Investigation on factors affecting drug delivery using polymers and phospholipids 3

For thispurpose the complexation of haloperidol with two derivatives of β-CD RM β-CD and HP β-CD at pH 5 were studied by the phase solubility method.. To elucidate the influence of pH of

Haloperidol (HP) is practically insoluble in water and has a basic pK of 8.3 (Lim etal., 2006) To increase the solubility, pH of 2.5-3.8 and 2.5-4.5 are used for injectionand oral dosage forms respectively However such acidic solutions can causeirritation in the site of injection (Loukas et al., 1997).

Current approaches to soubilize water-insoluble drugs are complex formation withcyclodextrins, liposomes, microemulsion-based drug delivery systems andsupersaturation Cyclodextrins (CDs) are attractive candidates for increasing theaqueous solubilities of lipophilic drugs They are cyclic oligosaccharides of D-glucopyranose units in the shape of cones, each with an outer hydrophilic surface and

an inner hydrophobic cavity The solubilization effect of CDs is due to the formation

of a non-covalent water soluble inclusion complex, therefore drug-CD complexes are

easily dissociated and in equilibrium with free drug (Loukas et al., 1997; Liu et al.,2003) Due to the solubility, hygroscopicity and toxicity concerns of CDs, they weremodified and examples are hydroxypropyl β-CDs (HP β-CDs) and randomlymethylated β-CD (RM β-CDs) (Liu et al., 2003; Gibaud et al., 2005; Murthy et al.,2004).

CD derivatives can influence the solubilities of drugs (Loukas et al., 1997; Liu et al.,2003; Sigurðardóttir and Loftsson 1995) They were also reported to decrease localirritation (Amdidouche et al., 1994; Hoshino et al., 1989; Ventura et al., 2006) as well

as stabilize photosensitive drugs (Godwin et al., 2006) Some investigators reportedthat CDs increased the skin permeation rates of drugs by extracting the lipid from theskin (Bently et al., 1997; Okamoto et al., 1986; Uekama et al., 1982; Vianna et al.,1998) while others reported that CDs did not show any enhancing effect on the fluxrates of drugs through the skin (Larrucea et al., 2001; Shaker et al., 2003; Williams etal., 1998)

The pH of the vehicle influences the solubility and partitioning of the drug into theskin, implying that the ionized and unionized moieties of a drug influence itssolubility and partitioning through stratum corneum and hence affect the skinpermeation (Hadgraft and Valenta 2000; Sridevi and Diwan 2000a and b) Wagner’sgroup reported that pH values of donor and receptor compartments influence skin pHand change the permeability of the drug (Wagner et al., 2003) On the contrary,Sznitowska’s team reported that there were no significant differences in permeability

of hydrocortisone in the pH range of 1-10, and only extreme pH values affectedpermeability (Sznitowska et al., 2001; Thune et al., 1988).

To further understand the effect of CD and pH, the aim of the present work is toinvestigate the solubility and permeation of a drug from CD inclusions For thispurpose the complexation of haloperidol with two derivatives of β-CD (RM β-CD and

HP β-CD) at pH 5 were studied by the phase solubility method Molecular modelingwas conducted using DM β-CD (Dimethyl--cyclodextrin) and HP β-CD Surfacetension and contact angle measurements were carried out to further elucidate theeffect of CDs on the permeability of HP though human epidermis The effect ofconcentrations of RM β-CD alone and then combined with limonene on the skinpermeation were studied To elucidate the influence of pH of the donor phase on skinpermeability, further experiments using phosphate buffer at pH 5 in the donorcompartment alone and also in combination with RM β-CD were carried out Then,

RM β-CD was added to the receptor solution to maintain a sink condition, while thedonor compartment consisted of solutions of HP in RM β-CD or propylene glycol

3.2 Materials and Methods

3.2.1 Materials

Haloperidol (HP) was purchased from Sigma, Singapore cyclodextrin (HP β-CD) (degree of substitution of about 0.6) and randomlymethylated-β-cyclodextrin (RM β-CD) (degree of substitution of about 1.8) were kindgifts from Roquette (Lestrem, France) and Wacker (Burghausen, Germany),

3.2.2 HPLC Analysis

HP concentration was quantified by HPLC from Shimadzu (Kyoto, Japan) 2010A.The analysis was carried out using a reversed-phase Waters Symmetry Shield column(3.5 m, 3.0 mm  100 mm) Mobile phase was a 55:45 volume ratio of acetonitrileand 0.05M phosphate buffer adjusted to pH 3 using phosphoric acid, flowing at a rate

of 0.4 ml/min UV detection at wave length 254 nm, injection volume 100 μL gave aretention time of 5 min Standard solutions of HP (0.05 - 2 μg/ml) were prepared in0.03% v/v lactic acid (Lim et al., 2006)

