Nanoparticulate Assembly of Mannuronic Acid-and Guluronic Acid-Rich Alginate: Oral Insulin Carrier and Glucose Binder

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Nanoparticulate Assembly of Mannuronic Acid- and Guluronic Acid-Rich Alginate: Oral Insulin Carrier and Glucose Binder AMINAH KADIR,1,2 MOHAMMAD TARMIZI MOHD MOKHTAR,1,2 TIN WUI WONG1,2 1 2

Non-Destructive Biomedical and Pharmaceutical Research Centre, Universiti Teknologi MARA, Puncak Alam, Selangor 42300, Malaysia Particle Design Research Group, Faculty of Pharmacy, Universiti Teknologi MARA, Puncak Alam, Selangor 42300, Malaysia

Received 22 July 2013; revised 31 August 2013; accepted 12 September 2013 Published online 8 October 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23742 ABSTRACT: The relationship of high and low molecular weight mannuronic acid (M)- and guluronic acid (G)-rich alginate nanoparticles as oral insulin carrier was elucidated. Nanoparticles were prepared through ionotropic gelation using Ca2+ , and then in vitro physicochemical attributes and in vivo antidiabetic characteristics were examined. The alginate nanoparticles had insulin release retarded when the matrices had high alginate-to-insulin ratio or strong alginate–insulin interaction via O–H moiety. High molecular weight M-rich alginate nanoparticles were characterized by assemblies of long polymer chains that enabled insulin encapsulation with weaker polymer–drug interaction than nanoparticles prepared from other alginate grades. They were able to encapsulate and yet release and have insulin absorbed into systemic circulation, thereby lowering rat blood glucose. High molecular weight G- and low molecular weight M-rich alginate nanoparticles showed remarkable polymer–insulin interaction. This retarded the drug release and negated its absorption. Blood glucose lowering was, however, demonstrated in vivo with insulin-free matrices of these nanoparticles because of the strong alginate-glucose binding that led to intestinal glucose retention. Alginate nanoparticles can be used as oral insulin carrier or glucose binder in the treatment of diabetes as a function of its chemical composition. High molecular weight M-rich alginate nanoparticles are a suitable vehicle for future development into oral C 2013 Wiley Periodicals, Inc. and the American Pharmacists Association J Pharm Sci 102:4353–4363, 2013 insulin carrier.  Keywords: alginate; diabetes mellitus; dissolution; insulin; intestinal absorption; nanoparticles; oral drug delivery; polymeric drug carrier; polymeric drug delivery system; protein delivery

INTRODUCTION Alginate, a water-soluble polysaccharide, is commonly isolated from brown algae, such as Laminaria hyperborea, Ascophyllum nodosum, and Macrocystis pyrifera.1,2 Its chain is made of homopolymeric regions of $-D-mannuronic acid (M) blocks and "-L-guluronic acid (G) blocks, interdispersed with alternating "-L-guluronic and $-D-mannuronic acid blocks (Fig. 1). The viscosity, binding affinity for cations, gelation, aqueous solubility, mechanical strength, swelling capacity, and bioadhesiveness of alginate are dictated by its uronic acid composition and molecular weight.2–8 The viscosity of alginate increases with its molecular weight and is a function of polymer conformation. The mannuronate unit is reported to have a lower affinity for calcium ion (Ca2+ ) than guluronate moiety.1,2 Alginate is a nontoxic, biodegradable, nonimmunogenic, and biocompatible polymer.5,9–13 It possesses several inherent biological effects such as anticholesterolaemic, antihypertensive, antidiabetic, antiobesity, antimicrobial, anticancer, antihepatotoxicity, wound healing, anticoagulation, and coagulation activities.2 The biological effects of alginate have been indicated to be attributed to its structural assembly and physicochemical features. The depolymerized or low molecular weight alginate has been found to be able to control the development of diabetes. Diabetes mellitus is an endocrine disease associated with the disorders of carbohydrate metabolism brought about by deficiency in insulin secretion, insulin resistance, or both.14,15 The epidemiology study indicates that hyperglycemia Correspondence to: Tin Wui Wong (Telephone: +60-3-32584691; Fax: +60-332584602; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 102, 4353–4363 (2013)  C 2013 Wiley Periodicals, Inc. and the American Pharmacists Association

is the primary cause of diabetes. The primary mode of treatment of type I diabetes is exogenous insulin administration by means of subcutaneous route. Following chronic subcutaneous insulin administration, lipoatrophy or lipohypertrophy, however, tends to surface at the injection sites. The injection mode has also brought about pain, hypoglycemia, and risk of infection at injection site. Oral administration of insulin has been the alternative goal of many researchers.14,15 Lately, polymeric nanoparticles have been decorated with pH-sensitive moiety, bioadhesive, and/or water-insoluble materials to promote insulin delivery in gastrointestinal tract.15 The overview of various strategies applied in design of oral insulin nanoparticles denotes the significance of mucosa adhesiveness of drug delivery system. A prolonged adhesion of dosage form in the intestinal tract could translate to cumulative release and absorption of insulin transcellularly, paracellularly, and/or via the Peyer’s patches. It overcomes the absorption limit of insulin. Nanoparticles, being small in physical dimension, exhibit a large specific surface area. Their nanometric feature enhances particulate mucoadhesion and facilitates oral insulin delivery. The latter can also be promoted through sustaining drug release from nanoparticles to allow timely insulin absorption without premature drug contact with the gastrointestinal milieu.16 Low molecular weight alginate has been widely used as the matrix substance of oral insulin nanoparticles.17–20 Though low molecular weight alginate inherently possesses antidiabetic activity,2 the matrix made of such polymer releases drugs at a fast rate.5 Alginate’s uronic acid composition can critically affect drug release characteristics of a matrix as a function of matrix pore, crack, lamination, and diffusion barrier formation.2,21 The construction of oral insulin nanoparticles has nonetheless

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

4353

4354

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Figure 1. Chemical structures of alginate characterized by block sequences of (a) GG, (b) MM, and (c) MG.

yet to consider from the viewpoint of uronic acid composition of alginate. On the note of sustained drug release and mucoadhesion of a dosage form are envisaged to be greater with an increase in polymer molecular weight,5,22 this study aims to evaluate high molecular weight M-rich and G-rich alginates from the perspective of their capacity as an oral insulin nanoparticulate carrier. In vitro drug release and in vivo antidiabetic characteristics of these nanoparticles are examined against those of low molecular weight alginates to identify functional oral insulin carrier and their structural influences on blood glucose lowering attribute.

