1-s2.0-S0378517398001690-main

International Journal of Pharmaceutics 172 (1998) 3370

Non-ionic surfactant based vesicles (niosomes) in drug delivery

Ijeoma F. Uchegbu a,*, Suresh P. Vyas b

a Department of Pharmaceutical Sciences, Uni6ersity of Strathclyde, Royal College, George St, Glasgow G1 1XW, UKb Department of Pharmaceutical Sciences, Dr Harisingh Gour Uni6ersity, Sagar 470003, MP, India

Received 1 May 1998; accepted 4 May 1998

Abstract

The self assembly of non-ionic surfactants into vesicles was first reported in the seventies by researchers in thecosmetic industry. Since then a number of groups world wide have studied non-ionic surfactant vesicles (niosomes)with a view to evaluating their potential as drug carriers. This article presents a summary of the achievements in thefield to date. Niosomes may be formed form a diverse array of amphiphiles bearing sugar, polyoxyethylene,polyglycerol, crown ether and amino acid hydrophilic head groups and these amphiphiles typically possess one to twohydrophobic alkyl, perfluoroalkyl or steroidal groups. The self assembly of surfactants into niosomes is governed notonly by the nature of the surfactant but by the presence of membrane additives, the nature of the drug encapsulatedand the actual method of preparation. Methods of niosome preparation and the number of different morphologiesthat have been identified are detailed. The influence of formulation factors on niosome stability is also examined asare methods to optimise drug loading. In vivo these systems have been evaluated as immunological adjuvants,anti-cancer:anti-infective drug targeting agents and carriers of anti-inflammatory drugs. Niosomes have also beenused in diagnostic imaging. Efforts to achieve transdermal and ophthalmic drug delivery with some formulations arealso discussed. 1998 Elsevier Science B.V. All rights reserved.

Keywords: Non-ionic surfactants; Niosomes; Drug delivery; Self-assembly

1. Introduction

Non-ionic surfactant based vesicles (niosomes)are formed from the self-assembly of non-ionicamphiphiles in aqueous media resulting in closed

bilayer structures (Fig. 1). The assembly intoclosed bilayers is rarely spontaneous (Lasic, 1990)and usually involves some input of energy such asphysical agitation or heat. The result is an assem-bly in which the hydrophobic parts of themolecule are shielded from the aqueous solventand the hydrophilic head groups enjoy maximumcontact with same. These structures are analogous

* Corresponding author. Tel.: 44 141 5483895; fax: 44141 5526443; e-mail: [email protected]

0378-5173:98:$ - see front matter 1998 Elsevier Science B.V. All rights reserved.

PII S0378-5173(98)00169-0

I.F. Uchegbu, S.P. Vyas : International Journal of Pharmaceutics 172 (1998) 337034

Fig. 1. Schematic representation of a niosome, hy-drophilic head group, --hydrophobic tail.

compounds as alternatives to phospholipids. Nio-somes were first reported in the seventies as afeature of the cosmetic industry (Vanlerberghe etal., 1972; Handjani-Vila et al., 1979) but havesince been studied as drug targeting agents. Thischapter reviews the relevant data on these systemsgenerated in our laboratories and those of otherswith the emphasis on the steps leading to thedevelopment of these systems as drug carriers.Areas to be covered are: non-ionic surfactantself-assembly, niosome preparation, toxicologystudies, specialised systems, stability and examplesof specific applications.

It is hoped that this chapter will introduce newresearchers to this topic and more importantlyoffer the industrial community an idea of thepotential utility of these systems as drug carriers.

The ultimate identity of any niosomal systemand hence its properties are determined by thefactors listed in Fig. 2. It is thus obvious that allthese variables must be carefully controlled in thedesign of a niosomal drug delivery system. Invari-ably drug delivery design leads should always betaken from the host biology (Fig. 3). Examples ofthis include the use of niosomes to target the liverand spleen in leishmaniasis (Baillie et al., 1986), as

to phospholipid vesicles (liposomes) and are ableto encapsulate aqueous solutes and serve as drugcarriers. The low cost, greater stability and resul-tant ease of storage of non-ionic surfactants (Flo-rence, 1993a) has lead to the exploitation of these

Fig. 2. Factors influencing niosome physical chemistry.

I.F. Uchegbu, S.P. Vyas : International Journal of Pharmaceutics 172 (1998) 3370 35

Fig. 3. Schematic flow diagram detailing the stages involved in drug delivery design.

particulate uptake by the liver and spleen is aknown fact. A further example is found in anti-cancer drug targeting with niosomes (Rogerson etal., 1988; Uchegbu et al., 1995) which exploits thespecific vascular architecture of tumour tissue.

It is hoped that more specific forms of targetingthat incorporate molecular recognition elementsmay be undertaken once a correlation is madebetween the nature of the niosome surface and theresulting biological response. The biological re-sponse to polyoxyethylene (Blume and Cevc,1990) coated liposomes, i.e. their reduced liverand spleen uptake has been exploited for thetargeting of niosomes to tumours for example(Uchegbu et al., 1995, 1996a). Advantageouslyniosomes may be constructed from a variety ofhydrophilic head groups (Fig. 4) and it is likelythat a more specific correlation between niosomesurface chemistry and niosome pharmacodynam-ics:pharmacokinetics may eventually emerge. Thedesign of a hypothetical niosomal system basedon a recognised biological objective would thenlead either to tailored chemical synthesis or (inindustrial settings where a more conservative ap-proach prevails) the selection of tools from anexisting database of approved pharmaceutical ex-cipients (Fig. 3). Design of the drug deliverysystem would then be followed by stability andbiological testing. Invariably yet more questionswill arise from this process, but a systematic and

rational evaluation scheme such as that outlinedin Fig. 3 will ultimately lead to a richer under-standing of the capabilities of particulate drugdelivery with non-ionic surfactants.

2. Factors governing the self assembly ofnon-ionic surfactants into niosomes

2.1. Non-ionic surfactant structure

Theoretically niosome formation requires thepresence of a particular class of amphiphile andaqueous solvent. In certain cases cholesterol isrequired in the formulation and vesicle aggrega-tion for example may be prevented by the inclu-sion of molecules that stabilise the system againstthe formation of aggregates by repulsive steric orelectrostatic effects. An example of steric stabilisa-tion is the inclusion of Solulan C24 (a cholesterylpoly-24-oxyethylene ether) in doxorubicin (DOX)sorbitan monostearate (Span 60) niosome formu-lations (Uchegbu et al., 1995). An example ofelectrostatic stabilisation is the inclusion of dicetylphosphate in 5(6)-carboxyfluorescein (CF) loadedSpan 60 based niosomes (Yoshioka et al., 1994).

Previous accounts have listed the types of non-ionic surfactants that are known to form vesicles(Ozer et al., 1991; Florence, 1993b; Uchegbu,Florence, 1995). Such amphiphiles by definition


Fig. 4.


Fig. 4


Fig. 4. (Continued)


must possess a hydrophilic head group (Fig. 4)and a hydrophobic tail. The hydrophobic moi-ety may consist of one or two alkyl or pe-rfluoroalkyl (Fig. 5) groups or in certain casesa single steroidal (Fig. 6) group. The alkylgroup chain length is usually from C12C18(Handjani-Vila, 1990; Ozer et al., 1991; Flo-rence, 1993a; Bouwstra, Hofland, 1994; Yosh-ioka et al., 1994; Uchegbu, Florence, 1995).Molecules may possess one (Azmin et al., 1985;Baillie et al., 1985; Kiwada et al., 1985a,b;Rogerson et al., 1987, 1988; Stafford et al.,1988; Moser et al., 1989, 1990; Wallach, Philip-pot, 1993; Talsma et al., 1994; Yoshioka et al.,1994) two (Okahata et al., 1981; Azmin et al.,1985; Baillie et al., 1985; Rogerson et al., 1988;Chauhan, Lawrence, 1989; Schenk et al., 1990;Assadullahi et al., 1991; Polidori et al., 1994)or three (Tanaka, 1990) alkyl chains. Pe-rfluoroalkyl surfactants that form vesicles pos-sess chain lengths as short as C10 (Zarif et al.,1993; Guedj et al., 1994). Additionally crownether amphiphiles bearing a steroidal (Fig. 6)(Echegoyen et al., 1988), C14 alkyl (Montserratet al., 1980) or C16 alkyl (Darwish, Uchegbu,1997) hydrophobic unit have been shown toform vesicles.

While the number of hydrophobic permuta-tions is at present limited, there have been awide variety of hydrophilic head groups in vesi-cle forming surfactants (Fig. 4) and it is in thisarea of vesicle forming surfactant design that

considerable scope for new formulations stillexist. The two portions of the molecule may belinked via ether, amide or ester bonds (Fig. 4).

We have observed that a parameter like thehydrophilic lipophilic balance (HLB) is a goodindicator of the vesicle forming ability of anysurfactant. With the sorbitan monostearate(Span) surfactants, a HLB number of between4 and 8 was found to be compatible with vesi-cle formation (Yoshioka et al., 1994; Uchegbu,Florence, 1995). The guidance offered by theHLB number is useful as apart from the theo-retical methods of estimating HLB number inwhich the relative proportions of both the hy-drophilic and hydrophobic portions of themolecule are assessed, practical methods ofHLB number determination have been reported(Trapani et al., 1995). These studies may beuseful in the evaluation of new classes of com-pounds for their vesicle forming ability. Thewater soluble detergent polysorbate 20 (Fig. 7)also forms niosomes in the presence of choles-terol (Saettone et al., 1996; Santucci et al.,1996; Carafa et al., 1998). This is tucci et al.,1996; Carafa et al., 1998). This is despite thefact that the HLB number of this compound is16.7 and it appears on first inspection to betoo hydrophilic to form a bilayer membrane.However with an optimum level of cholesterol,it seems that niosomes are indeed formed frompolysorbate 20 (Santucci et al., 1996)

