The first part of this article (PJ, August 28, p309) discussed the use of liposomes, non-ionic surfactant vesicles and solid nanoparticles and microparticles. Parenteral drug delivery technologies discussed in this second part include emulsions, prodrugs, implants and needle-free injections.
Emulsions are dispersions of one liquid inside a second liquid, the liquids being immiscible. Such dispersions (oil in water or water in oil) are stabilised by emulsifiers which coat the droplets and prevent droplet coalescence by either reducing interfacial tension or creating a physical repulsion between droplets.176 Emulsions are usually used as a means of administering aqueous insoluble drugs by dissolution of the drugs within the oil phase177-180 or to prevent drug hydrolysis or drug uptake by infusion sets.181 It is recommended that emulsions destined for the intravenous (IV) route have a submicron droplet size,182 although emulsions with a droplet size of 10mm have been used parenterally.
|
Figure 9: Targeted emulsion droplet bearing covalently linked antibodies |
Emulsions may also be used to reduce drug toxicity. Use of a water in oil emulsion of amphotericin deoxycholate, as opposed to a solution, reduces the incidence and severity of renal impairment and chills in patients while still maintaining the antifungal efficacy of the drug.186
In summary, although emulsions are chiefly seen as vehicles for administering aqueous insoluble drugs, oil in water emulsion formulations with a submicron droplet size may be used for drug targeting by the attachment of targeting ligands to the droplet surface.
|
Cyclodextrins (Figure 10) are water-soluble cyclic carbohydrate compounds with a hydrophobic cavity due to the specific orientation of the glucosidic substituents. These compounds form inclusion complexes with hydrophobic guest molecules (reviewed by Albers, 1995187), endowing such molecules with aqueous solubility. Only the modified cyclodextrins,188 such as hydroxypropyl b-cyclodextrin189,190 and sulphobutyl b-cyclodextrin,191 are regarded as safe for parenteral use.
|
Figure 10: b-cyclodextrins. Sulphobutyl b-cyclodextrin, R = CH2CH2CH2CH2SO3Na or H (average degree of sulphobutyl substitution = 4). Hydroxypropyl b-cyclodextrin, R = CH2CHOHCH3 or H (varying degrees of substitution at the 2, 3 and 6 positions) |
Cyclodextrin inclusion complexes encapsulated within liposomes195 reduce drug urinary clearance although it is not clear how the liposomal encapsulation affects drug blood levels.
The use of prodrugs in cancer chemotherapy as a means of targeting relatively toxic compounds to specific areas of pathology is enjoying renewed activity. Two of the technologies being evaluated at present are antibody directed enzyme prodrug therapy (ADEPT) and the use of polymeric prodrugs (or polymer drug conjugates as they are more commonly known).
Antibody directed enzyme prodrug therapy (ADEPT) The principle behind the ADEPT approach is shown schematically in Figure 11. Basically, an antibody-enzyme conjugate is administered intravenously, localises in tumour tissue and subsequently activates an administered prodrug predominantly within such tumours.196 Prodrug activation occurs on the cell surface or in the extracellular fluid, which is in marked contrast to the polymeric prodrug approach where prodrug activation occurs intracellularly (see below)119 The appearance of the active drug after prior administration of the antibody-enzyme conjugate to patients confirms the feasibility of the ADEPT approach.86,196,197 To promote specificity, a non-human, eg, bacterial, enzyme such as carboxypeptidase G2 is used to activate the prodrug.196
![]() |
![]() |
Figure 11: Principle behind ADEPT approach to drug targeting. Step 1 — injection of antibody enzyme conjugate; step 2 — activation of the prodrug |
|
However, such enzymes may promote an immune response,198 necessitating the administration of immunosuppressants.199 Another method of reducing the immune response seen with the bacterial enzyme is to use a mutant human form of carboxypeptidase A.200 Unfortunately, this enzyme is unstable in vivo and hence there is little tumour reduction when it is used as an antibody conjugate with a methotrexate prodrug.200
