Article pubs.acs.org/molecularpharmaceutics
Direct Observation of Interactions of Silk-Elastinlike Protein Polymer with Adenoviruses and Elastase Se-Hui Jung,† Joung-Woo Choi,‡ Chae-Ok Yun,‡ Sun Hwa Kim,† Ick Chan Kwon,† and Hamidreza Ghandehari*,†,§,∥,⊥ †
Center for Theragnosis, Biomedical Research Institute, Korea Institute of Science and Technology, Hwarangno 14-gil 5, Seongbuk-gu, Seoul 136-791, Korea ‡ Department of Bioengineering, College of Engineering, Hanyang University, Seoul, Korea § Department of Pharmaceutics and Pharmaceutical Chemistry, ∥Department of Bioengineering, and ⊥Center for Nanomedicine, Nano Institute of Utah, University of Utah, Salt Lake City, Utah 84112, United States ABSTRACT: Silk-elastinlike protein polymer (SELP) hydrogels have been investigated for sustained local delivery of adenoviral gene carriers to solid tumors. These polymers degrade in the presence of proteases such as elastase. A detailed understanding of the interaction of SELPs with viruses and their degradation in the presence of elastase can provide useful information about mechanisms of sustained gene delivery from these systems. In this work, we investigated the interactions of SELPs with adenoviruses (Ads) and elastase using atomic force microscopy. We observed that viral particles interacted strongly with SELP networks formed by crosslinking of nanofibers. The presence of viruses contributed to enhanced network formation. Incubation of Ad with SELPs in the liquid state induced close packing of the viral colony. Morphological changes of SELP networks cleaved by enzymatic interaction with elastase were investigated. SELP-415K fiber networks were more responsive to temperature changes and were slowly degraded by elastases compared to SELP-47K, a SELP analogue with shorter elastin units in the monomer repeat. These studies provide insight into the influence of SELP structure on degradation and potential mechanisms of increased viral stability. KEYWORDS: atomic force microscopy, silk-elastinlike protein polymers, adenovirus, elastase
■
INTRODUCTION Silk-elastinlike protein polymers (SELPs) make up a family of recombinant polymers made of repeating units of silk-like (GAGAGS) and elastin-like (GVGVP) blocks.1−5 With the appropriate sequence and composition, SELPs form crosslinked networks nucleated by hydrogen bonding of silk units. The density of these cross-links depends on the amino acid sequence of the repeating units, the ratio of silk to elastin blocks, the polymer concentration, and the incubation time, among other factors.6−8 Cross-linked networks of SELPs have utility in matrix-mediated adenoviral gene delivery.9 It has been demonstrated that the extent and duration of transgene expression of adenoviral vectors are both increased when they are embedded in SELP matrices, while localized delivery is achieved.10,11 The presence of adenoviruses (Ads) influences the network properties of SELPs,12 while enhancing viral viability.13 Little is known about the molecular mechanisms of the interaction between SELP polymer strands and viral particles. Previously, we had examined the influence of polymer structure and concentration on the extent of polymer degradation in the presence of elastases.10 In addition, we had examined the influence of the presence and absence of adenoviruses on network properties as a function of polymer structure and concentration.12 However, there was limited © 2015 American Chemical Society
visual evidence of how adenoviruses affect network properties and the extent to which elastases lead to fragmentation of polymer strands and fibers as a function of polymer structure. One way to gain visual insight into the extent of network formation and investigating factors that influence the interactions of SELPs with viral particles or proteases is atomic force microscopy (AFM). AFM imaging of SELPs has provided direct observation of nanofiber nucleation.14 In this article, the interaction of SELP nanofibers with adenoviral particles observed by AFM imaging is reported. The physicochemical properties and degradation of SELPs are assessed by the numbers of silk and elastin units within a monomer segment. We selected SELP-47K and -415K (Figure 1), which have the same number of silk units but different numbers of elastin units, to evaluate the influence of polymer structure on degradation and interaction with the viruses. Additionally, the degradation of SELPs in the presence of elastase as a function of polymer structure and enzyme concentration is documented. Received: Revised: Accepted: Published: 1673
January 24, 2015 April 6, 2015 April 16, 2015 April 16, 2015 DOI: 10.1021/acs.molpharmaceut.5b00075 Mol. Pharmaceutics 2015, 12, 1673−1679
