Novel ionically crosslinked acrylamide-grafted poly(vinyl alcohol)/sodium alginate/sodium carboxymethyl cellulose pH-sensitive microspheres for delivery of Alzheimer's drug donepezil hydrochloride: Preparation and optimization of release conditions

Abstract

In this work, the graft copolymer, poly(vinyl alcohol)-grafted polyacrylamide (PVA-g-PAAm), was synthesized and characterized by Fourier transform infrared spectroscopy, differential scanning calorimetry, and elemental analysis. Microspheres of PVA-g-PAAm/sodium alginate (NaAlg)/sodium carboxymethyl cellulose (NaCMC) were prepared by the emulsion-crosslinking method and used for the delivery of an Alzheimer's drug, donepezil hydrochloride (DP). The release of DP increased with the increase in drug/polymer ratio (d/p) and PVA-g-PAAm/NaAlg/NaCMC ratio, while it decreased with the increase in the extent of crosslinking. The optimum DP release was obtained as 92.9% for a PVA-g-PAAm/NaAlg/NaCMC ratio of 1/2/1, d/p ratio of 1/8, and FeCl₃ concentration of 7% (w/v).

Keywords::

• crosslinking
• drug delivery systems
• donepezil hydrochloride
• graft copolymer
• microspheres

Introduction

For decades, polymeric structures have been used for pharmaceutical applications, particularly to provide controlled release of drugs. Polymer–drug systems may also be useful in protecting the drug from biological decomposition prior to its release. The development of such devices begins with the use of non-biodegradable polymers which depend on the diffusion process, and subsequently progresses to the use of biodegradable polymers in which swelling and erosion take place (Arifin et al. 2006).

Natural polymers such as methyl cellulose, sodium alginate (NaAlg), and chitosan have been the preferred polymers due to their biodegradability and biocompatibility. On the other hand, there are also some synthetic polymers that exhibit biocompatibility under the physiological conditions used in controlled release studies (Ramesh Babu et al. 2008). Poly(vinyl alcohol) (PVA) has been extensively used in controlled release studies due to its semicrystalline and hydrophilic nature (Sullad et al. 2010), biocompatibility and perfect mechanical properties to obtain stable microspheres (Álvarez et al. 2012). However, PVA is a highly hydrophilic polymer and has weak stability in water; therefore, the hydrophilicity of PVA must be controlled for its use (Şanlı et al. 2007). This problem can be overcome through grafting (Kumbar and Aminabhavi 2002, Şanlı et al. 2007), blending (Bajpai and Sharma 2005), and crosslinking (Şanlı et al. 2009) techniques. Among these, the grafting technique can be considered as a practical tool for the preparation of new NaAlg/sodium carboxymethyl cellulose (NaCMC) microspheres with PVA.

NaAlg is a natural polysaccharide obtained from marine brown algae and composed of β-D-mannuronic acid and α-L-guluronic acid residues in varying proportions. Soluble NaAlg can be crosslinked with divalent or polyvalent cations to form an insoluble Alg (Mandal et al. 2010, Chan et al. 2002). CMC, a water-soluble polysaccharide possessing both hydroxyl and carboxylate groups, is an inexpensive polysugar that exerts strong interactions with metal ions (Butun et al. 2011). Metal–polymer composite materials mainly have been utilized in the development of sustained release drug delivery systems (Alkhatib et al. 2006). Alg and CMC seem to be most commonly used in the preparation of polymer–metal composites (Mandal et al., 2010, Chan et al. 2002, Alkhatib et al. 2006, Kim et al. 2012, Taha et al. 2008, Sungur 1999, Rastogi et al. 2007, Xiao et al. 2009, Zheng et al. 2004, Şanlı et al. 2008, Bajpai and Sharma 2004, Silva et al. 2006, Hosny and al-Helw 1998). Kim et al. (2012) developed Fe³⁺-crosslinked Alg/CMC beads and investigated the release profile of albumin from the beads under simulated gastrointestinal conditions (pH 1.2, 4.5, and 7.4). They found that the blend beads were more effective at controlling the release of protein therapeutics. Sungur (1999) investigated the controlled release of erythromycin from CMC beads crosslinked with ferric ions, and applied different techniques to obtain a suitable delivery system. Sungur found minimal release in the stomach (simulated by pH 2.0) and a high release in the intestinal systems (simulated by pH 7.5), by applying a combination of wet and dry coating, drying and secondary crosslinking steps. Xiao et al. (2009) prepared Fe–CMC/PVA double network microparticles crosslinked via ionic crosslinking and freezing–thawing technology, and examined the stability of a protein in acidic and neutral media. They have found that the particles were special pH-sensitive vehicles with zero release in acidic medium and exhibited pH-responsive release behavior.

Donepezil is a piperidine-based reversible acetylcholinesterase inhibitor (AChEI) and the most prescribed pharmacological agent for the treatment of Alzheimer's disease (AD) (Choi et al. 2012). AD is a progressive, neurodegenerative disease characterized by memory loss, impaired visuospatial skills, poor judgment, language deterioration, and indifferent attitude, but preserved motor function (Liu et al. 2005). Donepezil, galantamine, and rivastigmine are the three AChEIs most widely used to treat AD (Christodoulou et al. 2006). However, the large fluctuations in plasma concentration levels after oral administration of these drugs have been associated with gastrointestinal adverse effects including diarrhea, nausea, and vomiting (Choi et al. 2012). For this reason, it is important to control delivery of the drug to improve its bioavailability. One way to overcome gastrointestinal problems is to encapsulate the drug in polymeric microspheres. The DP release profiles from microparticles (Zhang et al. 2007), tablets (Yan et al. 2010), and smectite clays (Park et al. 2008) have been reported by researchers. Zhang et al. (2007) prepared poly(D,L-lactide-co-glycolide) microparticles by an oil–water emulsion solvent evaporation technique and investigated the controlled release of donepezil from the microparticles. Donepezil-loaded microparticles released the donepezil within 28 days in water, but showed a slower release in phosphate buffer solution (pH 7.4). Yan et al. (2010) prepared non-bitter orally disintegrating tablets of DP for enhanced patient compliance, by preparing microspheres with different ratios of drug and Eudragit^® EPO using a spray drying method, and reported that the EPO-based drug-loaded microspheres neither decreased the bioavailability nor delayed the release of DP.

