapeutic benefit. One of ordinary skill in the art of gene therapy would reasonably expect a beneficial level of gene expression specific to a disease based on a sufficient degree of marker gene outcome. In the hemophilia example, if the protein level produced by marker gene expression is comparable in amount to 2% of the normal level of factor VIII, it is reasonable to expect that the gene encoding factor VIII will also be expressed at a similar level. Thus, reporter genes or marker genes generally serve as a useful paradigm for intracellular protein expression.

In view of the important role of the liver in metabolism (e.g. lipoprotein metabolism in various hypercholesterolaemias) and in the secretion of circulating proteins (e.g. coagulation factors in hemophilia), the liver is one of the most important target tissues for gene therapy. In addition, acquired conditions such as chronic hepatitis (e.g. hepatitis b virus infection) and cirrhosis are common and may also be treated by polynucleotide-based liver therapy. Many diseases or conditions that affect or are affected by the liver are likely to be treated by the knockdown (inhibition) of gene expression in the liver. These liver diseases and conditions may be selected from the group consisting of: liver cancer (including hepatocellular carcinoma HCC), viral infections (including hepatitis), metabolic disorders (including hyperlipidemia and diabetes), fibrosis, and acute liver injury.

The amount (dose) to be administered of the delivery peptide and the RNAi polynucleotide conjugate can be determined empirically. We have shown that gene expression is efficiently knocked down using 0.1 to 10mg siRNA conjugate per kg animal weight and 5 to 60mg delivery peptide per kg animal weight. Preferred amounts in mice are 0.25 to 2.5mg/kg siRNA conjugate and 10 to 40mg/kg delivery peptide. More preferably, about 12.5 to 20mg/kg of delivery peptide is administered. It is easy to increase the amount of RNAi polynucleotide conjugate because it is generally non-toxic at larger doses.

As used herein, in vivo means occurring within an organism, and more specifically refers to a process that is performed in or on the living tissue of an intact, living multicellular organism (animal), such as a mammal, rather than a partial or dead organism.

Examples

Example 1 melittin synthesis. All melittin peptides were prepared using standard peptide synthesis techniques in the art. Suitable melittin peptides can be all L-type amino acids, all D-type amino acids (flip). The melittin peptide sequence may be inverted (reversed) regardless of the L-type or D-type.

Example 2 melittin modification.

Amino terminal modification of melittin derivatives. In dH₂A solution of CKLK-melittin (20mg/ml), TCEP-HCl (28.7mg/ml, 100mM) and MES-Na (21.7mg/ml, 100mM) was prepared in O. CKLK-melittin (0.030mmol, 5ml) was reacted with 1.7 molar equivalents of TCEP-HCl (0.051mmol, 0.51ml) in a 20ml scintillation vial and stirred at room temperature for 30 min. MES-Na (2ml) and water (1.88ml) were then quantitatively added to give final concentrations of melittin and 20mM MES-Na, 10 mg/ml. The pH was checked and adjusted to pH 6.5-7. In dH₂Preparation of NAG-PEG in O₂-a solution of Br (100 mg/ml). Adding NAG-PEG₂-Br (4.75 equiv., 0.142mmol, 0.61ml) and the solution was stirred at room temperature for 48 h.

Alternatively, Cys-melittin (0.006mmol, 1ml) was reacted with 1.7 molar equivalents of TCEP-HCl (0.010mmol, 100. mu.l) in a 20ml scintillation vial and stirred at room temperature for 30 min. MES-Na (400. mu.l) and water (390. mu.l) were quantitatively addedl) to give a final concentration of 10mg/ml melittin and 20mM MES-Na. The pH was checked and adjusted to pH 6.5-7. In dH₂Preparation of NAG-PEG in O₈-a solution of maleimide (100 mg/ml). Adding NAG-PEG₈Maleimide (2 eq, 0.012mmol, 110 μ l), and the solution was allowed to stir at room temperature for 48 hours.

In Luna10 u C1821.2 × 250mm column purification sample buffer A: H₂O0.1% TFA; and buffer B: MeCN, 10% isopropanol, 0.1% TFA. The flow rate was 15ml/min, 35% A to 62.5% B in 20 minutes.

Other amino-terminal modifications were performed using similar means. The carboxyl-terminal modification was performed by substituting melittin with amino-terminal cysteine with melittin peptide with carboxyl-terminal cysteine.