For ADEPT to be effective clinically, the prodrug should be non-toxic and the enzyme should locate only at tumour sites. Tumour associated antigens are rarely confined to tumour tissue alone199 and hence some enzyme presence in normal tissue is experienced in practice. Unwanted enzyme may be removed from non-tumour tissue by the administration of an anti-enzyme antibody.86 A reduction in the toxic effects seen with ADEPT86 may be achieved if the active drug has a short half life, although not to such an extent as to prevent tumour cell kill. The active drug must thus be rapidly cleared or inactivated after production in vivo. The inactivation of residual active drug within the vascular compartment by the administration of specific inactivating agents has been proposed.199
In patients, the IV administration of an antibody-enzyme conjugate, followed 36 to 48 hours later by an IV dose of a galactosylated anti-enzyme antibody, and finally a benzoic acid mustard prodrug, results in temporary regression of advanced disease.86 The galactosylated anti-enzyme antibody conjugate was administered to enable enzyme clearance from non-tumour sites via the liver galactose receptor.86 Unfortunately the long plasma half life of the active drug causes myelosuppression due to active drug migration from tumour sites to the bone marrow.86
Active drug will also have an effect on tumour cells devoid of tumour-specific antigen but in close proximity to tumour cells bearing such antigens, an effect known as the bystander effect (Figure 11).
In summary, proof of principle studies of the ADEPT approach have been conducted in the clinic, although problems such as the immunogenicity of the non-human enzyme and the long half life of the active drug leading to toxic sequelae still remain to be addressed.
Polymeric prodrugs (polymer drug conjugates) Drug delivery with polymeric prodrugs, first envisioned 25 years ago,201 involves the use of an active substance and possibly a targeting moiety, both linked via spacers to a water-soluble polymeric backbone (Figure 12). From this basic blueprint a number of polymer drug conjugates for cancer chemotherapy have been synthesised with cleavable drug polymer linkers. These include soluble polymeric prodrugs of daunorubicin,202 doxorubicin,203 cisplatin 204 and 5-fluorouracil,205 to name but a few. Polymer drug conjugates accumulate selectively within tumour tissue,203,206 leaking through the disorganised vasculature in a similar manner to that shown in Figure 2 for liposomes (PJ, August 28, p310). Clearance from tumour tissue is delayed due to the poor lymphatic drainage. Tumour accumulation of polymer drug conjugates has been termed the enhanced permeation and retention effect.207 On IV administration the conjugate is taken up by tumour cells and the active drug released intracellularly.119 It is not clear whether passive tumour targeting of polymer drug conjugates is influenced by the nature of the polymer backbone. However, passive tumour targeting is increased with increase in polymer molecular weight.208 Most of the polymeric backbones explored are prepared from non-biodegradable materials. Obviously, biodegradable polymers will be more acceptable, although care must be taken to ensure that biodegradation does not hamper the accumulation of conjugates in tumour tissue.
|
Figure 12: Polymer drug conjugates |
Actively targeted drug polymer conjugates bearing galactose target hepatocytes212,213 and antibodies specific for an ovarian carcinoma cell line214 bind to ovarian carcinoma cells. Unfortunately, galactose-bearing polymer drug conjugates do not target hepatic carcinomas to any significant extent and preferentially locate in the normal liver tissue of mice215 and patients.216 This is evidence that the galactose receptor is not sufficiently upregulated in neoplastic tissue. It is possible that this liver targeting strategy may thus find a home in the treatment of non-neoplastic diseases.
In summary, polymer drug conjugates have progressed from an elegant scientific concept to the clinic and may result in a new form of therapeutics for routine use.