Fig. 4. Hydrophilic head groups found in vesicle forming surfactants: (a) glycerol head groups (Vanlerberghe et al., 1972;Handjani-Vila et al., 1979; Ribier et al., 1984; Azmin et al., 1985; Baillie et al., 1985, 1986; Rogerson et al., 1987; Kerr et al., 1988;Rogerson et al., 1988; Stafford et al., 1988; Cable, 1989; Florence, Baillie, 1989; Moser et al., 1989; Lesieur et al., 1990; Moser etal., 1990; Seras et al., 1992; Uchegbu et al., 1992; Seras et al., 1994; Uchegbu et al., 1994; Arunothayanun et al., 1996; Bernard etal., 1996; Seras et al., 1996; Uchegbu et al., 1996a,b; Dimitrijevic et al., 1997; Uchegbu, Duncan, 1997) (b) ethylene oxide headgroups (Okahata et al., 1981; Chauhan, Lawrence, 1989; Tanaka, 1990; Hofland et al., 1992; Yoshida et al., 1992; Wallach,Philippot, 1993; Hofland et al., 1994; Niemec et al., 1994; Schreier, Bouwstra, 1994; Talsma et al., 1994; Lawrence et al., 1996;Vanhal et al., 1996; Gianasi et al., 1997) (c) crown ether head groups (Montserrat et al., 1980; Echegoyen et al., 1988; Darwish,Uchegbu, 1997) (d) polyhydroxy head groups (Assadullahi et al., 1991) (e) sugar head groupsamino acids 1,2,4&5galactosederivatives, 3glucose derivative (Zarif et al., 1993; Guedj et al., 1994; Polidori et al., 1994) (f) sugar head groups 1galactose,2mannose, 3glucose, 4 lactose, 5 and 6sorbitan esters, 7maltose, 8sucrose (Kiwada et al., 1985a,b; Schenk et al.,1990; Chandraprakash et al., 1993; Naresh et al., 1993; Reddy, Udupa, 1993; Udupa et al., 1993; Parthasarathi et al., 1994; Polidoriet al., 1994; Yoshioka, Florence, 1994; Yoshioka et al., 1994; Jain and Vyas, 1995a,b; Uchegbu et al., 1995; Yoshioka et al., 1995;Murdan et al., 1996; Naresh, Udupa, 1996; Duncan et al., 1997; Gianasi et al., 1997; Uchegbu, Duncan, 1997).


Fig. 5. Vesicle forming fluorinated surfactant (Zarif et al.,1993).

Fig. 7. Polysorbate 20.

have in turn been used to provide computersimulations of the three dimensional form of theproposed surfactant vesicles. Whether such nio-some modelling software will be used as a pre-dictive tool in the future is unknown at thepresent time. However a software generated fea-ture that will be particularly useful for vesiculardrug delivery is a database incorporating all therelevant experimental data on surfactants whichwill in turn predict the vesicle forming be-haviour of hypothetical compounds or com-pound mixtures.

In time it is envisaged that such an ExpertSystem will emerge that will hold details notonly on the optimum conditions required to in-duce vesicle formation on an amphiphile by am-

Although a particular membrane surfactantmay be chosen by combining the hydrophilicmoieties given in Fig. 4 with an appropriate hy-drophobic group. Established molecules mayalso be chosen from those mentioned in a fewearlier reviews (Ozer et al., 1991; Florence,1993b; Bouwstra, Hofland, 1994; Uchegbu, Flo-rence, 1995). Some of these surfactants such asthe Span and Brij surfactants are already estab-lished pharmaceutical excipients.

Recently neutron reflectivity data obtainedfrom glycerol and polyoxyethylene surfactantmonolayers has been used to calculate the areaper molecule and degree of monolayer hydration(Barlow et al., 1995). These calculated values

Fig. 6. Vesicle forming crown ether surfactants 1cholestanyl derivative, 2cholesteryl derivative (Echegoyen et al., 1988).


phiphile basis but also on predicted vesicle size,drug-loading capabilities and expected pharma-cology. Ideally such a system should be inter-faced with regulatory, intellectual property,chemical assay and compound physical stabilitydatabases.

2.2. Membrane additi6es

Unfortunately the prediction of vesicle form-ing ability is not a simply a matter of HLBnumbers and chemical structure and variousother factors come into play. It is generally ac-cepted that the parameters for self-assemblylaid own by Israelachvili (Israelachvili, 1985) inwhich a critical packing parameter (CPP) wasdefined, largely hold true today.

CPP6:lca0

where 6hydrophobic group volume, lc thecritical hydrophobic group length and a0 thearea of the hydrophilic head group (Fig. 8). ACPP of between 0.5 and 1 indicates that thesurfactant is likely to form vesicles (Is-raelachvili, 1985). A CPP of below 0.5 (indicat-ing a large contribution from the hydrophilichead group area) is said to give spherical mi-celles and a CPP of above 1 (indicating a largecontribution from the hydrophobic group vol-ume) should produce inverted micelles, the lat-ter presumably only in an oil phase, orprecipitation would occur.

Often various additives must be included inthe formulation in order to prepare stable nio-somes. Fig. 9 illustrates the number of differentmorphologies, all with different permeabilityand stability properties, which may be formedby the manipulation of the membrane formingagents in a typical system (Uchegbu et al.,1996b). The most common additive found inniosomal systems is cholesterol. Thus in caseswhere a mixture of surfactants or cholesterol isused to prepare niosomes, the operational CPPvalues will be those of the entire components.

The bilayer membrane is an ordered structureand may exist in the gel state (Lb or Lb%thelatter indicating a situation where the alkylchains are tilted at a slight angle with respectto the plane of the bilayer) or the liquid crys-talline statesometimes called the lamellarphase (La). The difference between these twophases is the degree of order, with the gel statebeing the most ordered structure and the liquidcrystal state being less ordered. In the liquidcrystal state there is lateral diffusion of bilayermaterial whereas in the gel state the alkylchains are crystallised or otherwise less mobile.

For any system the liquid crystalline state ex-ists at a higher temperature than the gel state.The increase in temperature (T) although yield-ing an increase in the enthalpy term (DH) alsoresults in an increase in entropy (DS) and thusa lowering of the free energy (DG) of the sys-tem and it is the application of heat that is thedriving force for this transition. The phasetransition for hexadecyl diglycerol ether(C16G2): Solulan C24 niosomes (91:9) (Uchegbuet al., 1997) and 1,2-dialkyl glycerol poly-oxyethylene ether surfactants (Lawrence et al.,1996) has been recorded using differential scan-ning calorimetry and on C16G2 and hexadecylpoly-5-oxyethylene ether (C16EO5) using phasefluorimetry (Ribier et al., 1984). With the lattertechnique the mobility of a membrane fluores-cent probe is monitored as a function of tem-perature (Ribier et al., 1984). This change fromthe gel to liquid phase (analogous to melting)has been adequately documented for liposomalsystems (New, 1990).

Fig. 8. Schematic representation of an amphiphile, a0hy-drophilic head group area, 6hydrophobic chain volume,lchydrophobic chain length.


Fig. 9. The hexadecyl diglycerol (C16G2)-cholesterol-Solulan C24 ternary phase diagram. Total surfactant:lipid content60 mM.Region 1polyhedral niosomes (5 mm), region 2spherical, helical and tubular niosomes (210 mm), region 3discomes (1260mm), large niosomes (1240 mm), spherical, helical and tubular niosomes (210 mm), region 4discomes and possibly Mixedmicelles, region 5cholesterol crystals, region 6spherical niosomes (210 mm) and region 7a clear liquid probably consistingof mixed micelles, region 8mixed micelles formed on elevation of temperature (Uchegbu et al., 1996a).

Cholesterol is known to abolish the gel to liquidphase transition of liposomal (New, 1990) andniosome systems (Fig. 10) (Cable, 1989) resultingin niosomes that are less leaky (Rogerson et al.,1987). Span 60 niosomes prepared without choles-terol formed a gel and only on the addition ofcholesterol was a homogenous niosome dispersionobtained (Yoshioka et al., 1994). Cholesterol isthus usually included in a 1:1 molar ratio in mostformulations. However even after the addition ofcholesterol, the intrinsic phase transition be-haviour of vesicle forming surfactants still influ-ences the properties of the dispersion: notably themembrane permeability, encapsulation efficiency,bilayer rigidity, ease of rehydration of freeze driedniosomes, toxicity, etc. all of which are discussedbelow.

Often niosomes must be stabilised by the addi-tion of a charged molecule to the bilayer such as

dicetyl phosphate as mentioned above. Dicetylphosphate prevents the aggregation of C16G2 nio-somes (Cable, 1989) and is also added to C16G2niosomes encapsulating haemoglobin in order toachieve an electrophoretic mobility similar to thatof erythrocytes (Moser et al., 1989).

2.3. Nature of the encapsulated drug

Another factor often overlooked is the influenceof an amphiphilic drug on vesicle formation.While sorbitan monostearate (Span 60) niosomescontaining dicetyl phosphate formed homogenousdispersions when encapsulating CF (Yoshioka etal., 1994) this system formed an aggregated dis-persion when encapsulation of the amphipathicdrug DOX was attempted (Uchegbu, 1994). Asteric stabiliser Solulan C24 (poly-24-oxyethylene


Fig. 10. Differential scanning calorimetry (DSC) scans ofniosomes containing the mol% C16G2 shown and cholesterol(Cable, 1989).

of hexadecyl diglycerol ether (C16G2) niosomes ina pH dependent manner (Cable, 1989), an indica-tion that the amphipathic drug is incorporated inthe vesicle membrane. From the foregoing, it isclear that Israelachvilis CPP of a potential nio-some system must take into account the presenceof amphipathic or hydrophobic drugs as boththese substances will be incorporated into thevesicle membrane.

2.4. Surfactant and lipid le6els

The level of surfactant:lipid used to make nio-somal dispersions is generally 1030 mM (12.5%w:w) (Okahata et al., 1981; Baillie et al., 1985;Lesieur et al., 1990; Seras et al., 1992; Uchegbu etal., 1992; Zarif et al., 1993; Lawrence et al., 1996;Saettone et al., 1996; Santucci et al., 1996; Seraset al., 1996; Uchegbu et al., 1996a). Altering thesurfactant:water ratio during the hydration stepmay affect the systems microstructure (Tanaka,1990) and hence the systems properties. Howeverincreasing the surfactant:lipid level also increasesthe total amount of drug encapsulated, as dis-cussed below, although highly viscous systemsresult, if the level of surfactant:lipid is too high.

2.5. Temperature of hydration

The hydrating temperatures used to make nio-somes should usually be above the gel to liquidphase transition temperature of the system.

