CA2141216A1 - Targeting of liposomes to the blood-brain barrier - Google Patents

Abstract

The present invention is directed to the targeting of drugs, peptides, and proteins to the brain. Liposomes containing the compound to be delivered are coupled to antibody binding fragments which bind to receptors on the vascular endothelium of the blood-brain barrier. The targeted receptors include transferrin receptor, insulin receptor, IGF-I and IGF-II receptors, among others. Also provided are pharmaceutical compositions and methods for administering the drugs to a subject in need by means of the receptor-targeted immunoliposomes.

Description

~ 2141216 TARGETING OF LIPOSOMES TO THE BLCOD-BRAIN BARRIER

Bac~lo~l,d of the Invention Thi8 invention relates generally to the delivery of drugs into the brain by transcytosis across the blood-brain barrier. More particularly, the invention relates to the targeting of drug-containing liposomes to the blood-brain barrier by means o~ antibodies and b~n~ fragments thereof which bind to receptors on capillary endothelial c~lls of the brain.
Delivery of many drugs, particularly nonlipophilic drugs, to the mammalian brain i8 restricted by the blood-brain barrier. The barrier i8 due in part to tight intercellular ~unctions between brain capillary endothelial cells and prevents the passive movement of many substances from the blood to the brain. The brain endothelial cells lack continuous gaps or ch~n~el~ connecting the luminal and abluminal membranes which would otherwise allow the passage of blood-borne molecules into the brain tissue. Instead, the pres~nce of specific transport systems within the capillary endothelial cells, such as those for insulin, amino acids, glucose and transferrin, assure the controlled transport of compounds ~cç~sAry to the functioning of the brain.
Various strategies have been developed to deliver drugs into the brain that would not otherwise be able to cross the blood-brain barrier. Commonly, although quite undesirably, an intraventricular catheter is surgically implanted to deliver a drug directly into the brain. Not only does thiæ involve an invasive procedure that itself is potentially harmful to the patient, but the drug delivered by wo 9~ 6 ~ ~. PCT/US93/069 ~

this means is only super~icially distributed within the brain.
Other strategies have been devised to circumvent the blood-brain barrier. Some pharmacologic approaches involve chemically converting hydrophilic drugs ii~to more lipid-soluble forms 80 they are more easily transported across thebarrier. Another approach involves transiently opening the barrier by infusing hypertonic substAn~ intra-arterially to allow pAsc~e of hydrophilic drugs, although hypertonic subs~nc~ may be associated with toxicity and even damage the barrier.
More recently, it has been proposed to link hydrophilic neuropharmaceutical agents ~uch as neuropeptides to other peptides which are themselves capable of crossing the blood-brain barrier by transcytosi~. U.S. Patent No.
4,801,575 to Pardridge describes the preparation of chimeric peptides by coupling or con~ugating the pharmaceutical agent to a transportable peptide. The chimeric peptide purportedly p~re- acro~s the barrier via receptors for the transportable peptide. Transportable peptides, or vectors, mentioned as suitable for coupling to the pharmaceutical agent include insulin, transferrin, insulin-like ~rowth factors I and II, basic albumin and prolactin. However, as noted in Pardridge et al., J. Pharm. ~Yp. Therap. 259:66-70 (1991), the use of insulin as a vector to the blood-bra~n barrier may be limited by hypoglycem~a side effects and by rapid clearance of the peptide by the liver, and transferrin may be limited as a vector by its high con~entration in the plasma. Friden et al., Proc. Natl. Acad. Sci. USA 88:4771-4775 (1991), have suggested that drugs conjugated to anti-transferrin receptor antibodies cross the blood-brain barrier.
It has also been suggested that liposomes can ~nh~nce drug delivery to the brain across the blood-brain barrier. See, e.g., Umezawa and Eto, Biochem. Biophys. Res.
Comm. 153:1038-1044 (1988); Chan et al., Ann. Neurol. 21:540-547 (1987); Laham et al., Life Sciences 40:2011-2016 (1987);
and Yagi et al., J. Appl. Biochem. 4:121-125 (1982).
Liposomes are small vesicles (usually submicron-sized) comprised of one or more concentric bilayers of phospholipids 2 1 4 1 2 1 6 f PCT/US93/06942 ~eparated by aqueous compartments. Although liposomes have been reported to ~hAnce the uptake of certain drugs into the brain after intravenous in~ection (Chan et al., and Laham et al., ~hid.)~ it ha~ more recently been ~hown that liposomes do not cross the blood-brain barrier. Schackert et al., Select.
cer Therapeut. 5:73-79 (1989) and Nicklus et al., Riochim.
Riophys. Acta 1124:7-12 (1992). As noted in Micklus et al., lh~., liposomeg circulating in the plasma are ultimately taken up by the liver, digested and the lipid components released and redistributed to other organs. See also, Derkcen et al., Bioc~im. Biophys. Acta 971:127-136 (1988). The radioactivity found in the brain following intravenous in~ection of labeled liposomes of Umezawa and Eto, ~upra, and others may in fact be derived from digested lipids and not from the intact liposomes themselves.
Many drugs, particularly neuropept~des, growth factors and the like, are rapidly degraded in the systemic circulation prior to reaching the brain, or minimally penetrate the blood-brain barrier. Many of these peptides, if effectively del~vered to the brain in small quantities, could have a dramatic therapeutic effect. See, e.g., R~n~c and Xastin, Am. J. Physlol. 259:El-ElO (1990). Presently, however, large amounts must be administered to allow entry of even minimal levels required for pharmacologic action.
What is n~e~?~ in the art is an effective means to deliver pharmaceutical agents such as neu~o~e~ides across the blood-brain barrier and into brain tissue. The methods and pharmaceutical compositions thereof would desirably provide uniform introduction of the pharmaceutical agent throughout the brain and present as little risk to the patient as possible. Quite surprisingly, the present invention fulfills these and other related needs.

PCT/U593/06g4 Summary of the Tnvention The present invention prov~des ~mmunoliposomes and pharmaceutical compositions thereof capable of targeting pharmacological com~ounds to the brain. Liposomes are coupled to an antibody binding fragment such as Fab, F(ab')2, Fab' or a single antibody chain polypeptide~which binds to a receptor molecule present on the vascular çndothelial cells of the mammalian blood-brain barrier. Typically the antibody binding fragment i5 prepared from a monoclonal antibody. The receptor i~ preferably of the brain peptide transport system, such as the transferrin receptor, insulin receptor, IGF-I or IGF-2 receptor. The antibody b~ n~ fragment is preferably coupled by a covalent bond to the liposome.
Pharmacological comro~m~ which eæpecially benefit from this invention are those which are typically poorly transported across the blood-brain ~arrier, ~uch as hydrophilic peptides, and which are often highly potent in small quantities in the bra~n. These include, for example, peptide neulo~G~l~ic factor~, neurotransmitters and neuromodulators, e.g., beta-endorph~n~ and ~nkerh~lins. For optimal delivery and to lessen the possibility of harmful side effects as a result of the liposome, the diameter of the liposome is less than 1 micron and typically smaller than about 0.45 microns.
In other aspects the invention provides methods for targeting a pharmacological com~oul.d to the blood-brain barrier of a mammal. A liposome cont~ning the pharmacological compo~n~ is coupled to an antibody b~ n~ i n~
fragment which binds to a receptor molecule present on vascular endothelial cells of the mammalian blood-brain barriers, such as the transferrin receptor, insulin receptor, IGF-I or IGF-2 receptor. The pharmaceutical-containing immunoliposome is then administered to the sub~ect in an 3s amount sufficient to effectively treat or prevent the disorder. The preparation is usually administered parenterally, e.g., intravenously or intraarterially.