Compounds for modified Cys-melittin or melittin-Cys:

PEG maleimide

n is an integer from 1 to 120 (PEG molecular weight up to about 5kDa)

NAG-PEG maleimide

N-NAG-PEG bromoacetamide

During peptide synthesis, peptides with acetyl, dimethyl, stearoyl, myristoyl and PEG amino or carboxyl terminal modifications on the resin rather than terminal cysteine residues were produced using methods typical in the art.

Example 3 masking agent synthesis.

Ph labile masking agent: synthesizing a space stabilizer CDM-PEG and a targeting group CDM-NAG (N-acetylgalactosamine). To a solution of CDM (300mg, 0.16mmol) in 50mL of dichloromethane were added oxalyl chloride (2g, 10 weight equivalents) and dimethylformamide (5. mu.L). The reaction was allowed to proceed overnight, after which excess oxalyl chloride and dichloromethane were removed by rotary evaporation to yield CDM acid chloride. The acid chloride was dissolved in 1mL of dichloromethane. To this solution was added 10mL of dichloromethane containing 1.1 molar equivalents of polyethylene glycol monomethyl ether (average MW 550; for CDM-PEG) or (aminoethoxy) ethoxy-2- (acetylamino) -2-deoxy- β -D-galactopyranoside (i.e., aminobisethoxy-ethyl NAG; for CDM-NAG), and pyridine (200 μ l, 1.5 equivalents). The solution was then stirred for 1.5 hours. The solvent was then removed and the resulting solid was dissolved in 5mL of water and purified using reverse phase HPLC with a 0.1% TFA water/acetonitrile gradient.

CDM

Universal disubstituted maleic anhydride masking agents

R1 comprises a neutral ASGPr ligand. The masking agent is preferably uncharged.

CDM-PEG

R is methyl or ethyl, and n is an integer from 2 to 100. Preferably, the PEG contains 5 to 20 ethylene units (n is an integer from 5 to 20). More preferably, the PEG contains 10 to 14 ethylene units (n is an integer from 10 to 14). PEG can have variable length and average length of 5 to 20 or 10 to 14 ethylene units. Alternatively, the PEG may be monodisperse, uniform or discrete; with for example exactly 11 or 13 ethylene units.

CDM-NAG

n is an integer of 1 to 10. As indicated above, a PEG spacer may be located between the anhydride group and the ASGPr ligand. Preferably the PEG spacer contains 1 to 10 ethylene units.

Alternatively, an alkyl spacer may be used between the anhydride and N-acetylgalactosamine.

CDM-NAG (alkyl spacer)

n is an integer of 0 to 6.

Other spacers or linkers may be used between the anhydride and N-acetyl-galactosamine. However, hydrophilic, neutral (preferably uncharged) spacers or linkers are preferred.

B. Protease (peptidase) -cleavable masking agents. Melittin peptides may also be reversibly modified using specific enzyme cleavable linkers. These enzymatically cleavable linkers employ a dipeptide attached to an amidobenzyl-activated carbonate moiety. Targeting compounds such as asialoglycoprotein receptor ligands are linked to melittin peptides via peptidase-cleavable dipeptide-amidobenzyl carbamate linkages via the reaction of activated carbonates with peptide amines. Enzymatic cleavage of the dipeptide removes the targeting ligand from the peptide and triggers an elimination reaction that causes regeneration of the peptide amine. The following enzymatically cleavable linkers were synthesized:

NAG-Ala-Cit-PABC-PNP

NAG-Glu-Gly-PABC-PNP

NAG-PEG4-Phe-Cit-PABC-PNP

NAG-PEG7-Phe-Cit-PABC-PNP

the dipeptides Glu-Gly, Ala-Cit, Phe-Cit ("Cit" is the amino acid citrulline) are shown. Other combinations of amino acids may be allowed. Furthermore, 3 to 5 amino acids may be used as a linker between the amidobenzyl group and the targeting ligand. In addition, other activated carbonates known in the art readily replace the para-nitrophenols used in the above compounds.

Example 4 reversible modification/masking of melittin.

