Microencapsulated Solid Lipid Nanoparticles as a Hybrid Platform for

Aug 15, 2017 - Related Content: Preparation of the HIV Attachment Inhibitor BMS-663068. Part 9. Active Pharmaceutical Ingredient Process Development a...
7 downloads 3 Views 7MB Size
Download PDF
Article pubs.acs.org/molecularpharmaceutics

Microencapsulated Solid Lipid Nanoparticles as a Hybrid Platform for Pulmonary Antibiotic Delivery Diana P. Gaspar,†,§ Maria Manuela Gaspar,§ Carla V. Eleutério,§ Ana Grenha,⊥,¶ Mateo Blanco,‡ Lídia M. D. Gonçalves,§ Pablo Taboada,‡ António J. Almeida,§ and Carmen Remuñań -López*,† †

Nanobiofar Group, Pharmacology, Pharmacy and Pharmaceutical Technology Department, Faculty of Pharmacy and ‡Colloids and Polymers Physics Group, Condensed Matter Physics Department, Faculty of Physics, University of Santiago de Compostela-Campus Vida, 15782 Santiago de Compostela, Spain § Research Institute for Medicines (iMed.ULisboa), Faculty of Pharmacy, Universidade de Lisboa, Av. Professor Gama Pinto, 1649-003 Lisbon, Portugal ⊥ Centre for Biomedical Research (CBMR) and ¶Centre for Marine Sciences (CCMar), University of Algarve, Faculty of Sciences and Technology, Campus de Gambelas, 8005-139 Faro, Portugal ABSTRACT: Solid lipid nanoparticles (SLN) containing rifabutin (RFB), with pulmonary administration purposes, were developed through a technique that avoids the use of organic solvents or sonication. To facilitate their pulmonary delivery, the RFB-loaded SLN were included in microspheres of appropriate size using suitable excipients (mannitol and trehalose) through a spray-drying technique. Confocal analysis microscopy showed that microspheres are spherical and that SLN are efficiently microencapsulated and homogeneously distributed throughout the microsphere matrices. The aerodynamic diameters observed an optimal distribution for reaching the alveolar region. The dry powder’s performance during aerosolization and the in vitro drug deposition were tested using a twin-impinger approach, which confirmed that the microspheres can reach the deep lung. Isothermal titration calorimetry revealed that SLN have higher affinity for mannitol than for trehalose. Upon microsphere dissolution in aqueous media, SLN were readily recovered, maintaining their physicochemical properties. When these dry powders reach the deep lung, microspheres are expected to readily dissolve, delivering the SLN which, in turn, will release RFB. The in vivo biodistribution of microencapsulated RFB-SLN demonstrated that the antibiotic achieved the tested organs 15 and 30 min post pulmonary administration. Their antimycobacterial activity was also evaluated in a murine model of infection with a Mycobacterium tuberculosis strain H37Rv resulting in an enhancement of activity against M. tuberculosis infection compared to nontreated animals. These results suggest that RFB-SLN microencapsulation is a promising approach for the treatment of tuberculosis. KEYWORDS: aerodynamic properties, deep lung deposition, microencapsulated SLN, pulmonary administration, rifabutin, solid lipid nanoparticles

1. INTRODUCTION

these features could further decrease systemic concentrations and reduce the adverse effects.4 Nanoparticles of different compositions have been proposed to deliver drugs to the lung.1,5,6 Solid lipid nanoparticles (SLN) can be produced without using organic solvents and can incorporate hydrophobic or hydrophilic drugs, thus fulfilling the requirements for an ideal nanoparticulate drug carrier system for the pulmonary route.7 Although SLN have already been efficiently aerosolized, reaching the deep lung region,8 some instability phenomena common to nanosized systems,

Pulmonary tuberculosis (TB) is the most frequent form of TB. In the lung, Mycobacterium tuberculosis is engulfed within alveolar macrophage.1 Nowadays, effective chemotherapy of TB involves daily oral administration of two or more drugs for a period of six months or longer.2 Moreover, most orally delivered anti-TB drugs currently in use fail to reach high drug concentrations in the lung since drugs can be degraded before reaching their target site.3 Hence, delivering drugs to the infected alveolar macrophages through pulmonary administration may offer the therapeutic benefit of the drug being directly targeted to infected sites, which confers a shorter lag time before onset of action. This direct targeting can also reduce the therapeutic dose as well as treatment schedule. All © XXXX American Chemical Society