In this study, we report the development of novel, biologically compatible and pH-sensitive crosslinked microspheres of NaAlg/NaCMC and poly(vinyl alcohol)-grafted polyacrylamide (PVA-g-PAAm) blends for oral administration. For this purpose, AAm was grafted onto PVA and then blended with NaAlg and NaCMC to prepare PVA-g-PAAm/NaAlg/NaCMC microspheres which were crosslinked using different crosslinking agents. When exposed to the highly acidic (pH 1.2) environment of the stomach, the NaAlg microspheres may transform to the insoluble alginic acid form causing a reduction in their degree of crosslinking, and then degrade after arriving in the colon. PVA-g-AAm was used to increase the strength of the NaAlg/NaCMC microspheres so that after passing through the stomach, they can remain for a longer time in the intestinal pH conditions. The particle size, entrapment efficiency, microsphere yield, equilibrium swelling degree (ESD) of the microspheres, and DP release rates were studied at pH of 1.2, 6.8, and 7.4. The effects of the type of crosslinker, PVA-g-PAAm/NaAlg/NaCMC (w/w/w) ratio, crosslinker concentration, time of exposure to the crosslinker, and drug/polymer (d/p) ratio on the release of DP were investigated and discussed.

Experimental

Materials

NaAlg (medium viscosity) and NaCMC (medium viscosity) were purchased from Sigma Chemical Co. (USA). PVA and AAm were supplied by Merck (Darmstadt, Germany). The molecular weight and degree of saponification of PVA were 72,000 and greater than 98%, respectively. Light liquid paraffin was supplied by Aklar Chemistry Co. (Turkey). Glutaraldehyde (25% w/w), dimethyl sulfoxide (DMSO), FeCl₃, CaCl₂, AlCl₃, ZnCl₂, HCl, Span 85 (a commercial surfactant), n-hexane, acetone, and NaH₂PO₄ were all purchased from Merck (Darmstadt, Germany) and used as received. A gift sample of DP was provided by Sanovel Co. (Turkey).

Synthesis of the graft copolymer of PVA with AAm

The graft copolymer of PVA with AAm was synthesized at 40°C. The aqueous solution of PVA was degassed under a slow stream of nitrogen gas for 15 min. After that, AAm solution was added to this PVA solution and mixed for 15 min under a slow stream of nitrogen gas. At the end of this period, ceric ammonium nitrate was added slowly to the reaction mixture at the required concentration. A continuous supply of nitrogen was maintained throughout the reaction period. The grafting reaction was carried out for 3 h at a temperature of 40°C. At the end of 3 h, the reaction solution was added into an excess amount of acetone to precipitate the polymer, which was then filtered through suction and dried in a vacuum oven (Medcenter, Einrichtungen GmbH, Germany) at 60°C. The polymer was dissolved in DMSO and filtered to remove the homopolymer of AAm. The dissolved graft copolymer was reprecipitated in an excess amount of acetone and vacuum-dried at 40°C for 48 h.

The percentage of AAm in the graft copolymer was calculated from the elemental analysis (LECO CHNS-932 (USA) C, H, N analyzer) of PVA-g-PAAm as 34.32% (w/w) using Eq. (1), where A% is the mass% of AAM in the copolymer, M_m is the molar mass of the monomer, M_e is the atomic weight of the element for which the elemental analysis was performed, and a is the mass % of an element (N%) in the copolymer other than carbon and hydrogen.

(1) $$A\%=\frac{a}{M_e}\times M_m$$(1)

On the other hand, copolymerization efficiency was calculated according to the following formula (Eq. 2), where N and N₀ are the nitrogen contents (%) found from the elemental analysis of the graft copolymer and the theoretical nitrogen content (%) of the graft copolymer, respectively.

(2) $$\text{Grafting}\text{efficiency}(\%)=\frac{N}{N_0}\times 100$$(2)

Viscometric measurements

The viscosity measurements of PVA and PVA-g-PAAm solutions were determined using an Ubbelohde viscometer in water thermostated at 25°C. The intrinsic viscosity was determined by extrapolating the linear portion of the reduced viscosity versus concentration plot. The viscosity average molar mass, was calculated using the Mark–Houwink–Sakurada (MHS) relation:

(3) $$[\text{η}]=k{\left(\overline{M_{\text{v}}}\right)}^{\text{α}}$$(3)

The values of the MHS parameters, k and α, for PVA, were taken from the literature, assuming the constants did not change considerably with grafting (k = 1.4 × 10^{− 3} dL/g, α = 0.6] (Brandrup and Immergut 1975).

