Accepted Manuscript

Silica-coated calcium pectinate beads for colonic drug delivery

Ali Assifaoui, Fréderic Bouyer, Odile Chambin, Philippe Cayot

PII: S1742-7061(12)00581-8

DOI: http://dx.doi.org/10.1016/j.actbio.2012.11.031

Reference: ACTBIO 2498

To appear in: Acta Biomaterialia

Received Date: 13 June 2012

Revised Date: 15 October 2012

Accepted Date: 27 November 2012

Please cite this article as: Assifaoui, A., Bouyer, F., Chambin, O., Cayot, P., Silica-coated calcium pectinate beads

for colonic drug delivery, Acta Biomaterialia (2012), doi: http://dx.doi.org/10.1016/j.actbio.2012.11.031

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers

we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and

review of the resulting proof before it is published in its final form. Please note that during the production process

errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1  

Silica-coated calcium pectinate beads for colonic drug delivery 1 

Ali Assifaoui a,b,*, Fréderic Bouyer c, Odile Chambin a,b, Philippe Cayot b 2 

a Department of Pharmaceutical Technology, School of Pharmacy, Université de 3 

Bourgogne, 7 bd Jeanne d’Arc 21079 Dijon, France 4 

b UMR PAM Université de Bourgogne/AgroSup Dijon, PAPC team, 1 5 

Esplanade Erasme 21000 Dijon, France 6 

c Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS - 7 

Université de Bourgogne, 9 Av. Alain Savary 21078 Dijon, France 8 

* Email: ali.assifaoui@u-bourgogne.fr 9 

Tel.: + 33 380393214 Fax.: +33 380393300 10 

Keywords 11 

Pectin; silica-coating; hybrid beads, controlled release 12 

Abstract 13 

The aim of this work is to develop novel organic-inorganic hybrid beads for 14 

colonic drug delivery. For this purpose, calcium pectinate beads with 15 

theophylline are prepared by a cross-linking reaction between amidated low-16 

methoxyl pectin and calcium ions. Then beads are covered with silica starting 17 

from tetraethyoxysilane (TEOS) by a sol-gel process. The influence of TEOS 18 

concentration (0.25, 0.50, 0.75 and 1.00 M) during the process is studied in 19 

order to modulate the thickness of the silica layer around the pectinate beads and 20 

2  

thus to control the drug release. The interactions between silica coating and 21 

organic beads are weak according to the physicochemical characterizations. A 22 

good correlation between physicochemical and in-vitro dissolution tests is 23 

observed. Beyond 0.25 M of TEOS, the silica layer is thick enough to act as a 24 

barrier to water uptake and to reduce the swelling ratio of the beads. In the 25 

meantime, the drug release is delayed. Silica-coated pectinate beads are then 26 

promising candidates for sustained drug delivery systems. 27 

1. Introduction 28 

Natural biopolymers are of considerable interest as drug carriers due to their 29 

good biocompatibility, nontoxicity and controlled release properties [1]. Among 30 

these biopolymers, pectin is an important polysaccharide widely explored as the 31 

matrix for drug delivery. Due to its ability to be degraded by the colonic 32 

microflora, pectin matrix can be appropriate to efficient colon drug delivery [2-33 

7]. Pectin is composed of long sequences of partially methyl-esterified (1-4)-34 

linked �-D-galacturonic residues (known as ‘smooth regions’ or 35 

homogalacturonan), interrupted by defects of other neutral sugars such as D-36 

xylose, D-glucose, L-rhamnose, L-arabinose and D-galactose (known as non 37 

gelling ‘hairy’ region) [8]. Previous studies [4-6, 9] have used pectin with a low 38 

degree of esterification (DE<50%) to manufacture pectinate beads by ionotropic 39 

gelation using calcium or zinc ions as cross-linking agent. It was shown that 40 

drug release from calcium pectinate beads is due to solvent penetration into the 41 

3  

calcium pectinate network, followed by ion exchange between calcium and 42 

potassium and/or sodium ions from the dissolution media. It was also 43 

demonstrated that calcium pectinate beads swell in the dissolution media, and 44 

disintegrate afterwards to quickly target the colon [2, 3, 6]. The drug release 45 

mechanism for these dosage forms can be described as swelling-erosion 46 

controlled process. 47 

Amorphous silica can be found in many drug formulations and it is well 48 

tolerated when administrated both orally or topically. Silica matrices synthesised 49 

by sol–gel process are suitable materials for bioencapsulation. They are usually 50 

obtained by polymerization process of silica alkoxydes (tetraethyoxysilane, 51 

TEOS or tetramethyoxysilane, TMOS) or inorganic precursors (Na2SiO3 sodium 52 

metasilicate) [10-16]. Amorphous silica-based materials are non-toxic and 53 

biologically compatible [13, 16, 17]. They do not swell in aqueous or organic 54 

solvents preventing a too rapid release of entrapped biomolecules [13]. Silica 55 

xerogels are very porous and degrade by hydrolysis of the siloxane bonds on the 56 

surface through the matrix, when water penetrates the porous structure resulting 57 

in mass loss of the silica device [18]. The release process of drugs entrapped in 58 