Received: March 5, 2017 Revised: July 2, 2017 Accepted: July 7, 2017

A

DOI: 10.1021/acs.molpharmaceut.7b00169 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Vaz Pereira, S.A. (Portugal). Lung surfactant (Curosurf) was a generous gift from Angelini Farmacêutica, Lda. (Portugal). Phosphate buffered saline (PBS, pH 7.4) was from Invitrogen. Bodipy 630/650-X was from Molecular Probes (Netherlands). Purified water was produced by inverse osmosis (Millipore, Elix 3) with a 0.45 μm pore filter. Middlebrook 7H9 broth and 7H11 agar, BACTO Middlebrook and albumin-dextrosecatalase enrichments were from Difco Laboratories (Detroit, USA). MTS assay reagents were from Promega (WI, USA). All other reagents were of analytical grade and were used without further purification. 2.1.2. Animals. BALB/c mice (6−8 weeks old, 25−30 g) were acquired from the Gulbenkian Institute of Science (Oeiras, Portugal). The animals were maintained with free access to commercial food and water ad libitum under standard hygiene conditions. All animal experiments were conducted according to the local ethical committees of the Faculty of Pharmacy and of the Institute of Molecular Medicine in agreement with the EU Directive (2010/63/UE) and Portuguese laws (DR 113/2013, 2880/2015, and 260/2016). 2.1.3. Mycobacterial Strains. For in vitro studies, the M. avium strain DSMZ 44157 from Leibniz-Institute, Germany was used. For in vivo studies, the M. tuberculosis strain H37Rv was used. Quantification of inocula were performed as described previously by Gaspar et al.14 2.2. Methods. 2.2.1. Preparation of RFB-Loaded SLN. RFB-loaded glyceryl dibehenate and glyceryl tristearate SLN were prepared using a hot high shear homogenization (HSH), as previously described.15 Briefly, 300 mg of the lipid phase was melted at a temperature 10 °C above its m.p.; RFB was dissolved in the melted lipid (drug concentration was calculated to obtain theoretical contents of 10.0% and 7.3% (w/w) relative to glyceryl dibehenate and glyceryl tristearate, respectively). A hot aqueous phase consisting of Tween 80 in purified water (at 2% and 7% (w/v) for glyceryl dibehenate and glyceryl tristearate SLN, respectively) was added into the lipid phase and homogenized using a high-shear laboratory mixer (Silverson SL2, UK) at 12 300 rpm for 10 min. The SLN dispersions were allowed to cool in an ice bath, with mild stirring for 5 min. Each formulation was carried out in triplicate (n = 3). 2.2.2. Characterization of RFB-Loaded SLN. Size and zeta potential of RFB-loaded SLN were analyzed by photon correlation spectroscopy (PCS) using a Zetasizer Nano S and Z (Malvern Instruments, UK). Briefly, samples were diluted (1:100) with filtered (0.45 μm) purified water, placed in polystyrene cuvettes, and their size measured at a 173° scattering angle at 25.0 ± 0.1 °C. Surface charge was characterized to determine the mobility of SLN (previously diluted with filtered purified water and placed in a appropriate cuvette) under the influence of an electric field. Particle size distribution and morphology were also analyzed by atomic force microscopy (AFM). Noncontact AFM in air was performed in a XEI-100 instrument from Park Systems decoupled XY (50 Å ∼ 50 μm2) and Z-scanners (12 μm). SLN formulations were diluted (1:250) in filtered purified water and placed on cleaved muscovite mica for ∼20 min. Then samples were washed and allowed to air-dry at room temperature, mounted on magnetic discs, and visualized by using microfabricated crystal silicon probes with a spring constant of 40 N/m and a resonant frequency of 300 kHz (ACTA, AppNano). To avoid imaging artifacts, the scan direction was varied to ensure a true image. The images were obtained from at least three macroscopically separated areas on each sample. All

