Background of the Invention The present invention relates to new cationic lipids and intermediates therefor as well as their use for delivering polyanionic polymers into eukaryotic and prokaryotic cells.

Lipids that are useful for transfecting nucleic acids into cells and their use have been described (WO 94/05624; WO 94/00569; WO 93/24640; WO 91/16024; WO 90/11092; U.S. Patent Nos. 5,283,185, 5,171,687 and 5,286,634; Feigner et al L Biol. Sci. (1994) 2£9:2550; Nabel et al Proc. Natl. Acad. Sci. H .S.A (1993) 9Q:11307; Nabel et al Hum. Gene Ther. (1992) 3:649); Remy et al Bioconjugate Chem. (1994) 5:647-654.

Objects of the Invention

It is an object of the invention to provide cationic lipids and intermediates for making such lipids.

Another object of the invention is to provide cationic lipids that are suitable for delivering polyanionic polymers such as nucleic acids into cells in the presence of serum or blood.

Another object of the invention is to provide cationic lipids that are suitable for efficiently delivering polyanionic polymers such as nucleic acids into cells using cells at a cell confluency of about 50% to 100%.

Another object of the invention is to provide cationic lipids that are suitable for efficiently delivering a large amount of polyanionic polymers such as nucleic acids or peptides into cells.

Another object of the invention is to provide cationic lipids having improved pharmacological or other properties such as, improved storage stability, reduced toxicity or increased efficacy in the presence of serum.

Summary of the Invention The invention is directed to cationic lipids and intermediates in their synthesis having the structures F and G.

wherein each R ¹ is independently alkyl (10-22 C, i.e., having 10 to 22 carbon atoms) or a mono unsaturated alkenyl (10-22 C) group; each R ⁴ is independently hydrogen or is a protecting group; R5 is H, CH ₃ , CH(CH ₃ ) ₂ , CH ₂ CH(CH ₃ ) ₂ , CH(CH ₃ )CH ₂ CH ₃ , CH ₂ C ₆ H ₅ , CH2C ₆ H ₄ -P-OH, CH ₂ OH, CH(OH)CH ₃ , (CH ₂ ) ₄ NHR4, (CH ₂ ) ₃ NHR4 or

(CH ₂ ) ₃ NR4C(NH)NHR4;

R ¹¹ is alkyl; alkyl substituted with halogen, hydroxyl, S(0)2, O-alkyl or O- aryl; alkenyl; alkenyl substituted with halogen, hydroxyl, S(0)2, O-alkyl or O-aryl; aryl; aryl substituted with halogen, hydroxyl, alkyl, or alkyl substituted with halogen, hydroxyl, S(0)2, O-alkyl or O-aryl; heteroaryl; heteroaryl substituted with halogen, hydroxyl, alkyl, or alkyl substituted with halogen, hydroxyl, S(0)2,

O-alkyl or O-aryl; N(H)alkyl; N(alkyl)2 wherein each alkyl group is independently chosen; O-alkyl; O-aryl; alkylaryl; O-alkylaryl; alkylheteroaryl; or

O-alkylheteroaryl; R ¹⁵ is alkyl; alkyl substituted with halogen, hydroxyl, S(0)2, O-alkyl or O- aryl; alkenyl; alkenyl substituted with halogen, hydroxyl, S(0)2, O-alkyl or O-aryl; aryl; aryl substituted with halogen, hydroxyl, alkyl, or alkyl substituted with halogen, hydroxyl, S(0)2, O-alkyl or O-aryl; heteroaryl; heteroaryl substituted with halogen, hydroxyl, alkyl, or alkyl substituted with halogen, hydroxyl, S(0)2, O-alkyl or O-aryl; N(H)alkyl; N(alkyl)2 wherein each alkyl group is independently chosen; O-alkyl; O-aryl; alkylaryl; O-alkylaryl; alkylheteroaryl; or

O-alkylheteroaryl;

L is an integer having a value of 2, 3, 4, 5, 6 or 7;

m is an integer having a value of 0 or 1; n is an integer having a value of 0 or 1; provided that when L is 3 or 4, m is 0, n is 1, each R ¹ is independently alkyl (12-22C) or alkenyl (12-22C), and R ¹⁵ together with the C(O) to which it is linked, is a protecting group, then R is hydrogen; provided that when L is 3, m is 1, n is 1, each R ¹ is independently alkyl (12- 22C) or alkenyl (12-22C), and R ¹⁵ together with the C(O) to which it is linked, is a protecting group, then R ⁴ is hydrogen; and provided that when each R ¹ is independently alkyl (12-22C) or alkenyl (12- 22C), and R ¹¹ is a protecting group, then R is hydrogen; and the salts, solvates, resolved and unresolved enantiomers and diastereomers thereof.

Detailed Description of the Invention As used herein, and unless modified by the immediate context:

1. The terms alkyl and alkenyl mean linear, branched, normal and alicyclic hydrocarbons. Alkyl includes methyl, ethyl, propyl, cyclopropyl, cyclobutyl, isopropyl, n-, sec-, iso- and tert-butyl, pentyl, isopentyl, 1-methylbutyl, 1-ethylpropyl, neopentyl, and t-pentyl. Ranges of carbon atoms for a given group, such as alkyl (1-4C), mean alkyl groups having 1, 2, 3 or 4 carbon atoms can be present at the indicated position. Similarly, a group specified as alkyl (1- 8C), means alkyl groups having 1, 2, 3, 4, 5, 6, 7 or 8 carbon atoms can be present at the indicated position. Thus, for example, the terms alkyl and alkyl (1-8C) includes n-hexyl, cyclohexyl and structural and stereoisomers of n-hexyl. 2. Alkenyl means branched, normal or cyclic hydrocarbons containing at least 1 (generally 1, 2 or 3) cis or trans oriented conjugated or unconjugated double bond, including allyl, ethenyl, propenyl, isopropenyl, 1-, 2- and 3-butenyl, 1- and 2-isobutenyl and the like.

3. Aryl or heteroaryl means a resonant cyclic or fused polycyclic ring structure containing at least one 3-6 membered ring containing ring atoms that are solely carbon or carbon and one, two or three N-, S- or O- heteroatoms, including for example phenyl, 2- and 4-imidazolyl, 2-, 4- and 5-oxazolyl, 2-, 3-, 4- or 5-isoxazolyl, 2-, 3-, 4- or 5-furazanyl, 2-, 4- and 5-thiazolyl, 3-, 4- and 5- isothiazolyl, 3- and 4-pyrazolyl, 2-, 3- and 4-pyridinyl or 2-, 4- and 5-pyrimidinyl, 1-, 2-, 3- or 4- azetidine, 2-, 3-, 4-, or 5-thiophene, 2-, 3-, 4-, or 5-furanyl, 1-, 2-, 3-, 4- , or 5-pyrrolyl and analogs thereof in which a double bond has been shifted, e.g. 2H -pyrrole, or has been saturated, e.g. 2-pyrrolinyl or 3-pyrazolinyl. A nitrogen atom in a heteroaryl structure can be present as -N= or as -NH-. Aryl and

heteroaryl groups will generally comprise one, two or three 5-membered or 6- membered ring structures.

As used herein, halogen includes chlorine, fluorine, bromine and iodine. Groups such as heteroaryl having 1, 2, or 3 ring O, N or S atoms, do not include obviously unstable combinations such as peroxides.

Exemplary R ¹ have the structures -(CH )i ₉ CH ₃ , -(CH2)i CH ₃ , -(CH ₂ )i5CH ₃ , -(CH2)l4CH3, -(CH ₂ )ι ₃ CH ₃ , -(CH ₂ )ι ₂ CH3, -(CH ₂ )ιιCH ₃ , -(CH ₂ )ιθCH3, -(CH ₂ )5CH=CH(CH ₂ )7CH ₃ and -(CH ₂ ) ₈ CH=CH(CH ₂ ) ₇ CH ₃ . The alkenyl species are in a cis or trans configuration at the double bond. In general, each R ¹ on a given molecule will have the same structure, although they may be different. R ¹ can comprise, for example, alkyl (12-16C) groups. R ¹ typically is a normal alkane such as n-Ci8H ₃₇ , n-Ci6H ₃₃ or n-Ci ₄ H ₂ α.

Ordinarily, in compounds of structures F and G_, R ⁴ is H, although R is an amine protecting group in intermediates used to synthesize cationic lipids to be formulated with polyanions such as nucleic acids. When R is a protecting group then any conventional protecting group is useful. See for example Green et al., (op. cit.) and further discussion in the schemes.

R ⁵ typically is H, but also ordinarily may be CH ₃ , CH(CH ₃ ) ₂ , CH ₂ CH(CH ₃ ) ₂ , CH(CH ₃ )CH ₂ CH ₃ or CH ₂ C ₆ H ₅ . In some cases, R ¹⁵ or R ¹¹ , together with the group to which they are attached, e.g., when linked to an adjacent C(O) group, can be an amine protecting group. Thus, R ¹⁵ or R ¹¹ , together with the group to which they are attached, can comprise a protecting group or they may comprise a group that is not a protecting group. For example, where R ¹⁵ is -0-C(CH ₃ ) ₃ and is linked to C(O), both groups together comprise t-BOC when n in structure F is one. An alkyl-aryl R ¹⁵ or R ¹¹ can, for example, together with the C(O) group to which they are linked, comprise the protecting group FMOC, which contains a tricyclic aryl group. In such cases, the protecting group formed by R ¹¹ or R ¹⁵ and the group to which they are linked, will usually be differentially removable from R that are present on intermediates used to synthesize the final molecule.

However, where R ¹¹ or R ¹⁵ and the group to which they are linked, are the same as R ⁴ , the final molecule can be obtained by partial deprotection and separation of partially and fully deprotected species. In some cases, an R group such as t-BOC is more labile at some positions than at others. For example, t- BOC at R ⁴ in structure F compounds where L is 3, m is 1, n is 1 and R ⁵ is hydrogen, is more acid labile than t-BOC at R ¹⁵ -C(0)- and monoprotected species at R ¹⁵ -C(0)- is obtained by acid hydrolysis in 2-4N HC1 in organic solvent at room temperature for about 2 hours. Complete removal of t-BOC at R ¹⁵ -C(0)- is

accomplished by incubation in 2-4N HC1 at room temperature for about 16-24 hours.

Exemplary R ¹¹ and R ¹⁵ include (1) alkyl (1-8C), (2) alkyl (1-8C) substituted with halogen, hydroxyl, ether or S(0) ₂ ("substituted alkyl"), (3) alkenyl (1-8C), (4) alkenyl (1-8C) substituted with halogen, hydroxyl, ether or S(0) ₂ ("substituted alkenyl"), (5) aryl, (6) aryl substituted with halogen, hydroxyl, alkyl or substituted alkyl ("substituted aryl"), (7) heteroaryl, (8) heteroaryl substituted on a ring carbon with halogen, hydroxyl, alkyl or substituted alkyl, (9) heteroaryl that has 1, 2, or 3 ring carbons replaced with O, N or S, but where no more than a total of four O, N and S atoms are present, (10) N-alkyl (1-8C), (11) N-substituted alkyl (1- 8C), (12) O-alkyl (1-8C), (13) O-substituted alkyl (1-8C), (14) alkyl-aryl (1-4C alkyl), (15) alkyl-heteroaryl (1-4C alkyl), (16) alkyl-substituted aryl (1-4C alkyl), (17) alkyl¬ heteroaryl (1-4C alkyl) substituted on a ring carbon with halogen, hydroxyl, alkyl or substituted alkyl, (18) alkyl-heteroaryl (1-4C alkyl) that has 1, 2, or 3 ring carbons replaced with O, N, NH or S (including species where no more than a total of four or a total of no more than five O, N and S atoms are present).