silica xerogels is found to be diffusion controlled [16, 19]. 59 

During the last decade, the combination of natural biopolymers and inorganic 60 

compounds to form controlled drug release systems has incited a great interest 61 

due to a wide range of potential applications [20]. These novel hybrid dosage 62 

4  

forms can provide a delayed drug release and can maintain the drug bioactivity 63 

which may improve the therapeutic efficiency. It has been found that many 64 

biopolymers such as alginate [21, 22], chitosan [11, 20, 23], carrageenan [12], 65 

gelatine [24] can form hybrids with precursors of silica or titania prepared by the 66 

sol-gel process. To the best of our knowledge, no study using pectin to form 67 

hybrid dosage forms has been reported. The development of silica-coated 68 

alginate beads can be applied to in-vitro protein synthesis by encapsulating cell-69 

free transcriptional and translational machinery [22]. Silica coating can be a 70 

protective shield for beads making them more resistant to chemical and 71 

environmental stresses. In hybrid membranes, the use of TEOS can limit 72 

chitosan swelling by the formation of cross-linked structures between the 73 

biopolymer and the inorganic moiety [20]. The interaction between both 74 

chitosan and TEOS has occurred in the presence of DMSO in aqueous HCl 75 

solution. Previous work [12] has shown that carrageenan-TEOS association can 76 

preserve the texture of the hydrogel upon supercritical drying and prevent the 77 

collapse of the gel when the solvent is removed. 78 

Among these biopolymers, pectin can potentially form organic–inorganic hybrid 79 

dosage forms using the sol-gel reaction. The aim of this work is to prepare 80 

calcium pectinate beads containing theophylline as a drug model and to coat the 81 

beads with a silica layer formed by hydrolysis and condensation of 82 

tetraethyoxysilane (TEOS) in order to control the drug release. The interactions 83 

between the pectinate core and the silica shell are characterized by FT-IR 84 

5  

spectroscopy. Thermogravimetric analysis (TGA) is used to evaluate the 85 

inorganic amount in the hybrid beads. Sorption isotherm and swelling 86 

measurements are carried out to investigate the interactions between beads and 87 

water. In order to evaluate the impact of silica coating on the dissolution 88 

profiles, in-vitro release experiments in simulated intestinal fluid (SIF) are 89 

performed. 90 

2. Materials and methods 91 

2.1. Materials 92 

Amidated low-methoxyl pectin (Unipectine OF 305 C) is a gift from Cargill 93 

France. Before use, the pectin is purified using the alcohol-precipitation 94 

procedure [25]. Theophylline is purchased from Sigma-Aldrich (Germany), 95 

Tetraethyoxysilane (TEOS) from Fluka (France). All other chemicals are of 96 

analytical reagent grade and used as received. 97 

2.2. Preparation of calcium pectinate beads (CPGT-ref) 98 

Calcium pectinate beads are produced by ionotropic gelation method using 99 

calcium ions as a cross-linking agent. Four grams of purified pectin (pKa=3.5) 100 

are dispersed in 100 mL of acetate buffer (acetic acid/sodium acetate) at pH 5 to 101 

deprotonate the pectin carboxylic groups. Four grams of theophylline (drug 102 

model) is added to this dispersion and stirred until it is homogenized. The 103 

dispersion is added drop-wise, at an average rate of 2 mL/min, using a nozzle of 104 

0.8 mm inner diameter, into 100 mL of a gently agitated solution of CaCl2 105 

6  

(10 g/L ; pH 5). The gelled beads are formed instantaneously and are allowed to 106 

cure for 10 minutes in the cross-linking solution. Then they are filtered and 107 

washed with deionised water. Finally, the beads are dried at 37 °C in the oven 108 

for 24 hours. Beads are referred CPGT-ref. 109 

2.3. Preparation of silica-coated pectinate beads by sol-gel process (CPGT-110 

SG) 111 

Calcium pectinate beads are immersed in a pre-hydrolysed Tetraethyoxysilane 112 

(TEOS) solution at pH 2 for 30 minutes. After filtration, the wet beads are 113 

introduced in a (tris-hydroxymethyl)-aminomethane / HCl) solution noted Tris-114 

buffer (pH 7.6) for 30 minutes to induce silica condensation. The two steps are 115 

made under stirring at room temperature. After the condensation step, beads are 116 

washed and dried at 37 °C for 24 hours. In this study, different concentrations of 117 

TEOS (0.25, 0.5, 0.75 and 1.00 M) are used to prepare silica-coated beads. 118 

These beads are referred CPGT-SG0.25, CPGT-SG0.50, CPGT-SG0.75 and 119 

CPGT-SG1.00, respectively. 120 

2.4. Physicochemical and morphological characterizations 121 

Structural characteristics of beads with and without silica coating are 122 

investigated by FTIR spectroscopy (Bruker IFS 28) equipped with a DTGS 123 

detector and working in a spectral range of 400-4000 cm−1, with a resolution of 124 

2 cm−1. A total of 32 scans are collected to obtain a high signal-to-noise ratio. 125 