such as particle−particle interactions and poor delivery efficiency due to exhalation of low-inertia nanoparticles, could not be precluded. In previous studies dealing with polymeric nanoparticles intended for inhalation, their microencapsulation in micron-sized powder carriers using suitable spray-drying techniques has been proposed.5,9 This strategy enhanced the nanoparticles handling and aerosolization through improvement of powder stability and aerodynamic properties, leading to an efficient pulmonary delivery. Dry powders have demonstrated to provide higher drug absorption than their liquid counterparts.10 The so formed microparticles present a suitable aerodynamic diameter (daer) and lung deposition. Aerosol powders with suitable aerodynamic properties must be composed by particles that are neither too small, hazarding exhalation, nor too large, which are likely to sediment in the upper airways. These characteristics are crucial for enabling nanoparticles to be phagocytized by alveolar macrophages where M. tuberculosis resides and proliferates. The optimal daer for reaching the alveolar region has been described to be between 1 and 5 μm.11 Besides, the nanoparticle-containing microparticles must readily deliver the primary nanoparticles upon depositing in the lung interstitial fluid, which should retain their benefits as highly potent carriers for biopharmaceuticals. This feature is crucial for nanoparticles being phagocytised by the alveolar macrophages where the Mycobacterium tuberculosis harbors.12,13 For this purpose, the spray-drying process typically employs a wide range of inert pharmaceutical excipients, which are approved by the regulatory authorities and act as drying adjuvants that protect the structural integrity of the nanoparticles, avoiding coalescence upon exposure to high temperature during the spray-drying process and during subsequent storage.13 The aim of this study was to obtain microencapsulated SLN powders adequate for the pulmonary administration of the antiTB drug rifabutin (RFB). Herein, we report the preparation of two types of RFB-loaded SLN based on glyceryl dibehenate and glyceryl tristearate (biocompatible and biodegradable solid lipids exhibiting high melting point (m.p.)). Thereafter, these SLN were microencapsulated in mannitol and trehalose microspheres by spray-drying to acquire adequate morphological and aerodynamic properties for lung deposition. Structural and aerodynamic characterization of these micropowders loaded with SLN was performed using scanning electron microscopy (SEM), confocal laser scanning microscopy (CLSM), and aerodynamic sampling using a twinimpinger. Physical characterization of the powders was performed by isothermal titration calorimetry (ITC). SLN recovery from microspheres and RFB release from microencapsulated SLN were also assessed. In vitro antimycobacterial activity of RFB formulations was evaluated by a (3-(4,5dimethylthiazol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfophenyl)-2H-tetrazolium) (MTS) reduction assay. The in vivo profile of microencapsulated RFB-SLN was analyzed following 15 and 30 min post pulmonary administration. Their antimycobacterial activity was also evaluated in a murine model of infection with a M. tuberculosis strain H37Rv.

2. EXPERIMENTAL SECTION 2.1. Materials. 2.1.1. Chemicals. RFB was purchased from CHEMOS GmbH (Germany). Glyceryl tristearate, mannitol, and D-(+)-trehalose dihydrate were acquired from SigmaAldrich (Spain). Glyceryl dibehenate was a kind gift from Gattefossé (France). Tween 80 (polysorbate 80) was from J. B

DOI: 10.1021/acs.molpharmaceut.7b00169 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

