- Email: [email protected]
Stability and cell uptake of calcium carbonate templated insulin microparticles
Acta Biomaterialia 10 (2014) 1423–1430
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Stability and cell uptake of calcium carbonate templated insulin microparticles S. Schmidt a,b, K. Uhlig a, C. Duschl a, D. Volodkin a,⇑ a b
Fraunhofer Institut für Biomedizinische Technik, Am Mühlenberg 13, 14476 Potsdam-Golm, Germany Universität Leipzig, Institut für Biochemie, Johannisalle 21–23, 04103 Leipzig, Germany
a r t i c l e
i n f o
Article history: Received 20 August 2013 Received in revised form 23 October 2013 Accepted 14 November 2013 Available online 22 November 2013 Keywords: Microparticles Micromechanics Biopharmaceuticals Hard templating Microgels
a b s t r a c t Therapeutic proteins are an integral part of today’s pharmaceutical practice, but they still present challenges from the drug delivery point of view. In this work, a new approach is studied based on hard templating for fabrication of microparticles composed of pure insulin, which may enable effective delivery, for instance pulmonary delivery. The approach is both simple and versatile: the protein particles are prepared by selective precipitation into porous CaCO3 microtemplates, followed by full decomposition of the template at the isoelectric point of the protein (pH 5.2). Control over the main material parameters (mechanical properties, porosity, morphology and stability at physiological conditions) are critical for the envisioned application in drug delivery. It is demonstrated that these critical parameters can be significantly tuned by a slight final pH variation around the isoelectric point (pH range 4–6) and by the denaturation degree of insulin. Electrostatic interactions and inter-protein crosslinking in the protein particles as well as their internal structure are considered, to explain the variation in the particle properties. The particle property parameters are explored using atomic force microscopy, optical microscopy and circular dichroism spectra. Finally, phagocytic clearance of the protein particles in vitro was studied to explore possible enhancements in particle fabrication to improve the efficiency of insulin delivery by inhalation. Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction Biopharmaceuticals, i.e. molecules that are based on sugars, nucleic acids or proteins, hold great promise for advanced medical therapies. To date, virtually all existing biopharmaceuticals are based on proteins or peptides [1,2]. The significance of proteinbased drugs is, to a large extent, due to techniques such as genetic engineering, which allow virtually any class of pharmaceutically active proteins to be produced on a large scale, including hormones, vaccines, cytokines, growth factors, antibodies or antibiotics [3]. However, administration and handling of protein-based drugs pose strict limits to the development of new therapies because of the inherent side effects, e.g. inflammatory response, fast degradation, low retention time, limited control over the release profile and drug bioactivity. Therefore, future developments in biopharmaceuticals are directed towards microscale engineering of protein-based therapeutic compounds to reduce the abovementioned limitations [4–7]. One obvious approach would be formulating protein-based drugs into particulate form [8–10]. By controlling parameters such as size, density and mechanical properties, one can modulate the release profile and increase protein ⇑ Corresponding author. E-mail address: [email protected] (D. Volodkin).
mechanical stability and stability against degradation. Advanced particulate systems may also include stealth properties and specific release in response to controlled stimuli in vivo [5]. Classic techniques toward microparticulate proteins are lyophilization [11], spray drying [12] or fabrication of pure protein crystals [13]. Protein aggregates can be stabilized by polymers to form microparticles as well [14–16]. However, although these techniques yield microparticulate proteins, the prepared particles are often polydisperse, e.g. in size and density. This hampers control over the release profile of the active component and other relevant parameters such as mechanical properties, colloidal stability and retention time. As an alternative, templating techniques are employed to generate materials with tailored properties and improved uniformity of the particles’ material parameters: size, morphology, composition and porosity [17,18]. This work focuses on so-called ‘‘hard templating’’, and the general fabrication procedure can be summarized by two steps: (1) a ‘‘hard’’ but porous inorganic or polymeric template particle is infiltrated with the desired protein species; (2) the inorganic particle is subsequently removed, yielding precise replicates of the inverse template morphology initially. Using this approach, the material parameters of interest can be controlled simply by selection of the template type. For example, the mechanical properties or therapeutic payload can be conveniently tuned when varying the
1742-7061/$ - see front matter Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actbio.2013.11.011
1424
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
template porosity. Templating via porous CaCO3 vaterite particles as hard templates has been identified as a suitable approach for the preparation of fragile biomolecular particles, owing to the template removal in mild conditions [19]. Other template systems often involve hash conditions, e.g. pyrolysis or hydrofluoric acid etching for the removal of polymer [17] or silica templates [20]. Moreover, by controlling the salt concentration and/or stirring speed and time during synthesis, the morphology of CaCO3 microtemplates and thus the size of the resulting templated protein particles can be controlled in a wide range [21]. Furthermore, CaCO3 particles can be obtained with a narrow size distribution and well-defined internal morphology, making them useful as templates for biomolecular species, in particular [22–24]. Most importantly, the CaCO3 templates can be removed in slightly acetic conditions (pH < 6) or by mild treatment with complexing agents such as EDTA. Because of these advantages, templating by porous sacrificial microparticles composed of CaCO3 is now a well-established strategy also for the fabrication of various microparticulate systems, including capsules made of synthetic polymers [25], biopolymers [26,27] and single-component capsules [28,29]. Recently, it was shown that pure, monodisperse protein particles can also be obtained by CaCO3 templating without any additional stabilizers [30,31]. Such unmodified pure protein systems are not stabilized by chemical crosslinkers or other components, but the protein particles are stable as a result of physical inter-protein interaction at the isoelectric point. For many biomedical applications, such additive-free proteins are favored over heterogeneous systems, because additional components, e.g. crosslinkers, may hamper the activity of the protein or lead to undesired side effects, not to mention drug approval issues. The preparation of these pure protein particles makes use of the differential solubility of the protein payload and the CaCO3 matrix, depending on the pH of the solution (Fig. 1). Previous work highlighted the synthesis of pure insulin microparticles [30] and the mechanism of precipitation in the porous CaCO3 matrix [31]. The present work also focuses on insulin-based microparticles because of the growing demand for improved insulin therapies due to an increasing numbers of patients suffering from diabetes. The aim of this work is to fabricate microparticulate insulin, employing well-controlled material properties (size, structure, stability and mechanics) that are important for the application as a drug. For example, it is known that the size and stiffness of drug carriers may strongly affect the release characteristics (e.g. circulation time) [32] and also their immune response [33]. These parameters may also affect particle adhesion on surfaces and thus colloidal stability, shelf life and the ability to target specific tissues. Consequently, the mechanical properties, porosity, de-
gree of swelling, size, stability and release profile are controlled by adjusting the pH and degree of insulin denaturation. For example, circular dichroism (CD) spectra indicate large differences in activity and solution stability when insulin is stored at high pH (>10) greatly affecting the mechanics and stability of the particles. On a change in pH, the particles exhibit different surface charge, and the swelling behavior is also changed, thus indicating changes in stability in physiological conditions. The results demonstrate new design strategies for protein particle synthesis with tunable material properties. Via atomic force microscopy (AFM) and optical microscopy, these properties are next tested in different media. After characterization of the material properties, the uptake and degradation of the particles by macrophages is tested. The idea is to test the rate of particle uptake as a function of the particle size in order to identify morphologies with the smallest uptake propensity to improve delivery and activity of the drug in vivo. 2. Experimental section 2.1. Particle preparation FITC-labeled (I2383) and unlabeled insulin (I5500) from bovine pancreas with a content of 0.5% zinc were purchased from Sigma– Aldrich (Germany). CaCO3 particles were prepared as described previously [31]. Controlled denaturation of insulin was conducted in solution at pH 11 via addition of NaOH and incubation at 4 °C for 48 h. Non-denatured particle samples were prepared by immediate processing of fresh insulin solutions at pH 9.5. Insulin particle preparation was based on hard templating, as described in detail previously [31]. Briefly, first, at pH > 9 (adjusted by NaOH/HCl), the CaCO3 templates (5 lm in diameter) were dispersed in protein (insulin) solution. The initial protein/CaCO3 mass ratio was 15%. Then the pH was decreased by controlled addition of 50 mM acetic acid buffer (pH 5.2) via a peristaltic pump at a slow rate, resulting in gradual insulin precipitation and dissolution of the templates (Fig. 1). The pH titration was finished after 45 min, after reaching a pH of 5.2, if not stated otherwise. After synthesis, the particles were purified by 1 day dialysis (Float-A-Lyser G2 dialysis tubes, cut-off 3.5 kDa, SpectraPor, USA) in 2 liters of 10 mM acetic acid pH 5.2. 2.2. Characterization of material parameters Optical microscopy was conducted on an Olympus IX 71 equipped with a Zeiss HRm camera. Phase contrast and fluorescence images were collected with a 20 Objective (PLN 20XPH, N.A. 0.4, Olympus).
Fig. 1. Fabrication of insulin microspheres by templating via CaCO3 particles. Starting with the empty templates, the insulin solution is added at pH 9.5. Then, the pH is gradually decreased by addition of acetic acid buffer, resulting in selective precipitation that is completed at pH 6.5. Significant dissolution of the template particles starts at the same pH and is typically continued up to the isoelectric point of insulin at pH 5.2 to dissolve CaCO3 completely. Under these conditions, the particle hydrophobicity is maximum and the particle collapses. Scale bar: 2 lm.