A. Modified with a maleic anhydride-based masking agent. Prior to modification, 5 x mg of a disubstituted maleic anhydride masking agent (e.g., CDM-NAG) was lyophilized from 0.1% aqueous glacial acetic acid. To the dried disubstituted maleic anhydride masking agent was added 0.2 × mL isotonic glucose solution containing × mg melittin and 10 × mg HEPES free base. After the anhydride was completely dissolved, the solution was incubated at room temperature for at least 30 minutes prior to administration to the animals. The di-substituted maleic anhydride masking agent reacts with the peptide to yield:

wherein R is melittin and R1 comprises an ASGPr ligand (e.g., NAG). The anhydride carboxyl groups generated in the reaction between the anhydride and the polymeric amine exhibit the expected charge of about 1/20 (Rozema et al Bioconjugate Chemistry 2003). Thus, the membrane active polymer is effectively neutralized rather than converted to a highly negatively charged polyanion.

B. Modified with a masking agent cleavable by a protease. 1-10mg/mL peptide 1 x mg peptide and 10 x mg HEPES base were masked by the addition of 2-6 x mg amine-reactive p-nitrophenyl carbonate or N-hydroxysuccinimide carbonate derivatives containing a protease-cleavable substrate of NAG. The solution was then incubated at Room Temperature (RT) for at least 1 hour, followed by injection into animals.

Example 5 siRNA. The siRNA had the following sequence:

factor VII-rodents

A sense: (Chol) -5 'GfcAFaAfgGfcGfccCfaAfcUfcAf (invdT)3' (Seq ID97)

Antisense: 5 'pdTsgfaGfuUfgGfcaFcCfcCfuUfuGfccdTsdT 3' (Seq ID98)

Or

Sense 5 'GGAUfCfAUfCfUfCfAAGUfCfUfUfACfdTdT 3' (Seq ID99)

Antisense 5 'GUfAAGACfUfUfGAUfGAUfCfCfdTdT 3' (Seq ID100)

Factor VII = primate

Sense (chol) -5 'uuAGGfuUfGfuGfaAffGfAfGfCfuCfaGf (invdT)3' (Seq ID101)

Antisense 5 'pCfsUfgAfgCfuCfcAUfcAfCfaAfcdTsdT 3' (Seq ID102)

ApoB siRNA:

Sense (cholC6SSC6) -5 'GGAAUCuAuuAuuGAUCcAsA 3' (Seq ID103)

Antisense 5 'uuGGAUcAAAuAuAAGAuUCcscU 3' (Seq ID104)

siLUC

Sense (chol)5 '-uAuCfuFafCfuGfaGfuAfcUfuCfgAf (invdT) -3' (Seq ID105)

Antisense 5 '-UfcgaafUfaCfuCfaGfcGcGfuAfaGfdTdT-3' (Seq ID106)

Lower case =2' -O-CH₃Substitution

s = phosphorothioate linkage

F =2' -F substitution after nucleotide

D =2' -deoxy before nucleotide

In thatRNA synthesis was performed on solid phase by conventional phosphoramidate chemistry on Oligopilot100(GE Healthcare, fleburg, germany) and with Controlled Pore Glass (CPG) as solid support.

Example 6.siRNA targeting molecule conjugates.

A. Synthesizing GalNAc cluster. The GalNAc cluster polynucleotide targeting ligand was synthesized as described in U.S. patent publication 20010207799.

GalNAc cluster-siRNA conjugates. As shown in figure 2 and described below, the GalNAc cluster of example 6A above was conjugated to siRNA.

(1) Compound 1(150mg, 0.082mmol, fig. 2) was dissolved in anhydrous methanol (5.5ml) and 42 μ L of sodium methoxide (25% in methanol) was added. The mixture was stirred at room temperature under argon atmosphere for 2 hours. Equal amounts of methanol and parts of the anion exchange material Amberlite (Amberlite) IR-120 were added to give a pH of about 7.0. Amberlite was removed by filtration. With Na₂SO₄The solution was dried and the solvent was removed under reduced pressure. Compound 2 was obtained in quantitative yield as a white foam. TLC (SiO)₂Dichloromethane (DCM)/MeOH5:1+0.1% CH₃COOH)：R_f2=0.03, detection was with a solution of MeOH containing sulfuric acid (5%) followed by heating. ESI-MS, direct injection, negative ion mode; [ M-H ]]^-1_Computing：1452.7；[M-H]^1-_{Measured in fact}：1452.5。

(2) Compound 2(20mg, 0.014mmol, FIG. 2) was co-evaporated with pyridine and dichloromethane. The residue was dissolved in anhydrous DMF (0.9ml) and a solution of N-hydroxysuccinimide (NHS) in DMF (1.6mg, 0.014mmol) was added while stirring under an argon atmosphere. A DMF solution (3.2mg, 0.016mmol) containing N, N' -Dicyclohexylcarbodiimide (DCC) was added slowly at 0 ℃. The reaction was allowed to warm to room temperature and stirred overnight. Compound 3 was used for conjugation with RNA without further purification.