Exemplary R ¹¹ and R ¹⁵ have the structures -CH ₃ , -CH ₂ CH ₃ , -CH ₂ CH CH ₃ , -C (CH3)CH3r -CH ₂ CH ₂ CH ₂ CH _3/ -CH(CH ₃ )CH ₂ CH ₃ , -CHCH^CH^CHs, -C(CH ₃ ) ₃ , -CH ₂ C(CH ₃ ) ₃ , -C6H5, -0-C(CH ₃ ) _3/ -0-CH(CH ₃ ) ₂ , -0-CH ₃ , -0-CH ₂ CH ₃ , -0-CH ₂ CH ₂ CH ₃ , -0-CH(CH ₃ )CH ₂ CH ₃ , -0-CH ₂ CH ₂ CH ₂ CH ₃ , -CF ₃ , -CH2-0-CHS, -CH ₂ -S(0) ₂ -CH ₃ , -NH(CH ₃ ), -NH(CH ₂ CHs), -NH(CH ₂ CH ₂ CH ₃ ), -NH(CH ₂ CH ₂ CH ₂ CH ₃ ), -NHC(CH ₃ ) ₃ , -N(alkyl) ₂ (including -N(CH ₃ ) ₂ , -N(CH ₂ CH ₃ ) ₂ , -N(CH ₂ CH ₂ CH ₃ ) ₂ , -N(CH ₂ CH ₂ CH ₂ CH ₃ ) ₂ , -N[C(CH ₃ ) ₃ ] ₂ ) and unsaturated analogs thereof and the like. In compounds where R ¹¹ or R ¹⁵ is -N(alkyl) ₂ , both alkyl groups will usually be the same, although they can differ. Exemplary R ¹¹ or R ¹⁵ are alkyl (having 1, 2, 3, 4, 5, 6, 7 or 8 carbon atoms), O-alkyl (having 1, 2, 3, 4, 5, 6, 7 or 8 carbon atoms), alkyl (C3-6), O-alkyl (C3-6), and the like.

Generally, R ¹¹ and R ¹⁵ groups together with the C(O) group to which they are linked are removable by hydrolysis, but it is also within the scope of this invention that these groups be stable to hydrolysis, e.g. in mildly acidic conditions such as pH 3-6.

In embodiments of this invention, compositions containing compounds of structures E or G_ are free of otherwise identical compounds in which the basic groups are unsubstituted with Rϋ, R-$, or R^ groups, as the case may be.

Alternatively, compositions containing compounds of structures F or G_ or free of otherwise identical compounds which do not contain any amino protecting substitutents. In other embodiments, the compositions contain less than about

1%, 0.5% or 0.1% by weight of such unsubstituted analogues in relation to the weight of the substituted congener.

L generally is normal propyl or butyl, but any normal alkyl having from 2 to 7 carbon atoms is satisfactory.

The variable m typically is 1.

The variable n generally is 1.

The usual embodiments are those of structure E where R is H, m is 1, and R ⁵ is as described immediately above.

The compounds of the invention include enriched or resolved optical isomers at any or all asymmetric atoms. For example, the chiral centers apparent from the depictions are provided as the chiral isomers or racemic mixtures. Both racemic and diasteromeric mixtures, as well as the individual optical isomers isolated or synthesized, substantially free of their enantiomeric or diastereomeric partners, are all within the scope of the invention. The racemic mixtures are separated into their individual, substantially optically pure isomers through well-known techniques such as, for example, the separation of diastereomeric salts formed with optically active adjuncts, e.g., acids or bases followed by conversion back to the optically active substances. In most instances, the desired optical isomer is synthesized by means of stereospecific reactions, beginning with the appropriate stereoisomer of the desired starting material. Methods and theories used to obtain enriched and resolved isomers have been described (see for example, J. Jacques et al, "Enantiomers, Racemates and Resolutions." Kreiger, Malabar, Florida, 1991).

Cationic lipids having the structure A, B. or C_

wherein R ² has the structure

wherein R ⁶ is NHR ⁴ , CH ₂ NHR or NR ⁴ C(NH)NHR ⁴ , are suitable compounds for efficiently transfecting polyanions into cells.

Intermediates that are useful for synthesizing these cationic lipids have the structure D or E

wherein R ⁷ is hydrogen or a protecting group, provided that for compounds containing both R ⁴ and R ⁷ , such as D or E wherein R ⁵ consists of a group such as (CH2) ₃ NR C(NH)NHR ⁴ , R ⁷ can be removed without removing R .

Positions designated "*" are carbon atoms having substituents in the R, S or RS configuration.

The invention lipids F and G_ form a complex with polyanions such as nucleic acids or peptides at least through attraction between the positively charged lipid and the negatively charged polyanion. The complexes may comprise multilamellar or unilamellar liposomes or other particles.

Hydrophobic interactions between the cationic lipids and the hydrophobic substituents in the polyanion such as aromatic and alkyl groups may also facilitate complex formation. The invention lipids efficiently deliver nucleic acids or peptides into cells in the presence of serum and thus are suitable for use in vivo or ex vivo.

The lipids and their intermediates are synthesized as shown in the schemes below.

SCHEME 1

o o

_R A- _OH R ⁶ QH _{I R} A _{0Rβ R} 5 Coupling Reagent R5

(DCC)

1 2

3

O remove R ^{7 H} 2 , _Ni ι>

-_ ^ | N(H ) ₂

R

remove R ⁴

The unprotected carboxyl group shown in structure 1 of scheme 1, is activated by reacting 1 , where R ⁷ is a protecting group, with a suitable activating group (R ⁸ OH) in the presence of a coupling reagent such as dicyclohexylcarbodiimide (DCC), diisopropylcarbodiimide (DIC), N- hydroxybenzotriazolephosphoryl chloridate (BOP-C1), isobutyl chloroformate, N-N'-carbonyldiimidazole or the like to form the activated ester group in 2 using previously described methods (see, for example, J. March, editor, Advanced Organic Chemistry, Wiley & Sons, third edition, 1985, p 348-351). Any amine

groups at R ⁵ will also be protected with an R protecting group. Such activating groups linked to the amino acid (-OR ⁸ ) function as a leaving group for synthesizing or 11. Suitable R ⁸ OH activating groups include N- hydroxysuccinimide (NHS), p-nitrophenol, pentachlorophenol, pentafluorophenol and the like. The activated ester is reacted with approximately one equivalent each of a secondary amine and a tertiary amine such as ethyldiisopropyl-amine or triethylamine (TEA) to yield structure 3_. The amino protecting group R ⁷ is removed to yield 4 which is coupled with a compound of structure 7 to yield 5, a protected lipid. The R amino protecting group(s) present on the protected lipid J5 is removed to yield the free lipid.

Compounds of structure 2 shown in scheme 1 wherein R ⁵ contains an amine group(s) will also have an R amine protecting group in synthesizing 5 so as to avoid forming adducts at the R ⁵ amino groups. In these compounds, the R ⁷ amine protecting group and the R amine protecting group will be different so that R ⁷ can be removed without removing R , i.e., R ⁷ and R ⁴ are different and can be differentially removed from a given molecule. In general, any R present in 2 will be the same as R in 2.

The R and R ⁷ amine protective groups will be selected from groups that have been described (see for example, T.W. Greene et al, editors, Protective Groups in Organic Chemistry, second edition, 1991, Wiley, p 309-405, p 406-412 and p 441-452). Amine protective groups such as benzyl carbamate (Cbz), p- methoxybenzyl carbamate (Moz), p-bromobenzyl carbamate, 9-fluorenylmethyl carbamate (FMOC), or 2,4-dichlorobenzyl will be used where an acid stable protective group is desired while protective groups such as t-butyl carbamate (t- BOC) or 1-adamantyl carbamate (Adoc) will be used where a base stable or nucleophile stable group is desired. Protective groups will be used to protect amine groups when coupling reactions are carried out such as in conversion of 1 to 2 or . to 5. Protective groups such as 2-(2'- or 4'-pyridyl)ethyl carbamate will be used where a group stable to catalytic palladium-carbon hydrogenation or to trifluoroacetic acid is desired. The R ⁴ group can thus be selected from a diverse group of known protective groups as needed. Exemplary R ⁴ and R ⁷ groups that can be present in 2 include the following pairs of protective groups. Other suitable R ⁴ and R ⁷ pairs are determined experimentally by routine methods using the relative reactivity information of the different amine protective groups described by Greene et al, supra at p 406-412 and p 441-452. Acceptable exemplary R ⁴ and R ⁷ pairs are shown below.

Lipids containing cholesterol are synthesized as shown in scheme 2.

SCHEME 2

chol.

TEA, CH ₂ CI ₂ 11

Cholesteryl chloroformate (Aldrich, Cat. No. C7,700-7) is coupled to ethylenediamine in organic solvent (CH ₂ C1 ₂ ) at 0-24° C to obtain 8_. S. is converted to the protected lipid intermediate 2 by reaction with 7 or is converted to 11 by reaction with 2. The protected lipid intermediates £ , H and 13 are deprotected as shown in scheme 2 to obtain free lipids 1Q_ and 14 or the partially deprotected lipid (when an amine group is present at R ⁵ ) 12.

Lipid intermediates of structure 11 contain both R and R ⁷ amine protective groups. As described for intermediates in scheme 1 that contain both R ⁴ and R ⁷ , the protective groups will be differentially removable from the protected lipid intermediate, H. The same pairs of differentially removable amine protective groups can be used as described. The R ⁴ group present in 7 will generally be the same as R present in 2 which permits removing all R from 13 using a single set of conditions.

Intermediates of structure HN(R ^! ) ₂ can be synthesized by reacting an acyl chloride of structure C1C(0)R ⁹ wherein R ⁹ is alkyl (11-21 C) or mono unsaturated alkenyl (11-21 C), with H ₂ NR ^! to obtain the intermediate HN(R ¹ )[C(0)R ⁹ ] which is reduced (using, for example, lithium aluminum hydride) to yield HN(R ¹ ) ₂ . The acyl chlorides are obtained by reaction of the free fatty acid with, for example,

oxalyl chloride, SOCl ₂ or PC1 ₃ . The H ₂ NR ¹ intermediate is obtained by reacting C1C(0)R ⁹ with ammonia gas (at about 0° C). In addition, many R ⁹ C(0)C1 chlorides and H2NR ¹ amines are available commercially (Aldrich Chemical, Kodak, K&K Chemicals).

Cationic lipids of structure F are synthesized as shown in schemes 3 and 3A and as described in detail in example 1, or by modification of the synthesis route used to prepare exemplary compounds 23-25.

SCHEME 3

IS /=\

HN(R ¹ ) ₂ >- VcH ₂ OCONHCH ₂ CON(R ¹ ) ₂

2Q : R ¹ = n-Cι ₈ H37; 21 : R ¹ = n-Cι ₂ H ₂ 5; 22 : R = n-Cι ₄ H ₂ 9

/=\ ^| -1 ₂ , &U h'SI t -CH ₂ OCONHCH ₂ CON(R ¹ ) ₂ *- H ₂ NCH ₂ CON(R ¹ ) ₂

^- 2Q 10% Pd/C, rt 23. : R ¹ = n-Cι ₈ H ₃ 7; 24 : R ¹ = n-Cι ₂ H ₂ 5; 25 : R ¹ = n-Cι H ₂₉

SCHEME 3A

1) coupling reagent, NHS

2) H ₂ NCHR ⁵ C(0)N(R ¹ ) ₂

remove R ⁷

remove R ⁴

Partially protected compound 6J) is fully protected using protecting group R ⁴ . R4 as shown in the first step represents an unreacted form of the protecting group. Thus, where R ⁴ is t-BOC, the compound used in the first step would be (t- BOC) ₂ 0 or if R ⁴ is FMOC, the compound used could be the chloride (FMOC-Cl). Such protecting groups have been described (Greene, ibid). The starting material 60 is obtained commercially, e.g. compound 4j> of example 21, or can be prepared from L-, D- or DL-arginine using known protecting groups R ⁴ and R ⁷ according to standard methods (Greene, ibid). 61 is coupled with NH ₂ CH ₂ C(0)N(R ¹ ) ₂ to yield the protected lipid £2 using a coupling reagent such as DCC, DIC or the like. The R ⁷ protecting group is then removed from __l to yield __ , followed by reaction of _a3_ with R ¹⁵ C(0)C1 to obtain the protected lipid compounds 6_4- The protected lipids are then deprotected to yield the final product 6J which can be conveniently obtained as a salt including a hydrogen halide (HC1, HBr, etc) or other biologically compatible or nontoxic salt. Partially deprotected intermediates can be obtained by the use of protecting groups that are removed under different conditions, e.g., t-BOC and CBZ. Partially deprotected

intermediates can also be obtained from a compound having two or more of a single type of protecting group by conducting a deprotection reaction for a period of time that is insufficient to fully deprotect the molecule and separating the different partially deprotected species by e.g., HPLC or TLC. Synthesis of analogous compounds containing a 4 methylene chain instead of the 3 methylene chain shown in scheme 3A, is accomplished using lysine as a starting material and reacting α-amino protected lysine with protected thiourea to obtain the analog of δ . This compound is then suitably protected and used as shown. Cationic lipids of structure G_ are synthesized as shown in scheme 4.