3. Niosome preparation

The formation of vesicular assemblies requiresthe input of some form of energy (Lasic, 1990)and all the experimental methods surveyed consistof the hydration of a mixture of the surfactant:lipid at elevated temperature followed by optionalsize reduction to obtain a colloidal dispersion.This is followed by the separation of the unen-trapped drug from the entrapped drug by eithercentrifugation, gel filtration or dialysis. Only onemethod (Novasome) could be found in the litera-ture on the preparation of niosomes on an indus-trial scale (Wallach, Philippot, 1993). This

cholesteryl ether) must be added to the formula-tion to ensure a homogenous formulation devoidof aggregates (Uchegbu, 1994). DOX (Fig. 11) hasbeen shown to alter the electrophoretic mobility

Fig. 11. Doxorubicin.


involves the injection of the melted surfactants:lipids into a large volume of well-agitated heatedaqueous solutions. Although a method involvingthe addition of an aqueous solution to a solidmixture of lipids and surfactants (Handjani-Vilaet al., 1979) is said to be suitable for the handlingof large quantitieskilograms of dispersions.

3.1. Hydration techniques

The more commonly used laboratory methodsof niosome preparation and drug loading iden-tified in the literature are listed below.

1. The injection of an organic solution of sur-factants:lipids in an aqueous solution of the drugto be encapsulated which is heated above theboiling point of the organic solvent (ether injec-tion) (Baillie et al., 1985).

2. The formation of a surfactant:lipid film bythe evaporation of an organic solution of surfac-tants:lipids. This film is then hydrated with asolution of the drug (hand shaking) (Azmin et al.,1985; Baillie et al., 1985). This method was previ-ously described by Bangham and others (Bang-ham et al., 1965) for the preparation of liposomes.

3. The formation of an oil in water (o:w) emul-sion from an organic solution of surfactants:lipidsand an aqueous solution of the drug. The organicsolvent is then evaporated to leave niosomes dis-persed in the aqueous phase. In some cases, a gelresults which must be further hydrated to yieldniosomes. (reverse phase evaporation) (Kiwada etal., 1985a), previously described by Szoka andPapahadjopoulos (Szoka and Paphadjopoulos,1978) for the preparation of liposomes.

4. The injection of melted lipids:surfactantsinto a highly agitated heated aqueous phase inwhich presumably the drug is dissolved (Wallach,Philippot, 1993) or the addition of a warmedaqueous phase dissolving the drug to a mixture ofmelted lipids and hydrophobic drug (Niemec etal., 1994).

5. The addition of the warmed aqueous phaseto a mixture of the solid lipids:surfactants (Hand-jani-Vila et al., 1979).

Methods 4 and 5 do not require the use oforganic solvents, which are expensive, difficult toremove in their entirety and hazardous.

6. Niosomes may also be formed from a mixedmicellar solution by the use of enzymes(Chopineau et al., 1994). A mixed micellar solu-tion of C16G2, DCP, polyoxyethylene cholesterylsebacetate diester (PCSD) converts to a niosomedispersion when incubated with esterases. PCSDis cleaved by the esterases to yield poly-oxyethylene, sebacic acid and cholesterol. Choles-terol in combination with C16G2 and DCP thenyields C16G2 niosomes.

7. The homogenisation of a surfactant:lipidmixture followed by the bubbling of nitrogen gasthrough this mixture (Talsma et al., 1994). Appar-ently the homogenisation step may be omittedfrom the procedure with out affecting particlesize, although a longer bubbling time wasrequired.

3.2. The reduction of niosome size

Niosomes prepared as described above are usu-ally in the micron size range (Handjani-Vila et al.,1979; Azmin et al., 1985) although some of themethods (Baillie et al., 1985; Wallach, Philippot,1993; Talsma et al., 1994) produce niosomes inthe sub-micron (:300 nm) size range.

Often a size reduction step must be incorpo-rated into the niosome production procedure,subsequent to the initial hydration step as vesiclesize has an important bearing on vesicle biodistri-bution. For example sub-200 nm phospholipidvesicles have been shown to avoid splenic but notliver uptake (Litzinger et al., 1994). A reduction invesicle size may be achieved by a number ofmethods.1. Probe sonication (Azmin et al., 1985; Baillie et

al., 1985) which yields C16G3 niosomes in the100140 nm size range.

2. Extrusion through 100 nm Nucleopore filters(Stafford et al., 1988) which yields sodiumstibogluconate C16G3 niosomes in the 140 nmsize range.

3. In some instances the combination of sonica-tion and filtration (220 nm Millipore filter)has been used to achieve DOX loaded Span 60niosomes in the 200 nm size range (Uchegbu etal., 1995).


Fig. 12. The effect of microfluidization at a pressure of 0.55mpa and a flow rate of 48 ml min1 on the size of C16G3,cholesterol, dicetyl phosphate (47.5:47.5:5) niosomes. One cy-cle35 s.

Fig. 14. The effect of total amount of surfactant:lipid usedwith the fixed level of DOX (6.9 mmoles) on the size andencapsulation efficiency of doxorubicin C16G2 niosomes. C16G2, cholesterol, Solulan C24 (45:45:10)-niosome size, C16G2, cholesterol, Solulan C24 (45:45:10)-niosome encapsula-tion efficiency, C16G2, cholesterol, Solulan C24, dicetylphosphate (42.75:42.75:9.5:2)-niosome size, C16G2,cholesterol, Solulan C24, dicetyl phosphate (42.75:42.75:9.5:2)-niosome encapsulation efficiency. Figures in parenthesis repre-sent % encapsulation values.4. The achievement of sub-50 nm sizes is possible

by the use of a microfluidizer (Fig. 12).5. High-pressure homogenisation also yields vesi-

cles of below 100 nm in diameter althoughdrug loading is ultimately sacrificed to achievethis small size (Fig. 13).

3.3. Drug loading optimisation

3.3.1. Units for the reporting of drug loadAs expected drug loading is a crucial factor in

the formulation of niosome delivery systems.However before a discussion on drug loading canbegin, it is important to emphasise that due careand attention must be paid to the units used toquote drug-loading values. For example, drug-loading values are often quoted as the % drugencapsulated. However for these values to haveany meaning the initial drug, surfactant:lipid ratiomust be stated. A simple study in which theamount of DOX encapsulated was measured as afunction of the initial level of surfactant:lipid,showed that this initial surfactant:lipid ratio de-termines ultimately the % encapsulation (Fig. 14).It was found that although the % encapsulationvalues steadily increased the final ratio of drug tosurfactant:lipid decreased steadily. In a similarstudy a surfactant:lipid concentration rangingfrom 501000 mM showed no change in the finalmolar ratio of CF to surfactants:lipids althoughthe % encapsulation increased steadily (Yosh-ioka et al., 1994).

Fig. 13. The effect of homogenisation of doxorubicin multil-amellar niosomes at 60 mpa pressure on niosome size andencapsulation efficiency, C16G2, cholesterol, Solulan C24(45:45:10)-niosome size, C16G2, cholesterol, Solulan C24(45:45:10)-niosome encapsulation efficiency, C16G2,cholesterol, Solulan C24 (30:30:40)-niosome size.


Another value often quoted in the literature isthe litres of drug solution encapsulated per moleof surfactant:lipid (l mol1). This value assumesno change in the concentration of the encapsulat-ing solvent from the hydration step through to theseparation and analysis steps, yet evidenceabounds to show that niosome membranes arepermeable to low molecular weight compoundsand do release encapsulated solutes with time. Itis conceivable that this release will be increasedonce a sufficient concentration gradient isachieved across the vesicle membrane, providingthe vesicle membrane is not absolutely imperme-able. It is obvious that a sufficient concentrationgradient will be achieved during the separationprocedure and thus drug solute leakage will com-mence before the start of analysis for drugencapsulation.

Clearly encapsulation efficiency once given in %encapsulation must be qualified with details onthe initial ratio of drug to surfactant lipid. In ouropinion the most useful value to any formulatorwill be the ratio of drug to surfactant in the finalformulation in (g g1) or (mol mol1). This givesadequate information on the level of excipientthat must be administered at each dose level.

3.3.2. Effect of bilayer constituentsThe chemical nature of the niosome membrane

may be manipulated to increase drug loading, byaltering the nature of the hydrophilic head groupand:or the hydrophobic moiety. When a series ofC16-sugars were examined the encapsulation effi-ciency for the aqueous solute [14C]sucrose fol-lowed the trend glucose\mannose\galactose\lactose (Kiwada et al., 1985b). These differencesmay be attributed to the different levels of hydra-tion of these sugars, a parameter that would affecttheir vesicle forming abilities. The encapsulationof DOX N (2-hydroxypropyl) methacrylamide co-polymer (PK1) in C16G2 niosomes was twice ashigh as in hexadecyl poly-5-oxyethylene (C16EO5)niosomes under identical conditions of prepara-tion (Uchegbu, Duncan, 1997). The decreasedmembrane fluidity identified in C16G2 bilayermembranes when compared with C16EO5 bilayermembranes (Ribier et al., 1984) is thought to beresponsible for this increased entrapment

(Uchegbu, Duncan, 1997).The nature of the hydrophobic alkyl chain af-

fects the encapsulation efficiency of CF (Yoshiokaet al., 1994) and DOX (Uchegbu, Florence, 1995)by unsonicated Span surfactant niosomes. Span60 (C18) and Span 40 (C16) gave the greatestencapsulation efficiency for CF niosomes andwere the least leaky niosomes due to the fact thatthese Span surfactants had the highest phase tran-sition temperature (Yoshioka et al., 1994). Alsounsonicated Span 60 niosomes gave the greatestencapsulation efficiency for DOX (Uchegbu, Flo-rence, 1995). With sonicated DOX Span surfac-tant niosomes the encapsulation efficiencyfollowed the trend C18\C16\C12. (Uchegbu,Florence, 1995). In the same study sorbitanmonoleate was found to have the lowest encapsu-lation efficiency. A leaky membrane due to unsat-uration in the oleic side chain (Yoshioka et al.,1994) was thought to be the cause of this. Theencapsulation efficiency of CnEO5 niosomes forPK1 was higher for C18EO5 than C16EO5 nio-somes and the former were also more rigid andless deformable (Uchegbu, Duncan, 1997). Theseeffects are attributed to the higher membranephase transition temperature (Bouwstra, Hofland,1994) for C18 systems when compared with C16systems. The ease of rehydration of freeze-driedniosomes also decreased as expected with increasein surfactant hydrophobicity.