2.2.5. Characterization of Microspheres Size, Morphology, and Moisture Content. Particle size and morphology of the spray-dried powders were assayed using scanning electron microscopy (SEM). The dry powder samples were mounted on an aluminum stub using a double-side adhesive tape, covered with a 200 nm thick gold−palladium film using an Emitech K550 (London, UK) sputter coater, and analyzed using a Zeiss Evo LS15 (UK) microscope, which worked at an accelerating voltage of 20 kV and at various amplifications. The SEM micrographs were used to calculate the mean particle size, which was estimated as the Feret’s diameter and measured as the mean of 300 particle measurements (n = 300).5 The moisture content of the spray-dried powders was determined by calculating the loss of weight upon drying from 25 to 105 °C, using an electronic moisture balance (Shimadzu, EB-280 MOC, Japan). The percentage of the initial weight that was lost during the heating process was ascribed to the moisture content of the powders. Measurements were carried out in triplicate (n = 3). 2.2.6. Structural Characterization of SLN-Loaded Microspheres Using CLSM. Inner structures of SLN-loaded microspheres were observed by CLSM, using an AOBS SP5X microscope (Leica GmbH, Germany), which collects images using different detectors for fluorescent signals upon irradiation with a white light laser. Small aliquots of the dry powders composed of coumarin-6-labeled SLN encapsulated in Bodipylabeled excipient (mannitol and trehalose) microspheres were placed on a glass slide, with an immersion oil drop to avoid the displacement of particles during their observation. Laser excitation wavelengths of 470 and 633 nm were used to scan the powder, and fluorescent emissions from coumarin-6 (emission λ = 480−555 nm) and Bodipy (emission λ = 650−700 nm) were collected using separate channels. Images were acquired with a magnification of 63× by using an oil immersion objective (HCX PL APO CS). The gray scale images obtained from each scan were pseudocolored green (coumarin-6) and red (Bodipy) and subsequently overlapped (LAS AF, Leica Confocal Software, Leica GmbH, Germany) to obtain a multicolored image. 2.2.7. Determination of Dry Powders Flow Properties. The real densities (real ρ) of the dry powders were measured using a helium pycnometer (AccuPyc 1330, Micrometrics Ltd., Dunstable, UK) at room temperature (n = 3). The apparent densities or tapped densities (apparent ρ) were determined under defined volumes before and after settling (Stav 2003, JEF Germany) a known weight of each dry powder in a 10 mL test tube after 1250 taps (n = 3). The flow properties of the dry powders were evaluated through the Carr’s index and Hausner ratio using tapped and bulk ρ values. The daer, Carr’s index, and Hausner ratio of spray-dried particles were calculated based on the following eqs 4, 5, and 6, respectively:

images were processed using procedures for plane-fit and flattened in the WSxM 4.0 Develop 11.4 software without any filtering. Surface topography and roughness were determined from 10 Å to 10 μm2 images. All experiments were performed at room temperature. 2.2.3. Determination of Association Efficiency and Drug Loading of RFB. Nonassociated RFB was separated from the SLN by size exclusion chromatography using Sephadex G-25/ PD-10 columns. Associated RFB to SLN was calculated after dissolving the SLN with acetonitrile followed by a centrifugation (30 min at 12 300 rpm) step, promoting the lipid phase precipitation. After centrifugation, the encapsulated RFB remained in the supernatant, and it was measured by UV− visible spectrophotometry, at λmax of 320 nm, using a microplate spectrophotometer reader (FLUOstar Omega, BMGLabtech, Germany). The supernatant of unloaded SLN was used as basic correction. All experiments were carried out in triplicate (n = 3). The encapsulation efficiency (EE) and drug loading (DL) of RFB were obtained using the following equations: EE (%) =

DL (%) =

Wloaded drug Wtotal drug Wloaded drug Wlipid

× 100 (1)

× 100 (2)

where Wtotal drug is the weight of total drug used, Wloaded drug is the weight of encapsulated drug (detected in the supernatant), and Wlipid is the weight of the lipid vehicle. 2.2.4. Preparation of Dry Powders Containing SLN. The RFB-SLN were suspended in aqueous solutions of the spraydrying excipients (20% and 22% w/v of mannitol and 15% and 17% w/v of trehalose, for glyceryl dibehenate and glyceryl tristearate SLN, respectively) at a 1:3 SLN:excipient mass ratio. Dry powders were obtained by spray-drying the suspensions (laboratory-scale spray-dryer; Büchi Mini spray-dryer, B-290, Switzerland) under the following conditions: two fluids external 0.7 mm nozzle; feed rate varying from 2.3−8.5 mL/min (depending on the spray-dried suspension); air flow of 400 L/h and inlet temperature (Tinlet) of 103 ± 2 °C. The aspirator rates were kept constant at 100% and 70% for SLN suspensions in mannitol and trehalose, respectively. These conditions resulted in an outlet temperature (Toutlet) of 63−70 °C. The produced powders were collected and stored in a desiccator at room temperature until further use. For CLSM studies, the SLN and excipients were previously stained with fluorescent labels to allow visualization. SLN were labeled with coumarin-6 during their preparation (40 μL of a coumarin-6 solution was added to the melted lipid). Mannitol and trehalose were stained with the fluorophore Bodipy, which was added to the spray-drying excipient solution (500 μL of a 1 mg/mL solution of Bodipy in DMSO) and kept under magnetic stirring for 2 h before spraydrying. The technique production yields (PY) were determined by gravimetry by comparing the weight of the resultant spraydrying powders (microspheres) (Wmicrospheres) and the weight of the initial total solids (RFB−SLN + spray-drying excipients) (Winitial total solids), as follows: PY (%) =