MATERIALS AND METHODS Materials Four grades of M- and G-rich sodium alginates (ISP, Waterfield, Tadworth) were employed as the matrix polymer of nanoparticles (Table 1) with calcium chloride dihydrate (Fluka, Prague, Czech Republic) as cross-linking agent and bovine insulin (Sigma–Aldrich, St. Louis, Missouri) as drug. Sodium hydroxide and hydrochloric acid (Merck, Darmstadt, Germany) were chemicals used in pH adjustment of processing materials. Ultrafiltration membrane (polyethersulfone; Millipore Corporation, Billerica, Massachusetts) with nominal molecular weight limits of 100 and 10 kDa was used to harvest high and low molecular weight alginate nanoparticles respectively. Sodium hydroxide and potassium dihydrogen phosphate (Merck) were employed in preparation of phosphate buffer pH 6.8 USP for use as dissolution medium. Acetonitrile (Merck), trifluoroacetic acid (BDH, Poole, England), and ultrapure water processed at 18 M were utilized to prepare the mobile phase for highperformance liquid chromatography (HPLC) analysis. Streptozotocin (Sigma–Aldrich), citric acid monohydrate (Merck), tri-sodium citrate (Fisher Scientific, Loughborough, Leicestershire, UK), and glucose (IDS Manufacturing Sdn Bhd, Shah Alam, Selangor, Malaysia) were chemicals employed in diabetic induction of rats. Other materials used in in vivo experiments included sodium chloride, methanol, and diethyl ether (Merck).

Formulation Variables of Nanoparticles Insulin nanoparticles were prepared through ionotropic gelation of alginate solution with calcium chloride. The influences of alginate concentration, calcium chloride concentration, and pH of processing liquid on particle size, zeta potential, insulin association efficiency, and content of matrix were examined where applicable. High molecular weight G-rich alginate was used as the model polymer to streamline the working range of formulation variables. It was envisaged that formulation conditions applicable to viscous sample would be appropriate for other alginate grades (Table 1). The formulation conditions that produced particles over a size range encompassing nano- and micro-dimensions were selected.

Alginate and Calcium Chloride Concentration Twenty gram of aqueous solution containing 0.025%, 0.050%, or 0.100% (w/w) alginate had its solution pH adjusted to 4. It was then added dropwise using a microsyringe (BD Ultra-FineTM Needle, Becton, Dickinson and Company, Franklin Lakes, New Jersey) with a flow rate of 0.194 ± 0.021 g/s into 50 g of 0.003%, 0.006%, 0.012%, 0.025%, or 0.050% (w/w) calcium chloride solution under high-speed magnetic stirring at 1000 rpm and 25◦ C, similar to the methods adopted by Racovita et al.23 The stirring was continued for an additional period of 15 min after the last addition of alginate solution. The dispersion was then subjected to particle size and zeta potential analysis. Triplicates were conducted and results averaged.

pH of Processing Liquid An accurately weighed amount of insulin (3 mg) was dissolved in 20 g 0.0050 M hydrochloric acid solution. The insulin solution was then added dropwise into 90 g of alginate solution at pH 4 under high-speed magnetic stirring at 1000 rpm and 25◦ C to provide an aqueous mixture carrying 20 mg alginate and 3 mg insulin. A diluted alginate solution (90 g instead of 20 g liquid) was used to prepare alginate–insulin aqueous mixture. The reason was the large molecular weight alginates exhibited a high-interaction propensity with insulin solution. A simple mixing of 0.100% (w/w) alginate and 0.015% (w/w) insulin in

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

DOI 10.1002/jps.23742

30.007 ± 7.399 52.082 ± 15.755 23.703 ± 2.416 17.392 ± 2.840 0.382 9.726 4.830 1.950 ± ± ± ± 20.284 36.095 14.081 9.857 68.89 ± 9.15 72.00 ± 4.68 44.94 ± 8.97 43.56 ± 1.91 1.19 0.61 1.17 0.25 ± ± ± ± 8.99 9.39 5.86 5.68

20 g liquid was accompanied by the formation of precipitates in spite of under high-speed magnetic stirring at 1000 rpm. When necessary, the pH of insulin solution, alginate solution, and aqueous alginate–insulin mixture was adjusted. The aqueous alginate–insulin mixture was processed into nanoparticles through cross-linking using 50 g of 0.012% (w/w) calcium chloride solution. The dispersion was subjected to particle size and zeta potential analysis. It was also prepared and harvested by means of ultrafiltration using stirred ultrafiltration cell (Model 8050; Millipore Corporation) under nitrogen current at 245 rpm of stirring and drying at 4◦ C for 48 h. Collection by centrifugation technique was omitted to avoid excessive particle caking. The nanoparticles were subjected to insulin association efficiency and content analysis. Triplicates were conducted and results averaged. Preparation of Nanoparticles Nanoparticles, made of alginates with different molecular weights and uronic acid compositions, were prepared with pH of insulin solution, alginate solution, and aqueous alginate– insulin mixture being kept at 2.40, 4.00, and 3.00, respectively. The aqueous alginate–insulin mixture was subsequently processed into nanoparticles using 50 g of 0.012% or 0.016% (w/w) calcium chloride solution. The particle size, zeta potential, morphology, molecular interaction, insulin dissolution, association efficiency, and content of nanoparticles were analyzed. Triplicates were conducted and results averaged. Size and Zeta Potential

− 31.83 − 32.10 − 22.50 − 23.00

± ± ± ±

0.29 0.00 0.36 0.40

4355

The particle size and zeta potential of nanoparticles were measured by means of photon correlation spectroscopy using Malvern Zetasizer Nano ZS 90 (Malvern Instruments Ltd., Malvern, Worcestershire, UK) at 25◦ C in quartz cell and zeta potential cell with a detect angle of 90◦ , respectively. Triplicates were conducted and results averaged.