SCHEME 4

)R ¹¹ )C(NH)NHR ⁷

1) coupling reagent, NHS

2) H ₂ NCHR ⁵ C(0)N(R ¹ ) ₂

remove R ⁴ and R ⁷

Compound 60 is reacted with R ^π C(0)Cl to obtain 66, followed by substitution with the R ¹ substituted amine to obtain 67. The protecting groups present on 6.7 are deprotected in two steps by standard methods to obtain 68, either R ⁴ or R ⁷ first. Intermediates monoprotected with R ⁴ or R ⁷ are isolated using conventional methods. Compounds analogous to £0 but containing a single species of protecting group, such as t-BOC, at both protected positions can also be used as shown in scheme 4, and protecting group removal in the last step would be accomplished in one step for both groups using the same reaction condition.

To the extent any compound of this invention cannot be produced by one of the foregoing schemes other methods will be apparent to the artisan referring to conventional methods and the relevant teachings contained herein (see for instance Liotta et al. "Compendium of Organic Synthesis Methods" (John Wiley & Sons, New York), Vol. 1, Ian T. Harrison and Shuyen Harrison, 1971; Vol. 2, Ian T. Harrison and Shuyen Harrison, 1974; Vol. 3, Louis S. Hegedus and Leroy

Wade, 1977; Vol. 4, Leroy G. Wade, Jr., 1980; Vol. 5, Leroy G. Wade, Jr., 1984; and Vol. 6, Michael B. Smith; March, J., "Advanced Organic Chemistry, Third Edition", (John Wiley & Sons, New York, 1985); Saul Patai, "The Chemistry of the Amino Group. Volume 4" (Interscience, John Wiley & Sons, New York, 1968); Saul Patai, "Supplement F. The Chemistry of Amino, Nitroso and Nitro Compounds and their Derivatives, Parts 1 and 2" (Interscience, John Wiley & Sons, New York, 1982); as well as "Comprehensive Organic Synthesis. Selectivity, Strategy & Efficiency in Modern Organic Chemistry. In 9 Volumes", Barry M. Trost, Editor-in-Chief (Pergamon Press, New York, 1993 printing).

Exemplary invention cationic lipids and their synthetic intermediates have the structures

wherein R ¹ and R have the meanings given, and R ³ is a substituted cholesteryl moiety having the structure

The salts include pharmaceutically or physiologically acceptable non-toxic salts of these compounds. Such salts may include those derived by combination of appropriate cations such as alkali and alkaline earth metal ions or ammonium and quaternary amino ions with the acid anion moiety of the phosphate or phosphorothioate acid group present in polynucleotides. In addition salts may be formed from acid addition of certain organic and inorganic acids with basic centers of the purine, specifically guanine, or pyrimidine base present in polynucleotides. Suitable salts of the invention cationic lipids include acid addition salts such as HC1, HBr, HF, HI, H ₂ Sθ ₄ , and trifluoroacetate. The salts may be formed from acid addition of certain organic and inorganic acids, e.g., HC1, HBr, H2Sθ4 ₇ amino acids or organic sulfonic acids, with basic centers, typically amines, or with acidic groups. The compositions herein also comprise compounds of the invention in their un-ionized, as well as zwitterionic forms. Cationic lipid-polyanionic polymer complexes are formed by preparing lipid particles consisting of either (1) an invention cationic lipid or (2) an invention cationic lipid-colipid mixture, followed by adding a polyanionic polymer to the lipid particles at about room temperature (about 18 to 26° C). Ordinarily the invention lipid is a lipid of structures F or G, in which R is H, i.e. in which the terminal amino is substituted with R ¹¹ or R ¹⁵ as the case may be. If the R ¹¹ or R ¹⁵ group is not stable to long-term storage, e.g. t-BOC at low pH, then one may select an R ¹¹ or R ¹⁵ group that is less susceptible to hydrolysis (e.g. methyl, isopropyl, O-methyl or O-isopropyl), or the lipid is formulated under conditions that will enhance storage stability. In the case of t-BOC-substituted compounds this will be adjustment of the storage pH to neutral or mildly alkaline conditions, e.g., pH 7-8. In general, conditions will be chosen that are not conducive to deprotection when R ¹¹ or R ¹⁵ are protecting groups. The mixture is then allowed to form a complex over a period of about 10 min to about 20 hours, with about 15 to 60 min most conveniently used. The complexes may be formed over a longer period, but additional enhancement of transfection efficiency will usually not be gained by a longer period of complexing. A

phospholipid such as DOPE is generally used as a colipid with the invention lipids to enhance the efficiency of polyanion delivery into cells. Additional colipids that are suitable for preparing lipid complexes with the invention cationic lipids are dimyristoylphosphatidylethanolamine, dipalmitoylphosphatidyl-ethanolamine, palmitoyloleoylphosphatidylethanolamine, cholesterol, distearoylphosphatidylethanolamine, phosphatidylethanolamine covalently linked to polyethylene glycol and mixtures of these colipids.

The invention cationic lipids can be used without a colipid but are preferably used with a colipid to obtain a high transfection efficiency. The optimal cationic lipid:colipid ratio for a given invention cationic lipid is determined by mixing experiments to prepare lipid mixtures for complexing with a polyanion using cationic lipid:colipid ratios between about 1:0 and 1:10. Methods to determine optimal cationic lipid:colipid ratios have been described (see, for example, Feigner, infra). Each lipid mixture is optionally tested using more than one nucleic acid-lipid mixture having different nucleic acid:lipid molar ratios to optimize the nucleic acid:lipid ratio.

Suitable molar ratios of invention lipid olipid are about 0.1:1 to 1:0.1, 0.2:1 to 1:0.2, 0.4:1 to 1:0.4, or 0.6:1 to 1:0.6. Lipid particle preparations containing increasing molar proportions of colipid were found to deliver increasing amounts of nucleic acid to transfected cells as the proportion of colipid increased.

In addition, the lipids of structures F or G_ can be used together in admixture, or different embodiments of either or both structures are used in admixture, with or without colipid. The amount of polyanion, present as an oligonucleotide, delivered to a representative cell by at least some of the lipids was found to be significantly greater than the amount delivered by commercially available transfection lipids {Lipofectin™ (Gibco/BRL), Transfectam™ (Promega) or Lipofectamine™ (Gibco/BRL)} for all cell lines where comparisons were made. The difference in transfection efficiency between lipids described herein and commercially available lipids is usually most pronounced when the transfections are done in the presence of medium containing serum. The amount of polyanion delivered into cells was estimated to be about 2- to 100-fold greater for lipids 3J 3JL and 3J> based on the observed fluorescence intensity of transfected cells after transfection using a fluorescently labeled oligonucleotide. The cationic lipids described herein also transfect some cell types that are not detectably transfected by commercial lipids, particularly where the transfection is conducted in the presence of serum.

The cationic lipids described herein also differed from commercially available lipids by efficiently delivering a polyanion (oligonucleotide) into cells in tissue culture over a range of cell confluency from about 50 to 100%. Most commercially available lipids require cells that are at a relatively narrow confluency range for optimal transfection efficiency. For example, Lipofectin™ requires cells that are 70-80% confluent for transfecting the highest proportion of cells in a population. The invention lipids could be used to transfect cells that are about 10-50% confluent, but toxicity of the lipids was more pronounced, relative to that seen using cells that are about 50-100% confluent. In general, the invention lipids transfected cells that were about 60-100% confluent with minimal toxicity and optimal efficiency. Confluency ranges of 60-95% or 60-90% are thus convenient for transfection protocols with most cell lines in tissue culture.

The invention cationic lipids complexed with an oligonucleotide, were used to transfect cells in tissue culture. Similar lipids complexed with plasmid DNA were used to transfect cells with RNA or DNA in mice and the RNA and the DNA encoded gene products were expressed in the transfected cells.

Liposomes or complexes consisting of the invention cationic lipids and a colipid are conveniently prepared by first drying the lipids in solvent (usually chloroform) under reduced pressure (spin vac in 1.5 mL polypropylene tubes for small volumes (about 100 μL) or rotovap in round bottom flasks for larger volumes, e.g. 10 mL in 100 mL flask). The lipids are then hydrated and converted to liposomes or lipid complexes by adding water or low ionic strength buffer (less than about 200 mM total ion concentration) followed by agitating (by vortexing and/or sonication) and/or freeze/thaw treatments.

The invention lipid-polyanion complexes are believed to form micelles or liposomes of about 40 to 600 nm in diameter. Light scattering experiments using a 1:1 molar ratio of cationic lipid and DOPE colipid (prepared by sonication as described in Example 5 below) showed two peaks corresponding to particles of about 66 nm and about 260 nm in diameter, with lesser amounts of particles above, below and between these sizes.

The invention cationic lipids 3_P_ or 3JL and DOPE colipid (1:1) were prepared by sonication and then filtered using 200, 100 or 50 nm filters to obtain particles less than about 200 nm in diameter, less than about 100 nm and less than about 50 nm. Transfection efficiency using 200 nm filtered or 100 nm filtered preparations were the most efficient (and equally efficient) with regard to both the proportion of cells transfected and the amount of nucleic acid delivered per cell. The 50 nm filtered preparations were about 40-50% as efficient as the

preparations containing larger particles. The size of the particles or micelles in a given preparation will vary depending on the preparation method. Sonicating cationic lipid-colipid mixtures will provide smaller micelles and vortexing will provide larger micelles (Feigner T. Biol. Chem. (1994) 2^9:2550-2561). Such micelles are believed to transfer the polyanion into the cytoplasm of a eukaryotic or prokaryotic cell by pinocytosis, endocytosis and/or by direct fusion with the plasma membrane.

The complexes of this invention are then used to transfect one or more cell lines or administered in vivo to animals to determine the efficiency of transfection obtained with each preparation. The invention lipid-colipid complexes in medium were used at a concentration between about 0.5 and 20 μg/mL, with typical transfections using about 1.0 to 15 μg/mL of lipid. If the nucleic acid encodes a polypeptide, it also may comprise a selectable (such as neomycin phosphotransferase) or detectable (such as β-galactosidase) marker or gene that will serve to allow measuring or estimating the efficiency of transfection. Polyanions will have at least two negative charges per molecule to facilitate complex formation with the positively charged cationic lipid. Polyanions typically will have at least 5 charges per molecule or at least 10 charges per molecule. In general, oligonucleotides will have at least 2, 4, 6, 10, 14 or more charges to facilitate complex formation. However, nucleic acids containing two, three, four or more adenine residues (e.g., 4 to 10, 5 to 15 or 6 to 35 adenine residues) and no charges associated with the internucleotide linkages can be transfected with invention cationic lipids containing arginine (such as or 3.5) because arginine and adenine form noncovalent complexes with each other.

As used herein, polynucleotide means single stranded or double stranded DNA or RNA, including for example, oligonucleotides (which as defined herein, includes DNA, RNA and oligonucleotide analogs) and plasmids. In general, relatively large nucleic acids such as plasmids or mRNAs will carry one or more genes that are to be expressed in a transfected cell, while comparatively small nucleic acids, i.e., oligonucleotides, will comprise (1) a base sequence that is complementary (via Watson Crick or Hoogsteen binding) to a DNA or RNA sequence present in the cell or (2) a base sequence that permits oligonucleotide binding to a molecule inside a cell such as a peptide, protein or glycoprotein. Exemplary RNAs include ribozymes and antisense RNA sequences that are complementary to a target RNA sequence in a cell.

Polynucleotides include single stranded unmodified DNA or RNA comprising (a) the purine or pyrimidine bases guanine, adenine, cytosine,

thymine and/or uracil; (b) ribose or deoxyribose; and (c) a phosphodiester group that linkage adjacent nucleoside moieties. Polynucleotides include oligonucleotides which typically comprise 2 to about 100 or 3 to about 100 linked nucleosides. Typical oligonucleotides comprise size ranges such as 2-10, 2-15, 2-20, 2-25, 2-30, 2-50, 8-20, 8-30 or 2-100 linked nucleotides.