~ 21gl216 W094/02178 ~ PCT/US93/06942 Description of the Specific ~mbodiments The present invention provides comp~sitions and methods for targeting drugs to and across the blood-brain barrier. The compositions, including pharmaceutical compositions, are comprised of a liposome, a pharmaceutical aqent inten~e~ to be transported to the brain, and an antibody molecule or b1n~n~ fragment thereof which binds to a receptor present on vascular endothelium cells of the brain.
Preferably the receptor is transferrin receptor and the antibody which binds to the receptor is a b~n~n~ fragment which lacks some or all of the Fc portion of the molecule to minimize clearance of the composition by the reticuloendothelial system (RES).
Liposomes are used in the present invention to carry the pharmaceutical agent of interest to the blood-brain barrier for ultimate transport across the barrier and into the brain. As used herein and as reCo~n~ 7-e~ in the art, liposomes include any synthetic (i.e., not naturally occurring) structure compos~ of lipid bilayers that enclos~ ~ volume.
Liposomes include emulsions, foams, micelles, insoluble monolayer~, liquid cryctals, phospholipid dicpersions, lamellar layers and the like. In the liposome preparations of the invention, the pharmaceutical agent to be delivered across the blood-brain barrier is incorporated as part of a liposome in conjunction with the targeting molecule.
The receptor to which the antibody fragment/liposome is directed may be one or several receptors present in vascular endothelium of the brain. The receptor should be sufficiently accessible to be bound by the antibody fragment coupled to the liposome. Receptors which are present in - higher quantities in the brain capillaries than others are preferred for enhAncing levels of transport across the blood-brain barrier. Receptors which may be targeted include, for example, transferrin receptor, insulin receptor, insulin-like growth factors I and II (IGF-I and IGF-II) receptors and other receptors of the brain peptide transport system, as generally described in Pardridge, Peptide Drug Delivery to the Brain, 2~ 6 PCT/U593/069 Raven Press, New York, NY (1991), and U.S. Patent 4,801,575, which are incorporated herein by reference, the glucose transport receptor and the like.
Preferably the receptor targeted~by antibody b~ n~
fragment i8 transferrin ~c~or. Transferrin receptor is selectively enriched on the endothelium of the brain microvascular endothelium of a variety of mammals, including humans, and is the primary pathway for iron to enter the brain. Iron-loaded transferrin, an 80 Kd glycoprotein, the principal iron transport protein in the circulation, undergoes transcytosis through the blood-brain barrier via the transferrin receptor. The structure and function of the transferrin receptor have been described in Seligman, Prog.
Hematol. 13:131-147 (1983), which i8 incorporated by reference herein. Transferrin receptor is believed to pro~ect largely into the extracellular space. It can be isolated and purified from brain t1~ using well known prooedu~e_. The pur~fication, cloning, expression of human in~ulin ~ec~or and the production of monoclonal an~iho~ies thereto are described in U.S. Patent No. 4,761,371, which i8 i~ L~orated herein by reference.
The targeted receptor purified from brain or ex~pr~ by recombinant DNA techn~ques may be used to produce antibodies, and preferably monoclonal ant~ho~es, which may then be used to produce the antibody fragments useful in the present invention. The ant~ho~1es are made to the receptor molecule or a portion of the receptor molecule, such as that domain or domains which contributes to ligand b~n~in~.
Preferably, the antibody will bind to an extracellular portion of the receptor which is proximate to the b~n~n~ site of the receptor's native ligand, e.g., transferrin or insulin. In more preferred emho~iments the antibody does not substantially block or compete with the binding of ligand to the receptor.
The production of monoclonal anti ho~ ~ es to transferrin receptor, including the OX-26 monoclonal antibody described in the Examples below, has been described in Jeffries et al., Immunology 54:333-341 (1985), which is incorporated herein by reference. Other monoclonal antibodies WO94/02178 2 1 ~ 1 2 1 6 PCT/US93/06942 to transferrin receptor have been described. See, e.q., the ATCC Catalogue of Cell Lines and Hybridomas, 7th ed., 1992, describing HB21, a murine IgGl antibody which precipitates human transferrin receptor; HB84, a murine ~gG2a monoclonal antibody to human transferrin receptor; and TIB219 and TIB220, rat anti-murine transferrin receptor ant1hoA~es. Monoclonal antibodie~ to the human insulin receptor have also been reported, including cell lines available from the ATCC, such as ATCC HB175 (J. Biol. Chem. 258:6561-6566 (1983)) and ATCC
CRL 1827 (Curr. Top. Cell. Reg. 27:83-94 (1985)), which are incorporated herein by reference.
Methods for the production of monoclonal antibodies are well known and may be accomplished by, for example, immunizing the animal with cells which express the receptor, substantially purified receptor, or fragments thereof as ~cl~se~ above. Antibody producing cells obtained from immunized animals are immortalized and ~cre~n~, or screened first for the production of antibody which binds to the receptor protein and then immortalized.
As an antibody fragment used in the present invention will generally lack the immunogenic Fc portion, the nececcity of using human monoclonal antibodies i8 ~ubstantially avoided. However, it may be desirable to transfer antigen b1n~n~ regions of the non-human ant~h~ies, e.g., the hypervariable regions, to human framework regions by recombinant DNA techniques to produce substantially human Fab' molecules. Such methods are generally known in the art and are described in, for example, EPO publications 173,494 and 239,400, which are incorporated herein by reference. The production of single polypeptide chain binding molecules, also referred to as single chain ant;ho~ies, by recombinant DNA
tec~n~ques is described in detail in U.S. Patent No.
4,946,778, which is incorporated herein by reference.
The antibody to the brain receptor molecule may be used intact or, more preferably, as a binding fragment thereof. By binding fragment is meant that the fragment retains an ability to specifically bind to an epitope of the target molecule of interest, such as transferrin or insulin 2~ 6 8 receptor, for example. Antibody binding fragments include at least the hypervariable or complementarity determining region (CDR) ~ituated in an appropriate framework to produce a conformation which binds to the antigenic determinant. These b~1ng fragments include, but are not limited to, Fv, Fab, F(ab')2, Fab', and single polypeptide chain b~ molecules.
Antibody b~ ng fragments can be produced from intact antibody molecules by a variety of proce~res well known to those of skill in the art. For example, antibodies are usually fragmented by partial digestion with p~F~n to produce two Fab fragments. Pepsin treatment can be used to produce the F(ab')2 fragments where the two antigen binding domains are still bound together, i.e., as a multivalent binding molecule. Further reduction of the F(ab')2 fragment with, e.g., dithiothreitol can be used to separate the antigen binding domains and produce two F(ab') fragments.
Ihe preparation of the various antibody fragments i8 described in detail in, e.g., Harlow and Lane, Antibodies A T~horatorv Manual, Cold Spring Harbor Laboratory Pres~, Cold Spring Harbor, NY 1988 pp. 626-631, and Parham, in Cel 1~ r Immunolo~v, 4th ed., (ed. D.M. Weir), vol. 1, chap. 14, Blackwell Scientific Publ., CA 1986, both of which are incorporated herein by reference. The polypeptide ch~ i n~ of the desired fragments may also be produced by a variety of recombinant DNA techn~gues and assembled intra- or extracellularly.
