Ag

HAa

b

a

ARRAA

KTAE

1

pitsadabca[dtva

h0

International Journal of Biological Macromolecules 78 (2015) 102–111

Contents lists available at ScienceDirect

International Journal of Biological Macromolecules

j ourna l ho me pa g e: www.elsev ier .com/ locate / i jb iomac

lginate gel-coated oil-entrapped alginate–tamarindum–magnesium stearate buoyant beads of risperidone

riday Beraa,∗, Shashank Boddupalli a, Sridhar Nandikondaa, Sanoj Kumara,mit Kumar Nayakb

Department of Industrial Pharmacy, Gokaraju Rangaraju College of Pharmacy, Bachupally, Hyderabad 500090, IndiaDepartment of Pharmaceutics, Seemanta Institute of Pharmaceutical Sciences, Mayurbhanj, Odisha 757086, India

r t i c l e i n f o

rticle history:eceived 22 January 2015eceived in revised form 21 March 2015ccepted 1 April 2015vailable online 8 April 2015

eywords:amarind gumlginatemulsion gel beads

a b s t r a c t

A novel alginate gel-coated oil-entrapped calcium-alginate–tamarind gum (TG)–magnesium stearate(MS) composite floating beads was developed for intragastric risperidone delivery with a view to improv-ing its oral bioavailability. The TG-blended alginate core beads containing olive oil and MS as low-densitymaterials were accomplished by ionotropic gelation technique. Effects of polymer-blend ratio (sodiumalginate:TG) and crosslinker (CaCl2) concentration on drug entrapment efficiency (DEE, %) and cumula-tive drug release after 8 h (Q8h, %) were studied to optimize the core beads by a 32 factorial design. Theoptimized beads (F–O) exhibited DEE of 75.19 ± 0.75% and Q8h of 78.04 ± 0.38% with minimum errorsin prediction. The alginate gel-coated optimized beads displayed superior buoyancy and sustained drugrelease property. The drug release profiles of the drug-loaded uncoated and coated beads were best fit-

ted in Higuchi kinetic model with Fickian and anomalous diffusion driven mechanisms, respectively. Theoptimized beads yielded a notable sustained drug release profile as compared to marketed immediaterelease preparation. The uncoated and coated Ca-alginate–TG–MS beads were also characterized by SEM,FTIR and P-XRD analyses. Thus, the newly developed alginate-gel coated oil-entrapped alginate–TG–MScomposite beads are suitable for intragastric delivery of risperidone over a prolonged period of time.

© 2015 Elsevier B.V. All rights reserved.

. Introduction

Over the past few decades, an exclusive attention has beenaid to the development of alginate-based hydrogel beads as

ntragastric multiple unit floating drug delivery systems due toheir excellent intrinsic properties (biocompatibility, mucoadhe-ion, acid stability and ease of surface manipulation) [1–3]. Sodiumlginate is an anionic linear polysaccharide containing �-1,4-linked-mannuronic acid and �-1,4-linked l-glucuronic acid residuesrranged randomly along the chains. It has unique property ofeing instantaneously gelled when contacted with multivalentations (e.g., Ca+2, Zn+2 etc.), which has long been employed as

facile method to fabricate alginate-based drug delivery carriers4–6]. Unfortunately, floating alginate beads are suffered from poorrug entrapment and initial burst drug release with impaired ini-

ial and long-term buoyancy [7]. To minimise such inconveniences,arious additives like low-density oils, effervescent agents and suit-ble polymer-blends were incorporated. Low-density oils play a

∗ Corresponding author. Tel.: +91 8977726256.E-mail address: hriday.bera1@gmail.com (H. Bera).

ttp://dx.doi.org/10.1016/j.ijbiomac.2015.04.001141-8130/© 2015 Elsevier B.V. All rights reserved.

pivotal role in buoyancy and impose a hydrophobic barrier towardsthe drug escaping from the matrices, which results in increaseddrug trapping efficiency with prolonged release behaviour [3,7].Pioneering research in this field provided convincing evidencethat the amalgamation of magnesium stearate (MS), a low-densitymaterial, into the oil-entrapped alginate beads could furtherimprove their buoyant properties [3]. In recent years, various nat-ural biopolymers blended alginate beads are also being developed,which conferred improved retention potential and release profileof the encapsulated bioactive molecules [4,8].

Tamarind gum (TG) is a naturally occurring galactoxyloglucanisolated from seed kernel of Tamarindus indica. The backbone of TGconsists mainly of (1-4)-�-d-glucan residues substituted with �-d-xylopyranose and �-d-galactopyranosyl (1-2)-�-d-xylopyranoselinked (1-6) to glucose units [5,9]. It exhibits outstanding bioadhe-sivity and broad pH tolerance with excellent drug holding capacity.The exciting and promising features of TG encouraged our researchgroup to develop groundnut oil-entrapped TG-alginate floating

beads for gastroretentive diclofenac sodium delivery [10]. How-ever, the developed alginate–TG composite beads could displaynonhindered diffusion of the entrapped oil and surface associ-ated drug molecule due to their highly porous structure, leading

iologic

t[

bdpciflawitdlwobrm

oCrcbaa[bteefcewmbmjpwwp

2

2

CF

TC

H. Bera et al. / International Journal of B

o an initial burst drug release with poor floatation capabilities10,11].

