- Email: [email protected]
Characterization of a dextran-budesonide prodrug for inhalation therapy
Accepted Manuscript Characterization of A dextran-budesonide prodrug for inhalation therapy
Robert C. Waters, Günther Hochhaus PII: DOI: Reference:
S0928-0987(18)30529-3 https://doi.org/10.1016/j.ejps.2018.11.038 PHASCI 4772
To appear in:
European Journal of Pharmaceutical Sciences
Received date: Revised date: Accepted date:
17 July 2018 20 October 2018 29 November 2018
Please cite this article as: Robert C. Waters, Günther Hochhaus , Characterization of A dextran-budesonide prodrug for inhalation therapy. Phasci (2018), https://doi.org/10.1016/ j.ejps.2018.11.038
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Characterization of A Dextran-Budesonide Prodrug for Inhalation Therapy Robert C. Watersa and Günther Hochhausa
Department of Pharmaceutics, College of Pharmacy, University of Florida, Gainesville, FL
T
a
CR
IP
32610
US
KEYWORDS: Budesonide, dextran, prodrug, pharmacokinetics, asthma, esterase, modeling
*Address correspondence to:
PT
ED
M
AN
Günther Hochhaus University of Florida College of Pharmacy Department of Pharmaceutics 1600 SW Archer Road Gainesville, FL 32610 e-mail: [email protected]
AC
CE
Acknowledgment: This special issue is a wonderful way to recognize the scientific achievements of my colleague Hartmut Derendorf. Our careers strongly intertwined, having spent the last 30 years at the same institution and having collaborated on many projects. The development of the Department during his tenure was impressive.
1
ACCEPTED MANUSCRIPT ABSTRACT: Reducing the dosing frequency of corticosteroids may increase compliance and increase pulmonary targeting. The objective of this study was to evaluate whether a high molecular weight dextran-budesonide conjugate might be suitable for pulmonary slow release of the
T
otherwise fast absorbed budesonide. An array of dextran-spacer-budesonide conjugates was
IP
prepared that differed in the molecular weight of dextran (20 kDa or 40 kDa) and the length of
CR
the dicarboxylic spacer (succinic, glutaric, and adipic anhydride). The conjugates were characterized for identity by proton nuclear magnetic resonance (1H NMR) and Fourier-
US
transform infrared spectroscopy (FTIR), the degree of dextran-hydroxyl conjugation, purity, and
AN
physiological activation (release of budesonide). The 40 kDa dextran-succinate-budesonide conjugate was formulated as a dry powder for pulmonary delivery and characterized for particle
M
size distribution, particle morphology, and aerodynamic particle size. The degree of substitution
ED
(grams of budesonide in 100 grams of conjugate) ranged from 4 to 10% for all six dextranspacer-budesonide conjugates. Incubation at 37oC and pH 7.4 in phosphate buffered saline
PT
resulted in release of 25 -75% of the conjugated budesonide over an 8-hour period with the rate
CE
of release increasing with molecular weight of dextran and the length of the spacer. Modeling of the concentration time profiles of the released budesonide and budesonide-21-hemisucinate in
AC
phosphate buffered saline, suggested that budesonide is generated either directly or via the budesonide-21-hemisucinate pre-cursor. Data also suggested that the rate of budesonide generation likely depends on the position of budesonide on the dextran molecule. Spray-drying the 40 kDa dextran-succinate-budesonide produced respirable particles of the conjugate with a mass median aerodynamic particle size (MMAD) of 4 micrometers. The slow generation of
2
ACCEPTED MANUSCRIPT budesonide from the chemical delivery system might further improve the pharmacological
AC
CE
PT
ED
M
AN
US
CR
IP
T
profile of budesonide.
3
ACCEPTED MANUSCRIPT INTRODUCTION: Inhaled corticosteroids remain a cornerstone for the therapy of asthma. Budesonide is an orally inhaled non-halogenated glucocorticoid steroid for maintenance treatment of asthma that reduces the asthma related inflammation caused by cytokines, chemokines, enzymes, and cell
T
adhesion molecules (Hubner, 2005). The compliance to therapy, generally given twice a day,
IP
has been documented to be poor, especially in the elderly (Bozek, 2012). Reducing the dosing
CR
frequency to once a day by providing the drug throughout the day in a slow release fashion may increase compliance. Long pulmonary residence times after inhalation have also been shown to
US
be beneficial for improving pulmonary efficacy and targeting (Hochhaus, 1997).
AN
We were interested in designing slow release formulations of corticosteroids by applying chemical drug delivery approaches using budesonide, a corticosteroid absorbed relatively fast
M
into the systemic circulation (Hubner, 2005; Naikwade, 2009; Tunek, 1997) as a model drug.
ED
Budesonide dextran conjugates have been previously evaluated for targeted delivery to the GI tract, with the goal of releasing the corticosteroid to areas of the disease in patients with Crohn’s
PT
disease. While these prodrugs upon oral delivery are activated through dextranases enzyme
CE
present in the colonic microflora, a slow chemical activation has been described for the situation where the specific enzymes are not present. Budesonide dextran conjugates, because of the
AC
higher hydrophilicity and higher molecular weight is assumed to be only slowly absorbed and the prodrug should stay in the lung for a longer period of time. The prodrug providing a depot of inert drug should provide a slow pulmonary release of budesonide over time due to chemical hydrolysis of the ester bonds. High-molecular weight, hydrophilic dextran conjugates of budesonide were synthesized and their release of budesonide was investigated. A dextran conjugate of budesonide was
4
ACCEPTED MANUSCRIPT formulated for inhalation with lactose by spray-drying and the resultant physical characterization
AC
CE
PT
ED
M
AN
US
CR
IP
T
assessed.
5
ACCEPTED MANUSCRIPT MATERIALS AND METHODS:
Materials Budesonide was provided as a gift from Nanotherapeutics, Inc. (Alachua, FL). Dextran
T
(average molecular weight distribution 35 – 50 kDa) was obtained from Advance Scientific &
IP
Chemicals (Fort Lauderdale, FL). Dextran (average molecular weight distribution 15 – 25 kDa)
CR
was obtained from Sigma-Aldrich Company, Ltd. (St. Louis, MO). Succinic anhydride, glutaric anhydride, adipic anhydride, carbonyldiimidazole (1, 1’), 4-dimethylaminopyridine (4-DMAP),
US
sodium dodecyl sulfate (SDS), citric acid, potassium monobasic phosphate, potassium phosphate
AN
dibasic, acetone, dimethyl sulfoxide, ethyl acetate, phosphate buffered saline (10X), HPLC Grade acetonitrile and methanol, dimethyl sulfoxide-d6, hydrochloric acid, sodium hydroxide,
M
absolute ethanol, and triethylamine were obtained from Fisher Scientific (Pittsburgh, PA).
ED
Lactose monohydrate was obtained from DMV-Fonterra Excipients GmbH Co KG (Germany).
PT
Synthesis of budesonide-21-hemiester Conjugates (Intermediates)
CE
Budesonide-21-hemisuccinate (BHS) and budesonide-21-hemigluterate (BHG) were synthesized as previously described (McLeod, 1993; Pang, 2002; Varshosaz, 2009, 2010) with
AC
modifications to the purification process. Budesonide-21-hemiadipate (BHA) was synthesized with a novel solvent system for the first time. Budesonide was dissolved in either acetone (BHS and BHG synthesis) or a chloroform-based solvent (BHA synthesis) with the anhydride form of the spacer (e.g., succinic anhydride, glutaric anhydride, adipic anhydride), and the 4-DMAP catalyst. The spacer and catalyst were added in a proportion of 30% more on a molar basis than budesonide to approach a 100% reaction with budesonide. The reaction took place at room
6
ACCEPTED MANUSCRIPT temperature for two days (BHS and BHG synthesis) and five days (BHA synthesis). The solvents were then evaporated under vacuum at room temperature. The material was then dissolved in ethyl acetate, followed by washing with hydrochloric acid (0.1 N) using a separation flask. The ethyl acetate was then evaporated by vacuum and the materials were collected. The
T
molecular structure of the various budesonide conjugates is presented in Figure 1. The identity
CR
and Fourier-transform infrared spectroscopy (FTIR).
IP
of the budesonide conjugates was confirmed by proton nuclear magnetic resonance (1H NMR)
US
Synthesis of Budesonide-Spacer-Dextran Conjugates (Prodrug)
AN
The budesonide-21-hemisuccinate, budesonide-21-hemigluterate, and budesonide-21hemiadipate were conjugated with dextran (20 kDa and 40 kDa) in dimethyl sulfoxide (DMSO)
M
under the conditions in which 27 moles of the dextran-hydroxyl groups were available for
ED
conjugation per a single mole of budesonide-21-hemiester. For each single mole of budesonide21-hemiester, an equivalent of 1.8 moles of carbonyldiimidazole (CDI) and 21.5 moles of
PT
triethylamine (TEA) were added to catalyze the reaction. The reaction took place at room
CE
temperature for at least 15 hours. The molecular structure of the various prodrug molecules is presented in Figure 1. Each prodrug was characterized for identity by 1H-NMR and IR
AC
Spectroscopy.
Identity of Intermediates and Prodrugs Each intermediate (BHS, BHG, BHA) and prodrug (budesonide-spacer-dextran conjugate) was assessed for their 1-dimensional proton (1H) spectra by NMR and infrared spectra by FTIR. The synthesized materials were assessed by NMR with a 5 mm TXI
7
ACCEPTED MANUSCRIPT CryoProbe, Bruker Avance II 600 Console, Magnex 14.1 T/54 mm AS Magnet at 37ºC. The materials were dissolved in d-DMSO (75 µL) and transferred to an NMR tube. For FTIR, 1 mg of sample was mixed with 110 mg of IR grade potassium bromide. The generated pellet was subsequently analyzed in a Mercury-Cadmium-Telluride (MCT) detector (128 scans, at 4 cm-1
T
resolution). Post-processing included an atmosphere correction by subtracting the contributions
Drug Content in Conjugates (Degree of Substitution)
CR
IP
of water vapor and carbon dioxide.
US
The budesonide content of each prodrug, expressed as the degree of substitution (DS,
AN
grams of budesonide per 100 grams of conjugate, ) was determined by previously described methods (Mehvar, 1999; Varshosaz, 2009). Five milliliters of 0.1 N sodium hydroxide and 3 mL
M
of methanol were mixed with 5 mg of each conjugate and incubated at room temperature for 15
ED
hours to fully release the budesonide and/or the budesonide-21-hemiester (intermediate). A 125 µL aliquot of hydrochloric acid (0.1 N) was added to a 125 µL aliquot of sample and analyzed in
PT
triplicate. As under alkaline conditions, the budesonide-21-hemiester is fully converted to
CE
budesonide and budesonide is subsequently fully converted to 16-alpha hydroxyprednisolone. The 16-alpha hydroxyprednisolone was determined by reversed-phase high performance liquid
AC
chromatography (HPLC) as described below. A calibration curve was prepared with alkaline degraded budesonide.
Purity of Prodrugs To probe for potential impurities, including residual budesonide and budesonide-21hemiester, 5 mg of dextran-spacer-budesonide was dissolved in 10 mL of hydrochloric acid
8
ACCEPTED MANUSCRIPT (0.1N) and assayed immediately by HPLC (see below) after neutralization of the sample with hydrochloric acid (0.1 N). A calibration curve was prepared similarly with budesonide and the budesonide-21-hemiester.
T
Chemical Stability of Prodrugs
IP
The chemical stability of the synthesized prodrugs (budesonide-spacer-dextran conjugate)
CR
was assessed in phosphate buffer saline (pH 7.4) containing 0.5% SDS at 37ºC by monitoring the generation of budesonide and the budesonide-21-hemiester. The budesonide-spacer-dextran
US
conjugate sample, (equivalent to 450 µg of budesonide based on the DS) was dissolved in pre-
AN
heated PBS and then placed on a magnetic stir/heat plate at 300 RPM. The study was conducted in triplicate. Samples were taken at 15, 30, 45, 60, and 90 minutes and 2, 4, 6, 8, 24, 48, 72, 96,
M
and 144 hours. Removed samples (125 µl) were diluted with equal volumes of 0.1 N
ED
hydrochloric acid and assayed by HPLC.