Oligonucleotides are usually linear with uniform polarity and, when regions of inverted polarity are present, such regions comprise no more than one polarity inversion per 10 nucleotides. One inversion per 20 nucleotides is typical. Oligonucleotides can also be circular, branched or double-stranded. Antisense oligonucleotides generally will comprise a sequence of about from 8 to 30 bases or about 8 to 50 bases that is substantially complementary to a cognate DNA or RNA base sequence present in the cell. The size of nucleic acid that is delivered into a cell using the invention lipids is limited only by the size of molecules that reasonably can be prepared and thus DNA or RNA that is 0.1 to 1 Kilobase (Kb), 1 to 20 Kb, 20 Kb to 40 Kb or 40 Kb to 1,000 Kb in length can be delivered into cells. Polynucleotides also include DNA or RNA comprising one or more covalent modifications. Covalent modifications include (a) substitution of an oxygen atom in the phosphodiester linkage of an polynucleotide with a sulfur atom, a methyl group or the like, (b) replacement of the phosphodiester group with a nonphosphorus moiety such as -0-CH ₂ -0-, -S-CH ₂ -0- or -0-CH ₂ -S-, and (c) replacement of the phosphodiester group with a phosphate analog such as -0-P(S)(0)-0-, -0-P(S)(S)-0-, -0-P(CH ₃ )(0)-0- or -O-P(NHR ^!0 )(O)-O- where R ¹ " is alkyl (Ci-6), or an alkyl ether (Cι_ ₆ ). Oligonucleotides include oligomers having a substitution at about 10% to 100% or 20 to 80% of the phosphodiester groups in unmodified DNA or RNA. Oligonucleotides include covalent modification or isomers of ribose or deoxyribose such as morpholino, arabinose, 2'-fluororibose, 2'-fluoroarabinose, 2'-0-methylribose or 2'-0-allylribose. Oligonucleotides and methods to synthesize them have been described (for example see: PCT/US90/03138, PCT/US90/06128, PCT/US90/06090, PCT/US90/06110, PCT/US92/03385, PCT/US91/08811, PCT/US91/ 03680, PCT/US91/06855, PCT/US91/01141, PCT/US92/10115, PCT/US92/10793, PCT/US93/05110, PCT/US93/05202, PCT/US92/ 04294, WO86/05518, WO89/12060, WO91/08213, WO90/15065, WO91/15500, WO92/02258, WO92/20702, WO92/20822, WO92/20823, U.S. Application Serial Nos. 07/864,873, 08/123,505 and 08/050,698, U.S. Patent No. 5,214,136 and Uhlmann Chem Rev (1990) 9_Q:543).

Oligonucleotides are usually linear with uniform polarity and, when regions of inverted polarity are present, such regions comprise no more than one polarity

inversion per 10 nucleotides. One inversion per 20 nucleotides is typical. Oligonucleotides can be circular, branched or double-stranded.

Linkage means a moiety suitable for coupling adjacent nucleomonomers and includes both phosphorus-containing moieties and non phosphorus- containing moieties such as formacetal, thioformacetal, riboacetal and the like. A linkage usually comprises 2 or 3 atoms between the 5' position of a nucleotide and the 2' or 3' position of an adjacent nucleotide.

A purine or pyrimidine base means a heterocyclic moiety suitable for incorporation into an oligonucleotide. It can be in the α or β anomer configuration. Purine or pyrimidine bases are moieties that bind to complementary nucleic acid sequences by Watson-Crick or Hoogsteen base pair rules. Bases need not always increase the binding affinity of an oligonucleotide for binding to its complementary sequence at least as compared to bases found in native DNA or RNA. However, such modified bases preferably are not incorporated into an oligomer to such an extent that the oligonucleotide is unable to bind to complementary sequences to produce a detectably stable duplex or triplex. Purine or pyrimidine bases usually pair with a complementary purine or pyrimidine base via 1, 2 or 3 hydrogen bonds. Such purine or pyrimidine bases are generally the purine, pyrimidine or related heterocycles shown in formulas 15 - if..

wherein R ³⁵ is H, OH, F, Cl, Br, I, OR ³⁶ , SH, SR ³ 6, NH2, or NHR ³⁷ ;

R ³ 6 is Ci - C6 alkyl (including CH3, CH2CH3 and C3H7), CH2CCH (2- propynyl) and CH2CHCH2; R ³⁷ is Cl - C6 alkyl including CH3, CH2CH3, CH2CCH, CH2CHCH2, C3H7;

RS8 is N, CF, CC1, CBr, CI, CR ³⁹ or CSR ³⁹ , COR ³⁹ ; . R ³⁹ is H, Ci - C9 alkyl, C2 - C9 alkenyl, C2 - C9 alkynyl or C7 - C9 aryl-alkyl unsubstituted or substituted by OH, O, N, F, Cl, Br or I including CH3, CH2CH3, CHCH2, CHCHBr, CH2CH2C1, CH2CH2F, CH2CCH, CH2CHCH2, C3H7, CH2OH, CH2OCH3, CH2OC2H5, CH2OCCH, CH2OCH2CHCH2, CH2C3H7, CH2CH2OH, CH2CH2OCH3, CH2CH2OC2H5, CH2CH2OCCH, CH2CH2OCH2CHCH2, CH2CH2OC3H7;

R 0 is N, CBr, CI, CC1, CH, C(CH ₃ ), C(CH ₂ CH ₃ ) or C(CH ₂ CH ₂ CH ₃ ); R ⁴¹ is N, CH, CBr, CCH ₃ , CCN, CCF3, CC≡CH or CC(0)NH2; R ⁴² is H, OH, NH2, SH, SCH3, SCH2CH3, SCH2CCH, SCH2CHCH2, SC3H7,

NH(CH3), N(CH3)2, NH(CH2CH3), N(CH2CH3)2, NH(CH2CCH), NH(CH2CHCH2), NH(C3H7) or halogen (F, Cl, Br or I);

R ⁴³ is H, OH, F, Cl, Br, I, SCH3, SCH2CH3, SCH2CCH, SCH2CHCH2, SC3H7, OR ¹⁶ , NH2, or NHR ³⁷ ; and R ⁴⁴ is O, S or Se.

Exemplary bases include adenine, cytosine, guanine, hypoxanthine, inosine, thymine, uracil, xanthine, 2-aminopurine, 2,6-diaminopurine, 5-(4- methylthiazol-2-yl)uracil, 5-(5-methylthiazol-2-yl)uracil, 5-(4-methylthiazol-2- yl)cytosine, 5-(5-methylthiazol-2-yl)cytosine and the like. Also included are alkylated or alkynylated bases having substitutions at, for example, the 5 position of pyrimidines that results in a pyrimidine base other than uracil, thymine or cytosine, i.e., 5-methylcytosine, 5-(l-propynyl) cytosine, 5- (l-butynyl)cytosine, 5-(l-butynyl)uracil, 5-(l-propynyl)uracil and the like. Base analogs and their use in oligomers have been described (see for example, U.S. AppUcation Serial No. 08/123,505; US92/10115; US91/08811; US92/09195; WO 93/10820; WO 92/09705; WO 92/02258; Nikiforov, T. T., et al, Tet Lett (1992) 22:2379-2382; Clivio, P., et al, Tet Lett (1992) 22:65-68; Nikiforov, T. T., et al, let Lett (1991) 22:2505-2508; Xu, Y.-Z., et al, Tet Lett (1991) 22:2817-2820; Clivio, P., et ^' al, Tet Lett (1992) 22:69-72; Connolly, B. A., et al, Nucl Acids Res (1989) 17:4957- 4974).

Nucleic acids complexed with the invention lipids may comprise nucleic acids encoding a therapeutic or diagnostic polypeptide. Examples of such polypeptides include histocompatibility antigens, cell adhesion molecules,

cytokines, antibodies, antibody fragments, cell receptor subunits, cell receptors, intracellular enzymes and extracellular enzymes or a fragment of any of these. The nucleic acids also may optionally comprise expression control sequences and generally will comprise a transcriptional unit comprising a transcriptional promoter, an enhancer, a transcriptional terminator, an operator or other expression control sequences.

Nucleopolymers (i.e., nucleic acids, oligonucleotides or oligonucleotide analogs) used to form complexes for transfecting a cell may be present as more than one expression vector or more than one oligonucleotide. Thus, 1, 2, 3 or more different expression vectors and/or oligonucleotides are delivered into a cell as desired. Expression vectors will typically express 1, 2 or 3 genes when transfected into a cell, although many genes may be present such as when a herpes virus vector or a yeast artificial chromosome is delivered into a cell. The ratio of each nucleopolymer in a lipid complex relative to each other can be selected as desired. Expression vectors that are introduced into a cell can encode selectable markers (E coli neomycin phosphotransferase, thymidine kinase from a herpesvirus, E coli xanthine-guanine phosphoribosyl-transferase, and the like) or biologically active proteins such as metabolic enzymes or functional proteins (such as immunoglobulin genes, cell receptor genes, cytokines (such as IL-2, IL-4, GM-CSF, γ-INF and the like), genes that encode enzymes that mediate purine or pyrimidine metabolism and the like).

Methods to prepare lipid-nucleic acid complexes and methods to introduce the complexes into cells in vitro and in vivo have been described (see for example, 5, 283,185; 5,171,678; WO 94/00569; WO 93/24640; WO 91/16024; Feigner T Biol Chem (1994) 2£9:2550; Nabel Proc Natl Acad Sci fUSA^ (1993) 9Q:11307;

Nabel Human Gene Ther (1992) 3:649; Gershon Biochem (1993) 22:7143; Strauss EMBO T (1992) 11:417).

Applications The invention lipids are useful for delivering polyanions, polypeptides or nucleopolymers into cells. The invention lipids can be used to deliver an expression vector into a cell for manufacturing or therapeutic use. The expression vectors can be used in gene therapy protocols to deliver a therapeutically useful protein to a cell or for delivering nucleic acids encoding molecules that encode therapeutically useful proteins or proteins that can generate an immune response in a host for vaccine or other immunomodulatory purposes according to known methods (see for example, 5,399^46, 5,336,615, WO 94/21807, WO 94/12629). The vector-transformed cell

can be used to produce commercially useful cell lines, such as a cell line for producing therapeutic proteins or enzymes (erythropoietin, and the like), growth factors (human growth hormone, and the like) or other proteins. The invention lipid-nucleic acid complexes can be used to construct cell lines for gene therapy applications in subjects such as humans or other species including murine, feline, bovine, equine, ovine or non human primate species. The invention lipids can be used in the presence of serum and will thus deliver polyanions into cells in tissue culture medium containing serum in vitro or in an animal in vivo. The invention lipids complexed with nucleopolymers can be used in antisense inhibition of gene expression in a cell by delivering an antisense oligonucleotide into the cell (see for example, Wagner Science (1993) 260:1510; WO 93/10820). Such oligonucleotides will generally comprise a base sequence that is complementary to a target RNA sequence that is expressed by the cell. However, the oligomer may regulate intracellular gene expression by binding to an intracellular nucleic acid binding protein (Clusel Nucl Acids Res (1993) 21:3405) or by binding to an intracellular protein or organelle that is not known to bind to nucleic acids (WO 92/14843). A cell that is blocked for expression of a specific gene(s) is useful for manufacturing and therapeutic applications. Exemplary manufacturing uses include inhibiting protease synthesis in a cell to increase production (i.e., reduce target protein degradation caused by the protease) of a protein for a therapeutic or diagnostic application. Exemplary therapeutic applications include inhibiting synthesis of cell surface antigens (histocompatibility antigens, such as MHC class II genes, and the like) to reduce rejection and /or to induce immunologic tolerance of the cell either after it is implanted into a subject or when the cell is transfected in vivo.

The following examples illustrate but do not limit the invention.

Example 1 A. Compound 19

Carbobenzyloxyglycine (1.05 g, 5 mmole) in THF (30 mL), N- hydroxysuccinimide (0.58 g, 5 mmole) and DCC (1.1 g, 5.3 mmole) were stirred at room temperature overnight. The precipitate was filtered off, washed with methylene chloride. The combined organic solution was concentrated to dryness, yielding compound 1£, and used for the next reaction without further purification.

B. Compounds 20-22

Compound 20: Compound IS was dissolved in methylene chloride (30 mL) containing TEA (1 mL), following by addition of distearylamine (2.6 g, 5 mmole). After 5 hr at room temperature, the reaction mixture was washed with water, dried over Na ₂ Sθ ₄ , concentrated. The residue was purified by flash column chromatography on silica gel, eluted with methylene chloride to afford the desired product, 20, as a colorless liquid 1.60 g; yield 45% (2 steps). NMR (CDCI ₃ ): δ 7.20-7.40 (m, 5 H), 5.85 (b, IH); 5.12 (s, 2H), 4.00 (d, 2 H), 3.15 (t, 2H), 3.31 (t, 2H), 1.42-1.60 (m, 4H), 1.10-1.40 (m, 60H), 0.88 (t, 6H).

Compound 21: yield 57% (2 steps). NMR (CDCI ₃ ): δ 7.20-7.40 (m, 5 H), 5.85 (t, IH); 5.12 (b, 2H), 4.00 (d, 2 H), 3.31 (t, 2H), 3.14 (t, 2H), 1.40-1.60 (m, 4H), 1.20-1.40 (m, 36H), 0.88 (t, 6H). Compound 22. NMR (CDCI ₃ ): δ 7.20-7.40 (m, 5 H), 5.82 (b, IH); 5.12 (b, 2H),

4.00 (d, 2 H), 3.31 (t, 2H), 3.14 (t, 2H), 1.40-1.60 (m, 4H), 1.10-1.40 (m, 44H), 0.88 (t, 6H).