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
2.3. AFM micromechanical studies The mechanics of the insulin particles were measured via colloidal probe AFM. Colloidal silica particles (Microparticles GmbH, Berlin, Germany) with a diameter of 30 lm served as force probes, glued to the apex of the AFM cantilevers with a spring constant of 0.03 N m 1 (CSC12, tipless, MicroMash, Estonia). Prior to attachment of the probe, the spring constant was determined according to the thermal noise [34] and the Sader [35] method. The cantilevers used in the measurements showed a deviation between both methods of <10%. The protein particles were allowed to sediment and adhere onto a polyethyleneimine-coated coverslip [36], by which they are immobilized. The AFM head was mounted on an optical microscope (IX71, Olympus, Japan). Using bright-field optics and also using the autofluorescence of the protein particles, the colloidal probe (37 lm in diameter) was positioned at the apex of the microgels in order to perform the AFM force measurement. The measurement was conducted using an approach speed of 1 lm s 1, applying peak forces of 80 nN. CD spectra were measured on a Jasco J-715 (Japan) spectrometer, using a quartz cuvette with an optical path length of 0.1 cm. The spectra were recorded in the wavelength interval from 190 to 250 nm, with 0.5-nm step resolution. Five scans were accumulated for one spectrum. All sample solutions were prepared in 10 mM phosphate buffer, pH 7.4. The background spectra of the pure buffer were subtracted. The insulin concentration was adjusted to 10 lM. To adjust the concentration of particle samples, the amount of particulate insulin was determined via UV–VIS spectroscopy. The measured CD signal was transformed to mean residue molar ellipticity [h] and divided by the number of peptide bonds to obtain [h]. UV–VIS spectroscopy was conducted with a Jasco V 400 spectrophotometer in 1 cm plastic cuvettes to compare the amount of insulin starting material with the amount of particulate insulin. 2.4. Cell assay The alveolar macrophages CRL-2192 (ATCC, Germany) were cultivated in Ham’s F12 medium with l-glutamine (Sigma–Aldrich, Germany) containing 1.5 g l 1 sodium bicarbonate, FCS (15%), penicillin/streptomycin (1%) (all Biochrom, Germany) at 37 °C and 5% CO2. To perform the cellular uptake of protein particles and dyed polystyrene microspheres (1 lm, 5 lm, 10 lm; Polysciences, USA) of various sizes, the macrophages were seeded in wells of Nunc chambers (Thermo Scientific, USA) with a concentration of 1 105 cells cm 2 and cultured overnight. The cells were incubated with particles for 5 h. Afterwards, the cell membrane was stained with a lipophilic dye, octadecyl rhodamine B (Invitrogen, USA), to visualize the particles engulfment. The assay was conducted on a Zeiss LSM 510 and 63 Objective (Zeiss Antiflex Neofluar, N.A. 1.2).
1425
insulin accumulates in the templates. Further decrease in pH then leads to dissolution of the CaCO3 and formation of pure protein particles. When the isoelectric point of insulin is reached (pH 5.2), the particle becomes hydrophobic and collapses, as shown earlier [31]. Here, the present authors now modulated the porosity and the degree of denaturation of the insulin particles via adjustment of the pH during preparation to allow control over their stability and release behavior. Overall, four different particle systems could be prepared: (1) non-denatured and compact; (2) non-denatured and swollen; (3) denatured and compact; and (4) denatured and swollen. Particle system (1) was obtained using the general preparation procedure, as explained above. Particles with a greater degree of swelling (2) and (4) were prepared by finalizing the titration at pH 5.7, well above the isoelectric point of insulin (pH 5.2). Measurements showed an effective zeta-potential of 10 mV (non-denatured) and 20 mV (denatured) at this pH. Therefore, these particles bear electrostatic charge as measured by zeta potential (Fig. 2). The charge prevents strong hydrophobic interaction thereby stabilizing the porous structure. As a result, these particles exhibit a higher degree of swelling (Fig. 3). The decreased zeta potential of the non-denatured particles compared with denatured particles in the pH range 5.5–6.0 could be caused by swelling that is stronger for non-denatured particles (swelling reduces the effective surface charge). Denatured particle systems (3) and (4) were obtained by treating the soluble insulin before titration in solution with a high pH value (pH 10.5 adjusted by NaOH) and storing for 2 days at 4 °C. At elevated pH, the cleavage and reshuffling of disulfide bond becomes prominent, leading to efficient denaturation of insulin [37].
3.2. Denaturation, swelling and stability of the particles in different media After preparation of the different particle systems, analytic methods were used to study the degree of denaturation and stability (release) of the particles in different media. CD measurements allowed a comparison of the degree of denaturation between untreated non-particulate (soluble) insulin and insulin particles prepared from denatured and non-denatured insulin. All measurements were conducted at pH 7.4 (phosphate buffer) and at the same insulin concentration, which was adjusted according to the UV-adsoption bands at 278 nm (Fig. 4A). As expected, the CD spectra showed almost complete denaturation for particles that were prepared with insulin incubated in high pH solutions. Denaturation of insulin can be readily explained by breakage of thiol bonds that stabilize the helical structures of the protein. Breakage and reshuffling of the thiol bonds is the principal mechanism for
3. Results and discussion 3.1. Sample preparation In this work, insulin was chosen as a model protein that can be readily synthesized into particulate form via pH titration in the presence of CaCO3 microtemplates (Fig. 1) [30]. The general preparation procedure is as follows: Starting from pH 9.5, solvated insulin is mixed with the CaCO3 templates 5 lm in diameter. Then, acetic acid buffer with a pH of 5.2 is slowly added to the mixture. As the pH of the insulin/CaCO3 particle mixture decreases, the solubility of insulin is reduced and it precipitates. Nucleation of precipitation is enhanced within the porous template, by which the
Fig. 2. Zeta potential indicates the pI of the particles at pH 5.2 and agrees with the pI of native insulin. Denatured compact particles (black) and non-denatured compact (gray) were initially prepared at pH 5.2, and the pH value has been adjusted further in the range 4–6.
1426
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
Fig. 3. (A) CD spectra indicate denaturation of insulin. Denatured particles were prepared by storing insulin at pH 10.5 before particle preparation. (B) Different solubility of non-denatured and denatured insulin particles in solution of physiological pH (7.4). The images were taken after 1, 5 and 10 min. Note that the non-denatured particles completely dissolved after 20 min (data not shown). Collapse at isoelectric point condition was observed for both particle types.
Fig. 4. UV–VIS measurements to determine the amount (yield) of insulin transformed into particulate and determination of the amount of released insulin in PBS solution after 60 min incubation. Solid lines in spectra (A) and (B) represent the amount of insulin applied for particle synthesis (bold solid lines) and the amount of insulin contained in the particles after synthesis (thin solid lines). Dashed lines in spectra (A) and (B) represent the amount of insulin before release in PBS (bold dashed lines) and the amount of released insulin in the supernatant of the PBS solution (thin dashed lines): (A) non-denatured insulin particles (gray); (B) denatured insulin particles (black); (C) summary of the UV data.
insulin denaturation at elevated pH [37]. The CD spectra also suggest a substantial decrease in reduction in a-helix [38] when the pure insulin sample is compared with the particle sample prepared with non-denatured insulin. In the experiment, a decrease in ellipticity between the pure insulin sample and the templated insulin of 35% (comparing the peaks at 210 and 200 nm) was always observed. This was, to some extent, unexpected, as special care was taken during particle synthesis to titrate insulin immediately from high pH conditions (denaturation) to the acidic regime (stable). An explanation for the pronounced degree of denaturation during particle synthesis might be protein adsorption to the interfaces of the charged calcium carbonate, which may result in irreversible change in the protein secondary structure [39]. It should be noted that denatured particles do not fully dissolve in PBS. Therefore, the non-dissolving protein aggregates were removed by centrifugation to study dissolved protein by CD exclusively. Partial denaturation of the particulate protein might be a reason for the bimodal release profile in PBS, as reported earlier [31]. An important parameter for pulmonary delivery of insulin is the size of the particles, which also controls the release rate of insulin [31]. As explained above, the present authors were able to control the size and degree of swelling of the particles, adjusting the final pH of the solution for both denatured and non-denatured insulin
particles. As shown in Fig. 3B, directly after titration to pH 5.7, the size of the particles is on the order of 5 lm, similar to the size of the CaCO3 templates. At this pH, the particles were stable over several weeks without changes in morphology. When the pH was decreased to 5.2 (isoelectric point), the particle diameter decreased by a factor of 2. This showed that the appropriate diameter for pulmonary delivery, typically 1–5 lm [40,41], can be conveniently adjusted by choosing the titration endpoint without the need to use templates of different sizes. Microscopic observation of the particle dissolution at pH 7.4 showed that non-denatured particles (1) and (3) dissolved readily within 20 min, as expected from previous release studies [31]. This is in accordance with the envisioned concept for release of insulin from particles in a slightly acidic storage solution (pH 5–6) to physiological pH of the lung (pH 7.4). However, denatured particle systems (2) and (4) did not completely dissolve at a pH of 7.4, but instead they swelled to a diameter of 8–10 lm. This suggests that the reshuffling of disulfide bonds at elevated pH has led to crosslinked insulin particles that can then be swollen and are stable to some extent at physiological pH. To analyze the release characteristics further, the insulin release from collapsed denatured and non-denatured particles stored at pH 5.2 was compared. The relative amount of insulin in the
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
supernatant of a centrifuged particle dispersion in storage buffer (pH 5.2) and PBS (pH 7.4) was measured using UV-VIS spectroscopy. This allows the amount of released insulin from non-denatured and denatured particles to be determined (dashed lines Fig. 4A and B). Owing to the good solubility of non-denatured insulin particles in PBS, the amount of released insulin reached 90%, whereas the denatured particles released only 25% of their insulin because of reduced solubility (Fig. 4C), in line with the results from optical microscopy (Fig. 3). As a result of the insolubility, the denatured particle component must be considered pharmaceutically inactive, as it does not bind to the respective kinase receptors that modulate glucose levels. In addition, the yield of the particle synthesis was determined by comparing the amount of insulin applied for synthesis with the amount of insulin contained in the particles after synthesis (solid lines Fig. 4A and B). Overall, the data show that 40% of the insulin was transformed to particulate form after the pH titration step (Fig. 4C). This means that 60% of the insulin was not precipitated during pH titration and stayed in the soluble phase. When a larger concentration of insulin and template particles is used before titration, the amount of insulin remaining in the soluble phase can be decreased, and the yield can be increased (data not shown). 3.3. Mechanical properties of insulin particles Micromechanical study using AFM is a well-established tool for studying the mechanical and adhesion properties of microparticles for biomedical applications [21,42–45]. So far, it has not been applied in great detail for the mechanical analysis of protein microparticles prepared by CaCO3 templating. However, the mechanical properties of the particles are crucial for the envisioned applications, as they play an important role in adhesion of the particles to surfaces, colloidal stability, circulation time in the vascular system, and ability to overcome mechanical barriers [32]. Therefore, the mechanical properties were analyzed first, comparing the deformability of swollen (porous) and compact insulin particles via colloidal probe AFM. As explained, the compact particle system was titrated from a high pH regime to isoelectric point conditions, where the particle collapsed (Fig. 1). For the swollen particle system, the titration was conducted only to pH 5.7, with the aim of stabilizing the porous template structure via electrostatic repulsion between negatively charged protein molecules. For colloidal probe AFM measurements, the insulin particles were first immobilized on a ‘‘sticky’’ PEI-coated glass slide in an acetic acid buffer (pH 4.0–5.8). Then the particles were deformed by applying 80 nN maximum force via colloidal probe AFM followed by retraction of the probe. After a dwell time of 60 s, another compression cycle was conducted. Typical force–deformation data
1427
Fig. 6. The elastic moduli of denatured and non-denatured collapsed insulin particles (stored at pH 5.2). At isoelectric point (pH 5.2) both particle systems have stiffnesses of several 100 kPa. When the pH is increased to 7.4, non-denatured particles start to dissolve (strong decrease in elastic modulus) and denatured particles swell (minor decrease in elastic modulus). Measurements at pH 7.4 were taken 2 min after increasing the pH from 5.2.