PT
Model Identification and Selection
CE
The non-enzymatic activation pathways for the generation of budesonide were further characterized by modeling the concentration time profiles of generated budesonide and
AC
budesonide-21-hemisuccinate from the 40 kDa dextran-succinate-budesonide conjugate (BSD40). The models allowed the direct generation of budesonide from BSD40 as well as the generation of budesonide-21-hemisuccinate from BSD40, which was further allowed to be cleaved to budesonide. Budesonide was further allowed to degrade, presumably to 16-alpha hydroxyprednisolone (see Table 1 and Figure 7 for model structures; and Table 1 for differential equations describing the models). A second model allowed generation of budesonide from
9
ACCEPTED MANUSCRIPT BSD40 to occur with two different rates. Similarly, budesonide-21-hemisuccinate was allowed to be generated with two first order processes from BSD40. This model assumed that hydrolysis rates might differ dependent on the position at which budesonide was coupled to the dextran backbone. The models were described by a series of differential equations which were solved
T
through Laplace transforms (Muller, 1999) with parameter estimates obtained with the solver
IP
function of Excel®. The solved governing time differential equations are in terms of generation
CR
and degradation rate constants and the initial conjugated budesonide content. The models were compared for best describing the concentration time profiles of
US
budesonide and budesonide-21-hemisuccinate observed during the chemical hydrolysis of the
AN
BSD40. The budesonide degradation (k23) and budesonide-21-hemisuccinate degradation to budesonide (k42) were calculated by non-linear curve fitting (Figure 7).
M
The model selection criteria was based on the use of the Akaike Information Criterion
ED
(AIC). The AIC was calculated with the number of observations and the residual sum of squares
PT
(Muller, 1999).
CE
Spray-dryer Setup and Formulation A modified spray-dryer was designed to operate at processing temperatures below
AC
commercially available spray-dryers. Feed air was heated using heating coils as it flows through silicon tubing prior to entering the glass vortex chamber. The glass spray drying vessel was shaped so that the fed air near the bottom of the vessel is spiraled towards the narrower top which vents to a Seven Stage Anderson Cascade Impactor (ACI). A glass nebulizer (Meinhard, Model TR-30-A3, Type A) was fitted to the bottom of the glass vessel to feed the spray-drying solution.
10
ACCEPTED MANUSCRIPT The budesonide-21-hemisuccinate-dextran, 40 kDa (BSD40) material was prepared for spray-drying by dry-mixing 100 mg of BSD40 with 300 mg of lactose then dissolving in 20 mL of hydrochloric acid (0.001 N) and methanol (9:1 ratio). The solution feed rate was 30 mL of sample per hour. The vacuum flow rate, vortex flow rate, and nebulizer air pressure were 20
T
liters per minute, 18 liters per minute, and 0.3 liters per minutes, respectively. The spray-drying
IP
began when the in-process inlet temperature (vortex air) was above 70ºC. Solids were collected
US
Particle Morphology and Size
CR
and pooled from Stages 2-7 of the attached ACI.
AN
Neat budesonide, the BSD40, and the spray-dried BSD40 formulated materials were characterized for particle size and morphology using a Scanning Electron Microscopy JEOL
M
6700 with a cold field emission gun (FEG). The sample preparation included direct adhesion to
ED
a standard 1” Aluminum SEM mount and sputter coated with gold in a precision etching and coating system (PECS), Gatan Model 681.
PT
The spray-dried BSD40 material was characterized for laser diffraction particle size
CE
distribution. Samples were prepared in isopropyl alcohol and characterized with a Beckman Coulter LS 13 320 Laser Diffraction Analyzer. Measurement conditions included a circulation
AC
speed of 6, preferable transmittance (T%) between 70 and 95%, and a sampling time of 10 and 2 for the laser and lamp, respectively.
Particle Aerodynamics The particle aerodynamics, such as the Mass Median Aerodynamic Diameter (MMAD) and fine particle fraction (FPF, < 8.06 µm) of the spray-dried BSD40 was assessed using a Next
11
ACCEPTED MANUSCRIPT Generation Impactor (NGI). A nominal 25 mg sample of the test material (sample size of 3) was added to blue capsules for use in a Spiriva Handihaler. The inhaler was locked into the mouth piece of the NGI and a vacuum of 60 liters per minute was pulled through the Impactor. The capsule was pierced with the inhaler and the test run for 4.0 seconds. The powder from each
T
stage was collected by adding 5 mL of sodium hydroxide (0.1 N, sufficient to convert
IP
budesonide into 16-alpha hydroxyprednisolone) and 3 mL of methanol. The degraded
CR
budesonide, 16-alpha hydroxyprednisolone, was measured from each stage by reversed-phased high-performance liquid chromatography (HPLC) as described below. The measured 16-alpha
US
hydroxyprednisolone was correlated back to the BSD40 content by the degree of substitution
AN
value.
ED
M
HPLC Method
A standard reversed-phase HPLC method routinely employed by the group for the evaluation of
PT
corticosteroid alcohols and hemisuccinates (Rohdewald, 1985) and adapted for budesonide
CE
within an HPLC/RIA assay method (Hochhaus, 1998) was used for the determination of the 16alpha hydroxyprednisolone, budesonide, and the budesonide-21-hemiester content on an Agilent
AC
1100 Series HPLC with an ultraviolet detector and data integration HPCore ChemStation Software. Separation of the test analyte was performed using a Restek HPLC column (250 x 4.6 mm) packed with C18 Hypersil (5 µm). The mobile phase flow rate was 1 mL per minute and eluent detection was performed at 244 nm. The ratio of organic solvent (acetonitrile) to phosphate buffer (0.025 M KH2PO4, pH 3.2) in the mobile phase was adjusted between analyzing 16-alpha hydroxyprednisolone (45:55) versus budesonide and budesonide-21hemiester (55:45). Each sample injection was 50 µL. The retention times of the 16-alpha 12
ACCEPTED MANUSCRIPT hydroxyprednisolone, budesonide, budesonide-21-hemisuccinate, budesonide-21-hemigluterate, and budesonide-21-hemiadipate were 4.0 minutes, 5.2 minutes, 5.8 minutes, 6.2 minutes, and 6.8 minutes, respectively. Calibration curves (0.5-20 µg/ml) were linear with correlation factors
IP
T
better than 0.9999. Sample runs included quality control samples.
Statistical Analysis
CR
Experimental data was compared using a Student’s t-test assuming a two-tailed
US
distribution and unequal variance. The differences were considered significant when a P-value
AC
CE
PT
ED
M
AN
was below 0.05.
13
ACCEPTED MANUSCRIPT RESULTS:
Synthesis of Intermediates and Prodrugs Hemi-esters of budesonide with succinic and glutaric acid were synthesized using a
T
previously described method (Varshosaz, 2009, 2010). To the author’s knowledge, the chemical
IP
synthesis of budesonide to adipic anhydride has not been previously published. The chemical
CR
identity of each intermediate was characterized by 1H-NMR (see Figure 2 and supplemental material) and FTIR. The 1H-NMR profile of the BHS, BHG, and BHA intermediates showed the
US
presence of a carboxylic acid group with a broad singlet chemical shift at 12.1 ppm, 12.0 ppm,
AN
and 12.0 ppm for BHS, BHG, and BHA, respectively.
The FTIR absorption results (cm-1) for the three intermediates showed the presence of a
M
carboxylic acid group by a broad peak representing a carboxylic alcohol group centered at 3475,
ED
3435, and 3478 for BHS, BHG, and BHA, respectively (see FTIR spectra in supplemental material). Also present was a peak at 1730, 1728, and 1728 representing a C=O carboxylic acid
PT
bond for the BHS, BHG, and BHA, respectively. These results suggest that a carboxylic acid
CE
group was chemically conjugated to the budesonide molecule. The synthesized budesonide hemi-esters were conjugated to two dextran polymers of
AC
either 20 kDa or 40 kDa using previously published methods resulting in a total of 6 prodrugs. The 1H-NMR profile of all six prodrugs showed the absence of a carboxylic acid proton chemical shift, while an unsaturated carbon proton was presented as a doublet for the budesonide C2-H (δ 6.1 ppm) and C1-H (δ 7.3 ppm) and a singlet for the C4-H (δ 5.9 ppm) atom (Figure 3). The FTIR absorption results (cm-1) for the budesonide-21-hemisuccinate-dextran, 20 kDa conjugate (BSD20), the budesonide-21-hemisuccinate-dextran, 40 kDa conjugate (BSD40), the
14
ACCEPTED MANUSCRIPT budesonide-21-hemigluterate-dextran, 20 kDa conjugate (BGD20), the budesonide-21hemigluterate-dextran, 40 kDa conjugate (BGD40), the budesonide-21-hemiadipate-dextran, 20 kDa conjugate (BAD20), and the budesonide-21-hemiadipate-dextran, 40 kDa conjugate (BAD40) prodrugs showed that the carboxylic acid alcohol peak was transferred from 3475
T
(BHS), 3435 (BHG), and 3478 (BHA) to an alcoholic peak of 3386 (BSD20), 3410 (BSD40),
IP
3411 (BGD20), 3385 (BGD40), 3397 (BAD20), and 3388 (BAD40). The carbon oxygen double
CR
bond of the carboxylic group was changed from 1730 (BHS) and 1728 (BHG and BHA) to 1735 (BSD20, BSD40, BGD40, BAD20, and BAD40), and 1740 (BGD20), which is in the absorption
AN
molecules were chemically conjugated to dextran.
US
range of the ester carbonyl group. These results suggest that the budesonide-21-hemiester
M
Prodrug Purity, Degree of Dextran-Hydroxyl Conjugation, and Chemical Stability
ED
The purification process based on the precipitation and rinsing of the six budesonide prodrugs with hydrochloric acid resulted in greater than 99.2% (w/w) purity for all six prodrugs
PT
when products were tested for budesonide and budesonide-21-hemiester as impurities.
CE
The degree of substitution (DS, Figure 4) was assessed by subjecting each of the prodrugs to alkaline conditions to catalyze the release of budesonide. The DS is directly
AC
correlated to the degree of dextran-hydroxyl conjugation. The highest degree of substitution and dextran-hydroxyl conjugation was for the 40 kDa dextran-budesonide conjugate with adipic acid as a spacer (BAD40). Overall, the DS increased linearly from the lowest molecular weight spacer (succinate) to the higher molecular weight spacer (adipate) for both the 20 kDa dextran and 40 kDa dextran prodrugs. The one exception was for the 20 kDa dextran-budesonide conjugate with succinic acid as a spacer (BSD20), where the DS was greater than the BSD40
15
ACCEPTED MANUSCRIPT material (Figure 4). Differences in substitution shown in Figure 4 were statistically significant (p<0.05). The total budesonide released from all prodrugs in phosphate buffer (pH 7.4) at 37ºC after 8 hours indicated the prodrugs containing gluterate and adipate released budesonide to a
T
higher extent at the sampling time-point of 8 hours than the prodrugs containing succinate
IP
(Figure 5). A more detailed evaluation of the activation step was performed for the budesonide-
CR
21-hemisuccinate-dextran 40 kDa derivative, the prodrug with the best characteristics for pulmonary delivery (see discussion), by monitoring the formed budesonide-hemisuccinate and
US
budesonide at physiological pH over a period of 144 hours. This allowed monitoring of the
AN
budesonide and hemisuccinate ester generation and also the degradation of budesonide. Peak budesonide concentrations were observed after about 20 hours, similarly to the ones for
ED
M
budesonide-21-hemisuccinate (Figure 7).