Article
Molecular Pharmaceutics
in distilled water were dropped on mica coated with a 10 μg/ mL SELP-47K solution. After the SELP solution had been added to the substrate, the samples were incubated in a humidity chamber for 3 h at 37 °C and then washed three times with distilled water. For investigation of time-dependent interactions of Ad with SELP fibers, 40 μL of a 1011 VPs/mL virus solution was dropped on SELP-coated mica. After incubation of the virus solution for the indicated times (5, 10, 30, 60, and 180 min) at 37 °C in a humidity chamber, the samples were washed three times with distilled water. Enzymatic Degradation of SELP Nanofibers. SELP nanofibers were formed by incubation with 40 μL of a 10 μg/ mL SELP-47K and SELP-415K solution on the freshly cleaved mica for 3 h in a humidity chamber at 37 °C. Enzymatic degradation of SELP nanofibers by elastase was performed according to the procedure described by the manufacturer (Elastin Product Company, Inc., Owensville, MO). Briefly, a 1 mg/mL elastase stock solution was prepared in acidic buffer [50 mM NaOAc (pH 5.0) and 0.1 M NaCl] to protect the enzyme from autolysis and proteolysis. To prepare various concentrations (0.010, 0.1, 1, and 10 μg/mL), the elastase stock solutions were diluted with chilled reaction buffer [0.1 M Tris (pH 7.5) and 0.5 M NaCl]. SELP nanofibers coated on a mica substrate were incubated with the elastase solution for 2 h in a humidity chamber at 37 °C. AFM Imaging. AFM imaging was performed at ambient temperature using a psia XE-100 AFM system (psia, Korea). Silicon cantilevers with a spring constant of 42 N/m were used for imaging. The morphology of the samples was imaged in tapping mode at a scanning rate of 0.1−1 Hz. Collected AFM images were analyzed with XEI version 1.7 (psia). The surface coverage of the SELP nanofiber was quantified using Image-Pro Plus version 4.5 (Media cybernetics, Bethesda, MD). The ratio of virus bound to SELP strands was calculated by dividing total
Figure 1. Amino acid sequences of SELP-47K and SELP-415K.
■
MATERIALS AND METHODS Sample Preparation. Recombinant SELP-47K and SELP415K copolymers were synthesized and characterized as described previously.15 Figure 1 shows the amino acid sequence of these copolymers. Oncolytic Ads (Ad-ΔB7-KOX) were generated as previously described.16 Ad-ΔB7-KOX is a cancer cell-specific replicating oncolytic Ad that expresses KOX. KOX is an artificial transcriptional repressor zinc-finger protein that strongly suppresses the expression of endothelial growth factor (VEGF). Ad-ΔB7-KOX was propagated in A549 cells followed by CsCl (Sigma, St. Louis, MO) density purification. The viral particle numbers were calculated from measurements of optical density at 260 nm (OD260), where 1 unit is equivalent to 1012 viral particles per milliliter. To prepare the samples for AFM imaging, 40 μL of SELP47K and SELP-415K solutions with various polymer concentrations (1, 10, 50, 100, 500, and 1000 μg/mL) in distilled water was dropped on freshly cleaved mica (Ted Pella, Inc.). Samples were incubated for 3 h in a humidity chamber at 37 °C and then washed three times with distilled water. To study interactions between SELP and Ad, 40 μL virus solutions at various concentrations (5, 10, 50, 100, and 500 × 109 VPs/mL)
Figure 2. Structural characterization of the SELP-47K polymer using tapping mode AFM. (A) Structure of SELP strands (not cross-linked). (B) Structure of SELP networks formed by cross-linking. A representative network formed by cross-linking is marked with an arrow. (C) Threedimensional structure of SELP networks and line profile of a single SELP strand. 1674
DOI: 10.1021/acs.molpharmaceut.5b00075 Mol. Pharmaceutics 2015, 12, 1673−1679
Article
Molecular Pharmaceutics
Figure 3. Formation of SELP networks by cross-linking as a function of polymer concentration: (A) SELP-47K and (B) SELP-415K. (C) Surface coverage of SELP networks that is dependent on polymer concentration. The surface coverage of the SELP nanofiber was quantified from each of three random fields using Image-Pro Plus version 4.5. The surface coverage represents the mean ± the standard deviation (n = 3).
of silk units.9 These cross-links are dependent on cure time, polymer concentration, and ionic strength. To structurally characterize SELP nanofibers, 40 μL of 10 μg/mL SELP-47K (Figure 1) was loaded on freshly cleaved mica and incubated for 3 h. When the droplet of the SELP-47K solution permeated through the natural crack of mica, SELP strands bound to a
numbers of viruses by the numbers of viruses bound to SELP strands from a selected image.