(3) Amino-modified RNA was synthesized. Use ofOligopilot100(GE Healthcare, Freuburg, Germany) and with controlled pore glass as solid support, RNA equipped with a C-6 amino linker at the 5' end of the sense strand was generated on a solid phase by standard phosphoramidite chemistry at 1215 μmol scale. RNA containing 2 '-O-methyl nucleotides was generated using the corresponding phosphoramidite, 2' -O-methylphosphide and TFA-hexylamino linker amide (amidite). Cleavage and deprotection and purification are achieved by methods known in the artAnd (Wincott F. et al, NAR1995,23,14, 2677-84).

Amino-modified RNA (purity: 96.1%) was characterized by anion exchange HPLC and identity was confirmed by ESI-MS ([ M + H ]]¹⁺_Computing：6937.4；[M+H]¹⁺_{Measured in fact}: 6939.0). The sequence is as follows:

5′-(NH₂C₆) GGAAUCuAuuAuuGAUCcAsA-3'; u and c: 2' -O-methyl nucleotide of the corresponding base, s: a thiophosphate.

(4) The GalNAc cluster is conjugated to RNA. RNA equipped with a C-6 amino linker at the 5' end (2.54. mu. mol) was lyophilized and dissolved in 250. mu.L sodium borate buffer (0.1mol/L sodium borate, pH8.5, 0.1mol/LKCl) and 1.1mL DMSO. After addition of 8 μ L N, N-Diisopropylethylamine (DIPEA), DMF solution containing compound 3 (theoretically 0.014mmol, FIG. 2) was added slowly to the RNA solution with continuous stirring. The reaction mixture was stirred at 35 ℃ overnight. The reaction was monitored using RP-HPLC (Resource RPC3ml, buffer: A: water with 100mM triethylammonium acetate (TEAA, 2.0M, pH7.0), B: 95% acetonitrile with 100mM TEAA, gradient: 5% B to 22% B, 20 Column Volumes (CV)). After precipitation of RNA at-20 ℃ using EtOH containing sodium acetate (3M), the RNA conjugate was purified using the conditions described above. The pure fractions were pooled and the desired conjugate compound 4 was precipitated using sodium acetate/EtOH, yielding a pure RNA conjugate. Conjugate 4 was isolated in 59% yield (1.50. mu. mol). The purity of conjugate 4 (purity: 85.5%) was analyzed by anion exchange HPLC and the identity was confirmed by ESI-MS ([ M + H ]]¹⁺_Computing：8374.4；[M+H]¹⁺_{Measured in fact}：8376.0)。

(5) Conjugate 4 (sense strand) anneals to the 2' -O-methyl modified antisense strand. The complementary strands were mixed in an equimolar solution in annealing buffer (20mM sodium phosphate, ph6.8, 100mM sodium chloride), heated in a water bath at 85-90 ℃ for 3 minutes, and cooled to room temperature over a period of 3 to 4 hours, thereby generating siRNA conjugates. Duplex formation was confirmed by native gel electrophoresis.

C. Hydrophobic group-siRNA conjugates.

(1) The siRNA is conjugated to an alkyl group. 5'-C10-NHS ester modified sense strand of siRNA (NHSC10-siRNA or COC9-siRNA) was prepared using 5' -carboxy modifier C10 amide from Glen Research (Va., U.S.A.). The activated RNA still attached to the solid support was used for conjugation with the lipophilic amines listed in table 1 below. 100mg of sense strand CPG (loaded with 60. mu. mol/g, 0.6. mu. mol of RNA) were mixed with 0.25mmol of the corresponding amine from Sigma Aldrich Chemie GmbH (Doffyn, Germany) or Fluka (Sigma-Aldrich, Switzerland Buchs).