Modeling of In vitro Chemical Stability Experiments
PT
Two budesonide prodrug activation models were evaluated for describing the
CE
concentration time profiles of budesonide esters and budesonide shown in Figure 7. These modeling exercises focused on the budesonide-21-hemisuccinate-dextran 40 kDa, the prodrug
AC
with the best characteristics for pulmonary delivery (see discussion). The initial rate constant estimates derived from incubations of the prodrug and various fragments in phosphate buffer (pH 7.4 and 37ºC) were used as a starting point for the least-squares non-linear curve fitting to the measured data. The non-linear (least squares) fitted concentration-time profiles of budesonide and budesonide-21-hemisuccinate for the mono-phasic model (Model 1) described the measured
16
ACCEPTED MANUSCRIPT data well over the first 8 hours but not over a longer period. The calculated non-linear (least squares) rate constants for the tested models are reported in Table 2. Evaluation of the models showed that a bi-phasic release of budesonide and budesonide21-hemisuccinate simulated the measured budesonide and budesonide-21-hemisuccinate release
T
closer according to compared concentration-time release profiles and AIC values than a single
IP
phasic model (as tested in the model determination). The fitted non-linear concentration-time
CR
profile was plotted with the measured budesonide and budesonide-21-hemisuccinate data, Figure 7. The simulated plots showed that the measured budesonide and budesonide-21-hemisuccinate
US
data could be predicted best with the bi-phasic model approach. The goodness of fit showed that
M
Formulation for Pulmonary Delivery
AN
the Model 2 had a lower AIC of -31.0 versus 56.7 for Model 1, Table 2.
ED
Particle size distribution and morphology: The budesonide prodrug was tested for particle size and morphology by SEM prior to spray-drying to obtain a physical reference of the
PT
particles. The budesonide prodrug particles (BSD40) were non-porous linear particles. The
CE
spray-dried budesonide prodrug resulted in a visually narrow particle size distribution with smooth solid spherical particles (Figure 6). The spray-dried budesonide prodrug particles had a
µm (d90).
AC
normal particle size distribution of 1.22 ± 0.04 µm (d10), 2.74 ± 0.06 µm (d50), and 6.44 ± 0.10
NGI Analysis: The aerosol dispersion performance of the spray-dried budesonide prodrug was tested using the Spiriva Handihaler through an NGI. The spray-dried material was characteristic of a dry-powder aerosol by quantifiable particle deposition on each stage of the NGI Impactor (Stage 1-7). The highest deposition was on Stage 2, which represents an
17
ACCEPTED MANUSCRIPT aerodynamic particle size of 4.46 µm. The aerosol performance parameters include the emitted dose (93.5 ± 2.4%), fine particle fraction (47.3 ± 17.9%), MMAD (4.04 ± 1.42 µm), and the geometric standard deviation (3.97 ± 3.37). The MMAD of the spray-dried budesonide prodrug
AC
CE
PT
ED
M
AN
US
CR
IP
T
formulation was less than 5 µm which is efficient for deep lung deposition.
18
ACCEPTED MANUSCRIPT DISCUSSION:
Budesonide is absorbed faster from the lung than other inhaled corticosteroids (Thorsson, 2001). Pharmacokinetic (PK) and pharmacodynamic (PD) simulations revealed that a slow
T
absorption of pulmonary deposited drug increases pulmonary selectivity (Hochhaus, 1997).
IP
Commercial preparations exhibiting slow pulmonary removal have utilized slow dissolution
CR
processes (fluticasone propionate, mometasone furoate) or extensive binding to lung components (long acting beta-2-adrenergic drugs) to achieve prolonged residence in the lung. We employed
US
here an alternative approach, that is based on the ability of high molecular weight dextran’s to be
AN
retained in the lung, while serving as a slow release depot for covalently attached drug. Dextran was selected as a high molecular weight backbone as it is non-toxic, non-immunogenic and
M
contains an abundance of hydroxyl groups available for conjugation with the primary 21-
ED
hydroxyl group of corticosteroids, if the appropriate spacers are used. The high molecular weight hydrophilic dextran backbone is the reason for a low permeability across cell membranes
PT
(Pang, 2007). Mehvar et al. demonstrated for dextran’s in the molecular range of 1000-40,000
CE
Dalton’s that the pulmonary absorption rate is inversely related to the molecular weight while a further increase did not further reduce the permeability (Mehvar, 1992). We included 20 kDa
AC
and 40 kDa dextran’s in conjunction with a range of spacers to provide a range of budesonide conjugates. We did not use dextran derivatives above the molecular weight above 40 kDa to ensure that the derivatives are cleared renally. As spacer length has been shown to affect chemical and enzymatic stability (Penugonda, 2008; Varshosaz, 2009, 2010, 2011), we included a range of spacers, namely succinic, glutaric, and adipic acid as spacers.
19
ACCEPTED MANUSCRIPT The chemical synthesis was a two-step process where first the intermediate product (budesonide-21-hemiester) was synthesized in the presence of a 4-dimethylaminopyridine (DMAP) catalyst. Using a standard method of conjugation (McLeod, 1993; Pang, 2002; Varshosaz, 2009, 2010), BHS and BHG intermediates were synthesized in acetone, while for the
T
budesonide hemi-adipate (BHA) synthesis, chloroform was used to prevent precipitation
IP
observed in acetone. The second part of the reaction was completed in DMSO, using standard
CR
conditions, and produced a gummy material that was further dried by vacuum drying to produce a dry crystalline material. NMR and FTIR confirmed the identity of the conjugates and HPLC
US
analysis suggested sufficient purity for the following studies.
AN
The amount of budesonide that was incorporated into the conjugate (degree of substitution, expressed as grams budesonide per 100 grams of conjugate) ranged from 4-10%
M
dependent on the backbone and spacer used (Figure 4). This agreed with values reported for
ED
other corticosteroid dextran conjugates (Mehvar, 2000; Rensberger, 2000; Penugonda, 2008). Overall, a higher degree of substitution was observed for the 40 kDa dextran backbone (Figure
PT
4). With one outlier representing BSD20, the degree of substitution was inversely proportional
CE
to the molecular weight of the spacer, similar to findings of Penugonda and collaborators for methylprednisolone dextran conjugates (Penugonda, 2008). Based on this substitution ratio for a
AC
dose of 50 µg of budesonide 500 µg of derivative might have to be inhaled. Dextran prodrugs can either be activated by pH dependent chemical ester hydrolysis or enzymatic esterase cleavage, e.g. by dextranases (McLeod, 1993). Preliminary experiments (data not shown) showed that carboxylesterase 1, the predominant esterase in the lung (Koitka, 2008; Zhang, 2014) and lung homogenate were hydrolyzing with similar smaller rates than pH dependent chemical ester hydrolysis in agreement with the CES1 enzyme typically binding to
20
ACCEPTED MANUSCRIPT substrates with small alcohol groups and large acyl groups (Hosokawa, 2008; Redinbo, 2003; Shimizu, 2014; Suzaki, 2013; Wang, 2011; Williams, 2011). This study therefore concentrated on the chemical activation of the prodrugs. Hydrolysis of the ester bond directly linking the drug molecule with the spacer results in
T
the generation of the pharmacologically active budesonide. Alternatively, chemical hydrolysis
IP
might cleave the ester connecting the spacer with the dextran backbone. In this case, the drug-
CR
21-hemiester (e.g. budesonide-21-hemiester) will be generated first followed by subsequent hydrolysis of the 21-hemi ester to budesonide (Larsen 1989). It cannot be assumed that
US
budesonide is stable in the incubation medium, therefore a slow degradation of budesonide had
AN
to be assumed. For example, the chemical in vitro degradation of budesonide results in an acetal splitting at the carbon 16 and 17 position leading to 16-alpha hydroxyprednisolone (Edsbacker,
M
1987).
ED
The activation of the various budesonide prodrugs, first evaluated over an 8-hour period at pH 7.4 and 37ºC, indicated that the total budesonide released increased as the spacer and
PT
dextran molecular weight increased (Figure 5). Similar results were reported by Penugonda for
CE
dextran-methylprednisolone conjugates (Penugonda, 2008). To better understand the complex degradation pathways after chemical hydrolysis, we
AC
applied structured multi-compartmental kinetic models to describe the time profiles of appearance and disappearance of budesonide and budesonide-21-hemiesters for the 40 kDa budesonide-succinate-dextran derivative (BSD40). We evaluated an array of potential models. Among these, a stepwise, sequential model (prodrug to budesonide hemisuccinate to budesonide to degradation prodrug) was unable to describe the data. Considering the nature of the prodrug using dicarboxylic ester functions for
21
ACCEPTED MANUSCRIPT coupling dextran and budesonide made it likely that both ester functions could be cleaved through hydrolysis generating budesonide-hemisuccinate or budesonide. Two models (Figure 7) evaluated the budesonide and budesonide-21-hemisuccinate concentration-time profiles. Both models allowed the parallel generation of budesonide with direct release and the budesonide-21-
T
hemisuccinate converting into budesonide, with the generated budesonide being slowly degraded
IP
over the 144-hour time-period. Considering the rates of generation of budesonide and its
CR
hemisuccinate (Table 2, rate constants in the range of 0.2 to 0.6 h-1, Model 2), at this time point all of the budesonide had long been released from the prodrug. Model 1 described the activation
US
of the prodrug with a single rate of release of budesonide and budesonide-21-hemisuccinate from
AN
the dextran backbone. Model 2 described the activation of the prodrug with two release rates for the generation of budesonide and budesonide-21-hemisuccinate from the dextran backbone (bi-
M
phasic approach). The bi-phasic approach (Model 2) concentration-time profile fit the measured
ED
data better with a lower AIC value. Whether, the position of budesonide conjugation on the dextran polymer may play a role in the bi-phasic model fit needs further investigations.
PT
Under physiological conditions (pH 7.4, 37°C) the prodrug with the smaller spacer (e.g.,
CE
succinic anhydride) had a relatively slower activation over the glutaric and adipic anhydride budesonide prodrugs (about 2-3 times slower, see Figure 5) and was therefore selected for
AC
incorporation into inhalable microspheres. The application of an inhaled corticosteroid dextran prodrug for asthma inhalation therapy could alter the relatively fast systemic absorption typically seen of a lipophilic steroid, such as budesonide, allowing for steady release of the active drug, hence improving its effectiveness. The budesonide-spacer-dextran prodrug has the potential of escaping mucociliary clearance due to its hydrophilicity.
22
ACCEPTED MANUSCRIPT To formulate the budesonide prodrug for lung deposition, the budesonide prodrug was spray-dried with lactose to produce respirable aerodynamic spherical particles. The lactose carrier when spray-dried was amorphous (data not shown) and dissolved upon disposition in the lung, thereby delivering the prodrug while reducing the effect of mucociliary clearance. Spray-
T
drying can influence particle morphology by the drying rate, the surface tension, the solubility of
IP
the drug in the feeding solution, and the viscosity of the feeding solution (Park, 2013). The
CR
modified spray-dryer was fabricated to collect the dried material onto an ACI where the particles are separated by size. The materials were collected and pooled from Stage 2-7 of the ACI for the
US
spray-dried budesonide prodrug with lactose materials and were characteristic of a normal and
AN
narrow (Span 1.9) particle size distribution with smooth surfaces (Figure 6). Particles having an aerodynamic particle size diameter between 5 and 10 microns will
M
enter the lungs and deposit by gravitational settling (sedimentation) predominantly in the middle-
ED
to-upper lung regions (Park, 2013). Particles with an aerodynamic particle size diameter of 2 to 5 microns can efficiently target the middle-to-lower lung regions whereas particles having an
PT
aerodynamic particle size less than 2 microns can efficiently target the smaller airways and deep
CE
lung region (Park, 2013). The MMAD results for the spray-dried budesonide prodrug were within the respirable particle size of less than 5 µm. The spray-drying process performed within
AC
an aqueous environment will require further optimization to prevent partial activation of the prodrug during the spray-drying process. Unfortunately, further optimization of the procedure was not possible as we ran out of prodrug. It is, however, very feasible that during the spraydrying procedure the spray-drying medium was not acetic enough to prevent partial ester hydrolysis at the elevated temperature.
23
ACCEPTED MANUSCRIPT CONCLUSIONS: Spray-drying a dextran based budesonide prodrug with lactose can produce respirable size particles for asthma inhalation therapy. The slow generation of budesonide from the chemical delivery system might further improve the pharmacological profile of budesonide. A
T
parallel activation of a budesonide prodrug can simulate the observed concentration-time profile
IP
of released budesonide and budesonide-21-hemisuccinate with simultaneous non-linear (least
CR
squares) curve fitting. To fully prevent degradation of the prodrug during spray-drying, the
AC
CE
PT
ED
M
AN
US
process will require further optimization.