0.016

0.012

169.67 ± 6.35 142.00 ± 1.73 149.67 ± 2.52 204.07 ± 9.87

High-Performance Liquid Chromatography

HG HM LG LM

Insulin Release at 24 h (%) Insulin Release at 8 h (%) Insulin Association Efficiency (%) Insulin Content (%)

LM 3.68 ± 0.02 6.899 × 105 ± 5.639 × 103 ManucolR LB

Zeta Potential (mV)

HM 5.26 ± 0.25 3.526 × 106 ± 1.825 × 105 ManucolR DMF

Particle Size (nm)

LG 3.43 ± 0.13 6.694 × 105 ± 5.492 × 103 ManugelR LBA

Calcium Chloride Concentration (%, w/w)

HG 6.92 ± 0.32 4.765 × 106 ± 1.808 × 105

63% guluronic acid, 37% mannuronic acid 63% guluronic acid, 37% mannuronic acid 40% guluronic acid, 60% mannuronic acid 40% guluronic acid, 60% mannuronic acid ManugelR DMB

DOI 10.1002/jps.23742

Alginate Grade

Label Polydispersity Index Molecular Weight (Da) Uronic Acid Content Grade

Table 1. The Grade of Alginate Used in Design of Oral Insulin Nanoparticles, and Size, Zeta Potential, Insulin Content, Association Efficiency, and Release Profiles of Nanoparticles Made of These Alginates

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

High-performance liquid chromatography analysis was conducted using Agilent 1200 Series (Agilent Technologies Inc., Waldbronn, Germany), which was equipped with UV-detector, vacuum degasser, quaternary pump, and automatic injector. The chromatographic separation was performed at 25◦ C by using Agilent Zorbax SB-C18 reversed-phase column (4.6 × 250 mm2 system packed with 5 × 10−6 m particles). The mobile phase constituted of: (a) 0.03% trifluoroacetic acid in 90% H2 O/10% acetonitrile and (b) 0.03% trifluoroacetic acid in 10% H2 O/90% acetonitrile. The chromatographic analysis was run in gradient mode for 5 min using a mobile phase that was constituted of 80:20 volume ratio of a–b, followed by isocratic run at 20:80 volume ratio of these mobile phase components for 10 min. All components of mobile phase were filtered through 0.45 :m Agilent membrane prior to use. The flow rate of mobile phase was kept at 0.5 mL/min and the volume of sample injection was 20 :L. The standard insulin solutions (0.1–0.4 mg/mL) were prepared using 0.4:1 solvent mixture of phosphate buffer pH 6.8 and 0.01 M hydrochloric acid solution. Detection wavelength of eluted analytes, stop time, and post time was 215 nm, 17.5, and 3 min, respectively. All test samples were prepared using the same solvent condition prior to HPLC assay. Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

4356

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Insulin Dissolution An accurately weighed amount of nanoparticles was placed in 10 mL of phosphate buffer pH 6.8 under sink condition at 37.0 ± 0.2◦ C and subjected to agitation at 50 strokes/min in a shaker bath (ST402; Nuve, Akyurt, Ankara, Turkey) to simulate intestinal conditions.24–26 At 8 h, 1 mL of aliquot was withdrawn and centrifuged at 122576 g for 75 min at 4◦ C. A volume of 0.4 mL supernatant was then sampled and acidified with 1 mL of 0.01 M hydrochloric acid solution. The sample was filtered through 0.45 :m polyvinylidene difluoride membrane (Durapore ; Millipore Corporation, Carrigtwohill, County Cork, Ireland) and had its insulin content analyzed by HPLC. The same experiment was repeated with fresh batches of nanoparticles and aliquots were withdrawn after 24 h of agitation. The percentage of insulin release was calculated with respect to the total drug content of nanoparticles. Triplicates were conducted for each experiment and results averaged. The insulin content and association efficiency of nanoparticles were determined by dispersing an accurately weighed 10 mg nanoparticles in 10 mL phosphate buffer pH 6.8 and magnetically stirred for 3 h at 1000 rpm and 25◦ C. The dispersion was then subjected to drug assay using the same protocol stated under the drug dissolution study. The insulin content was expressed as the percentage of drug encapsulated in a unit weight of nanoparticles. The insulin association efficiency was defined as percentage of insulin embedded in nanoparticles with reference to total insulin load processed. At least triplicates were conducted and results averaged. R

on the surfaces of a metallic stud. The droplet of nanoparticles was dried at 4◦ C for 24 h and subjected to platinum coating at a current intensity of 20 mA for 50 s by means of auto fine coater (JFC1600; Jeol) before examining under SEM at an accelerating voltage of 10 kV. Representative sections were photographed. Molecular Weight The molecular-weight characteristics of alginate were determined using gel permeation chromatography technique (1100 series; Agilent Technologies) by means of a refractive index detector.27 The weight-average molecular weight and polydispersity index were computed. A PL aquagel–OH mixed column (7.5 × 300 mm2 ; 8 :m; Agilent Technologies, Edinburgh, United Kingdom) was used with mobile phase consisted of 0.1% (w/w) sodium azide (Ajax Finechem, Seven Hills, New South Wales, Australia) dissolved in deionized water. The flow rate of mobile phase and column temperature was kept at 0.5 mL/min and 30◦ C, respectively. Dextrans (Sigma–Aldrich, Munich, Germany) with molecular weights of 50,000, 80,000, 150,000, 270,000, 410,000, 670,000, and 1400,000 Da were used as standards. One milligram per milliliter of standard or sample solution was filtered through a cellulose nitrate membrane (pore diameter = 0.45 :m; Sartorius, Germany) before analysis at an injection volume of 20 :L. The polydispersity index had a value equal to 1 for uniform chain lengths or greater than 1 for samples with heterogeneous molecular weights. At least triplicates were carried out for each batch of sample and the results averaged.