3.2.3 Molecular Modeling

Molecular modeling was carried out to elaborate the complexation modes

Dimethyl--cyclodextrin (DM β-CD) was adopted as a substitution for RM β-CD (degree ofsubstitution = 1.8) to facilitate the determination of the stable structure of thehaloperidol-RM β-CD complex, for RM β-CD is a mixture of different structures

HP β-CD, with a degree of substitution of 0.6 was used for the experimental study;four 2-hydroxypropyl groups were added on the primary hydroxyl groups of -cyclodextrin (Mura et al., 1995) The structures of haloperidol, DM β-CD and HP β-

CD were individually minimized by MMFF94s force field using software SYBYLversion 7.2 (Tripos Co., USA) The structures of haloperidol, DM β-CD and HP β-

CD were individually minimized by MMFF94s force field using software SYBYLversion 7.2 The HP molecule, in its favorable conformation was introduced into the

respective DM β-CD and HP β-CD cavities and the interaction energies werecomputed The most likely conformation of each complex was the one with thelowest interaction energy.

3.2.4 Phase Solubility Studies

Drug-CD inclusion complexes were prepared by adding an excess concentration of

HP (15 mg/ml), dissolved in water or buffer phosphate (pH 5), using RM β-CD and

HP β-CD solutions of different concentrations (0, 0.01, 0.05, 0.1, 0.2, 0.3 M) Thesuspensions were shaken on a horizontal rotary shaker in the absence of light for 7days and finally filtered through a membrane filter (Millipore filters®, 0.45 μm poresize, 25 mm diameter) to obtain clear solutions All samples were prepared intriplicates The concentrations of HP in the inclusion complexes were determined bythe HPLC assay

3.2.5 Surface Tension and Contact Angle Measurements

Surface tensions of RM β-CD and HP β-CD solutions and each formulation used inthe permeation study were measured using the method described previously in section2.2.3 The wettability of the excised human skin sample was determined by sessiledrop contact angle using a Rame-Hart 100 goniometer (USA) Drops were placed onthe surface using a micrometer with a flat tip needle For contact angle measurementsexcised human skin similar to that used in permeation studies was employed Thiswas done because of its availability and its potential for elucidation of the mechanism

of drug permeation studies Skin samples were prepared as those for the permeation

3.2.6 In vitro Skin Permeation Studies

Permeation studies of drug alone or as complexes with RM β-CD were performedusing a flow-through diffusion cell apparatus described earlier in section 2.2.8 Thedonor compartment was filled with 1 ml of formulations containing 2 mg/ml drug.The first receptor phase was isotonic phosphate buffer saline pH 7.4 (PBS) Samplesfrom the receptor phase were collected every 6 hour over a 30-h period, and theamount of HP permeated was analyzed by HPLC The steady state flux (J) wasestimated from the slope of the straight line portion of the cumulative HP absorbedagainst time profile Experiments were carried out in triplicates

The effect of RM β-CD on the skin permeation of HP was studied using two sets ofexperiments First a concentration dependent effect of RM β-CD (0, 0.01, 0.05, 0.1M) was studied Further, synergistic effect of RM β-CD in combination withlimonene 0.1% v/v in propylene glycol (PG) solution was investigated The effect of

pH on the permeability of the drug was studied at pH 5 The combined effects ofionization and 0.01 M RM β-CD were also observed In another set of experiments,PBS in the receiver solution was replaced by 0.01% w/v RM β-CD, while the donorcompartment consisted of HP in 0.01 M RM β-CD or PG solutions

3.3 Results and Discussion

3.3.1 Molecular Modeling

The hypothetical structures of the complexes formed by haloperidol and cyclodextrinsare presented in Fig 3.1 For the haloperidol-DM β-CD complex, the computed totalenergy is -40.7 kcals/mol and the steric energy is -30.587 kcals/mol For thehaloperidol-HP β-CD complex, total energy is -39.6 kcals/mol and the steric energy is-29.848 kcals/mol The differences in energy values indicate that the interactionbetween haloperidol and DM β-CD might be stronger than that of haloperidol and HPβ-CD

a (1) a (2)

Fig 3.1 (a) Hypothetical structure of the haloperidol-DM β-CD complex, and (b)

haloperidol-HP β-CD complex (1) Side view; (2) Side view with electron surface; (3) Top view; and (4) Top view with electron surface.