From the above, it appears that the less fluid thebilayer (higher the gel to liquid phase transitiontemperature) the higher the encapsulation effi-ciency. This is despite the fact that cholesterol,which presumably abolishes the phase transitionendotherm, is included in a 1:1 molar ratio in allthe above named formulations. The intrinsic mem-brane gel to liquid phase transition temperature hasa fundamental influence on the encapsulation effi-ciency and as discussed laterthe toxicity of thesesystems even when cholesterol is included in thebilayer. However other factors must also influencethe encapsulation of solutes as the examination ofa series of glucosides hydrated with solutions of[14C]sucrose showed that, C16-glucoside niosomeshad a higher encapsulation efficiency than both C14and C18 compounds (Kiwada et al., 1985a) forreasons that are not entirely apparent.


The inclusion of dicetyl phosphate in the for-mation of C16-glucoside niosomes was also foundto increase the encapsulation efficiency of[14C]sucrose (Kiwada et al., 1985b).

3.3.3. Methods of drug load enhancementVarious techniques may be used to optimise

drug load and this is especially important in in-dustrial settings where there is limited scope forthe chemical modification of excipients due toregulatory concerns. One such method is the de-hydrationrehydration vesicle (DRV) techniquefirst described by Kirby and Gregoriadis (Kirbyand Gregoriadis, 1984) which was found to in-crease the encapsulation efficiency of PK1 inC16G2 niosomes from 3.3 to 64.4% (Uchegbu,Duncan, 1997). Unfortunately niosome size wasalso doubled, increasing from 151 to 380 nm. Afinal PK1 to surfactant ratio of 0.3 was achievedwith these DRV formulations. It was noted thatthe ease of rehydration of these freeze-dried dis-persions was directly proportional to the phasetransition temperature of the non-ionic surfactant.Other methods used to maximise drug loadinginclude the use of pH gradients (Mayer et al.,1986). In this method a pH differential existsacross the niosome membrane with a lower pHinside the niosome. The amine drug is then addedexternal to the niosome and crosses the membranebarrier in the unionised state. Once inside theniosome the drug becomes protonated and is un-able to leave the niosome. The acid pH within theniosome interior thus acts as an intra-vesiculartrap. This method has been employed in the for-mulation of vincristine sulphate niosomes(Parthasarathi et al., 1994) using citrate buffer(pH 4.0) followed by the addition of vincristinesulphate and the upward adjustment of the pH to7.1. Once the pH has been adjusted upwards, theformulation is heated above the phase transitiontemperature (60C) of the membrane in order toincrease vesicle permeability. A similar methodutilising Tris buffered saline (TBS) has been re-ported for the loading of DOX C16G2 niosomes(Uchegbu et al., 1994).

An alternative intra-vesicular trap has also beendeveloped for DOX vesicles using ammonium sul-phate (Haran et al., 1993). The presence of am-

monium sulphate within the vesicles apparentlycauses DOX to form a gel within the vesicles.Using this method for DOX Span 60 niosomes,the entrapped ratio of drug to surfactants:lipidsincreased from 0.014 to 0.035 mol mol1

(Uchegbu et al., 1996b) and the niosomes werealso less leaky at 37C (Fig. 15).

While these remote loading procedures havebeen used for amine drugs, similar strategies havenot been used to increase the niosome encapsula-tion of acidic drugs.

3.4. Separation of entrapped material

The hydration of surfactant:lipid mixturesrarely leads to the entire drug being encapsulated,regardless of the drug loading optimisation stepstaken. It is thus often a requirement that unen-capsulated drug be removed by various means.Although it may be argued that the use of systemsin which half of the drug is encapsulated and halfis external to the niosome may eventually yieldsystems with a beneficial biphasic biodistributionprofile. This drug delivery system would give aninitial burst to initiate therapy followed by a

Fig. 15. The release of doxorubicin from niosomes preparedutilising a transmembrane proton gradient, niosomes wereprepared utilising: transmembrane proton gradient (200mg ml1 doxorubicin), transmembrane proton gradients(194 mg ml1 doxorubicin), transmembrane proton gra-dients (134 mg ml1 doxorubicin), transmembrane am-monium sulphate gradients (195 mg ml1 doxorubicin)(Uchegbu et al., 1996b).


Table 1The advantages and disadvantages of the different methods of separation of the entrapped from the unentrapped drug

AdvantagesSeparation Disadvantagesmethod

Exhaustive dialy- Suitable for large Extremely slow (524 h)vesicles\10 mmsis

Large volumes of dialysate required-(may not be suitable for drugs requiring spe-Suitable for highlyviscous systems cialised disposal)

Dilutes the niosome dispersionInexpensive

Centrifugation Fails to sediment the sub-micron niosomesQuick (30 min)(below 7000g)

Inexpensive instru- May lead to the destruction of fragile systemsmentationConcentrates theniosome dispersion

Expensive InstrumentationUltracentrifuga- Sediments all sizetion populations(150 000g)

Concentrates the Long centrifugation times (11.5 h)niosome dispersion

May lead to the destruction of fragile systemsMay lead to the formation of aggregates

Quick (45 minGel filtration Slow (12h when using Sepahrose 2B:4B for macromolecule separation)with Sephadex G50)

Gels are expensive if not reusedDilutes the niosome dispersionNot suitable for highly viscous formulationsNot suitable for formulations with a large particle size (\1020 mm)

sustained maintenance dose. This is demonstratedby the improved activity against Leishmania dono-6ani seen with alkyl polyglycerol or alkyl poly-oxyethylene based sodium stibogluconateniosomes when unentrapped drug was not re-moved when compared with niosomes in whichthe unentrapped drug had been removed(Williams et al., 1995). These former formulationswere also superior to the use of the free drug.

The methods that have been used for the re-moval of unentrapped material include:1. Exhaustive dialysis (Baillie et al., 1985, 1986).2. Separation by gel filtration (Sephadex G50)

(Uchegbu et al., 1994; Yoshioka et al., 1994).3. Centrifugation (7000g for 30 min) for DOX

C16G3 niosomes prepared by hand-shakingand ether injection methods (Rogerson et al.,1987),

4. Ultracentrifugation (150000g for 1.5 h) forPK1 niosomes (Duncan et al., 1997; Gianasi etal., 1997; Uchegbu, Duncan, 1997).

All these methods have their advantages anddisadvantages as given in Table 1. The choice ofmethod must take all these factors into accountand for industrial purposes it may be more worth-while to concentrate efforts and resources on theachievement of high levels of drug loading so asto avoid these separation steps altogether or toconsider systems in which the unentrapped drugserves as a specific priming dose as describedabove.

4. Osmotic activity

The osmotic activity of niosomal dispersions isestimated by a change in niosome size when dis-


persed in a hypertonic or hypotonic medium. Insome systems osmotic activity may not be ob-served because there is initial aggregation of thevesicles as the ions in the disperse phase shield theelectrostatic charge on the vesicle surface (Cable,1989). However the incorporation of steric sta-bilisers in the vesicle membrane such as poly-oxyethylene compounds prevents this aggregationand enables a study of the osmotic activity ofthese systems.

Niosomes prepared from C16G2 which containthe polyoxyethylene compound Solulan C24(Arunothayanun et al., 1996) and both C16G3 andC16C12G7 (Fig. 16) (Baillie et al., 1985) are osmot-ically active when cholesterol is incorporated intothe bilayer in a 1:1 molar ratio. Cholesterol freeniosomes were found to show less change in size(Baillie et al., 1985; Arunothayanun et al., 1996)due to the fact that these niosomes are morepermeable to solutes (Rogerson et al., 1987;Arunothayanun et al., 1996). Niosomes preparedfrom dialkyl glycerol polyoxyethylene ethers werereported not to be osmotically active at roomtemperature (Lawrence et al., 1996) although

some of the higher phase transition surfactantniosomes were osmotically active at elevated tem-perature (50C). The authors conclude that thislack of observed osmotic activity is due to thehydration of the polyoxyethylene head groups onthe membrane surface. Membrane permeabilitystudies were however not reported.

5. Toxicity studies

Unfortunately not too many niosome toxicitystudies abound in the literature, despite the factthat often the lack of a toxicological profile isfundamental to any regulatory objections. Morestudies examining the toxicology of these systemscan do nothing but advance the science and pre-sumably provide opportunities for commercial ex-ploitation. However the authors acknowledge thatdata demonstrating the drug delivery advantageof these systems over more established systemsmust first be produced to encourage the diversionof resources to toxicology studies.

In vitro studies on a ciliotoxicity model toestimate the toxicity of alkyl polyoxyethylene(CnEOx) niosomes on the nasal mucosa revealedthat an increase in alkyl chain length was accom-panied by a decrease in toxicity while an increasein the polyoxyethylene chain length caused anincrease in ciliotoxicity (Hofland et al., 1992). Ingeneral an increase in chain length increases thegel to liquid transition while an increase in thelength of the polyoxyethylene chain decreases thegel to liquid phase transition (Hofland et al.,1992). This study concluded that gel state nio-somes are less ciliotoxic than the liquid statevesicles. Although it is worthwhile to note that theauthors state that the HLB number had no influ-ence on the toxicity of the compounds, the morehydrophobic compounds were obviously lesstoxic.

Neither the length of the polyoxyethylene chainor the alkyl chain had any influence on the skintoxicity of alkyl polyoxyethylene niosomes as as-sessed by the cell proliferation of human kerati-nocytes in vitro (Hofland et al., 1991, 1992). Thelatter is considered to be a measure of the irrita-bility. However the nature of the linkage (ether or

Fig. 16. The osmotic shrinkage of alkylglycerol ether nio-somes, hexadecyl triglycerol ether (C16G3) (100), C16G3, cholesterol (50:50), hexadecyldodecylheptaglycerol ether (C16C12G7), cholesterol (50:50) in the pres-ence of NaI solution (Baillie et al., 1985).


ester) was a determining factor in this model andthe more labile ester bond was found to be moretoxic than the ether bond.