Once the desired antibody or, more preferably, an antibody binding fragment, has been obtained it is coupled to or incorporated into the liposome. Liposomes for use in the invention may be formed from st~n~rd vesicle-forming lipids, which generally include neutral and negatively charged phospholipids and a sterol, such as cholesterol. ~he selection of lipids is generally guided by consideration of, e-g., liposome size and stability in the bloodstream. Thus, the lipids which can be employed in the present invention include, e.g., cholesterol hemisuccinate and salts thereof, tocopherol hemisuccinate and salts thereof, a glycolipid, a phospholipid such as phosphatidylcholine, ~ ~lgl216 phosphatidylethanolamine (PE), phosphatidylserine, phosphatidylglycerol, phosphatidic acid, phosphatidylinositiol, sphingomyelin, and the like, alone or in combination. The phospholipids can be synthetic or derived from natural sources such ac egg or soy. Phosphatidylcholines having a variety of acyl chain yL~u~_ of varying ch~n length and degree of saturation are available or may be isolated or cynthesized by well known techniques. In general, less saturated phosphatidylcholines are more easily sizQd, particularly when the l~po~omes are sized below a~out 0.3 microns. The acyl chain compositions of phospholipid may also affect the stability of liposomes in the blood. Sterols such as cholesterol can be combined with the phospholipids. The phospholipids employed and the amount of ~terol present depends on a number of factors such as lipophilicity of any added pharmaceutical agent, the targeting antibody fragment, and required properties of the liposome. These factors are generally known to those skilled in the art.
A wide variety of methods for preparing liposomes suitable for targeting to the blood-brain barrier ~n the present invention are available. See, for example, Szoka et al., Ann. Rev. B~ophys. Bioeng. 9:467 (1980), Li~osome Technoloqy, ed. G. Gregoriadi~, CRC Press, Inc., Boca Raton, FL (1984), U.S. Patent Nos. 4,235,871, 4,501,728, 4,837,028 4,957,735 and 5,019,369, each of which is incorporated herein by reference. Since the targeted liposomes of the invention are not intended for uptake by the reticuloendothelial system, the lipid components are further selected to increa~e time in the bloodstream.
Following liposome preparation, the liposomes may be sized to achieve a desired size range and relatively narrow distribution of liposome sizes. For delivery to the brain, liposomes should generally be less than about 1.0 microns in size, more preferably about 0.2 to 0.45 microns, which allows 3s the liposome suspension to be sterilized by filtration. For sizing liposomes, a liposome suspension may be sonicated either by bath or probe down to small vesicles of less than about 0.05 microns in size. Homogenization may also be used, W ~ ~ PCT/US93/0694 which relies on shearing energy to fragment large liposomes into smaller ones. Homogenizers which may be conveniently used include microfluidizers pro~ce~ by Nicrofluidics o~
Boston, M~. In a typical homogenization p~o~ e, liposomes S are recirculated through a ~t~ rd emulsion homoge~ r until selected liposomes slzes, typically between 0.1 and O.5 microns, are observed. In both methods the particle size distribution can be monitored by conventional laser-beam particle size discrimination.
Extrusion of llposomes through a small-pore polycarbonate membrane or an asymmetric ceramic membrane is an effective method for reducing liposome sizes to a relatively well defined size distribution. Typically, the suspension is cycled through the membrane one or more times until the desired liposome size distribution is achieved. The liposomes may be extruded through successively smaller pore membranes to achieve a gradual reduction in liposome size.
Methods for coupling antibodies to liposomes generally involve either covalent crossl~n~n~ between a liposomal lipid and a native or modifiQd antibody (or bin~ng fragment thereof). In another approach, an antibody which has been covalently derivatized with a hydLo~hobic anchor, such as a fatty acid, is incorporated into a preformed lipid.
A variety of crosslinking agents can be employed to produce a covalent link between a lipid and the antibody. One protocol involves the derivatization of the free amino group of a PE with an amino reactive bifunctional crosslinker. The derivatized PE along with other l~pids are then used to form liposomes by the methods described above. Once incorporated into the liposomes the derivatized PE can be reacted with the antibody by using the second reactive site on the crossl~nk~ng reagent. The use of heterobifunctional reagents are particularly useful because homocoupling between liposomes or between antibodies is avoided. Heterobifunctional 3S crosslinkers which can be used include N-hydroxysuccinimidyl 3-(2-pyridythio) propionate (SPDP), m-maleimidobenzoyl-N-hydroxysuccinimde (MBS) and the like, and homobifunctional crosslinkers include toluene-2,4-diisocyanate (TDIC) and 1-~ 2141216 W O 94/02178 PC~r/~JS93/06942 ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDCI)).
Chemical cros~l~n~n~ tec~n~ques al80 include the use of glutaraldehyde. These and other regents are reviewed in Connor and Huang, Pharmac. Ther. 28:341-365 (1985), which is incorporated by reference herein.
The covalent attachment of antibody fragments to l~po~mes by crossl~n~ing provides the advantage of attaching a large amount of antibody b~n~n~ fragment to the liposome surface, with little or no loss of entrapped aqueous content using these methods. Care should be taken, however, that the crossl ~nk~ ng conditions are not so harsh that antibody bln~i ng is adversely affected, or that the liposomes and their contents are damaged. To avoid the orientation of antibody molecules which may render them ~nc~r~hle of b~ ng to antigen, a spacer arm between the antibody and the lipid may optionally be employed.
A non-crosslinking association of ant~h~A~es or b~nA~ng fragments thereof with liposomes can also be employed.
This approach allows more than one type of antibody b~ g specificity per liposome, e.g., ant~hoA~es or fragment~ to different antigenic determinants on the transferrin molecule or even ant~ho~es which bind to different receptors. The flexibility of the molecule minimizes steric hindrance which may block b~ q of the antibody to the antigen. Also, the reagents involved in the non-covalent methods are relatively mild and do not come into direct contact with the contents of the liposomes, thus avoiding possible damage to the liposome components.
In one method the non-covalent association of the antibody or fragment with the liposome can be accomplished by derivatizing the antibody component with a hydrophobic group.
The derivatized antibody is then incorporated into the vesicles by inserting the hydrophobic anchor into the bilayer.
Methods of derivatization include reacting the antibody with 3s the N-hydroxysuccinimide ester of palmitic acid (NHSP) in the presence of 1-2% detergent, typically deoxycholate, as described in Huang et al., Biochem. Biophys. Acta 716:140-150 WO94/o2l~ ~ 4~2 -6 PCT/US93/0694 ^ 12 (1982). Alternatively, palmitic acid chloranhydride can replace NHSP as the acylation reagent.
Another method to acylate the antibody involves derivatizing the head group of a PE o~ ~imilar molecule rendering it capable of reacting with a sulfhydryl yLou~. The antibody' 8 disulfide bonds are t~e~ reduced with dithiothreitol (DTT) and then incubated with the derivatized PE. In yet another method useful in the present invention, a free PE or similar molecule is crosslinked to the carboxylic ~-ou~s of an antibody by carbodiimide a described in Jansons and Mallett, Analyt. Biochem. 11:54-59 (1981), protecting the free amino yLOU~S of the antibody prior to the crosslinking to prevent homocoupling between antibody molecules.
Once derivatized, the antibody or binding fragment thereof can become associated into the liposomes by a variety of established protocols. For example, the derivatized antibody is mixed with the liposome in the pr6~encQ of detergent, and the detergent then removed by dialysis to form the immunoliposomes. Another method involves a modification of the reverse-phase evaporation vesicle preparation method, where the preformed vesicles are mixed with antibody (lipid:antibody = 10:1 w/w) in a deoxycholate solution in the pres~ce of ether, and both the detergent and ether are subsequently removed by extensive dialysis. See, e.g., Huang et al., in Meths. Enzymol. (1985). The preparation of immunoliposomes is also described in U.S. Patent No.