Preparation of the PVA-g-PAAm/NaAlg/NaCMC microspheres

A water-in-oil emulsion technique was utilized for the preparation of the microspheres followed by crosslinking with crosslinker. NaAlg, NaCMC (2% (w/v) of each in water), PVA-g-PAAm (8% (w/v) in water), and DP in various d/p ratios were mixed and stirred for 12 h to form a homogenous solution. The aqueous phase was emulsified in light liquid paraffin in a ratio of 1:5 (v/v) containing 2% (w/v) Span 85 by magnetic stirring (Nuve, Turkey) at 400 rpm for 15–60 min. Then, crosslinker solution was added slowly to this emulsion and stirred to assure efficient crosslinking. Microspheres were collected by filtration, washed with n-hexane, and then dried completely in an oven (Memmert, Germany) at 40°C. The non-loaded microspheres were prepared in a similar way without DP. Different variables, such as crosslinker type, PVA-g-PAAm/NaAlg/NaCMC ratio, d/p ratio, concentration of crosslinking agent, and time required for crosslinking were considered in the optimization of the formulation of microspheres, which are listed in Table II.

Swelling of the microspheres

ESDs of the crosslinked empty microspheres were determined gravimetrically by measuring the extent of their swelling in water and in solutions of pH 1.2, 6.8, and 7.4 at 37°C. To ensure complete equilibration, the samples were allowed to swell for 24 h. The excess surface-adhering liquid drops were removed by blotting. The swollen microspheres were weighed in an electronic balance (Sartorius, Korea), and then dried in an oven at 40°C until there was no change in the dried mass of the samples. The percent ESD was calculated as follows:

(4) $$\text{Equilibrium}\text{Swelling}\text{Degree}(\%)=\frac{(M_{\text{s}}-M_{\text{d}})}{M_{\text{d}}}\times 100$$(4)

where M_s and M_d are the masses of the swollen and dry microspheres, respectively.

Determination of DP content of the microspheres

A known mass of microspheres was crushed in an agate mortar and the polymeric powder was refluxed with 250 ml of distilled water for 4 h to ensure the complete extraction of DP from the microspheres. At the end of 4 h, precipitated polymer was filtered and the DP was analyzed using a TU-1880 Double Beam UV-VIS spectrophotometer at a wavelength of 270 nm, with distilled water as a blank. Actual DP loading was determined using a calibration curve. The entrapment efficiency (%) was then calculated as follows:

(5) $$\text{Entrapment}\text{Efficiency}(\%)=\left(\frac{\text{Actual}\text{DP}\text{loading}}{\text{Theoretical}\text{DP}\text{loading}}\right)\times 100$$(5)

Fourier transform infrared spectroscopy

Fourier transform infrared (FTIR) spectra of DP, microspheres, and polymer samples were recorded within the wavelength of 400–4000 cm^{− 1} using KBr disks and a Perkin-Elmer BX-II (Germany) spectrometer.

Differential scanning calorimetry

Thermal analysis was performed using a differential scanning calorimeter (DSC; Shimadzu, Japan). Sample weights ranged from 5 to 8 mg. They were heated from 30°C to 300°C at a heating rate of 10°C/min.

Scanning electron microscopy studies

Scanning electron microscopy (SEM) photographs were taken using a LEO 1430 VP (Germany) scanning electron microscope to examine the morphology and surface structure of the microspheres at the required magnification at room temperature. The microspheres were deposited on a brass holder and sputtered with a thin coating of gold under vacuum.

X-ray diffractometry

X-ray diffractometry (XRD) patterns of donepezil hydrochloride (DP), empty microsphere, and DP-loaded microspheres were recorded on an XRD-6000 model (Shimadzu) diffractometer equipped with Cu Kα radiation, at a generator voltage of 40 kV and a generator current of 40 mA. Dried microspheres were triturated to get fine powder before taking the scan. The samples were scanned using diffraction angle (2θ) ranging from 0^o to 60^o at the speed of 2 C^o/min.

Optical microscopy

Particle size of the microspheres was measured by optical microscopy (Olympus CH20BIMF200, Japan).

In vitro drug release

In vitro drug release from the microspheres was studied in 250 ml of HCl (pH 1.2) and phosphate buffer solutions (pH 6.8 and 7.4). The microspheres were incubated in a shaking water bath (Medline BS-21, Korea) at a speed of 100 rpm at 37°C. At 2-h intervals, the pH of the DP release medium was changed to 1.2, 6.8, and 7.4, respectively. At specified time intervals, the DP content was determined using an ultraviolet spectrophotometer at 270 nm. All experiments were performed in triplicate to minimize the variational error. The average values were used for further data treatment and plotting. Equal volumes of fresh HCl or phosphate buffer solution were added to the release medium to maintain the constant volume.

Results and discussion

Characterization of the graft copolymer and the microspheres

Graft copolymerization of PVA with AAm was achieved by the ceric ion-induced solution polymerization technique. The elemental analysis results, AAm % of PVA-g-PAAm, grafting efficiency, and the results of viscosity measurements, along with the viscosity average molecular masses, are presented in Table I. As it can be seen from the table, the grafting efficiency of AAm onto PVA was up to 99.8 with 34.32% of grafting. The increase in molecular mass and the intrinsic viscosity of PVA could be taken as the proof of the grafting. Figure 1 shows the scheme of the possible reaction between PVA and AAm.

Figure 1. Possible grafting reaction of AAm onto PVA.

Table I. Elemental analysis, AAm % of PVA-g-PAAm, grafting efficiency (%), intrinsic viscosity, and viscosity average molar mass of polymers.