Recently, a flurry of scientific investigations has developed theio-polymeric gel coated hydrogel beads as drug carriers, enablingrug delivery at desired rate by attenuating early swelling andremature drug release rate [12–14]. Therefore, ionotropicallyrosslinked bio-polymer coated alginate–TG blend beads contain-ng oil and MS as floating agents could be strategically designedor extended release gastroretentive drug delivery systems. A pro-onged drug release profile would be achieved from oil entrappedlginate–TG composite beads due to the hydrophobicity of MS andall effect of bio-polymeric coating [3,14]. To our knowledge, this

s the first report demonstrating a novel approach in sustaininghe release of drug in gastric environment, which is strategicallyifferent from our previous work [10,14]. It could also be specu-

ated that in addition to olive oil and MS, the bio-polymeric coatingould improve the buoyant properties of the beads by preventing

il leaching and imparting an air compartment in between coreeads and gel membrane [14]. However, no univocal study has beeneported such bio-polymers based floating systems working withultiple buoyancy control mechanisms yet.The current work illustrates an experimental demonstration

f the credibility of novel alginate gel-coated olive oil-entrappeda-alginate–TG–MS composite buoyant beads for intragastricisperidone delivery. Risperidone, an atypical antipsychotic medi-ation of benzisoxazole class, is often clinically recommended foripolar disorder and schizophrenia management [15]. It displaysn increased solubility at acidic pH relative to the solubility atlkaline pH with extremely variable elimination half-life (3–20 h)16,17]. Hence, a more reliable and predictable drug dissolution andioavailability can be ensured by extending the gastric residenceime of risperidone [14]. The oil-entrapped alginate–TG–MS beadsncapsulating risperidone as core were fabricated by ionotropicmulsion gelation technique. The beads were optimized by a 32

actorial design, where the effects of alginate to TG ratio and cal-ium chloride concentrations on responses, i.e., drug entrapmentfficiency (DEE, %) and cumulative drug release after 8 h (Q8h, %)ere studied. The release profiles of optimized beads (F–O) andarketed tablets (Riswel-4) were compared. Finally, the optimized

eads were coated with ionotropically crosslinked alginate gelembrane. The uncoated and alginate gel-coated beads were sub-

ected to various in vitro investigations. The initial, average and latehase drug diffusion coefficients and water penetration velocityere estimated for alginate–TG matrix based systems. The beadsere also examined for drug-excipients compatibility, surface mor-hology and drug crystallinity.

. Materials and methods

.1. Materials

Risperidone (Mylan Ltd., India), sodium alginate (S.D. Finehemicals Ltd., India), TG (Bhavna Gum Udhyog, India), MS (S.D.ine Chemicals Ltd., India), olive oil (relative density = 0.91 g/cm3,

able 1omposition of uncoated and alginate gel-coated oil-entrapped alginate–tamarind gum–

Ingredientsa F-1 F-2 F-3 F-4

Risperidone (mg) 300 300 300 300

Na-alginate to TG ratiob

(by weight)10:1 10: 1 10:1 5.5:1

CaCl2 solution (%, w/v) 15 15 15 10

Distilled water (ml) 50 50 50 50

Na-alginate coating gel (%, w/w) – – – –

a Each formulation contained 300 mg of magnesium stearate and 2.5% (w/v) of olive oilb Total weight of polymer was 900 mg in each formulation.

al Macromolecules 78 (2015) 102–111 103

Qualigens Fine Chemicals, India) and calcium chloride (CaCl2,Merck Ltd., India) were used. Riswel-4 (film coated tablets ofrisperidone) was commercially purchased from Sigmund ProMed-ica (India). All the chemicals and reagents were of analytical grade.

2.2. Preparation of oil-entrapped Ca-alginate–TG–MS compositebeads containing risperidone

Olive oil-entrapped alginate–TG–MS blend beads encapsulat-ing risperidone were accomplished by ionotropic emulsion gelationtechnique [2]. Briefly, the required amounts of sodium alginate andTG were dispersed in distilled water under constant magnetic stir-ring for 10 min. To it, risperidone (0.6%, w/v), olive oil (2.5%, v/v)and MS (0.6%, w/v) (Table 1) were introduced and the mixture wasagitated with a homogenizer operating at 5000 rpm for 15 min toensure emulsion stabilization. The drug-polymer in a ratio of 1:3was maintained for all the formulations. The resulting emulsionswere then extruded using 21 G needles into gently stirred gelationmedium of CaCl2 solution (5–15%, w/v) at room temperature. Theformed gel beads were allowed to stand in the solution for 20 min,collected by filtration and washed three times with distilled water.After drying overnight at room temperature, the beads were storedin the desiccators until used.

2.3. Experimental design

A 32 factorial design was employed to estimate the influencesof two independent variables viz., sodium alginate to TG ratio (X1)and CaCl2 concentration (X2) on the dependent variables like DEE(%) and Q8h (%) in simulated gastric fluid (pH 1.2). Each indepen-dent factor was varied at three levels (high, medium and low). Theexperimental trials were conducted on all the nine possible combi-nations, keeping all other parameters constant. (Table 2). Quadraticstatistical models including interactive and polynomial terms werederived to examine each response. The equation can be expressedas [18]:

Y = b0 + b1X1 + b2X2 + b3X1X2 + b4X21 + b5X2

2

where Y represents the response variable, bo denotes the arithmeticmean response of all the nine runs and bi is the estimated coeffi-cient for the independent factor Xi. X1 and X2 describe the individualeffects, X1X2 indicates interaction effects and polynomial terms (X2

1and X2

2 ) imply the nonlinearity of the model. One-way ANOVA anal-ysis was adopted to examine the significance of the models as wellas individual independent factors (p < 0.05).

2.4. Alginate-gel coating onto oil-entrapped alginate–TG–MScomposite beads containing risperidone

The optimized oil-entrapped composite beads loaded withrisperidone were placed in a 1.0% (w/w) aqueous dispersion ofsodium alginate and subsequently introduced into CaCl2 solution(5%, w/v) to deposit a polymeric coating onto the core beads. The

magnesium stearate beads containing risperidone.

F-5 F-6 F-7 F-8 F-9 F-O F-O (coated)

300 300 300 300 300 300 3005.5:1 5.5:1 1:1 1:1 1:1 2.149:1 2.149:1

10 10 5 5 5 14.76 14.7650 50 50 50 50 50 50– – – – – – 1.0

.

104 H. Bera et al. / International Journal of Biological Macromolecules 78 (2015) 102–111

Table 2Experimental plan of 32 full factorial design (coded values in bracket) with observed response values and various physical characteristics.