20_ : R1 = n-C ₁₈ H ₃ 7; 21 ■ R ¹ = n-Cι ₂ H ₂₅ ; 22 : R1 = n-C ₁₄ H ₂₉

C. Compounds 23-25

Compound 23.: Compound 20 (1.60 g, 2.24 mmole) in methylene chloride (7 mL), and ethanol (14 mL) was hydrogenated at 50 psi in the presence of 10% Pd/C for 8 hr. The catalyst was filtered off through a Celite pad. The filtrate was concentrated to afford a yellowish wax-like solid (1.27 g), 23.. NMR (CDC1 ₃ ): δ 3.51 (S, 2H), 3.48 (b, 2H); 3.30 (t, 2H), 3.15 (t, 2H), 1.45-1.51 (m, 4H), 1.10-1.40 (m, 60H), 0.88 (t, 6H).

Compound 24: yield 93%. NMR (CDCI3): δ 8.60 (b, 2H), 3.94 (S, 2H); 3.30 (t, 2H), 3.10-3.20 (m, 2H), 1.40-1.60 (m, 4H), 1.10-1.40 (m, 36H), 0.88 (t, 6H).

Compound 25: yield 94%. NMR (CDC1 ₃ ): δ 8.50 (b, 2H), 4.95 (b, 2H); 3.28 (b, 2H), 3.14 (b, 2H), 1.40-1.60 (m, 4H), 1.10-1.40 (m, 44H), 0.85 (t, 6H).

2£ : R ¹ = n-Cι ₈ H ₃₇ ; 24 : R ¹ = n-C ₁₂ H ₂₅ ; 25 : R1 = n-Cι ₄ H ₂₉

D. Compound 26 et-BOC-Ornithinel L-Ornithine.HCl (0.84 g, 5 mmole) was dissolved in H ₂ 0 (15 mL) containing NaOH (0.6 g, 15 mmole), followed by addition of a THF (15 mL) solution of (t-BOC) ₂ 0 (2.18 g, 10 mmole). The resulting mixture was stirred at room temperature overnight, then was neutralized with 1 N HC1 aqueous solution (pH 4), extracted with CH ₂ C1 ₂ (50 mL) three times. The combined organic phase was dried over Na ₂ Sθ ₄ , concentrated, affording the title compound 26_, 1.00 g, 60%.

E. Compounds 27-28

Compound 27: Compound ϋ (0.11 g. 0.33 mmole), 22 (0.192 g, 0.33 mmole) and DIC (46 mg, 0.36 mmole) were dissolved in CH ₂ CI ₂ (10 mL). After being stirred 16 hr at room temperature, the reaction was washed with water, dried, and concentrated. The residue was purified by flash column chromatography on silica gel, eluted with 3% CH ₃ OH in CH ₂ CI ₂ (product came off column at frs 4-11, 8 mL/fr), affording 0.2 g (69%) of product 27. MNR (CDCls): δ 7.04 (b, IH), 5.10 (m, IH), 4.65 (b, IH), 4.22 (b, IH), 4.02 (d, 2H), 3.30 (t, 2H), 3.10-3.25 (m, 4H), 1.10-1.60 (m, 86H), 0.88 (t, 6H).

Compound 2__ was prepared from compounds 26 and .24 in the same manner in a 26% yield. NMR (CDCI ₃ ): δ 7.05 (b, IH), 5.06 (b, IH), 4.62 (b, IH), 4.20 (b, IH), 4.02 (d, 2H), 3.34 (t, 2H), 3.10-3.20 (m, 4H), 1.20-1.60 (m, 62H), 0.88 (t, 6H).

t-BOC-HN

V ^• CONHCH ₂ CON(R ) ₂

2J> ₊ 2_3_ (or 24) _(C H ₂₎₃ NH-t-BOC

27 : R1 = n-Cι ₈ H ₃₇ ; 28 : R ¹ = n-C-| ₂ H ₂₅

F. Compound 29. (trifluoroacetic acid salt. TFA

Compound 2Z (0.2 g) was treated with trifluoroacetic acid (exc) at room temperature for 20 min. The reaction mixture was evaporated to dryness affording a wax-like off white solid, 0.19 g. NMR (DMSO-d ₆ ): δ 8.58 (t, IH), 8.18 (b, 2H), 7.75 (b, 2H), 3.95 (d, 2H), 3.90 (m, IH), 3.10-3.25 (m, 4H), 2.70-2.82 (m, 2H), 1.10-1.60 (m, 68H), 0.88 (t, 6H).

27: R ¹ = n-C ₁₈ H ₃₇ 29: R ¹ = n-C ₁₈ H ₃₇

Compound 27 (1.63 g; 1.82 mmole) was dissolved in 1,4-dioxane 5 mL and treated with 4 N HC1 in 1,4-dioxane (5 mL). After 1.5 hr at room temperature, the reaction mixture was concentrated and azeotroped with CH ₃ CN (5 times). The product weighed 1.30 g.

H. Compound 31 (HC1 salt

Compound 21 (HO salt), was prepared from compound 2S in the same manner as compound 30.

CONHCH ₂ CON(R ¹ ) ₂

.2HCI 30: R ¹ = n-C ₁₈ H ₃₇ 31: R ¹ = n-C ₁₂ H ₂₅

Example 2 A-l. Compound 32

L-t-BOC-Arg (t-BOC) ₂ -OH, L-Arginine protected with t-BOC group, was purchased from BACHEM Bioscience Inc., Cat. No. A2935 (0.50 g; 1.05 mmole),

compound 22 (0.58 g; 1 mmole), DIC (0.145 g; 1.15 mmole) were dissolved in CH ₂ C1 ₂ (30 mL). The reaction mixture was stirred at room temperature overnight, washed with H ₂ 0, dried over Na ₂ Sθ4, concentrated, and purified by flash column chromatography on silica gel, affording 0.74 g of product 22 in 72% yield. NMR (CDCI ₃ ): δ 9.20-9.50 (m, 2H), 7.20 (b, IH), 5.50 (d, IH), 4.23 (m, IH), 4.02 (m, 2H), 3.90 (t, 2H), 3.35 (m, 2H), 3.19 (t, 2H), 1.10-1.70 (m, 95H), 0.88 (t, 6H). A-2. Compound 33

Compound 22: yield 37.5%. NMR (CDC1 ₃ ): δ 9.15-9.50 (b, 2H), 7.20 (b, IH), 5.20 (d, IH), 4.23 (b, IH), 4.02 (dd, 2H), 3.89 (t, 2H), 3.26-3.31 (m, 2H), 3.14 (t, 2H), 1.60-1.70 (m, 4H), 1.10-1.60 (m, 75H), 0.88 (t, 6H).

2 HCI

24 hours room temperature

B-l. Compound 34 (HCI salt)

Compound 52 (0.6 g; 0.58 mmole) was dissolved in dioxane (4 mL) and treated with 4 N HCI in dioxane (4 mL) at room temperature for 1.5 hr. The

reaction mixture was concentrated, azeotroped with CH ₃ CN twice to yield 34. NMR (DMSO-Clό): δ 9.15-9.50 (m, 2H), 8.25 (6.4H), 3.80-4.05 (m, IH), 3.15-3.30 (m, 4H), 1.10-1.70 (m), 0.88 (t, 6H). B-2. Compound 35 (HCI salt) Compound 25 was synthesized in the same manner as 5 using 33 as a starting material. NMR (CDCI3): δ 4.10 (b, IH), 3.78 (m, 2H), 3.65 (m, 2H), 3.20- 3.35 (b, 4H), 1.10-1.70 (m), 0.88 (t, 6H).

Example 3 A. Compound 36

A pyridine solution (20 mL) of disterylamine (2.62 g; 5 mmole) and succinic anhydride (1.5 g; 15 mmole) was stirred at room temperature for 30 hr. The reaction mixture was concentrated to dryness. The residue was partitioned between CH ₂ C1 ₂ and 1 M TEAB (triethylammonium bicarbonate) aqueous solution. The organic phase was isolated, dried, concentrated to afford a yellowish wax-like of product, 3.38 g, 98% yield. NMR (CDC1 ₃ ): δ 3.20-3.340 (2t, 4H), 2.65 (s, 4H), 1.40-1.60 (m, 4H), 1.20-1.40 (m, 60H), 0.88 (t, 6H). B. Compound 37

To a THF solution (10 mL) of compound 2J. (0.62 g; 1 mmole), N- hydroxysuccimide (0.13 g; 1.1 mmole) and DMAP (20 mg) was added DIC (0.14 g; 1.1 mmole). The resulting solution was stirred at room temperature overnight. The reaction mixture was concentrated. The residue was then partitioned between CH ₂ C1 ₂ and water. The organic phase was isolated, dried, concentrated to give a wax-like product, 2Z 0.79 g (97%). C. Compound 38

A methylene chloride solution (10 mL) of compound _V7 (0.29 g; 0.27 mmole) and l,4-Bis(3-amino-propyl)piperazine (1.12g; 5.6 mmole) was stirred at room temperature overnight. The reaction mixture was washed twice with water, dried, concentrated to give 0.19 g of product, 2586% yield. NMR (CDCI ₃ ): δ 7.15 (t, IH), 3.29-3.35 (m, 6H), 2.75 (t, 4H), 2.65 (t, 2H), 2.32-2.58 (m, 14H), 1.20-1.80 (m, 68H), 0.88 (t, 6H).

(R ¹ ) ₂ NH 0 pyridine - (R ¹ ) ₂ NCOCH ₂ CH ₂ COOH 36

3Z H ₂ N(CH ₂ ) ₃ N N-(CH ₂ ) ₃ NHj

3δ (R ¹ ) ₂ NCOCH ₂ CH ₂ CONH(CH ₂ )3— N N-(CH ₂ ) ₃ NH ₂

Example 4 Compound 39: To CH ₂ CI ₂ (20 mL) about 0°C solution of ethylenediamine (2.170 g; 44.5 mmole) was added dropwise a CH ₂ CI ₂ solution (20 mL) of cholesteryl chloroformate (1.0 g; 2.2 mmole). After being stirred at room temperature for 1 hr, the reaction was washed with H ₂ O, saturated NaHC0 ₃ aqueous solution, dried and concentrated to give a white solid (1.03 g). NMR (CDC1 ₃ ): δ 5.37 (d, IH), 4.93 (m, IH), 4.54 (m, IH), 3.21 (q 2H), 2.81 (t, 2H), 2.20-2.45 (m, 2H). Compound 40:

A CH C1 ₂ (10 mL) solution of L-t-BOC-Arg(t-BOC) ₂ -OH (0.2 g; 0.42 mmole), compound 22 (0.2 g; 0.42 mmole) and DIC (58 mg; 0.46 mmole) was stirred at room temperature for 4 hrs. The reaction was worked-up as compound 32 and purified by flash column chromatography affording 0.31 g, 79%, of white powder product compound 40- NMR (CDCI ₃ ): δ 9.30-9.45 (m, 2H), 7.20 (b, IH), 5.90-6.0 (m, IH), 5.38 (m, 2H), 4.30-4.55 (b, 2H), 4.00 (m, 2H), 3.20-3.45 (m, 4H).

Compound 41:

Compound 41 was synthesized from compounds 26 and 39 in 52% yield.

NMR (CDCI3): δ 6.85 (b, IH), 5.38 (d, IH), 5.15-5.25 (m, 2H), 4.70-4.80 (b, IH), 4.40-

4.55 (m, IH), 4.20 (b, IH), 4.00 (m, IH), 3.00-3.40 (m, 6H).

Compound 42: Compound 42 was prepared from compound 40 in the same manner as compound 30.

Compound 43:

Compound 42 was prepared by treating compound 41 with 2N HCI in dioxane.

O II

CIC Chol. + NH ₂ CH ₂ CH ₂ NH ₂ ,NHC(O)-ch0l.

NH ₅ 39

41

40. NH ₂

.NHC(O)-chol.

HCI/dioxane CONH ^* 40. or 41 ►

(CH ₂ ) ₃ NH ₂ - HCI 43

or

NH ₂

,NHC(O)-chol.