Polymer	Reaction time (h)	% N	% C	% H	AAm % of PVA-g-PAAm	Grafting efficiency (%)	[η]
PVA-g-PAAm	3	6.76	51.32	8.17	34.32	99.8	1.1268	69,641
PVA	–	–	–	–	–	–	0.9266	50,266

FTIR spectra of PVA (A) and grafted PVA (B) are presented in Figure 2(I). A broad band around 3334 cm^{− 1} could be attributed to the O–H stretching vibration of the hydroxyl group of PVA. Similar O–H stretching (3406 cm^{− 1}) was seen in the grafted copolymer spectra, indicating that all the hydroxyl groups of PVA were not involved in the grafting. A sharp band at 1093 cm^{− 1} corresponds to C–O–C symmetrical stretching (in the acetyl group) present on the PVA backbone, because of the unhydrolyzed acetate groups of poly(vinyl acetate). The peak was due to the N–H stretching vibration of the primary amide overlapped with the O–H stretching vibrations. The band observed at 2941 cm^{− 1} represented the presence of C–H aliphatic stretching vibrations. The specific absorption bands at 1668 cm^{− 1} corresponding to the C = O group (amide band 1) and 1614 cm^{− 1} corresponding to the NH group (amide band 2) were found in the grafted copolymer spectra. Grafting was also confirmed by appearance of strong band at 1447 cm^{− 1} corresponding to C–N stretching (Şanlı et al. 2007, Kumbar and Aminabhavi 2002, Tudorachi and Lipsa 2006).

Figure 2. (I) FTIR results of PVA (A), PVA-g-PAAm (B); (II) FTIR results of NaAlg (A), NaCMC (B), PVA (C), DP (D), empty microspheres (E), DP-loaded microspheres (F).

FTIR spectra of NaAlg, NaCMC, PVA, DP, and empty and drug-loaded crosslinked PVA-g-PAAm/NaAlg/NaCMC microspheres are shown in Figure 2(II) (A, B, C, D, E, and F). The spectrum of NaAlg (Figure 2(II) (A)) showed peaks around 3447, 2929, 1616, 1418, and 1031 cm^{− 1} and the spectrum of NaCMC (Figure 2(II) (B)) showed peaks around 3445, 2919, 1616, 1423, and 1014 cm^{− 1}, indicating the stretching of O–H, aliphatic C–H, COO⁻ (asymmetric), COO⁻ (symmetric), and C–O–C, respectively (Bajpai et al. 2007, Rokhade et al. 2006, Pourjavadi et al. 2006). The spectrum of PVA was discussed in the previous paragraph. For DP, a sharp peak corresponding to the C = O stretching, C–N–C stretching, and C–H wagging bands were observed at 1684, 1502, and 1316 cm^{− 1}, respectively. The stretching vibrations of sp³ C–H were also seen at 2923 and 2858 cm^{− 1} (Park et al. 2008). DP-loaded microspheres showed no additional signals other than all the characteristic bands of the empty microsphere due to overlapping of DP-related signals. However, the bands in ∼3430 cm^{− 1} and ∼1608 cm^{− 1} being wider, indicated the chemical stability of the drug in the crosslinked microspheres.

DSC thermograms of PVA (A) and PVA-g-PAAm (B) are presented in Figure 3(I). As it can be seen from Figure 3(I), the T_g of the PVA used in this study was found to be 102°C, whereas the T_g of PVA-g-PAAm was found to be 148°C. This supports the modification of PVA by grafting with AAm that made the PVA thermally more stable, where similar results are available in literature (Şanlı et al. 2007, Kumbar et al. 2003, Kumbar and Aminabhavi 2003). DSC thermograms of DP (A), DP-loaded microspheres (B), and empty microspheres (C) are presented in Figure 3(II). DP showed a sharp endothermic peak at 230°C due to its melting. On the contrary, in case of DP-loaded microspheres, no characteristic peak was observed at 230°C, indicating the drug in the microspheres was molecularly dispersed.

Figure 3. (I) DSC thermograms of PVA (A), PVA-g-PAAm (B); (II) DSC thermograms of DP (A), DP-loaded microspheres (B), and empty microspheres (C).

XRD diffractograms of DP (A), empty microspheres (B), and DP-loaded microspheres (C) are presented in Figure 4. Here, DP revealed characteristic intense peaks between 2θ of 10^o and 60^o due to its crystalline nature. However, these peaks were not observed in the DP-loaded microspheres, indicating the dispersion of DP in polymer matrices at the molecular level.

Figure 4. XRD diffractograms of DP (A), empty microspheres (B), DP-loaded microspheres (C).

SEM micrographs of empty and DP-loaded microspheres are shown in Figure 5. As it can be seen from the figure, empty and DP-loaded microspheres were almost spherical in shape and empty microspheres showed a smooth surface, whereas DP-loaded microspheres showed roughness in the surface.

Figure 5. SEM photographs of empty PVA-g-PAAm/NaAlg/NaCMC microsphere (C_8,0)x5000 (A), DP-loaded PVA-g-PAAm/NaAlg/NaCMC microsphere (C₈)x4500 (B).

Particle size, entrapment efficiency, and yield value evaluation of the microspheres

The results of microsphere diameter, entrapment efficiency (%), and microsphere yield (%) measurements are displayed in Table II. Particle size was measured by optical microscopy. The microspheres formed had particle sizes ranging from 42.92 μm to 123.53 μm in diameter. The size of the microspheres changed with the crosslinker type, PVA-g-PAAm/NaAlg/NaCMC (w/w/w) ratio, d/p ratio, crosslinker concentration, and the crosslinking time. Diameters of the microspheres prepared with FeCl₃ and AlCl₃ were larger than those prepared with CaCl₂ and ZnCl₂. Moreover, diameters of the microspheres prepared with GA were the largest among the all microspheres.

Table II. Preparation conditions, entrapment efficiency (%), microsphere yield (%), and microsphere size (μm) of the DP-loaded microspheres.