Code Factors with normalized levels Responses Diameter (mm)a Density (gm/cm3)a Floating lag time (min)a

Na-alginate to TGratio (by weight)(X1)

CaCl2(%, w/v) (X2)

DEE (%)a Q8h (%)a

F-1 10:1 (+1) 15 (+1) 70.74 ± 1.04 84.64 ± 0.46 1.32 ± 0.08 0.92 ± 0.05 7.25 ± 0.19F-2 10:1 (+1) 10 (0) 56.33 ± 0.56 86.25 ± 1.24 1.39 ± 0.14 0.88 ± 0.12 6.40 ± 0.45F-3 10:1 (+1) 5 (−1) 50.70 ± 2.01 89.23 ± 1.38 1.55 ± 0.13 0.75 ± 0.15 5.84 ± 0.55F-4 5.5:1 (0) 15 (+1) 74.88 ± 0.61 83.23 ± 0.39 1.42 ± 0.07 0.82 ± 0.11 6.25 ± 0.35F-5 5.5:1 (0) 10 (0) 59.53 ± 0.45 86.01 ± 0.46 1.51 ± 0.09 0.73 ± 0.06 5.68 ± 0.38F-6 5.5:1 (0) 5 (−1) 53.92 ± 0.69 87.28 ± 0.51 1.78 ± 0.13 0.66 ± 0.12 5.82 ± 0.45F-7 1:1 (−1) 15 (+1) 78.70 ± 0.72 73.17 ± 2.26 1.51 ± 0.05 0.74 ± 0.15 5.88 ± 0.86F-8 1:1 (−1) 10 (0) 63.91 ± 1.13 79.04 ± 0.54 1.63 ± 0.12 0.71 ± 0.09 5.23 ± 0.30F-9 1:1 (−1) 5 (−1) 56.62 ± 0.88 81.76 ± 1.52 1.92 ± 0.15 0.60 ± 0.10 5.12 ± 0.43

Experimental valuesF–O 2.149:1 14.76 75.19 ± 0.75 78.04 ± 0.38 1.47 ± 0.14 0.77 ± 0.06 5.85 ± 0.43

Predicted values76.91 77.17

% error# +2.24 −1.12F–O (coated) 2.149:1 14.76 73.53 ± 1.04 75.43 ± 1.34 1.70 ± 0.11 0.54 ± 0.08 4.05 ± 0.25

b

bw

2

cl2slt

D

2

bJu

�

w(r

2

d6da

2

aa

a Mean ± S.D; n = 3.Error (%) = [(Predicted value − actual value)/Predicted value] × 100.

eads were left to harden in the CaCl2 solution for 10 min, washedith distilled water, and dried overnight at room temperature [14].

.5. Estimation of DEE

An accurately weighed quantity (100 mg) of the uncoated andoated dry beads was crushed and dissolved in 500 ml of simu-ated gastric fluid (pH 1.2) by stirring with magnetic stirrer for4 h. The drug content in the filtered supernatant was estimatedpectrophotometrically (Shimadzu/UV-1700, Japan) at 272 nm fol-owing suitable dilution. The DEE was then determined accordingo the following relationship [18]:

EE (%) = Actual drug contentTheoretical drug content

× 100

.6. Beads size and density measurement

The average diameters of 100 beads were measured for eachatch using digital slide callipers (CD-6 CS, Mitutoyo Corporation,apan). The densities of the beads were subsequently calculatedsing the following equations [7]:

= M

Vand V = 4

3�r3

here �, M, V and r indicate the density (g/cm3), weight (g), volumecm3) and radius (cm) of the Ca-alginate–TG–MS composite beads,espectively.

.7. Scanning electron microscopy (SEM)

The uncoated and alginate gel-coated beads containing risperi-one were examined by a scanning electron microscope (JSM360A, JOEL, Tokyo, Japan) equipped with a secondary electronetector. The samples were coated with a thin layer of gold andnalysed at an accelerating voltage of 20 kV.

.8. In vitro buoyancy testing

The in vitro floating property of the oil-entrapped Ca-lginate–TG–MS composite beads was estimated according to

previously described procedure [19]. Fifty beads were placed in

900 ml of simulated gastric fluid (pH 1.2) under stirring at a speedof 50 rpm at 37 ± 0.5 ◦C. The time required for the beads to rise tothe surface of the test medium and the total duration of floatationwere recorded to estimate the floating lag time and duration offloatation, respectively.

2.9. Fourier transform-infrared (FTIR) spectroscopy

The pure drug, sodium alginate, TG, placebo beads anddrug-loaded uncoated and coated beads were analyzed by FTIRspectroscopy (PerkinElmer, USA) to assess the drug-excipientscompatibility. The samples were prepared by KBr pellet methodand scanned over a fixed wavelength range (400–4000 cm−1).

2.10. Powder X-ray diffraction (P-XRD)

The drug crystallinity was estimated using a powder X-raydiffractometer (Bruker-AXS D8) with a CuK� radiation detector,working at 40 kV voltage and 30 mA input current. The sampleswere examined over the 2� range of 10–80◦ with a scan angularspeed of 4◦/min.

2.11. In vitro drug release study

The in vitro dissolution study was performed using USP Type-II dissolution apparatus (Electrolab dissolution tester, TDT-08) in900 ml of simulated gastric fluid (pH 1.2) at 37 ± 0.5 ◦C [18]. Thepaddles rotation was set at 50 rpm. An equivalent weight of theuncoated and coated oil-entrapped alginate–TG–MS compositebeads containing 25 mg of risperidone was introduced into the dis-solution medium. At regular time intervals, 5 ml of aliquots werewithdrawn and replaced with an equal volume of fresh mediumto maintain the sink condition. Collected samples were thereafteranalyzed by using a UV-Visible Spectrophotometer (Shimadzu/UV-1700, Japan) at 272 nm.

2.12. Kinetics modelling and mechanism of drug release

The in vitro release data were fitted into various releaseequations and kinetic models like zero-order (Q = kt + Q0), first-order (Q = Q0ekt), Higuchi (Q = kt1/2), Hixson–Crowell (Q 1/3 = kt +Q 1/3

0 ) and Korsmeyer–Peppas (Q = ktn) model [5,8]. Q denotes

iologic

tvnrntp

csp

D

ri

2

cwsa(

S

wt

m

V

wo3

2

EaM

3

3b

csaafc

H. Bera et al. / International Journal of B

he amount of drug released at time t, Q0 represents the initialalue of Q and k is the release rate constant. The diffusion expo-ent (n) of the Korsmeyer–Peppas model demonstrates the drugelease mechanism. When n is ≤0.45, the Fickian diffusion phe-omenon dominates. The n values between 0.45 and 0.89 indicatehe anomalous transport. The values of n ≥ 0.89 refer case-II trans-ort mechanism [20].