The parenteral administration of niosomes usu-ally proceeds via the intravenous route and assurfactants are used in the formulation, it is im-portant to conduct haemocompatibility studies.Less than 5% haemolysis was detected in vitroafter 5 h of incubation of C16G2 and Span 60niosomes (containing 10 mol% Solulan C24asoluble surfactant) with rat erythrocytes at a simi-lar surfactant level as would be achieved immedi-ately after the administration of 10 mg kg1

DOX in the form of PK1 niosomes (Uchegbu,Duncan, 1997). This level of haemolysis is notconsidered significant as less than 2% of the in-jected dose of these particular C16G2 niosomes isstill present in the plasma 5 h after dosing(Uchegbu, Duncan, 1997) and less than 2% of theinjected dose of Span 60 niosomes is present inthe plasma 4 h after dosing (Uchegbu et al.,1995). By all indications Span 60 niosomes con-taining 10 mol% Solulan C24 are not haemotoxic.This soluble surfactantSolulan C24 was alsofound to be toxic to Caco-2 cells in vitro (Dimitri-jevic et al., 1997). However when incorporatedinto niosomes (10 mol%), this toxicity was drasti-cally reduced. There was an increase in the toxic-ity when the level of Solulan C24 in niosomes wasincreased above 10 mol% (Dimitrijevic et al.,1997) due to the fact that above 10 mol%, thesoluble surfactant Solulan C24 is not incorporatedinto the membrane of C16G2 niosomes and is thuspresent in solution as monomers or micelles.

The intraperitoneal injection of DOX C16G2niosomes led to an inflammatory response in thelung and ultimate fatalities within 24 h (Uchegbuet al., 1994). This response was dose related andnot found on the administration of these nio-somes by the intravenous route. The effect wasalso not observed after the intraperitoneal injec-tion of empty (not drug loaded) C16G2 niosomesor DOX solution (Uchegbu et al., 1994). It ispossible that DOX niosomes are transportedaway from the peritoneum by the lymphatics viathe thoracic duct allowing a higher dose in themain veins emptying into the heart (Uchegbu etal., 1994). However the exact reason for this

response is unclear as although DOX is a knowncardio-toxic agent (Calabresi and Chabner, 1996),heart levels of DOX were similar irrespective ofwhether the drug was administered intraperi-toneally or intravenously (Uchegbu et al., 1994).The intraperitoneal administration of methotrex-ate niosomes also results in a large percentage ofthe dose (56%) being found in the thoracic lymph3 h after dosing compared with 12% of the dosewhen methotrexate was administered in solution(Jain and Vyas, 1995a).

The toxicity of certain non-ionic surfactantsmay be modulated by the incorporation into nio-somes. An example is the modulation of the toxi-city of free Solulan C24 to Caco-2 cell monolayersby the incorporation of this soluble surfactant inniosomes (Dimitrijevic et al., 1997) or the lessweight loss and bone marrow suppression ob-served on the intraperitoneal administration ofsucrose ester niosomes in comparison with thatobserved on the administration of sucrose esteralone (Schenk et al., 1990).

One of the main aims of these specialised drugdelivery systems is to modulate drug toxicity. Thishas been achieved with certain niosomal formula-tions. The encapsulation of vincristine within nio-somes of unspecified composition reduced theneurological toxicity, diarrhoea and alopecia asso-ciated with the intravenous administration of vin-cristine and increased vincristine anti-tumouractivity in S-180 sarcoma and Erlich ascites mousemodels (Parthasarathi et al., 1994). However bonemarrow suppression was similar after the adminis-tration of free or niosomal drug (Parthasarathi etal., 1994).

6. Specialised systems

6.1. Vesicle in water in oil systems

Span surfactant niosomes have been dispersedin an oil in water emulsion to yield a vesicle inwater in oil system v:w:o using the same surfac-tant that was used to make the niosomes (Yosh-ioka, Florence, 1994). The release of CF fromthese systems followed the trend v:w:oBwater inoil (w:o) emulsionsBniosome dispersions. The


difference between the v:w:o and w:o formula-tions was minimal. The release of CF encapsu-lated within these niosomes was influenced by theemulsion oil following the trend, isopropyl myris-tate\octane\hexadecane and by the nature ofthe surfactant, following the trend Span 20\Span 40\Span 60. Span 80 v:w:o systems had arather faster release rate due to the unsaturationin the oleyl alkyl chain, which leads to the forma-tion of a more leaky membrane.

Span 60 was found to cause the formation of agel in the oil phase, which the authors attribute tothe crystallisation of Span 60 within the oil phase.The net result is an extremely slow release ratefrom the Span 60 v:w:o formulation (Yoshioka,Florence, 1994). These gelled Span 60 systemsmay be stabilised by incorporation of polysorbate20 (Murdan et al., 1996) and the resultant Span60 v:w:o organogels (oil phasehexadecane)were found to have a temperature dependant re-lease profile when CF was encapsulated within theniosomes. The release rate was highest at 37Cwhen the gel microstructure showed the presenceof tubulespresumably aqueous water channelsalong which CF is transported and slowest at60C when the gel transforms to a recognisablev:w:o system. At this elevated temperature thewater channels present in the gel state transformto water droplets within which niosomes are con-tained. The slow rate of CF transport at 60C waspresumed to be due to the presence of the oilphase completely surrounding the water dropletthrough which CF must traverse.

An explanation for this gel formation is soughtin the phase transition behaviour of Span 60. Atthe elevated temperature (60C) which exceeds theSpan 60 membrane phase transition temperature(50C) (Yoshioka et al., 1994), it is assumed thatSpan 60 surfactant molecules are self assembled toform a liquid crystal phase. This liquid crystalphase stabilises the water droplets within the oil.However below the phase transition temperaturethe gel phase persists and it is likely that themonolayer stabilising the water droplets collapsesand Span 60 precipitates within the oil. This Span60 precipitate thus immobilises the liquid oil toform a gel. Water channels are subsequentlyformed when the w:o droplets collapse.

This explanation is plausible as the aqueousvolume marker CF was identified within theseelongated water channels and non-sphericalaqueous droplets were formed within the gel(Murdan et al., 1996).

These v:w:o systems have been further evalu-ated as immunological adjuvants as discussedbelow.

6.2. Niosomes in hydroxypropyl methyl cellulose

A transdermal flurbiprofen (1% w:v) formula-tion has been prepared from Flurbiprofen Span60, cholesterol, DCP (46:50:4) niosomes incorpo-rated within a hydroxypropyl methyl cellulosesemi-solid base containing 10% glycerine (Reddy,Udupa, 1993). The in vitro characterisation of theformulation is not given although this formula-tion was evaluated in a rat inflammation model asdiscussed below.

6.3. Discomes

The solubilisation of C16G2 niosomes by Solu-lan C24 results in the formation of the discomephase (Uchegbu et al., 1992). This phase consistsof giant vesicles of 60 mm in diameter whichencapsulate aqueous solutes such as CF. Theselarge vesicles were found to be of two types largevesicles that appear ellipsoid in shape and largevesicles that are truly discoid (Uchegbu et al.,1996a). These morphologies were confirmed byconfocal laser scanning microscopy and are onlyformed in a very specific region of the C16G2ternary phase diagram (Fig. 9) namely regions 3and 4. The discomes found in region 3 coexistwith small spherical, helical and tubular niosomes(210 mm) which are found in a neighbouringregionregion 2. While discomes found in region4 do not co-exist with small spherical, helical andtubular vesicle (Uchegbu et al., 1996b). On heat-ing the discome dispersion identified in region 3,the large discomes are seen to disappear leavingonly the spherical:helical and tubular structures(Uchegbu, Florence, 1995). This is a reversibleprocess and the discomes reform on cooling al-though the encapsulated aqueous solute is lost tothe disperse phase by this heating and cooling


Fig. 17. The release of 5(6)-CF from discomes prepared fromC16G2, cholesterol, Solulan C24 (50:35:15). The temperaturewas altered as shown by the arrows.

6.4. Polyhedral niosomes

Polyhedral niosomes (Fig. 18) (Uchegbu et al.,1997) are formed in low cholesterol regions of theC16G2, cholesterol, Solulan C24 ternary phase dia-gram (Fig. 9). Polyhedral niosomes also encapsu-late aqueous solutes such as CF. The vesiclemembrane is in the gel phase (La) (Uchegbu et al.,1997), meaning that the hydrocarbon chains enjoyminimum mobility. This gives the vesicles theunusual angular shape. On heating these vesiclesabove the phase transition temperature (43C),the angular shape is lost and a spherical morphol-ogy is observed (Fig. 19a, b, c) (Bernard et al.,1996), which on cooling results in an alteredmorphology (Fig. 19d). It appears that the heat-ing and cooling cycle causes irreversible changesto the membrane.

Polyhedral niosomes were found to be ther-moresponsive (Fig. 20a) (Uchegbu et al., 1997).Above 35C, there was an increase in the releaseof CF from these niosomes even though the poly-hedral shape was preserved until these vesicleswere heated to 50C. Since Solulan C24 free poly-hedral niosomes do not exhibit this thermore-sponsive behaviour (Uchegbu et al., 1997) it wasthought to be due to a decrease in the interactionof the polyoxyethylene compoundSolulan C24with water at this temperature (due to decreasedhydrogen bonding) as identified by viscometry(Bernard et al., 1996). This observed thermore-sponsive behaviour was used to design a reversiblethermoresponsive controlled release system (Fig.20b). Thermoresponsive liposomal systems whichrely on the changing membrane permeabilitywhen the system transfers from the gel state (La)to the liquid crystal state (Lb) (Ono et al., 1994)are not reversible. This is not unexpected as thereis a definite alteration of the membrane character-istics on proceeding through a cooling and heat-ing cycle across the phase transition temperature(Fig. 19).

It is proposed that these thermoresponsivepolyhedral dispersions may be used in dermatol-ogy as they are extremely viscous (due to thepolyhedral niosome shape) (Bernard et al., 1996)and also due to the fact that at 30C, (skin surfacetemperature lies between 26 and 30C) they are

cycle (Fig. 17). Once the discomes are destroyedthe release of CF is once again slowed and repre-sents the release from the remaining small spheri-cal, tubular and helical niosomes.

It is proposed that the system described byregion 3 (Fig. 9) may prove useful in ophthalmicdelivery as the initial instillation of the formula-tion in to the eye would result in the slow destruc-tion of the discomes and release of an initial burstdose in accordance with the kinetics shown in Fig.17. The remaining small spherical, helical andtubular vesicle would then release the rest of thedose slowly to the eye. The large size of thediscomes means that clearance from the eye willbe slowed down and the destruction of the dis-comes at 37C results in the release of the encap-sulated contents taking place over several min(Fig. 17) which would in theory allow the dose toenjoy an increased residence time within the eye.