24
ACCEPTED MANUSCRIPT CONFLICT OF INTEREST DECLARATION The authors declare that they have no competing interests
Acknowledgments: This work was in part supported by a Department of Pharmaceutics
AC
CE
PT
ED
M
AN
US
CR
IP
T
unrestricted research funds.
25
ACCEPTED MANUSCRIPT REFERENCES: Bozek, A., Warkocka-Szoltysek, B., Filipowska-Gronska, A., Jarzab, J., 2012. Montelukast as an Add-on Therapy to Inhaled Corticosteroids in the Treatment of Severe Asthma in Elderly Patients. Journal of Asthma 49, 530-534.
T
Chanteux, H., Rosa, M., Delatour, C., Prakash, C., Smith, S., Nicolas, J., 2014. In Vitro Hydrolysis and Transesterification of CDP323, an alpha4beta1/alpha4beta7 Integrin Antagonist Ester Prodrug. Drug Metabolism and Disposition 42, 153-161.
CR
IP
Edsbacker, S., Andersson, P., Lindberg, C., Ryrfeldt, A., Thalen, A., 1987. Metabolic Acetal Splitting of Budesonide. Drug Metabolism and Disposition 15, 412-417.
US
Hochhaus G, Möllmann H, Derendorf H, Gonzalez-Rothi RJ. Pharmacokinetic /pharmacodynamic aspects of aerosol therapy using glucocorticoids as a model. J Clin Pharmacol. 1997;37:881–92.
AN
Hochhaus G, Froehlich P, Hochhaus R, Möllmann A, Derendorf, Möllmann H. A selective HPLC/RIA for the determination of budesonide, J. Pharm. Biomed. Anal. 17, 1235-1242 (1998)
M
Hosokawa, M., 2008. Structure and Catalytic Properties of Carboxylesterase Isozymes Involved in Metabolic Activation of Prodrugs. Molecules 13, 412-431.
ED
Hubner, M., Hochhaus, G., Derendorf, H., 2005. Comparative Pharmacology, Bioavailability, Pharmacokinetics, and Pharmacodynamics of Inhaled Glucocorticosteroids. Immunology Allergy and Clinics of North America 25, 469-488.
CE
PT
Jung, Y.J., Lee, J.S., Kim, H.H., Kim, Y.T., Kim, Y.M., 1998. Synthesis and Properties of Dextran-5-aminosalicylic Acid Ester as a Potential Colon-specific Prodrug of 5-Aminosalicylic Acid. Arch. Pharm. Res. 21, 179-186.
AC
Koitka, M., Hochel, J., Obst, D., Rottmann, A., Gieschen, H., Borchert, H., 2008. Determination of rat serum esterase activities by an HPLC method using S-acetylthiocholine iodide and pnitrophenyl acetate. Analytical Biochemistry 381, 113-122. Larsen, C., 1989. Dextran prodrugs - structure and stability in relation to therapeutic activity. Advanced Drug Delivery Reviews 3, 103-154. Lee, Y., Kim, I.H., Kim, J., Yoon, J., Shin, Y.H., Jung, Y., Kim, Y.M., 2011. Evaluation of dextran-flufenamic acid ester as a polymeric colon-specific prodrug of flufenamic acid, an antiinflammatory drug, for chronotherapy. Journal of Drug Targeting 19, 336-343. McLeod, A.D., Friend, D.R., Tozer, T.N., 1993. Synthesis and chemical stability of glucocorticoid-dextran esters: potential prodrugs for colon-specific delivery. International Journal of Pharmaceutics 92, 105-114.
26
ACCEPTED MANUSCRIPT Mehvar, R., Shepard, T.L., 1992. Molecular-weight-dependent pharmacokinetics of fluoresceinlabeled dextrans in rats. J. Pharm. Sci. 81 (9), 908-912. Mehvar, R., 1999. Simultaneous analysis of dextran-methylprednisolone succinate, methylprednisolone succinate, and methylprednisolone by size-exclusion chromatography. Journal of Pharmaceutical and Biomedical Analysis 19, 785-792.
IP
T
Mehvar, R., Dann, R.O., Hoganson, D.A., 2000. Kinetics of hydrolysis of dextranmethylprednisolone succinate, a macromolecular prodrug of methylprednisolone, in rat blood and liver lysosomes. J. Control Release, 68(1): 53-61.
CR
Muller, S., Hutson, A., Arya, V., Hochhaus, G., 1999. Assessment of Complex Peptide Degradation Pathways via Structured Multicompartmental Modeling Approaches: The Metabolism of Dynorphin A1-13 and Related Fragments in Human Plasma. Journal of Pharmaceutical Sciences 88, 938-944.
AN
US
Naikwade, S.R., Bajaj, A.N., Gurav, P., Gatne, M.M., Soni, P.S., 2009. Development of Budesonide Microparticles Using Spray-Drying Technology for Pulmonary Administration: Design, Characterization, In Vitro Evaluation, and In Vivo Efficacy Study. American Association of Pharmaceutical Scientists 10, 993-1012.
ED
M
Naikwade, S.R., Bajaj, A.N., Gurav, P., Gatne, M.M., Soni, P.S., 2009. Development of Budesonide Microparticles Using Spray-Drying Technology for Pulmonary Administration: Design, Characterization, In Vitro Evaluation, and In Vivo Efficacy Study. American Association of Pharmaceutical Scientists 10, 993-1012.
PT
Pang, Y., Zhang, Y., Zhang, Z., 2002. Synthesis of an enzyme-dependent prodrug and evaluation of its potential for colon targeting. World Journal of Gastroenterology 8, 913-917.
CE
Pany, Y., Sakagami, M., Byron, PR., 2007. Insulin self-association: effects on lung disposition kinetics in the airways of the isolated perfused rat lung (IPRL). Pharmaceutical Research 24(9), 1636-44.
AC
Penugonda, S., Kumar, A., Agarwal, H.K., Parang, K., Mehvar, R., 2008. Synthesis and in vitro characeterization of novel dextran-methylprednisolone conjugates with peptide linkers: effects of linker length on hydrolytic and enzymatic release of methylprednisolone and its peptidyl intermediates. J. Pharm. Sci. 97(7), 2649-2664. Piao, H., Cho, H., Oh, E., Chung, S., Shim, C., Kim, D., 2009. Budesonide Microemulsions for Enhancing Solubility and Dissolution Rate. Journal of Korian Pharmaceutical Sciences 39, 417422. Redinbo, M.R., Bencharit, S., Potter, P.M., 2003. Human carboxylesterase 1: from drug metabolism to drug discovery. Biochemical Society Transactions 31, 620-624.
27
ACCEPTED MANUSCRIPT Rohdewald P, Rehder J, Drehsen G, Hochhaus G, Derendorf H, Möllmann H. Simultaneous determination of glucocorticoid alcohols, their succinates and hydrocortisone in plasma. Journal of pharmaceutical and biomedical analysis. 1985;3:566–73. Shimizu, M., Fukami, T., Ito, Y., Kurokawa, T., Kariya, M., Nakajima, M., Yokoi, T., 2014. Indiplon is Hydrolyzed by Arylacetamide Deacetylase in Human Liver. Drug Metabolism and Disposition 42, 751-758.
CR
IP
T
Suzaki, Y., Uemura, N., Takada, M., Ohyama, T., Itohda, A., Morimoto, T., Imai, H., Hamasaki, H., Inano, A., Hosokawa, M., Tateishi, M., Ohashi, K., 2013. The effect of carboxylesterase 1 (CES1) polymorphisms on the pharmacokinetics of oseltamivir in humans. Eur. J. Clin. Pharmacol. 69, 21-30.
US
Thorsson, L., Edsbacker, S., Kallen, A., Lofdahl, C-G., 2001. Pharmacokinetics and systemic activity of fluticasone via Diskus® and pMDI, and of budesonide via Turbuhaler®. British J. Clin. Pharmcol. 52 (5), 529-538.
AN
Tunek, A., Sjodin, K., Hallstrom, G., 1997. Reversible Formation of Fatty Acid Esters of Budesonide, an Antiasthma Glucocorticoid, in Human Lung and Liver Microsomes. Drug Metabolism and Disposition 25, 1311-1317.
ED
M
Varshosaz, J., Emami, J., Ahmadi, F., Tavakoli, N., Minaiyan, M., Fassihi, A., Mahzouni, P., Dorkoosh, F., 2011. Preparation of budesonide-dextran conjugates using glutarate spacer as a colon-targeted drug delivery system: in vitro / in vivo evaluation in induced ulcerative colitis. Journal of Drug Targeting 19, 140-153.
CE
PT
Varshosaz, J., Emami, J., Fassihi, A., Tavakoli, N., Minaiyan, M., Ahmadi, F., Mahzouni, P., Dorkoosh, F., 2010. Effectiveness of budesonide-succinate-dextran conjugate as a novel prodrug of budesonide against acetic acid-induced colitis in rats. International Journal of Colorectal Disease 25, 1159-1165.
AC
Varshosaz, J., Emami, J., Tavakoli, N., Fassihi, A., Minaiyan, M., Ahmadi, F., Dorkoosh, F., 2009. Synthesis and evaluation of dextran-budesonide conjugates as colon specific prodrugs for treatment of ulcerative colitis. International Journal of Pharmaceutics 365, 68-75. Varshosaz, J., Emami, J., Tavakoli, N., Minaiyan, M., Rahmani, N., Dorkoosh, F., 2012. Development and Evaluation of a Novel Pellet-Based Tablet System for Potential Colon Delivery of Budesonide. Journal of Drug Delivery 2012, 1-7. Wang, J., Williams, E.T., Bourgea, J., Wong, Y.N., Patten, C.J., 2011. Characterization of Recombinant Human Carboxylesterases: Fluorescein Diacetate as a Probe Substrate for Human Carboxylesterase 2. Drug Metabolism and Disposition 39, 1329-1333. Williams, E.T., Bacon, J.A., Bender, D.M., Lowinger, J.J., Guo, W., Ehsani, M.E., Wang, X., Wang, H., Qian, Y., Ruterbories, K.J., Wrighton, S.A., Perkins, E.J., 2011. Characterization of
28
ACCEPTED MANUSCRIPT the Expression and Activity of Carboxylesterases 1 and 2 from the Beagle Dog, Cynomolgus Monkey, and Human. Drug Metabolism and Disposition 39, 2305-2313.
AC
CE
PT
ED
M
AN
US
CR
IP
T
Zhang, C., Xu, Y., Zhong, Q., Li, X., Gao, P., Feng, C., Chu, Q., Chen, Y., Liu, D., 2014. In Vitro Evaluation of the Inhibitory Potential of Pharmaceutical Excipients on Human Carboxylesterase 1A and 2. PLOS ONE 9, e93819.
29
ACCEPTED MANUSCRIPT TABLES Table 1
Kinetic models describing the activation pathway of the budesonide prodrug. Budesonide prodrug (P1), budesonide (P2), 16-alpha hydroxyprednisolone (P3), and budesonide-21-hemisuccinate (P4) are represented in the model flow diagrams. The differential equations, assuming first order processes are also shown for each model. Model 2 splits the prodrug pool into two independent fractions (P1a and P1b) with separate hydrolysis constants. Governing Time Differential Equation
IP
T
Model
ED
M
AN
US
CR
Model 1
AC
CE
PT
Model 2
30
ACCEPTED MANUSCRIPT Degrading and generating rate constants from a budesonide prodrug as determined by a non-linear (least-squares) curve-fitting (free float all k) method using a single-phase (Model 1) and bi-phasic (Model 2) prodrug activation model approach. The model fit was determined by using the Akaike Information Criterion (AIC) for the combined release of budesonide B(t) and budesonide-21hemisuccinate L(t).