TABLE 1 lipophilic amines for forming hydrophobic group-siRNA conjugates

Numbering	Lipophilic amines	mg	mmol	Solvent(s)
					2	N-hexylamine	25	0.25	1mL CH2Cl2
3	Dodecyl amine	50	0.25	0.55mL CH3CN，0.45mL CH2Cl2
					4	Octadecamine	67	0.25	1mL CH2Cl2
5	Didecylamine	74	0.25	1mL CH2Cl2
					6	Bis (dodecylamine)	88	0.25	1mL CH2Cl2
7	Dioctadecylamine	67	0.12	0.45mL CH2Cl20.45mL of cyclohexane

The mixture was shaken at 40 ℃ for 18 hours. RNA was cleaved from the solid support and washed with aqueous ammonium hydroxide (NH)₃33%) at 45 ℃ overnight, remove the 2' -protecting group with TEA × 3HF at 65 ℃ for 3.5 hours by RP-HPLC (Resource RPC3ml, buffer: A: water containing 100mM TEAA, B: 95% CH containing 100mM TEAA₃CN, gradient: from 3% B to 70% B, 15 column volumes; number 7 is an exception: from 3% B to 100% B, 15 column volumes).

To generate siRNA from an RNA single strand, equimolar amounts of complementary sense and antisense strands were mixed in annealing buffer (20mM sodium phosphate, pH 6.8; 100mM sodium chloride), heated at 80 ℃ for 3 minutes, and cooled to room temperature over a period of 3 to 4 hours. siRNA against factor VII mRNA was characterized by gel electrophoresis.

(2) siRNA conjugation to cholesterol-siRNA-cholesterol conjugates were synthesized using standard methods in the art. Cholesterol may be attached to the 5 'or 3' end of the sense strand or antisense strand of the siRNA. Preferably to the 5' end of the sense strand of the siRNA. siRNA-cholesterol can also be prepared following siRNA synthesis using an RNA strand containing a reactive group (e.g., a thiol, amine, or carboxyl group) using methods standard in the art.

In vivo siRNA delivery

Example 7 RNAi polynucleotides are administered in vivo and delivered to hepatocytes. RNAi polynucleotide conjugates and masked melittin peptides were synthesized as described above. Mice of 6 to 8 weeks of age (strain C57BL/6 or ICR, each about 18 to 20g) were obtained from Harlan Sprague Dawley (indianapolis, indiana). Mice were housed for at least 2 days prior to injection. Feeding was performed ad libitum with Harlan Teklad rodent diet (Harlan, madison, wisconsin). 0.2mL of the delivery peptide solution and 0.2mL of the siRNA conjugate were injected into the tail vein of the mice. For simultaneous injection of the delivery peptide and siRNA, the siRNA conjugate was added to the modified peptide prior to injection and the full amount was injected. The composition is soluble and non-aggregating under physiological conditions. The solution was injected into the tail vein by infusion. Injection into other blood vessels (e.g., retroorbital injection) is expected to be equally effective.

From Charles River (Wilmington, Mass.) 175 to 200g Wistar Han rats were obtained. Rats were housed for at least 1 week prior to injection. The injection volume for rats is typically 1 ml.

And (4) measuring the content of the serum ApoB. Mice were fasted for 4 hours (16 hours for rats) before serum was collected by mandibular bleeding. For rats, blood was collected from the jugular vein. Serum ApoB protein content was determined by standard sandwich ELISA methods. Briefly, polyclonal goat anti-mouse ApoB antibody and rabbit anti-mouse ApoB antibody (biodesign international) were used as capture and detection antibodies, respectively. HRP-conjugated goat anti-rabbit IgG antibody (Sigma) was then administered to bind the ApoB/antibody complex. The absorbance of the colorimetric coloration of tetramethyl-benzidine (TMB, Sigma) was then measured by a TecanSafire2 (Austria, Europe) microplate reader at 450 nm.

Plasma factor VII (F7) activity measurement. Plasma samples from animals were prepared following standard procedures by collecting blood (9 volumes) (either by submandibular exsanguination for mice; or by jugular vein exsanguination for rats) into microcentrifuge tubes containing 0.109mol/L sodium citrate anticoagulant (1 volume). F7 activity in plasma was measured chromogenic using the BIOPHEN VII kit (Hyphen BioMed/Anira, Meison, Ohio) following the manufacturer's recommendations. The absorbance of the colorimetric color development was measured at 405nm using a Tecansafire2 microplate reader.

Example 8 delivery of ApoB siRNA in vivo knockdown of endogenous ApoB levels-dose response to melittin peptide following delivery of the peptide. Melittin was reversibly modified with CDM-NAG as described above. Then a prescribed amount of melittin was co-injected with 200 μ g ApoB siRNA-cholesterol conjugate. The effect on ApoB levels was determined as described above.