7  

The measurements are done in KBr pellets, which are a mixture of 200 mg of 126 

KBr dried at 120 °C and 3 mg of studied sample. 127 

The organic/inorganic ratios in the beads are measured by thermal gravimetric 128 

analysis (TGA) (SDT 2960, TA instruments, Inc.) with a heating rate of 129 

10°C/min from room temperature to 800 °C under O2 flow. 130 

The surface and cross section morphologies of the beads are observed by 131 

scanning electron microscopy using an Hitachi SU-1510 with an accelerating 132 

voltage of 20 KeV. To evaluate the thickness of the silica layer, the beads are 133 

immersed in an epoxy resin (ESCIL, France) and polished to evidence the 134 

organic core and the inorganic shell. The surface and the cross sections are 135 

observed using the electron backscattering mode (EBS). The elemental analysis 136 

of the cross sections is done by energy dispersive X-ray spectroscopy (EDX). 137 

2.5. Water vapour sorption isotherm 138 

For all the bead formulations, water vapour sorption isotherms (25 °C) are 139 

realized as a function of relative humidity (RH) from 0 % to 85 %, using a 140 

controlled atmosphere microbalance Autosorp (Biosystems SA, Couternon, 141 

France). The samples are allowed to equilibrate until there is no noticeable 142 

weight change (±0.5 mg). The time to reach the equilibrium state is determined 143 

for each bead formulation. It is well-known that the equilibrium time depends on 144 

the temperature and relative humidity. It takes a longer time to equilibrate 145 

sample at higher relative humidity and low temperature than sample at lower 146 

8  

relative humidity and higher temperature. The water content in the beads at the 147 

equilibrium (g/100g of dried basis (d.b.)), X, is studied as a function of the water 148 

activity, aw (=RH/100). For each formulation, two replicates are performed. 149 

The sorption behaviour of the beads is analyzed using the Guggenheim, 150 

Anderson and de Boer (GAB) model. This model is used instead of the BET 151 

model as the range of water activity is between 0.1 and 0.8. Indeed the BET 152 

model is limited to water activities values lower than 0.4 [26, 27]. Three 153 

constants are usually obtained from the GAB equation (1): Xm [g H2O /100g dry 154 

solid], the monolayer moisture content, c Guggenheim constant related to heat 155 

of sorption and k the constant related to heat of sorption of the multilayer. To 156 

obtain these three characteristic constants, the GAB equation is linearised by the 157 

following equation: 158 

(1) 159 

Once Xm is known, the solid surface area of the sample, S (m2/g), could be 160 

determined according to equation (2). 161 

(2) 162 

where NA is Avogadro’s number (6.023 x 1023), Mw is the molecular weight of 163 

water (18 g/mol) and Sw is the area of a molecule of water (~0.125 x 10-164 

20 m2/molecule) [28]. The heat capacity, Qs (KJ/mol), at the monolayer can be 165 

calculated from the following equation (3): 166 

9  

(3) 167 

where R is the gas constant (8.314 J/mol.K), T is the temperature (K) and c and 168 

k are the energy constants obtained from the GAB equation. 169 

2.6. Swelling measurements 170 

An exact amount of pectinate beads (100 mg) with or without silica coating is 171 

placed in glass test tubes containing 10 mL of simulated intestinal fluid (SIF, 172 

pH 7.4). The SIF is prepared by mixing adequate amounts of 50 mM of KH2PO4 173 

and 30 mM of NaOH. Beads are then allowed to swell for a certain period of 174 

time at room temperature. The beads are periodically removed and drained on a 175 

filter paper in order to remove the excess of water. Then the change in weight is 176 

measured until mass equilibrium is achieved. The swelling ratio (% SR) is 177 

calculated using the following equation (4): 178 

(4) 179 

where, wt is the weight of the beads at a specific time point, and w0 the initial 180 

weight of the dry beads. 181 

2.7. In-vitro drug release 182 

In-vitro theophylline release is measured at 37 ± 0.5 °C by paddle dissolution 183 

test (Sotax AT7, France) with a stirring rate of 50 rpm. It was demonstrated that 184 

pectinate beads are rather stable in simulated gastric fluid (pH = 1.2) but they 185 

are completely degraded by colonic pectinolytic enzymes at pH 7.4 [2,4,9]. In 186 

10  

this study, we focus on the drug release in simulated intestinal fluid (pH 7.4). 187 

Dissolution media samples are withdrawn at various time intervals for 14 hours. 188 

The theophylline release is assayed using a spectrophotometer (UVIKON XS 189 

SECOMAM, Serlabo UVK-LAB, France) at �=271 nm. The experiments are 190 

carried out in triplicate; therefore, only mean values with standard deviation 191 

error bars are reported. The in-vitro drug release data from beads with and 192 

without silica coating are fitted using Weibull model according to the following 193 

equation (5): 194 

(5) 195 

where Qt is the amount of drug released at time t, Q0 is the initial amount of drug 196 

into the bead. According to the Weibull model, d, characterizes the curve as 197 

either exponential (d=1) or sigmoidal (d>1). �d is the time parameter (provides 198 

information about the overall rate of the process). It represents the time interval 199 

to release 63.2 % of the drug present in the dosage form [29]. 200 

3. Results 201 

3.1. SEM observations 202 

SEM micrographs of the surface and cross sections of the beads are presented in 203 

Figure 1. Silica-coated beads exhibit a rather smooth surface compared to that of 204 