Code	Crosslinker type	PVA-g-PAAm /NaAlg/ NaCMC (w/w/w)	Crosslinker concentration (%) (w/v)	Crosslinking time (min)	Drug/polymer ratio (w/w)	Entrapment efficiency (%)	Microsphere yield (%)	Microspheres size (μm)
D1	AlCl3	1/2/1	7.0	30	1/4	20.27 ± 1.42	99.13 ± 1.49	97.50 ± 2.50
E1	ZnCl2	1/2/1	7.0	30	1/4	23.31 ± 1.32	81.04 ± 4.24	80.11 ± 1.83
F1	CaCl2	1/2/1	7.0	30	1/4	40.39 ± 3.13	72.89 ± 2.49	68.02 ± 0.53
G1	GA	1/2/1	7.0	30	1/4	26.08 ± 0.67	50.65 ± 1.90	123.53 ± 0.40
C1	FeCl3	1/2/1	7.0	30	1/4	84.77 ± 1.20	76.05 ± 2.81	88.33 ± 1.52
C2	FeCl3	2/2/1	7.0	30	1/4	74.39 ± 2.24	71.65 ± 1.07	88.61 ± 1.27
C3	FeCl3	3/2/1	7.0	30	1/4	75.38 ± 0.56	69.53 ± 1.02	42.92 ± 1.19
C4	FeCl3	1/2/1	3.0	30	1/4	73.72 ± 1.73	66.93 ± 2.21	58.33 ± 2.88
C5	FeCl3	1/2/1	5.0	30	1/4	82.40 ± 3.89	74.28 ± 3.99	76.33 ± 1.52
C6	FeCl3	1/2/1	10.0	30	1/4	86.46 ± 3.96	> 100	73.33 ± 3.33
C7	FeCl3	1/2/1	7.0	15	1/4	75.68 ± 0.34	67.83 ± 1.40	94.17 ± 2.20
C8	FeCl3	1/2/1	7.0	60	1/4	69.61 ± 0.88	58.21 ± 0.41	57.64 ± 3.23
C9	FeCl3	1/2/1	7.0	30	1/2	36.64 ± 3.60	50.28 ± 0.25	90.83 ± 3.53
C10	FeCl3	1/2/1	7.0	30	1/8	98.74 ± 2.31	88.82 ± 0.66	85.83 ± 3.81

In the formulations, the diameter of the microspheres decreased with increasing PVA-g-PAAm content and crosslinking time. However, as the crosslinker concentration and d/p ratio were increased, the diameter of microspheres increased. These results are in accordance with the literature (Rastogi et al. 2007, Kumbar et al. 2003). Rastogi et al. (2007) prepared NaAlg microspheres for the delivery of isoniazid. They observed an increase in the particle size with the increase in the concentration of the crosslinker; however, a concentration of crosslinker above 10% caused the formation of irregular lumps due to extensive crosslinking of the guluronic acid units of NaAlg. Kumbar et al. (2003) prepared microspheres of PAAm-g-chitosan crosslinked with GA, and reported that as the percent of drug loading increased, the diameter of the microspheres increased, which might be due to the filling of the free volume of the microspheres.

The percentage of microsphere yield and entrapment efficiency might change depending on the preparation conditions of the microspheres. The results of entrapment efficiency (%) and microsphere yield (%) are presented in Table II. As it can be seen from the table, the microspheres prepared with FeCl₃ had a higher percentage of entrapment efficiency than the microspheres prepared with other crosslinkers. Entrapment efficiency might be related to the crosslinking mechanism of ions with NaAlg and NaCMC. The Ca²⁺ and Zn²⁺ cations are divalent and their bonding with Alg and CMC occurs in a planar two-dimensional manner which is usually described by the “egg-box” model (Taha et al. 2008, Bajpai and Sharma 2004). In contrast, the trivalent cations, Fe³⁺ and Al³⁺, are expected to form a three-dimensional bonding structure (Kim et al. 2012, Hosny and al-Helw 1998). Although both ions are divalent, the percentage of entrapment efficiency of the microspheres prepared with Ca²⁺ ions was higher than that of the microspheres prepared with Zn²⁺. Similar results were stated by Taha et al. (2008). Taha et al. prepared zinc-crosslinked Alg and calcium-crosslinked Alg beads for the release of chlorpheniramine maleate as a model drug and found that calcium crosslinking accessed higher drug loading levels compared to zinc crosslinking. They believed that this difference was related to the fact that calcium is a hard electrophile capable of forming quicker electrostatic attractive interactions with algin's carboxylates compared to the softer zinc ions.

A decrease in entrapment efficiency (%) and microsphere yield (%) was observed with an increasing amount of PVA-g-PAAm in the microspheres, which is due to the formation of a loose network that allows the leaching out of more drug particles during the production stage of the microspheres (Rokhade et al. 2006). In the microspheres, as the d/p ratio increased from 1/8 to 1/2, the entrapment efficiency and microsphere yield decreased from 98.74 to 36.64% and from 88.82 to 50.28%, respectively. This phenomenon may be explained as follows: when the d/p ratio increases, the polymer traps less DP and thus entrapment efficiency decreases (Şanlı et al. 2008). With increasing time of exposure to the crosslinking agent, the entrapment efficiency (%) decreases and this decreasing trend might be attributed to the high aqueous solubility of DP resulting in loss of the drug during crosslinking, washing, and filtering processes (Rastogi et al. 2007, Şanlı et al. 2008). As can also be seen from Table II, an increase in the crosslinker concentration increased the entrapment efficiency. Similar results were stated by Rokhade et al. (2007). Rokhade et al. (2007) prepared interpenetrating polymer network microspheres of chitosan and methylcellulose for controlled release of theophylline and reported such a trend due to higher extent of crosslinking, resulting in the formation of a more rigid network, and this caused retention of more drug particles during the microsphere preparation.