■
RESULTS AND DISCUSSION Formation of SELP Networks. Self-assembly network formation of SELPs occurred by cross-links due to association 1675
DOI: 10.1021/acs.molpharmaceut.5b00075 Mol. Pharmaceutics 2015, 12, 1673−1679
Article
Molecular Pharmaceutics
Figure 4. Dose-dependent interactions of Ad with SELP fibers. Oncolytic Ad solutions with indicated concentrations (5, 10, 50, 100, and 500 × 109 VPs/mL) were added to the mica substrate coated with 10 μg/mL SELP-47K. Then, samples were imaged by tapping mode AFM.
networks that were looser than those of SELP-47K, because of the longer distance between the cross-linking sites in silk blocks of SELP-415K. These results correlate with previous observations that SELP-47K provided hydrogel networks denser than those of SELP-415K.15 At similar polymer concentrations, the surface coverage of SELP-415K was 1.5−2.9-fold higher than that of SELP-47K in the range of 10−500 μg/mL (Figure 3C), where the average width and height of SELP-415K (width of 54.2 ± 4.8 nm, height of 1.57 ± 0.23 nm, n = 20) were greater than those of SELP-47K (width of 43.4 ± 2.8 nm, height of 0.95 ± 0.08 nm, n = 20). This is due to the higher swelling ratio of SELP-415K resulting in a higher capacity of binding with water molecules due to longer elastin units.15 Interactions of SELP Nanofibers with Ad. Recombinant SELPs have been used for sustained local delivery of various therapeutics, including adenoviral gene carriers. In such an application, aqueous solutions of SELPs are mixed with the adenoviruses at room temperature, followed by injection where they form hydrogels at body temperature. Previously, we reported that Ad interacts with SELP nanofibers, although the interaction between them did not critically influence the swelling ratio of the resulting hydrogels at high polymer concentrations.12 Hexon, a major capsid protein found in Ad, has richly acidic amino acid residues such as glutamate and aspartate that can participate in electrostatic interaction with the positively charged lysine residues in the SELP polymer backbone. While the details of electrostatic interactions between hexon proteins and SELP nanofibers are not clearly understood, these interactions can potentially play a role in prolonging the biological activity of Ad by stabilization of the virus envelope due to the interaction with the silk or elastin units as well as stabilization of hydrogel networks. The study of interactions between SELP nanofibers and Ad is important for understanding the mechanism of release of viral particles from SELP matrices and the resulting spatiotemporal control over delivery of Ad. To investigate these interactions, oncolytic Ad aliquots at the indicated concentrations (5, 10, 50, 100, and 500 × 109 VPs/mL) were dropped to the mica substrate coated with 10 μg/mL SELP-47K in a humidity chamber for 3 h at 37 °C. While negatively charged virus
mica surface by electrostatic interactions between the surface and SELP strands, as observed previously,17 and SELP strands interacted with each other but did not cross-link (Figure 2A). On the other hand, when the droplet was well-maintained, a SELP network was formed by cross-linking between SELP strands (Figure 2B, arrow). This suggests that interactions between the mica surface and SELP strands are stronger than those between SELP strands, and network formation requires a sufficient amount of SELP fibers for relatively longer periods. The SELP fibers had a width of 40 nm, a height of 1 nm, a length between 400 and 500 nm, and a twist with a half-pitch of ∼25 nm (Figure 2C). To investigate the effect of SELP concentration on network formation, the mica substrate was treated with various concentrations of a SELP-47K and SELP-415K solution (polymer structures shown in Figure 1). The surface coverage by SELP-47K nanofibers increased as the polymer concentration increased (Figure 3A). While networks of SELP-47K were not observed at a concentration of 1 μg/mL, they were observed at concentrations of >10 μg/mL, suggesting that this concentration is suitable for the investigation of