The initial diffusion coefficients (DI), average diffusion coeffi-ient (DA) and late diffusion coefficients (DL) from the swellableystems could be estimated according to the following modifiedower law expressions [21,22]:

Mt

M�= 4

(DIt

�l2

)0.5;

A = 0.049l2

t1/2; and

Mt

M�= 1 −

(8

�2

)exp

[−�2DLt

l2

]

The Mt and M� represent drug released at time ‘t’ and at equilib-ium, respectively, ‘l’ denotes the thickness of the sample and t1/2

s the time required for 50% release of the drug.

.13. Swelling index

The swelling behaviour of the uncoated and alginate gel-oated beads loaded with risperidone was examined in terms of %eight gain in simulated gastric fluid (pH 1.2) at 37 ± 0.5 ◦C. The

wollen beads were periodically withdrawn from the test mediand weighed after removal of excess surface water. Swelling index%) was then calculated as [5]:

welling Index = Wt − W0

W0× 100

here Wt refers the weight of the beads at time t and W0 denoteshe weight of the beads before immersion.

The water penetration velocity into the matrices was also deter-ined from the % swelling data using following equation [20]:

= 12�A

× dw

dt

here V represents penetration velocity, dw/dt indicates the slopef % swelling versus time curve, � denotes the density of water at10 K and A is the surface area of a bead.

.14. Statistical analysis

The statistical optimization was carried out employing Design-xpert 8.0.6.1 software (Stat-Ease Inc., USA). All the measured datare expressed as mean ± standard deviation and analyzed usingedCalc® software.

. Results and discussion

.1. Preparation of oil-entrapped Ca-alginate–TG–MS compositeeads containing risperidone

The drug carrier systems comprising a blend of ionic polysac-harides exhibited desirable functional properties like highertability, excellent drug encapsulation, improved mucoadhesivity

nd sustained drug release rate [23]. Herein, novel oil-entrappedlginate–TG–MS composite beads encapsulating risperidone wereabricated by ionotropic gelation technique utilizing CaCl2 asrosslinking agent. When the emulsions containing drug, polymers,

al Macromolecules 78 (2015) 102–111 105

olive oil and MS were introduced into a CaCl2 solution, the inter-actions between the carboxylate anions of alginate guluronateunits and the positively charged calcium ions resulted in theformation of cross-linked networks (i.e., ‘egg-box’ model) andyielded the gelled spheres, instantaneously [3]. In this process,ionotropically crosslinked alginate gels might restrict the fluidityof TG molecules to enhance their conglomeration. In addition, thehydrogen bonding and electrostatic interactions [11] between twobio-polymers paved the way for the formation of oil-entrappedCa-alginate–TG–MS composite gel beads loaded with risperidone.

3.2. Optimization

A total nine trial formulations were accomplished accord-ing to a two factors and three levels factorial design. Differentexperimental trial formulations and the observed responses aredisplayed in Table 2. The software derived model equation relat-ing DEE (%) as response became: DEE (%) = +59.66 − 3.58X1 +10.51X2 − 0.51X1X2 + 0.19 X2

1 + 4.47X22 ; [F-value = 1700.29;

R2 = 0.9993; p < 0.0001]. The model equation illustrating Q8h (%)as response was: Q8h (%) = +85.96 + 4.36X1 − 2.87X2 + 1.00X1X2 −3.21X2

1 − 0.60X22 ; [F-value = 62.28; R2 = 0.9811; p < 0.0001].

The simplified model equations after eliminating insignificantterms (p > 0.05) on the basis of ANOVA results (Table 3) were: DEE(%) = +59.73 − 3.58X1 + 10.51X2 − 0.51X1X2 + 4.53X2

2 ; [F-value =2119.20; R2 = 0.9992; p < 0.0001] and Q8h (%) = +85.76 + 4.36X1 −2.87X2 + 1.00X1X2 − 3.41X2

1 ; [F-value = 72.60; R2 = 0.9765;p < 0.0001].

Software generated three-dimensional response surface andcorresponding two-dimensional contour plots demonstrated theeffects of the independent factors on the investigated responses(Fig. 1). The response surface plots as a function of two independentfactors at a time are valuable to examine their main and interac-tion effects on responses concurrently [7]. The response surface andthe corresponding contour plots describing DEE implied that thepolymer blends with low alginate contents and greater CaCl2 con-centrations led to an improvement in drug entrapment (Fig. 1aand b, respectively). Moreover, the response surface and the cor-responding contour plots portraying Q8h indicated that the drugrelease rate increased with increasing alginate contents in the poly-mer blend and decreased with increasing CaCl2 concentrations(Fig. 1c and d, respectively). The contour plots relating DEE andQ8h varied in a nonlinear pattern, indicating a nonlinear relation-ship between independent factors and response parameters (Fig. 1band d, respectively).

A numerical optimization technique based on the criterion ofdesirability was adopted to explore the optimal composition of theindependent factors that maximize DEE and minimize Q8h. Thedesirable ranges of responses were restricted to 75 ≤ DEE ≤ 85%and 70 ≤ Q8h ≤ 80%. An optimal formulation setting having high-est desirability near to 1.0 was selected among different settingsrecommended by Design Expert 8.0.6.1 software. The selectedoptimal values of alginate to TG ratio and CaCl2 concentrationwere 2.149:1 (w/w) and 14.76% (w/v), respectively. The opti-mized oil-entrapped Ca-alginate–TG–MS composite beads loadedwith risperidone (F–O) showed DEE of 75.19 ± 0.75% and Q8h of78.04 ± 0.38% with minimum prediction errors (Table 2). In short,the experimental findings were in excellent agreement with themodel-based predictions, which conferred the predictability andvalidity of the model.

3.3. Coating on oil-entrapped alginate–TG–MS blend beads

containing risperidone

The Ca-alginate–TG–MS composite beads were expected to dis-play accelerated drug release behaviour because of their highly

106 H. Bera et al. / International Journal of Biological Macromolecules 78 (2015) 102–111

Table 3A Summary of ANOVA analysis for the response variables.