TABLE 2 comparison of ApoB Activity inhibition in normal liver cells in mice treated with ApoB-siRNA-cholesterol conjugate and CDM-NAG with melittin peptide reversibly inhibited by CDM-PEG.

^aKnockdown relative to animals injected with isotonic glucose

Example 9 knock down of endogenous factor VII levels in vivo following ApoB siRNA and melittin delivery peptide delivery in rats. Designated melittin was reversibly modified with 5 x CDM-NAG as described above. Subsequently, the indicated amounts of melittin (in milligrams per kilogram of animal weight) were co-injected with 3mg/kg cholesterol-factor VII siRNA. The effect on factor VII levels was determined as described above.

Table 3 inhibition of factor VII activity in normal liver cells in rats treated with factor VII-siRNA-cholesterol conjugate and melittin reversibly inhibited by CDM-NAG.

^aMg peptide per kg animal weight

^bKnockdown relative to animals injected with isotonic glucose

Example 10 knock down of endogenous ApoB levels in vivo following ApoB siRNA delivery with melittin delivery peptide in mice, type L melittin versus type D melittin. Melittin was reversibly modified with CDM-NAG as described above. Then a prescribed amount of melittin was co-injected with 50 μ g ApoB siRNA-cholesterol conjugate. The effect on ApoB levels was determined as described above.

Table 4 inhibition of ApoB activity in normal liver cells in mice treated with ApoB-siRNA cholesterol conjugate and melittin peptide specifying reversible inhibition by CDM-NAG.

Example 11 knock-down of endogenous ApoB levels in vivo following ApoB siRNA and melittin delivery peptide delivery in mice, normal versus inverted (inverted) sequences. Melittin was reversibly modified with CDM-NAG (5 ×) as described above. Followed by co-injection of the indicated amount of melittin with the indicated amount of ApoB siRNA-cholesterol conjugate. The effect on ApoB content was determined as described above.

Table 5 inhibition of ApoB activity in normal liver cells in mice treated with ApoB-siRNA cholesterol conjugate and melittin peptide specifying reversible inhibition by CDM-NAG.

^a-reverse turn = normal melittin amino acid sequence inverted and all amino acids are D-amino acids (glycine (G) is achiral)

Example 12 knock down of endogenous ApoB levels in vivo following ApoB siRNA and melittin delivery peptide delivery in mice, the extent of melittin modification. Melittin was reversibly modified with the indicated amount of CDM-NAG as described above. Then 50. mu.g of melittin was co-injected with 100. mu.g of ApoB siRNA-cholesterol conjugate. The effect on ApoB levels was determined as described above.

Percent melittin amine modification was determined by TNBS analysis of free amines on the peptide. Mu.g of peptide were pipetted into a 96-well transparent plate (NUNC96) containing 190. mu.L of 50mM BORAX buffer (pH9) and 16. mu.g of TNBS. The sample was reacted with TNBS at room temperature for about 15 minutes and then measured A on a Safire microplate reader₄₂₀. The% amine modification was calculated as follows: (A)_Control-A_{Sample (I)})/(A_Control-A_{Blank space}) × 100 amine modified by more than 80% provides optimal melittin masking and activity.

Table 6 inhibition of ApoB activity in normal liver cells in mice treated with ApoB-siRNA cholesterol conjugate and melittin reversibly modified with CDM-NAG to the indicated extent.

^aDetermination by TNBS analysis

Example 13 knock-down of endogenous ApoB levels in vivo following ApoB siRNA and melittin delivery peptide delivery in mice, melittin peptide derivatives. Melittin peptides with the indicated sequence were reversibly modified with CDM-NAG (5 ×) as described above. Followed by co-injection of the indicated amount of melittin with the indicated amount of ApoB siRNA-cholesterol conjugate. The effect on ApoB levels was determined as described above.

Table 7 inhibition of ApoB activity in normal liver cells in mice treated with ApoB-siRNA cholesterol conjugate and melittin peptide specifying reversible inhibition by CDM-NAG.

