Macromolecular confinement of therapeutic protein in polymeric particles for controlled release : insulin as a case study

1Laboratory for Pharmaceutical Biotechnology, CCS, Federal University of Rio de Janeiro, UFRJ, Rio de Janeiro, RJ, Brazil. 2Department of Pharmaceutical Sciences, Institute of Biological and Health Sciences, Rural Federal University of Rio de Janeiro, UFRRJ, Seropédica, RJ, Brazil, 3Laboratory for Enzymology and Metabolism Control, LabECoM, Federal University of Rio de Janeiro, UFRJ, Rio de Janeiro, RJ, Brazil, 4Laboratory for Cell Supramolecular Structure Prof. Hertha Meyer, IBCCF, Federal University of Rio de Janeiro, UFRJ, Rio de Janeiro, RJ, Brazil, 5Laboratory for Galenic Pharmaceutical Technology, LADEG, Federal University of Rio de Janeiro, UFRJ, Rio de Janeiro, RJ, Brazil


INTRODUCTION
The World Health Organization (WHO) estimates that more than 180 million people worldwide are diabetic and this health problem is responsible for 5% of all deaths each year (Wild et al., 2004;Roglic et al., 2005).The Diabetes Mellitus (DM) is a group of chronic diseases which occur when pancreas does not produce enough insulin or there is a resistance to this hormone, and in some cases are observed production and action of insulin deficient condition increasing blood sugar level and risk of microvascular damage (retinopathy, nephropathy and neuropathy) (American Diabetes Association, 2013;WHO, 2008).
Insulin is a polypeptide hormone synthesized in the β-cells of Langerhans islets in the pancreas.This hormone is composed of two chains linked by disulfide bonds, chain A with 21 amino acids and chain B with 30 amino acids.Hormone replacement therapy using exogenous insulin is necessary for patients with DM, especially for patients with Type 1 DM.However, a perfect glycemic control is still a difficult task despite the number of insulin formulations in market with different pharmacokinetics characteristics (Veiseh et al., 2015).
Therefore, the developments of pharmaceutical compositions, which reduce the number of insulin shots, are needed for both better glycemic control and for improving patient's quality of life.There are many approaches for sustained release development, but not all methods are applicable to biopharmaceuticals due to its high lability.The most common methods used in pharmaceutical biotechnology are: nano and microencapsulation, using solid matrixes as PLGA, PLA and PCL; liposomes,, which uses liquid matrixes such as phosphatidylcholine; derivatization with polymers, such as polyethylene glycol; the use of biocompatible gels, such as Pluronic F27; or the combination of two or more of these techniques (Calceti et al., 2004;Almeida, Souto, 2007;Liu et al., 2007a;Liu et al., 2007b;Huynh et al., 2009;Rastogi, Anand, Koul, 2009).
Several approaches have been proposed to make an insulin sustained release system but so far no medicine containing micro or nanoencapsuladed insulin has been available commercially.In our work, we ought to develop a system containing insulin confined into polymeric microparticles, which could be used for the sustained release of insulin in vivo.In this study, the PCL was the polymer of choice because it is of lowcost, biocompatible, biodegradable and it has been using in various formulations (Benoit, Baras, Gillard, 1999;Lamprecht et al., 1999;Sinha et al., 2004;Hasan et al., 2007;Wang et al., 2007).We report here a simple method based on double emulsification followed by simple vacuum solvent extraction for the production of PCL microparticles comprising bioactive insulin.

Preparation of insulin microparticles
PCL microparticles were prepared by a modified double-emulsion (water-in-oil-in-water) solvent evaporation/extraction technique (Benoit, Baras, Gillard, 1999;Damgé et al., 2009).Briefly, 400 mL regular-acting human insulin formulation (Humulin R ® ) and 100 mL of 5 % PVA aqueous solution was emulsified in a mixed organic solution (5 mL of dichloromethane) containing PCL (100 mg) by an Ultraturrax ® at 10,000 rpm for 1 min.Thereafter, the primary emulsion was poured into 25 mL of the 1.5 % PVA aqueous solution and homogenized with magnetic stirrer for 10 minutes.The resulting emulsion was stirred with a magnetic stirrer in low speed and the dichloromethane was purged out by evaporation under reduced pressure.The particles were washed out by centrifuging for 15 min at 14.000 x g at 4 o C and resuspending the particles in 1.5% PVA aqueous solution for three times.