interaction of SELP nanofibers with Ad and elastase. SELP networks were distributed evenly on the mica surface at concentrations of 10 and 50 μg/mL. Additionally, SELP networks with a highdensity mimicking spider web and network cores were formed at concentrations of >100 μg/mL. The formation of SELP networks along the z-axis as well as the x- and y-axes was also observed, which suggests the critical concentration of SELP47K for network formation is 1 mg/mL. A dose-dependent increase in surface coverage by network formation was also observed with SELP-415K (Figure 3B). While only shorter SELP nanofibers were observed at 1 μg/mL, the length of SELP nanofibers was increased as a function of polymer concentration, suggesting these constructs are elongated vertically as well as cross-linked by interaction between adjacent nanofibers. This is consistent with earlier reports that network formation is predominantly influenced by polymer concentration rather than polymer structure.12 However, the length of networks of SELP-415K was shorter, and each nanofiber of SELP-415K loosely interacted with adjacent fibers resulting in 1676
DOI: 10.1021/acs.molpharmaceut.5b00075 Mol. Pharmaceutics 2015, 12, 1673−1679
Article
Molecular Pharmaceutics
Figure 5. Time-dependent interactions of Ad with SELP fibers. Oncolytic Ad solutions of 1 × 1011 VPs/mL were treated on a mica substrate coated with SELP-47K for the indicated times (5, 10, 30, 60, and 180 min). Total viral particles and those bound to SELP polymers were quantified from each of three random fields. Data represent the mean ± the standard deviation (n = 3).
Figure 6. Interactions of SELP-47K with Ad as a function of polymer concentration in the liquid state. The mixtures of the 1 × 1011 VPs/mL oncolytic Ad solution and SELP-47K polymer with indicated concentrations (1, 10, 100, and 1000 μg/mL) were treated on freshly cleaved mica for 3 h in a humidity chamber at 37 °C. Samples were imaged by tapping mode AFM.
reported that tapping by AFM can act as a mechanical stimulus for new amyloid growth by providing broken amyloid fibers on mica.14 Thus, we can expect that the Brownian motion of Ad can stimulate further network formation and result in stabilization of the polymer network at lower polymer concentrations. To investigate time-dependent interactions between SELP nanofibers and Ad, 1 × 1011 VPs/mL Ad solutions were dropped on the the mica substrate coated with 10 μg/mL SELP-47K in a humidity chamber for the indicated times at 37 °C. Samples were subsequently imaged by tapping mode AFM. The probability of viral particles bound to SELP nanofibers was expressed as the ratio of the number of viral particles bound to fibers and total numbers of viral particles from AFM images.
particles were not attached to the mica substrate that also has a negative surface charge (data not shown), a dose-dependent increase of the number of viral particles was observed (Figure 4). Most of the viral particles were attached to the SELP nanofibers, suggesting the presence of viral coat−SELP interactions. The density of SELP nanofibers also increased depending on the concentrations of Ad, while the concentration of the polymer solution for coating of SELP nanofibers on mica was fixed. This is likely due to the Brownian motion of Ad that can act as a stimulus for growth of the SELP network. The induction of further growth of nanofibers by macroscopic stimuli, including agitation, sonication, and shear flow, which can break mature amyloid fibers and provide these fragments as seeds, has been reported.14 At the molecular level, it was 1677
DOI: 10.1021/acs.molpharmaceut.5b00075 Mol. Pharmaceutics 2015, 12, 1673−1679
Article
Molecular Pharmaceutics
Figure 7. Enzymatic degradation of SELPs as a function of elastase concentration. The elastase solutions at various concentrations (0.010, 0.1, 1, and 10 μg/mL) were dropped on (A) SELP-47K or (B) SELP-415K nanofibers coated on a mica substrate for 2 h in a humidity chamber at 37 °C. Morphological changes caused by enzymatic degradation were investigated using tapping mode AFM.