are depicted in Fig. 5, showing force curves from the first and tenth deformation cycles. Indeed, the force curves for the swollen/porous particles (titrated only to pH 5.7) show a clear shift in the range 75–150 nm in the contact point between probe and sample when comparing the first and tenth deformation cycle (Fig. 5A). This shift is the plastic deformation of the particle after repeated compression with 80 nN force. However, compact particles (titrated to pH 5.2, pI conditions) do not deform under this force (Fig. 5B), therefore no shift of the contact point was observed, indicating the absence of plastic deformation. As a result, not only the size but also the deformability of the particles can be controlled by slightly varying the endpoint pH during particle synthesis (Fig. 5C). It should be noted, however, that the softer, supposedly porous particles prepared at pH 5.7 deformed irreversibly. In terms of material stiffness, the porous particles (titrated to pH 5.7) show low elastic moduli (fitted by Hertzian model) in the first compression cycle on the order of 15 ± 9 kPa, which increases to several hundred kilopascal, similar to the elastic modulus of the collapsed particles titrated to pH 5.2 (Fig. 6). As show in Fig. 6, denaturation does not affect the mechanical properties of the protein particles in isoelectric point conditions. However, denaturation significantly influences particle stability at physiological pH. Here, the elastic modulus of non-denatured particles quickly drops below 1 kPa as a result of dissolution of the particles, whereas only a minor change in elastic modulus modulus is observed for denatured particles. In order to investigate the mechanical stability of the particles further in different media, the elastic moduli of completely
Fig. 5. The degree of plastic deformation depends on the preparation history. (A) Freshly prepared particles at pH 5.7 deform plastically, whereas (B) completely collapsed particles at pH 5.2 show no permanent deformation between the compression cycles. (C) Complete data set shows the average amount of plastic deformation as calculated from the shift in contact point between the first and tenth compression cycles. The particles were composed of non-denatured insulin.
1428
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
non-denatured particles quickly dissolved (Fig. 3), therefore exhibiting very low elastic moduli shortly after addition of the buffer. The elastic moduli for denatured particles decreased at pH 7.4 as well, but only by a factor of two compared with pH 5.2. This again showed that the denatured particles stay intact, but undergo swelling at physiological pH, decreasing their elastic modulus. Thus, pH variation may significantly affect both morphological (size) and mechanical properties of the insulin particles. Both can be at the same time modulated by the degree of denaturing. This may enable tuning of the stability, release characteristics and resistance to mechanical barriers of the protein particles delivered to a body. Fig. 7. Mechanical properties depend on pH; charged particles show larger moduli. Elastic moduli of the non-denatured collapsed particles between pH 4.0 and 6.0 (black). Preparation history and degree of denaturation strongly affect the mechanical properties at physiological pH; denatured particles at pH 7.4 (gray), non-denatured particles at pH 7.4 (light gray), both measured 2 min after pH adjustment.
collapsed particles were determined via colloidal probe AFM at different pH values (Fig. 7). In the pH range between 4.0 and 5.9 (acetic acid buffer), only moderate changes by a factor of two of the elastic modulus were observed. Quite surprisingly, the lowest elastic modulus was obtained in isoelectric point conditions for the collapsed particles. This seems in contradiction to the previous measurement at pH 5.2 and 5.7 on the collapsed and swollen/porous particle system (Fig. 5), which showed greater stiffness for collapsed particles at pH 5.2. However, here one is looking at particles that were already collapsed in isoelectric point conditions before performing AFM measurements at different pH values. From the deformation measurement, it was seen that the compact structure of the particle is irreversible also for pH values shifted from the isoelectric point condition. Therefore, the particles do not swell in the pH range 4.0–5.9, as confirmed by optical microscopy observations (Supporting Information S1). However, the zeta potential measurement showed that these systems are charged (Fig. 2). Therefore, the decrease in the elastic modulus towards the isoelectric point condition is explained by the absence of electrostatic repulsion between protein molecules within the particle. However, when changing the buffer system in order to mimic physiological pH (phosphate buffer), swelling of the particles and a strong decrease in the elastic modulus were observed accordingly. When the pH was changed from 5.2 to 7.4 via addition of phosphate buffer, the
3.4. Phagocytosis of insulin particles with different size Phagocytic clearance in the lung can be considered significant for the deactivation of pulmonary delivered protein particles [46]. If taken up by macrophages, the insulin particles will not be delivered to a bloodstream, this is why the uptake is considered a negative process for delivery and must be avoided. This is particularly important for prolonged delivery, when the time is sufficient for the macrophages to take the protein particles up and to activate further immune processes. In general, once a bead or pathogen is recognized by macrophages, the internalization into the cell takes 30 min [47], which is still in the time-frame of insulin release from the particles [31], especially for more stable, partially denatured insulin particles. Therefore, a rough test was performed on the recognition and uptake of the insulin particles by alveolar macrophages. In addition, it is known that recognition and uptake are strongly dependent on the size of the particles [48,49]. For practical reasons, insulin particles of different sizes were used (1, 3 and 8 lm), which are completely denatured (stable at pH 7.4) and show only small changes in mechanical properties (400 kPa) and size when transferred from the acidic storage solution to the cell medium at pH 7.4. This tested size range reflects the currently accessible range for hard templating with CaCO3 particles and is a favorable range for efficient pulmonary delivery [31]. For comparison, the uptake of polystyrene latex beads of similar size was investigated. The results show that all particle systems and the polystyrene control particles were taken up, regardless of the size of the particles (Fig. 8). A major fraction of particles were taken up already after 2 h of incubation. The invariance of uptake with respect to the size of the particles could be expected, as a significant
Fig. 8. Uptake of insulin (first row) and polystyrene micro particles (second row) of different diameters (from left to right column: 1, 3 and 8 lm for insulin and 1, 5 and 10 lm for polystyrene particles). All particles were engulfed by phagocytes after 5 h of incubation. Scale bar: 10 lm.
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
reduction of uptake is observed only for particles that are smaller than 100 nm [49]. Therefore, the results show that pure protein particles suited for pulmonary delivery (>1 lm in diameter) should release the active component within a few minutes (burst release) in order to avoid phagocytic clearance. In the case of the insulin system presented here, this can be achieved by particles composed of non-denatured insulin, as they dissolved completely (within 20 min), i.e. before they could be engulfed by phagocytes. However, for longer-acting insulin formulations, the dissolution process occurs within hours. Thus, when delivered by pulmonary routes, the activity of such systems could indeed be impaired by phagocytic clearance, as shown in the present experiment. In order to slow phagocytic clearance, much smaller (<0.1 lm) or larger (>20 lm) particle systems should be prepared. At the moment, however, CaCO3 templating is not suited for this size range, and pulmonary delivery routes as well as toxicity of such particle systems also have to be carefully analyzed [48]. As an alternative, the particles should be prepared to adapt so-called stealth properties such that recognition by macrophages is reduced. A typical strategy is the modification of the active component with polyethylene glycol (PEG). As shown in previous work hard CaCO3 templating with PEG, it is possible to prepare particles with very low non-specific interaction [50,51]. Therefore, future work will focus on such hybrid protein/ PEG particle systems.
4. Conclusion Pure insulin microparticles were prepared by hard templating on porous CaCO3 using a single trigger: pH (titration of insulin– CaCO3 mixture with pH from 9.5 to 5.2). The size and porosity of the pure insulin particles can be controlled by slight variation in the final pH (from 4 to 6) value around the insulin isoelectric point (5.2) as a result of extra charge doping. Compact and mesoporous protein particles possess different mechanical properties—elastic and plastic deformation, respectively—and deviate up to two orders of magnitude in elastic modulus. pH-induced (pH 10.5) denaturation of insulin was used to prepare protein particles with significantly different stability at physiological conditions. Thus, partial denaturing of the protein may be used to adjust the release profile in physiological conditions. However, it should be noted that the denatured part of the protein via randomization of disulfide bridges leads to insoluble and pharmaceutically inactive insulin. Therefore, it might be advantageous to modulate the solubility and release profile of insulin by other means, e.g. by co-precipitation with crosslinked polymers [14,51,52]. The in vitro test on alveolar macrophages showed that phagocytic clearance remains an issue for longer-acting particle formulations in the size range 1–10 lm. But such longer acting formulations are especially interesting from the medical therapy point of view. Therefore, future work will focus on the preparation of protein/PEG formulations prepared by hard CaCO3 templating. This approach is the most promising route, as previous work on the morphology [50,53] and application [51,52] of templated PEG particles has already shown. Considering the large amount of preceding work on CaCO3 templated protein particles, the next logical step is in vivo testing, e.g. with the well-established insulin particles presented here.
Acknowledgements DVV thanks the Alexander von Humboldt Foundation for support (AvH Fellowship and Sofja Kovalevskaja Program). The authors thank Dr. Jürgen Hartmann for the SEM imaging.