Wmicrospheres Winitial total solids

× 100

daer(μm) = dgeo

ρe λρs

(4)

where dgeo is the particle geometric diameter, ρs is 1 g/cm , ρe is the effective particle density in the same unit as ρs, and λ is the dynamic shape factor of the particle. This shape factor is theoretically defined as 1 for spherical particles. 3

Carr’s index (%) =

(3) C

100 × (V0 − Vf ) V0

(5)

DOI: 10.1021/acs.molpharmaceut.7b00169 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics Hausner ratio =

V0 Vf

(TSLI) (TSI, Copley Instruments, UK). Briefly, the TSLI equipment was divided in four impingement chambers, corresponding to (a) device + capsules (D+C); (b) mouth + throat (M+T); (c) medium compartment (MC), and (d) lung. According to the Ph. Eur. 9th Edition (2.9.18. Preparation for inhalation: aerodynamic assessment of fine particles), 7 and 30 mL of ethanol were added to MC and lung compartments, respectively. The content of each capsule was discharged through a flow rate of 60 ± 5 L/min during 15 s (5 s × 3). The lower compartment, that mimics the lung and consequently the alveoli, was schemed to have an effective mean daer particle cutoff size of ∼6.4 μm for an air flow rate of 60 L/min. The TSLI was disassembled, and all the compartments were carefully washed with ethanol and separately collected to conical flasks up to final volume of 50 mL. Samples were filtered and RFB content in each compartment determined by spectrophotometry using a microplate reader (FLUOstar Omega, BMG Labtech, Germany) at 320 nm (n = 3). 2.2.10. SLN Recovery from Microspheres. The microspheres were incubated in 10 mM isotonic PBS pH 7.4 with 0.1% of lung surfactant (Curosurf) under mild magnetic stirring at 37 °C for 90 min. After this time, the mean particle size and zeta potential of the recovered SLN were analyzed using a Zetasizer Nano S and Z (Malvern Instruments, UK), respectively, as described in detail in section 2.2.2. The morphology was visualized by transmission electron microscopy (TEM). The morphological analysis of SLN recovered from microspheres was conducted by TEM. SLN were stained with phosphotungstic acid at 2% (w/v) for 2 min and mounted on copper racks covered by a carbon membrane to allow the visualization on a JEOL Microscoper (JEM 2010, Japan) at 120 kV. The microphotographs were obtained using a Gatan Orius camera. 2.2.11. In Vitro RFB Release from Microencapsulated RFBLoaded SLN. RFB release from microspheres was evaluated by incubating the SLN (∼0.75 mg) and the microspheres (100 ± 5 mg) in a medium comprised by 10 mM isotonic PBS pH 7.4 with 0.1% of lung surfactant (Curosurf), under horizontal agitation at 37 °C. At suitable time intervals, samples were taken and centrifuged (high speed centrifuge; Allegra 64R centrifuge, Beckman Coulter) at 30 000 × g during 30 min at 4 °C. The amounts of released RFB in the supernatants were measured by spectrophotometry at 320 nm in a microplate reader (FLUOstar Omega, BMG Labtech, Germany) (n = 3). 2.2.12. In Vitro Activity of RFB Formulations against M. avium Strain DSMZ 44157 by MTS Assay. The susceptibility of M. avium strain to RFB formulations was assessed by adding 10 μL of inoculum at a concentration of 109 colony-forming units (CFU)/mL to 96-well plates containing the formulations diluted in 7H9 broth with albumin-dextrose-catalase supplement (six replicate samples). RFB formulations under study, at a concentration of 10 μg/mL, were added to the first well making serial two-fold dilutions up to 12 wells with 7H9 broth supplemented. Wells only with the supplemented 7H9 broth were used as negative controls. After inoculation, the plates were incubated for 10 days at 37 °C and 5% CO2 in a humidified atmosphere. After the incubation period, 20 μL of the MTS was added to all wells, plates were incubated at 37 °C for 3 h, and the absorbances were measured at 490 nm in a microplate reader Model 680 (Bio-Rad, CA, USA). 2.2.13. In Vivo Fate of RFB Formulations. 2.2.13.1. Biodistribution Studies. Mice received inhaled RFB formulations using a simple apparatus based on a 15 mL centrifuge tube where the dry powder was inserted. A small hole was formed in