CPGT-ref. Cracks observed in different beads come from the preparation of the 205 

samples. Figure 1 shows that the thickness of the silica layer increases with the 206 

11  

initial TEOS concentration. The silica layer is about 115 ± 20 µm, 122 ± 21 µm 207 

and 157±10 µm for CPGT-SG0.5, CPGT-SG0.75 and CPGT-SG1.00, 208 

respectively. The elemental analysis obtained by EDX shows that the layer 209 

surrounding the beads is composed qualitatively by oxygen, carbon and silicon. 210 

The silicon content for both CPGT-SG0.75 and CPGT-SG1.00 is similar and 211 

about 10 atomic % (at. %) while it is about 5 (at. %) for CPGT-SG0.50. The 212 

core of the beads with and without silica coating is mainly composed by carbon 213 

and oxygen. It is noted that about 1 (at. %) of silicon is present in the core for 214 

silica-coated beads indicating a diffusion of silicon inside the core of the beads 215 

during the impregnation step. 216 

3.2. Physicochemical characterizations 217 

FTIR spectra for different beads are presented in Figure 2. In all cases, a large 218 

band observed around 3440 cm−1 corresponds to the vibration of -OH groups. 219 

The absorbance of pectin functional groups is located between 1800 – 1400 cm-220 

1. The band at about 1718 cm-1 can be assigned to C=O stretching vibration of 221 

methyl esterified carboxylic group. According to the literature [30], the methyl 222 

ester vibration band in pectin is observed at 1740 cm-1. The wavemumber shift 223 

from 1740 to 1718 cm-1 is due to the impact of calcium binding [31]. The degree 224 

of esterification of pectin carboxylate groups affects the gelling properties by 225 

eliminating negative charges that may sterically hinder the formation of chains. 226 

Bands observed at 1667, 1566 and 1444 cm-1 correspond to primary amide 227 

12  

groups, NH2 from amide function and OCH2 deformation of methyl ester, 228 

respectively. The region of 1200-1000 cm−1, which contains skeletal C-O and C-229 

C vibration bands of glycosidic bonds and pyranoid ring are considered as the 230 

‘fingerprint’ region that is specific to a polysaccharide compound [30]. When 231 

the TEOS concentration is increased, no significant changes in the wavenumbers 232 

of pectin functional groups (primary amide and methyl ester groups) are 233 

observed. The interaction between carboxylate on the alginate and 234 

aminopropyltrimethoxysilane (APTMS), which is used as an inorganic 235 

precursor, is mainly electrostatic [22]. FTIR spectra (data not shown) do not 236 

present any significant difference in wavenumbers between CPGT-ref and 237 

CPGT-SG0.25. However, for silica-coated beads prepared with 0.50 and 1.00 M 238 

of TEOS (Figure 2), two broad peaks at 1090 and 465 cm−1 characteristics of Si-239 

O-Si bands confirm the formation of an organic/inorganic composite [18, 32]. 240 

Furthermore, the intensity of these two peaks increases with the concentration of 241 

TEOS, suggesting an increase of the silica/pectinate ratio in the beads. 242 

The thermal decomposition of the various beads from TG analysis is presented 243 

in Figure 3. Below 150 °C the weight loss is due to adsorbed water in the beads. 244 

It is about 8% whatever the type of beads. Between 150 °C and 550 °C, the 245 

decomposition of organic compounds (pectin and theophylline) is responsible 246 

for the weight loss. It is very difficult to separate the decomposition of these two 247 

organic compounds. Indeed, anhydrous theophylline melts at 276 °C and 248 

decomposes up to 315 °C (Theophylline TG data not shown) [33]. The thermal 249 

13  

decomposition of pectin showed two sharp losing stages occurring at 200-250 

298 °C and 300-463 °C (data not shown) which correspond respectively to a 251 

primary and secondary decarboxylation involving the acid side group and a 252 

carbon in the ring [34]. As TEOS content increases from 0 to 1 M the organic 253 

loss decreases from 77 % to 44 %. It is important to observe that the amount of 254 

residue in the beads without silica coating (CPGT-ref) is about 15 %. This may 255 

correspond to the presence of calcium introduced as cross-linking agent during 256 

the beads formation and to the presence of free calcium chloride entrapped in 257 

the beads. 258 

The water vapour sorption isotherms at 25 °C are presented in Figure 4. The 259 

sorption isotherms are type II, which is typical of finely divided nonporous or 260 

macroporous solids and hydrophilic polymers. In all type of beads, the water 261 

content is increased with the water activity. It is also noted that the affinity for 262 

water molecules is decreased when the TEOS content is increased (Figure 4). 263 

At aw = 0.85, the water content is above 38% w/w for beads without silica 264 

coating (CPGT-ref). For silica-coated beads, two types of behaviour are 265 

observed: below 0.50 M of TEOS, the water content is above 21 % w/w; 266 

whereas, above 0.50 M, the water content is above 18 % w/w. The analysis of 267 

different isotherms using GAB Equation can provide information about the 268 

water monolayer, Xm, the solid surface area, S, and the heat capacity, Qs, at this 269 