CONH'

(CH ₂ ) ₃ NHC(NH)NH ₂ -3 HCI 42

Example 5 Preparation of lipid-nucleic acid complexes and cell transfection. Lipids were prepared for complexing with nucleic acids by drying a cationic lipid-colipid mixture in CHC1 ₃ under argon. The molar ratio of cationic lipid to colipid was adjusted by adding appropriate amounts of each lipid in CHC1 ₃ together prior to drying. Usually 10 mL of water was added to a 100 mL round bottom flask containing a dried film of lipid and colipid. The lipids consisted of a mixture of an invention cationic lipid and the colipid DOPE (dioleylphosphatidylethanolamine). Sterile-filtered water or a low ionic strength

aqueous buffer such as physiological saline, TE (10 mM rris, 1 mM EDTA, pH 7-8) or Ringer's solution was then added to the dried lipids to obtain a lipid suspension at 1 mg lipid/ mL followed by a 10 minute bath sonication (Ultra Sonik 100, NEY) at room temperature to suspend the lipids in the flask. The suspended lipids (10 mL) were then sonicated 5 times for 15 seconds per sonication at 0-4° C with about 30-60 seconds between pulses. Sonication was usually conducted in a 15 mL polypropylene culture tube. A probe sonicator (Sonifier 250, Branson Ultrasonics) was used at maximum power for the ice bath sonication and for each 15 second pulse. The lipid suspension was optionally filtered or centrifuged (2000 rpm, 10 minutes at 0-4° C) to remove large particulate matter and the resulting lipid suspensions were kept at 4° C until used.

Lipid nucleic acid complexes were prepared by mixing (a) 10% v/v (usually 100 μL) oligonucleotide in Optimem™ (Bethesda Research Labs, Inc.) at room temperature (about 19-24° C) with (b) 10% v/v lipid suspension in Optimem™ at room temperature and allowing the mixture to stand for about 15 minutes, followed by adding (c) 80% v/v Optimem™ without serum or 80% v/v tissue culture medium (Modified Essential Medium (MEM), Dulbecco's Modified Eagle's Medium (DMEM), etc) containing 0-80% fetal bovine serum (Hyclone, Inc.). The lipid and nucleic acid solutions could also be prepared using DMEM without serum, a low ionic strength (less than about 200 mM total ion concentration) aqueous medium lacking phosphate (TE, etc) or water in place of Optimem™. Typically, 100 μL each of Optimem™ containing oligonucleotide (100-500 pmole) and Optimem™ containing added lipid was mixed with 800 μL of serum free Optimem™ or with 800 μL of tissue culture medium containing 10-80% FBS. Generally, 5 or 10 μL (5 or 10 μg) of lipid was added to 95 or 90 μL of Optimem™ to give a final 1 mL transfection preparation containing 5 or 10 μg of lipid and 250 pmole of oligonucleotide. The lipid-nucleic acid mixture was held at room temperature for about 15 min hours before adding medium to give the final transfection preparation volume (usually 1 mL) and warming to 37° C and immediately applying onto the cells. The transfection mixture was left on the cells for 2 to 24 hours (at 37° C), with a 6 hour time period commonly used.

Cells were typically transfected using 1 mL of transfection mixture per well in a 6-well plate. Cells at about 50-100% confluency were used. The efficiency of transfection was relatively uniform over this confluency range for the invention cationic lipids. Typical transfections with the invention lipids used cells at about 60-100% confluency. Other cationic lipids (Lipofectin™, Transfectam™ or Lipofectamine™) were used according to manufacturers

instructions and cells were thus at recommended confluency for transfection with these lipids. Cells at a lower confluency (at least about 10% confluent) could usually be transfected with oligonucleotides using the invention cationic lipids, but more cell toxicity was observed relative to cells at 50-100% confluency. The increased toxicity is believed to be due, at least in part, to the high level of oligonucleotide that was delivered into the cells.

Nucleic acids typically used were 15-mer phosphorothioate oligonucleotides that were labeled with fluorescein at the 5' or 3' terminus using a linker of structure NH ₂ (CH ₂ ) ₆ -0-P(0)(OH)-0- 5' linked to the oligonucleotide at the 5' or 3' hydroxyl group. The oligonucleotide designated ODNl was labeled at the 5' terminus and had a base sequence complementary to a region of the HSV strain 17 DNA polymerase gene at coordinates 421 to 436 (McGeoch J Gen Virol (1988) £>9_:1531). There was no known sequence that was complementary to ODNl or ODN2 in the tested cells, except where ODN2 without the fluorescent label (ODN2A) was used to inhibit T antigen synthesis in an antisense assay. The oligonucleotide designated ODN2 was labeled at the 3' terminus and consisted of the same bases and the same base sequence as the oligonucleotide designated TAg-15 (Wagner Science (1993) 2^0:1510-1513, at page 1511, Table 1). The oligonucleotides were used at a concentration of 250 or 500 nM (i.e., 250 to 500 pmole) for typical 1 mL transfections.

Cells for transfection experiments were generally seeded onto 25 mm diameter glass cover slips in 6-well plates 12-24 hours prior to transfection and were grown under conditions recommended for each cell line. The cells were generally grown on cover slips to facilitate analysis by fluorescence microscopy. The cells were washed twice with tissue culture medium containing serum or with serum-free Optimem™, followed by adding 1 mL of the lipid-nucleic acid mixture at room temperature. The cells were incubated with the lipid-nucleic acid mixture at 37° C for 2 to 24 hours, followed by washing the cells twice in medium containing serum. The cell washes were done by removing most of the medium from the wells, but leaving a thin layer, and adding about 1 mL of fresh medium for each wash. Removing all medium from the cells between each wash resulted in increased cell toxicity relative to washes with a thin layer of medium present at all times. Transfection efficiency and lipid toxicity was examined by fluorescence microscopy essentially as described (Wagner et al Science (1993) 260:1510) shortly after the lipid-nucleic acid complex was removed. Transfected cell lines were A10 (rat smooth muscle, ATCC No. CRL 1476), CV-1 (African green monkey kidney, ATCC No. CCL 70), NIH 3T3 (murine fibroblast), HeLa (human cervical carcinoma, ATCC No. CCL 2), SK-OV-3

(human ovary carcinoma, ATCC No. HTB 77), HL60 (human lymphoma, ATCC No. CCL 240), HUVEC (human primary umbilical endothelial cells), ECL (human endothelial cells, ATCC No. CRL 1730), NHEK (normal human epidermal keratinocyte, Clonetics™ No. CC-2603), L7 (murine connective tissue cells), Caov-3 (human ovary carcinoma, ATCC No. HTB 75), SK-BR-3 (human breast adenocarcinoma, ATCC No. HTB 30), Rat-2 (rat fibroblast, ATCC CRL 1764), MCF7 (human breast adenocarcinoma, ATCC No. HTB 22) and Caco-2 (human colon adenocarcinoma cells, ATCC No. HBT 37). HUVEC cells did not tolerate exposure to PBS and were washed in medium containing serum (DMEM with 10% FBS) prior to adding lipid-nucleic acid mixture. NHEK cells were grown in KGM™ medium (Clonetics™ No. CC-3001) according to supplier's instructions. L7 cells, derived from clone 929 (ATCC No. CCL 1), were stably transfected with a dexamethasone responsive β-galactosidase gene and were grown in EMEM with 10% horse serum and nonessential amino acids or in DMEM with 10% fetal bovine serum.

Example 6 A10 and CV-1 cells were transfected for 18 hours in medium containing

10% FBS (fetal bovine serum) using 1 mL of lipid complexed with 250 pmole ODNl for 15 minutes prior to addition to the cells on cover slips in 6-well plates. The cells were observed by fluorescence microscopy immediately after the lipid- ODN1 complex was removed to determine the proportion of cells that were transfected as shown by nuclear staining by the fluorescent labeled oligonucleotide ODNl. The results showed that both 2P_ and 25 efficiently transfected both A10 and CV-1 cell lines in the presence of serum with little or no toxicity to the cells.

% cells transfected

Compound A10. CV-1 CV-1 toxicity

Lipofectin™* 90 84 +**

211 95 85 +

25 85 70 - 44*** <10 <5_ _{= :}

* N-[l-(2,3-dioleyloxy)propyl]-N,N,N-trimethyl ammonium bromide-DOPE mixture; all lipids were used at 15 μg/mL; all lipids were used as a 1:1 molar mixture with DOPE as the colipid

** (+) low toxicity with no dead cells apparent and occasional change in cell morphology; (++) significant toxicity with > 40% of cells dead or dying; (-) no significant toxicity observed

*** compound having structure C. except that one R ¹ was hydrogen and the other R ¹ was n-Cι ₂ H ₂ 5.

Example 7 SK-OV-3 cells were transfected for 18 hours in medium containing either

40% or 80% FBS using 1 mL of lipid complexed with 500 pmole ODNl for 15 minutes prior to addition to the cells on cover slips in 6-well plates. The cells were observed by fluorescence microscopy immediately after the lipid-ODNl complex was removed to determine the proportion of cells that were transfected as shown by nuclear staining by the fluorescent labeled oligonucleotide ODNl.

The results showed that 20. efficiently transfected SK-OV-3 cells in the presence of serum. 80% serum approximates the serum protein and ion conditions that exist in vivo in mammalian circulation. 3J) is thus suitable for transfecting cells under physiological conditions.

% SK-OV-3 lipid transfected % serum

Lipofectin™ 15-20 40

2Q 76 40

25 <10 40

Lipofectin™ 0 80 2fi 72 80

55 Q 8J2

* all lipids were used as a 1:1 molar mixture with DOPE as colipid

Example 8 CV-1 cells were transfected for 18 hours and HL60 cells were transfected for

5 hours in medium containing different amounts of FBS using a transfection preparation consisting of 1 mL of 2Q complexed with 500 pmole ODNl for 15 minutes prior to addition to the cells on cover slips in 6-well plates. The cells were observed by fluorescence microscopy 5 hours after the lipid-ODNl complex was removed to determine the proportion of cells that were transfected as shown by nuclear staining by the fluorescent labeled oligonucleotide ODNl. The results showed that 20. efficiently transfected the cells in the presence of reduced levels of

lipid. The toxicity noted for HL60 cells may have been at least partly due to the high level of ODNl that was delivered to the cells.

* (-) transfection not done ** significant toxicity noted

Example 9 HUVEC cells were transfected for 5 hours in medium containing different amounts of FBS using 1 mL transfection preparations consisting of of 25 (1:1 mixture with DOPE) or Lipofectin™ complexed with 500 pmole ODNl for 15 minutes prior to addition to the cells on cover slips in 6-well plates. The cells were observed by fluorescence microscopy to determine the proportion of cells that were transfected as shown by nuclear staining. The results showed that 30 transfected HUVEC cells at a modest efficiency.

μg/mL % HUVEC lipid transfected lipid % serum

5 1

10 1

5 1

10 5

15.

Example 10 EC1 cells were transfected for 5 or 6 hours and NHEK cells were transfected for 24 hours in DMEM containing 10% FBS or serum free Optimem™ using 1 mL of 2 _ complexed with 250 pmole ODNl or Lipofectin™ complexed with 250 pmole ODNl for 15 minutes prior to addition to the cells on cover slips in 6-well

plates. The cells were observed by fluorescence microscopy after the lipid-ODNl complex was removed to determine the proportion of cells that were transfected as shown by nuclear staining. The results showed that 20 efficiently transfected ECl and NHEK cells. Fluorescence observations indicated that high levels of ODNl were delivered into both ECl and NHEK cells with little to moderate toxicity.

* 6 hour transfection ** Lipofectamine™ (Gibco/BRL; 2,3-dioleyloxy-N-[2(sperminecarboxamido)- ethyl]-N,N-dimethyl-l-propaminium trifluoroacetate) *** 5 hour transfection

Example 11

L7 cells were transfected for 18 hours in medium containing 20% FBS using 1 mL of 20. or Lipofectin™ complexed with 250 pmole ODNl for 15 minutes prior to addition to the cells on cover slips in 6-well plates. The cells were observed by fluorescence microscopy after the lipid-ODNl complex was removed to determine the proportion of cells that were transfected as shown by nuclear staining. The results showed that 20. efficiently transfected L7 cells with little or no toxicity.

Example 12 SK-BR-3, Caco-2, Caov-3 and MCF7 cells were transfected with ODNl using 24 which was prepared as described in Example 5 using a 1:1 molar ratio of DOPE colipid. The sonicated lipid-colipid mixture (1 mL of a 2.5, 5 or 10 μg/mL suspension) was mixed with 250 pmole ODNl and allowed to stand for 15 minutes at room temperature for 15 minutes before adding 1 mL of the lipid- ODNl complex to the cells. The lipid-ODNl complex was removed from the cells after 6 hours by washing the cells twice in PBS, followed by observing the cells in serum-containing medium immediately after removing the lipid-ODNl complex.