(6)

where V0 is the unsettled apparent volume, and Vf is the final tapped volume. 2.2.8. Isothermal Titration Calorimetry (ITC) of SLN and Spray-Drying Excipients. To determine the energetics of the interactions of SLN and excipients during the spray-drying process, isothermal titration calorimetry measurements were performed using a VP-ITC titration microcalorimeter (MicroCal Inc., Northampton, MA) at 25 °C. In a typical experiment, the reference cell was filled with buffer solution, whereas the sampling one was filled with the excipient aqueous solution, that is, mannitol at 1160 or 1280 mM, and trehalose at 397 or 450 mM, for injection onto glyceryl dibehenate and glyceryl tristearate SLN nanoparticle solutions, respectively, and thermostatized at 25 °C. Each experiment consists of an initial 2 μL injection followed by 55 subsequent 5 μL injections. The nanoparticles suspension (64.7 and 31.4 mM for glyceryl dibehenate and glyceryl tristearate, respectively) was subsequently mixed at 286 rpm without foaming during the set of injections, which were spaced in 400 s intervals to enable the heat signal to return to equilibrium. The experimental recorded heats of interactions expressed as the injection heat normalized by the nanocarriers concentration added per each injection (Q*) were represented as a function of the SLN to excipient molar ratio. Control experiments, that is, titration of SLN solutions onto buffer ones, were performed to subtract potential heat sources derived from SLN dilution. All experiments were performed at least in duplicate, and the reproducibility was within ±3%.16 Experimental heat data were modeled according to the well-known two identical binding sites model using Affinimiter software (S4SD S.L., Spain), in which the experimental recorded heat per injection, Q*, can be expressed as: Q * = MV (n1Θ1ΔH1 + n2 Θ2ΔH2)

(7)

where M is the excipient concentration, V is the cell volume, ni is the binding stoichiometry, Θi is the fractional sites of macromolecule occupied by ligand, ΔHi is the interaction enthalpy, and the subindices 1 and 2 rise for the two sets of sites. From the former equation, the association constants (K1 and K2) can be expressed in terms of the occupied fractional sites and the unbound ligand concentration, [X], as follows: K1 =

Θ1 Θ2 and K 2 = (1 − Θ1)[X ] (1 − Θ2)[X ]

with [X ] = X − M(n1Θ1 + n2 Θ2)

(8)

In addition, the free energy (ΔGi) and entropy (ΔSi) changes upon SLN-excipient interactions can be derived from the modelization of experimental heat data through the wellknown thermodynamic relationships: ΔGi = − R gT ln K iΔGi = ΔHi − T ΔSi

(9)

where Rg is the universal gas constant. 2.2.9. In Vitro Deposition of Dry Powders Using a TwinStage Liquid Impinger. Hard capsule shells made of hydroxypropyl methylcellulose (HPMC) were manually filled up with 40.0 ± 0.2 mg of each dry powder and introduced in a Rotahaler device. The aerodynamic properties of inhalable dry powders were assessed through a twin-stage liquid impinger D

DOI: 10.1021/acs.molpharmaceut.7b00169 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Figure 1. AFM images of the optimized RFB-glyceryl dibehenate and RFB-glyceryl tristearate SLN: (A) height image and (B) cross-section height profiles performed to quantitatively measuring the width of the particles.

the bottom of the tube to allow the powder delivery to mouse. A manual pump connected to the upper part of the tube allowed the production of a turbulent air stream for fluidizing the powder. Each mouse was restrained in a 50 mL tube where a small hole in the bottom was made. This lower part of the 50 mL tube was connected to the 15 mL tube using a baby bottle teat. The tested formulation was RFB-glyceryl dibehenate SLN microencapsulated in mannitol and RFB-mannitol microspheres as control. Forty milligrams of each dry powder was put in the device, and each animal received the formulation by inhalation during 2 min. Fifteen and thirty minutes following administration (four animals per time point), mice were anesthetized with isoflurane, blood was collected from the orbital sinus into heparinized tubes, and it was stored at −30 °C. Animals were then sacrificed by cervical dislocation, and lung, spleen, and liver were removed and stored at −70 °C. 2.2.13.2. RFB Extraction from Blood and Tissues. RFB levels in blood and tissues were quantified by high-performance liquid chromatography (HPLC) following an extraction procedure previously described.17 In brief, 500 μL of blood was incubated with 250 μL of potassium dihydrogen phosphate buffer 0.05 M and sodium acetate 0.05 M (pH adjusted to 4.0 with acetic acid). The drug extraction was performed twice with 1 mL of a dichloromethane/isooctane mixture (2:3, v/v) under mechanical agitation for 15 min at room temperature, and centrifugation during 10 min at 1200 × g (Beckman Instruments, Inc.). Approximately 100 mg samples of spleen, lung, and liver tissues were also subjected to an extraction step with 2300 μL of dichloromethane/isooctane mixture by agitation for 30 min, followed by a centrifugation step as above-mentioned. The organic extracts were allowed to evaporate under a nitrogen stream. The precipitates were dissolved in 500 μL of mobile phase, filtered, and injected into the HPLC system. To evaluate the efficiency of the extraction method, a known amount of RFB was added to blood and