On heating (\35C) the discomes formed inregion 4 (Fig. 9), a clear isotropic solution isobtained thought to consist of mixed micelles.Hydrophobic drugs such as paclitaxel may besolubilised by this system and no precipitation ofthe drug was observed on heating the formulationabove 35C (Uchegbu et al., 1996a). This pacli-taxel formulation could be stored freeze dried.


Fig. 18. a and b (4400) three dimensional reconstruction from a series of confocal laser scanning micrographs of polyhedralniosomes prepared from a mixture of C16G2, Solulan C24 (91:9) (a) viewed from the top; (b) viewed from the side; and (c) (1000)fluorescence micrographs of polyhedral niosomes prepared as described above and stored for 36 days, bar10 mm (Uchegbu et al.,1997).

non-thermoresponsive and thus are capable ofreleasing their encapsulated contents once eitherthe ambient temperature increased to 35C (e.g. inthe use of photoprotective agents) or the skintemperature was raised (e.g. in inflammation).

7. Niosome stability

It would be unwise not to include a separatediscussion of niosome stability in this review al-though it must be borne in mind that all thematerial presented above relate to or have a directinfluence on the stability of niosomal dispersions.

A stable niosome dispersion must exhibit aconstant particle size and a constant level of

entrapped drug. There must be no precipitation ofthe membrane components, which are to a largeextent not insoluble in aqueous media. Ideallythese systems should be stored dry for reconstitu-tion by nursing staff or by the patient and whenrehydrated should exhibit dispersion characteris-tics that are similar to the original dispersion.Freeze dried dispersions of Span 60 niosomesencapsulating the DOX polymer conjugate PK1have been prepared (Gianasi et al., 1997) but onreconstitution only 50% of the drug is encapsu-lated. As mentioned previously, this may not nec-essarily be a draw back for a drug deliverysystem. The solubilisation of a model hydropho-bic drug paclitaxel in discomes remained un-changed after storage at 4C as a freeze dried


Fig. 19. (500) fluorescence micrographs of polyhedral niosomes prepared from C16G2, Solulan C24 (91:9) (a) at 28C; (b) onheating to 50C; (c) on cooling to 40C; and (d) on cooling to 35Cpolyhedral vesicles are seen to reform and also an apparentsplintering of the vesicles is observed (arrows) (Bernard et al., 1996) (bar20 mm).

formulation for 10 days (Uchegbu et al., 1996b).This was assessed by analysing the amount ofpaclitaxel still solubilised after rehydration ofthese formulations.

As insufficient data abounds on the storage offreeze dried niosomal dispersions, the rest of thediscussion will be limited to the stability dataavailable on niosomes in suspension.

The persistence of bilayer self-assembly intoclosed spheres was observed in niosome disper-sions prepared from C16G3, cholesterol in a 1:1molar ratio (5% dicetyl phosphate) (Fig. 21)and C16G2, cholesterol in a 1:1 molar ratio (5%stearylamine) after storage for 84 months at ambi-ent temperature. Niosomes were prepared by hy-drating a thin film using the hand shaking methodand were presumably originally in the 110 mm

size range. The observed difference in the natureof these particles after this length of time was theformation of near perfect spherical shaped nio-somes (Fig. 21). Fig. 22 illustrates DOX niosomesof a similar morphology which had been preparedfrom C16G2, cholesterol, Solulan C24 (40:40:10)by high pressure homogenisation (600 bar10cycles). These particles had been stored for 84days and had an original particle size of 70 nm.The particle size of this dispersion increased ap-proximately 250-fold. Sonicated niosomes pre-pared from these surfactant mixtures, which hadalso been stored for 84 months, revealed thepresence of small needle shaped crystalline mate-rial, indicating a precipitation of the bilayer mate-rial. It appears that the original size of theformulation has an effect on the stability of the


Fig. 20. (a) The release of CF from exhaustively dialysed polyhedral niosomes. The temperature of the surrounding medium wasaltered as shown by the arrows, C16G2, Solulan C24 (91: 9), C16G2, Solulan C24 (95:5) (Uchegbu et al., 1997). (b) Therelease of nicotinamide adenine dinucleotide from polyhedral niosomes prepared from C16G2, Solulan C24 (91:9) at 24C.Arrowspoints at which the temperature of the surrounding medium was raised to 37C for 10 min.

system. This is in keeping with thermodynamictheory as the smaller niosomes require a higherinput of energy and thus contain more excessenergy and an inherently greater instability thanthe larger niosomes prepared by hand shaking.

A further example of how the method of vesicleformation has important bearing on the stabilityof these systems (Engberts, Hoekstra, 1995) is thefact that vesicles prepared by the solvent injectionmethod (ethanol) are found to have an additionalphase transition due to the presence of residualethanol. A number of membrane properties relyon the temperature of the main phase transitionsuch as the membrane permeability and rigidity.The introduction of membrane defects due to thepresence of residual ethanol may destabilise thesedrug delivery systems.

Vesicles prepared from C16G3, Solulan C24(50:50) which had also been stored for 84 monthswere also seen to be morphologically stable (Fig.23). The large size of these vesicles coupled withtheir angular morphology indicates that these areanalogous to the discomes formed in 3 and 4 ofthe C16G2, cholesterol, Solulan C24 ternary phasediagram (Fig. 9). Discomes prepared from C16G2,cholesterol, Solulan C24 (50:15:35) were morpho-logically stable 12 months after preparation

(Uchegbu et al., 1996a). Polyhedral niosomes vi-sualised after storage for 36 days (Fig. 18c) alsoshowed an unaltered morphology (Uchegbu et al.,1997).

It is interesting that the surfactant molecules inall the systems described above were still selfassembled into closed bilayers after prolongedperiods and Figs. 21 and 23 represent the firstrecord of niosomes and to our knowledge vesiclesof any kind that had been stored for over 7 years!However although this data pointing to a persis-tence of surfactant self assembly is impressive, notmuch is revealed from the above on the stabilityof a drug delivery system in all its aspects afterthis length of time.

7.1. Influence of the surfactant:lipid nature

The choice of membrane surfactant determinesthe nature of the membrane and ultimately affectsthe stability of the system. The leakiness of CFloaded Span surfactant niosomes was found tofollow the trend Span 80BSpan 20BSpan 40BSpan 60 (Yoshioka et al., 1994) and was deter-mined by the degree of membrane fluidity Theincorporation of cholesterol into these niosomalsystems also decreases the leakiness of the mem-brane (Rogerson et al., 1987, 1989).


7.2. Influence of the encapsulated drug

The encapsulated drug could also be the majordeterminant of the fate of any niosomal system.75% of the drug polymer conjugate (PK1) re-mained encapsulated within the vesicles 28 daysafter storage as the vesicle suspension at 4 and at25C (Gianasi et al., 1997). Vesicle size was alsofound to remain unchanged (Gianasi et al., 1997).The encapsulation of a polymer obviously leads toa more stable system as the membrane is suffi-ciently impermeable to this macromolecule.

The physical nature of the encapsulated mate-rial also affects stability. DOX loading into vesi-cles using an ammonium sulphate gradient is saidto lead to the formation of a gel within the

vesicles (Haran et al., 1993). Niosomes loadedusing this technique were also less leaky (Fig. 15)(Uchegbu et al., 1996b).

7.3. Temperature of storage

The temperature of storage of these dispersionsmust be controlled as a change in the temperatureof the system often leads to a change in thefundamental nature of the system (Tanaka, 1990)or an increase in the release of an encapsulatedsolute (Santucci et al., 1996; Uchegbu et al., 1997)a property which may be exploited to construct athermoresponsive system (Fig. 20b).

7.4. Detergents

High concentrations of detergents (soluble sur-factants) are incompatible with niosomal systemsand cause eventual solubilisation of the vesicles toform mixed micelles and a host of intermediateaggregates (Lesieur et al., 1990; Seras et al., 1992;Uchegbu et al., 1992; Serascansell et al., 1996;Uchegbu et al., 1996a). This solubilisation hasbeen studied for a few formulations and the de-struction of C16G2 niosomes by octyl glucosideappears to proceed via the build up of a criticallocalised concentration of octyl glucosidemolecules within the niosome membrane beforemicellisation can occur (Seras et al., 1994). Thesolubilisation C16G2 niosomes by Solulan C24 hasbeen shown to proceed via the formation of dis-comes which are then converted into mixed mi-celles (Uchegbu et al., 1992).

7.5. Thermodynamic considerations

Hydrated bilayer systems such as liposomes andniosomes are not deemed to be thermodynami-cally stable and are thought to represent ametastable state in that the vesicles possess anexcess of energy (Lasic, 1990). These particles arethus predicted to transform into bilayer stackswith time. To produce a system with maximalstability thus requires that these predicted trans-formations be slowed down to such an extent asto produce a product with a reasonable shelf life.In effect the colloidal dispersion must not show

Fig. 21. (500) micrograph of C16G3, cholesterol, dicetylphosphate (47.5:47.5:5) niosomes that had been stored for 84months (bar20 mm).


Fig. 22. (1000) micrographs of doxorubicin C16G2, cholesterol, Solulan C24 (40:40:10) niosomes that had been prepare by handshaking and high pressure homogenisation and stored for 84 days. The original particle size of this dispersion was 79 nm. Thepresence of doxorubicin which is preferentially located at the membrane is confirmed by (a) the fluorescence and (b) the dark colour(bar10 mm).

vesicle aggregation, fusion or swelling. The en-trapped material must remain entrapped andshow no incompatibility with the bilayermaterials.

7.6. Stability enhancement

Methods to enhance the stability of these nio-somes are also found in the literature. Decreasingthe air water interface may prevent the crystallisa-tion of these self assembled surfactant monomers(Engberts, Hoekstra, 1995) and it may be possibleto stabilise niosomes by a variety of methods suchas the addition of polymerised surfactants to theformulation, the use of membrane spanning lipidsand the interfacial polymerisation of surfactantmonomers in situ (Florence, 1993a).

The inclusion of a charged molecule in thebilayer shifts the electrophoretic mobility makingit positive with the inclusion of stearylamine andnegative with the inclusion of DCP and alsoprevents niosome aggregation (Cable, 1989). Inaddition, as mentioned above, the entrapment ofhydrophobic drugs (Uchegbu et al., 1996b) ormacromolecular prodrugs (Gianasi et al., 1997)also increases the stability of these dispersions.

8. The evaluation of niosomes as drug deliveryagents

Although pharmaceutical niosome formulationshave yet to be commercially exploited, a numberof studies have demonstrated the potential ofniosomes in drug delivery.

