Multilayered Magnetic Nanobeads for the Delivery of Peptides Molecules Triggered by Intracellular Proteases.

In this work, the versatility of layer-by-layer technology was combined with the magnetic response of iron oxide nanobeads to prepare magnetic mesostructures with a degradable multilayer shell into which a dye quenched ovalbumin conjugate (DQ-OVA) was loaded. The system was specifically designed to prove the protease sensitivity of the hybrid mesoscale system and the easy detection of the ovalbumin released. The uptake of the nanostructures in the breast cancer cells was followed by the effective release of DQ-OVA upon activation via the intracellular proteases degradation of the polymer shells. Monitoring the fluorescence rising due to DQ-OVA digestion and the cellular dye distribution, together with the electron microscopy studying, enabled us to track the shell degradation and the endosomal uptake pathway that resulted in the release of the digested fragments of DQ ovalbumin in the cytosol.

Introduction

Magnetic nanoparticles are an effective example of how nanoscale materials can improve medical field research by providing new diagnostic and therapeutic solutions.[1−4] Colloidal suspensions of iron oxide nanoparticles (IONPs) are indeed FDA approved contrast agents in magnetic resonance imaging, detectable as T2 and T2* transversal relaxivity detection modes.[5] Still, IONPs can serve as heat mediators in so-called magnetic hyperthermia (MH). In recent years, MH apparatus and associated IONPs (Nano-Cancer therapy) have received approval as medical devices after a clinical trial that was conducted in Germany by Magforce company.[6,7] This trial showed an overall survival of 23 months in glioblastoma patients when injecting intratumorally IONP and associating MH to radiotherapy.[8]

When using IONPs for delivering drugs to solid tumors and metastases, it would be desirable to inject intravenously NPs and exploit their response to magnetic field gradients to accumulate them at the tumor. However, single IONPs have small magnetic moments that hardly respond to magnetic field gradients applicable to a human patient. A way to improve the magnetic response of IONPs is to cluster them into mesoscale entities (100–400 nm), here named “magnetic nanobeads”, in which several IONPs are grouped together.[9] In these controlled assemblies, the cumulative response of the single IONPs within the same assembly gives a much higher magnetic moment than individual IONPs. This results in a faster magneto-phoresis mobility with respect to single nanoparticles when exposed to the same field gradient.[10,11] Moreover, the beads keep the same superparamagnetic behavior as the single nanoparticles, with no tendency to aggregate in the absence of a magnetic field.

Several groups have reported different procedures to control the condensation of several superparamagnetic nanoparticles within a single template material of different compositions (polymer, silica, or fat droplets matrix).[12−26] Also, some of us have developed a method for the synthesis of nanobeads with control over core and surface properties, and these nanobeads were used as platforms on which to add additional polymeric shell and provided more advanced features.[23,24] For example, by decorating our magnetic nanobeads with a thermoresponsive polymer we were able to control drug adsorption and release it by increasing the temperature solution from 37 to 47 °C.[27] On the other hand, by tuning the electrostatic interactions between the polymeric shell of the nanobeads and oligonucleotides (siRNA), we could deliver a still functionally active anti-GFP siRNA to living cells to silence protein expression.[23]

The rational of this study is to demonstrate the advantages that derive from the combination of a mesoscale magnetic structure with a layer-by-layer (LbL) assembly of a biodegradable multilayer shell. A LbL assembly allows for the formation of multilayer shells around spherical charged templates through the alternating adsorption of polyanions and polycations.[28,29] The composition and properties of the resulting structures can be tailored by employing polymers with different functions as well as stimuli-responsive nanomaterials (i.e., silver, gold, and magnetic nanoparticles) as layer constituents.[30−35] The magnetic nanobeads used in this work offer the unique advantage of a superparamagnetic system with enhanced magnetic response and these features are merged with the advantages of LbL technology. In recent years, a variety of magnetic-responsive carriers have been developed but they cover different size ranges than the one used in the present work. For instance, it has been reported the adsorption of magnetic NPs in between the layer shells of hollow microcapsules thus providing LbL micrometer capsules.[36,37] In other work, the LbL deposition of polyelectrolyte multilayer shells directly around individual magnetic NPs[38−41] of few nanometers have been developed.