1429
Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Fig. 8 are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/j.actbio.2013. 11.011. Appendix B. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2013. 11.011. References [1] Brock A. Protein drugs: global markets and manufacturing technologies. BCC Research 2008; Report no.: BIO021C. [2] Leader B, Baca QJ, Golan DE. Protein therapeutics: a summary and pharmacological classification. Nat Rev Drug Discov 2008;7:21–39. [3] Walsh G. Biopharmaceuticals: biochemistry and biotechnology. 2nd ed. Chichester: Wiley-Blackwell; 2003. [4] Parveen S, Misra R, Sahoo SK. Nanoparticles: a boon to drug delivery, therapeutics, diagnostics and imaging. Nanomed Nanotechnol Biol Med 2012;8:147–66. [5] Putnam D. Polymers for gene delivery across length scales. Nat Mater 2006;5:439–51. [6] Becker AL, Johnston APR, Caruso F. Layer-by-layer-assembled capsules and films for therapeutic delivery. Small 2010;6:1836–52. [7] Yan Y, Such GK, Johnston APR, Best JP, Caruso F. Engineering particles for therapeutic delivery: prospects and challenges. ACS Nano 2012;6:3663–9. [8] Szlachcic A, Zakrzewska M, Otlewski J. Longer action means better drug: tuning up protein therapeutics. Biotechnol Adv 2011;29:436–41. [9] Yang SX, Yuan WE, Jin T. Formulating protein therapeutics into particulate forms. Expert Opin Drug Deliv 2009;6:1123–33. [10] Tan ML, Choong PFM, Dass CR. Recent developments in liposomes, microparticles and nanoparticles for protein and peptide drug delivery. Peptides 2010;31:184–93. [11] Wang W. Lyophilization and development of solid protein pharmaceuticals. Int J Pharm 2000;203:1–60. [12] Maa YF, Nguyen PA, Sweeney T, Shire SJ, Hsu CC. Protein inhalation powders: spray drying vs spray freeze drying. Pharm Res 1999;16:249–54. [13] Basu SK, Govardhan CP, Jung CW, Margolin AL. Protein crystals for the delivery of biopharmaceuticals. Expert Opin Biol Ther 2004;4:301–17. [14] Balabushevich NG, Pechenkin MA, Shibanova ED, Volodkin DV, Mikhalchik EV. Multifunctional polyelectrolyte microparticles for oral insulin delivery. Macromol Biosci 2013;13:1379–88. [15] Volodkin DV, Balabushevitch NG, Sukhorukov GB, Larionova NI. Model system for controlled protein release: pH-sensitive polyelectrolyte microparticles. STP Pharm Sci 2003;13:163–70. [16] Volodkin DV, Balabushevitch NG, Sukhorukov GB, Larionova NI. Inclusion of proteins into polyelectrolyte microparticles by alternative adsorption of polyelectrolytes on protein aggregates. Biochem-Moscow 2003;68:236–41. [17] Li Q, Retsch M, Wang J, Knoll W, Jonas U. Porous networks through colloidal templates. In: Broekmann PDKHSCA, editor. Templates in chemistry III. Berlin: Springer; 2009. p. 135–80. [18] Thomas A, Goettmann F, Antonietti M. Hard templates for soft materials: creating nanostructured organic materials. Chem Mater 2008;20:738–55. [19] Schmidt S, Volodkin D. Microparticulate biomolecules by mild CaCO3 templating. J Mater Chem B 2013;1:1210–8. [20] Wang YJ, Caruso F. Nanoporous protein particles through templating mesoporous silica spheres. Adv Mater 2006;18. 795-+. [21] Schmidt S, Behra M, Uhlig K, Madaboosi N, Hartmann L, Duschl C, et al. Mesoporous protein particles through colloidal CaCO3 templates. Adv Funct Mater 2013;23:116–23. [22] Petrov AI, Volodkin DV, Sukhorukov GB. Protein–calcium carbonate coprecipitation: a tool for protein encapsulation. Biotechnol Prog 2005;21: 918–25. [23] Volodkin DV, Larionova NI, Sukhorukov GB. Protein encapsulation via porous CaCO3 microparticles templating. Biomacromolecules 2004;5:1962–72. [24] Sukhorukov GB, Volodkin DV, Gunther AM, Petrov AI, Shenoy DB, Mohwald H. Porous calcium carbonate microparticles as templates for encapsulation of bioactive compounds. J Mater Chem 2004;14:2073–81. [25] Pechenkin MA, Mohwald H, Volodkin DV. PH- and salt-mediated response of layer-by-layer assembled PSS/PAH microcapsules: fusion and polymer exchange. Soft Matter 2012;8:8659–65. [26] Volodkin D, Skirtach A, Möhwald H. Bioapplications of light-sensitive polymer films and capsules assembled using the layer-by-layer technique. Polym Int 2012;61:673–9. [27] Volodkin D, Skirtach A, Mohwald H. LbL films as reservoirs for bioactive molecules. In: Borner HG, Lutz JF, editors. Bioactive surfaces. Berlin: Springer; 2011. p. 135–61.
1430
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
[28] Fujii A, Maruyama T, Ohmukai Y, Kamio E, Sotani T, Matsuyama H. Crosslinked DNA capsules templated on porous calcium carbonate microparticles. Colloids Surf A Physicochem Eng Asp 2010;356:126–33. [29] Han YY, Tong WJ, Zhang YY, Gao CY. Fabrication of chitosan single-component microcapsules with a micrometer-thick and layered wall structure by stepwise core-mediated precipitation. Macromol Rapid Commun 2012;33:326–31. [30] Volodkin DV, von Klitzing R, Möhwald H. Pure protein microspheres by calcium carbonate templating. Angew Chem Int Ed 2010;49:9258–61. [31] Volodkin DV, Schmidt S, Fernandes P, Larionova NI, Sukhorukov GB, Duschl C, et al. One-step formulation of protein microparticles with tailored properties: hard templating at soft conditions. Adv Funct Mater 2012;22:1914–22. [32] Merkel TJ, Jones SW, Herlihy KP, Kersey FR, Shields AR, Napier M, et al. Using mechanobiological mimicry of red blood cells to extend circulation times of hydrogel microparticles. Proc Natl Acad Sci USA 2011;108:586–91. [33] Goodridge HS, Reyes CN, Becker CA, Katsumoto TR, Ma J, Wolf AJ, et al. Activation of the innate immune receptor Dectin-1 upon formation of a ‘phagocytic synapse’. Nature 2011;472:U471–541. [34] Hutter JL, Bechhoefer J. Calibration of atomic-force microscope tips. Rev Sci Instrum 1993;64:1868–73. [35] Sader JE. Frequency response of cantilever beams immersed in viscous fluids with applications to the atomic force microscope. J Appl Phys 1998;84:64–76. [36] Schmidt S, Fernandes PAL, De Geest BG, Delcea M, Skirtach AG, Mohwald H, et al. Release properties of pressurized microgel templated capsules. Adv Funct Mater 2011;21:1411–8. [37] Brange J, Langkjaer L. Chemical stability of insulin. 3. Influence of excipients, formulation, and pH. Acta Pharm Nord 1992;4:149–58. [38] Pocker Y, Biswas SB. Conformational dynamics of insulin in solution. Circular dichroic studies. Biochemistry 1980;19:5043–9. [39] Mollmann SH, Jorgensen L, Bukrinsky JT, Elofsson U, Norde W, Frokjaer S. Interfacial adsorption of insulin – conformational changes and reversibility of adsorption. Eur J Pharm Sci 2006;27:194–204. [40] Khafagy E-S, Morishita M, Onuki Y, Takayama K. Current challenges in noninvasive insulin delivery systems: a comparative review. Adv Drug Deliv Rev 2007;59:1521–46. [41] Patton JS, Byron PR. Inhaling medicines: delivering drugs to the body through the lungs. Nat Rev Drug Discov 2007;6:67–74.
[42] Fernandes PAL, Schmidt S, Zeiser M, Fery A, Hellweg T. Swelling and mechanical properties of polymer gels with cross-linking gradient. Soft Matter 2010;6:3455–8. [43] Delcea M, Schmidt S, Palankar R, Fernandes PAL, Fery A, Möhwald H, et al. Mechanobiology: correlation between mechanical stability of microcapsules studied by AFM and impact of cell-induced stresses. Small 2010;6:2858–62. [44] Palankar R, Pinchasik B-E, Schmidt S, De Geest BG, Fery A, Mohwald H, et al. Mechanical strength and intracellular uptake of CaCO3-templated LbL capsules composed of biodegradable polyelectrolytes: the influence of the number of layers. J Mater Chem B 2013;1:1175–81. [45] Fery A, Weinkamer R. Mechanical properties of micro- and nanocapsules: single-capsule measurements. Polymer 2007;48:7221–35. [46] Edwards DA, Ben-Jebria A, Langer R. Recent advances in pulmonary drug delivery using large, porous inhaled particles. J Appl Physiol 1998;85:379–85. [47] Blanchette CD, Woo YH, Thomas C, Shen N, Sulchek TA, Hiddessen AL. Decoupling internalization, acidification and phagosomal–endosomal/ lysosomal fusion during phagocytosis of InlA coated beads in epithelial cells. PLoS One 2009:4. [48] Renwick LC, Donaldson K, Clouter A. Impairment of alveolar macrophage phagocytosis by ultrafine particles. Toxicol Appl Pharmacol 2001;172: 119–27. [49] Zauner W, Farrow NA, Haines AMR. In vitro uptake of polystyrene microspheres: effect of particle size, cell line and cell density. J Control Release 2001;71:39–51. [50] Behra M, Schmidt S, Hartmann J, Volodkin DV, Hartmann L. Synthesis of porous PEG microgels using CaCO3 microspheres as hard templates. Macromol Rapid Commun 2012;33:1049–54. [51] Behra M, Azzouz N, Schmidt S, Volodkin DV, Mosca S, Chanana M, et al. Magnetic Porous Sugar-Functionalized PEG Microgels for Efficient Isolation and Removal of Bacteria from Solution. Biomacromolecules 2013;14:1927–35. [52] Pussak D, Ponader D, Mosca S, Ruiz SV, Hartmann L, Schmidt S. Mechanical carbohydrate sensors based on soft hydrogel particles. Angew Chem Int Ed 2013;52:6084–7. [53] Pussak D, Behra M, Schmidt S, Hartmann L. Synthesis and functionalization of poly(ethylene glycol) microparticles as soft colloidal probes for adhesion energy measurements. Soft Matter 2012;8:1664–72.