4,957,735, which is incorporated herein by reference.
Preferably the antibody or bin~ing fragment thereof is attached to the liposome component after the liposome is prepared.
A variety of drugs or other pharmacologically active agents are suitable for incorporation into the liposome and thus delivery to the blood-brain barrier. The use of liposomes as drug delivery vehicles is extensively reviewed in Gregoriadis, (ed.), Liposomes as Drug Carriers: Recent Trends and Progress, John Wiley & Sons, NY (1988), which is incorporated by reference herein. Therapeutic agents which may be delivered include protein neurotrophic factors, e.g., W094/02178 ~ ~ PCT/US93/06942 nerve growth factor, to treat brain in~ury and degenerative disease~, enzymes to replace those lost through genetic defects causing metabolic storage diseases, e.g., Tay-Sachs disease, neurotransmitters and neuromodulators, e.g., dopamine and beta-endorphin for treating Par~n~on~ 9A~ ~
co~tions associated with pain, disorders of movement or co~n1tion and behavior, antibiotics, chemotherapeutic agents, diagnostic imaging agents, etc. Particularly useful in the present invention are drugs such as peptides which have specific effects in the brain but which poorly cross into the brain normally, and have no or little effect in other organ~.
These com~oun~s are often readily broken down in the bloodstream, and thus benefit from the protection provided by encapsulation in a liposome of the present invent~ on. When targeted to a brain transport ~ystem as described herein, even a small amount crossing into the brain can provide a desired pharmacological effect.
One general class of drugs include water-soluble, liposome-permeable com~o~,~s which are characterized by a tendency to partition preferentially into the a~
compartments of the liposome suspension, and to equil~brate, over time, between the inner liposomal spaces and outer bulk phase of the suspension. Representative drugs in this class include propranolol, ibuprofin, gentamicin, tobramycin, penicillin, theophylline, bleomycin, etoposide, n-acetyl cysteine, verapamil, vitamins, and radio-opaque and particle-emitter agents, such as chelated metals. Because of the tendency of these agents of equilibrate with the aqueous composition of the medium, it is preferred to store the liposome composition in lyophilized form, with rehydration shortly before administration. Alternatively, the composition may be prepared in concentrated form, and diluted shortly before administration.
A second general class of drugs are those which are water-soluble, but liposome-impermeable. For the most part, these are peptide or protein molecules, such as peptide hormones, enzymes, enzyme inhibitors, apolipoproteins, and higher molecular weight carbohydrates characterized by long-wog4/o21~8~ ~2 ~6 1~ PCT/US93/069 term stability of encapsulation. Representative compounds inth~s class include nerve growth factor, interferon (~, ~ or y), oxytoc~n, vasopressin, insulin, interleukins (e.g., IL-l, IL-2, etc.), ~uperoxide dismutase, tumor necro~is factor, somatostatin, thyrotropin releasing hormone, and macrophage colony stimulating factor, among~others. Peptide molecules and hormones which are typical~ very potent, and thus even small amounts crossing the blood-brain barrier can affect function, are particularly useful in targeted methods and compositions of the present invention. They include, for example, the ~-endorphins and analogues, the ~nkerhAlins (relatively lipid soluble) such as Leu- and Met-enkçph~lins and analogues, melanocyte-stimulating hormone and melanocyte-stimulating hormone inhibitory factor, ~-casomorphin, glucagon, delta sleep-inducing peptide and others, some of which are disc~ in R~nkc and Kastin, supra, incorporated by reference herein.
A third class of drugs are lipophilic molecules which tend to partition into the lipid bilayer phase of the liposome, and which are therefore ~soçiated with the liposomes predominantly in a membrane-entrapped form. The drugs in this class are defined by an oil/water partition coefficient, as measured in a st~n~rd oil/water mixture such as octanol/water, of greater than 1 and preferably greater than about 5. Representative drugs include prostagl~A i n~
amphotericin B, methotrexate, cis-platin and derivatives, vincristine, vinblastine, progesterone, testosterone, estradiol, doxorubicin, epirubicin, beclomethasone and esters, vitamin E, cortisone, dexamethasone and esters, betamethasone valerete and other steroids, etc.
Following treatment to remove free drug and/or targeting antibody molecule, the liposome suspension is brought to a desired concentration in a pharmaceutically acceptable carrier for administration to the patient. The pharmaceutical compositions are intended for parenteral, topical, oral or local administration, but preferably they are administered parenterally, e.g., intravenously, intraarterially or intramuscularly. Thus, this invention 21~1216 WO94/02178 ~ PCT/US93/06942 provides compositions for parenteral administration targeted for the blood-brain barrier which comprise a solution of a selected pharmaceutical cont~ne~ in liposome associated with an antibody molecule or b~n~n~ fragment thereof which binds to a brain receptor mol~ec~ule, preferably the transferrin receptor molecule, dissolved in a pharmaceutically acceptable carrier, preferably an aqueous carrier. A variety of aqueous carriers may be used, e.g., water, buffered water, 0.4%
saline, 0.3~ glycine, and the like, including glycoproteins for e~h~c~ stability, such as albumin, lipoprotein, globulin, etc. These compositions may be sterilized ~e~h~ques referred to above or produced under sterile conA~tions. The resulting aqueous solution~ may be packaged for use or filtered under aseptic conditions and lyophilized, lg the lyophilized preparation being combined with a æterile aqueous solution prior to administration. The compositions may contain pharmaceutically acceptable auxiliary substances as required to a~L~imate physiological conditions, such an pH ad~usting and buffering agents, tonicity ad~usting agents and the like, for example, sodium acetate, sodium lactate, sodium chloride, potassium chloride, calcium chloride, etc.
The concentration of the liposomes in these formulations can vary widely, i.e., from less than about 0.5~, usually at or at least about 1% to as much as 15 or 20% by weight and will be selected primarily by fluid volumes, v1~cosities, etc., in accordance with the particular mode of administration selected. Actual methods for preparing parenteràlly administrable liposomes formulations will be known or apparent to those skilled in the art and are described in detail in, for example, Remington's Pharmaceutical Science, 17th ed., Mack Publishing Company, Easton, PA (1985), which in incorporated herein by reference, and Gregoriadis (1988), su~ra.
The dose and the route of administration and the carrier used may vary based on the disease being treated and in view of known treatments for such diseases. Amounts effective for a particular use will depend on the severity of the disease and the weight and general state of the patient WO9~/0217~2~6 16 PCT~US93/069 being treated. It must be kept in mind that the materials of the present invention may be employed in 6erious disQase states, that i8, life-threa~en~g or potentially life threatening ~ituations. In such case~, in view of the ability of the liposomes to target the blood-brain barrier and the minimal immunogenicity of the antibo~dy b1n~ fragments, it i8 possible and may be felt desira~ie by the treating physician to administer substantial ~Y~eC~- of these compositions.
The following examples are provided by way of illustration, not limitation.