SK-BR-3 2.5 60-70

5.0 >90 1P ) >90 ±

* (-), no observed toxicity; (±), slight toxicity with no dead cells apparent, (+), moderate toxicity with <10% dead cells, (+++), all cells dead

Example 13

The optimal molar ratio of cationic lipid: colipid was determined using 2 . and DOPE colipid. and DOPE in CHC1 ₃ were mixed to obtain various molar ratios of the 2 lipids. The lipids were dried under vacuum, resuspended in deionized water by 10 freeze-thaw cycles, and filtered through a 100 nm Lipofast™ filter (Avestin, Inc.) according to manufacturers instructions to obtain a lipid suspension consisting of particles less than about 100 nm in diameter. CV-1 cells were transfected for 24 hours using 1 mL of 10 μg/mL lipid and 250 pmole ODN2 and observed immediately after removing the transfection mixture by washing twice with DMEM containing 10% FBS. The efficiency of ODN2 transfection into the cells increased steadily from a 80-90% transfection efficiency at a 54.DOPE ratio of 1:2.5 to a very high efficiency (-100%) at a 2.5:1 ratio. The transfection efficiency with respect to the amount of ODNl delivered into cells was highest for the 2.5:1 transfection based on a qualitative estimate of the fluorescence intensity observed in the cells after transfection. The transfection efficiency of the 24:DOPE (1:2) preparation delivered about the same amount of oligonucleotide to CV-1 cells as Lipofectin™ when used according to manufacturer's instructions. No toxicity or slight toxicity was observed after transfection with the tested preparations.

Example 14

The transfection efficiency of 24 was compared with Lipofectin™ and Transfectam™ (Promega, No. E1231) by determining the proportion of transfected cells as observed by fluorescence using ODN-1. Transfection preparations were as follows: #1 was 24 cationic lipid with DOPE colipid (1:1 molar ratio); #2 was Lipofectin™ with DOPE (1:1 molar ratio); #3 was Transfectam™ without colipid; #4 was Transfectam™ with DOPE (1:1 molar ratio). The cells were transfected for 6 hours using 5 μg of lipid with 250 pmole of ODN-1 and viewed immediately after the cells were washed.

cell line

* - A transfection in Optimem™ without serum; B - transfection in

DMEM with 15% FBS; C - transfection in DMEM with 10% FBS.

The transfection efficiency of 2 was also compared to that obtained using

Lipofectin™ and Transfectam™ by measuring the transfection efficiency (ODN2) in CV-1 cells using transfection times from 2 to 24 hours. Transfection preparations #l-#4 were used in medium containing 10% FBS and gave the following results.

The reduced efficiency of transfection with Lipofectin™ at the 6 hour time point was probably due to transfection of cells that were at about 90-95% confluency, which was more confluent than the 70-80% confluency recommended by the manufacturer. Relatively low cell toxicity was observed in all of the transfections (0-10% dead cells observed after transfection) except with transfection #3 at 24 hours, where about 50% of the cells were dead.

Transfection preparations #1 and #2 were used to transfect CV-1 cells for 6 hours as described except that the transfection medium contained either 25% FBS or 50% FBS. #1 transfected >95% of CV-1 cells in medium containing 25% FBS and 71% in 50% FBS while #2 transfected 21% of the cells in 25% FBS and 0-5% in 50% FBS. The transfection efficiency using 24 in the presence of high levels of serum was thus relatively high.

Example 15 CV-1 cells were transfected with a different molar ratios of 25 and DOPE colipid for 24 hours using 250 pmole of ODN2 and 1 μg of lipid-colipid complex per 1 mL transfection. DMEM with 10% FBS was used in the transfection preparations. 55 was found to transfect about 2- to 10-fold more oligonucleotide into transfected cells than 24 at all tested lipid:colipid ratios. No toxicity or slight toxicity (less than about 5% dead cells after transfection) was observed after transfection with the tested preparations.

55 was tested as a 2:1 25.DOPE preparation using 0.5, 1 or 2.5 μg of lipid per 1 mL transfection (CV-1 cells transfected for 24 hours using 250 pmole ODN2). Greater than 95% of the cells were transfected at all 3 lipid concentrations, although the amount of oligonucleotide delivered into cells was reduced by about 10-fold in the transfection using 0.5 μg of lipid preparation relative to the transfections that used 1.0 or 2.5 μg of lipid. No toxicity or slight toxicity (less

than about 5% dead cells after transfection) was observed after transfection with the tested preparations.

Example 16 CV-1 cells were transfected for 24 hours with RNA encoding chloramphenicol acetyltransferase (CAT). The transfections used 2.5 μg of RNA, 5 μg of i and DMEM with 10% FBS or Optimem™ with no serum. CAT activity was measured by immunofluorescence staining using fluorescent labeled anti-CATantibodies (5',3' Inc.). The cells were observed as previously described (Wagner, Science (1993) 2≤0_:1510). The results showed 1-2% of cells expressed CAT while no control cells transfected with lipid lacking RNA expressed CAT.

Example 17

Transfection preparations #5 (25 and DOPE at a 1:1 molar ratio; 5 μg lipid complex), #6 (5 μg 24 without DOPE colipid), #7 (42 and DOPE at a 1:1 molar ratio; 5 μg lipid complex) and #8 (42 and DOPE at a 1:1 molar ratio; 5 μg lipid complex) were used to transfect CV-1 cells for 48 hours using 250 pmole ODN2. Cells were observed by immunofluorescence immediately after removing the transfection lipids. Transfection #5 resulted in >95% of cells transfected and no visible toxicity, with a very high fluorescence intensity in nearly all nuclei, indicating that high levels of ODN2 were delivered into the cells. Transfection #6 similarly resulted in >95% of cells transfected and no visible toxicity. Transfections #7 and #8 resulted in 44% and 32% of cells transfected respectively, with no visible toxicity.

Example 18 Two plasmid DNAs, one encoding the CAT enzyme (Invitrogen, No. V790-20) and the other encoding β-galactosidase (Clontech, No. 6177-1), were transfected into cells in Swiss Webster mice by injecting 50 μL of lipid-plasmid complex intradermally, followed by assaying for CAT and β-galactosidase activity at the site of injection 2 days after injection. The transfection preparations consisted of lipid-nucleic acid complexes as follows:

#9, 20 μg of a 1:1 24:DOPE complex with 0.5 μg of each plasmid DNA in 1 mL of Optimem™ without serum; #10, 20 μg of a 1:1 24:DOPE complex with 1.5 μg of each plasmid DNA in 1 mL of Optimem™ without serum;

#11, 20 μg of Lipofectin™ with 0.5 μg of each plasmid DNA in 1 mL of Optimem™ without serum;

#12, 20 μg of Lipofectin™ with 1.5 μg of each plasmid DNA in 1 mL of

Optimem™ without serum; #13, 20 μg of DC-cholesterol with 0.5 μg of each plasmid DNA in 1 mL of Optimem™ without serum; #14, 20 μg of DC-cholesterol with 1.5 μg of each plasmid DNA in 1 mL of Optimem™ without serum; and

#15, 20 μg/mL of a 1:1 2 DOPE complex with no plasmid DNA in 1 mL of Optimem™ without serum. Transfection preparations containing 24 were complexed with the plasmids for 15 to 30 minutes prior to intradermal injection. DC-cholesterol (3-β-[N-(N',N'- dimethylaminoethane)carbamyl]cholesterol) was prepared and used as described (Gao Biochem Biophys Res Commun (1991) 179:280). A 6 mm diameter biopsy of epidermis and dermis was removed from the mice at each injection site. The tissue from each site was homogenized in 0.25 M Tris pH 7.4, 0.1% Triton X-100, centrifuged and the tissue pellet used to prepare a 150 μL cell extract essentially as described (Gorman Mol Cell Biol (1982) 2:1044). The cell extracts were assayed for CAT and β-galactosidase activity according to published methods (Gorman, ibid.; Miller Experiments in Molecular Genetics (1972) CSH press, p 352). β- galactosidase activity was measured using a fluorimeter to measure conversion of 4-methylumbelliferyl-β-D-galactopyranoside to the flourescent dreivative 4- methylumbelliferone (Prince Proc Natl Acad Sci (1987) 54:156). The error inherent in the assay due to variation in the protein concentrations of the cell extracts was estimated to be less than about ±15%. No signs of toxicity were visually observed at the injection site with any of the transfection preparations. The following results were obtained.

* - percent acetylation of ¹⁴ C labeled chloramphenicol using

50 μL of cell extract ** - fluorescence units using 50 μL of cell extract

Similar methods can be used to transfect oligonucleotides into cells in animals. For example, inhibiting gene expression using an antisense oligonucleotide in cells can be shown by adding an oligonucleotide having a base sequence complementary to CAT or β-galactosidase to a transfection preparation such as #9 or #10. Inhibition of gene expression is then measured by determining the expression of the gene in the presence of oligonucleotide compared to expression that is obtained using transfection preparations lacking antisense oligonucleotide. The nontargeted gene is used as an internal control to detect nonspecific inhibition of transcription that may be caused by the oligonucleotide.

Example 19

ODN2A, which was ODN2 lacking the fluorescent label at the 3' terminus, was used as an antisense oligonucleotide to inhibit T antigen expression in CV-1 cells using the protocol previously described (Wagner Science (1993) 260:1510- 1513). CV-1 cells were transfected for 18 hours using 5 μg of 24 (1:1 molar ratio with DOPE colipid) and then cell nuclei were microinjected with T antigen plasmid using a needle concentration of 0.003 μg/mL of T antigen plasmid (ibid at page 1513, first column, note 5) and 0.025 μg/mL of β-galactosidase plasmid. The cells were scored for T antigen expression 4.5 hours after removing the transfection preparation from the cells. The cells were immunostained for β- galactosidase expression as an internal control for microinjected cells and the proportion of cells expressing both T antigen and β-galactosidase was determined. This manner of inhibiting gene expression using this oligomer was previously observed (Wagner, ibid) and 24 thus did not affect the nature of the biological activity previously observed with ODN2A. The results showed that transfection preparations containing 2! efficiently delivered the antisense oligomer into the cells without interfering with the capacity of the oligonucleotide to inhibit gene (protein) expression by binding to the target RNA sequence in the cell. Transfection using 24 compared to transfection using Lipofectin™ indicated that 24 was about 10-fold more efficient in delivering ODN2A into the cells (i.e., about 10-fold less ODN2A was needed to inhibit T antigen expression to an equal extent).

Example 20

Lipid complexes consisting of 25 :DOPE (2:1 molar ratio) were prepared as follows. Two molar equivalents of 25 in CHC1 ₃ and one equivalent of DOPE in CHC1 ₃ was dried under reduced pressure by rotovap at room temperature.

Deionized water was then added to give 1 mg/mL of total lipid. 10 mL of water containing 10 mg of lipid was vortexed for 5 minutes in a 100 mL round bottom flask. 2 mL aliquots were then used to prepare lipid suspensions for complexing with ODNl. The first preparation was obtained by directly using the vortexed 2 mL aliquot to prepare complexes with ODNl. The second preparation was obtained by 6 cycles of freezing on dry ice and thawing in a 37°C water bath (Baxter, Dura Bath). The third preparation was obtained by placing a culture tube containing the lipid into a Ultra Sonik 100 (NEY) sonicator and the preparation was sonicated for 30 minutes at room temperature using the probe tip at maximum power. The fourth preparation was obtained by 6 cycles of freezing on dry ice and thawing in a 37°C water bath, followed by filtering the suspension through a Lipofast™ 100 nm filter according to the manufacturer's instructions to obtain complexes less than about 100 nm in diameter. The fifth preparation was obtained by sonication for 10 minutes in a sonication bath followed by 5 sonication pulses for 15 seconds each with a probe sonicator (Sonifier 250, Branson Ultrasonics). Each preparation was used to transfect CV-1 cells for 6 hours using 1 μg lipid and 250 nmole ODNl in 1 mL of medium with 10% FBS. All preparations efficiently transfected CV-1 cells, with some variation observed in the amount of ODNl delivered to cells and /or some variation in the proportion of cells that were transfected.