Examination of the literature reveals that onintravenous administration of niosomes thehighest drug levels are found in the liver (Azminet al., 1985; Baillie et al., 1986; Ozer et al., 1991;Uchegbu et al., 1995). However there were excep-tions. When DOX 850 nm C16G3 niosomes wereadministered, DOX liver levels were not signifi-cantly different from the administration of DOXsolution (Rogerson et al., 1988) and serum levelsthough low (0.5% of the administered dose 10min after dosing) were higher for the niosomeformulation. The cause of this non liver uptake isnot apparent although smaller DOX niosomes arefound to accumulate in the liver following intra-venous administration (Uchegbu et al., 1995).Liver accumulation is also avoided when DOXvesicles are prepared from palmitoyl muramicacid, cholesterol, Solulan C24 (45:45:10)(Uchegbu, 1998). With these muramic acid based


vesicles 5% of the intravenously administered dosewas found in the liver 5 h after dosing, comparedwith 18% of the intravenously administered dosestill present in the liver at this time point when thedrug was administered in the form of sorbitanmonostearate, cholesterol, Solulan C24 (45:45:10)niosomes (Uchegbu, 1998). Splenic uptake is notavoided with the use of palmitoyl muramic acidvesicles (Uchegbu, 1998).

The intravenous administration of iopromideC16G3 and C16C12G7 niosomes containing steary-lamine and extruded through a 220 nm filterresulted in these niosomes being found predomi-nantly in the kidneys (Erdogan et al., 1996). Theauthors conclude that the incorporation of thesurface positive charge (stearylamine) enablestargeting to the kidneys.

In vitro studies reveal that niosomes preparedfrom ester surfactants were found to be degradedto a larger extent by esterases which are normallypresent in the plasma although they resist degra-dation by phospholipase A2 which readily de-

grades dipalmitoylphosphatidylcholine liposomes(Florence, Baillie, 1989). The use of ester surfac-tants will thus affect the stability of niosomes inthe plasma and ultimately the biodistribution ofthe drug in niosome formulations. The intraperi-toneal administration of niosomes (Span 85based) (Jain and Vyas, 1995a) results in targetingto the lymphatics while the C16G2 DOX formula-tion (Uchegbu et al., 1994) on intraperitonealadministration acts as a depot within the peri-toneum, eventually peaking in the plasma 2 hafter dosing.

The only conclusion that may be drawn fromthese studies is that issues such as niosome surfacenature, in vivo stability and vesicle size all con-tribute to the observed biodistribution on par-enteral administration. The nature of theencapsulated drug may also affect the chemistryof the niosome surface and in this way influencebiodistribution

An additional factor to be taken into accountare the type of assay methods used in variousreports. Methods that detect clinically relevantdrug concentrations such as high performanceliquid chromatography (HPLC) with specific de-tection parameters are more discerning thanmethods measuring radioactivity which detectmetabolic fragments in addition to the parentdrug.

The transdermal (Reddy, Udupa, 1993;Hofland et al., 1994; Schreier, Bouwstra, 1994)and oral routes (Azmin et al., 1985; Yoshida etal., 1992; Rentel et al., 1996) have also beenexplored and are discussed in greater detail below.

It must also be stated that drug metabolism invivo is also altered by encapsulation within nio-somes (Azmin et al., 1985; Kerr et al., 1988;Rogerson et al., 1988; Al-Angary and Halbert,1992; Uchegbu et al., 1995)

8.1. Anti-infecti6e agents

Indeed one of the earliest diseases for whichniosomal formulations proved particularly benefi-cial was from the antiparasitic class, specifically inthe treatment of experimental leishmaniasis (Bail-lie et al., 1986). The intravenous administration ofsodium stibogluconate C16G3 niosomes or

Fig. 23. (250) micrograph of vesicles prepared from C16G3,Solulan C24 (50:50) stained with toludine blue and store for 84months. These vesicles are analogous to the discomes found inregion 4 (Fig. 9) of the C16G2-cholesterol-Solulan C24 ternaryphase diagram (bar40 mm).


dipalmitoyl phosphatidylcholine (DPPC) lipo-somes both containing 30 and 20% cholesterol,respectively, resulted in higher liver levels of anti-mony when compared with the administration ofthe drug in solution (Baillie et al., 1986). Theseniosomes were prepared by the ether injectionmethod and thus are presumed to be in the 300nm1 mm size range. There was no apparentdifference in antimony levels when liposomes orniosomes were administered. When the liver para-site burden was assessed in this study it was foundthat niosomal sodium stibogluconate was signifi-cantly more active in reducing parasite burdenthan the free drug and the effect observed aftermultiple dosing suggests that niosomal formula-tions act as a depot within the liver (Baillie et al.,1986). In addition there appeared to be no differ-ence in activity when a series of other non-ionicalkyl glycerol ether and ester surfactants wereused providing antimony dose levels were in ex-cess of 40 mg per mouse (Hunter et al., 1988).With one formulation (C16G3, cholesterol 70:30)at antimony doses below 40 mg per mouse therewas an inexplicable increase in the liver parasiteburden (Hunter et al., 1988) not previously seen inearlier studies (Baillie et al., 1986), indicative of avariation in response with this particular formula-tion. This increase in parasite burden was notobserved when cholesterol levels in C16G3 nio-somes were increased to 50% (Hunter et al., 1988).The use of specified niosome formulations at aparticular dose level to reduce liver Leishmaniadono6ani parasite burdens thus appears provenand can be attributed to the rapid uptake of theseparticulate formulations by the liver on intra-venous administration. However the failure ofthese formulations to eradicate splenic and bonemarrow parasite burdens would allow for relapseof the disease (Collins et al., 1993) leading to theconclusion that formulations capable of deliveryto spleen and bone marrow regions are required(Carter et al., 1988). Niosomes prepared frompolyoxyethylene surfactants C16EO2, C16EO4,C16EO6 and the polyglycerol surfactant C18G3,however did suppress parasite burdens in thespleen when compared with controls (Williams etal., 1995). C16EO6 niosomes also suppressed theproliferation of bone marrow parasites (Williams

et al., 1995). All niosomes used against leishma-niasis in these more recent studies contained inaddition to the main surfactant 40% cholesteroland 10% dicetyl phosphate.

Rifampicin, an anti-tuberculosis agent, encap-sulated within Span 85 (sorbitan tri-oleate) basedniosomes in the 815 mm size range were foundto accumulate in the lung of mice (Jain and Vyas,1995b) thus offering the possibility of improvedanti-tuberculosis therapy.

8.2. Anticancer drugs

8.2.1. MethotrexateWhen methotrexate 100 nm C16G3 niosomes

containing either 47.5 or 30% cholesterol wereadministered intravenously or orally higher levelsof the drug were found in the livermore so forthe formulations administered by the intravenousroutewith serum levels higher than when thedrug was administered in solution (Azmin et al.,1985). A 23-fold increase in the area under theplasma level time curve was observed when Span60 4.5 mm methotrexate niosomes were adminis-tered by the intravenous route to tumour bearingmice (Chandraprakash et al., 1993) a fact at-tributed to the large size of these niosomes. Thearea under the plasma level time curve was in-creased 100-fold when methotrexate Span 60 nio-somes were administered following macrophageactivation with muramyl dipeptide-gelatin conju-gates and methotrexate tumoricidal activity wasalso increased (Chandraprakash et al., 1993). Ear-lier studies with methotrexate had found thatincreased levels of methotrexate were foundacross the blood brain barrier with C16G3 nio-somes administered by the oral or intravenousroute (Azmin et al., 1985). Unfortunately drugdelivery to the brain with niosomes is not an areathat has enjoyed extensive study and the potentialof these formulations to deliver drugs across theblood brain barrier while indicated by these re-sults still awaits confirmation and possibleexploitation.

Metastitial cancer of the lymphatic system hasbeen targeted by the administration of 8 mm Span85 methotrexate niosomes intraperitoneally (Jainand Vyas, 1995b). High levels of methotrexate


Fig. 24. (a) Mean plasma levels (9S.E.M.) of doxorubicin after the intravenous administration of doxorubicin (10 mg kg1) Span60, cholesterol, Solulan C24 (45:45:10) niosomes to female tumour bearing mice,niosome-associated doxorubicin,doxoru-bicin released from the niosomes in vivo (Uchegbu et al., 1995). (b) Mean% of plasma doxorubicin (9S.E.M.) still encapsulated inniosomes after intravenous administration of doxorubicin (10 mg kg1) Span 60, cholesterol, Solulan C24 (45:45:10) niosomes tofemale NMRI mice (Uchegbu et al., 1995).

were found in the thoracic lymph following nioso-mal administration by this route when comparedwith administration via the intravenous route andthe administration of the free drug via the peri-toneal route.

8.2.2. DOXDOX administered in 850 nm C16G3 niosomes

to tumour bearing mice resulted in increased tu-mour, serum and lung levels but not in an in-crease in liver levels of DOX (Rogerson et al.,1988). DOX 240 nm Span 60 niosomes increasedplasma, liver and tumour levels (Uchegbu et al.,1995). Lung levels were also increased. Fractiona-tion of plasma samples over a Sepharose 2Bcolumn revealed that 90% of plasma DOX wasstill encapsulated within Span 60 niosomes up to 4h after intravenous dosing, falling to 50% by 24 h(Fig. 24) (Uchegbu et al., 1995). Initial plasmalevels and liver levels have been found to be thehighest with this formulation than any of theother DOX formulations studied (Fig. 25a, b)(Uchegbu et al., 1996a). Tumoricidal activity wasincreased with different DOX niosome formula-tions as measured by decreased proliferation ofthe S180 sarcoma in NMRI mice (Fig. 26)(Rogerson et al., 1988) and terminal mean tumour

weight of a MAC 15A tumour in NMRI mice(Uchegbu et al., 1996a). However studies involv-ing a human lung (Kerr et al., 1988) or humanovarian xenograft (Uchegbu et al., 1996a) re-vealed that in these latter models niosomal formu-lations had no advantage over the free drug.

It is well known that good clinical anti-tumouractivity cannot be predicted by the use of animalmodels but it is clear from the above that thechoice of animal model may prove crucial at theearly in vivo screening stage (Uchegbu et al.,1996a). It is advocated that a large number ofanimal models are chosen to evaluate deliverysystems containing these broad spectrum anti-can-cer agents such as DOX. The apparent dis-crepancy in response offered by the mouse andxenograft tumour models may lie in differences intumour anatomy:biochemistry.