tissues removed from mice that did not receive RFB formulations and then subjected to the same previously described extraction method. 2.2.13.3. Determination of RFB by HPLC. Plasma and tissue levels of RFB were quantified by HPLC as previously reported using a Beckman equipment with autoinjector and a diodearray detector at 275 nm (Beckman Instruments, Inc.).17 A Purospher STAR RP-8 end-capped (5 μm) LiChroCART 250− 4.6 analytical column (Merck, Darmstadt, Germany) was used. The mobile phase was composed by 0.05 M potassium dihydrogen phosphate and 0.05 M sodium acetate (pH adjusted to 4.0 with acetic acid)-acetonitrile (53:47, v/v) with a flow rate of 1 mL/min at 25 °C. 2.2.13.4. Preparation of Standard Solutions for HPLC. The RFB standard solutions for HPLC were prepared as previously reported.17 Briefly, a stock RFB solution (100 μg/mL) was dissolved in the mobile phase, and from that, serial dilutions were done. Three seven-point calibration curves, ranging from 0.25−5.0 μg/mL with a loop of 100 μL, were prepared and enabled the antibiotic quantification in blood and tissues. A stock solution of 1.0 μg/mL was frozen and injected together with the analyzed samples to determine the precision of the HPLC concentrations of RFB in samples and controls from their peak area concentration response. 2.2.14. Biological Evaluation in a M. tuberculosis Infection Model: Experimental Infection, Treatment, and Bacterial Counts. Infection in BALB/c mice was induced by intravenous injection with 5 × 104 CFU of M. tuberculosis H37Rv. Treatment with RFB-loaded SLM microspheres started 2 weeks after infection induction, and mice received the tested formulation (RFB-glyceryl dibehenate SLN microencapsulated in mannitol) using the same apparatus described in section 2.2.13.1. Forty milligrams of dry powder was put in the device, and each mouse received the formulation by inhalation during 2 min. Administrations were performed in a total number of 10 E

DOI: 10.1021/acs.molpharmaceut.7b00169 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics Table 1. Physical and Aerodynamic Properties of SLN-Loaded Dry Powdersa formulations A B C D

process yield (%) 60.5 47.0 52.4 56.4

± ± ± ±

2.5 1.1 7.0 2.5

Feret’s diameter (μm, n = 300) 4.87 5.54 4.75 4.50

± ± ± ±

1.61 1.79 1.10 1.60

real ρ (g/cm3) 1.47 1.46 1.46 1.50

± ± ± ±

0.00 0.00 0.00 0.00

apparent ρ (g/cm3) 0.62 0.59 0.48 0.45

± ± ± ±

aerodynamic diameter (μm)

0.01 0.01 0.02 0.04

4.75 5.16 4.56 4.33

± ± ± ±

0.52 0.88 0.75 0.81

Carr’s index (%) 38.7 38.3 40.3 36.0

± ± ± ±

4.6 0.6 1.5 3.0

Hausner ratio 1.6 1.6 1.7 1.6

± ± ± ±

0.1 0.0 0.0 0.1

moisture residual content (%) 0.57 0.01 1.67 1.66

± ± ± ±

0.02 0.01 0.04 0.05

a

Glyceryl dibehenate SLN microencapsulated in mannitol (Formulation A), glyceryl tristearate SLN microencapsulated in mannitol (Formulation B), glyceryl dibehenate SLN microencapsulated in trehalose (Formulation C), and glyceryl tristearate SLN microencapsulated in trehalose (Formulation D) (mean ± SD, n = 3).