T
Model 2 0.2076 0.0517 0.6211 0.0163 0.0023 0.0108 -31.0
AC
CE
PT
ED
M
AN
US
Model 1 0.0841 N/A 0.1192 N/A 0.0050 0.0117 56.7
IP
Rate Constant (h-1) k12a k12b k14a k14b k42 k23 AIC
CR
Table 2
31
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Robert C. Waters, Günther Hochhaus PII: DOI: Reference:
S0928-0987(18)30529-3 https://doi.org/10.1016/j.ejps.2018.11.038 PHASCI 4772
To appear in:
European Journal of Pharmaceutical Sciences
Received date: Revised date: Accepted date:
17 July 2018 20 October 2018 29 November 2018
Please cite this article as: Robert C. Waters, Günther Hochhaus , Characterization of A dextran-budesonide prodrug for inhalation therapy. Phasci (2018), https://doi.org/10.1016/ j.ejps.2018.11.038
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Characterization of A Dextran-Budesonide Prodrug for Inhalation Therapy Robert C. Watersa and Günther Hochhausa
Department of Pharmaceutics, College of Pharmacy, University of Florida, Gainesville, FL
T
a
CR
IP
32610
US
KEYWORDS: Budesonide, dextran, prodrug, pharmacokinetics, asthma, esterase, modeling
*Address correspondence to:
PT
ED
M
AN
Günther Hochhaus University of Florida College of Pharmacy Department of Pharmaceutics 1600 SW Archer Road Gainesville, FL 32610 e-mail: [email protected]
AC
CE
Acknowledgment: This special issue is a wonderful way to recognize the scientific achievements of my colleague Hartmut Derendorf. Our careers strongly intertwined, having spent the last 30 years at the same institution and having collaborated on many projects. The development of the Department during his tenure was impressive.
1
ACCEPTED MANUSCRIPT ABSTRACT: Reducing the dosing frequency of corticosteroids may increase compliance and increase pulmonary targeting. The objective of this study was to evaluate whether a high molecular weight dextran-budesonide conjugate might be suitable for pulmonary slow release of the
T
otherwise fast absorbed budesonide. An array of dextran-spacer-budesonide conjugates was
IP
prepared that differed in the molecular weight of dextran (20 kDa or 40 kDa) and the length of
CR
the dicarboxylic spacer (succinic, glutaric, and adipic anhydride). The conjugates were characterized for identity by proton nuclear magnetic resonance (1H NMR) and Fourier-
US
transform infrared spectroscopy (FTIR), the degree of dextran-hydroxyl conjugation, purity, and
AN
physiological activation (release of budesonide). The 40 kDa dextran-succinate-budesonide conjugate was formulated as a dry powder for pulmonary delivery and characterized for particle
M
size distribution, particle morphology, and aerodynamic particle size. The degree of substitution
ED
(grams of budesonide in 100 grams of conjugate) ranged from 4 to 10% for all six dextranspacer-budesonide conjugates. Incubation at 37oC and pH 7.4 in phosphate buffered saline
PT
resulted in release of 25 -75% of the conjugated budesonide over an 8-hour period with the rate
CE
of release increasing with molecular weight of dextran and the length of the spacer. Modeling of the concentration time profiles of the released budesonide and budesonide-21-hemisucinate in
AC
phosphate buffered saline, suggested that budesonide is generated either directly or via the budesonide-21-hemisucinate pre-cursor. Data also suggested that the rate of budesonide generation likely depends on the position of budesonide on the dextran molecule. Spray-drying the 40 kDa dextran-succinate-budesonide produced respirable particles of the conjugate with a mass median aerodynamic particle size (MMAD) of 4 micrometers. The slow generation of
2
ACCEPTED MANUSCRIPT budesonide from the chemical delivery system might further improve the pharmacological
AC
CE
PT
ED
M
AN
US
CR
IP
T
profile of budesonide.
3
ACCEPTED MANUSCRIPT INTRODUCTION: Inhaled corticosteroids remain a cornerstone for the therapy of asthma. Budesonide is an orally inhaled non-halogenated glucocorticoid steroid for maintenance treatment of asthma that reduces the asthma related inflammation caused by cytokines, chemokines, enzymes, and cell
T
adhesion molecules (Hubner, 2005). The compliance to therapy, generally given twice a day,
IP
has been documented to be poor, especially in the elderly (Bozek, 2012). Reducing the dosing
CR
frequency to once a day by providing the drug throughout the day in a slow release fashion may increase compliance. Long pulmonary residence times after inhalation have also been shown to
US
be beneficial for improving pulmonary efficacy and targeting (Hochhaus, 1997).
AN
We were interested in designing slow release formulations of corticosteroids by applying chemical drug delivery approaches using budesonide, a corticosteroid absorbed relatively fast
M
into the systemic circulation (Hubner, 2005; Naikwade, 2009; Tunek, 1997) as a model drug.
ED
Budesonide dextran conjugates have been previously evaluated for targeted delivery to the GI tract, with the goal of releasing the corticosteroid to areas of the disease in patients with Crohn’s
PT
disease. While these prodrugs upon oral delivery are activated through dextranases enzyme
CE
present in the colonic microflora, a slow chemical activation has been described for the situation where the specific enzymes are not present. Budesonide dextran conjugates, because of the
AC
higher hydrophilicity and higher molecular weight is assumed to be only slowly absorbed and the prodrug should stay in the lung for a longer period of time. The prodrug providing a depot of inert drug should provide a slow pulmonary release of budesonide over time due to chemical hydrolysis of the ester bonds. High-molecular weight, hydrophilic dextran conjugates of budesonide were synthesized and their release of budesonide was investigated. A dextran conjugate of budesonide was
4
ACCEPTED MANUSCRIPT formulated for inhalation with lactose by spray-drying and the resultant physical characterization
AC
CE
PT
ED
M
AN
US
CR
IP
T
assessed.
5
ACCEPTED MANUSCRIPT MATERIALS AND METHODS:
Materials Budesonide was provided as a gift from Nanotherapeutics, Inc. (Alachua, FL). Dextran
T
(average molecular weight distribution 35 – 50 kDa) was obtained from Advance Scientific &
IP
Chemicals (Fort Lauderdale, FL). Dextran (average molecular weight distribution 15 – 25 kDa)
CR
was obtained from Sigma-Aldrich Company, Ltd. (St. Louis, MO). Succinic anhydride, glutaric anhydride, adipic anhydride, carbonyldiimidazole (1, 1’), 4-dimethylaminopyridine (4-DMAP),
US
sodium dodecyl sulfate (SDS), citric acid, potassium monobasic phosphate, potassium phosphate
AN
dibasic, acetone, dimethyl sulfoxide, ethyl acetate, phosphate buffered saline (10X), HPLC Grade acetonitrile and methanol, dimethyl sulfoxide-d6, hydrochloric acid, sodium hydroxide,
M
absolute ethanol, and triethylamine were obtained from Fisher Scientific (Pittsburgh, PA).
ED
Lactose monohydrate was obtained from DMV-Fonterra Excipients GmbH Co KG (Germany).
PT
Synthesis of budesonide-21-hemiester Conjugates (Intermediates)
CE
Budesonide-21-hemisuccinate (BHS) and budesonide-21-hemigluterate (BHG) were synthesized as previously described (McLeod, 1993; Pang, 2002; Varshosaz, 2009, 2010) with
AC
modifications to the purification process. Budesonide-21-hemiadipate (BHA) was synthesized with a novel solvent system for the first time. Budesonide was dissolved in either acetone (BHS and BHG synthesis) or a chloroform-based solvent (BHA synthesis) with the anhydride form of the spacer (e.g., succinic anhydride, glutaric anhydride, adipic anhydride), and the 4-DMAP catalyst. The spacer and catalyst were added in a proportion of 30% more on a molar basis than budesonide to approach a 100% reaction with budesonide. The reaction took place at room
6
ACCEPTED MANUSCRIPT temperature for two days (BHS and BHG synthesis) and five days (BHA synthesis). The solvents were then evaporated under vacuum at room temperature. The material was then dissolved in ethyl acetate, followed by washing with hydrochloric acid (0.1 N) using a separation flask. The ethyl acetate was then evaporated by vacuum and the materials were collected. The
T
molecular structure of the various budesonide conjugates is presented in Figure 1. The identity
CR
and Fourier-transform infrared spectroscopy (FTIR).
IP
of the budesonide conjugates was confirmed by proton nuclear magnetic resonance (1H NMR)
US
Synthesis of Budesonide-Spacer-Dextran Conjugates (Prodrug)
AN
The budesonide-21-hemisuccinate, budesonide-21-hemigluterate, and budesonide-21hemiadipate were conjugated with dextran (20 kDa and 40 kDa) in dimethyl sulfoxide (DMSO)
M
under the conditions in which 27 moles of the dextran-hydroxyl groups were available for
ED
conjugation per a single mole of budesonide-21-hemiester. For each single mole of budesonide21-hemiester, an equivalent of 1.8 moles of carbonyldiimidazole (CDI) and 21.5 moles of
PT
triethylamine (TEA) were added to catalyze the reaction. The reaction took place at room
CE
temperature for at least 15 hours. The molecular structure of the various prodrug molecules is presented in Figure 1. Each prodrug was characterized for identity by 1H-NMR and IR
AC
Spectroscopy.
Identity of Intermediates and Prodrugs Each intermediate (BHS, BHG, BHA) and prodrug (budesonide-spacer-dextran conjugate) was assessed for their 1-dimensional proton (1H) spectra by NMR and infrared spectra by FTIR. The synthesized materials were assessed by NMR with a 5 mm TXI
7
ACCEPTED MANUSCRIPT CryoProbe, Bruker Avance II 600 Console, Magnex 14.1 T/54 mm AS Magnet at 37ºC. The materials were dissolved in d-DMSO (75 µL) and transferred to an NMR tube. For FTIR, 1 mg of sample was mixed with 110 mg of IR grade potassium bromide. The generated pellet was subsequently analyzed in a Mercury-Cadmium-Telluride (MCT) detector (128 scans, at 4 cm-1
T
resolution). Post-processing included an atmosphere correction by subtracting the contributions
Drug Content in Conjugates (Degree of Substitution)
CR
IP
of water vapor and carbon dioxide.
US
The budesonide content of each prodrug, expressed as the degree of substitution (DS,
AN
grams of budesonide per 100 grams of conjugate, ) was determined by previously described methods (Mehvar, 1999; Varshosaz, 2009). Five milliliters of 0.1 N sodium hydroxide and 3 mL
M
of methanol were mixed with 5 mg of each conjugate and incubated at room temperature for 15
ED
hours to fully release the budesonide and/or the budesonide-21-hemiester (intermediate). A 125 µL aliquot of hydrochloric acid (0.1 N) was added to a 125 µL aliquot of sample and analyzed in
PT
triplicate. As under alkaline conditions, the budesonide-21-hemiester is fully converted to
CE
budesonide and budesonide is subsequently fully converted to 16-alpha hydroxyprednisolone. The 16-alpha hydroxyprednisolone was determined by reversed-phase high performance liquid
AC
chromatography (HPLC) as described below. A calibration curve was prepared with alkaline degraded budesonide.
Purity of Prodrugs To probe for potential impurities, including residual budesonide and budesonide-21hemiester, 5 mg of dextran-spacer-budesonide was dissolved in 10 mL of hydrochloric acid
8
ACCEPTED MANUSCRIPT (0.1N) and assayed immediately by HPLC (see below) after neutralization of the sample with hydrochloric acid (0.1 N). A calibration curve was prepared similarly with budesonide and the budesonide-21-hemiester.
T
Chemical Stability of Prodrugs
IP
The chemical stability of the synthesized prodrugs (budesonide-spacer-dextran conjugate)
CR
was assessed in phosphate buffer saline (pH 7.4) containing 0.5% SDS at 37ºC by monitoring the generation of budesonide and the budesonide-21-hemiester. The budesonide-spacer-dextran
US
conjugate sample, (equivalent to 450 µg of budesonide based on the DS) was dissolved in pre-
AN
heated PBS and then placed on a magnetic stir/heat plate at 300 RPM. The study was conducted in triplicate. Samples were taken at 15, 30, 45, 60, and 90 minutes and 2, 4, 6, 8, 24, 48, 72, 96,
M
and 144 hours. Removed samples (125 µl) were diluted with equal volumes of 0.1 N
ED
hydrochloric acid and assayed by HPLC.