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Stability and cell uptake of calcium carbonate templated insulin microparticles S. Schmidt a,b, K. Uhlig a, C. Duschl a, D. Volodkin a,⇑ a b
Fraunhofer Institut für Biomedizinische Technik, Am Mühlenberg 13, 14476 Potsdam-Golm, Germany Universität Leipzig, Institut für Biochemie, Johannisalle 21–23, 04103 Leipzig, Germany
a r t i c l e
i n f o
Article history: Received 20 August 2013 Received in revised form 23 October 2013 Accepted 14 November 2013 Available online 22 November 2013 Keywords: Microparticles Micromechanics Biopharmaceuticals Hard templating Microgels
a b s t r a c t Therapeutic proteins are an integral part of today’s pharmaceutical practice, but they still present challenges from the drug delivery point of view. In this work, a new approach is studied based on hard templating for fabrication of microparticles composed of pure insulin, which may enable effective delivery, for instance pulmonary delivery. The approach is both simple and versatile: the protein particles are prepared by selective precipitation into porous CaCO3 microtemplates, followed by full decomposition of the template at the isoelectric point of the protein (pH 5.2). Control over the main material parameters (mechanical properties, porosity, morphology and stability at physiological conditions) are critical for the envisioned application in drug delivery. It is demonstrated that these critical parameters can be significantly tuned by a slight final pH variation around the isoelectric point (pH range 4–6) and by the denaturation degree of insulin. Electrostatic interactions and inter-protein crosslinking in the protein particles as well as their internal structure are considered, to explain the variation in the particle properties. The particle property parameters are explored using atomic force microscopy, optical microscopy and circular dichroism spectra. Finally, phagocytic clearance of the protein particles in vitro was studied to explore possible enhancements in particle fabrication to improve the efficiency of insulin delivery by inhalation. Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction Biopharmaceuticals, i.e. molecules that are based on sugars, nucleic acids or proteins, hold great promise for advanced medical therapies. To date, virtually all existing biopharmaceuticals are based on proteins or peptides [1,2]. The significance of proteinbased drugs is, to a large extent, due to techniques such as genetic engineering, which allow virtually any class of pharmaceutically active proteins to be produced on a large scale, including hormones, vaccines, cytokines, growth factors, antibodies or antibiotics [3]. However, administration and handling of protein-based drugs pose strict limits to the development of new therapies because of the inherent side effects, e.g. inflammatory response, fast degradation, low retention time, limited control over the release profile and drug bioactivity. Therefore, future developments in biopharmaceuticals are directed towards microscale engineering of protein-based therapeutic compounds to reduce the abovementioned limitations [4–7]. One obvious approach would be formulating protein-based drugs into particulate form [8–10]. By controlling parameters such as size, density and mechanical properties, one can modulate the release profile and increase protein ⇑ Corresponding author. E-mail address: [email protected] (D. Volodkin).
mechanical stability and stability against degradation. Advanced particulate systems may also include stealth properties and specific release in response to controlled stimuli in vivo [5]. Classic techniques toward microparticulate proteins are lyophilization [11], spray drying [12] or fabrication of pure protein crystals [13]. Protein aggregates can be stabilized by polymers to form microparticles as well [14–16]. However, although these techniques yield microparticulate proteins, the prepared particles are often polydisperse, e.g. in size and density. This hampers control over the release profile of the active component and other relevant parameters such as mechanical properties, colloidal stability and retention time. As an alternative, templating techniques are employed to generate materials with tailored properties and improved uniformity of the particles’ material parameters: size, morphology, composition and porosity [17,18]. This work focuses on so-called ‘‘hard templating’’, and the general fabrication procedure can be summarized by two steps: (1) a ‘‘hard’’ but porous inorganic or polymeric template particle is infiltrated with the desired protein species; (2) the inorganic particle is subsequently removed, yielding precise replicates of the inverse template morphology initially. Using this approach, the material parameters of interest can be controlled simply by selection of the template type. For example, the mechanical properties or therapeutic payload can be conveniently tuned when varying the
1742-7061/$ - see front matter Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actbio.2013.11.011
1424
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
template porosity. Templating via porous CaCO3 vaterite particles as hard templates has been identified as a suitable approach for the preparation of fragile biomolecular particles, owing to the template removal in mild conditions [19]. Other template systems often involve hash conditions, e.g. pyrolysis or hydrofluoric acid etching for the removal of polymer [17] or silica templates [20]. Moreover, by controlling the salt concentration and/or stirring speed and time during synthesis, the morphology of CaCO3 microtemplates and thus the size of the resulting templated protein particles can be controlled in a wide range [21]. Furthermore, CaCO3 particles can be obtained with a narrow size distribution and well-defined internal morphology, making them useful as templates for biomolecular species, in particular [22–24]. Most importantly, the CaCO3 templates can be removed in slightly acetic conditions (pH < 6) or by mild treatment with complexing agents such as EDTA. Because of these advantages, templating by porous sacrificial microparticles composed of CaCO3 is now a well-established strategy also for the fabrication of various microparticulate systems, including capsules made of synthetic polymers [25], biopolymers [26,27] and single-component capsules [28,29]. Recently, it was shown that pure, monodisperse protein particles can also be obtained by CaCO3 templating without any additional stabilizers [30,31]. Such unmodified pure protein systems are not stabilized by chemical crosslinkers or other components, but the protein particles are stable as a result of physical inter-protein interaction at the isoelectric point. For many biomedical applications, such additive-free proteins are favored over heterogeneous systems, because additional components, e.g. crosslinkers, may hamper the activity of the protein or lead to undesired side effects, not to mention drug approval issues. The preparation of these pure protein particles makes use of the differential solubility of the protein payload and the CaCO3 matrix, depending on the pH of the solution (Fig. 1). Previous work highlighted the synthesis of pure insulin microparticles [30] and the mechanism of precipitation in the porous CaCO3 matrix [31]. The present work also focuses on insulin-based microparticles because of the growing demand for improved insulin therapies due to an increasing numbers of patients suffering from diabetes. The aim of this work is to fabricate microparticulate insulin, employing well-controlled material properties (size, structure, stability and mechanics) that are important for the application as a drug. For example, it is known that the size and stiffness of drug carriers may strongly affect the release characteristics (e.g. circulation time) [32] and also their immune response [33]. These parameters may also affect particle adhesion on surfaces and thus colloidal stability, shelf life and the ability to target specific tissues. Consequently, the mechanical properties, porosity, de-
gree of swelling, size, stability and release profile are controlled by adjusting the pH and degree of insulin denaturation. For example, circular dichroism (CD) spectra indicate large differences in activity and solution stability when insulin is stored at high pH (>10) greatly affecting the mechanics and stability of the particles. On a change in pH, the particles exhibit different surface charge, and the swelling behavior is also changed, thus indicating changes in stability in physiological conditions. The results demonstrate new design strategies for protein particle synthesis with tunable material properties. Via atomic force microscopy (AFM) and optical microscopy, these properties are next tested in different media. After characterization of the material properties, the uptake and degradation of the particles by macrophages is tested. The idea is to test the rate of particle uptake as a function of the particle size in order to identify morphologies with the smallest uptake propensity to improve delivery and activity of the drug in vivo. 2. Experimental section 2.1. Particle preparation FITC-labeled (I2383) and unlabeled insulin (I5500) from bovine pancreas with a content of 0.5% zinc were purchased from Sigma– Aldrich (Germany). CaCO3 particles were prepared as described previously [31]. Controlled denaturation of insulin was conducted in solution at pH 11 via addition of NaOH and incubation at 4 °C for 48 h. Non-denatured particle samples were prepared by immediate processing of fresh insulin solutions at pH 9.5. Insulin particle preparation was based on hard templating, as described in detail previously [31]. Briefly, first, at pH > 9 (adjusted by NaOH/HCl), the CaCO3 templates (5 lm in diameter) were dispersed in protein (insulin) solution. The initial protein/CaCO3 mass ratio was 15%. Then the pH was decreased by controlled addition of 50 mM acetic acid buffer (pH 5.2) via a peristaltic pump at a slow rate, resulting in gradual insulin precipitation and dissolution of the templates (Fig. 1). The pH titration was finished after 45 min, after reaching a pH of 5.2, if not stated otherwise. After synthesis, the particles were purified by 1 day dialysis (Float-A-Lyser G2 dialysis tubes, cut-off 3.5 kDa, SpectraPor, USA) in 2 liters of 10 mM acetic acid pH 5.2. 2.2. Characterization of material parameters Optical microscopy was conducted on an Olympus IX 71 equipped with a Zeiss HRm camera. Phase contrast and fluorescence images were collected with a 20 Objective (PLN 20XPH, N.A. 0.4, Olympus).
Fig. 1. Fabrication of insulin microspheres by templating via CaCO3 particles. Starting with the empty templates, the insulin solution is added at pH 9.5. Then, the pH is gradually decreased by addition of acetic acid buffer, resulting in selective precipitation that is completed at pH 6.5. Significant dissolution of the template particles starts at the same pH and is typically continued up to the isoelectric point of insulin at pH 5.2 to dissolve CaCO3 completely. Under these conditions, the particle hydrophobicity is maximum and the particle collapses. Scale bar: 2 lm.