with similar average cross-sectional widths. Overall, particle sizes obtained using AFM images are similar to those established from the PCS analysis. The full characterization of these SLN has been reported elsewhere.15 3.2. Dry Powder Preparation and Characterization. Pulmonary administration of therapeutic SLN is an attractive concept. However, their direct aerosolization may result in poor delivery efficiency due to exhalation of the low-inertia nanoparticles.19−21 The inclusion of nanoparticles in a microparticulated system became a promising alternative to deliver them to the deep lung5,9 while avoiding the stability problems of liquid formulations. In this context, previously optimized RFB-loaded SLN were microencapsulated using a spray-drying process with two different inert and FDAapproved excipients, the polyol mannitol and the disaccharide trehalose, at a 1:3 SLN:excipient ratio. The used excipients are known to enhance aerosolization performance as well as to reduce moisture content. The SLN microencapsulation process was performed in a one-step spray-drying process after an initial screening of different SLN/excipient mass ratios, excipient concentrations and spray-drying conditions, seeking for the production of dry powders with the most suitable characteristics for pulmonary delivery. Hence, Tinlet, which in turn determines Toutlet, needs to be higher than the boiling point of water to evaporate the excipient drops, but it cannot be too high to avoid SLN melting. Therefore, Tinlet was established to be 103 °C. In addition, a screening of spray-drying conditions (e.g., aspirator and pump) using several mannitol and trehalose concentrations indicated that the optimum excipient concentration for glyceryl tristearate SLN (22% and 17% w/w for mannitol and trehalose, respectively) was higher than those used for glyceryl dibehenate SLN (20% and 15% w/w for mannitol and trehalose, respectively). This is due to the fact that glyceryl tristearate is a lipid that presents a lower m.p. than glyceryl dibehenate, and thus, SLN based on glyceryl tristearate require a larger quantity of excipient to protect them from the high temperatures during the spray-drying process. The spray-drying processes presented acceptable PY values (approximately 50% and higher, as can be seen in Table 1) for both SLN, in line with data reported elsewhere for spray-drying using similar excipients5,10 in spite of the limitation of the low spray-drying Tinlet used in this study. It should be emphasized that lab-scale spray-drying processes typically present lower PY values, in comparison to large-scale spray-drying processes, due to higher wall deposits since air residence times and radial distances from the atomizer to the drying chamber wall are shorter.22 Successful delivery of inhaled particles is governed by the deposition pattern. This is mainly controlled by particle size and density that strongly influence the dispersion and sedimentation properties of the powder.23,24 The daer along

for 2 weeks. Two days after the last administration, mice were sacrificed, and liver, spleen, and lung were aseptically extracted, homogenized, and diluted in 0.04% Tween 80 and incubated in Middlebrook 7H11 agar medium supplemented with OADC for CFU countings. One group of mice infected and nontreated was sacrificed at the beginning of treatment and another one at the end. The CFU counts in liver, spleen, and lung were also analyzed. Colonies were counted after 21 days of incubation at 37 °C with 5% CO2. The growth index was calculated from CFU counts using the difference between the log10 CFU at the end and at the beginning of treatment.18 2.2.15. Statistical Analysis. All the experiments were performed with at least three replicates. The data were expressed by mean ± SD and treated with the GraphPad Prism statistic software (GraphPad Software, San Diego, CA). The confidence interval was 95% (p < 0.05). The groups were compared by performing analysis of variance (one-way ANOVA) followed by Dunnet’s post hoc test, and the significant differences between groups were determined.

3. RESULTS AND DISCUSSION 3.1. SLN Formulation and Characterization. SLN have emerged as an alternative to liposomes and polymeric nanoparticles mainly due to their easy scale-up, good stability and biocompatibility, while also providing drug protection and sustained release.15 Herein, two different SLN formulations loaded with RFB were prepared using different lipids (glyceryl dibehenate and glyceryl tristearate), and Tween 80 as the surfactant component, using a hot HSH technique.15 Both SLN formulations presented particle size distributions within the nanometer range (108 ± 5 nm for RFB-glyceryl dibehenate SLN and 191 ± 7 nm for RFB-glyceryl tristearate SLN) with suitable polydispersity index values (