PT
Model Identification and Selection
CE
The non-enzymatic activation pathways for the generation of budesonide were further characterized by modeling the concentration time profiles of generated budesonide and
AC
budesonide-21-hemisuccinate from the 40 kDa dextran-succinate-budesonide conjugate (BSD40). The models allowed the direct generation of budesonide from BSD40 as well as the generation of budesonide-21-hemisuccinate from BSD40, which was further allowed to be cleaved to budesonide. Budesonide was further allowed to degrade, presumably to 16-alpha hydroxyprednisolone (see Table 1 and Figure 7 for model structures; and Table 1 for differential equations describing the models). A second model allowed generation of budesonide from
9
ACCEPTED MANUSCRIPT BSD40 to occur with two different rates. Similarly, budesonide-21-hemisuccinate was allowed to be generated with two first order processes from BSD40. This model assumed that hydrolysis rates might differ dependent on the position at which budesonide was coupled to the dextran backbone. The models were described by a series of differential equations which were solved
T
through Laplace transforms (Muller, 1999) with parameter estimates obtained with the solver
IP
function of Excel®. The solved governing time differential equations are in terms of generation
CR
and degradation rate constants and the initial conjugated budesonide content. The models were compared for best describing the concentration time profiles of
US
budesonide and budesonide-21-hemisuccinate observed during the chemical hydrolysis of the
AN
BSD40. The budesonide degradation (k23) and budesonide-21-hemisuccinate degradation to budesonide (k42) were calculated by non-linear curve fitting (Figure 7).
M
The model selection criteria was based on the use of the Akaike Information Criterion
ED
(AIC). The AIC was calculated with the number of observations and the residual sum of squares
PT
(Muller, 1999).
CE
Spray-dryer Setup and Formulation A modified spray-dryer was designed to operate at processing temperatures below
AC
commercially available spray-dryers. Feed air was heated using heating coils as it flows through silicon tubing prior to entering the glass vortex chamber. The glass spray drying vessel was shaped so that the fed air near the bottom of the vessel is spiraled towards the narrower top which vents to a Seven Stage Anderson Cascade Impactor (ACI). A glass nebulizer (Meinhard, Model TR-30-A3, Type A) was fitted to the bottom of the glass vessel to feed the spray-drying solution.
10
ACCEPTED MANUSCRIPT The budesonide-21-hemisuccinate-dextran, 40 kDa (BSD40) material was prepared for spray-drying by dry-mixing 100 mg of BSD40 with 300 mg of lactose then dissolving in 20 mL of hydrochloric acid (0.001 N) and methanol (9:1 ratio). The solution feed rate was 30 mL of sample per hour. The vacuum flow rate, vortex flow rate, and nebulizer air pressure were 20
T
liters per minute, 18 liters per minute, and 0.3 liters per minutes, respectively. The spray-drying
IP
began when the in-process inlet temperature (vortex air) was above 70ºC. Solids were collected
US
Particle Morphology and Size
CR
and pooled from Stages 2-7 of the attached ACI.
AN
Neat budesonide, the BSD40, and the spray-dried BSD40 formulated materials were characterized for particle size and morphology using a Scanning Electron Microscopy JEOL
M
6700 with a cold field emission gun (FEG). The sample preparation included direct adhesion to
ED
a standard 1” Aluminum SEM mount and sputter coated with gold in a precision etching and coating system (PECS), Gatan Model 681.
PT
The spray-dried BSD40 material was characterized for laser diffraction particle size
CE
distribution. Samples were prepared in isopropyl alcohol and characterized with a Beckman Coulter LS 13 320 Laser Diffraction Analyzer. Measurement conditions included a circulation
AC
speed of 6, preferable transmittance (T%) between 70 and 95%, and a sampling time of 10 and 2 for the laser and lamp, respectively.
Particle Aerodynamics The particle aerodynamics, such as the Mass Median Aerodynamic Diameter (MMAD) and fine particle fraction (FPF, < 8.06 µm) of the spray-dried BSD40 was assessed using a Next
11
ACCEPTED MANUSCRIPT Generation Impactor (NGI). A nominal 25 mg sample of the test material (sample size of 3) was added to blue capsules for use in a Spiriva Handihaler. The inhaler was locked into the mouth piece of the NGI and a vacuum of 60 liters per minute was pulled through the Impactor. The capsule was pierced with the inhaler and the test run for 4.0 seconds. The powder from each
T
stage was collected by adding 5 mL of sodium hydroxide (0.1 N, sufficient to convert
IP
budesonide into 16-alpha hydroxyprednisolone) and 3 mL of methanol. The degraded
CR
budesonide, 16-alpha hydroxyprednisolone, was measured from each stage by reversed-phased high-performance liquid chromatography (HPLC) as described below. The measured 16-alpha
US
hydroxyprednisolone was correlated back to the BSD40 content by the degree of substitution
AN
value.
ED
M
HPLC Method
A standard reversed-phase HPLC method routinely employed by the group for the evaluation of
PT
corticosteroid alcohols and hemisuccinates (Rohdewald, 1985) and adapted for budesonide
CE
within an HPLC/RIA assay method (Hochhaus, 1998) was used for the determination of the 16alpha hydroxyprednisolone, budesonide, and the budesonide-21-hemiester content on an Agilent
AC
1100 Series HPLC with an ultraviolet detector and data integration HPCore ChemStation Software. Separation of the test analyte was performed using a Restek HPLC column (250 x 4.6 mm) packed with C18 Hypersil (5 µm). The mobile phase flow rate was 1 mL per minute and eluent detection was performed at 244 nm. The ratio of organic solvent (acetonitrile) to phosphate buffer (0.025 M KH2PO4, pH 3.2) in the mobile phase was adjusted between analyzing 16-alpha hydroxyprednisolone (45:55) versus budesonide and budesonide-21hemiester (55:45). Each sample injection was 50 µL. The retention times of the 16-alpha 12
ACCEPTED MANUSCRIPT hydroxyprednisolone, budesonide, budesonide-21-hemisuccinate, budesonide-21-hemigluterate, and budesonide-21-hemiadipate were 4.0 minutes, 5.2 minutes, 5.8 minutes, 6.2 minutes, and 6.8 minutes, respectively. Calibration curves (0.5-20 µg/ml) were linear with correlation factors
IP
T
better than 0.9999. Sample runs included quality control samples.
Statistical Analysis
CR
Experimental data was compared using a Student’s t-test assuming a two-tailed
US
distribution and unequal variance. The differences were considered significant when a P-value
AC
CE
PT
ED
M
AN
was below 0.05.
13
ACCEPTED MANUSCRIPT RESULTS:
Synthesis of Intermediates and Prodrugs Hemi-esters of budesonide with succinic and glutaric acid were synthesized using a
T
previously described method (Varshosaz, 2009, 2010). To the author’s knowledge, the chemical
IP
synthesis of budesonide to adipic anhydride has not been previously published. The chemical
CR
identity of each intermediate was characterized by 1H-NMR (see Figure 2 and supplemental material) and FTIR. The 1H-NMR profile of the BHS, BHG, and BHA intermediates showed the
US
presence of a carboxylic acid group with a broad singlet chemical shift at 12.1 ppm, 12.0 ppm,
AN
and 12.0 ppm for BHS, BHG, and BHA, respectively.
The FTIR absorption results (cm-1) for the three intermediates showed the presence of a
M
carboxylic acid group by a broad peak representing a carboxylic alcohol group centered at 3475,
ED
3435, and 3478 for BHS, BHG, and BHA, respectively (see FTIR spectra in supplemental material). Also present was a peak at 1730, 1728, and 1728 representing a C=O carboxylic acid
PT
bond for the BHS, BHG, and BHA, respectively. These results suggest that a carboxylic acid
CE
group was chemically conjugated to the budesonide molecule. The synthesized budesonide hemi-esters were conjugated to two dextran polymers of
AC
either 20 kDa or 40 kDa using previously published methods resulting in a total of 6 prodrugs. The 1H-NMR profile of all six prodrugs showed the absence of a carboxylic acid proton chemical shift, while an unsaturated carbon proton was presented as a doublet for the budesonide C2-H (δ 6.1 ppm) and C1-H (δ 7.3 ppm) and a singlet for the C4-H (δ 5.9 ppm) atom (Figure 3). The FTIR absorption results (cm-1) for the budesonide-21-hemisuccinate-dextran, 20 kDa conjugate (BSD20), the budesonide-21-hemisuccinate-dextran, 40 kDa conjugate (BSD40), the
14
ACCEPTED MANUSCRIPT budesonide-21-hemigluterate-dextran, 20 kDa conjugate (BGD20), the budesonide-21hemigluterate-dextran, 40 kDa conjugate (BGD40), the budesonide-21-hemiadipate-dextran, 20 kDa conjugate (BAD20), and the budesonide-21-hemiadipate-dextran, 40 kDa conjugate (BAD40) prodrugs showed that the carboxylic acid alcohol peak was transferred from 3475
T
(BHS), 3435 (BHG), and 3478 (BHA) to an alcoholic peak of 3386 (BSD20), 3410 (BSD40),
IP
3411 (BGD20), 3385 (BGD40), 3397 (BAD20), and 3388 (BAD40). The carbon oxygen double
CR
bond of the carboxylic group was changed from 1730 (BHS) and 1728 (BHG and BHA) to 1735 (BSD20, BSD40, BGD40, BAD20, and BAD40), and 1740 (BGD20), which is in the absorption
AN
molecules were chemically conjugated to dextran.
US
range of the ester carbonyl group. These results suggest that the budesonide-21-hemiester
M
Prodrug Purity, Degree of Dextran-Hydroxyl Conjugation, and Chemical Stability
ED
The purification process based on the precipitation and rinsing of the six budesonide prodrugs with hydrochloric acid resulted in greater than 99.2% (w/w) purity for all six prodrugs
PT
when products were tested for budesonide and budesonide-21-hemiester as impurities.
CE
The degree of substitution (DS, Figure 4) was assessed by subjecting each of the prodrugs to alkaline conditions to catalyze the release of budesonide. The DS is directly
AC
correlated to the degree of dextran-hydroxyl conjugation. The highest degree of substitution and dextran-hydroxyl conjugation was for the 40 kDa dextran-budesonide conjugate with adipic acid as a spacer (BAD40). Overall, the DS increased linearly from the lowest molecular weight spacer (succinate) to the higher molecular weight spacer (adipate) for both the 20 kDa dextran and 40 kDa dextran prodrugs. The one exception was for the 20 kDa dextran-budesonide conjugate with succinic acid as a spacer (BSD20), where the DS was greater than the BSD40
15
ACCEPTED MANUSCRIPT material (Figure 4). Differences in substitution shown in Figure 4 were statistically significant (p<0.05). The total budesonide released from all prodrugs in phosphate buffer (pH 7.4) at 37ºC after 8 hours indicated the prodrugs containing gluterate and adipate released budesonide to a
T
higher extent at the sampling time-point of 8 hours than the prodrugs containing succinate
IP
(Figure 5). A more detailed evaluation of the activation step was performed for the budesonide-
CR
21-hemisuccinate-dextran 40 kDa derivative, the prodrug with the best characteristics for pulmonary delivery (see discussion), by monitoring the formed budesonide-hemisuccinate and
US
budesonide at physiological pH over a period of 144 hours. This allowed monitoring of the
AN
budesonide and hemisuccinate ester generation and also the degradation of budesonide. Peak budesonide concentrations were observed after about 20 hours, similarly to the ones for
ED
M
budesonide-21-hemisuccinate (Figure 7).