^aMu.g peptide per mouse

^bMu.g siRNA per mouse

dMel = melittin peptide with D-amino acids

Example 14 knock-down of endogenous ApoB levels in vivo, enzyme cleavable masking agents following ApoB siRNA and melittin delivery peptides delivery in mice. Melittin was reversibly modified with a specified amount of an enzymatically cleavable masking agent as described above. Then 200 to 300 μ g of the masked bee toxin was co-injected with 50 to 100 μ g of ApoB siRNA-cholesterol conjugate. The effect on ApoB levels was determined as described above. Peptidase-cleavable dipeptide-amidobenzyl carbamate modified melittin is an effective siRNA delivery peptide. Preferably, the D-melittin peptide is used in combination with an enzymatically cleavable masking agent. Although achieving the same degree of target gene knockdown requires more polypeptide, the therapeutic index is not changed or improved (compared to masking the same peptide with CDM-NAG) because the peptide masking is more stable.

Table 8 inhibition of factor VII activity in normal liver cells in mice treated with factor VII-siRNA cholesterol conjugates and G1L-melittin (type D) (Seq ID7) reversibly inhibited by designated enzyme cleavable masking agents.

^aMasking doses of each melittin amine used in the masking reaction.

Example 15 knock-down of endogenous ApoB levels in vivo after ApoB siRNA and melittin delivery peptides delivery in mice, amine modified melittin peptides. Melittin peptides containing the indicated PEG amino-terminal modifications were synthesized as described above. The PEG amino-terminally modified melittin peptide was then reversibly modified with 5 × CDM-NAG as described above. Then a prescribed amount of melittin was co-injected with 100 to 200 μ g of ApoB siRNA-cholesterol conjugate. The effect on ApoB levels was determined as described above. The addition of PEG with a size below 5kDa reduces the toxicity of melittin peptides. Amino-terminal modification with PEG above 5kDa resulted in reduced efficacy (data not shown).

Table 9 inhibition of ApoB activity in normal liver cells in mice treated with ApoB-siRNA cholesterol conjugates and melittin peptides reversibly inhibited by designated CDM-NAG.

Example 16 other melittin derivative sequences with membrane activity are known.

TABLE 10 melittin peptides with membrane activity.

Example 17 factor VII knockdown in primates following delivery of factor VII siRNA by melittin delivery peptide. NAG-PEG2-G1L melittin was masked by reaction with 10 xCDM-NAG as described above. The G1L melittin was masked by reaction with 5 x CDM-NAG as described above. On day 1, 1mg/kg of either the masked NAG-PEG2-G1L melittin, 1mg/kg of the masked G1L melittin or 3mg/kg of the masked G1L melittin was co-injected with 2mg/kg of cholesterol-factor VII siRNA into primates-eating crab macaques (Cynomolgus macaque) (Macaca fascicularis) (male, 3.0 to 8.0 kg). 2ml/kg were injected into the saphenous vein using a 22 to 25 gauge intravenous catheter. As a control, another group of primates was co-injected with 10mg/kgG1L melittin and 2mg/kg control siRNA cholesterol-luciferase siRNA. At the indicated time points (indicated in fig. 3 to 5), blood samples were drawn and analyzed for factor VII and toxicity markers. Blood was collected from the femoral vein and primates were fasted overnight prior to all blood collection. Blood tests for Blood Urea Nitrogen (BUN), alanine Aminotransferase (ALT), aspartate Aminotransferase (AST) and creatinine (creatinine) were performed on a Cobas Integra400(Roche Diagnostics) according to the manufacturer's recommendations. Factor VII levels were determined as described above. Significant knockdown of factor VII was observed at peptide doses of less than 1 mg/kg. No significant toxicity was observed at a peptide dose of 10 mg/kg. Thus, the therapeutic index of the masked melittin peptides is between 5 and 10.

Example 18 ApoB knockdown in primates following ApoB siRNA delivery by melittin delivery peptide. The G1L melittin was masked by reaction with 5 x CDM-NAG as described above. On day 1, 2mg/kg of the masked G1L melittin was co-injected into primate cynomolgus macaques (cynomolgus monkeys) with 2mg/kg of cholesterol-ApoB siRNA. At the indicated time points (table 11), blood samples were drawn and analyzed for ApoB protein levels and toxicity markers. Blood tests for Blood Urea Nitrogen (BUN), alanine Aminotransferase (ALT), aspartate Aminotransferase (AST) and creatinine were performed on a cobas integra400(Roche Diagnostics) according to the manufacturer's recommendations. ApoB levels were determined as described above. No increase in BUN, creatinine, or AST was observed. Only a brief slight increase in AST was observed on day 2 (1 day after injection). Knockdown of ApoB reached near 100% on day 11 and remained low for 31 days.