Source Sum of squares d.f.a Mean square F value p-value

(a) For drug entrapment efficiency (%) (Quadratic model)Model 802.73 5 160.55 1700.29 <0.0001 (S)X1-SA:TG 76.76 1 76.76 812.89 <0.0001 (S)X2-CaCl2 663.18 1 663.18 7023.57 <0.0001 (S)X1X2 1.04 1 1.04 11.02 0.0160 (S)X2

1 0.096 1 0.096 1.02 0.3516 (NS)X2

2 53.28 1 53.28 564.30 <0.0001 (S)Residual 0.57 6 0.094Total 803.29 11

(b) For cumulative drug release after 8 h (%) (Quadratic model)Model 203.29 5 40.66 62.28 <0.0001 (S)X1-SA:TG 113.97 1 113.97 174.57 <0.0001 (S)X2-CaCl2 49.48 1 49.48 75.79 0.0001 (S)X1X2 4.00 1 4.00 6.13 0.0481(S)X2

1 27.48 1 27.48 42.09 0.0006 (S)X2

2 0.96 1 0.96 1.47 0.2708 (NS)Residual 3.92 6 0.65Total 207.21 11

p t and

pfai

Fa

value > 0.05 was considered as statistically significant. S and NS indicate significana d.f., degrees of freedom.

orous structure [11]. To address such shortcomings, the optimizedormulation (F–O) was coated with an ionotropically crosslinkedlginate gel membrane, which could in turn reduce their permeabil-ty and offer sustained drug release. The optimized oil-entrapped

ig. 1. Response surface 3D plots (a) and contour plots (b) describing the effects of Na-algnd contour plots (d) illustrating the effects of Na-alginate to TG ratio and CaCl2 concentr

not significant, respectively.

alginate–TG–MS composite beads (F–O) were exposed to an aque-ous dispersion of sodium alginate and subsequently introducedinto the CaCl2 solution. The interactions between the negativelycharged carboxylic groups of alginate molecules and the positively

inate to TG ratio and CaCl2 concentration on DEE (%); response surface 3D plots (c)ation on Q8h (%).

iologic

cib

3

c(tritterbcemecaltpmctt

3

1uitfvlaiscblcoatta

3

rbbtaoosn

H. Bera et al. / International Journal of B

harged calcium ions could lead to the deposition of an ionotrop-cally crosslinked alginate gel-membrane around the optimizedeads [5].

.4. DEE

The DEE of the uncoated oil-entrapped alginate–TG–MSomposite beads containing risperidone exhibited high DEE78.70 ± 0.72%–50.70 ± 2.01%) (Table 2), which could be attributedo the oil and MS inclusions. The incorporation of olive oil mightesult in greater partitioning of risperidone, a hydrophobic drug,nto the oil phase [7]. The oil and MS could also create a tight barriero impede the drug diffusion to the external media during prepara-ion, leading to enhanced drug loading capacity of the beads [3]. Thentrapment efficiency was highly influenced by the polymer blendatio and CaCl2 concentrations. Increasing TG contents in polymerlend resulted in greater drug encapsulation in the polymer matri-es. The higher levels of TG in the blend could produce low viscositymulsions and obstruct the passage of drug molecules, whichight improve the drug trapping efficiency [5]. Moreover, the drug

ntrapment was progressively increased with increasing CaCl2 con-entrations. This might be due to a greater degree of cross-linking,chieved by the ionic interactions between polymers and higherevels of CaCl2. In contrast, the beads prepared using lower concen-ration of CaCl2 solution might have an increased number of largerores due to insufficient cross-linking, which in turn could aug-ent the drug leaching with decreased drug encapsulation [5]. The

oated beads exhibited lower drug encapsulation as compared tohe uncoated optimized beads (F–O) (Table 2), possibly because ofhe drug diffusion during the coating process.

.5. Bead size and density

The average diameter of the uncoated beads ranged from.32 ± 0.08 to 1.92 ± 0.15 mm (Table 2). The particle size of thencoated beads notably increased with increasing TG contents

n the formulations. The inclusion of high levels of TG could leado increased viscosity of the emulsions, which could favour theormation of the beads with a larger particle diameter. The obser-ation was in line with the results reported in the previousiterature [9]. In contrast, the particle size of the formulatedlginate–TG–MS beads decreased as the CaCl2 concentrationncreased. This could possibly be explained by a greater extent of gelhrinkage at increasing levels of crosslinking [5]. The alginate gel-oated beads displayed larger particle diameters than the uncoatedeads (F–O). The risperidone-loaded uncoated beads exhibited

ower density (0.92 ± 0.05–0.60 ± 0.10 g/cm3) (Table 2), whichould be due to the incorporation of the low-density materials likelive oil and MS into the matrices [3]. The decline in density wasssociated with decreased sodium alginate and CaCl2 concentra-ions. The density of the alginate gel-coated beads was lower thanhe uncoated optimized beads (F–O) due to their enlarged surfacerea after coating.

.6. Surface morphology

The exterior surface and cross-sectional morphologies of theisperidone-loaded uncoated and coated beads were investigatedy SEM analysis (Fig. 2). Both uncoated and alginate gel-coatedeads displayed a spherical morphology. Under scanning elec-ron microscope, the outer surface of the uncoated beads (F–O)ppeared rough and fibrous with small pores evenly distributed

ver entire matrices (Fig. 2a and b). The pores could form as a resultf the migration of water molecules during the drying process andubsequent shrinkage of the polymeric gel [9]. In comparison, algi-ate gel-coated beads portrayed rough and compact outer surface

al Macromolecules 78 (2015) 102–111 107

without pores (Fig. 2c and d). The cross-sectional view of the coatedbeads depicted a sponge-like porous structure, in which the oil wasentrapped (Fig. 2e and f). The uneven size of the pores could be dueto the partial coalescence of the oil droplets during the gelationprocess [24]. The cross-sectional morphology of the coated beadsalso exhibited that a thin-film membrane was deposited onto thesurface of the oil-entrapped alginate–TG–MS composite beads ascoating (Fig. 2e and f).