EXAMPLE I
Production of Ant~-Transferrin Rece~tor Fab'-Tar~eted Tmmunoliposomes This Example describes the production of Fab'-coupled immunoliposomes for targeting to the transferrin receptor.

A. Preparation of Liposomes Liposomes were prepared with a Heat Systems W 380 cup horn sonicator (Farmington, NY) by a method modified from Barrow and Lentz, Biochim. Biophys. Acta, 597:92-99 (1980).
L-a-Phosphatidylcholine (PC) and N-t4-p-malimido phenyl butyryl] dioleoyl phosphatidylethanolamine (MPB-PE) were obtained from Avanti Polar Lipids, Inc. (Pelham, AL), and cholesterol (Chol) was pur~h~se~ from Sigma Chemical Co. (St.
Louis, M0). Aliguots of chloroform (or chloroform/methanol) solution containing the appropriate lipids (PC/Chol/MPB-PE
lipids at a molar ratio of 10/5/1), were pipetted into 5-ml glass test tubes. ~14C]-labeled dipalmitoylphosphatidylcholine (DPPC) (purchased from Amersham, Arlington Heights, IL), was added to give a final aqueous concentration of 0.5 ~Ci/ml. Solvents werQ evaporated under nitrogen, and solvent traces were removed under vacuum.

~ W094/02178 2 1 4 1 2 1 6 PCT/US93/06942 Lipid films were di~persed in 0.9% (w/v) saline (or Hepes-buffered 0.9% saline), transferred to 10 ml polycarbonate test tubes in 1-2 ml aliquots, and sonicated - continuously at 50-C for 2-3 min to visual clarity. Liposomes were ch~ckeA for diameter size by filtration through 0.45 micron-diameter cellulose acetate filters (Millipore, Boston, MA) . ", ' B. Preparation of Antibody Fragments Ftab')2 fragments were prepared from "OX-26"
anti-transferrin receptor monoclonal antibody (purchA~ from Bioproducts for Science, Inc. (TnA~n~rolis, IN) by Loftrand Lab LTD (Gaithereburg, ~D). Nonspecific rabbit IgG F(ab')2 was purch~e~ from Organon Te~n~k~-Cappel (Durham, NC). Fab' fragments were generated from OX-26 F(ab')2 (10 mg/ml) and nonspecific rabbit F(ab')2 fragments (10 mg/ml) by reduction with 20 mM DTT (Sigma) in 20 mM ~ S (pH 6.5) for 60 min.
Fab' fragments were separated from ~cefiC DTT on minicolumn~
of Sephadex G-25 (Pharmacia Fine Chemicals, U~ , Sweden), made from the barrels of 3-ml plastic syringes according to the te~hn~que of Fry et al., Anal. B~ochem. 90:809-815 (1987).
One-half ml aliquots of reduced Fab' were placed on previously centrifuged minicolumns and centrifuged at 1000 rpm for 10 min. The degree of reduction was determined by the method of Ellman, Arch. Biochem. 8iophys. 82:70-77 (1959).

C. Coupling of Fab' Fragments to Liposomes Freshly reduced Fab' fragments were added to equivalent volumes of freshly prepared liposomes (in 20 mM
Hepes-buffered saline, pH 6.5) and mech~nically stirred at room temperature overnight.
Fab'-coupled liposomes were isolated from the reaction mixture by flotation on a metrizamide (Sigma) gradient by the method of Heath et al., Biochim. Biophys. Acta 640:66-81 (1981). Briefly, 1 ml of liposomal suspension was mixed with 1 ml of 1 M metrizamide solution in 0.9% saline in a 5 ml nitrocellulose centrifuge tube. This solution was gently overlaid with 3 ml of 0.3 M metrizamide in 0.9% saline ~ ~ 4~2~6 18 and topped with 0.5 ml of 0.9% saline. The solution was centrifuged at 40,000 rpm for 1 h in a L8-80M ultracentrifuge (Beckman Instrument~, Palo Alto, CA). The liposomes were pipetted off the gradient, and the met`rizamide was removed on a SerhA~ G-25 min~column as previous~y described.
The protein co~ntrations of OX-26 anti-transferrin IgG Fab', and of nonspecific rabbit IgG Fab' (control) associated with the related liposomal formulations were estimated colorimetrically (Sigma Protein Assay Kit) to be 60 and 140 ~g/ml, respectively.
Fab'-coupled liposomal formulations were not tested for receptor site bin~in~. However, con~ugates prepared from SATA-modified proteins have been shown to retain enzyme activity or function (~c~ et al., Anal. Bioche~l 132:68-73 (1983)).