Differential Scanning Calorimetry Differential scanning calorimetry (DSC) thermograms were obtained using a differential scanning calorimeter (Pyris 6 DSC; Perkin Elmer, Norwalk). Three milligram of a sample was crimped in a standard aluminum pan and heated from 30◦ C to 380◦ C at a heating rate of 10◦ C/min under constant purging of nitrogen at 40 mL/min. The characteristic peak temperature and enthalpy values of endotherm and exotherm were recorded. At least triplicates were carried out for each batch of sample and results averaged. Fourier Transform Infrared Spectroscopy Two milligram of a sample was mixed with 78 mg potassium bromide [Fourier transform infrared spectroscopy (FTIR) grade; Aldrich, Steinheim, North-Rhine-Westphalia]. The mixture was ground into a fine powder using an agate mortar before compressing into a disc. Each disc was scanned at a resolution of 4 cm−1 over a wavenumber region of 450–4000 cm−1 using a FTIR spectrometer (Spectrum RX1 FTIR system; Perkin Elmer). The characteristic peaks of infrared transmission spectra were recorded. At least triplicates were carried out for each batch of sample and results averaged. Morphology The morphology of nanoparticles was examined using highresolution scanning electron microscopy (SEM) technique (JSM-6360LA; Jeol, Akishima, Tokyo, Japan). A dispersion of nanoparticles was rigidified by osmium tetroxide (Fluka Analytical), followed by washing through repetitive supernatant removal by centrifugation at 1946 g for 10 min at 25◦ C and addition of fresh batches of deionized water. A drop of washed nanoparticles was then placed onto an aluminum tape adhered

Alginate–Glucose Interaction Study The state of alginate–glucose interaction was examined using FTIR technique. Fifty milligram of alginate was wet mixed with 1 :L of deionized water with or without the addition of 50 mg glucose. Two milligram of this sample was then blended with potassium bromide and subjected to FTIR spectroscopy testing. At least triplicates were carried out for each batch of sample and the results averaged.

In Vivo Determination of Blood Glucose and Insulin Concentrations Male Sprague–Dawley rats with individual body weight of 280 ± 30 g (Genetic Improvement and Farm Technologies Sdn Bhd, Petaling Jaya, Selangor, Malaysia) were used in this study. The animals were housed in bioBubble softwall enclosures (The Colorado Clean Room Company, Colorado). They were acclimatized for 7 days under standard environmental condition at an ambient temperature of 25 ± 2◦ C and a relative humidity of 65 ± 5% on a 12-h light/dark cycle. The rats were given free access to standard pelletized food (GoldCoin Enterprise, Port Klang, Selangor, Malaysia) and deionized water ad libitum. Experimental diabetes was induced in rats according to a modified protocol using streptozotocin solution prepared in 0.1 M citrate buffer pH 4.5 and sterilized by filtration through 0.22 :m MILLEX GP Syringe Driven Filter Unit (Millipore Corporation, Bedford, Massachusetts).28 Following 15 h of fasting, each rat was administered with a single intraperitoneal injection of the freshly prepared streptozotocin solution at the dose of 55 mg/kg. In avoidance of severe hypoglycemia subsequent to streptozotocin administration, the rats were given 5% (w/v) glucose solution orally in the form of feeding liquid for a

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

R

R

DOI 10.1002/jps.23742

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

period of 24 h. The rats with diabetic symptoms such as polydipsia and polyuria as well as fasting blood glucose concentration higher than 13.75 mmol/L after 14 days of streptozotocin injection were selected for experiment. These rats were subjected to 12 h of fasting prior to experiment. All experiments were conducted in accordance to the university ethics regulations adapting the international guidelines (OECD Environment, Health and Safety) on the conduct of animal experimentation. Groups of 10 rats were randomly identified. Following fasting, these groups of rats were fed intraduodenally with insulin solution, blank, and insulin-loaded nanoparticles at a drug dose of 14 IU/kg with deionized water as dispersant. Positive control was rats receiving 3 IU/kg of insulin by subcutaneous administration. The blood was withdrawn from the tail vein of diabetic rats immediately prior to (0 h) and after the administration of test samples for the first 8 h (2, 4, 6, and 8 h) or 12–24 h (12, 16, 20, and 24 h). The blood glucose concentration of the rats was examined using the glucometer (Ascensia Elite; Bayer Corporation, Brussel, Belgium). The reduction extent of blood glucose concentration was defined as difference between the blood glucose concentration at a specified time and at 0 h to that of blood glucose concentration at 0 h, expressed in the unit of percentage. In the case of blood insulin concentration, the blood collected from the tail vein of rats at 12 h was allowed to clot and subsequently centrifuged at 1946 g for 20 min at 25◦ C to obtain the serum sample. The serum sample was preserved at −20◦ C until further analysis using the rat insulin enzyme immunoassay kit (A05105-96 WELLS; SPI-BIO, Montigny le Bretonneux, France) with insulin concentration levels read using spectrophotometer plate reader (BioTek Instrument, Winooski, Vermont) at the wavelength of 405 nm.

RESULTS AND DISCUSSION Alginate formed nanoparticles via ionotropic gelation with soluble Ca2+ . In the present investigation, high molecular weight G-rich alginate was used in the first phase of experiments on nanoparticle formation and insulin encapsulation. Alginate and Calcium Chloride Concentration In either experiments with alginate concentration or Ca2+ concentration as the formulation variable, the use of a low-weight ratio of alginate to Ca2+ resulted in the formation of large particles (Table 2; ANOVA, p < 0.05). Inferring from the reduction in zeta potential values with low-weight ratios of alginate to Ca2+ , Table 2.