In Table II, microsphere yield (%) of the C6 formulation was found to be 108%. The increase from 7% to 10% (w/v) of concentration of FeCl₃ solution caused the formation of irregular lumps due to extensive crosslinking which could prevent complete removal of the crosslinker from the microspheres.

Effect of crosslinker type on DP release

The effect of crosslinker type on the cumulative release of DP is shown in Figure 6. The DP release rate from the microspheres crosslinked with Fe³⁺ ions was faster than that of the other microspheres. This observation might be related to the crosslinking mechanism of ions with NaAlg and NaCMC. As described in the previous section, calcium and zinc cations are divalent and their bonding to Alg and CMC is expected to occur in a planar two-dimensional manner (Taha et al. 2008, Bajpai and Sharma 2004). In contrast, the trivalent cations, Fe³⁺ and Al³⁺, are expected to form a three-dimensional bonding structure (Kim et al. 2012, Hosny and al-Helw 1998). Aiedeh et al. (2007) prepared caffeine-containing chitosan diacetate (CDA) matrices ionotropically crosslinked with Al³⁺, Zn²⁺, and Ca²⁺. They have reported that the most probable explanation for this interesting behavior was based on the electrophilic softness/hardness of the three metallic cations. In another study, similar results were found by Taha et al. (2008), who investigated the effects of sodium lauryl sulfate (anionic surfactant) on ionotropically crosslinked Alg beads with zinc and calcium ions. They have stated that calcium ions are hard electrophiles and therefore, interact with carboxylates via electrostatic attractions, which tend to be readily hydrated and thus broken under aqueous conditions. On the other hand, softer zinc ions tend to form covalent-like coordinate bonds with carboxylates, which are more resistant to hydration.

Figure 6. Effect of crosslinker type on the DP release. PVA-g-PAAm/NaAlg/NaCMC: 1/2/1, d/p: 1/4, exposure time to crosslinker: 30 min, concentration of crosslinker: 7%.

A slow release was observed from the microspheres crosslinked with GA. These results were also supported by swelling measurements shown in Table III.

Table III. ESD of the empty microspheres.

Code	Water	pH 1.2	pH 6.8	pH 7.4
C1,0	24.72 ± 0.59	61.99 ± 0.75	477.22 ± 2.95	718.39 ± 5.57
E1,0	61.33 ± 0.31	45.73 ± 0.49	299.44 ± 4.56	440.61 ± 1.04
F1,0	152.33 ± 4.51	91.07 ± 0.17	eroded	eroded
G1,0	24.73 ± 2.36	19.84 ± 3.17	38.31 ± 2.87	48.91 ± 1.57
C2,0	57.29 ± 2.98	201.16 ± 1.20	712.13 ± 1.32	862.33 ± 1.52
C3.0	71.10 ± 0.70	219.53 ± 1.80	747.4 ± 1.44	896.69 ± 1.23
C4,0	29.02 ± 1.70	66.49 ± 3.24	782.67 ± 2.82	716.66 ± 7.32
C5,0	26.87 ± 0.63	65.40 ± 1.55	762.03 ± 1.65	712.03 ± 2.84
C6,0	22.55 ± 1.02	60.48 ± 1.93	352.83 ± 1.71	525.21 ± 1.20
C7,0	156.43 ± 1.69	357.40 ± 1.44	492.53 ± 2.33	856.96 ± 6.37
C8,0	35.68 ± 2.18	62.92 ± 3.60	328.02 ± 2.07	495.84 ± 0.89

FeCl₃ was preferred as the crosslinking agent in the rest of the study due to the high entrapment efficiency.

Effect of PVA-g-PAAm/NaAlg/NaCMC ratio on DP release

In vitro release of DP from crosslinked PVA-g-PAAm/NaAlg/NaCMC microspheres was studied in gastric (2 h), input intestinal (2 h), and intestinal (2 h) pH conditions at 37°C. Figure 7 presents the data of cumulative release of DP from the microspheres with different PVA-g-PAAm/NaAlg/NaCMC ratios at pH of 1.2, 6.8, and 7.4, respectively. It can be seen from the figure that an increase in PVA- g-PAAm/NaAlg/NaCMC ratio from 1/2/1 to 3/2/1 caused an increase in the DP release from the microspheres. These results could be attributed to the increase in the hydrophilic character of the microspheres due to the presence of PVA-g-PAAm. The release results were also supported by the swelling measurements given in Table III. As it can be seen from Table III, the swelling degree of the PVA-g-PAAm/NaAlg/NaCMC microspheres increased with increasing amount of PVA-g-PAAm in the microspheres. However, as Table III shows, the ESDs (%) of some microspheres at pH 1.2 (C_1,0–C_8,0) were higher than that of water. This could be attributed to the possible hydrolysis of amide groups of PVA-g-PAAm at pH 1.2, which leads to high ESD% of PVA-g-AAm/NaAlg/NaCMC microspheres at this pH (Table III) (Asman et al. 2008).

Figure 7. Effect of PVA-g-PAAm/NaAlg/NaCMC ratio on the DP release. d/p: 1/4, exposure time to FeCl₃: 30 min, concentration of FeCl₃: 7%.

In the study on the controlled release of cefadroxil from pH-sensitive IPN microgels based on chitosan, by Krishna Rao et al. (2006), AAm-g-PVA and hydrolyzed AAm-g-PVA were studied. They have reported an increase in the swelling of the blends due to incorporation of hydrophilic PVA segments along with PAAm chains into the IPN matrix. Similar results were also found in our previous study (Şanlı et al. 2008, Şanlı et al. 2009), where diltiazem hydrochloride (DT) containing PVP/NaAlg blend microspheres were prepared (Şanlı et al. 2008) to achieve a controlled drug release profile suitable for oral administration. We have reported that DT release from the microspheres increased with an increase in the PVP/NaAlg ratio. Also, in the other study (Şanlı et al. 2009), it was found that as the chitosan/PVA ratio in the microspheres decreased, drug release increased.