EXAMPLE II
T~calization in the Brain of Immunoliposomes Targeted to Transferrin Receptor This Example demonstrates that liposomes covalently coupled to Fab' fragments can bind to transferrin ~e~ors in brain capillaries. Rats were used in these studies to demonstrate localization of the antibody fragment-targeted liro~Qmes. Rats have a brain capillary system that, although smaller overall than the capillary system found in a human brain, is not substantially different. Further, radiolabeled OX-26 monoclonal antibody has been shown to bind isolated brain capillaries from rats and humans to approximately the same extent. Pardridge et al., J. Pharmacol. Ex~. Ther.
259:66-70 (l99l).

A. Intracarotid Infusion in Rats Three-month-old male Sprague-Dawley rats (Charles River Br~e~ing Laboratories, Lakeview, NJ) weighing approximately 200-300 g were given free access to food and ~ 2141216 water and acclimated for at lea~t 2 days to the Laboratory of Neurosc~ences animal room, where temperature, humidity and light cycles (6-18 h) were controlled. Experiments were conducted consistent with NIH guidelines, under a protocol approved by the NICHD Animal Care and Use Committee.
Intraarterial liposomal administration was undertaken to maximize liposome delivery and b~ ng to the ipsilateral cerebral vasculature. The advantage of this administration route occurs only during the initial passage of liposomes through ipsilateral brain. Liposomes then enter the systemic circulation and show pharmacokinetic behavior ~imilar to liposomes administered intravenously (Greig, N. in Implications of the Blood-Rrain Barrier and ~ts Manipulation, vol. 1: "Basic Science Aspects, n (Neuwelt, E., ed.), pp. 311-368, Plenum Press, New York, NY (1989), which i8 incorporated herein by reference).
Rats anesthetized by i.p. in~ection of 50 mg/kgNa-pentobarbital (Ayerst, New York, NY) had sodium heparin-filled (100 I.U. in isotonic saline) catheters (size PE-50) inserted ~nto the right external carotid and femoral arteries, according to the procedure of Rapoport et al., J. Phvsio~. 238:R421-R431 (1980).
Liposomal formulations containing trace levels (0.S
~Ci/ml) of 14C]-labeled DPPC were administered by infusion into the carotid artery catheter at a rate of 0.2 ml/min for 5 min (infusion pump No. 944, Harvard Instruments, Millis, MA).
The carotid bifurcation was observed to ensure that the infusate passed cephalad into the internal carotid artery and not in a retrograde fashion down the common carotid artery.
By li nki ng the carotid catheter to a pressure transducer (Statham Instruments, Oxnard, CA), carotid artery pressure was monitored to ensure that hypertensive blood-brain barrier opening did not occur. After 1 or 10 min, the rats were killed by decapitation, and brain, kidneys, liver, and spleen were removed from each animal and were weighed. Plasma, brain (cerebrum ipsilateral and contralateral to infusion si~e, and cerebellum), other organ samples, and aliquots of liposomal formulations were measured for radioactivity (Tri-Carb 2200 CA
.

WOg4/02178 ~ 6 PCT/US93/069 ~

Liquid Scintillation Analyzer, Packard Instruments, Downers Grove, IL) according to the techn~gue of Greig et al., J.
Pharmaaol. ~D. Ther. 245:1-7 (1988), which is incorporated herein by reference.
B. Calculations Radioactivity in organ samples was calculated as mean DPM/total DPM injected. Volumes of di~tribution (Vd) for tl4C]-DPPC-labeled liposomal formulations were calculated as follows: Vd-DPM per gram wet ~c~ weight/DPM per ml plasma x 100 (% wet wt). In the calculation of radioactivity distributions, blood volume was assumed to be 6.5% of the total body weight (Altman, Blood and Other BodY Fluids, Biological ~n~hook Serie~, Fed. Amer. Soc. Exp. Biol., Wash., D.C. (1961)). Ti~sue counts were corrected for counts present in residual blood (0.01 in brain, 0.1 in spleen, kidneys, and liver (~ki~-) Assuming the liposomal diameter to be approximately 1000 A, the lip~d bilayer thickness 37 A (M~r~n and Huang, Ann. N.Y. Acad. Sci. 308:29-48 (1978)), and the lipid (liposomal) density 0.9 g/cm3, the number of l~posomes bound to the blood-brain barrier and the surface area occupied by bound liposomes were estimated. Capillary surface area in rat brain was estimated as 135 cm2/g wet weight brain (Altman, supra).
Dunnett's test was carried out to compare two means (Miller, R. Simultaneous Statistical Inferences pp. 67-87, McGraw-Hill, New York, Ny (1966)). A paired t-test was used to compare the perfused with unperfused sides of the brain of each rat. The value for statistical significance in all tests was P < 0.05.
The organ distributions of t14C]-DPPC-labeled liposomal formulations following right-sided carotid infusion in 10 rats for each study are summarized In Table I. OX-26 liposomes showed a 50~ greater association with brain than the corresponding nonspecific Fab' controls at 1 and 10 min.
Levels of radioactivity in brain of rats infused with OX-26 liposomes were statistically greater than in controls at 1 min ~ 2141216 21 ` J
(P < 0.05, Dunnett's tect) but not at lO min. Although brain levels were essentially llnch~nged at lO min, statistical significance was not achieved because of greater variability Ac~ociated with the data at lO min.
S ~eft-right differences in cerebrum distribution values in individual animals were assessed by paired t-test.
Although values for cerebrum ipsilateral to infusion (right) were greater than those of contralateral (left) cerebrum, these did not reach statistical signif~c~nc~. OX-26 and OX-26 control liposomes had similar liver uptakes; however, the spleen uptake of OX-26 liposomes was approximately 300~ of controls at l and lO min. OX-26 liposomes cleared from plasma more rapidly than controls. There was insignificant uptake of all liposomal formulations by kidney. Recoveries of all liposomal formulations averaged 75 + 4% ISEM) of the in~ected dose.
Vd values, radioactivity per g wet weight of each organ divided by radioactivity per ml of plasma, normalize variability in plasma pharmaco~netics and liposomal plasma clearA~cec among rats. The Vd values of the liposomal formulations (Table II) for oX-26 in brain were significantly greater than their respective controls at both l and lO min, by approximately 2-fold at l min and by 6-8-fold at lO min (p ~ 0.05, Dunnett's test). Although the Vd value for right cerebrum was up to 3.5 fold greater than left cerebrum, this did not reach ætatistical significance (paired t-test). At l and lO min, high levels of OX-26 liposomes were present in liver and spleen, indicating liposomal accumulation in these organs. Vd values of all formulations in brain tissue did not increase between l and lO min, indicating that liposomes did not measurably cross the blood-brain barrier.
The foregoing results indicate that liposomes covalently coupled to antibody binding fragments, and particularly Fab' fragments can bind to transferrin receptors in brain capillaries. Bound liposomes were estimated to be approximately 0.03% of the injected dose or 0.025 pmol of liposomes. Control liposomes associated with rat brain were subtracted from total brain uptake of OX-26 liposomes because ~ 22 PCT/US93/069 brain radioactivity level5 are slightly elevated (above background) after intracarotid infusion (Micklus et al., Riochim. Biophys. Acta 1124:7-12 (19923). Transferrin receptor density in rat brain has been estimated to be about 0.02 pmole per gram of brain (Pardridge et al., J. Pharmacol.
~. Ther. 259:66-70 (1991)), generally equivalent to 0.034 pmole per brain for the rats use~ in the present experiments, or approximately 1 liposome per~ 1.3 receptors. OX-26 liposomal brain uptake remained constant rather than increasing after 1 min, suggesting that essentially all the receptor sites were occupied after 1 min. This result is consistent with the observation that OX-26 antibodies show significant binding to capillaries at 5 min after in;ection (Friden et al., Proc. Natl. Acad. SC1. USA 88:4771-4775 (1991)). No increased brain liposomal uptake was observed during the brief 10 min period used in this ~tudy.
Comparison of intra-animal left to right cerebrum distributions and Vd values of our liposomal preparations showed no statistical difference, indicating that receptor sites are rapidly saturated in both cerebral hemispheres.
This suggests that similar receptor site saturation could be achieved by (1) intravenous administration of liposomes and/or (2) a reduction in the liposomal dose.
The capillary surface area in rat brain cortex ~s about 135 cm2 per gram of brain (wet wt), or 230 cm2 for rats in the present study; therefore, it was estimated that about 0.7~ of brain capillary surface was covered with liposomes.
Liposomal binding to brain can be increased to cover all of the capillary surface by upregulating the transferrin receptor or targeting other receptors, e.g., in a 270 g rat a~ used in this study, this would represent 3.8 pmol of liposomal brain uptake.
Transport of the Fab'-targeted liposome across the blood-brain barrier is demonstrated in experiments performed according to the brain capillary depletion procedure as described in Pardridge et al., J. Pharmacol. Exp. Therap.
259:66-70 (1991), and Triguero et al., J. Neurochem. 54:1882-1888 (1990), which are incorporated herein by reference.