4357

the formation of large particles was presumed to be a result of excessive cross-linkages formed between alginate and Ca2+ . It gave rise to both intra- as well as inter-particulate aggregation, thereby leading to a growth in particle size. At a constant weight ratio of alginate to Ca2+ , a more concentrated polymer system tended to give rise to the formation of larger particles particularly when a low alginate to Ca2+ weight ratio was used (Table 2; alginate–Ca2+ weight ratio = 2:1, p < 0.05). Using 0.1% (w/w) alginate solution, a solution concentration of 0.012% (w/w) calcium chloride dihydrate led to the formation of nanoparticles. Further increases in Ca2+ concentration brought about a sharp rise in particle size beyond the nanoparticulate ranges. Lowering the Ca2+ concentration likewise produced larger nanoparticles because of the inadequate cross-linking and densification of alginate nanostructure that were reflected by nanoparticles formed at high zeta potential values. pH of Processing Liquid Encapsulation of insulin in alginate nanoparticles can be mediated by means of hydrogen bonding and electrostatic interaction.17,29 The latter was affected by the microenvironmental pH of insulin and alginate molecules. pH modulation can be one approach used to maximize insulin association efficiency in alginate nanoparticles. Using 0.012% (w/w) calcium chloride dihydrate solution and 0.1% (w/w) alginate solution with a pH of 4.00, it was found that lowering the insulin solution pH to 2.40 produced the smallest nanoparticles with the lowest zeta potential values and the highest insulin association efficiency as well as insulin content (Table 3a). Insulin is characterized by an isoelectric point of 5.30.30 Lowering pH of an insulin solution below 5.30 would mean a conversion of its –NH2 moiety into –NH3 + . This then facilitated the electrostatic interaction between insulin and alginate, and insulin encapsulation by alginate. A further reduction in insulin solution pH to a value below 2.40 may give rise to a higher insulin association efficiency and content. Nonetheless, it was omitted from practice because of the instability of insulin at a pH value below 2.00.31 Keeping the pH of insulin solution at 2.40, the alginate solution that was characterized by a pH of 3.50 produced nanoparticles with the highest levels of insulin association efficiency and content (Table 3b). The alginate has a pKa value of approximately 3.50.2 In a medium with pHs lower than 3.50, it exhibited a low level of ionization at carboxylic acid moiety.

The Effects of Alginate and Calcium Chloride Concentrations on Size and Zeta Potential of Nanoparticles

Alginate Concentration (%, w/w)

Calcium Chloride Concentration (%, w/w)

Alginate–Calcium Chloride Weight Ratio

Particle Size (nm)

0.100

0.050 0.025 0.012 0.006 0.003

2:1 2:0.5 2:0.24 2:0.120 2:0.060

2830.00 738.33 257.00 445.67 431.00

± ± ± ± ±

52.92 193.20 110.08 110.11 64.16

− 21.20 − 24.53 − 34.44 − 41.73 − 50.30

± ± ± ± ±

0.26 3.44 0.21 0.70 0.87

0.050

0.050 0.025 0.012 0.006

2:2 2:1 2:0.5 2:0.24

4586.67 354.00 257.67 282.33

± ± ± ±

341.96 223.42 175.28 102.40

− 18.57 − 22.97 − 27.43 − 29.20

± ± ± ±

0.35 0.15 0.75 0.53

0.025

0.025

1:1

2279.67 ± 458.69

DOI 10.1002/jps.23742

Zeta Potential (mV)

− 20.70 ± 0.44

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

4358

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Table 3. Effects of pH of (a) Insulin Solution, (b) Alginate Solution, and (c) Alginate–Insulin Mixture on Size, Zeta Potential, Insulin Content, and Association Efficiency of Nanoparticles (a) Insulin Solution pH

Alginate Solution pH

Particle Size (nm)

2.40 2.75 4.50 4.80

4.00

Insulin Solution pH

Alginate Solution pH

Alginate–Insulin Mixture pH

2.40

2.50 3.00 3.50 4.00

2.58 2.90 3.00 3.50

Insulin Solution pH

Alginate Solution pH

Alginate–Insulin Mixture pH

2.40

3.50 4.00 4.50 5.00

3.00

164.33 246.33 256.67 373.33

± ± ± ±

10.12 125.86 53.12 294.57

Zeta Potential (mV) − 31.57 − 32.97 − 35.10 − 35.43

± ± ± ±

Insulin Content (%)

0.31 0.31 0.61 0.50

5.50 4.01 0.30 0.76

± ± ± ±

0.46 0.56 0.01 0.12

Insulin Association Efficiency (%) 42.17 30.74 2.27 5.80

± ± ± ±

3.51 4.29 0.06 0.88

(b) Particle Size (nm) 167.00 174.00 158.00 164.33

± ± ± ±

Zeta Potential (mV)