Here, we used our magnetic nanobeads of size defined in between the ones so far developed (MNBs, 126 ± 22 nm size as measured by TEM) as a core template on which to set an LbL assembly of a biodegradable multilayer shell hosting a protease-activated fluorophore. As layer constituents, we used oppositely charged synthetic polyamino acids, namely poly-l-glutamic acid (PGA) and poly-l-arginine (PARG).[35,42] As a protease-activated substrate, we used dye quenched ovalbumin conjugate (DQ-OVA), a protein consisting of naturally mannosylated OVA extensively labeled with fluorochrome BODIPY (excitation/emission maxima ∼503/512 nm).[43] This protein represents an ideal cargo for studying the intracellular processing and release of molecules embedded within carriers because the protein digestion of DQ-OVA by proteases leads to cleaved peptide fragments that begin to fluoresce.[43−46]

The LbL method enables quick drug encapsulation within a thin and biodegradable multilayer shell and its further release upon intracellular enzymatic activity. Thus, the hybrid system presented here is a model of a protease-triggered drug vector displaying superparamagnetic behavior and an average size within the optimal range for exploitation in living systems.[47,48]

Results and Discussion

Preparation and Characterization of Multilayered Magnetic Nanobeads Loaded with DQ-OVA

Negatively charged MNBs were used as templates for an LbL assembly of multilayer biodegradable shells made with polycation poly-l-arginine hydrochloride (PARG) and polyanion poly-l-glutamic acid sodium salt (PGA) polymer (Figure a). In particular, the MNB samples used in this work, prepared according to a procedure recently described,[24] have a size of approximately 100 nm as measured by TEM (Figure a). First, a layer of PARG was deposited by adding 1 mL of PARG (2 mg mL–1, 50 mM NaCl, pH 6.5) to the MNBs. Following 1 h of incubation, the MNBs@PARG were washed two times with 3 mL of ultrapure water to remove the unbound PARG. The magnetic behavior of the MNBs allowed for collecting them by application of an external magnetic field (Figure a,b), without the need of centrifugation. This aspect is highly valuable, since repetitive centrifuge steps, required in LbL protocols to remove excess polymers following each incubation and washing step, are known to reduce colloidal stability, increasing particle aggregation and leading to the poor applicability of multilayered nanoparticles, especially in cellular studies.[49] Next, a layer of PGA was deposited by adding 1 mL of PGA (2 mg mL–1, 50 mM NaCl, pH 6.5) to the MNBs@PARG pellet. After two washing steps, a layer of DQ-OVA (20 μg mL–1, in PBS, pH 6.5) was added. Three additional layers were added to obtain MNBs with the following multilayer shell: (PARG/PGA)(DQ-OVA/PARG)(PGA/PARG).

Figure 1

(a) Schematic depiction of the LbL assembly of a biodegradable multilayer shell around magnetic nanobeads. Negatively charged nanobeads (MNBs) serve as magnetic templates onto which biodegradable polycations (PARG) and polyanions (PGA) are alternatively adsorbed via electrostatic interactions. Excess polyelectrolytes are removed by the magnetic collection of the MNBs and washing them in ultrapure water. As a third step, a layer of the cationic DQ-OVA is deposited, followed by the LbL of three additional layers to obtain a stable multilayer shell. Shell architecture: (PARG/PGA)(DQ-OVA/PARG)(PGA/PARG). (b) Photographs of a colloidal suspension of MNBs before (top) and after (bottom) accumulation to the magnet. (c) Upon the proteolytic digestion of the multilayer shells and DQ-OVA, quenching is relieved and green fluorescence appears. Beads and polymers are not drawn to scale. Only a few layers of polyelectrolytes are shown for clarity.

Figure 4

TEM images of MNBs (a, d) before and (b, e) after LbL and DQ-OVA adsorption. (c, f) MMNBs after the protease treatment. Arrowheads highlight the damage to the polymer shell following its exposure to proteases. Scale bars (a–c) 100 and (d–f) 50 nm.

Accordingly, the outermost surface of the multilayered MNBs (MMNBs) is positive in order to enhance the cellular uptake, as described elsewhere.[50−52] The alternating adsorption of the oppositely charged polyelectrolytes and DQ-OVA was monitored by means of z-potential analyses. As shown in Figure a, PARG and PGA regularly interchanged during the LbL assembly process. The z-potential of the MNBs shifted from −44 ± 1 to +40 ± 1 mV following the deposition of the first layer of PARG. As expected, the addition of anionic PGA reversed the charge of the sample to a negative zeta potential value (−59 ± 3 mV). Notably, the adsorption of DQ-OVA resulted in a slight reduction in the negative charge while full charge reversal occurred with the subsequent addition of PARG. The final configuration of the MNBs had five oppositely charged layers and a net positive charge.