DOX is a cardiotoxic drug and it is thus benefi-cial to observe that heart levels of the drug wereslightly decreased with niosomal formulations(Kerr et al., 1988; Rogerson et al., 1988) and thatcholesterol free niosomes decreased heart levels toa larger extent than the cholesterol containinganalogues (Rogerson et al., 1988). Ultimate drugtargeting requires that toxicity be minimised andthat therapeutic benefit be maximised. In this


respect the administration of DOX in niosomesdoes offer some benefit. However drugmetabolism was increased on the administrationof DOX niosomes (Kerr et al., 1988; Rogerson etal., 1988; Uchegbu et al., 1995) due to the in-creased bioavailability of the drug.

8.2.3. DOX N(2-hydroxypropylmethacrylamide)copolymer conjugate

A DOX PK1 copolymer in which DOX isbound to N(2-hydroxypropyl methacrylamide) by

Fig. 26. The growth in the mass of implanted tumour as afunction of time after the intravenous injection of () phos-phate buffered saline pH 7.4, () doxorubicin solution (5 mgkg1), () doxorubicin (5 mg kg1) C16G3 (100) niosomes,() doxorubicin (5 mg kg1) C16G3, cholesterol (50:50).

Fig. 25. Murine tissue levels of doxorubicin following theintravenous administration of doxorubicin niosomes, () boxfillC16G3 (100) niosomes, size850 nm, (X) box fillC16G3, cholesterol (50:50) niosomes, size850 nm, (fl) boxfillSpan 60, cholesterol, Solulan C24 (45:45:10), size235nm (a) liver and (b)plasma (Uchegbu et al., 1996a).

a peptidyl spacer (Duncan et al., 1989) has beendesigned to enable DOX targeting to tumours.Following fluid phase pinocytic uptake of thisdrug polymer conjugate the active drug will becleaved off by lysosomal enzymes (Duncan andSpreafico, 1994) and become available intracellu-alrly to exert its pharmacological effect. Recentlythis polymeric prodrug has been encapsulatedwithin sorbitan ester, alkyl polyoxyethylene etherand C16G2 niosomes (Duncan et al., 1997; Gianasiet al., 1997; Uchegbu, Duncan, 1997).

In vitro, this drug delivery system was found tobe stable in plasma and no free DOX was de-tected although free DOX was detected when thissystem was incubated with a lysosomal enzymepreparation (Gianasi et al., 1997). This indicatesthat free DOX would only be released in vivoonce the formulation was exposed to lysosomalenzymes. In vivo, 420 nm C16G2 PK1 niosomesyielded non-detectable plasma levels of DOX buthigh spleen levels of PK1 and low spleen levels ofDOX (Uchegbu, Duncan, 1997). Liver levels ofDOX were found to rise steadily with time until atthe 24 h time point, 25% of the administered dose


was present in the liver (Fig. 27). This is thehighest DOX liver level we are currently aware ofwith any drug targeting strategy at the 24 h timepoint. This unusual liver profile is due to the factthat large amounts of the encapsulated polymerprodrug are taken up by the liver and DOX isthen slowly released by enzymatic cleavage. It ispossible that this polymer in niosome formulationmay act as a depot within the liver giving rise toa controlled release of an active substance in adefined organ. This formulation may be useful inthe treatment of hepatic metastasis

8.2.4. Other agentsVincristine Span 40 niosomes increased the vin-

cristine anti-tumour activity in S-180 sarcoma andErlich ascites bearing mice (Parthasarathi et al.,1994). Span 60 bleomycin niosomes also increasedthe tumoricidal activity of bleomycin in these twotumour models (Naresh, Udupa, 1996) Howeverthe intraperitoneal administration of sucrosepalmitate stearate ester (HLB7) niosomes wasfound to result in no apparent increase in life spanwhen mice bearing a P388 murine leukaemia wereadministered glyoxal-bis-guanyl hydrazone encap-sulated within these niosomes (Schenk et al.,1990).

8.2.5. Anti-cancer niosomes and the futureNiosomes are proving useful agents in the

targeting of anti-cancer drugs although problemssuch as the choice of suitable animal tumourmodels still remain unresolved.

8.3. Anti-inflammatory agents

Diclofenac niosomes reportedly prepared frompolysorbate 60, cholesterol and DCP (22:73:5)and 3 mm in size were found to reduce the inflam-mation in rats with carageenan induced pawoedema on intraperitoneal administration to agreater extent than the free drug (Naresh et al.,1993). This increase in activity is a direct result ofan observed increase in the area under the plasmatime curve. In a similar study involving flurbipro-fen Span 60 niosomes (Reddy, Udupa, 1993) therewas also increased bioavailability and an in-creased reduction of carageenan induced rat pawoedema when the niosomal formulation was ap-plied topically in a hydroxypropyl methyl cellu-lose semisolid base containing 10% glycerine ororally as a suspension in saline.

8.4. Diagnostic imaging with niosomes

Apart from the use of niosomes as various drugcarriers one report in the literature details theevaluation of these systems as diagnostic agents.C16G3 and C16C12G7 niosomes containing choles-terol and stearylamine encapsulating the radio-opaque agent iopromide were found toconcentrate in the kidneys on intravenous admin-istration (Erdogan et al., 1996). This kidneytargeting was attributed to the presence of thepositive charge on the niosome surface althoughno neutral control niosomes were used in thisstudy. The highest kidney iopromide concentra-tion was found with the C16G3 niosomes whichare said to form a less fluid bilayer than theC16C12G7 niosomes (Erdogan et al., 1996). Al-though the niosome formulation enhanced theopacity of this contrast agent, low levels of encap-sulation of the contrast agent were a problemwith this system and clinically relevant enhance-ment of opacity was not achieved in this study.

Fig. 27. The biodistribution of free doxorubicin after theintravenous administration of PK1 (a doxorubicin polymerconjugate) C16G2, niosomes (5 mg kg

1 doxorubicin) to maleBALB:C mice, (solid) fillheart, (open) fillspleen, (X)fill liver.


8.5. Niosomal formulations as 6accine adju6ants

A number of surfactants have documented im-munostimulatory properties (Hilgers et al., 1989)and have been used in emulsion vaccine adju-vants. The adjuvanticity of niosomes preparedfrom 1-mono-palmitoyl glycerol, cholesterol,dicetyl phosphate5:4:1 has been demonstratedin mice, on subcutaneous administration ofbovine serum albumin (Brewer and Alexander,1992) ovalbumin (Brewer et al., 1996) or a syn-thetic peptide containing a known T-cell epitope(Brewer et al., 1996). The encapsulation of theantigen was determined to be crucial to the adju-vanticity (Brewer and Alexander, 1992). The sameniosome system has also been shown to act as avaccine adjuvant when administered intraperi-toneally to severe combined immunodeficiencymice reconstituted with peripheral bloodlymphocytes (PBL-SCID mice) (Walker et al.,1996). This mouse model is designed to mimic thehuman response to an antigen challenge (Walkeret al., 1995). A vesicle in water in oil (v:w:o)emulsion (Yoshioka, Florence, 1994) preparedfrom Span 80 and cotton seed oil has been evalu-ated as an immunological adjuvant using themodel antigen tetanus toxoid (Yoshioka et al.,1995). An increased secondary response (level ofIgG1) was observed when the v:w:o formulationwas administered by the intramuscular route incomparison with the vesicle formulation and thefree antigen

8.6. Niosomes and the oral route

The oral delivery of drugs using niosomal for-mulations was first demonstrated by a study in-volving 100 nm methotrexate C16G3 niosomes(Azmin et al., 1985). Significantly higher levels ofmethotrexate were found in the serum, liver andbrain of PKW mice orally administered a nioso-mal formulation when compared with the admin-istration of the free drug. It thus appears thatthere is enhanced drug absorption with these nio-somal formulations.

Recently more gut labile compounds such asproteins (Rentel et al., 1996) have been adminis-tered by this route and the intestinal transport of

a niosomal formulation of the peptide 9-desglyci-namide 8-arginine vasopressin has been studied invitro (Yoshida et al., 1992). In one of the firststudies of its kind sucrose ester niosomes loadedwith ovalbumin were found to cause a modest butsignificant increase in the level of specific antibod-ies after oral administration (Rentel et al., 1996).

8.7. Transdermal drug deli6ery

Although the emergence of niosomes into thepharmaceutical arena was the result of activity inthe cosmetic industry, it was only fairly recentlythat the transdermal delivery of drugs with nio-somes was seriously considered. The enhanceddelivery through the stratum corneum of niosomeencapsulated drugs has been observed (Reddy,Udupa, 1993; Niemec et al., 1994; Schreier,Bouwstra, 1994; Vanhal et al., 1996) and it there-fore remains to elucidate the mechanism of thisdelivery, especially as the stratum corneum isconsidered to be a particularly impermeable bar-rier (Junginger et al., 1991). Small (100 nm) vesic-ular structures have been observed between thefirst and second layer of human corneocytes 48 hafter incubation with niosomes prepared fromdodecyl alcohol polyoxyethylene ether andcholesterol (Junginger et al., 1991). Penetration byniosomes of this upper layer appears plausible asthese layers are only loosely packed (Junginger etal., 1991). However the same study reports thepresence of vesicular structures in the deeperseemingly inaccessible areas of the skin and con-cludes that there was a reorganisation of theniosome membrane into individual monomerswhich on arriving at these deeper layers reformedinto niosomes (Junginger et al., 1991). In vitrostudies on the transdermal penetration (passagethrough pieces of excised stratum corneum) ofoestradiol using high phase transition sucrose es-ter niosomes or C18EO7 niosomes and low phasetransition C12EO7 niosomes revealed the latter tobe better transdermal carriers (Vanhal et al.,1996). C12EO7 micelles were ineffective as drugcarriers in this study. The higher flexibility ofthese bilayers is said to be responsible for thisimproved transdermal penetration (Vanhal et al.,1996) a similar explanation to that given for the


transdermal delivery of material by transfersomes(flexible liposomes) (Cevc et al., 1995; Cevc, 1996;Cevc et al., 1996). Reducing the cholesterol con-tent of these niosomes also increased the transder-mal delivery of oestradiol (Vanhal et al., 1996).

The migration of cyclosporin A from c

1-s2.0-S0378517398001690-main

Documents

Transcript of 1-s2.0-S0378517398001690-main