1.00 2.65 9.17 10.12

− 23.90 − 32.07 − 33.50 − 31.57

± ± ± ±

0.87 0.49 0.10 0.31

Insulin Content (%) 6.19 5.88 8.18 5.50

± ± ± ±

0.16 0.11 0.98 0.46

Insulin Association Efficiency (%) 47.49 45.07 62.75 42.17

± ± ± ±

1.20 0.87 7.53 3.51

(c) Particle Size (nm) 158.00 169.67 178.00 178.67

± ± ± ±

Zeta Potential (mV) − 33.50 − 31.83 − 28.27 − 30.13

9.17 6.35 1.73 2.52

There was a lack of carboxylate moiety to interact electrostatically with the amine functional group of insulin, thereby resulting low insulin encapsulation levels. In a medium of higher pHs, deprotonation of amine functional group of insulin could take place at its interface with alginate molecules. This negated electrostatic interaction propensity of alginate with insulin and gave rise to low degrees of drug encapsulation. Mixing of insulin solution of pH 2.40 and alginate solution of pH 3.50 was characterized by a final solution pH of 3.00 (Table 3b). Keeping the final solution pH at 3.00, it was found that size, zeta potential, insulin association efficiency, and insulin content of nanoparticles were not markedly affected by the solution pH of alginate (Table 3c), except that the use of alginate solution of pH 4.00 provided a marginal improvement in insulin association efficiency and insulin content of nanoparticles. As such, the subsequent experiments were conducted using insulin solution of pH 2.40, alginate solution of pH 4.00, and 0.012% (w/w) calcium chloride dihydrate solution with final alginate–insulin solution pH adjusted to 3.00 (Table 1). Preparation of Nanoparticles Four grades of alginate were employed in nanoparticle preparation, in vitro insulin release, in vivo blood glucose lowering, and insulin absorption studies. They were high molecular weight G-rich (HG) and M-rich (HM) alginates, and low molecular weight G-rich (LG) and M-rich (LM) alginates. Similar protocol of nanoparticle preparation as HG was adopted for all alginate grades except a higher Ca2+ concentration was needed to enable the formation of LG and LM nanoparticles (Table 1).

In Vitro Analysis Regardless of alginate grade, encapsulation of insulin in alginate nanoparticles was accompanied by physicochemical inter-

± ± ± ±

0.10 0.29 0.38 0.29

Insulin Content (%) 8.18 8.99 8.07 8.36

± ± ± ±

0.98 1.19 0.57 0.18

Insulin Association Efficiency (%) 62.75 68.89 61.87 64.12

± ± ± ±

7.53 9.15 4.34 1.39

action involving specific domains of alginate matrix. This was reflected by a consistent reduction in enthalpy value of DSC exotherms at high-temperature regimes between 230◦ C and 250◦ C of blank nanoparticles upon insulin loading (Fig. 2I). The exothermic enthalpy values of blank HG, HM, LG, and LM nanoparticles were reduced from −635.8 ± 34.8, −553.8 ± 32.9, −551.3 ± 1.0, and −442.3 ± 10.8 J/g to −501.7 ± 48.6, −453.3 ± 18.1, −318.6 ± 10.1, and −332.2 ± 23.4 J/g, respectively, in cases of insulin-loaded nanoparticles. Inferring from the thermograms of unprocessed polymer (Fig. 2I),21 it denoted a change in physicochemical environment of alginate beyond which could be brought about through the weight gain from insulin. The event of alginate–insulin interaction in nanoparticles was also reflected in FTIR spectroscopy analysis where the peaks ascribing to C=O moiety of polymer between 1600 and 1750 cm−1 coalesced upon insulin loading (Fig. 2II). Figure 3 shows that the alginate nanoparticles were round. The mean particle sizes of HG, HM, LG, and LM nanoparticles were mainly found between 140 and 210 nm (Table 1). LG and LM were constituted of small molecular chains of alginate and were expected to exhibit poor physical entanglement and linkages. A high-solution concentration of Ca2+ was needed to realize the nanoparticle formation. This could then partly translate to reduced zeta potential of matrices when compared with those prepared using the high molecular weight alginates. Despite the fact that a high-solution concentration of Ca2+ was used in cross-linking reaction, insulin leaching was possible during high-speed mixing process. This gave rise to low insulin association efficiency and insulin content to the low molecular weight alginate nanoparticles. Low molecular weight alginate nanoparticles exhibited a slower drug release propensity than high molecular weight alginate nanoparticles (Table 1).32

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

DOI 10.1002/jps.23742

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

4359

Figure 2. Profiles of (II) DSC thermograms and (II) FTIR spectra of (a) insulin, (b) HG, (c) HM, (d) LG, (e) LM; blank nanoparticles of (f) HG, (g) HM, (h) LG, (i) LM; insulin-loaded nanoparticles of (j) HG, (k) HM, (l) LG, and (m) LM. T, temperature (◦ C); H = enthalpy (J/g).

Pearson correlation study of HG, HM, LG, and LM nanoparticles indicated that the extent of insulin release could be governed by the insulin content of the nanomatrices (8 h: r = 0.863, p = 0.137; 24 h: r = 0.836, p = 0.164). Inferring from insulin content study, a higher polymer fraction was available in a unit weight of low molecular weight alginate nanomatrix to retard insulin release. High molecular weight G-rich nanoparticles released less insulin than the corresponding HM sample within a period of 24 h (Table 1; p < 0.05). With reference to the remarkable rise in endothermic temperatures of blank HG nanoparticles upon insulin loading, DSC analysis indicated that the strength of alginate–insulin interaction was greater when HG was used as the matrix polymer (Fig. 2I). Further characterization by means of FTIR analysis indicated that HG alginate–insulin interaction was mediated through an increase in binding strength via the O–H moiety, in addition to that of C=O functional group (Fig. 2II).21 The polymer–drug interaction via O–H moiety was suggested by a reduction in FTIR wavenumber of blank HG nanoparticles from 3468.2 ± 5.3 to 3413.0 ± 5.4 cm−1 in insulinDOI 10.1002/jps.23742

loaded nanoparticles. This sustained the release of insulin from HG nanoparticles. Other possible reason could be a greater level of reactivity of HG than HM alginate for Ca2+ and densely cross-linked HG nanoparticles were produced.2 Nonetheless, such conclusion cannot be made as preliminary atomic absorption spectrophotometry assay of nanoparticles indicated that these matrices only contained extremely low levels of Ca2+ below which can be determined accurately. In comparison to HM alginate, the HG alginate was known to be characterized by long-buckled chains with low polymer chain flexibility and reduced solute diffusivity.33 This could be one factor that aided to retard the insulin release from HG nanoparticles. Unlike high molecular weight alginate nanoparticles, LM instead of LG nanoparticles were inclined to exhibit a greater degree of drug release retardation (Table 1). Similar to HG nanoparticles, the sustained drug release property of LM nanoparticles was attributed to strong alginate–insulin binding as indicated by the rise in DSC endothermic temperature of blank nanoparticles upon insulin loading (Fig. 2I). With reference to Pearson correlation analysis, the molecular weight