14  

monolayer according respectively to Equations (1), (2) and (3). All these 270 

parameters are listed in Table 1. 271 

Water molecules are easily adsorbed into calcium pectinate beads (CPGT-ref) to 272 

form water vapour monolayer with a low heat capacity (3.7 KJ/mol). For silica-273 

coated beads, the heat capacity is higher (5 to 6 KJ/mol) indicating that the 274 

adsorption of water molecules is more difficult. It is noted that solid surface area 275 

and water vapour monolayer are significantly lower than for CPGT-ref. Thus, 276 

silica coating reduces the water adsorption site which induces a reduction of 277 

water vapour monolayer, and increases the corresponding heat capacity. 278 

However, the sorption isotherms for silica-coated beads at various TEOS 279 

concentration are similar. 280 

At fixed water activity, the equilibrium time required to have a constant weight 281 

is different for the different kind of beads. This equilibrium time increases with 282 

water activity as shown in the Figure 4 (inset). The difference in equilibrium 283 

time between CPGT-ref and silica-coated beads is reduced for high water 284 

activity. For each water activity, the equilibrium time is increased when the 285 

TEOS concentration is increased. We can conclude that water molecules need 286 

more time to be adsorbed to silica-coated beads than CPGT-ref. 287 

3.3. Swelling measurements 288 

The swelling behaviour of calcium pectinate beads with and without silica 289 

coating in SIF medium is different and is influenced by TEOS concentration 290 

15  

(Figure 5). Two phases are observed: a swelling phase with a maximum of 291 

400 % and 600 % for CPGT-ref and CPGT-SG0.25 respectively, followed by 292 

rapid erosion, occurring between 1h30 and 2 h of contact with SIF medium. 293 

When the TEOS concentration is higher than 0.25 M, the swelling behaviour is 294 

lower, and no erosion is observed in the studied range time. After 3 hours of 295 

contact between beads and SIF, the swelling ratio for CPGR-SG 0.50; 0.75 and 296 

1.00 is 351, 232 and 93 % respectively (Figure 5). The silica coating may act as 297 

a barrier for water uptake, which is in accordance with results from water vapour 298 

sorption isotherms. 299 

3.4. In-vitro drug release tests 300 

Drug release from beads with and without silica coating is studied in SIF (Figure 301 

6). It is clearly observed that increasing the concentration of TEOS in the 302 

soaking solution induces a delay in the theophylline release. For beads without 303 

silica coating and with low TEOS concentration (0.25 M), the drug is 304 

completely released after 3 hours of dissolution. While at the same time, beads 305 

prepared with 1.00 M of TEOS released only 50 % of the theophylline. For 306 

beads with and without silica coating, the drug release exhibited a sigmoidal 307 

shape, which can be fitted using Weibull model (5). Sigmoidal release profile is 308 

characterized by slower release at the initial stage followed by increased release 309 

at the later stage. Parameters from the profile release modelisation with Weibull 310 

model are presented in Table 2. 311 

16  

The time parameter �d, is increased from 1.11 to 3.35 hours for CPGT-ref 312 

CPGT-SG1.00, respectively. The d parameter, which characterizes the shape of 313 

the curve is higher than 1 for all types of beads indicating a sigmoidal release. 314 

Previous studies showed that the interaction between calcium ions and pectin 315 

chain is the main mechanism that controls the sigmoidal release of indomethacin 316 

from pectin matrix [35]. It is noted that the use of phosphate buffer can induce 317 

calcium chelation. Indeed, during the in-vitro dissolution test CaHPO4,2H2O 318 

complex can be formed [2]. This complex which is easily soluble in water, acts 319 

as a pumping effect of calcium ions form the beads and then increases drug 320 

release. The use of silica coat may reduce this effect protecting thus the 321 

diffusion of calcium from the bead to the dissolution medium. 322 

4. Discussion 323 

Silica-coated pectinate beads are synthesized by a two-step procedure: calcium 324 

beads are prepared by ionotropic gelation, and coated with pre-hydrolysed silica 325 

followed by condensation of TEOS. During the cross-linking step, it is 326 

considered that all carboxylate groups of the pectin are neutralized by calcium 327 

ions. The other functional groups (methyl ester and primary amide groups) also 328 

contribute to the gel formation. The presence of methyl ester groups affects the 329 

gelling properties by eliminating negative charges which may sterically hinder 330 

the formation of chain aggregates [33]. The amide group is suggested to play a 331 

significant role in this interaction by the formation of strong inter and intra 332 

17  

molecular hydrogen bonds [33]. Beads have then been impregnated for 30 333 

minutes in a pre-hydrolyzed TEOS at pH=2. This step induces a diffusion of a 334 

small amount of silica to the core of the beads as observed by EDX analysis. 335 

During the impregnation step hydroxyl from TEOS hydrolysis may interact with 336 

some free functional groups of pectin by hydrogen interaction. FT-IR 337 

spectroscopy showed that silica coating did not modify pectin functional groups 338 

(Figure 2). Two broad peaks at 1090 and 465 cm−1 are assigned to Si-O-Si bond. 339 

These two vibration bands confirmed the presence of silica in the hybrid beads. 340 