A PVA-g-PAAm/NaAlg/NaCMC ratio of 1/2/1 was preferred in the rest of the study due to more controlled DP release and high cumulative release at a pH of 7.4.

Effect of the concentration of FeCl₃ and crosslinking time on the DP release

The release of DP from the microspheres depends on a number of physical and chemical parameters, including those related directly to the release medium, the release conditions (pH and temperature), preparation conditions, and those resulting from the changes in the characteristics of the microspheres. One of the most effective ways of changing the release rate of microspheres is to change the crosslinking density of the matrix, by employing varying concentrations of the crosslinking agent and time of exposure to crosslinking agent. The effect of concentration of FeCl₃ on the release of DP was investigated by varying the concentration of FeCl₃ solution from 3% to 10% (w/v). The results are given in Figure 8(I), which clearly indicate that as the FeCl₃ concentration increased (3% to 10%), the release of DP from the microspheres decreased. Maximum DP release was found to be 94.07% for the microspheres prepared with FeCl₃ concentration of 3%. However, since the burst effect was high at this concentration, a FeCl₃ concentration of 7% was preferred in the rest of the study.

Figure 8. (I) Effect of FeCl₃ concentration on DP release. PVA-g-PAAm/NaAlg/NaCMC ratio: 1/2/1, d/p: 1/4, exposure time: 30 min; (II) Effect of exposure time to FeCl₃ on DP release. PVA-g-PAAm/NaAlg/NaCMC ratio: 1/2/1, d/p: 1/4, concentration of FeCl₃: 7%.

Another way of changing the crosslink density of the microspheres is to change the time of exposure to the crosslinking solution. For this purpose, exposure time to FeCl₃ was changed from 15 min to 60 min and the results are presented in Figure 8(II). Figure 8(II) indicates that increasing crosslinking time decreased the release rate, and the release of DP became more controlled.

An increase in the concentration of FeCl₃ and exposure time resulted in an increase in the crosslink density of the microspheres, which gave rise to a tighter network structure. Consequently, free volume decreased, and permeation of water molecules and diffusion of drug molecules became more complicated. Similar observations were reported in some studies in the literature (Şanlı et al. 2008, 2009, Kurkuri and Aminabhavi 2004, Agnihotri and Aminabhavi 2006). Şanlı et al. (2009) prepared chitosan/PVA blend microspheres for salicylic acid. They have reported that as the concentration of GA and crosslinking time increased, salicylic acid release decreased. Agnihotri and Aminabhavi (2006) investigated release of the capecitabine from chitosan-poly(ethylene oxide-g-AAm) hydrogel microspheres and reported that drug release at higher amounts of GA was slower than that at lower amounts of GA due to high crosslinking density. DP release results were also supported by swelling measurements. As it can be seen from Table III, as the exposure time to FeCl₃ and concentration of FeCl₃ increased, a decrease in ESD occurred because of the increase in crosslinking density. Similar results were reported in some of the studies in the literature (Şanlı et al. 2007, 2008, Kumbar and Aminabhavi 2003, Agnihotri and Aminabhavi 2004, Mundargi et al. 2008). Mundargi et al. (2008) studied controlled release of ampicillin from starch-based tableted microspheres crosslinked with epichlorohydrin (EPI). They have reported that as the amount of EPI increased from 4 to 8 ml, equilibrium water uptake decreased significantly from 586% to 380% due to the formation of a rigid network structure at a higher concentration of the crosslinking agent.

An exposure time of 30 min was preferred in the rest of the study.

Effect of d/p ratio on the DP release

Another parameter affecting DP release from microspheres could be the effect of d/p ratio. The effect of the d/p ratio on the release of DP is shown Figure 9. As it can be seen from the figure, the cumulative DP release from the microspheres with d/p ratio of 1/2 was higher than that of microspheres with 1/4 and 1/8 d/p ratio. This is obvious because the release rate becomes quite high with the high amount of drug molecules in the matrix. Ramesh Babu et al. (2006) investigated the controlled release of ibuprofen (IB) from pH-sensitive interpenetrating network microgels of NaAlg/acrylic acid. They have reported that formulations containing the highest amount of drug (75%) displayed faster and higher release rates than those formulations containing a small amount of IB. A similar finding was reported for 5-fluorouracil-loaded semi-IPN microspheres from NaAlg and N-isopropylAAm by Mallikarjuna Reddy et al. (2008).

Figure 9. Effect of d/p ratio on DP release. PVA-g-PAAm/NaAlg/NaCMC ratio: 1/2/1, concentration of FeCl₃: 7%, exposure time to FeCl₃: 30 min.

pH-responsive characteristics of microspheres

The role of pH in the degree of swelling of polymeric structures is of a great significance since it affects release characteristics of the polymeric structures. The degree of swelling in the ionically crosslinked carboxyl group containing anionic polymers decreases in an acidic medium due to the formation of hydrogen bonds between the COOH and OH groups. However, in phosphate solutions of pH 6.8 and 7.4, the carboxyl groups are partially ionized and electrostatic repulsion increases the degree of swelling (Kim et al. 2012). However, it is clear from Figures 6–9 that PVA-g-PAAm/NaAlg/NaCMC microspheres have demonstrated a higher release of DP in the acidic pH of 1.2 than in the medium of pH 6.8 and 7.4. The drug–polymer interactions are an important factor governing drug release from the polymeric structures (Şanlı et al. 2009). Interactions between –O– or –NR₃ group of DP and the –OH, –NH₂, or –COOH groups of PVA-g-PAAm/NaAlg/NaCMC blend microsphere may also affect release behavior.