In other a~pects, OX-26 Fab'-coupled liposomes showed ~ignificant liver and spleen uptakes at l and lO min after infusion. OX-26 liposome3 demonstrated an unu~ually high spleen uptake, with 6~ of their total radioactivity present in the spleen after l min, and 12% after lO min. This enhanced uptake i8 likely due to the many transferrin receptors in spleen or on red blood cell~, which are rapidly cleared by the spleen after b~ n~ ~n~ to OX-26 liposomes. This suggests that OX-26 liposomes can be used to effectively target therapeutic agents to the spleen.

EXAMPLE III
~posomes Tarqeted to Tnsulin ~eceptor This Example compares the ability of insulin-targeted liposomes to distribute labeled formulations in brain when compared to control preparations. Al~ho~h the brain distribution of insulin-coupled liposomes wa~ not statistically different from the ~onL~ol formulations, Vd values, which adjust data for plasma clearance, were higher for insulin liposomes in right cerebrum than controls at l and lO min, suggesting that a moderate binAin~ to brain capillaries does occur.

Formulations The tl4C]-DPPC-labeled liposomal formulation~
evaluated in this study consisted of 8 ~mol/ml suspensions of PC/Chol/MPB-PE lipids (molar ratio, lO/S/l), covalently coupled to proteins. The protein concentrat~ons of insulin, of oX-26 anti-transferrin IgG Fab', and of nonspecific rabbit IgG Fab' (control) associated with the related liposomal formulations were estimated colorimetrically (Sigma Protein Assay Kit) to be 250, 60 and 140 ~g/ml, respectively.

wog4/~7~ 4~ZI6 24 PCT/US93/0694 A. Preparation of Thiolated Insulin Insulin (porcine, ~odium) and N-succinimidyl-S-acetyl-thioacetate (SATA) and were obtained from Cal Biochem Corp. (LaJolla, CA). Insulin was thiolated with SATA
by a modification of the method of -~nC~n et al., ~n~l.
RiQchem 132:68-73 (1983). Twenty ~l of SATA in dimethyl formamide (DMF) (100 mg/ml) was reacted per 1 ml of insulin solution (10 mg/ml insulin + S ~1 5 N NaOH) in deionized water for 30 min. Thioacetylated insulin was deacetylated py reaction with 100 ~1 (per ml insulin solution) of aqueous 0.5 M hydroxylamine (pH 6.8) for 30 min. Thiolated insulin was separated from low molecular weight solutes on minicolumns of Sephadex G-25, made from the barrels of 3-ml plastic syringes according to the ~echn~que of Fry et al., ~nal. Biochem~
90:809-815 (1978). One-half ml aliquots of th~olated insulin solution wer~ placed on previously centrifuged minicolumns and centrifuged at 1,000 rpm for 10 min. After addition of 0.2 ml of ~aline, this centrifugation stop was repeated. The extent of thiolation was determined by the method of Ellman, supra.
B. Coupling of Insulin to Liposomes Freshly thiolated insulin was added to an equivalent volume of freshly prepared liposomes (in 20 mM Hepes-buffered saline, pH 7.5) and mechAn~cally stirred at room temperature overnight. Insulin-coupled liposomes were isolated from the reaction mixture by flotation on a metrizamide gradient as described in Example I above.
The protein concentration of insulin associated with the liposomal formulation was estimated colorimetrically to be 250 ~g/ml. Insulin-coupled liposomal formulations were not tested for receptor site binding. However, conjugates prepared from SATA-modified proteins have been shown to retain enzyme activity or function (~ln~n et al., su~ra).
Animal studies were performed in accordance with the procedures described in Example II above.
The results, shown in Table I and Table II, indicate that insulin and insulin control liposomes had comparable distributions in brain and, similarly, there were no WO94/02178 2 1 ~ 1 2 1 6 PCT~US93/06942 significant riqht-left cerebrum differences following administration of either preparation. Insulin liposomes cleared from plasma more rapidly than control liposomes and had greater liver and spleen uptakes.
Vd values of insulin liposomes in right cerebrum J were significantly higher than control values at l and lO min by approximately 2-fold (P < 0.05, ~nnett'~ test). Values in left cerebrum, however, were not different from controls, nor were there left-right differences in cerebrum Vd values.
The brain distribution of insulin-coupled liposomes was not statistically different from the control formulations, suggesting that they do not significantly bind to rat brain capillary. However, Vd values, which ad~ust data for plasma clearance, for insulin liposomes in right cerebrum (ipsilateral to liposome infusion) were higher than controls at l and lO min, suggesting that a moderate b~nA1ng to brain capillaries does occur. Possible reasons for such moderate b~n~ng are (l) low insulin bin~1ng affinity between liposome and capillary receptor (2) partial inactivation of the insulin during the liposomal coupling reaction, (3) low insul~n receptor density at the cerebral capillary, and (4) competition between insulin liposomes with endogenous insulin for brain capillary receptor sites.