Particle size distribution and morphology
Particle size distribution (mean diameter and polydispersity index) of microparticles was determined at 25 o C by dynamic laser light scattering (Shimadzu SALD-2201, Shimadzu, Japan) over the range of 0.005 to 1,000 mm, in a 1.0 cm quartz cell after proper dilution in water.Particle distribution and morphology were also evaluated by conventional optical microscopy with particles dispersed in water over siliconized glass slide.Several viewing fields were registered and particle size measurement was performed using ImageJ software (version 1.42, National Institutes of Health, USA), computed from a set comprising a total of 10,000 particles randomly selected.Results are representative of measurements performed with three microparticles preparations (n=3), to confirm mean diameter obtained by DLS (data not show).

Morphological analysis by scanning electron microscopy (SEM)
The morphological analysis of the resulting microparticles was performed using Scanning Electron Microscopy (SEM).In brief, a suspension of the microparticles was dispersed onto a glass slide and plunge frozen in liquid nitrogen and freeze-dried.Dried samples were then coated with 10 nm gold layer at room temperature and observed in a JEOL6340 Field Emission SEM operating at 5.0 kV.

Insulin encapsulation efficiency and yield
The total amount of insulin entrapped within PCL microparticles was accessed by measuring the amount of nonentrapped protein in the external aqueous solution recovered after centrifugation and washing of the microparticles.Protein loading efficiency was expressed as the percentage of insulin entrapped with respect to the theoretical value (Equation 1).The yield was calculated using the particles after washing and lyophilization.The value was the percentage of mass recovered as follow (Equation 2).Protein was quantified with the Bradford method (Bradford, 1976), using insulin for standard curve.Insulin for the standard curve was quantified by absorbance measurements at 280 nm (Gill, Von Hippel, 1989).

In vitro kinetic release assays
The washed microparticles were dispersed in 20 mL of PBS pH 7.4 and 0.02% NaN 3 , distributed in centrifugal microtubes containing 1.0 mL of the microparticles suspension each and incubated statically at 37 °C (Guerreiro et al., 2012).This assay was performed in sink conditions with the theoretical maximum insulin concentration below 0.05 mg/mL.At varying time intervals, samples were centrifuged for 30 minutes at 20,000 g at 4 o C, and the supernatant were assayed for total protein released using the Comassie colorimetric method (Bradford, 1976).

Pharmacological evaluation in vivo
Pharmacological evaluation of the hypoglycemic effect of insulin was performed as previously reported (Finotelli et al., 2010) with Swiss male mice (8 weeks; 27 g ± 1 g), housed in a temperature-controlled room with a 12 h light-dark cycle and given free access to water and normal chows.Type 1 diabetes was induced by one intraperitoneal (ip.) injection of STZ (200 mg/ kg) dissolved in fresh citrate buffer (100 mM, pH 4.5).After STZ administration (5 days), blood was drawn from mice by tail snip and glucose levels were measured using a glucometer (Accu-Chek® Active -Roche).After STZ treatment, animals, which show more than 300 mg/ dL, were considered diabetics.Mice were separated in three groups: i) control: receiving placebo MP (n= 3); ii) receiving soluble regular human insulin (n=5; 5 U/kg); and iii) test: receiving insulin-loaded microparticles (n=5; 10 U/kg).Food was suspended 4 h before the experiments, and all mice were fasted throughout the experiments and kept at 22±2 o C. Mice were administrated by subcutaneous route (s.c.) using a standard 29 gauge needle (BD TM ).The glycemia was monitored following the recommendations of the Mouse Metabolic Phenotyping Center (MMPC) Consortium from the National Institutes of Health (NIH) (Ayala et al., 2010) by using whole blood from the tail tip of conscious, unrestrained mice using pre-calibrated point-ofcare glucometers (Accu-Chek® Active, Roche Diagnostics, Germany; Serial No GN08146937) before injection and at repeated intervals after administration.The pharmacological evaluation over a prolonged time was not possible due to the limitations in the extent of animal starvation.This protocol was approved by the Institutional Bioethics Committee on Animal Care and Experimentation at UFRJ.