Analysis of the kinetic results

Solvent sorption by a polymeric microsphere depends mechanistically on the diffusion of water molecules into the gel matrix and subsequent relaxation of macromolecular chains of the microsphere (Bajpai and Sharma 2005). Release results were analyzed by fitting the fraction release data M_t/M_∞ using the empirical equations proposed by Peppas (1985),

(6) $$k{t}^n=\left(\frac{M_{\text{t}}}{M_{\infty }}\right)$$(6)

where k is a constant, characteristic of the drug–polymer system, and n is an empirical parameter characterizing the release mechanism; Mt is the amount of DP released at time t and M_∞ is the drug released at equilibrium time. Fickian release is defined by an initial t^1/2 time dependence of the fractional release for slabs, cylinders, and spheres. Analogously, Case-II transport is defined by an initial linear time dependence of the fractional release for all geometries (Ritger and Peppas 1987). A value of n = 0.5 indicates the Fickian transport (mechanism), while n = 1 is of Case II or non-Fickian transport (swelling controlled). The intermediary values ranging from 0.5 to 1.0 are attributed to the anomalous transport (Babu et al. 2007). The least-squares estimations of the fractional release data along with the estimated correlation coefficient values, r, are presented in Table IV. According to these data, the n value ranged from 0.437 to 1.685, indicating that DP release from the microspheres has shown mainly anomalous transport (n = 0.5–1.0), in which drug release is governed by coupling of the drug diffusion and polymer relaxation processes.

Table IV. D, k, n, r values calculated from Eqs. (6) and (7).

Code	D × 10^12 (cm^2/s)	k (min^− n)	n	r	Diffusion mechanism
C1	2.67	0.1380	0.682	0.9732	Anomalous transport
D1	2.07	0.1325	0.437	0.9262	Fickian transport
E1	2.38	0.0848	0.706	0.9911	Anomalous transport
F1	1.50	0.0960	0.626	0.9792	Anomalous transport
G1	0.02	0.0013	0.950	0.9831	Anomalous transport
C2	2.87	0.1173	0.663	0.9790	Anomalous transport
C3	30.39	0.1453	0.845	0.9987	Anomalous transport
C4	3.28	0.1356	0.869	0.9992	Anomalous transport
C5	3.24	0.1349	0.729	0.9838	Anomalous transport
C6	1.72	0.1754	0.544	0.9426	Anomalous transport
C7	4.54	0.1235	0.766	0.9896	Anomalous transport
C8	0.07	0.0138	1.086	0.9790	Case II
C9	3.48	0.0329	1.005	0.9986	Case II
C10	1.64	0.0021	1.685	0.9804	Case II

The diffusion coefficient value, D, was calculated using the equations as follows (Rokhade et al. 2007):

(7) $$D={\left(\frac{r\text{θ}}{6M_{\infty }}\right)}^2\text{π}$$(7)

where θ is slope of the linear portion of the plot of M_t/M_∞ versus t^1/2, r is the radius of the dry microspheres, and M_∞ is the maximum drug release. The data reported in Table IV show the relationship between D and the crosslinker type, PVA-g-PAAm/NaAlg/NaCMC ratio, exposure time, concentration of FeCl₃, and d/p ratio. It was noticed that the diffusion coefficient values showed an increase from 2.67 × 10^{− 12} to 30.39 × 10^{− 12} cm²/s with increasing PVA-g-PAAm/NaAlg/NaCMC ratio, which was also in agreement with the release results. The D value decreased from 2.67 × 10^{− 12} to 1.72 × 10^{− 12} cm²/s with increasing FeCl₃ concentration. An increase in exposure time to FeCl₃ decreased the diffusion coefficient, which was also in agreement with the release results. The D values also increased with increasing d/p ratio. Similar results were reported by many other workers (Rokhade et al. 2007, Ramesh Babu et al. 2006, Kulkarni et al. 1999). Rokhade et al. (2007) prepared IPN microspheres of chitosan and methylcellulose for controlled release of theophylline. They have observed that diffusion coefficient values were higher in the case of microspheres crosslinked with 5 ml of GA than those crosslinked with 10 ml of GA.

Conclusions

PVA-g-PAAm/NaAlg/NaCMC microspheres were successfully prepared by the emulsification crosslinking method and used for the delivery of DP. The microspheres formed were characterized using SEM, DSC, XRD, and FTIR techniques. DSC, XRD, and FTIR studies indicated that DP was molecularly dispersed in the microspheres. Microsphere size, entrapment efficiency, microsphere yield, and ESD of the microspheres were determined and it was found that these parameters were affected by the various processing parameters, such as PVA-g-PAAm/NaAlg/NaCMC ratio, type of crosslinker, crosslinker concentration, time of exposure to the crosslinker, and d/p ratio. The release of DP increased with the increase in d/p and PVA-g-PAAm/NaAlg/NaCMC ratio while it decreased with the increase in the extent of crosslinking. The optimum DP release was obtained as 92.9% for PVA-g-PAAm/NaAlg/NaCMC ratio of 1/2/1, d/p ratio of 1/8, and crosslinker concentration of 7% (w/v) at the end of 6 h. Diffusion constants and ESDs of all the formulations were found to be consistent with the release results. The n values calculated for the release of drug from microspheres indicated that DP release mainly displayed anomalous transport.

Declaration of interest

The authors report no declarations of interest. The authors alone are responsible for the content and writing of the article.

The authors are grateful to the Gazi University Scientific Research Foundation for their support and to Sanovel Company for the supply of the drug (Donepezil hydrochloride).