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

4360

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Figure 3. Surface morphology of (a) HG, (b) HM, (c) LG, and (d) LM nanoparticles.

characteristics of polymer chains per se did not markedly govern the insulin release profiles of HG, HM, LG, and LM nanoparticles (molecular weight: 8 h: r = 0.651, p = 0.349; 24 h: r = 0.613, p = 0.387; polydispersity index: 8 h: r = 0.492, p = 0.508; 24 h: r = 0.448, p = 0.551). Stronger binding of insulin to LM than LG alginate could be facilitated by the formation of dense matrix by short and linear LM alginate.1

In Vivo Analysis In vivo studies show that subcutaneous insulin generally provided blood glucose lowering effects over at least a period of 6 h beyond which the effect subsided, with reference to the subjects administered with pure insulin solution (Figs. 4a and 4b; 2–6 h: p < 0.05). Among all tested nanoparticles, only HM nanoparticles demonstrated blood glucose lowering effects after prolonged administration for 12 h (HM nanoparticles vs. insulin solution and blank HM nanoparticles: p < 0.05). A prolonged duration was required to enable blood glucose lowering effect to manifest, and this was probably because of alginate nanoparticles releasing small insulin fractions with time. HG nanoparticles did not exhibit marked blood glucose lowering effects when compared with their drug-free nanoparticulate sample (p = 1.00). The quantum of insulin release could have governed the blood glucose lowering capacity of alginate nanoparticles. Using low molecular weight alginate nanoparticles, there was no significant reduction in the blood glucose

levels of rats in spite of an extended administration was given (12–24 h: p = 0.99–1.00). Interestingly, insulin-free LM and HG nanoparticles had their blood glucose lowering property exhibited with reference to subjects administered with pure insulin solution (12–20 h: p < 0.05). The fiber in seaweed had been reported to be able to delay glucose absorption from gastrointestinal tract.34 Figure 5 showed that both HG and LM alginate could interact with glucose in an aqueous milieu via the O–H moiety. The mixing of HG or LM with glucose in an aqueous milieu was characterized by the existence of FTIR peaks at wavenumbers intermediate to those of HG/LM and glucose, which were less observable in the cases of HM and LG. Both low and high molecular weight alginates were known to inhibit glucose absorption from the small intestine.35 Formulating alginate in the form of nanoparticles led to low molecular weight LM and high molecular weight HG alginate demonstrated blood glucose lowering property. Inferring from DSC studies, the high chemical-binding affinity of HG and LM nanoparticles might have translated to strong glucose-binding ability of these matrices in gastrointestinal tract. The absorption of glucose could be hindered by these matrices thereby leading to blood glucose lowering effects. Such observation was suggested as there was no marked elevation in the blood insulin levels of rats administered with insulin-free nanoparticles (Fig. 4c). On the contrary, blood glucose lowering effects of HM nanoparticles were

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

DOI 10.1002/jps.23742

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

4361

Figure 4. Blood glucose lowering profiles at (a) the first 8 h and (b) 12–24 h; and insulin levels at (c) 12 h of rats receiving insulin solution, blank, and insulin-loaded nanoparticles of HG, HM, LG, and LM orally. Positive control received subcutaneous insulin at 3 IU/kg. Extent of blood glucose lowering of * positive control > insulin solution (p < 0.05), @ HM nanoparticles > insulin solution and blank HM nanoparticles (p < 0.05), # blank LM and HG nanoparticles > insulin solution (p < 0.05).

accompanied by a tendency to have rats characterized by high blood insulin concentrations.

CONCLUSIONS High molecular weight M-rich alginate nanoparticles demonstrate blood glucose lowering effects in diabetic rats through their ability to encapsulate, release, and have insulin absorbed

DOI 10.1002/jps.23742

into systemic circulation. Insulin-free HG and LM alginate nanoparticles are able to lower blood glucose levels probably via their strong chemical-binding affinity that translates to glucose retention in intestinal tract. HM alginate is the best polymer candidate for use in design of oral insulin carrier for prolonged action over 24 h. This study identifies the primary choice of alginate grade that can serve as a platform for subsequent studies in nanoparticle decoration.

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

4362

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Figure 5. Fourier transform infrared spectra of (a) glucose, (b) HG, (c) HG–glucose, (d) HM, (e) HM–glucose, (f) LG, (g) LG–glucose, (h) LM, and (i) LM–glucose in aqueous milieu.

ACKNOWLEDGMENTS

REFERENCES

The authors wish to express their heart-felt thanks to Nanofund, Ministry of Science, Technology and Innovation, and Ministry of Higher Education of Malaysia (0141903) for facility and grant support.

1. Gombotz WR, Wee SF. 1998. Protein release from alginate matrices. Adv Drug Deliv Rev 31:267–285. 2. Wong TW. 2011. Alginate graft copolymers and alginate-co-excipient physical mixture in oral drug delivery. J Pharm Pharmacol 63:1497– 1512.