The number of viral particles bound to fibers as well as total viral particles increased depending on incubation time, and most observed Ads were attached to SELP nanofibers by electrostatic interaction between Ad and the fibers. The ratios of viral particles bound to fibers were 74 and 90% at 5 and 180 min, respectively (Figure 5). The round-shaped morphologies of most Ad attached to SELP nanofibers were kept for 30 min. However, the deformation of round-shaped virus occurred in 15% of total virus over 1 h. One possible cause for deformation of Ad is multiple strong interactions of Ad with several solid SELP strands bound to the two-dimensional surface. On the contrary, when mixed solutions of Ad and SELP at the indicated concentrations (1, 10, 100, and 1000 μg/mL) were dropped on the mica substrate, viral particles were aggregated depending on the SELP concentration by interaction of several Ads with SELPs without deformation (Figure 6). Further, SELP networks were not observed up to 100 μg/mL and evidently observed at 1000 μg/mL, which were formed between viral particles rather than on the mica surface. This result suggests that SELPs prefer Ad to the mica surface, and SELP network
formation can be influenced by Ad at low polymer concentrations. When a mixed solution of Ad and SELP was dropped on mica after incubation for 1 week at room temperature, viral particles were closely packed (data not shown) by interaction between Ads and SELPs. This closely packed structure of viruses may reduce the extent of denaturation induced by the external environment. Enzymatic Degradation of SELP Nanofibers by Elastase. SELPs provide spatiotemporal control over viral gene delivery9 and are known to degrade in the presence of elastases.10 To gain insight into the influence of elastase concentration and polymer structure on the degradability of these polymers, morphological changes in SELP nanofibers in the presence of elastases were investigated. The chilled elastase solutions at the indicated concentrations (0.01, 0.1, 1, and 10 μg/mL) were applied to the mica substrate coated with 10 μg/ mL SELP-47K and -415K in a humidity chamber for 2 h at 37 °C. The morphology of SELP-47K nanofibers changed as a function of elastase concentration (Figure 7A). Smaller and larger aggregates were observed at 10 and 100 ng/mL elastase, 1678
DOI: 10.1021/acs.molpharmaceut.5b00075 Mol. Pharmaceutics 2015, 12, 1673−1679
Article
Molecular Pharmaceutics
(3) Teng, W.; Huang, Y.; Cappello, J.; Wu, X. Optically transparent recombinant silk-elastinlike protein polymer films. J. Phys. Chem. B 2011, 115, 1608−1615. (4) Frandsen, J. L.; Ghandehari, H. Recombinant protein-based polymers for advanced drug delivery. Chem. Soc. Rev. 2012, 41, 2696− 2706. (5) Cappello, J.; Crissman, J.; Dorman, M.; Mikolajczak, M.; Textor, G.; Marquet, M.; Ferrari, F. Genetic engineering of structural protein polymers. Biotechnol. Prog. 1990, 6, 198−202. (6) Dandu, R.; Von Cresce, A.; Briber, R.; Dowell, P.; Cappello, J.; Ghandehari, H. Silk-elastinlike protein polymer hydrogels: Influence of monomer sequence on physicochemical properties. Polymer 2009, 50, 366−374. (7) Dinerman, A. A.; Cappello, J.; Ghandehari, H.; Hoag, S. W. Swelling behavior of a genetically engineered silk-elastinlike protein polymer hydrogel. Biomaterials 2002, 23, 4203−4210. (8) Megeed, Z.; Cappello, J.; Ghandehari, H. Genetically engineered silk-elastinlike protein polymers for controlled drug delivery. Adv. Drug Delivery Rev. 2002, 54, 1075−1091. (9) Gustafson, J. A.; Ghandehari, H. Silk-elastinlike protein polymers for matrix-mediated cancer gene therapy. Adv. Drug Delivery Rev. 2010, 62, 1509−1523. (10) Gustafson, J.; Greish, K.; Frandsen, J.; Cappello, J.; Ghandehari, H. Silk-elastinlike recombinant polymers for gene therapy of head and neck cancer: From molecular definition to controlled gene expression. J. Controlled Release 2009, 140, 256−261. (11) Greish, K.; Araki, K.; Li, D.; O’Malley, B. W., Jr.; Dandu, R.; Frandsen, J.; Cappello, J.; Ghandehari, H. Silk-elastinlike protein polymer hydrogels for localized adenoviral gene therapy of head and neck tumors. Biomacromolecules 2009, 10, 2183−2188. (12) Dandu, R.; Ghandehari, H.; Cappello, J. Characterization of structurally related