Data analysis
Data analysis of insulin release in vitro and insulin effects over plasma glycemia was performed to obtain the rates by non-linear least squares fitting, using double exponential equation as follow: Eq. 3 where C t and C 0 are the concentration of insulin or glucose at time t and zero, respectively; A and k are the event amplitude (concentration changes) and kinetic rates for phases 1 and 2, respectively.

Statistical analysis
Statistical analysis was performed with the t-test.The significance of the differences in the pharmacological evaluation was calculated and values of p<0.05 were considered to be significant.The analyses were performed using the SigmaStat as implemented in SigmaPlot 11 (Systat Software Inc).

Characterization of the microparticles
We have prepared insulin-loaded microparticles using PCL and a simple w/o/w emulsification system.Our preparation yielded microparticles with average size of 9.8+0.6 mm, as evaluated by dynamic light scattering measurements (Figure 1).
Optical microscopy imaging shows uniform, individualized spherical microparticles, with average sizes of 10.8 ± 4.6 mm (Figure 2), in agreement with DLS measurements.The particles have micrometric sizes, as demonstrated by the morphological evidences shown in the optical microscopy, because this data reveal compartmentalized microparticles.The insulin concentration in supernatant of the washing pool indicates efficiency of 64.5±6.8%.After freeze-dried samples were weighed and yield calculated for the process was 88.4±4.2%.

In vitro release
Figure 3 shows the release kinetics in vitro for microencapsulated insulin formulation.In this result was possible to observe a burst effect with about 50% of insulin being released in the first 2 hours.The kinetics of in vitro release was shown biphasic with an initial burst effect in the first 2 h and a second phase of release, slower, approximately four days.

In vivo release
Were performed in vivo experiments using diabetic mice to evaluate the activity of insulin released over time.Figure 4 shows that the formulation was able to reduce blood glucose levels of mice for 24 hours significantly different from that observed in the group of mice that received the commercial formulation of insulin R.
The development of formulations with insulin sustained release systems have been previously reported in the literature (Liu et al., 2007a(Liu et al., , 2007b;;Damgé et al., 2009;Han et al., 2009;Huynh et al., 2009;Kim et al., 2009;Klingler, Müller, Steckel, 2009;Ungaro et al., 2009).Almost recent studies use nanoparticles to increase the insulin effect.However, it was reported the nanoparticles may move from injection sites to the blood vessels leading to the capture of these by the liver and spleen and consequent loss of action (De Jong, Borm, 2008).Despite of all efforts, only a few systems for insulin pulmonary delivery reached the market, and one of these was removed after showed some side effects such as the formation of amyloid deposits in the lungs of patients (Lasagna-Reeves et al., 2010).The encapsulation efficiency of 64.5% is lower than the expected for the method of double-emulsion and solvent extraction using emulsion type w/o/w since other authors using a similar methodology obtained higher efficiency (Damgé, Maincent, Ubrich, 2007;Peng et al., 2012).However, the encapsulation efficiency per se is not an exclusion criterion in the development pipeline, once assured proper pharmacological profile and maintenance of physical-chemical stability of the peptide in the formulation.In preliminary studies we tried to use dialyzed insulin, but the formulations without excipients (metacreol, glycerol etc.) did not show satisfactory data (data not shown) indicating that our method, to keep these excipients, was able to better preserve this hormone.
Regarding the release profile in vitro was observed that the preparation have a release kinetic over three days, and the first day of a release of most of the material.The kinetics of release stresses of the first 2 hours was also observed in experiments using diabetic mice, and the hypoglycemic effect of the formulation was more than 24 hours of experiment.Interestingly, despite having a release kinetics in vitro faster than those described by other authors working with similar formulations developed here, our in vivo release data shows a long-lasting action, suggesting a more effective stabilization.We believe that this stability is due in part because of the use of preformulated regularacting insulin from commercial source (insulin and adjuvants) to prepare the microparticles, since experiments using non-formulated insulin (without excipients) did not result in satisfactory encapsulation efficiency as well as in an erratic release profiles (data not shown).These data demonstrate that the encapsulation efficiency and quality of the resulting product for protein products in a biologically active form is dependent of the formulation process, including stabilization of the proteic active pharmaceutical ingredient.In fact, glycerol and metacresol are important excipients for the stability of commercial insulin formulations.
Glycerol is commonly used as an osmotic agent, but this adjuvant also can prevent the formation of high-molecular-weight protein and aggregates in the   compositions with insulin (Brange, Langkjaer, 1992).
Phenol derivatives are used because of their antimicrobial effects as preservatives, but they are capable of reducing deamidation processes, which are the main route of degradation of insulins during storage (Whittingham et al., 1998).In additional, phenolic derivates can stabilize interactions between monomers and insulin dimers and hexameric conformation (Crommelin, Sindelar, Meibohm, 2008).In this work we have not investigated the effects of the excipients alone since both are important for the stability of the compositions because they act in different ways, but further studies may be carried out in order to evaluate the stability of the insulins in polymer matrices in the presence of pharmaceutical adjuvants.
Another important finding in our study is the reduction in blood glucose in animals treated in the first points of the test.Thus, although high dose of 10 IU versus 5 IU in control group, the microencapsulated formulation was able to prevent episodes of hypoglycemia in diabetic mice, a major side effect of commercial insulin formulations.Another important application of this system might be the prevention of amyloid insulin formation, since studies reporting insulin aggregates at the application site are frequent (Swift et al., 2002;Nagase, 2009;Lasagna-Reeves et al., 2010).In contrast, polymer particles similar to those developed in this work have been used to inhibit markedly amyloidogenic proteins (Cabaleiro-Lago et al., 2010;Guerreiro et al., 2012).