This result agrees with TGA and SEM observations. When TEOS concentration 341 

increased the amount of residue corresponding to the inorganic matter, the 342 

thickness of the layer also increased (Figures 1 and 3). 343 

For CPGT-SG0.25, the amount of inorganic matter after calcination is close to 344 

that of CPGT-ref suggesting that the amount of silica coating is negligible. 345 

Swelling behaviour (figure 5) and drug release profile (figure 6) for CPGT-346 

SG0.25 are similar to that of CPGT-ref. We can suppose that below this TEOS 347 

concentration, silica layer is not efficient to control drug release. Time 348 

parameter, �d, final loss and swelling ratio obtained from Weibull modelisation, 349 

TGA and swelling measurements, respectively, are plotted against TEOS 350 

concentration in Figure 7. An increase of TEOS content induces a quasi linear 351 

increase of the time parameter, �d. This may confirm the efficiency of silica 352 

coating to delay drug release. This delay can be attributed to the increase of the 353 

thickness and the amount of the silica layer coating the bead. The silica coating 354 

18  

method could be used to modulate drug release as a function of the therapeutic 355 

needs. 356 

Water sorption results show that silica-coated beads are less sensitive to water 357 

adsorption than uncoated beads. The silica coating is also responsible of the high 358 

stability over time of hybrid beads. Indeed beads have been stored at room 359 

temperature and under controlled relative humidity (RH =33%) for 6 months 360 

without microbial growth. This may be explained by the low water content in 361 

these aged beads. Moreover, SEM observations do not show significant changes 362 

in the morphology after aging whereas few changes in weight loss profiles 363 

between freshly prepared and aged silica-coated beads are observed by TGA. At 364 

this relative humidity, it can be concluded that silica coating maintain the 365 

stability of silica-coated beads. 366 

For high water activities, hybrid beads are also less sensitive to water than beads 367 

without silica coating. The study of time required to achieve the equilibrium 368 

showed that the silica coating slowed the water adsorption into beads. It is noted 369 

that sorption isotherms could not differentiate the various silica-coated beads 370 

whereas swelling studies are discriminant for beads with different TEOS 371 

concentration. Indeed as TEOS concentration is increased the swelling ratio is 372 

decreased (Figure 7). When TEOS concentration is higher than 0.5 M, no 373 

erosion is observed for hybrid beads. Previous studies concluded that the silica 374 

coating in silica coated alginate beads, could act as a protective layer keeping 375 

19  

the beads from the disintegration of the alginate core [22]. According to Figure 376 

7, a good correlation between these three methods is observed. Below 0.25 M of 377 

TEOS the silica layer did not cover the whole bead surface. Beyond this TEOS 378 

concentration, the features of the silica layer allowed the decrease of the 379 

swelling ratio and the delay of the drug release. 380 

5. Conclusion 381 

Novel silica-coated calcium pectinate beads containing theophylline are 382 

successfully prepared with tetraethyoxysilane (TEOS) as a silica precursor using 383 

an impregnation/condensation procedure of the pectinate beads. Due to weak 384 

interactions between the hydrolyzed TEOS species and pectin, silica not only 385 

covered the beads but also condensed inside the beads. It was demonstrated that 386 

TEOS concentration must be higher than 0.25 M to form a protective silica layer 387 

around the pectin beads against water uptake and swelling that ensures a better 388 

stability over time. Moreover this work clearly demonstrates that the TEOS 389 

concentration in the impregnation solution can modulate the release profile. The 390 

higher TEOS concentration is, the slower theophylline release is. Weibull model 391 

showed that this phenomena is quasi linear which allows to predict the beads 392 

dissolution profile from the bead formulation. It is also established that the silica 393 

layer surrounding beads act as a water barrier which limits bead swelling, and so 394 

subsequent drug release. 395 

Acknowledgments 396 

20  

The authors would like to thank M. L. Léonard for her technical support in TGA 397 

and SEM observations. Thanks are due to A E. J. Laukamp and A. Corpet for 398 

her preliminary tests on the formulation and the in-vitro release tests. 399 

400 

21  

401 

References: 402 

[1] Coviello T, Matricardi P, Marianecci C, Alhaique F. Polysaccharide 403 

hydrogels for modified release formulations. J Controlled Release 2007;119:5-404 

24. 405 

[2] Assifaoui A, Chambin O, Cayot P. Drug release from calcium and zinc 406 

pectinate beads: Impact of dissolution medium composition. Carbohydr Polym 407 

2011;85:388-93. 408 

[3] Das S, Ng K-Y. Colon-specific delivery of resveratrol: Optimization of 409 

multi-particulate calcium-pectinate carrier. Int J Pharm 2010;385:20-8. 410 

[4] Dhalleine C, Assifaoui A, Moulari B, Pellequer Y, Cayot P, Lamprecht A, et 411 

al. Zinc-pectinate beads as an in vivo self-assembling system for pulsatile drug 412 

delivery. Int J Pharm 2011;414:28-34. 413 

[5] El-Gibaly I. Oral delayed-release system based on Zn-pectinate gel (ZPG) 414 

microparticles as an alternative carrier to calcium pectinate beads for colonic 415 

drug delivery. Int J Pharm 2002;232:199-211. 416 

[6] Munjeri O, Collett JH, Fell JT. Hydrogel beads based on amidated pectins 417 

for colon-specific drug delivery: the role of chitosan in modifying drug release. J 418 