Amphiphilic block copolymers such as the Pluronics (polyoxyethylene polyoxypropylene block copolymers) self-assemble into polymeric micelles (Figure 13). For drug delivery purposes, hydrophobic drugs may be solubilised within the core of the micelle217,218 or, alternatively, conjugated to the micelle-forming polymer.219 Although micelles are rather dynamic systems which continuously exchange units between the micelle structure and the free units in solution,176 polyoxyethylene polyaspartic acid micelles are sufficiently stable in the blood to alter the pharmacokinetics of solubilised drug.220 They thus circulate for prolonged periods221,222 and deliver more of the drug to tumour tissue when compared with administration of the drug in solution.222 Pluronic micelles solubilising epirubicin and doxorubicin increase the tumoricidal activity of these anticancer drugs218 and polyoxyethylene polyaspartic acid micelle-forming block copolymers bearing covalently attached doxorubicin at the polyaspartic acid end reduce the toxicity of doxorubicin in vivo.219
|
Figure 13: Polymeric micelles — A = drug solubilised in hydrophobic micelle core; B = drug covalently linked to hydrophobic portion of polymer chain; C = polymeric micelle carrying antibodies attached to hydrophilic portion of polymer molecule |
Implants are used as depot formulations either to limit high drug concentrations to the immediate area surrounding the pathology223-226 or to provide sustained drug release for systemic therapy.227-229 Clinically, implant systems have been used in situations where chronic therapy is indicated, such as hormone replacement therapy229 and chemical castration in the treatment of prostate cancer.227
Parenteral implants may take the form of highly viscous liquids224 or semi-solid230 formulations, both of which may be injected with a needle. Alternatively, implants may be in the form of tiny rods impregnated with drug substances223 or a liquid which gels in situ.231-233 In situ forming gels either gel when the polymer solubilising solvent diffuses away from the injection site, leaving the polymer in contact with an aqueous environment in vivo,231,232,234 or gel on cooling after being injected at an elevated temperature.233 In situ forming gels may be used to prepare sustained release formulations of oligonucleotides234 and non-steroidal anti-inflammatory agents.231
Implants are prepared from a variety of polymeric materials, such as polysaccharides,235 polylactic acid co-glycolic acid,224,231,236 and the non-biodegradable methacrylates.223,227 Biodegradable materials, such as polylactic acid co-glycolic acid, are of course preferred as this removes the need for surgical removal of the implant after treatment has ended. However, non-biodegradable materials do provide therapeutic levels of drug for up to one year in vivo.227
Solid implants tend to have biphasic release kinetics, with an initial burst of drug followed by a slower release. The initial burst is usually due to the release of drug deposited on the surface of the implant236 although zero order release kinetics may be achieved by, for example, coating the implant with a drug impermeable material.228 Overall drug release may be controlled by varying polymer composition.236 An increase in the level of lactic acid in a polylactic acid co-glycolic acid copolymer236 tends to retard drug release and an increase in the polymer molecular weight also tends to retard drug release and prolong drug effects in vivo.237
Drug release may also be controlled by various stimuli, such as electrical stimuli in polyelectrolyte systems. This allows the fabrication of pulsatile delivery systems such as the electrically triggered release of insulin from polydimethylaminopropylacrylamide gels.238 Solid implants avoid the peak levels associated with the administration of the drug in solution,224,231,239 thus limiting the toxic effects associated with the free drug.
Implants are used for the delivery of anticancer agents as they are able to confine potentially toxic anticancer drugs to tumour sites and also allow sustained drug release. A viscous gelatin solution224 or galactoxyloglucan gel240 of mitomycin C administered intraperitoneally decreases peritoneal224,240 and plasma224,240 clearance. Tumoricidal activity of mitomycin C against a peritoneal ascites model was thus increased with use of the gelatin formulation in mice due to the presence of a depot of the gel in the peritoneum.224 Additionally, the intraperitoneal use of a 5-fluorouracil polyorthoester implant improves the tumoricidal activity of the drug in mice, although the use of high doses of the implant is associated with drug related toxicity.241 A biodegradable copolymer of polylactic acid polymandelic acid polyhydroxyphenylacetic acid also provides therapeutic levels of the anticancer drug estramustine for as long as 10 weeks.242
Implants may also be used to achieve high drug doses in traditionally inaccessible areas, such as the central nervous system,226 bone tissue224 and beyond the blood-retinal barrier.236 Ethylene vinyl acetate copolymer dexamethasone intracranial implants achieve high drug levels in the brain without correspondingly increased plasma levels226 and a polylactic acid co-glycolic acid gentamicin bone implant is superior to an intramuscular injection of the drug at eradicating bone infections in a canine model.224 Additionally, polylactic acid co-glycolic acid scleral implants of ganciclovir for the treatment of cytomegalovirus infection maintained sufficient therapeutic levels of the drug in the vitreous humour and retina/choroid236 for three to five months.