EXAMPLE IV
r~posomes Targeted Via Antibody Fraqments to Insulin Receptor A liposome is prepared according to Example I, incorporating a transportable neuropeptide to be targeted to the brain of an animal. A monoclonal antibody to human insulin receptor is digested to prepare Fab' fragments which are coupled to the liposome as described in Example I. The resulting targeting immunoliposome is combined with sterile saline to provide a physiologically acceptable solution, and is administered to the patient or ~ubject parenterally.

Wo94/02178 PCT/US93/069 4~2 ~6 26 EXAMPLE V
osomes Targeted Via Antibody Fraqments to IGF-I Receptor A liposome i8 prepared according to Example I, incorporating a pharmacological com~oul.d to be targeted for delivery to the brain of a patie~t or sub~ect. A monoclonal antibody to human IGF-I receptor ls digested as described in Example I to prepare Fab' fragments. The Fab' fragment are coupled to the l~o^ome a~ described in Example I. The resulting targeting immunolipo-Q~? is combined with ~teril~
saline to provide a phyciologically acceptable solution, and is administered to the patient or ~ub~ect parenterally.

WO 94/02178 2 7 2 1 4 1 2 1 6 PC,lUS93,06942 d ~ ~ O ~, ~ ` O

O P
2 ~ ~ ~ ~ ~ ~ ~ u ~- ~ ~ ~ ~ t` o ~ o o o o ~ o ~ 7 8 ~ . . . o .
, I o ,~ u O o O a S 3!

O O O O 1 ~1 ~ o ~ ~
3 ~2 ~ . . O . O

O ~ ~ O O C: --I a 3~ a 0O~O ~0~O
~ ~ ~ o a a ~ -- ~ I! a a O ¦ a '' 3 ~ ~

WO 94/02178 2 8 PCr/US93/0694~
~ 2~

~ N ~ ~ ~ N U

g~ N 1` 1~
O ~ O
C ~ ~ O
O ,~ ~ , C O t~
o ~ ~' U~ O

H C ~ ~1 ~ ~ C o ~ o 'I ~ ;
a ~ " ' ' - ' ~ 5 v o o ~ o O u 5 0 ,~o o ~ ~ O O ~ O O O _ O

9 8 jOg :~

2 1~ 2 c WO94/02178 21 41 Z I 6 ` PCT/US93/~942 It i8 evident from the above results that the ~ub~ect invention provides antibody b~n~tng fragment-targeted liposomal formulations useful in the delivery of therapeutic~
and diagno~tic compo~n~ across the blood-brain barrier. In particular, the pr~-^nce of antibody b~nA~ng fragment~ in a liposome which are specific for brain transport receptor, such as transferrin receptor, insulin receptor, IGF-I or IGF-II
receptor provides an effective means to e~h~ce the uptake of certain drugs, such as hydrophilic drugs, peptide~ and proteins to the brain. The antibody bin~ng fragment-modified liposomal formulations may be pro~-~c~ by a variety of means.
Thus, increased efficacy, convenience and economics of lower dosages and less frequent administration to achieve therapeutic levels of drugs in the brain are among the lS advantages provided by compositions and methods of the present invention.
Although the foregoing invention has been described in some detail by way of illustration and example for ~u.~o~es of clarity of unders~ ng, it will be obvious that certain rh~es and modifications may be practiced within the scope of the app nd~ claims.

Claims (20)

WHAT IS CLAIMED IS:

1. A composition which comprises a liposome coupled to an antibody binding fragment which binds to a receptor molecule present on vascular endothelial cells of a mammalian blood-brain barrier.

2. The composition of claim 1, wherein the antibody binding fragment is Fab, F(ab')2, Fab' or a single antibody chain polypeptide.

3. The composition according to claim 1, wherein the antibody fragment binds to a receptor of the brain peptide transport system.

4. The composition according to claim 3, wherein the receptor is transferrin receptor, insulin receptor, IGF-I
or IGF-2 receptor.

5. m e composition according to claim 1, wherein the antibody binding fragment is coupled by a covalent bond to the liposome.

6. The composition according to claim 1, wherein the liposome contains a pharmaceutical compound.

7. The composition of claim 6, wherein the pharmacologic compound is a hydrophilic neuropeptide.

8. The composition according to claim 6, further comprising a physiologically acceptable carrier.

9. The composition according to claim 1, wherein the diameter of the liposome is less than 1 micron.

10. The composition according to claim 9, wherein the diameter of the liposome is smaller than about 0.45 microns.

11. The composition of claim 1, wherein the antibody binding fragment is Fab, F(ab')2 or Fab' which binds specifically to transferrin receptor.

12. The composition of claim i, wherein the antibody binding fragment is prepared from a monoclonal antibody.

13. A pharmaceutical composition which comprises a antibody binding fragment which binds to a receptor molecule present on vascular endothelial cells of a mammalian blood-brain barrier, wherein the antibody binding fragment is coupled to a liposome which contains a pharmaceutical compound intended for treatment of a brain-associated disorder, and a pharmaceutically acceptable carrier.

14. The pharmaceutical composition of claim 13, wherein the antibody binding fragment is Fab, F(ab')2, Fab' or a cingle antibody chain polypeptide.

15. The pharmaceutical composition of claim 14, wherein the receptor is transferrin receptor, insulin receptor, IGF-I or IGF-2 receptor.

16. A method for targeting a pharmacological compound for delivery to the blood-brain barrier of a mammal, which comprises:
administering to the mammal in need of the pharmacological compound a composition which comprises a liposome containing the pharmacological compound coupled to an antibody binding fragment which binds to a receptor molecule present on vascular endothelial cells of the mammalian blood-brain barrier.

17. The method of claim 16, wherein the composition is administered intravenously or intraarterially.

18. The method of claim 16, wherein the antibody binding fragment binds specifically to transferrin receptor, insulin receptor, IGF-I or IGF-2 receptor

19. The method of claim 11, wherein the antibody binding fragment is Fab, F(ab')2, Fab' or a single antibody chain polypeptide.

20. The method of claim 16, wherein the receptor is the transferrin receptor.