Controlled Release 1997;46:273-8. 419 

[7] Murata Y, Miyashita M, Kofuji K, Miyamoto E, Kawashima S. Drug release 420 

properties of a gel bead prepared with pectin and hydrolysate. J Controlled 421 

Release 2004;95:61-6. 422 

22  

[8] Guillotin SE, Bakx EJ, Boulenguer P, Schols HA, Voragen AGJ. 423 

Determination of the degree of substitution, degree of amidation and degree of 424 

blockiness of commercial pectins by using capillary electrophoresis. Food 425 

Hydrocolloids 2007;21:444-51. 426 

[9] Sriamornsak P, Puttipipatkhachorn S, Prakongpan S. Calcium pectinate gel 427 

coated pellets as an alternative carrier to calcium pectinate beads. Int J Pharm 428 

1997;156:189-94. 429 

[10] De Zárate DO, Bouyer F, Zschiedrich H, Kooyman PJ, Trens P, Iapichella 430 

J, et al. Micromesoporous Monolithic Al-MSU with a Widely Variable Content 431 

of Aluminum Leading to Tunable Acidity. Chem Mater 2008;20:1410-20. 432 

[11] Ayers MR, Hunt AJ. Synthesis and properties of chitosan-silica hybrid 433 

aerogels. J Non-Cryst Solids 2001;285:123-7. 434 

[12] Boissière M, Tourrette A, Devoisselle JM, Di Renzo F, Quignard F. 435 

Pillaring effects in macroporous carrageenan-silica composite microspheres. J 436 

Colloid Interface Sci 2006;294:109-16. 437 

[13] Lovino M, Cardinal MF, Zubiri DBV, Bernik DL. Electronic nose 438 

screening of ethanol release during sol–gel encapsulation: A novel non-invasive 439 

method to test silica polymerisation. Biosens Bioelectron 2005;21:857-62. 440 

[14] Musgo J, Echeverría JC, Estella J, Laguna M, Garrido JJ. Ammonia-441 

catalyzed silica xerogels: Simultaneous effects of pH, synthesis temperature, and 442 

ethanol:TEOS and water:TEOS molar ratios on textural and structural 443 

properties. Microporous Mesoporous Mater 2009;118:280-7. 444 

23  

[15] Teoli D, Parisi L, Realdon N, Guglielmi M, Rosato A, Morpurgo M. Wet 445 

sol–gel derived silica for controlled release of proteins. J Controlled Release 446 

2006;116:295-303. 447 

[16] Wu Z, Joo H, Gyu Lee T, Lee K. Controlled release of lidocaine 448 

hydrochloride from the surfactant-doped hybrid xerogels. J Controlled Release 449 

2005;104:497-505. 450 

[17] Morpurgo M, Teoli D, Pignatto M, Attrezzi M, Spadaro F, Realdon N. The 451 

effect of Na2CO3, NaF and NH4OH on the stability and release behavior of sol–452 

gel derived silica xerogels embedded with bioactive compounds. Acta 453 

Biomaterialia 2010;6:2246-53. 454 

[18] Brinker CJ, Scherer GW. Sol-gel science: The Physics and Chemistry of 455 

Sol–Gel Processing: Academic Press, San Diego; 1990. 456 

[19] Kortesuo P, Ahola M, Kangas M, Yli-Urpo A, Kiesvaara J, Marvola M. In 457 

vitro release of dexmedetomidine from silica xerogel monoliths: effect of sol-gel 458 

synthesis parameters. Int J Pharm 2001;221:107-14. 459 

[20] Uragami T, Katayama T, Miyata T, Tamura H, Shiraiwa T, Higuchi A. 460 

Dehydration of an Ethanol/Water Azeotrope by Novel Organic−Inorganic 461 

Hybrid Membranes Based on Quaternized Chitosan and Tetraethoxysilane. 462 

Biomacromolecules 2004;5:1567-74. 463 

[21] Boissière M, Allouche J, Chanéac C, Brayner R, Devoisselle J-M, Livage J, 464 

et al. Potentialities of silica/alginate nanoparticles as Hybrid Magnetic Carriers. 465 

Int J Pharm 2007;344:128-34. 466 

24  

[22] Lim SY, Kim K-O, Kim D-M, Park CB. Silica-coated alginate beads for in 467 

vitro protein synthesis via transcription/translation machinery encapsulation. J 468 

Biotech 2009;143:183-9. 469 

[23] Enescu D, Hamciuc V, Ardeleanu R, Cristea M, Ioanid A, Harabagiu V, et 470 

al. Polydimethylsiloxane modified chitosan. Part III: Preparation and 471 

characterization of hybrid membranes. Carbohydr Polym 2009;76:268-78. 472 

[24] Smitha S, Mukundan P, Krishna Pillai P, Warrier KGK. Silica-gelatin bio-473 

hybrid and transparent nano-coatings through sol-gel technique. Mater Chem 474 

Phys 2007;103:318-22. 475 

[25] Yapo BM. Pectin quantity, composition and physicochemical behaviour as 476 

influenced by the purification process. Food Res Int 2009;42:1197-202. 477 

[26] Brunauer S, Emmett PH, Teller E. Adsorption of Gases in Multimolecular 478 

Layers. J Am Chem Soc 1938;60:309-19. 479 

[27] Timmermann EO, Chirife J, Iglesias HA. Water sorption isotherms of foods 480 

and foodstuffs: BET or GAB parameters? J Food Eng 2001;48:19-31. 481 

[28] McClellan AL, Harnsberger HF. Cross-sectional areas of molecules 482 

adsorbed on solid surfaces. J Colloid Interface Sci 1967;23:577-99. 483 

[29] Costa P, Sousa Lobo JM. Modeling and comparison of dissolution profiles. 484 