3.3.2 Solubility Studies

The solubilities of HP in phosphate buffer of pH 5 solutions with and without RM

β-CD or HP β-β-CD are presented in Fig 3.2 This pH was selected as it is the same pH

of the skin and may therefore minimize skin irritation The highest increase in drugsolubility occurred for RM β-CD, indicating that this oligosaccharide complexed

more of the drug than HP β-CD Molecular modeling supports the results obtainedfor solubility profile such that the increase in solubility was due to greater interaction

of HP with DM β-CD The solubilization profile in Fig 3.2 is linear for allformulations indicating the formation of a 1:1 complex irrespective of the ionization

of the drug (Loukas et al., 1997) More solubilization was achieved when the drugwas in its degree of ionized form in RM β-CD, resulting in a 128-fold increase of theintrinsic solubility of the drug When phase solubility experiments were performedwith CD in the presence of buffer, the change in solubility was higher than in thepresence of CD alone, indicating a synergistic effect Methylated CDs have beenobserved to have larger cavity volumes than HP β-CD Consequently, RM β-CD caneasily accommodate the hydrophobic drugs such as HP (Torque et al., 2004)

Fig 3.2 Phase solubility of haloperidol in CD solutions (n=3).

3.3.3 Surface Tension and Contact Angle Measurments

The surface tensions of aqueous solutions of different concentrations of RM β-CD and HP β-CD are shown in Fig 3.3 A remarkable change in the surface tension of

pure water occurred when RM β-CD or HP β-CD was added, indicating that these

systems have effect on the surface tensions of pure water The methylatedcompounds showed the most reduction (Evrard et al., 2004; Thompson 1997) FromFig 4.3 it is evident that surface tension reached a constant value after certainconcentration suggesting the formation of super molecular aggregates of RM β-CDand HP β-CD (Leclercq et al., 2007; Binkowski-Machut et al., 2006) Critical micelleconcentration values were determined from the sharp changes in the slope of thesurface tension versus log [CD] plot CMC values for RM β-CD and HP β-CD were4.4 mM and 3.2 mM respectively The aqueous solution of naturally occurring β-CDdoes not have any surface activity (Leclercq et al., 2007; Lu et al., 1997)

45 55 65 75

Fig 3.3 Surface tension of RM β-CD and HP β-CD (n=3).

Based on above results, a possible mechanism for the formation of large micelleassemblies was deduced as shown in Fig 3.4 Particle size analysis, as observed bythe light scattering, supported this hypothesis (Binkowski-Machut et al., 2006).However as compared with conventional surfactants these aggregates do not behave

as micellar systems This maybe due to the short length of the hydrophobic chain inthe CD structure (Leclercq et al., 2007; Lombardo et al., 2004).

Fig 3.4 Schematic aggregation of CD.

Surface tension and contact angle values of the solutions used are stated in Table 3.1

It was observed that surface tension of water (70.3 ± 0.25 mN/m) decreases withincrease in the concentration of RM β-CD The interfacial tension of RM β-CD 0.05and 0.1 M were of similar values, 54.8 ± 0.31 and 54.1 ± 0.22 (mN/m), respectively,however surface tension of RM β-CD 0.01M was 57.5 ± 0.68 mN/m Phosphatebuffer solutions had lower surface tension of 59.2 ± 0.42 mN /m, addition of RM β-

CD further decreased the surface tension to 56.5 ± 0.03 mN/m Addition of limonenedid not demonstrate any decrease in interfacial tension of the PG (36.2 ± 0.06 mN/m)when compared to pure PG solutions (36.2 ± 0.21 mN/m), indicating that limonenedoes not possess any surface active effect (see Table 3.1)

Table 3.1 Surface tension and contact angle values of the solutions (n=3).

Formulations Surface Tension (mN/m) ± SD Contact angle ± SD

as their interfacial values did not differ much

3.3.4 In vitro Skin Permeation Studies

RM β-CDs were used in permeation studies due to their significant effect on thesolubility of HP compared with HP β-CDs Fig 3.5a shows the effect of differentmolar ratios of RM β-CD on the permeation of HP RM β-CD concentrations usedfor skin permeation were in the ranges of below (0.05 and 0.1 M) and above the CMC