adenovirus-laden silk-elastinlike hydrogels. Journal of Bioactive and Compatible Polymers 2008, 23, 5−19. (13) Hatefi, A.; Cappello, J.; Ghandehari, H. Adenoviral gene delivery to solid tumors by recombinant silk-elastinlike protein polymers. Pharm. Res. 2007, 24, 773−779. (14) Varongchayakul, N.; Johnson, S.; Quabili, T.; Cappello, J.; Ghandehari, H.; Solares Sde, J.; Hwang, W.; Seog, J. Direct observation of amyloid nucleation under nanomechanical stretching. ACS Nano 2013, 7, 7734−7743. (15) Haider, M.; Leung, V.; Ferrari, F.; Crissman, J.; Powell, J.; Cappello, J.; Ghandehari, H. Molecular engineering of silk-elastinlike polymers for matrix-mediated gene delivery: Biosynthesis and characterization. Mol. Pharmaceutics 2005, 2, 1339−1350. (16) Kang, Y. A.; Shin, H. C.; Yoo, J. Y.; Kim, J. H.; Kim, J. S.; Yun, C. O. Novel cancer antiangiotherapy using the VEGF promotertargeted artificial zinc-finger protein and oncolytic adenovirus. Mol. Ther. 2008, 16, 1033−1040. (17) Hwang, W.; Kim, B. H.; Dandu, R.; Cappello, J.; Ghandehari, H.; Seog, J. Surface induced nanofiber growth by self-assembly of a silk-elastin-like protein polymer. Langmuir 2009, 25, 12682−12686.
respectively. These aggregations likely occurred by binding of fragments degraded by elastase to nanofibers. Enzymatic cleavage of networks by elastase were observed at 1 and 10 μg/mL SELP-47K. These results indicate that the morphology of SELP networks was preserved in the presence of up to 100 ng/mL elastase, although partial degradation was observed. Dose-dependent changes in SELP-415K networks caused by elastase were also observed as shown in Figure 7B. Interestingly, morphological changes in SELP-415K networks were induced by chilled reaction buffer used to protect autoproteolysis. Several strands of SELP nanofibers bound longitudinally with each other and resulted in the formation of thicker fibers. SELP-415K strands coiled probably because of the enhanced temperature sensitivity of these polymers due to longer elastin units as reported previously,15 while the effect of the change in temperature on the morphology of SELP-47K fibers was negligible because of the presence of shorter elastin units. The cleavage of SELP-415K networks started at 1 μg/mL elastase, but networks that consisted of longer and thicker nanofibers were not cleaved at this concentration. One possible explanation is that the cold temperature induced formation of thicker fibers by longitudinal interaction that can contribute to steric hindrance and cause the degradation of SELP-415K to be slower than that of SELP-47K.
■
CONCLUSIONS In summary, this study provides visual evidence of the interactions of SELPs with adenoviruses using AFM. The SELP strands interacted with viral particles, which contributed to the growth of the three-dimensional network. The interaction of SELPs with Ad in the liquid state induced the formation of a densely packed viral colony through the SELP scaffolds. Morphological changes and degradation of SELP networks in the presence of elastase were observed. These results provide molecular evidence of the degradation of SELPs and their interaction with adenoviral gene carriers.
■
AUTHOR INFORMATION
Corresponding Author
*Address: 5205 SMBB, 36 S. Wasatch Dr., Salt Lake City, UT 84112-5001. E-mail:
[email protected]. Phone: (801) 587-1566. Fax: (801) 581-6321. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the Global Innovative Research Center program of the National Research Foundation of Korea (2012K1A1A2A01055811) and by the Intramural Research Program (Global RNAi Carrier Initiative) of the Korean Institute of Science and Technology, a grant from the National Institutes of Health (R01CA107621, H.G.), and the National Research Foundation of Korea (2010-0029220, C.-O.Y.).
■
REFERENCES
(1) Teng, W.; Cappello, J.; Wu, X. Physical crosslinking modulates sustained drug release from recombinant silk-elastinlike protein polymer for ophthalmic applications. J. Controlled Release 2011, 156, 186−194. (2) Qiu, W.; Huang, Y.; Teng, W.; Cohn, C. M.; Cappello, J.; Wu, X. Complete recombinant silk-elastinlike protein-based tissue scaffold. Biomacromolecules 2010, 11, 3219−3227. 1679
DOI: 10.1021/acs.molpharmaceut.5b00075 Mol. Pharmaceutics 2015, 12, 1673−1679