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
2.3. AFM micromechanical studies The mechanics of the insulin particles were measured via colloidal probe AFM. Colloidal silica particles (Microparticles GmbH, Berlin, Germany) with a diameter of 30 lm served as force probes, glued to the apex of the AFM cantilevers with a spring constant of 0.03 N m 1 (CSC12, tipless, MicroMash, Estonia). Prior to attachment of the probe, the spring constant was determined according to the thermal noise [34] and the Sader [35] method. The cantilevers used in the measurements showed a deviation between both methods of <10%. The protein particles were allowed to sediment and adhere onto a polyethyleneimine-coated coverslip [36], by which they are immobilized. The AFM head was mounted on an optical microscope (IX71, Olympus, Japan). Using bright-field optics and also using the autofluorescence of the protein particles, the colloidal probe (37 lm in diameter) was positioned at the apex of the microgels in order to perform the AFM force measurement. The measurement was conducted using an approach speed of 1 lm s 1, applying peak forces of 80 nN. CD spectra were measured on a Jasco J-715 (Japan) spectrometer, using a quartz cuvette with an optical path length of 0.1 cm. The spectra were recorded in the wavelength interval from 190 to 250 nm, with 0.5-nm step resolution. Five scans were accumulated for one spectrum. All sample solutions were prepared in 10 mM phosphate buffer, pH 7.4. The background spectra of the pure buffer were subtracted. The insulin concentration was adjusted to 10 lM. To adjust the concentration of particle samples, the amount of particulate insulin was determined via UV–VIS spectroscopy. The measured CD signal was transformed to mean residue molar ellipticity [h] and divided by the number of peptide bonds to obtain [h]. UV–VIS spectroscopy was conducted with a Jasco V 400 spectrophotometer in 1 cm plastic cuvettes to compare the amount of insulin starting material with the amount of particulate insulin. 2.4. Cell assay The alveolar macrophages CRL-2192 (ATCC, Germany) were cultivated in Ham’s F12 medium with l-glutamine (Sigma–Aldrich, Germany) containing 1.5 g l 1 sodium bicarbonate, FCS (15%), penicillin/streptomycin (1%) (all Biochrom, Germany) at 37 °C and 5% CO2. To perform the cellular uptake of protein particles and dyed polystyrene microspheres (1 lm, 5 lm, 10 lm; Polysciences, USA) of various sizes, the macrophages were seeded in wells of Nunc chambers (Thermo Scientific, USA) with a concentration of 1 105 cells cm 2 and cultured overnight. The cells were incubated with particles for 5 h. Afterwards, the cell membrane was stained with a lipophilic dye, octadecyl rhodamine B (Invitrogen, USA), to visualize the particles engulfment. The assay was conducted on a Zeiss LSM 510 and 63 Objective (Zeiss Antiflex Neofluar, N.A. 1.2).
1425
insulin accumulates in the templates. Further decrease in pH then leads to dissolution of the CaCO3 and formation of pure protein particles. When the isoelectric point of insulin is reached (pH 5.2), the particle becomes hydrophobic and collapses, as shown earlier [31]. Here, the present authors now modulated the porosity and the degree of denaturation of the insulin particles via adjustment of the pH during preparation to allow control over their stability and release behavior. Overall, four different particle systems could be prepared: (1) non-denatured and compact; (2) non-denatured and swollen; (3) denatured and compact; and (4) denatured and swollen. Particle system (1) was obtained using the general preparation procedure, as explained above. Particles with a greater degree of swelling (2) and (4) were prepared by finalizing the titration at pH 5.7, well above the isoelectric point of insulin (pH 5.2). Measurements showed an effective zeta-potential of 10 mV (non-denatured) and 20 mV (denatured) at this pH. Therefore, these particles bear electrostatic charge as measured by zeta potential (Fig. 2). The charge prevents strong hydrophobic interaction thereby stabilizing the porous structure. As a result, these particles exhibit a higher degree of swelling (Fig. 3). The decreased zeta potential of the non-denatured particles compared with denatured particles in the pH range 5.5–6.0 could be caused by swelling that is stronger for non-denatured particles (swelling reduces the effective surface charge). Denatured particle systems (3) and (4) were obtained by treating the soluble insulin before titration in solution with a high pH value (pH 10.5 adjusted by NaOH) and storing for 2 days at 4 °C. At elevated pH, the cleavage and reshuffling of disulfide bond becomes prominent, leading to efficient denaturation of insulin [37].
3.2. Denaturation, swelling and stability of the particles in different media After preparation of the different particle systems, analytic methods were used to study the degree of denaturation and stability (release) of the particles in different media. CD measurements allowed a comparison of the degree of denaturation between untreated non-particulate (soluble) insulin and insulin particles prepared from denatured and non-denatured insulin. All measurements were conducted at pH 7.4 (phosphate buffer) and at the same insulin concentration, which was adjusted according to the UV-adsoption bands at 278 nm (Fig. 4A). As expected, the CD spectra showed almost complete denaturation for particles that were prepared with insulin incubated in high pH solutions. Denaturation of insulin can be readily explained by breakage of thiol bonds that stabilize the helical structures of the protein. Breakage and reshuffling of the thiol bonds is the principal mechanism for
3. Results and discussion 3.1. Sample preparation In this work, insulin was chosen as a model protein that can be readily synthesized into particulate form via pH titration in the presence of CaCO3 microtemplates (Fig. 1) [30]. The general preparation procedure is as follows: Starting from pH 9.5, solvated insulin is mixed with the CaCO3 templates 5 lm in diameter. Then, acetic acid buffer with a pH of 5.2 is slowly added to the mixture. As the pH of the insulin/CaCO3 particle mixture decreases, the solubility of insulin is reduced and it precipitates. Nucleation of precipitation is enhanced within the porous template, by which the
Fig. 2. Zeta potential indicates the pI of the particles at pH 5.2 and agrees with the pI of native insulin. Denatured compact particles (black) and non-denatured compact (gray) were initially prepared at pH 5.2, and the pH value has been adjusted further in the range 4–6.
1426
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
Fig. 3. (A) CD spectra indicate denaturation of insulin. Denatured particles were prepared by storing insulin at pH 10.5 before particle preparation. (B) Different solubility of non-denatured and denatured insulin particles in solution of physiological pH (7.4). The images were taken after 1, 5 and 10 min. Note that the non-denatured particles completely dissolved after 20 min (data not shown). Collapse at isoelectric point condition was observed for both particle types.
Fig. 4. UV–VIS measurements to determine the amount (yield) of insulin transformed into particulate and determination of the amount of released insulin in PBS solution after 60 min incubation. Solid lines in spectra (A) and (B) represent the amount of insulin applied for particle synthesis (bold solid lines) and the amount of insulin contained in the particles after synthesis (thin solid lines). Dashed lines in spectra (A) and (B) represent the amount of insulin before release in PBS (bold dashed lines) and the amount of released insulin in the supernatant of the PBS solution (thin dashed lines): (A) non-denatured insulin particles (gray); (B) denatured insulin particles (black); (C) summary of the UV data.
insulin denaturation at elevated pH [37]. The CD spectra also suggest a substantial decrease in reduction in a-helix [38] when the pure insulin sample is compared with the particle sample prepared with non-denatured insulin. In the experiment, a decrease in ellipticity between the pure insulin sample and the templated insulin of 35% (comparing the peaks at 210 and 200 nm) was always observed. This was, to some extent, unexpected, as special care was taken during particle synthesis to titrate insulin immediately from high pH conditions (denaturation) to the acidic regime (stable). An explanation for the pronounced degree of denaturation during particle synthesis might be protein adsorption to the interfaces of the charged calcium carbonate, which may result in irreversible change in the protein secondary structure [39]. It should be noted that denatured particles do not fully dissolve in PBS. Therefore, the non-dissolving protein aggregates were removed by centrifugation to study dissolved protein by CD exclusively. Partial denaturation of the particulate protein might be a reason for the bimodal release profile in PBS, as reported earlier [31]. An important parameter for pulmonary delivery of insulin is the size of the particles, which also controls the release rate of insulin [31]. As explained above, the present authors were able to control the size and degree of swelling of the particles, adjusting the final pH of the solution for both denatured and non-denatured insulin
particles. As shown in Fig. 3B, directly after titration to pH 5.7, the size of the particles is on the order of 5 lm, similar to the size of the CaCO3 templates. At this pH, the particles were stable over several weeks without changes in morphology. When the pH was decreased to 5.2 (isoelectric point), the particle diameter decreased by a factor of 2. This showed that the appropriate diameter for pulmonary delivery, typically 1–5 lm [40,41], can be conveniently adjusted by choosing the titration endpoint without the need to use templates of different sizes. Microscopic observation of the particle dissolution at pH 7.4 showed that non-denatured particles (1) and (3) dissolved readily within 20 min, as expected from previous release studies [31]. This is in accordance with the envisioned concept for release of insulin from particles in a slightly acidic storage solution (pH 5–6) to physiological pH of the lung (pH 7.4). However, denatured particle systems (2) and (4) did not completely dissolve at a pH of 7.4, but instead they swelled to a diameter of 8–10 lm. This suggests that the reshuffling of disulfide bonds at elevated pH has led to crosslinked insulin particles that can then be swollen and are stable to some extent at physiological pH. To analyze the release characteristics further, the insulin release from collapsed denatured and non-denatured particles stored at pH 5.2 was compared. The relative amount of insulin in the
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
supernatant of a centrifuged particle dispersion in storage buffer (pH 5.2) and PBS (pH 7.4) was measured using UV-VIS spectroscopy. This allows the amount of released insulin from non-denatured and denatured particles to be determined (dashed lines Fig. 4A and B). Owing to the good solubility of non-denatured insulin particles in PBS, the amount of released insulin reached 90%, whereas the denatured particles released only 25% of their insulin because of reduced solubility (Fig. 4C), in line with the results from optical microscopy (Fig. 3). As a result of the insolubility, the denatured particle component must be considered pharmaceutically inactive, as it does not bind to the respective kinase receptors that modulate glucose levels. In addition, the yield of the particle synthesis was determined by comparing the amount of insulin applied for synthesis with the amount of insulin contained in the particles after synthesis (solid lines Fig. 4A and B). Overall, the data show that 40% of the insulin was transformed to particulate form after the pH titration step (Fig. 4C). This means that 60% of the insulin was not precipitated during pH titration and stayed in the soluble phase. When a larger concentration of insulin and template particles is used before titration, the amount of insulin remaining in the soluble phase can be decreased, and the yield can be increased (data not shown). 3.3. Mechanical properties of insulin particles Micromechanical study using AFM is a well-established tool for studying the mechanical and adhesion properties of microparticles for biomedical applications [21,42–45]. So far, it has not been applied in great detail for the mechanical analysis of protein microparticles prepared by CaCO3 templating. However, the mechanical properties of the particles are crucial for the envisioned applications, as they play an important role in adhesion of the particles to surfaces, colloidal stability, circulation time in the vascular system, and ability to overcome mechanical barriers [32]. Therefore, the mechanical properties were analyzed first, comparing the deformability of swollen (porous) and compact insulin particles via colloidal probe AFM. As explained, the compact particle system was titrated from a high pH regime to isoelectric point conditions, where the particle collapsed (Fig. 1). For the swollen particle system, the titration was conducted only to pH 5.7, with the aim of stabilizing the porous template structure via electrostatic repulsion between negatively charged protein molecules. For colloidal probe AFM measurements, the insulin particles were first immobilized on a ‘‘sticky’’ PEI-coated glass slide in an acetic acid buffer (pH 4.0–5.8). Then the particles were deformed by applying 80 nN maximum force via colloidal probe AFM followed by retraction of the probe. After a dwell time of 60 s, another compression cycle was conducted. Typical force–deformation data
1427
Fig. 6. The elastic moduli of denatured and non-denatured collapsed insulin particles (stored at pH 5.2). At isoelectric point (pH 5.2) both particle systems have stiffnesses of several 100 kPa. When the pH is increased to 7.4, non-denatured particles start to dissolve (strong decrease in elastic modulus) and denatured particles swell (minor decrease in elastic modulus). Measurements at pH 7.4 were taken 2 min after increasing the pH from 5.2.