Example 21 Compound 46:

A CH ₂ C1 ₂ (20 mL) solution of 45. (Sigma, L-N ^α -t-BOC-N ^ω -CBZ- Arginine; 0.41 g; 0.1 mmole, Sigma), (BOC) ₂ 0 (0.24 g; 1.1 mmole) and TEA (0.4 g; 4 mmole) was refluxed for 3 hrs and then at room temperature for 5 hours. The reaction was concentrated and purified by flash column chromatography affording 0.27 g, 72%, of compound 15. NMR (DMSO-d ₆ ): δ 9.17 (b, 2H), 7.30-7.40 (m, 5H), 6.40 (b, IH), 5.05 (s, 2H), 3.60-3.90 (m, 3H), 1.30-1.75 (2s+m, 2x9H + 4H). Compound 47:

A CH ₂ C1 ₂ (10 mL) solution of 46 (260 mg; 0.51 mmole) was activated with NHS (65 mg; 0.56 mmole) and 1,3-dicyclohexylcarbodimide (DCC) (126 mg; 0.61 mmole) at room temperature for 30 min. Activated compound 46. (0.51 mmole) was then reacted with 25 (NH ₂ CH ₂ C(0)N(R ^! ) ₂ where both R ¹ were n-Ci4H ₂ 9) (0.24 g; 0.51 mmole) in the presence of TEA (0.1 g; 1.0 mmole) at room temperature for 3 hours. The reaction mixture was quenched with water. The solid was filtered off and washed with CH ₂ C1 ₂ . The filtrate was washed with water, dried, concentrated and purified by flash column chromatography,

affording 0.37 g of 47 (75%). NMR (CDC1 ₃ ): δ 9.30-9.45 (m, 2H), 7.10-7.41 (m, 6H), 5.39 (d, IH, J=4.5 Hz), 5.15 (s, 2H), 4.20 ( , IH), 3.80-4.05 (m, 4H), 3.45 (m, IH), 3.20- 3.38 (m, 2H), 3.09 (t, 2H, J=7.8 Hz), 1.05-2.00 (m+2s, 70H), 0.88 (t, 6H, J=6.4 Hz). Compound 48: _\7 was dissolved in ethanol (20 mL) and hydrogenated at 50 p.s.i. in the presence of 10% Pd/C for 40 hours. The catalyst was filtered off and the filtrate containing S was concentrated to dryness. NMR (DMSO-d6): δ 8.60 (b, 3H), 7.79 (b, IH), 6.99 (d, IH, J=8.4 Hz), 3.75-4.0 (m, 2H), 3.50-3.60 (m, IH), 1.00-1.70 (m, 70H), 0.88 (t, 6H, J=6.4 Hz). Compound 49:

A CH ₂ C1 ₂ (5 mL) solution of 48 (50 mg; 0.15 mmole), acetic anhydride (16 mg; 0.15 mmole) and TEA (31 mg; 0.31 mmole) was stirred at room temperature for 1 hour. The reaction mixture was poured into 10% citric acid aqueous solution, dried, and purified by flash column chromatography, affording 50 mg (78%) of 49. NMR (CDCI ₃ ): δ 10.40 (b, IH), 9.26 (b, IH), 7.09 (b, IH), 5.27 (d, IH,

J=8.1 Hz), 4.20-4.22 (m, IH), 4.00 (s, 2H), 3.87-3.95 (m, 2H), 3.30 (t, 2H, J=8.0 Hz), 3.14 (t, 2H, J=7.6 Hz), 2.13 (s, 3H), 1.10-2.0 (m, 70H), 0.87 (t, 6H, J=6.4 Hz). Compound 50:

A CH ₂ C1 (5 mL) solution of 45 (51 mg; 0.15 mmole), isopropyl chloroformate (170 μL of a 1 M solution in toluene; 0.17 mmole) and TEA (40 mg; 0.39 mmole) was stirred at room temperature for 1 hour. The reaction mixture was poured into 10% citric acid aqueous solution, dried, and purified by flash column chromatography, affording 57.3 mg (79%) of 5Q. NMR (CDC1 ₃ ): δ 9.44 (b, IH), 9.27 (b, IH), 7.16 (b, IH), 5.34 (d, IH, J=8.1 Hz), 4.80-4.90 (m, IH), 4.18- 4.20 (m, IH), 4.00 (s, 2H), 3.80-3.98 (m, 2H), 3.30 (t, 2H, J=8.3 Hz), 3.14 (t, 2H, J=7.6 Hz), 1.0-2.0 (m, 76H), 0.87 (t, 6H, J=6.3 Hz). Compound 51:

A CH ₂ C1 ₂ (5 mL) solution of 4 (51.3 mg; 0.15 mmole), distilled trimethylacetyl chloride (21 mg; 0.17 mmole) and TEA (40 mg; 0.39 mmole) was stirred at room temperature for 1 hour. The reaction mixture was poured into 10% citric acid aqueous solution, dried, and purified by flash column chromatography, affording 61 mg (84%) of 51. NMR (CDC1 ₃ ): δ 10.30 (b, IH), 9.16 (b, IH), 7.03 (b, IH), 5.02 (d, IH, J=7.6 Hz), 4.15-4.20 (m, IH), 3.80-4.10 (m, 4H), 3.30 (t, 2H, J=7.8 Hz), 3.14 (t, 2H, J=7.5 Hz), 1.0-2.0 (m, 79H), 0.88 (t, 6H, J=6.4 Hz). Compound 52:

49 (47 mg) was treated with 2N HCI in 1,4-dioxane (3 mL) at room temperature for 3 hours. The reaction mixture containing was concentrated to dryness, azeotroped with CH ₃ CN twice and dried under vacuum overnight.

NMR (DMSO-d6): δ 12.10 (s, H), 9.00 (b, H), 8.60 (b, IH), 8.25 (b, 3H), 3.80-4.10 (m, 3H), 2.15 (s, 3H), 1.10-1.90 (m, 52H), 0.88 (t, 6H). Compound 53:

5fl (50 mg) was treated with 2N HCI in 1,4-dioxane (3 mL) at room temperature for 3 hours. The reaction mixture containing 52 was concentrated to dryness, azeotroped with CH ₃ CN twice and dried under vacuum overnight. NMR (DMSO-d6 + HCI): δ 11.20 (s, Nδ-H), 8.20-8.98 (m, N«-H ₃ , N ^® -H _2/ N-H, N ^§ - H), 4.93 (m, IH), 1.10-1.85 (m, 76H), 0.88 (t, 6H). Compound 54:

51 (52 mg) was treated with 2N HCI in 1,4-dioxane (3 mL) at room temperature for 3 hours. The reaction mixture containing 52 was concentrated to dryness, azeotroped with CH ₃ CN twice and dried under vacuum overnight. NMR (DMSO-d6 + HCI): δ 11.42 (s, Nδ-H),9.54 (b, Nδ-H), 9.15 (b, Nω-H), 8.89 (b, Nω-H), 8.64 (t, N-H), 8.26 (b, N< -H ₃ ), 3.82-4.10 (m, 3H), 3.18-3.40 (m, 6H), 1.10-1.85 (m, 79H), 0.88 (t, 6H).

t-BOC-NH 45

H _2l 10% Pd/C 1 CONHCH ₂ C(O)N(R ) ₂

47

(CH ₂ ) ₃ N(t-BOC)C(NH)NH ₂

(CH ₂ ) ₃ N(t-BOC)C(NH)NHC(O)R ¹ⁱ

3: R ¹⁵ = CH ₃ ; 5p_: R ¹⁵ = OCH(CH ₃ ) ₂ ; 51: R ¹⁵ = C(CH ₃ ) ₃

NH ₂ 52 - 54

.2HCI

2N HCI/dioxane -CONHCH ₂ C(O)N(R ¹ ) ₂ la-si

(CH ₂ ) ₃ NHC(NH)NHC(O)R ¹⁵

52: R ¹⁵ = CH ₃ ; 53: R ¹⁵ = OCH(CH ₃ ) ₂ ; 51: R ¹⁵ = C(CH ₃ ) ₃

Example 22 Lipids 2 _/ 52, 2 _/ 51 and 55 were each prepared using a 2:1 lipid:DOPE molar ^' ratio as follows. Two molar equivalents of each lipid in CHC1 ₃ and one equivalent of DOPE in CHC1 ₃ was dried under reduced pressure by rotovap at room temperature. Deionized water was then added to give 1 mg/mL of total lipid. 10 mL of water containing 10 mg of lipid was vortexed for 5 minutes in a 100 mL round bottom flask. 2 mL aliquots were then used to prepare lipid suspensions for complexing with ODNl. Each preparation was obtained by 6 cycles of freezing on dry ice and thawing in a 37°C water bath.

CV-1 cells at about 80% confluency were transfected for 6 hours in DMEM medium containing 10% FBS using 1 mL of lipid complexed with 250 pmole ODNl for 15 minutes prior to addition to the cells on cover slips in 6-well plates. The cells were observed by fluorescence microscopy immediately after the lipid- ODNl complex was removed to determine the proportion of cells that were transfected as shown by cytoplasmic and nuclear staining by the fluorescent labeled oligonucleotide ODNl. The results showed that the tested compounds efficiently transfected CV-1 cell lines in the presence of serum with little or no toxicity to the cells.

% cells

Compound lipid (μg) transfected toxicity

25** 1 40 2.5 85 +

5.0 100 +

10.0 60 + +

55 1 40 2.5 75 +

5.0 100 +

10.0 40 + +

* (+) low toxicity with no dead cells apparent and occasional change in cell morphology; (++) significant toxicity with > 40% of cells dead or dying; (-) no significant toxicity observed ** cells were relatively uniformly stained

Compounds 25, 2 and 55 were used to transfect CV-1 cells with a 21-mer antisense phosphorothioate-linked oligonucleotide analog containing a base sequence complementary to a 21 base region near the middle of the cdc-2 kinase gene mRNA. The oligonucleotide contained 5-(l-propynyl)cytosine in place of cytosine and 5-(l-propynyl)uracil in place of uracil. A control oligonucleotide analog having a mismatched base sequence was also used to check for any non¬ specific effect on cdc-2 kinase levels in the transfected cells caused by the oligonucleotide analog. A control containing only lipid was also used as a control. The transfections were conducted essentially as described in Example 6 using 2.5 μg and 5.0 μg of lipid per transfection with an oligonucleotide analog concentration of 30 nM. Protein levels were examined by Western blot analysis. The results indicated that transfections using the cdc-2 antisense oligonucleotide analog efficiently reduced cdc-2 protein levels when transfected into the cells with each of the tested lipids. Control transfections using lipid alone or using lipid with the mismatched oligonucleotide analog did not affect cdc-2 protein levels.

Example 23 Compound 55:

To a solution of 52 (0.365 g) in 2.75 mL of dry dioxane under a nitrogen atmosphere was added at room temperature 4N HCI (2.75 mL) in dioxane. The progress of the reaction was followed by reverse phase HPLC (Vydac C18 column, 4.6x250 mm; mobile phase 90% methanol, 10% 50 mM ammonium acetate; 1 mL/minute flow rate; column temperature 60° C; photodiode detector array). Compound 55 eluted with a peak at about 7.4 min. and 25 eluted with a peak at about 9.6 min. The reaction was stopped by removing HCI and the solvent under reduced pressure when the fully protected species, 22 _/ and its mono- deprotected derivatives were completely converted to a mixture of 25 and 55. The reaction mixture was concentrated in vacuo to a residue and the residue azeotroped two times with acetonitrile. A 0.125 g portion of the residue was purified on a C-8 reverse phase HPLC column (E. Merck) by eluting with a 50- 100% methanol-water gradient. Fractions corresponding to 25 and 55 were combined and concentrated in vacuo and the residue azeotroped twice in acetonitrile to give 0.021 g of an oil. The HCI salt of 25 and 55 was prepared by dissolving the residue in dioxane, adding about an equal volume of 4N HCI in dioxane and concentrating to a residue in vacuo. The resulting residue was azeotroped twice in acetonitrile.

35.55 and mono-deprotected derivatives of 22 were also separated by TCL using silica gel 60 (normal phase) and 65:25:4 chloroform:methanol: water. 55 had the lowest mobility (Rf = 0.25), 5 mobility was Rf = 0.54 and the mobility of mono-deprotected derivatives of 22 was 0.96. The compounds were detected by iodine vapor or by ninhydrin stain for 25 and 55-

25: 1HNMR (DMSO-d.6, 400 MHz) 8.06 (br s, IH), 3.89 (br s, 2H), 3.27 (br s, IH), 3.15 (m, 4H), 3.02 (br s, 2H), 1.6-1.4 (m), 1.31. (s, 9H), 1.20 (m), 0.82 (t, J=7.5 Hz, 6H). MS (FAB, m/e) = 723.6 (M+l), 623.6 (M+l-BOC).

Each of the cited works above is incorporated by reference in its entirety. Specifically cited sections or pages of the above cited works are incorporated by reference with specificity.