3.7. Buoyancy

The buoyancy of the uncoated and alginate gel-coated oil-entrapped alginate–TG–MS composite beads containing risperi-done is presented in Table 2. The buoyancy was significantlycorrelated with the mean density of these beads. The density ofall the formulations was less than that of simulated gastric fluid(i.e., 1.004 g/cm3) and consequently, they demonstrated floatationwith a short floating lag time (less than 8 min) (Table 2). Inclusionof oil in the uncoated gel beads resulted in the creation of numer-ous tiny pockets within the polymer matrices, which could playa pivotal role in buoyancy [7]. In addition, the incorporation of MScould decrease the density of the uncoated beads and impart floata-tion. The buoyancy can also be associated with swelling and matrixvolume expansion in aqueous media, leading to further reduce inbeads density [3]. The coated beads floated comparatively fasterthan the uncoated optimized beads (F–O) (Table 2). The superiorfloating ability of the alginate gel-coated beads might be due to theirlow densities and the presence of an air compartment in betweencoating membrane and uncoated beads [25]. All the uncoated andcoated beads remained buoyant throughout the study (>8 h).

3.8. FTIR and P-XRD analyses

The FTIR spectra of risperidone, sodium alginate, TG, drug-free and drug-loaded uncoated and coated optimized beads aredisplayed in Fig. 3. Pure risperidone exhibited various peaks char-acteristic of the N–O stretching, C O stretching, CH3 bending, CH3stretching, C–H asymmetric aromatic ring stretching vibration at1537, 1647, 1450, 2947, and 3068 cm−1, respectively. The distinctpeaks of sodium alginate were evident at 3445, 1643, 1417 and1130 cm−1, which were due to the stretching of –OH, –COO− (asym-metric), –COO− (symmetric), and C–O–C, respectively. In the FTIRspectrum of TG, the stretching vibration modes were observed at987, 1645, 2879, 3560 cm−1 for –HC O, –CH–OH, aliphatic C–H, and–OH groups, respectively [5]. The spectrum of drug-free uncoatedand coated beads exhibited all the characteristic peaks of bothalginate and TG. However, it was observed that the intensity ofthe absorption band of carboxyl groups decreased obviously anddrifted toward higher wave number, indicating the formation ofionic crosslinks between polymers and Ca+2 ions. A similar obser-vation was reported previously by Chae et al. [26]. The FTIR spectraof the uncoated and coated optimized beads (F–O) encapsulatingrisperidone depicted all the characteristic peaks of sodium alginate,TG, and risperidone without any significant shifting. Thus, FTIRanalysis ruled out the existence of any incompatibility betweenrisperidone and polymers employed to formulate uncoated andcoated oil-entrapped floating beads.

To ascertain the solid state transformation of risperidone, the P-XRD patterns of pure drug, placebo and drug-loaded uncoated andcoated beads were analysed (Fig. 4). P-XRD chart of risperidonedepicted characteristic peaks at 11.3◦, 14.2◦, 18.6◦, 19.3◦, 21.2◦,23.0◦, 28.8◦, 44.0◦, 64.4◦ and 77.5◦ (2�) that were sharp and intense,

illustrating its crystalline nature [7]. The intense peaks characteris-tics to the drug were not observed in drug-free uncoated and coatedbeads. In the P-XRD patterns of drug-loaded uncoated and coatedoptimized beads, the characteristic peaks corresponding to the drug

108 H. Bera et al. / International Journal of Biological Macromolecules 78 (2015) 102–111

Fig. 2. Scanning electron microphotographs of the uncoated and coated oil-entrapped Ca-alginate–TG–MS composite beads (F–O) depicting rough and fibrous surface of theu ads (c(

aphst

3

s(er

ncoated beads (a) with small pores (b), rough and compact surface of the coated bee) and high zoom (550×) (f).

ppeared almost at the same 2� values with attenuated diffractioneak intensities. It implied that risperidone and excipients mightave strongly interacted at molecular level, possibly drug yielded aolid solution in oil blended polymer matrices without solid stateransformation of the drug in the beads [27].

.9. Drug release

All the oil-entrapped alginate–TG–MS blend beads exhibited

ustained risperidone release over 8 h in simulated gastric fluidpH 1.2) (Fig. 5). The cumulative drug released from differ-nt batches of uncoated beads after 8 h was found within theange of 73.17 ± 2.26 to 89.23 ± 1.38% (Table 2). Compared to the

Fig. 3. The FTIR spectra of pure risperidone, alginate, TG, bla

) without pores (d); cross-sectional view of the coated beads with less zoom (95×)

developed beads, the marketed immediate release tablets (Riswel-4) demonstrated a more pronounced risperidone release in thedissolution media. After 1 h, the cumulative percentage of drugreleased from Riswel-4, optimized uncoated and alginate gel coatedbeads was approximately of 89%, 26% and 15%, respectively. More-over, the sustained drug release profile of the prepared beadswas comparable to the solid lipid nanoparticles based hydrogelsas reported by Silva et al. [28]. The incorporation of low densitymaterials (viz., MS and olive oil) into the emulsion gel beads could

modulate the drug release rate. The hydrophobic character of MSmight delay the matrix hydration, which in turn sustained the rateof risperidone release over an extended period of time [3]. Themost logical interpretation of the sustained drug release rate with

nk and drug-loaded uncoated and coated beads (F–O).

H. Bera et al. / International Journal of Biological Macromolecules 78 (2015) 102–111 109

and d

iistcpmsfrwi

Ffl

Fig. 4. The P-XRD patterns of pure risperidone, blank

nclusion of oil was that the drug remained saturated and dispersedn the oil pockets of the matrices to form a drug-oil dispersionystem [7]. The drug transportation from the uncoated beads tohe dissolution medium was expected to follow a two-step pro-ess. The drug could initially diffuse out of the oil pockets into theolymer matrices followed by transportation out of the polymeratrices into the dissolution medium [2]. Both independent factors

howed a significant influence on drug release. Increasing the massraction of alginate to the polymer blend gave rise to higher drug

elease rate. In acidic medium, the crosslinked Ca-alginate matrixas converted to alginic acid, leading to a substantial reduction

n gel strength with an accelerated rate of risperidone release. The

ig. 5. The in vitro drug release profiles of uncoated and alginate-gel coated oil-entrappeuid (pH 1.2). Results are presented as mean ± SD; SD denoted by error bars.