are depicted in Fig. 5, showing force curves from the first and tenth deformation cycles. Indeed, the force curves for the swollen/porous particles (titrated only to pH 5.7) show a clear shift in the range 75–150 nm in the contact point between probe and sample when comparing the first and tenth deformation cycle (Fig. 5A). This shift is the plastic deformation of the particle after repeated compression with 80 nN force. However, compact particles (titrated to pH 5.2, pI conditions) do not deform under this force (Fig. 5B), therefore no shift of the contact point was observed, indicating the absence of plastic deformation. As a result, not only the size but also the deformability of the particles can be controlled by slightly varying the endpoint pH during particle synthesis (Fig. 5C). It should be noted, however, that the softer, supposedly porous particles prepared at pH 5.7 deformed irreversibly. In terms of material stiffness, the porous particles (titrated to pH 5.7) show low elastic moduli (fitted by Hertzian model) in the first compression cycle on the order of 15 ± 9 kPa, which increases to several hundred kilopascal, similar to the elastic modulus of the collapsed particles titrated to pH 5.2 (Fig. 6). As show in Fig. 6, denaturation does not affect the mechanical properties of the protein particles in isoelectric point conditions. However, denaturation significantly influences particle stability at physiological pH. Here, the elastic modulus of non-denatured particles quickly drops below 1 kPa as a result of dissolution of the particles, whereas only a minor change in elastic modulus modulus is observed for denatured particles. In order to investigate the mechanical stability of the particles further in different media, the elastic moduli of completely
Fig. 5. The degree of plastic deformation depends on the preparation history. (A) Freshly prepared particles at pH 5.7 deform plastically, whereas (B) completely collapsed particles at pH 5.2 show no permanent deformation between the compression cycles. (C) Complete data set shows the average amount of plastic deformation as calculated from the shift in contact point between the first and tenth compression cycles. The particles were composed of non-denatured insulin.
1428
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
non-denatured particles quickly dissolved (Fig. 3), therefore exhibiting very low elastic moduli shortly after addition of the buffer. The elastic moduli for denatured particles decreased at pH 7.4 as well, but only by a factor of two compared with pH 5.2. This again showed that the denatured particles stay intact, but undergo swelling at physiological pH, decreasing their elastic modulus. Thus, pH variation may significantly affect both morphological (size) and mechanical properties of the insulin particles. Both can be at the same time modulated by the degree of denaturing. This may enable tuning of the stability, release characteristics and resistance to mechanical barriers of the protein particles delivered to a body. Fig. 7. Mechanical properties depend on pH; charged particles show larger moduli. Elastic moduli of the non-denatured collapsed particles between pH 4.0 and 6.0 (black). Preparation history and degree of denaturation strongly affect the mechanical properties at physiological pH; denatured particles at pH 7.4 (gray), non-denatured particles at pH 7.4 (light gray), both measured 2 min after pH adjustment.
collapsed particles were determined via colloidal probe AFM at different pH values (Fig. 7). In the pH range between 4.0 and 5.9 (acetic acid buffer), only moderate changes by a factor of two of the elastic modulus were observed. Quite surprisingly, the lowest elastic modulus was obtained in isoelectric point conditions for the collapsed particles. This seems in contradiction to the previous measurement at pH 5.2 and 5.7 on the collapsed and swollen/porous particle system (Fig. 5), which showed greater stiffness for collapsed particles at pH 5.2. However, here one is looking at particles that were already collapsed in isoelectric point conditions before performing AFM measurements at different pH values. From the deformation measurement, it was seen that the compact structure of the particle is irreversible also for pH values shifted from the isoelectric point condition. Therefore, the particles do not swell in the pH range 4.0–5.9, as confirmed by optical microscopy observations (Supporting Information S1). However, the zeta potential measurement showed that these systems are charged (Fig. 2). Therefore, the decrease in the elastic modulus towards the isoelectric point condition is explained by the absence of electrostatic repulsion between protein molecules within the particle. However, when changing the buffer system in order to mimic physiological pH (phosphate buffer), swelling of the particles and a strong decrease in the elastic modulus were observed accordingly. When the pH was changed from 5.2 to 7.4 via addition of phosphate buffer, the
3.4. Phagocytosis of insulin particles with different size Phagocytic clearance in the lung can be considered significant for the deactivation of pulmonary delivered protein particles [46]. If taken up by macrophages, the insulin particles will not be delivered to a bloodstream, this is why the uptake is considered a negative process for delivery and must be avoided. This is particularly important for prolonged delivery, when the time is sufficient for the macrophages to take the protein particles up and to activate further immune processes. In general, once a bead or pathogen is recognized by macrophages, the internalization into the cell takes 30 min [47], which is still in the time-frame of insulin release from the particles [31], especially for more stable, partially denatured insulin particles. Therefore, a rough test was performed on the recognition and uptake of the insulin particles by alveolar macrophages. In addition, it is known that recognition and uptake are strongly dependent on the size of the particles [48,49]. For practical reasons, insulin particles of different sizes were used (1, 3 and 8 lm), which are completely denatured (stable at pH 7.4) and show only small changes in mechanical properties (400 kPa) and size when transferred from the acidic storage solution to the cell medium at pH 7.4. This tested size range reflects the currently accessible range for hard templating with CaCO3 particles and is a favorable range for efficient pulmonary delivery [31]. For comparison, the uptake of polystyrene latex beads of similar size was investigated. The results show that all particle systems and the polystyrene control particles were taken up, regardless of the size of the particles (Fig. 8). A major fraction of particles were taken up already after 2 h of incubation. The invariance of uptake with respect to the size of the particles could be expected, as a significant
Fig. 8. Uptake of insulin (first row) and polystyrene micro particles (second row) of different diameters (from left to right column: 1, 3 and 8 lm for insulin and 1, 5 and 10 lm for polystyrene particles). All particles were engulfed by phagocytes after 5 h of incubation. Scale bar: 10 lm.
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
reduction of uptake is observed only for particles that are smaller than 100 nm [49]. Therefore, the results show that pure protein particles suited for pulmonary delivery (>1 lm in diameter) should release the active component within a few minutes (burst release) in order to avoid phagocytic clearance. In the case of the insulin system presented here, this can be achieved by particles composed of non-denatured insulin, as they dissolved completely (within 20 min), i.e. before they could be engulfed by phagocytes. However, for longer-acting insulin formulations, the dissolution process occurs within hours. Thus, when delivered by pulmonary routes, the activity of such systems could indeed be impaired by phagocytic clearance, as shown in the present experiment. In order to slow phagocytic clearance, much smaller (<0.1 lm) or larger (>20 lm) particle systems should be prepared. At the moment, however, CaCO3 templating is not suited for this size range, and pulmonary delivery routes as well as toxicity of such particle systems also have to be carefully analyzed [48]. As an alternative, the particles should be prepared to adapt so-called stealth properties such that recognition by macrophages is reduced. A typical strategy is the modification of the active component with polyethylene glycol (PEG). As shown in previous work hard CaCO3 templating with PEG, it is possible to prepare particles with very low non-specific interaction [50,51]. Therefore, future work will focus on such hybrid protein/ PEG particle systems.
4. Conclusion Pure insulin microparticles were prepared by hard templating on porous CaCO3 using a single trigger: pH (titration of insulin– CaCO3 mixture with pH from 9.5 to 5.2). The size and porosity of the pure insulin particles can be controlled by slight variation in the final pH (from 4 to 6) value around the insulin isoelectric point (5.2) as a result of extra charge doping. Compact and mesoporous protein particles possess different mechanical properties—elastic and plastic deformation, respectively—and deviate up to two orders of magnitude in elastic modulus. pH-induced (pH 10.5) denaturation of insulin was used to prepare protein particles with significantly different stability at physiological conditions. Thus, partial denaturing of the protein may be used to adjust the release profile in physiological conditions. However, it should be noted that the denatured part of the protein via randomization of disulfide bridges leads to insoluble and pharmaceutically inactive insulin. Therefore, it might be advantageous to modulate the solubility and release profile of insulin by other means, e.g. by co-precipitation with crosslinked polymers [14,51,52]. The in vitro test on alveolar macrophages showed that phagocytic clearance remains an issue for longer-acting particle formulations in the size range 1–10 lm. But such longer acting formulations are especially interesting from the medical therapy point of view. Therefore, future work will focus on the preparation of protein/PEG formulations prepared by hard CaCO3 templating. This approach is the most promising route, as previous work on the morphology [50,53] and application [51,52] of templated PEG particles has already shown. Considering the large amount of preceding work on CaCO3 templated protein particles, the next logical step is in vivo testing, e.g. with the well-established insulin particles presented here.
Acknowledgements DVV thanks the Alexander von Humboldt Foundation for support (AvH Fellowship and Sofja Kovalevskaja Program). The authors thank Dr. Jürgen Hartmann for the SEM imaging.