Modeling of In vitro Chemical Stability Experiments
PT
Two budesonide prodrug activation models were evaluated for describing the
CE
concentration time profiles of budesonide esters and budesonide shown in Figure 7. These modeling exercises focused on the budesonide-21-hemisuccinate-dextran 40 kDa, the prodrug
AC
with the best characteristics for pulmonary delivery (see discussion). The initial rate constant estimates derived from incubations of the prodrug and various fragments in phosphate buffer (pH 7.4 and 37ºC) were used as a starting point for the least-squares non-linear curve fitting to the measured data. The non-linear (least squares) fitted concentration-time profiles of budesonide and budesonide-21-hemisuccinate for the mono-phasic model (Model 1) described the measured
16
ACCEPTED MANUSCRIPT data well over the first 8 hours but not over a longer period. The calculated non-linear (least squares) rate constants for the tested models are reported in Table 2. Evaluation of the models showed that a bi-phasic release of budesonide and budesonide21-hemisuccinate simulated the measured budesonide and budesonide-21-hemisuccinate release
T
closer according to compared concentration-time release profiles and AIC values than a single
IP
phasic model (as tested in the model determination). The fitted non-linear concentration-time
CR
profile was plotted with the measured budesonide and budesonide-21-hemisuccinate data, Figure 7. The simulated plots showed that the measured budesonide and budesonide-21-hemisuccinate
US
data could be predicted best with the bi-phasic model approach. The goodness of fit showed that
M
Formulation for Pulmonary Delivery
AN
the Model 2 had a lower AIC of -31.0 versus 56.7 for Model 1, Table 2.
ED
Particle size distribution and morphology: The budesonide prodrug was tested for particle size and morphology by SEM prior to spray-drying to obtain a physical reference of the
PT
particles. The budesonide prodrug particles (BSD40) were non-porous linear particles. The
CE
spray-dried budesonide prodrug resulted in a visually narrow particle size distribution with smooth solid spherical particles (Figure 6). The spray-dried budesonide prodrug particles had a
µm (d90).
AC
normal particle size distribution of 1.22 ± 0.04 µm (d10), 2.74 ± 0.06 µm (d50), and 6.44 ± 0.10
NGI Analysis: The aerosol dispersion performance of the spray-dried budesonide prodrug was tested using the Spiriva Handihaler through an NGI. The spray-dried material was characteristic of a dry-powder aerosol by quantifiable particle deposition on each stage of the NGI Impactor (Stage 1-7). The highest deposition was on Stage 2, which represents an
17
ACCEPTED MANUSCRIPT aerodynamic particle size of 4.46 µm. The aerosol performance parameters include the emitted dose (93.5 ± 2.4%), fine particle fraction (47.3 ± 17.9%), MMAD (4.04 ± 1.42 µm), and the geometric standard deviation (3.97 ± 3.37). The MMAD of the spray-dried budesonide prodrug
AC
CE
PT
ED
M
AN
US
CR
IP
T
formulation was less than 5 µm which is efficient for deep lung deposition.
18
ACCEPTED MANUSCRIPT DISCUSSION:
Budesonide is absorbed faster from the lung than other inhaled corticosteroids (Thorsson, 2001). Pharmacokinetic (PK) and pharmacodynamic (PD) simulations revealed that a slow
T
absorption of pulmonary deposited drug increases pulmonary selectivity (Hochhaus, 1997).
IP
Commercial preparations exhibiting slow pulmonary removal have utilized slow dissolution
CR
processes (fluticasone propionate, mometasone furoate) or extensive binding to lung components (long acting beta-2-adrenergic drugs) to achieve prolonged residence in the lung. We employed
US
here an alternative approach, that is based on the ability of high molecular weight dextran’s to be
AN
retained in the lung, while serving as a slow release depot for covalently attached drug. Dextran was selected as a high molecular weight backbone as it is non-toxic, non-immunogenic and
M
contains an abundance of hydroxyl groups available for conjugation with the primary 21-
ED
hydroxyl group of corticosteroids, if the appropriate spacers are used. The high molecular weight hydrophilic dextran backbone is the reason for a low permeability across cell membranes
PT
(Pang, 2007). Mehvar et al. demonstrated for dextran’s in the molecular range of 1000-40,000
CE
Dalton’s that the pulmonary absorption rate is inversely related to the molecular weight while a further increase did not further reduce the permeability (Mehvar, 1992). We included 20 kDa
AC
and 40 kDa dextran’s in conjunction with a range of spacers to provide a range of budesonide conjugates. We did not use dextran derivatives above the molecular weight above 40 kDa to ensure that the derivatives are cleared renally. As spacer length has been shown to affect chemical and enzymatic stability (Penugonda, 2008; Varshosaz, 2009, 2010, 2011), we included a range of spacers, namely succinic, glutaric, and adipic acid as spacers.
19
ACCEPTED MANUSCRIPT The chemical synthesis was a two-step process where first the intermediate product (budesonide-21-hemiester) was synthesized in the presence of a 4-dimethylaminopyridine (DMAP) catalyst. Using a standard method of conjugation (McLeod, 1993; Pang, 2002; Varshosaz, 2009, 2010), BHS and BHG intermediates were synthesized in acetone, while for the
T
budesonide hemi-adipate (BHA) synthesis, chloroform was used to prevent precipitation
IP
observed in acetone. The second part of the reaction was completed in DMSO, using standard
CR
conditions, and produced a gummy material that was further dried by vacuum drying to produce a dry crystalline material. NMR and FTIR confirmed the identity of the conjugates and HPLC
US
analysis suggested sufficient purity for the following studies.
AN
The amount of budesonide that was incorporated into the conjugate (degree of substitution, expressed as grams budesonide per 100 grams of conjugate) ranged from 4-10%
M
dependent on the backbone and spacer used (Figure 4). This agreed with values reported for
ED
other corticosteroid dextran conjugates (Mehvar, 2000; Rensberger, 2000; Penugonda, 2008). Overall, a higher degree of substitution was observed for the 40 kDa dextran backbone (Figure
PT
4). With one outlier representing BSD20, the degree of substitution was inversely proportional
CE
to the molecular weight of the spacer, similar to findings of Penugonda and collaborators for methylprednisolone dextran conjugates (Penugonda, 2008). Based on this substitution ratio for a
AC
dose of 50 µg of budesonide 500 µg of derivative might have to be inhaled. Dextran prodrugs can either be activated by pH dependent chemical ester hydrolysis or enzymatic esterase cleavage, e.g. by dextranases (McLeod, 1993). Preliminary experiments (data not shown) showed that carboxylesterase 1, the predominant esterase in the lung (Koitka, 2008; Zhang, 2014) and lung homogenate were hydrolyzing with similar smaller rates than pH dependent chemical ester hydrolysis in agreement with the CES1 enzyme typically binding to
20
ACCEPTED MANUSCRIPT substrates with small alcohol groups and large acyl groups (Hosokawa, 2008; Redinbo, 2003; Shimizu, 2014; Suzaki, 2013; Wang, 2011; Williams, 2011). This study therefore concentrated on the chemical activation of the prodrugs. Hydrolysis of the ester bond directly linking the drug molecule with the spacer results in
T
the generation of the pharmacologically active budesonide. Alternatively, chemical hydrolysis
IP
might cleave the ester connecting the spacer with the dextran backbone. In this case, the drug-
CR
21-hemiester (e.g. budesonide-21-hemiester) will be generated first followed by subsequent hydrolysis of the 21-hemi ester to budesonide (Larsen 1989). It cannot be assumed that
US
budesonide is stable in the incubation medium, therefore a slow degradation of budesonide had
AN
to be assumed. For example, the chemical in vitro degradation of budesonide results in an acetal splitting at the carbon 16 and 17 position leading to 16-alpha hydroxyprednisolone (Edsbacker,
M
1987).
ED
The activation of the various budesonide prodrugs, first evaluated over an 8-hour period at pH 7.4 and 37ºC, indicated that the total budesonide released increased as the spacer and
PT
dextran molecular weight increased (Figure 5). Similar results were reported by Penugonda for
CE
dextran-methylprednisolone conjugates (Penugonda, 2008). To better understand the complex degradation pathways after chemical hydrolysis, we
AC
applied structured multi-compartmental kinetic models to describe the time profiles of appearance and disappearance of budesonide and budesonide-21-hemiesters for the 40 kDa budesonide-succinate-dextran derivative (BSD40). We evaluated an array of potential models. Among these, a stepwise, sequential model (prodrug to budesonide hemisuccinate to budesonide to degradation prodrug) was unable to describe the data. Considering the nature of the prodrug using dicarboxylic ester functions for
21
ACCEPTED MANUSCRIPT coupling dextran and budesonide made it likely that both ester functions could be cleaved through hydrolysis generating budesonide-hemisuccinate or budesonide. Two models (Figure 7) evaluated the budesonide and budesonide-21-hemisuccinate concentration-time profiles. Both models allowed the parallel generation of budesonide with direct release and the budesonide-21-
T
hemisuccinate converting into budesonide, with the generated budesonide being slowly degraded
IP
over the 144-hour time-period. Considering the rates of generation of budesonide and its
CR
hemisuccinate (Table 2, rate constants in the range of 0.2 to 0.6 h-1, Model 2), at this time point all of the budesonide had long been released from the prodrug. Model 1 described the activation
US
of the prodrug with a single rate of release of budesonide and budesonide-21-hemisuccinate from
AN
the dextran backbone. Model 2 described the activation of the prodrug with two release rates for the generation of budesonide and budesonide-21-hemisuccinate from the dextran backbone (bi-
M
phasic approach). The bi-phasic approach (Model 2) concentration-time profile fit the measured
ED
data better with a lower AIC value. Whether, the position of budesonide conjugation on the dextran polymer may play a role in the bi-phasic model fit needs further investigations.
PT
Under physiological conditions (pH 7.4, 37°C) the prodrug with the smaller spacer (e.g.,
CE
succinic anhydride) had a relatively slower activation over the glutaric and adipic anhydride budesonide prodrugs (about 2-3 times slower, see Figure 5) and was therefore selected for
AC
incorporation into inhalable microspheres. The application of an inhaled corticosteroid dextran prodrug for asthma inhalation therapy could alter the relatively fast systemic absorption typically seen of a lipophilic steroid, such as budesonide, allowing for steady release of the active drug, hence improving its effectiveness. The budesonide-spacer-dextran prodrug has the potential of escaping mucociliary clearance due to its hydrophilicity.
22
ACCEPTED MANUSCRIPT To formulate the budesonide prodrug for lung deposition, the budesonide prodrug was spray-dried with lactose to produce respirable aerodynamic spherical particles. The lactose carrier when spray-dried was amorphous (data not shown) and dissolved upon disposition in the lung, thereby delivering the prodrug while reducing the effect of mucociliary clearance. Spray-
T
drying can influence particle morphology by the drying rate, the surface tension, the solubility of
IP
the drug in the feeding solution, and the viscosity of the feeding solution (Park, 2013). The
CR
modified spray-dryer was fabricated to collect the dried material onto an ACI where the particles are separated by size. The materials were collected and pooled from Stage 2-7 of the ACI for the
US
spray-dried budesonide prodrug with lactose materials and were characteristic of a normal and
AN
narrow (Span 1.9) particle size distribution with smooth surfaces (Figure 6). Particles having an aerodynamic particle size diameter between 5 and 10 microns will
M
enter the lungs and deposit by gravitational settling (sedimentation) predominantly in the middle-
ED
to-upper lung regions (Park, 2013). Particles with an aerodynamic particle size diameter of 2 to 5 microns can efficiently target the middle-to-lower lung regions whereas particles having an
PT
aerodynamic particle size less than 2 microns can efficiently target the smaller airways and deep
CE
lung region (Park, 2013). The MMAD results for the spray-dried budesonide prodrug were within the respirable particle size of less than 5 µm. The spray-drying process performed within
AC
an aqueous environment will require further optimization to prevent partial activation of the prodrug during the spray-drying process. Unfortunately, further optimization of the procedure was not possible as we ran out of prodrug. It is, however, very feasible that during the spraydrying procedure the spray-drying medium was not acetic enough to prevent partial ester hydrolysis at the elevated temperature.
23
ACCEPTED MANUSCRIPT CONCLUSIONS: Spray-drying a dextran based budesonide prodrug with lactose can produce respirable size particles for asthma inhalation therapy. The slow generation of budesonide from the chemical delivery system might further improve the pharmacological profile of budesonide. A
T
parallel activation of a budesonide prodrug can simulate the observed concentration-time profile
IP
of released budesonide and budesonide-21-hemisuccinate with simultaneous non-linear (least
CR
squares) curve fitting. To fully prevent degradation of the prodrug during spray-drying, the
AC
CE
PT
ED
M
AN
US
process will require further optimization.