TABLE 11 inhibition of ApoB activity in normal liver cells in primates treated with ApoB-siRNA cholesterol conjugate and G1L melittin masked with CDM-NAG.

Claims (21)

1. A conjugate delivery system composition for delivering an RNA interference polynucleotide to a liver cell in vivo, comprising: RNAi-A and melittin- (L-Gal)_x

Wherein the content of the first and second substances,

melittin is melittin peptide:

l is a physiologically labile disubstituted maleate ester or dipeptide-amidobenzyl carbamate linkage,

gal comprises an asialoglycoprotein receptor (ASGPR) ligand,

x is an integer having a value greater than 80% of the number of primary amines on the plurality of melittin peptides,

RNAi is an RNA interference polynucleotide, and

a comprises a hydrophobic group having at least 20 carbon atoms or a galactose cluster containing 2-4 terminal galactose derivatives, wherein the galactose derivative consists of galactose or a galactose derivative having an affinity for an asialoglycoprotein receptor equal to or greater than the affinity of galactose for an asialoglycoprotein receptor.

2. The composition of claim 1, wherein the RNA interference polynucleotide is selected from the group consisting of: DNA, RNA, dsRNA, siRNA and microRNA.

3. The composition of claim 1, wherein the liver cells consist of hepatocytes.

4. The composition of claim 1, wherein the ASGPr ligand is reversibly linked to at least 90% of the primary amines on the plurality of melittin peptides.

5. The composition of claim 1, wherein said melittin peptide is selected from the group consisting of: seq.ID 1, seq.ID 7, seq.ID 11, seq.ID 51, seq.ID 57 and 58, seq.ID 92 and seq.ID 96.

6. The composition of claim 1, wherein the melittin peptide consists of D-amino acids.

7. The composition of claim 1, wherein L is disubstituted maleamate.

8. The composition of claim 1, wherein L is dipeptide-amidobenzyl carbamate.

9. The composition of claim 8, wherein the dipeptide is Glu-Gly, Ala-Cit, or Phe-Cit.

10. The composition of claim 7, further comprising polyethylene glycol (PEG) covalently attached to the amino terminus of the melittin peptide.

11. The composition of claim 7, further comprising an ASGPr ligand-PEG conjugate covalently attached to the amino terminus of the melittin peptide.

12. The composition of claim 8, further comprising polyethylene glycol (PEG) covalently attached to the amino terminus of the melittin peptide.

13. The composition of claim 8, further comprising an ASGPr ligand-PEG conjugate covalently attached to the amino terminus of the melittin peptide.

14. The composition of claim 1, wherein the ASGPr ligand is selected from the group consisting of: lactose, galactose, N-acetylgalactosamine (GalNAc), galactosamine, N-formylgalactosamine, N-propionylgalactosamine, N-N-butyrylgalactosamine and N-isobutyrylgalactosamine.

15. The composition of claim 1, wherein the RNAi is through a physiologically labile bond L²Is connected to A.

16. The composition of claim 15, wherein L²Orthogonal to L.

17. The composition of claim 1, wherein the galactose cluster consists of a trimer of N-acetylgalactosamine.

18. The composition of claim 1, wherein the hydrophobic group consists of cholesterol.

19. The composition of claim 1, wherein the composition is provided in a pharmaceutically acceptable carrier or diluent.

20. Use of a composition of claim 1 in the manufacture of a medicament for inhibiting expression of a gene expressed in a liver cell of a mammal, the composition comprising an RNAi polynucleotide having a sequence identical or close to identical to a partial sequence of the gene; the medicament is formulated for injection into the vessel of the mammal.

21. A method of making an RNA interference polynucleotide delivery composition, the method comprising:

a) formation of melittin peptide:

b) forming a plurality of uncharged masking agents each comprising an ASGPr ligand covalently linked to a disubstituted maleic anhydride or dipeptide amidobenzyl amine reactive carbonate;

c) modifying more than 80% of the primary amines on the population of melittin peptides with the masking agent of step b,

d) linking the RNA interference polynucleotide to a hydrophobic group having at least 20 carbon atoms or a galactose trimer;

e) the RNA interference polynucleotides and modified melittin peptides are provided in a solution suitable for in vivo administration.