Figure 2

(a) z-potential measurements reporting the charge alternation of MNBs after each stage of LbL coating. (b) DLS measurements of MNBs before (black curve) and after (gray curve) LbL processing. The dotted black curve refers to MNBs after the protease treatment.

The sequential adsorption of polyelectrolyte layers, as expected, resulted in the increase of the average hydrodynamic diameter of the nanobeads, as shown in the DLS profile of the MNBs before and after LbL processing (black and gray curve of Figure b, respectively). To be precise, the DLS size shifted from 240 ± 4 nm in the case of bare MNBs to 331 ± 12 nm after polyelectrolyte deposition.

Degradation of Multilayered Magnetic Nanobeads Loaded with DQ-OVA

Next, we studied the proteolytic degradation of the multilayered magnetic nanobeads (MMNBs) upon exposure to proteases by monitoring the fluorescence of DQ-OVA.

First, we estimated the DQ-OVA loading percentage (Figure b). To this end, we collected the feeding solution of the DQ-OVA after incubation with the MNBs (blue line) as well as the solution used in the following washing step (light blue line). The dark blue line refers to the starting solution of DQ-OVA used as a reference to measure the loading percentage. The three solutions were incubated with protease O.N. prior to being analyzed at the fluorimeter. Almost 90% of DQ-OVA administered was loaded into the multilayer shell (see the equation reported in the experimental section). Notably, in the term “DQ-OVA not loaded” we also included the fragments detected in the surnatant from the first washing step after DQ-OVA incubation (blue line), since this signal can be ascribed to DQ-OVA being weakly adsorbed on the MNBs and rapidly washed out. Indeed, after the second washing step (data not shown) no fluorescent signal could be detected, thus indicating that most of the DQ-OVA was loaded onto the MNBs.

Figure 3

(a) Schematic depiction of the loading of DQ-OVA into MMNBs followed by the protease treatment of the feeding solution (inset in panel b) and the MMNBs (inset in panel c), respectively, to generate fluorescent DQ-OVA fragments detectable by photoluminescence (PL) measurements. (b) Loading capacity of the MMNBs: PL spectra of DQ-OVA (dark blue curve), supernatant recovered after the incubation of the MMNBs with DQ-OVA solution (blue curve), and supernatant recovered after the first washing of the MMNBs post DQ-OVA incubation (light blue curve). (c) Release of digested DQ-OVA fragments: PL spectra of the free DQ-OVA solution (black curve) and MMNBs loaded with DQ-OVA (dark gray) after protease incubation. The two gray curves refer to the same sample that underwent two successive protease treatments in order to release and cleave all of the fluorescent molecules that were adsorbed into the layers or entrapped within the shells of the MMNBs. Excitation wavelength was set at 488 nm.

It is noteworthy that after the protease treatment the surface charge of the MNBs was −41.2 ± 0.6 mV, a value very close to that of the starting MNBs (see Figure ).

Next, DQ-OVA fragments were released into glass vials and evaluated (Figure c). MMNBs were incubated with protease (1 mL, 5 mg/mL in Tris-HCl buffer, pH 7.4) and kept at 37 °C (O.N.) under vigorous stirring. Then, the MNBs were collected by means of a magnet and the bulk solution was analyzed with a fluorescence spectrometer. The fluorescence spectra of Figure c refer to the cleaved green-fluorescent fragments released in solution: the graph reports the curves of the DQ-OVA released by the MNBs after two successive protease treatments. It is evident that after the first treatment (gray curve) the release was significantly higher than the one recorded following the second treatment (light gray curve), thus indicating that both the layers and the fluorophore were efficiently digested by the protease following the first treatment. We compared these curves with that of the initial DQ-OVA (dark gray curve) administered to the MNBs. To this end, a solution of DQ-OVA (20 μg mL–1) was incubated with the protease solution under the same conditions as the MNBs. Analyzing the three curves, we could estimate that almost 55% of the administered DQ-OVA was released from the MNBs after the protease treatment. The total amount of free fluorescent fragments detected after the first treatment was considerably higher than after the second ones, thus indicating that one enzymatic treatment was sufficient for cleaving the multilayer shells and most of the loaded ovalbumin.