CD concentrations The in vitro permeation of HP through human stratum corneumshowed a similar trend in the presence of both RM β-CD 0.1 M and 0.05 M, with alow drug penetration through the skin (p > 0.05) This is not surprising as it could bedue to the super molecular arrangement and of high concentrations of RM β-CD; theaggregations produced were too large to increase drug permeability However atlower concentrations of RM β-CD (0.01 M), an increase in permeation was observed(p < 0.01) and the flux rate was 2.7-fold higher than that of control (Table 3.2) Theflux is due to HP molecules which have not formed complexes with RM β-CD Anincrease in RM β-CD concentrations caused a decrease in dissociated HP molecules,therefore the amount of free drug available for permeation decreased (Shaker et al.,2003; Dias et al., 2003) From the phase solubility diagram, it is evident that theCDs are potent solubilizers, however it is important to use just enough CD to dissolvethe drug; addition of too much CD will decrease drug partitioning into the skin(Loftsson et al., 1991 and 1994; Felton 2002) It can be seen that the controvercialresults regarding the role of CD derivatives as penetration enhancers reported byother scientists may be related to their surface active behavior and depends on theconcentration of the CD At concentration above the CMC values the skin penetration

of the durg molecule may not be affected where as at lower concentrations the skinpermeation of the active compound may be significantly increased

CD has been used in combination with chemical enhancers (Larrucea et al., 2001;Maestrelli et al, 2005 and 2006; Zerrouk et al., 2006) and electroporation (Murthy etal., 2004) to increase the skin permeation rate of drugs Previous studies from ourlaboratories showed that limonene acts as a good penetration enhancer for the

delivery of HP (Lim et al., 2006) To investigate the possible use of CD as acandidate co-enhancer, additional tests were carried out by combining limonene with0.01 M RM β-CD As shown in Fig 3.5b and Table 3.2, the combination of 0.01 M

RM β-CD with 0.1% v/v limonene in PG solution improved percutaneous absorption

by 1.4 fold compared to the control, but not significantly (p > 0.05) However, thedrug permeation profiles were completely different and the lack of significantdifference between the flux in limonene and limonene RM β-CD solutions suggestshigh permeation enhancing effect of limonene that masks the effect of RM β-CD

Table 3.2 Flux value of HP across human epidermis (n=3).

when the buffer was used alone (p < 0.05) The higher flux is a result of increasedsolubility of the drug Synergistic effects can be achieved by adjusting the pH andconcentration of RM β-CD to obtain improved solubility and therefore skinpermeability of the drug (Sridevi and Diwan 2002 a and b; Singh et al., 2005) Bothionized and unionized moieties of drug molecules contributed to the total flux, (Jtot)which can be calculated using the following equation:

ion pion union

punion

(3-1)Where Kpunionand Kpion are the dependent drug permeabilities and Cunion and Cion arethe dependent concentrations of ionized and unionized moieties, respectively(Hadgraft and Valenta 2000) The results from the drug solubility profiles showed,that the ionized moieties have higher solubilities and therefore greater impact on thetotal flux of the drug Results from our studies suggested that the simultaneouspresence of both RM β-CD and phosphate buffer pH 5 at suitable concentrationscould be exploited to improve drug solubility and permeability resulting in enhancedpermeation of the drug though the skin

Physiological buffer solutions with pH 7.4 are normally placed in the receptor of theflow-though cell However for a less water-soluble drug, sink conditions andtherefore total flux of the drug may increase with changes in the receptor phase CDmolecule is not expected to cross the skin when used in the receptor compartment.The purpose of adding CD is to increase the solubility of the haloperidol that will begetting absorbed from the donor compartment to the receptor compartment.Maintaining sink condition will help increase the rate of drug permeation across the

skin Our recommendation to add CD to the receptor compartment would bettermimic the sink condition across in vitro skin delivery.

Our next approach was to study the possible promoting effect of the CDs in receptorsolutions Sclafani and co-workers reported that γ-CD in receiver solutionsinfluenced the permeation rate of progesterone (Sclafani et al., 1995) In our study,

RM β-CD (0.1% w/v) was chosen because of its significant increase on the solubility

of HP Formulations containing PG in donor compartments did not give significantlydifferent drug flux compared to control (p > 0.05) A substantial increase in HPpermeability was obtained by using 0.1% w/v RM β-CD as the receptor solution and

RM β-CD 0.01 M in donor compartment (Fig 3.5d) The flux was enhanced by a fold increase compared to that of the control (p < 0.05)

1.6-It was found that the critical contact angle value, which distinguishes betweenpenetration and non-penetration, helped to optimize the drug permeation rate Bycomparing Fig 3.5a and b, PG with lower contact angle, resulted in higher drugpermeation rate, compared with water Before any molecule can penetrate throughthe skin, it has to first adhere to the skin surface, thus the wetting rates and theviscosities of these formulations influence the penetration rate of the drug Thereforethe physicochemical characteristics of the vehicle and other ingredients informulation have great impact on skin permeation rate (Venkatraman and Gale 1998;Wokovich et al., 2006)