1429
Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Fig. 8 are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/j.actbio.2013. 11.011. Appendix B. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2013. 11.011. References [1] Brock A. Protein drugs: global markets and manufacturing technologies. BCC Research 2008; Report no.: BIO021C. [2] Leader B, Baca QJ, Golan DE. Protein therapeutics: a summary and pharmacological classification. Nat Rev Drug Discov 2008;7:21–39. [3] Walsh G. Biopharmaceuticals: biochemistry and biotechnology. 2nd ed. Chichester: Wiley-Blackwell; 2003. [4] Parveen S, Misra R, Sahoo SK. Nanoparticles: a boon to drug delivery, therapeutics, diagnostics and imaging. Nanomed Nanotechnol Biol Med 2012;8:147–66. [5] Putnam D. Polymers for gene delivery across length scales. Nat Mater 2006;5:439–51. [6] Becker AL, Johnston APR, Caruso F. Layer-by-layer-assembled capsules and films for therapeutic delivery. Small 2010;6:1836–52. [7] Yan Y, Such GK, Johnston APR, Best JP, Caruso F. Engineering particles for therapeutic delivery: prospects and challenges. ACS Nano 2012;6:3663–9. [8] Szlachcic A, Zakrzewska M, Otlewski J. Longer action means better drug: tuning up protein therapeutics. Biotechnol Adv 2011;29:436–41. [9] Yang SX, Yuan WE, Jin T. Formulating protein therapeutics into particulate forms. Expert Opin Drug Deliv 2009;6:1123–33. [10] Tan ML, Choong PFM, Dass CR. Recent developments in liposomes, microparticles and nanoparticles for protein and peptide drug delivery. Peptides 2010;31:184–93. [11] Wang W. Lyophilization and development of solid protein pharmaceuticals. Int J Pharm 2000;203:1–60. [12] Maa YF, Nguyen PA, Sweeney T, Shire SJ, Hsu CC. Protein inhalation powders: spray drying vs spray freeze drying. Pharm Res 1999;16:249–54. [13] Basu SK, Govardhan CP, Jung CW, Margolin AL. Protein crystals for the delivery of biopharmaceuticals. Expert Opin Biol Ther 2004;4:301–17. [14] Balabushevich NG, Pechenkin MA, Shibanova ED, Volodkin DV, Mikhalchik EV. Multifunctional polyelectrolyte microparticles for oral insulin delivery. Macromol Biosci 2013;13:1379–88. [15] Volodkin DV, Balabushevitch NG, Sukhorukov GB, Larionova NI. Model system for controlled protein release: pH-sensitive polyelectrolyte microparticles. STP Pharm Sci 2003;13:163–70. [16] Volodkin DV, Balabushevitch NG, Sukhorukov GB, Larionova NI. Inclusion of proteins into polyelectrolyte microparticles by alternative adsorption of polyelectrolytes on protein aggregates. Biochem-Moscow 2003;68:236–41. [17] Li Q, Retsch M, Wang J, Knoll W, Jonas U. Porous networks through colloidal templates. In: Broekmann PDKHSCA, editor. Templates in chemistry III. Berlin: Springer; 2009. p. 135–80. [18] Thomas A, Goettmann F, Antonietti M. Hard templates for soft materials: creating nanostructured organic materials. Chem Mater 2008;20:738–55. [19] Schmidt S, Volodkin D. Microparticulate biomolecules by mild CaCO3 templating. J Mater Chem B 2013;1:1210–8. [20] Wang YJ, Caruso F. Nanoporous protein particles through templating mesoporous silica spheres. Adv Mater 2006;18. 795-+. [21] Schmidt S, Behra M, Uhlig K, Madaboosi N, Hartmann L, Duschl C, et al. Mesoporous protein particles through colloidal CaCO3 templates. Adv Funct Mater 2013;23:116–23. [22] Petrov AI, Volodkin DV, Sukhorukov GB. Protein–calcium carbonate coprecipitation: a tool for protein encapsulation. Biotechnol Prog 2005;21: 918–25. [23] Volodkin DV, Larionova NI, Sukhorukov GB. Protein encapsulation via porous CaCO3 microparticles templating. Biomacromolecules 2004;5:1962–72. [24] Sukhorukov GB, Volodkin DV, Gunther AM, Petrov AI, Shenoy DB, Mohwald H. Porous calcium carbonate microparticles as templates for encapsulation of bioactive compounds. J Mater Chem 2004;14:2073–81. [25] Pechenkin MA, Mohwald H, Volodkin DV. PH- and salt-mediated response of layer-by-layer assembled PSS/PAH microcapsules: fusion and polymer exchange. Soft Matter 2012;8:8659–65. [26] Volodkin D, Skirtach A, Möhwald H. Bioapplications of light-sensitive polymer films and capsules assembled using the layer-by-layer technique. Polym Int 2012;61:673–9. [27] Volodkin D, Skirtach A, Mohwald H. LbL films as reservoirs for bioactive molecules. In: Borner HG, Lutz JF, editors. Bioactive surfaces. Berlin: Springer; 2011. p. 135–61.
1430
S. Schmidt et al. / Acta Biomaterialia 10 (2014) 1423–1430
[28] Fujii A, Maruyama T, Ohmukai Y, Kamio E, Sotani T, Matsuyama H. Crosslinked DNA capsules templated on porous calcium carbonate microparticles. Colloids Surf A Physicochem Eng Asp 2010;356:126–33. [29] Han YY, Tong WJ, Zhang YY, Gao CY. Fabrication of chitosan single-component microcapsules with a micrometer-thick and layered wall structure by stepwise core-mediated precipitation. Macromol Rapid Commun 2012;33:326–31. [30] Volodkin DV, von Klitzing R, Möhwald H. Pure protein microspheres by calcium carbonate templating. Angew Chem Int Ed 2010;49:9258–61. [31] Volodkin DV, Schmidt S, Fernandes P, Larionova NI, Sukhorukov GB, Duschl C, et al. One-step formulation of protein microparticles with tailored properties: hard templating at soft conditions. Adv Funct Mater 2012;22:1914–22. [32] Merkel TJ, Jones SW, Herlihy KP, Kersey FR, Shields AR, Napier M, et al. Using mechanobiological mimicry of red blood cells to extend circulation times of hydrogel microparticles. Proc Natl Acad Sci USA 2011;108:586–91. [33] Goodridge HS, Reyes CN, Becker CA, Katsumoto TR, Ma J, Wolf AJ, et al. Activation of the innate immune receptor Dectin-1 upon formation of a ‘phagocytic synapse’. Nature 2011;472:U471–541. [34] Hutter JL, Bechhoefer J. Calibration of atomic-force microscope tips. Rev Sci Instrum 1993;64:1868–73. [35] Sader JE. Frequency response of cantilever beams immersed in viscous fluids with applications to the atomic force microscope. J Appl Phys 1998;84:64–76. [36] Schmidt S, Fernandes PAL, De Geest BG, Delcea M, Skirtach AG, Mohwald H, et al. Release properties of pressurized microgel templated capsules. Adv Funct Mater 2011;21:1411–8. [37] Brange J, Langkjaer L. Chemical stability of insulin. 3. Influence of excipients, formulation, and pH. Acta Pharm Nord 1992;4:149–58. [38] Pocker Y, Biswas SB. Conformational dynamics of insulin in solution. Circular dichroic studies. Biochemistry 1980;19:5043–9. [39] Mollmann SH, Jorgensen L, Bukrinsky JT, Elofsson U, Norde W, Frokjaer S. Interfacial adsorption of insulin – conformational changes and reversibility of adsorption. Eur J Pharm Sci 2006;27:194–204. [40] Khafagy E-S, Morishita M, Onuki Y, Takayama K. Current challenges in noninvasive insulin delivery systems: a comparative review. Adv Drug Deliv Rev 2007;59:1521–46. [41] Patton JS, Byron PR. Inhaling medicines: delivering drugs to the body through the lungs. Nat Rev Drug Discov 2007;6:67–74.
[42] Fernandes PAL, Schmidt S, Zeiser M, Fery A, Hellweg T. Swelling and mechanical properties of polymer gels with cross-linking gradient. Soft Matter 2010;6:3455–8. [43] Delcea M, Schmidt S, Palankar R, Fernandes PAL, Fery A, Möhwald H, et al. Mechanobiology: correlation between mechanical stability of microcapsules studied by AFM and impact of cell-induced stresses. Small 2010;6:2858–62. [44] Palankar R, Pinchasik B-E, Schmidt S, De Geest BG, Fery A, Mohwald H, et al. Mechanical strength and intracellular uptake of CaCO3-templated LbL capsules composed of biodegradable polyelectrolytes: the influence of the number of layers. J Mater Chem B 2013;1:1175–81. [45] Fery A, Weinkamer R. Mechanical properties of micro- and nanocapsules: single-capsule measurements. Polymer 2007;48:7221–35. [46] Edwards DA, Ben-Jebria A, Langer R. Recent advances in pulmonary drug delivery using large, porous inhaled particles. J Appl Physiol 1998;85:379–85. [47] Blanchette CD, Woo YH, Thomas C, Shen N, Sulchek TA, Hiddessen AL. Decoupling internalization, acidification and phagosomal–endosomal/ lysosomal fusion during phagocytosis of InlA coated beads in epithelial cells. PLoS One 2009:4. [48] Renwick LC, Donaldson K, Clouter A. Impairment of alveolar macrophage phagocytosis by ultrafine particles. Toxicol Appl Pharmacol 2001;172: 119–27. [49] Zauner W, Farrow NA, Haines AMR. In vitro uptake of polystyrene microspheres: effect of particle size, cell line and cell density. J Control Release 2001;71:39–51. [50] Behra M, Schmidt S, Hartmann J, Volodkin DV, Hartmann L. Synthesis of porous PEG microgels using CaCO3 microspheres as hard templates. Macromol Rapid Commun 2012;33:1049–54. [51] Behra M, Azzouz N, Schmidt S, Volodkin DV, Mosca S, Chanana M, et al. Magnetic Porous Sugar-Functionalized PEG Microgels for Efficient Isolation and Removal of Bacteria from Solution. Biomacromolecules 2013;14:1927–35. [52] Pussak D, Ponader D, Mosca S, Ruiz SV, Hartmann L, Schmidt S. Mechanical carbohydrate sensors based on soft hydrogel particles. Angew Chem Int Ed 2013;52:6084–7. [53] Pussak D, Behra M, Schmidt S, Hartmann L. Synthesis and functionalization of poly(ethylene glycol) microparticles as soft colloidal probes for adhesion energy measurements. Soft Matter 2012;8:1664–72.