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013

DOI 10.1002/jps.23742

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

3. Tam SK, Bilodeau S, Dusseault J, Langlois G, Halle JP, Yahia LH. 2011. Biocompatibility and physicochemical characteristics of alginatepolycation microcapsules. Acta Biomater 7:1683–1692. 4. Skaugrud Ø, Hagen A, Borgersen B, Dornish M. 1999. Biomedical and pharmaceutical applications of alginate and chitosan. Biotechnol Genet Eng Rev 16:23–40. 5. George M, Abraham TE. 2006. Polyanionic hydrocolloids for the intestinal delivery of protein drugs: Alginate and chitosan—A review. J Control Release 114:1–14. 6. Vos PD, Faas MM, Strand B, Calafiore R. 2006. Alginate-based microcapsules for immunoisolation of pancreatic islets. Biomaterials 27:5603–5617. 7. Zimmermann H, Shirley SG, Zimmermann U. 2007. Alginate-based encapsulation of cells: Past, present and future. Curr Diab Rep 7:314– 320. 8. Draget KI, Taylor C. 2011. Chemical, physical and biological properties of alginates and their biomedical implications. Food Hydrocolloid 25:251–256. ´ ´ 9. Gonzalez-Rodr´ ıguez ML, Holgado MA, Sanchez-Lafuente C, Rabasco AM, Fini A. 2002. Alginate/chitosan particulate systems for sodium diclofenac release. Int J Pharm 232:225–234. 10. Chen S-C, Wu YC, Mi FL, Lin YH, Yu LC, Sung HW. 2004. A novel pH-sensitive hydrogel composed of N,O-carboxymethyl chitosan and alginate crosslinked by genipin for protein drug delivery. J Control Release 96:285–300. 11. Lee KY, Yuk SH. 2007. Polymeric protein delivery systems. Prog Polym Sci 32:669–697. 12. Yang L, Zhang B, Wen L, Liang Q, Zhang LM. 2007. Amphiphilic cholesteryl grafted sodium alginate derivative: Synthesis and selfassembly in aqueous solution. Carbohydr Polym 68:218–225. 13. Yao B, Ni C, Xiong C, Zhu C, Huang B. 2010. Hydrophobic modification of sodium alginate and its application in drug controlled release. Bioprocess Biosyst Eng 33(4):457–463. 14. Wong TW. 2009. Chitosan and its use in design of insulin delivery system. Rec Pat Drug Deliv Formulat 3(1):8–25. 15. Wong TW. 2010. Design of oral insulin delivery systems. J Drug Target 18(2):79–92. 16. Carino GP, Jacob JS, Mathiowitz E. 2000. Nanosphere based oral insulin delivery. J Control Release 65:261–269. 17. Sarmento B, Martins S, Ribeiro A, Veiga F, Neufeld R, Ferreira D. 2006. Development and comparison of different nanoparticulate polyelectrolyte complexes as insulin carriers. Int J Peptide Res Ther 12(2):131–138. 18. Sarmento B, Ribeiro A, Veiga F, Sampaio P, Neufeld R, Ferreira D. 2007. Alginate/chitosan nanoparticles are effective for oral insulin delivery. Pharm Res 24(12):2198–2206. 19. Reis CP, Veiga FJ, Ribeiro AJ, Neufeld RJ, Damge C. 2008. Nanoparticulate biopolymers deliver insulin orally elicit-

DOI 10.1002/jps.23742

4363

ing pharmacological response. Pharm Nanotechnol 97(12):5290– 5305. 20. Woitiski CB, Neufeld RJ, Veiga F, Carvalho RA, Figueiredo IV. 2010. Pharmacological effect of orally delivered insulin facilitated by multilayered stable nanoparticles. Eur J Pharm Sci 41:556– 563. 21. Radman EM, Wong TW. 2010. Effects of microwave on drug release responses of spray-dried alginate microspheres. Drug Dev Ind Pharm 36(10):1149–1167. 22. Kharenko EA, Larionova NI, Demina NB. 2009. Mucoadhesive drug delivery systems. Pharm Chem J 43(4):200–208. 23. Racovita S, Vasiliu S, Popa M, Luca C. 2009. Polysaccharide based on micro- and nanoparticles obtained by ionic gelation and their applications as drug delivery systems. Revue Roumaine de Chimie 54(9):709–718. 24. Bronagh UK, Owen IC. 2006. Effect of soluble filler on drug release from stearic acid based compacts. Int J Pharm 316:47–51. 25. Waree T, James W, Robert CS, Middaugh CR. 2003. Insulin containing polyethylenimine–dextran sulfate nanoparticles. Int J Pharm 255:139–151. 26. Harjoh N, Wong TW. 2011. Design of in situ dispersible and calcium cross-linked alginate pellets as intestinal-specific drug carrier by melt pelletization technique. J Pham Sci 100(6):2248–2257. 27. Shimariza Ashikin WHN, Wong TW, Law CL. 2010. Plasticity of hot air-dried mannuronate- and guluronate-rich alginate films. Carbohydr Polym 81:104–113. 28. Noor Shahida A, Wong TW, Choo CY. 2009. Hypoglycemic effect of quassinoids from Brucea javanica (L.) Merr (Simaroubaceae) seeds. J Ethnopharmacol 124(3):586–591. 29. Silva CM, Ribeiro AJ, Veiga F, Sousa A. 2006. Insulin release from alginate microspheres reinforced with dextran sulfate. Chem Ind Chem Eng Quart 12:40–46. 30. Sigma. 2001. Insulin from bovine pancreas. Sigma Prod. No. I5500, I6634, I1882, I4011 (CAS number: 11070-73-8):1–3. 31. Gualandi-signorini AM, Giorgi G. 2001. Insulin formulations—A review. Eur Rev Med Pharmacol Sci 5:73–83. 32. Moya ML, Morley M, Khanna O, Opara EC, Brey EM. 2012. Stability of alginate microbead properties in vitro. J Mat Sci Mat Med 23(4):903–912. 33. Amsden B, Turner N. 1999. Diffusion characteristics of calcium alginate gels. Biotechnol Bioeng 65(5):605–610. 34. Kim MS, Kim JY, Choi WH, Lee SS. 2008. Effects of seaweed supplementation on blood glucose concentration, lipid profile, and antioxidant enzyme activities in patients with type 2 diabetes mellitus. Nutr Res Pract 2(2):62–67. 35. Kimura Y, Watanabe K, Okuda H. 1996. Effects of soluble sodium alginate on cholesterol excretion and glucose tolerance in rats. J Ethnopharmacol 54:47–54.

Kadir, Mokhtar And Wong, JOURNAL OF PHARMACEUTICAL SCIENCES 102:4353–4363, 2013