rug-loaded uncoated and coated formulations (F–O).

beads containing comparatively higher amounts of TG could bindmore tightly to the water molecules to form a viscous gel structuredue to its greater hydrophilicity. This might blockade the poreson the surface of beads and eventually extend the drug releaserate. Moreover, the high levels of CaCl2 resulted in a sustainedrelease profile of risperidone. The higher concentrations of cross-linker (i.e., CaCl2) could produce a stiffer network with reducedfree volume spaces of the matrices, which would restrict the drugdiffusion through the buoyant beads [5]. The alginate gel-coated

beads demonstrated slower drug release than the uncoated beads(F–O). The coating membrane might act as a diffusion barrier toretard risperidone release from the oil-entrapped alginate–TG–MS

d Ca-alginate–TG–MS composite beads containing risperidone in simulated gastric

110 H. Bera et al. / International Journal of Biological Macromolecules 78 (2015) 102–111

Table 4Results of curve fitting of the in vitro drug-release profile, gel characteristic constant and diffusion coefficients of uncoated and alginate gel-coated Ca-alginate–TG–MScomposite beads containing risperidone in simulated gastric fluid (pH 1.2).

Code Correlation coefficient (R2) Releaseexponent (n)

Gelcharacteristicconstant (k)

Diffusion coefficient (cm2/min)

Zeroorder

Firstorder

Higuchi Hixson–crowell

Korsmeyer–Peppas

Initial (DI) Average (DA) Late time (DL)

F-1 0.868 0.961 0.981 0.951 0.989 0.35 37.84 4.18 × 10−6 6.23 × 10−6 5.65 × 10−6

F-2 0.865 0.957 0.978 0.950 0.985 0.34 39.11 4.76 × 10−6 7.12 × 10−6 6.66 × 10−6

F-3 0.873 0.975 0.984 0.967 0.987 0.36 40.54 6.77 × 10−6 9.41 × 10−6 9.75 × 10−6

F-4 0.902 0.961 0.988 0.960 0.984 0.40 33.34 4.79 × 10−6 5.31 × 10−6 6.34 × 10−6

F-5 0.913 0.969 0.987 0.971 0.980 0.41 34.43 6.08 × 10−6 6.42 × 10−6 8.33 × 10−6

F-6 0.902 0.948 0.979 0.958 0.965 0.37 36.34 8.07 × 10−6 8.77 × 10−6 11.24 × 10−6

F-7 0.925 0.968 0.986 0.963 0.979 0.44 26.67 4.29 × 10−6 4.03 × 10−6 5.32 × 10−6

F-8 0.935 0.976 0.990 0.976 0.982 0.45 28.93 6.13 × 10−6 5.61 × 10−6 7.81 × 10−6

0.43 30.64 8.82 × 10−6 8.32 × 10−6 11.59 × 10−6

0.39 33.17 4.91 × 10−6 5.46 × 10−6 6.14 × 10−6

0.56 22.92 6.88 × 10−6 5.87 × 10−6 7.91 × 10−6

ctp

HmcTodtttkp(oTtc(fiidTtcw[

3

cw1utstcarctem

Fig. 6. The swelling behavior of the uncoated and alginate-gel coated oil-entrapped

F-9 0.932 0.965 0.985 0.970 0.972

F-O 0.895 0.980 0.991 0.964 0.991

F–O (coated) 0.958 0.993 0.992 0.990 0.992

omposite beads. Moreover, the alginate gel coating could blockhe surface pores of the core beads (F–O), leading to reduced waterenetration through the micropores and slower drug release rate.

The in vitro release data was fitted into zero-order, first-order,iguchi, Hixson–Crowell and Korsmeyer–Peppas models to esti-ate the release kinetics. The results established that, in most

ases, the drug release rate obeyed Higuchi kinetic model (Table 4).he calculated n values (Table 4) implied that the drug releasef all the uncoated formulations was more likely to be Fickianiffusion driven. In Fickian diffusion, the rate of water penetra-ion into the polymer matrices would be significantly slower thanhe rate of polymer chain relaxation because of their low glass-ransition temperature (Tg) [21,29]. However, the drug releaseinetics was altered from Fickian diffusion to anomalous trans-ort upon coating with alginate gel, indicating that both erosionpolymer matrix relaxation) and drug diffusion contributed to theverall drug release mechanism of alginate gel-coated beads [27].he gel characteristic constants ‘k’ calculated from the intercepts ofhe ln Mt/M� versus ln t plots were found variable for uncoated andoated beads (Table 4) [21]. The values of late diffusion coefficientDL) were higher than the average (DA) and initial diffusion coef-cient (DI) (Table 4). A pronounced acceleration in drug diffusion

n the latter stages of release could be attributed to the polymerissolution and enlargement or relaxation of polymer chains [9].he coated beads demonstrated lower drug diffusion as comparedo the uncoated optimized beads (F–O). The alginate gel-coatingould restrict the mobility of polymer chains upon contact withater and act as a diffusion barrier for rapidly permeating drugs

12].

.10. Swelling

The swelling behaviour of the uncoated and alginate gel-oated optimized oil-entrapped alginate–TG–MS beads loadedith risperidone was investigated in simulated gastric fluid (pH

.2) as a function of time (Fig. 6). The swelling profile of thencoated optimized beads (F–O) displayed that the beads swelledo a lesser extent with gradual increments up to 8 h. The poorwelling of the uncoated beads (F–O) might be due to the facthat the –COOH groups of the alginate molecules, formed by de-rosslinking through the extraction of Ca2+ ions by Cl− ions of thecid solution, remained unionized and did not induce the chargeepulsion to increase the gel porosity and swelling of the matri-

es [20,27]. This could also be caused by decreased hydration ofhe rigid polymer matrices in the presence of olive oil and MS. Thextent of swelling was further reduced for alginate gel-coated opti-ized beads. Moreover, the slopes of % swelling versus

√time plots

Ca-alginate–TG–MS composite beads encapsulating risperidone (F–O) in simulatedgastric fluid (pH 1.2). Results are presented as mean ± SD; SD denoted by error bars.