Eur J Pharm Sci 2001;13:123-33. 485 

[30] Synytsya A, Copíková J, Matejka P, Machovic V. Fourier transform Raman 486 

and infrared spectroscopy of pectins. Carbohydr Polym 2003;54:97-106. 487 

25  

[31] Assifaoui A, Loupiac C, Chambin O, Cayot P. Structure of calcium and 488 

zinc pectinate films investigated by FTIR spectroscopy. Carbohydr Res 489 

2010;345:929-33. 490 

[32] Vinogradova E, Estrada M, Moreno A. Colloidal aggregation phenomena: 491 

Spatial structuring of TEOS-derived silica aerogels. J Colloid Interface Sci 492 

2006;298:209-12. 493 

[33] Bán M, Bombicz P, Madarász J. Thermal stability and structure of a new 494 

co-crystal of theophylline formed with phthalic acid. J Therm Anal Calorim 495 

2009;95:895-901. 496 

[34] Monfregola L, Bugatti V, Amodeo P, De Luca S, Vittoria V. Physical and 497 

Water Sorption Properties of Chemically Modified Pectin with an 498 

Environmentally Friendly Process. Biomacromolecules 2011;12:2311-8. 499 

[35] Wei X, Sun N, Wu B, Yin C, Wu W. Sigmoidal release of indomethacin 500 

from pectin matrix tablets: Effect of in situ crosslinking by calcium cations. Int J 501 

Pharm 2006;318:132-8. 502 

503 

504 

26  

Figure captions: 505 

Figure 1: SEM micrographs of the various beads (CPGT-ref, CPGT-SG0.50 and 506 

CPGT-SG1.00) and their cross-section using electron backscattering mode 507 

508 

Figure 2: FTIR spectra of different beads with and without silica coating 509 

510 

Figure 3: Thermogravimetric analysis of different types of beads 511 

512 

Figure 4: Isotherms of water vapour sorption at 25°C for different calcium 513 

pectinate beads. Inset: Equilibrium time versus water activity for all bead 514 

formulations 515 

516 

Figure 5: Swelling behaviour of beads with and without silica coating in SIF at 517 

ambient temperature 518 

519 

Figure 6: Theophylline release for beads with and without silica coat as function 520 

of the TEOS concentration (0, 0.25, 0.5, 0.75 and 1.00 M). The different lines 521 

serve as eyeguide 522 

523 

27  

Figure 7: Correlation between swelling ratio ( ), inorganic matter obtained by 524 

TG analysis ( ) and time parameter, �d, necessary to release 63% of the drug 525 

( ) in function of TEOS concentration 526 

527 

28  

528 

29  

529 

30  

530 

31  

531 

32  

532 

33  

533 

34  

534 

35  

Table captions: 535 

Table 1: Parameters from GAB model (range: 0.1 ≤aw≤0.8) for different types 536 

of beads. Where Xm is the monolayer moisture content, c is Guggenheim 537 

constant related to heat of sorption, k is the constant related to heat of sorption of 538 

the multilayer. S and Qs are the solid surface area and the heat capacity, 539 

respectively. 540 

541 

Table 2: Weibull parameters derived from the fraction drug released profiles. d, 542 

characterizes the sigmoidal curve, �d represents the time interval to release 543 

63.2 % of the drug present in the dosage form. 544 

36  

Table 1 545 

Xm(g/100g

d.b.)

c k Qs

(KJ/mol)

S (m2/g)

CPGT-ref 5.61 ± 0.06 4.40 ±

0.43

1.01 ±

0.00

3.7 ± 0.4 235

CPGT-

SG0.25

4.54 ± 0.07 10.65 ±

0.68

0.94 ±

0.00

5.7 ± 0.1 190

CPGT-

SG0.50 4.83 ± 0.06

9.04 ±

1.10

0.93 ±

0.01

6.3 ± 1.8 202

CPGT-

SG0.75 4.51 ± 0.20

8.14 ±

0.84

0.91 ±

0.01

5.0 ± 0.3 189

CPGT-

SG1.00 4.82 ± 0.12

15.07 ±

2.85

0.88 ±

0.01

6.4 ± 0.6 202

546 

37  

547 

Table 2 548 

�d (hour) d R2

CPGT-ref 1.11 ± 0.03 2.50 ± 0.15 0.997

CPGT-SG0.25 1.30 ± 0.06 2.18 ± 0.20 0.996

CPGT-SG0.50 2.00 ± 0.06 2.63 ± 0.07 0.999

CPGT-SG0.75 2.90 ± 0.05 2.56 ± 0.40 0.999

CPGT-SG1.00 3.35 ± 0.12 2.59 ± 0.10 0.997

549 

38  

550