CONCLUSION
Despite all efforts for developing intelligent drug delivery systems, the characteristics of complex peptides, like insulin, added to the specific needs for diabetes treatments make this a challenge that seems to have a distant end.Our work has shown a simple and scalable way to produce longer lasting insulin formulation with no hypoglycemic effect in the tested dose, which can improve the quality of the therapeutic intervention.

FIGURE 1 -
FIGURE 1 -Insulin-PCL particles size distribution.Particles were evaluated by DLS after dispersion in water.Mean particle diameter = 9.8 + 0.57 mm.Details in the Material and Methods section.FIGURE 2 -Morphological characterization of insulin-loaded PCL microparticles.Optical microscopy (A, B and C).Scanning Electron Microscopy (D).Details in the Material and Methods section.

FIGURE 2 -
FIGURE 1 -Insulin-PCL particles size distribution.Particles were evaluated by DLS after dispersion in water.Mean particle diameter = 9.8 + 0.57 mm.Details in the Material and Methods section.FIGURE 2 -Morphological characterization of insulin-loaded PCL microparticles.Optical microscopy (A, B and C).Scanning Electron Microscopy (D).Details in the Material and Methods section.

FIGURE 3 -
FIGURE 3 -Kinetics of human insulin release from PCL microparticles.A) Cumulative release in a linear time scale.Inset: detail view of kinetics at initial range.Solid lines correspond to best fitting using Equation 1. Results are depicted in Table I and II.B) Higuchi plot.Solid lines are solely to depict human insulin releasing phases.Data represents measurements in triplicate from kinetics performed with three microparticles batches.

FIGURE 4 -
FIGURE 4 -Pharmacological evaluation of MP-insulin in diabetic mices.Streptozotocin-induced mices were administrated s.c. with regular human insulin solution (closed symbols; n=5) or insulin-loaded PCL microparticles (open symbols, n=5) and the glucose level was evaluated.Symbols represent mean ± SD.Inset: glycemia change at initial time range.* p<0.05 (t-test).Lines are non-linear regression with Eq. 1.Details in the Material and Methods section.

TABLE I -
Kinetic parameters from the in vitro release of human insulin from the PCL microparticles.Parameters were obtained from data in Figure3

TABLE II -
Kinetic parameters from the pharmacological evaluation of the glycemic effect of MP-insulin.Parameters were obtained from data in Figure3