24
ACCEPTED MANUSCRIPT CONFLICT OF INTEREST DECLARATION The authors declare that they have no competing interests
Acknowledgments: This work was in part supported by a Department of Pharmaceutics
AC
CE
PT
ED
M
AN
US
CR
IP
T
unrestricted research funds.
25
ACCEPTED MANUSCRIPT REFERENCES: Bozek, A., Warkocka-Szoltysek, B., Filipowska-Gronska, A., Jarzab, J., 2012. Montelukast as an Add-on Therapy to Inhaled Corticosteroids in the Treatment of Severe Asthma in Elderly Patients. Journal of Asthma 49, 530-534.
T
Chanteux, H., Rosa, M., Delatour, C., Prakash, C., Smith, S., Nicolas, J., 2014. In Vitro Hydrolysis and Transesterification of CDP323, an alpha4beta1/alpha4beta7 Integrin Antagonist Ester Prodrug. Drug Metabolism and Disposition 42, 153-161.
CR
IP
Edsbacker, S., Andersson, P., Lindberg, C., Ryrfeldt, A., Thalen, A., 1987. Metabolic Acetal Splitting of Budesonide. Drug Metabolism and Disposition 15, 412-417.
US
Hochhaus G, Möllmann H, Derendorf H, Gonzalez-Rothi RJ. Pharmacokinetic /pharmacodynamic aspects of aerosol therapy using glucocorticoids as a model. J Clin Pharmacol. 1997;37:881–92.
AN
Hochhaus G, Froehlich P, Hochhaus R, Möllmann A, Derendorf, Möllmann H. A selective HPLC/RIA for the determination of budesonide, J. Pharm. Biomed. Anal. 17, 1235-1242 (1998)
M
Hosokawa, M., 2008. Structure and Catalytic Properties of Carboxylesterase Isozymes Involved in Metabolic Activation of Prodrugs. Molecules 13, 412-431.
ED
Hubner, M., Hochhaus, G., Derendorf, H., 2005. Comparative Pharmacology, Bioavailability, Pharmacokinetics, and Pharmacodynamics of Inhaled Glucocorticosteroids. Immunology Allergy and Clinics of North America 25, 469-488.
CE
PT
Jung, Y.J., Lee, J.S., Kim, H.H., Kim, Y.T., Kim, Y.M., 1998. Synthesis and Properties of Dextran-5-aminosalicylic Acid Ester as a Potential Colon-specific Prodrug of 5-Aminosalicylic Acid. Arch. Pharm. Res. 21, 179-186.
AC
Koitka, M., Hochel, J., Obst, D., Rottmann, A., Gieschen, H., Borchert, H., 2008. Determination of rat serum esterase activities by an HPLC method using S-acetylthiocholine iodide and pnitrophenyl acetate. Analytical Biochemistry 381, 113-122. Larsen, C., 1989. Dextran prodrugs - structure and stability in relation to therapeutic activity. Advanced Drug Delivery Reviews 3, 103-154. Lee, Y., Kim, I.H., Kim, J., Yoon, J., Shin, Y.H., Jung, Y., Kim, Y.M., 2011. Evaluation of dextran-flufenamic acid ester as a polymeric colon-specific prodrug of flufenamic acid, an antiinflammatory drug, for chronotherapy. Journal of Drug Targeting 19, 336-343. McLeod, A.D., Friend, D.R., Tozer, T.N., 1993. Synthesis and chemical stability of glucocorticoid-dextran esters: potential prodrugs for colon-specific delivery. International Journal of Pharmaceutics 92, 105-114.
26
ACCEPTED MANUSCRIPT Mehvar, R., Shepard, T.L., 1992. Molecular-weight-dependent pharmacokinetics of fluoresceinlabeled dextrans in rats. J. Pharm. Sci. 81 (9), 908-912. Mehvar, R., 1999. Simultaneous analysis of dextran-methylprednisolone succinate, methylprednisolone succinate, and methylprednisolone by size-exclusion chromatography. Journal of Pharmaceutical and Biomedical Analysis 19, 785-792.
IP
T
Mehvar, R., Dann, R.O., Hoganson, D.A., 2000. Kinetics of hydrolysis of dextranmethylprednisolone succinate, a macromolecular prodrug of methylprednisolone, in rat blood and liver lysosomes. J. Control Release, 68(1): 53-61.
CR
Muller, S., Hutson, A., Arya, V., Hochhaus, G., 1999. Assessment of Complex Peptide Degradation Pathways via Structured Multicompartmental Modeling Approaches: The Metabolism of Dynorphin A1-13 and Related Fragments in Human Plasma. Journal of Pharmaceutical Sciences 88, 938-944.
AN
US
Naikwade, S.R., Bajaj, A.N., Gurav, P., Gatne, M.M., Soni, P.S., 2009. Development of Budesonide Microparticles Using Spray-Drying Technology for Pulmonary Administration: Design, Characterization, In Vitro Evaluation, and In Vivo Efficacy Study. American Association of Pharmaceutical Scientists 10, 993-1012.
ED
M
Naikwade, S.R., Bajaj, A.N., Gurav, P., Gatne, M.M., Soni, P.S., 2009. Development of Budesonide Microparticles Using Spray-Drying Technology for Pulmonary Administration: Design, Characterization, In Vitro Evaluation, and In Vivo Efficacy Study. American Association of Pharmaceutical Scientists 10, 993-1012.
PT
Pang, Y., Zhang, Y., Zhang, Z., 2002. Synthesis of an enzyme-dependent prodrug and evaluation of its potential for colon targeting. World Journal of Gastroenterology 8, 913-917.
CE
Pany, Y., Sakagami, M., Byron, PR., 2007. Insulin self-association: effects on lung disposition kinetics in the airways of the isolated perfused rat lung (IPRL). Pharmaceutical Research 24(9), 1636-44.
AC
Penugonda, S., Kumar, A., Agarwal, H.K., Parang, K., Mehvar, R., 2008. Synthesis and in vitro characeterization of novel dextran-methylprednisolone conjugates with peptide linkers: effects of linker length on hydrolytic and enzymatic release of methylprednisolone and its peptidyl intermediates. J. Pharm. Sci. 97(7), 2649-2664. Piao, H., Cho, H., Oh, E., Chung, S., Shim, C., Kim, D., 2009. Budesonide Microemulsions for Enhancing Solubility and Dissolution Rate. Journal of Korian Pharmaceutical Sciences 39, 417422. Redinbo, M.R., Bencharit, S., Potter, P.M., 2003. Human carboxylesterase 1: from drug metabolism to drug discovery. Biochemical Society Transactions 31, 620-624.
27
ACCEPTED MANUSCRIPT Rohdewald P, Rehder J, Drehsen G, Hochhaus G, Derendorf H, Möllmann H. Simultaneous determination of glucocorticoid alcohols, their succinates and hydrocortisone in plasma. Journal of pharmaceutical and biomedical analysis. 1985;3:566–73. Shimizu, M., Fukami, T., Ito, Y., Kurokawa, T., Kariya, M., Nakajima, M., Yokoi, T., 2014. Indiplon is Hydrolyzed by Arylacetamide Deacetylase in Human Liver. Drug Metabolism and Disposition 42, 751-758.
CR
IP
T
Suzaki, Y., Uemura, N., Takada, M., Ohyama, T., Itohda, A., Morimoto, T., Imai, H., Hamasaki, H., Inano, A., Hosokawa, M., Tateishi, M., Ohashi, K., 2013. The effect of carboxylesterase 1 (CES1) polymorphisms on the pharmacokinetics of oseltamivir in humans. Eur. J. Clin. Pharmacol. 69, 21-30.
US
Thorsson, L., Edsbacker, S., Kallen, A., Lofdahl, C-G., 2001. Pharmacokinetics and systemic activity of fluticasone via Diskus® and pMDI, and of budesonide via Turbuhaler®. British J. Clin. Pharmcol. 52 (5), 529-538.
AN
Tunek, A., Sjodin, K., Hallstrom, G., 1997. Reversible Formation of Fatty Acid Esters of Budesonide, an Antiasthma Glucocorticoid, in Human Lung and Liver Microsomes. Drug Metabolism and Disposition 25, 1311-1317.
ED
M
Varshosaz, J., Emami, J., Ahmadi, F., Tavakoli, N., Minaiyan, M., Fassihi, A., Mahzouni, P., Dorkoosh, F., 2011. Preparation of budesonide-dextran conjugates using glutarate spacer as a colon-targeted drug delivery system: in vitro / in vivo evaluation in induced ulcerative colitis. Journal of Drug Targeting 19, 140-153.
CE
PT
Varshosaz, J., Emami, J., Fassihi, A., Tavakoli, N., Minaiyan, M., Ahmadi, F., Mahzouni, P., Dorkoosh, F., 2010. Effectiveness of budesonide-succinate-dextran conjugate as a novel prodrug of budesonide against acetic acid-induced colitis in rats. International Journal of Colorectal Disease 25, 1159-1165.
AC
Varshosaz, J., Emami, J., Tavakoli, N., Fassihi, A., Minaiyan, M., Ahmadi, F., Dorkoosh, F., 2009. Synthesis and evaluation of dextran-budesonide conjugates as colon specific prodrugs for treatment of ulcerative colitis. International Journal of Pharmaceutics 365, 68-75. Varshosaz, J., Emami, J., Tavakoli, N., Minaiyan, M., Rahmani, N., Dorkoosh, F., 2012. Development and Evaluation of a Novel Pellet-Based Tablet System for Potential Colon Delivery of Budesonide. Journal of Drug Delivery 2012, 1-7. Wang, J., Williams, E.T., Bourgea, J., Wong, Y.N., Patten, C.J., 2011. Characterization of Recombinant Human Carboxylesterases: Fluorescein Diacetate as a Probe Substrate for Human Carboxylesterase 2. Drug Metabolism and Disposition 39, 1329-1333. Williams, E.T., Bacon, J.A., Bender, D.M., Lowinger, J.J., Guo, W., Ehsani, M.E., Wang, X., Wang, H., Qian, Y., Ruterbories, K.J., Wrighton, S.A., Perkins, E.J., 2011. Characterization of
28
ACCEPTED MANUSCRIPT the Expression and Activity of Carboxylesterases 1 and 2 from the Beagle Dog, Cynomolgus Monkey, and Human. Drug Metabolism and Disposition 39, 2305-2313.
AC
CE
PT
ED
M
AN
US
CR
IP
T
Zhang, C., Xu, Y., Zhong, Q., Li, X., Gao, P., Feng, C., Chu, Q., Chen, Y., Liu, D., 2014. In Vitro Evaluation of the Inhibitory Potential of Pharmaceutical Excipients on Human Carboxylesterase 1A and 2. PLOS ONE 9, e93819.
29
ACCEPTED MANUSCRIPT TABLES Table 1
Kinetic models describing the activation pathway of the budesonide prodrug. Budesonide prodrug (P1), budesonide (P2), 16-alpha hydroxyprednisolone (P3), and budesonide-21-hemisuccinate (P4) are represented in the model flow diagrams. The differential equations, assuming first order processes are also shown for each model. Model 2 splits the prodrug pool into two independent fractions (P1a and P1b) with separate hydrolysis constants. Governing Time Differential Equation
IP
T
Model
ED
M
AN
US
CR
Model 1
AC
CE
PT
Model 2
30
ACCEPTED MANUSCRIPT Degrading and generating rate constants from a budesonide prodrug as determined by a non-linear (least-squares) curve-fitting (free float all k) method using a single-phase (Model 1) and bi-phasic (Model 2) prodrug activation model approach. The model fit was determined by using the Akaike Information Criterion (AIC) for the combined release of budesonide B(t) and budesonide-21hemisuccinate L(t).
T
Model 2 0.2076 0.0517 0.6211 0.0163 0.0023 0.0108 -31.0
AC
CE
PT
ED
M
AN
US
Model 1 0.0841 N/A 0.1192 N/A 0.0050 0.0117 56.7
IP
Rate Constant (h-1) k12a k12b k14a k14b k42 k23 AIC
CR
Table 2
31
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7