[20] conferred that the rate of swelling of the alginate gel-coatedbeads (49.95%/

√h) was lower than that of uncoated optimized

beads (F–O) (61.61%/√

h) in simulated gastric fluid. The water pen-etration velocities into the uncoated and coated matrices estimatedfrom the % swelling (weight gain) data were found 0.0045 and0.0028 cm/s, respectively. The alginate-coating could act as a bar-rier and impede the water imbibition into the matrices. Thus, thealginate gel-coated beads displayed decreased swelling relative tothe uncoated optimized beads encapsulating risperidone.

4. Conclusion

The present work described the credibility of novel algi-nate gel-coated oil-entrapped alginate–TG–MS composite buoyantbeads for intragastric risperidone delivery. The risperidone-loadedCa-alginate–TG–MS blend beads containing olive oil as coreaccomplished by ionotropic gelation method were optimized byemploying a 32 factorial design. These developed floating beadsexhibited excellent drug entrapping efficiency (DEE of 51–79%),appreciate buoyant ability (floating duration >8 h) with a min-imum buoyant lag time (<8 min), and sustained drug releaseprofile (Q8h of 73–89%) over a prolonged period of time. Theobserved responses of the optimized formulation (F–O) were inaccordance with the predicted values. The drug release profile ofthe optimized beads was notably sustained relative to the mar-keted tablets (Riswel-4). The optimized beads were further coated

with ionotropically crosslinked alginate gel membrane. The coatedbeads depicted slower drug release profile (Q8h of 75.43 ± 1.34%)endowed with improved buoyancy (floating duration >8 h, floatinglag time <5 min). This study also demonstrated that the alginate

iologic

gsrlbed

A

lSS

R

[[[

[[

[[

[

[[

[[[[[

[

[

H. Bera et al. / International Journal of B

el coating acted as a barrier which reduced the water uptake andustained the drug release rate. Overall, the beads were easily fab-icated without any cumbersome procedures requiring elevatedevels of technical expertise. The results obtained in this study coulde useful in the development of alginate-based hydrogel beads withnormous potential applicability in controlled bioactive moleculeelivery and tissue engineering.

cknowledgments

The authors gratefully acknowledge the financial support andaboratory facilities provided by Gokaraju Rangaraju Educationalociety, Hyderabad, India and Seemanta Institute of Pharmaceuticalciences, Mayurbhanj, India.

eferences

[1] C.H. Goh, P.W.S. Heng, L.W. Chan, Carbohydr. Polym. 88 (2012) 1–12.[2] R. Bera, B. Mandal, M. Bhowmik, H. Bera, S.K. Dey, G. Nandi, L.K. Ghosh, Sci.

Pharm. 77 (2009) 669–678.[3] J. Malakar, A.K. Nayak, Chem. Eng. Res. Des. 90 (2012) 1834–1846.

[4] A.K. Nayak, B. Das, R. Maji, Int. J. Biol. Macromol. 51 (2012) 1070–1078.[5] A.K. Nayak, D. Pal, Int. J. Biol. Macromol. 49 (2011) 784–793.[6] J. Malakar, P.K. Datta, S.D. Purakayastha, S. Dey, A.K. Nayak, Int. J. Biol. Macromol.

64 (2014) 181–189.[7] P.R. Guru, A.K. Nayak, R.K. Sahu, Colloids Surf. B 104 (2013) 268–275.

[[

[

al Macromolecules 78 (2015) 102–111 111

[8] A.K. Nayak, D. Pal, Int. J. Biol. Macromol. 59 (2013) 264–272.[9] A.K. Nayak, D. Pal, K. Santra, Carbohydr. Polym. 101 (2014) 220–230.10] A.K. Nayak, D. Pal, J. Malakar, Polym. Eng. Sci. 53 (2013) 238–250.11] Ji Zhang, Shengjun Xu, Zhaoli Du, Ke Ren, J. Compos. Mater. 45 (2011) 295–305.12] T. Sahasathian, N. Praphairaksit, N. Muangsin, Arch. Pharm. Res. 33 (2010)

889–899.13] B.K. Rasool, S.A. Fahmy, Acta Pharm. 63 (2013) 31–44.14] H. Bera, S.G. Kandukuri, A.K. Nayak, S. Boddupalli, Carbohydr. Polym. 120 (2015)

74–84.15] M.S. Muthu, S. Singh, Curr. Drug Deliv. 6 (2009) 62–68.16] A. Badshah, F. Subhan, K. Rauf, N.I. Bukhari, K. Shah, S. Khan, Z. Ahmed, I. Khan,

AAPS PharmSciTech 12 (2011) 525–533.17] M.J. Prieto, C.F. Temprana, N.E. del Rio Zabala, C.H. Marotta, V. Alonso Sdel, Eur.

J. Med. Chem. 46 (2011) 845–850.18] J. Malakar, A.K. Nayak, D. Pal, Int. J. Biol. Macromol. 50 (2012) 138–147.19] S.K. Jain, A.M. Awasthi, N.K. Jain, G.P. Agrawal, J. Control. Release 107 (2005)

300–309.20] R. Singh, S. Maity, B. Sa, Carbohydr. Polym. 106 (2014) 414–421.21] B. Singh, N. Sharma, Carbohydr. Polym. 74 (2008) 489–497.22] B. Singh, N. Sharma, Colloids Surf. B 82 (2011) 325–332.23] A.K. Nayak, D. Pal, S. Das, Carbohydr. Polym. 96 (2013) 349–357.24] Y. Murata, N. Sasaki, E. Miyamoto, S. Kawashima, Eur. J. Pharm. Biopharm. 50

(2000) 221–226.25] V. Iannuccelli, G. Coppi, M.T. Bernabei, R. Cameroni, Int. J. Pharm. 174 (1998)

47–54.26] T. Chae, H. Yang, V. Leung, F. Ko, T. Troczynski, J. Mater. Sci. Mater. Med. 24

(2013) 1885–1894.27] S. Mandal, S.K. Basu, B. Sa, Carbohydr. Polym. 82 (2010) 867–873.28] A.C. Silva, M.H. Amaral, E. Gonzalez-Mira, D. Santos, D. Ferreira, Colloids Surf.

B 93 (2012) 241–248.29] B. Singh, V. Sharma, D. Chauhan, Chem. Eng. Res. Des. 88 (2010) 997–1012.