The use of implant systems is often associated with inflammation at the site of implantation.30,243 The development of this inflammation is thought to be necessary for the processing of the delivery system.243
In summary, implant systems, although having the obvious drawback of requiring administration via mini-surgical procedures, offer a means of achieving high drug concentrations in areas that are usually inaccessible to peripherally administered drug. In addition, the high drug levels are sustained in these areas.
Needle-free injection devices operate by using compressed gas to accelerate a small jet of liquid244 or powder245 at high speed, causing it to penetrate the skin for subcutaneous, intradermal or intramuscular (IM) administration. Although used for mass vaccination for a number of years, these devices are currently being promoted as devices for the self-administration of parenteral drugs. The development of disposable pre-filled systems246 should go some way towards achieving this.
Needle-less injection technology should allow the painless injection of liquids and solids and also remove the risk of needle stick injury to health care professionals. Some reports indicate an increased level of bruising247,248 and general discomfort244,248 with jet injectors while others report a decreased level of general discomfort with the newer jet injectors.247,249-252 Children seem to find the jet injectors more psychologically acceptable and easier to use than needles.247
Jet injectors for administration of liquids include the Intraject,246 Biojector244 and Medijector253 systems, while solids may be administered using Powderject’s technology.245 Fortunately, needle-free devices appear not to affect drug disposition when administered subcutaneously to patients251,253,254 and drug responses are also largely unaffected when administration takes place by this route.254,255 However, a reduced bioavailability was observed after needle free (Biojector) IM injection of interferon b-1a when compared with IM administration with a needle.244 It appears that jet injectors are unable to administer the dose into the IM compartment efficiently and hence larger doses are required for effective IM injection with these devices.256 Despite this, there have been reports that indicate that the intracavernous administration of drugs for penile dysfunction may be effected by the use of jet injectors.257
Powderject technology is currently being used to develop intradermal DNA vaccines245,258 and a vaccine development collaboration agreement was recently signed by Glaxo Wellcome and Powderject to exploit this technology.259 This needle-less powder delivery device is able to elicit cytotoxic T-cell responses with much lower doses of DNA than is required by conventional needle injections.245
Needle-free injection devices currently being tested for the painless self-administration of drugs appear to be efficacious for subcutaneous and intradermal use. These devices also offer the possibility of improving patient responses to intradermally administered vaccines.
The past 10 years have witnessed a real explosion in the number of technologies available to control drug biodistribution. These have been exploited to produce particulate, soluble and implantable drug delivery systems. Passively targeted systems which accumulate in the desired area of the anatomy or pathology due to the intrinsic properties of the materials/systems have been identified, as have actively targeted systems in which targeting ligands or targeting antibodies allow the materials to be directed to specific body sites.
Drug delivery technologies such as those discussed above may be used to extend the patent life of drugs but more importantly to control drug delivery on parenteral administration. Some of the newer systems, such as liposomal doxorubicin, may soon be licensed for new indications. Additionally, the beginning of the next century may see some new formulation/drug delivery initiatives, such as the polymer drug conjugates and possibly the ADEPT systems, transformed into commercial products. Liposomes may also one day open the door to routine gene therapy in the clinic.
ACKNOWLEDGMENTS The author gratefully acknowledges the contribution of Dr Andreas Schätzlein, Department of Medical Oncology, University of Glasgow, in the preparation of the illustrations accompanying the two parts of this article, and Dr Laurence Tetley, EM Unit, IBLS, University of Glasgow, for his expertise in the production of Figure 7.
Dr Uchegbu is lecturer in pharmaceutics in the department of pharmaceutical sciences, University of Strathclyde
References 1 to 175 were published with part 1 of this article (PJ, August 28, p309-318)