Further evidence of protease degradation came from the TEM characterization of the MNBs (Figure ). Panels a and d of Figure show the structure of MNBs before the LbL deposition: the nanobeads, displayed an average size of 121 ±19 nm, a typical core–shell structure with an electron dense core, made of iron oxide NPs, and an electron lean polymer shell, made of poly(maleic anhydride-alt-1-octadecene). After the LbL deposition, the nanobeads preserved their original spherical shape (Figure , panels d and e) and size that is within the range considered appropriate for drug-delivery applications.[53−55] Indeed, it has been reported that nanoparticles with a diameter below 200 nm may accumulate at the tumor site and penetrate deeper into the inner regions of tumor lesions.[52,56,57] When compared to the TEM images of pristine MNBs (Figure a and b), the multilayer shells could not be appreciated, presumably due to the very low contrast of the polymers employed and to the number of added layers (only five layers).

Panels c and f of Figure report the morphology of the MNBs after protease activity. The edges of the polymer shell of the nanobeads appear rough and irregular as compared to both the pristine MNBs (Figures a–d, arrowheads) and MMNBs (Figures b-e), likely as a result of the protease attack. In addition, the DLS size of MNBs was greatly enlarged (see the dotted curve of Figure b) after protease incubation, likely due to the partial degradation of the polymer shell, resulting in the reduced stability of the nanobeads that tend to aggregate in large clusters.

Intracellular Degradation of Multilayered Magnetic Nanobeads Loaded with DQ-OVA and Cytosolic Release

The cellular uptake of DQ-OVA loaded MMNBs and the shell digestion were then assessed in vitro. To this end, MDA-MB-231 cells were incubated with the MMNBs and, at defined time points, the cells were fixed and imaged. The confocal (CLSM) images reported in Figure show a significant increase in the green fluorescence signal over time (the individual green and red channels at each time point are displayed in Figures S1–S6). At 24 h, small and green fluorescent spots became visible inside the cells (Figure , panels a–a.1). In agreement with previously published data on the proteolytic digestion of DQ-OVA loaded into different carrier systems,[44,46,58,59] these spots most likely correspond to an accumulation of the digestion products of DQ-OVA into the endosomal compartments, which are further processed intracellularly. Notably, a slight red fluorescence signal could be observed in some cell fragments that emitted a bright green fluorescence also (Figure S5, white arrows in panel f) which likely corresponds to the partially digested fragments of the DQ-OVA complex having largely separated DQ dyes (green fluorescence) or dyes in close proximity (ref fluorescence).[44]

Figure 5

Release of fluorescent peptide fragments. (Top) Schematic depiction of the internalization and degradation stages of MMNBs: (i) intact DQ-OVA loaded MMNBs being internalized; (ii) engulfment of intracellular vesicles (granular structures present in the cytosol); (iii) proteolytic digestion of the multilayer shells and the cleavage of DQ-OVA into small fragments connected by an enhancement of the green fluorescence; (iv) homogeneous diffusion of the cleaved fragments of DQ-OVA in the cytosol. (Bottom) Representative CLSM images of MDA-MB-231 cells incubated with MMNBs@(PARG/PGA)(DQ-OVA/PARG)(PGA/PARG) for 24 h (a-a.1). After 24h, the medium was replaced and the cells were kept in an MMNB-free medium up to 48 h (b–b.1), 96 h (c–c.1), and 120 h (d–d.1) at 37 °C. The increase in the green fluorescent signal indicates the ongoing degradation of DQ-OVA over the time. The overlay of the fluorescence (green channel) and bright-field channels is shown. Dashed boxes = zoomed area of the cell plate. Scale bars: 20 μm.

At 48 h, a significant increase in the green fluorescence signal was observed (Figure , panels b–b.1). The fluorescent images reported in Figure (panels b–d) clearly show that the green fluorescent DQ-OVA fragments, which originated from the enzymatic cleavage of the DQ-OVA entrapped in the protease-sensitive multilayer shells, were homogeneously distributed within the entire cell body after 48, 96, and 120 h. These results indicate that upon the enzymatic degradation of the multilayer shells, the entrapped DQ-OVA was cleaved and the originating peptides diffused in the cell cytosol. Further evidence of the full digestion and complete release of DQ-OVA was given by the images reported in Figure S6, in which cells with internalized but nonfluorescent MNBs could be observed (see Figure S6, yellow arrows in panel f). The presence of nonfluorescent MNBs indicates the digestion, release and intracellular distribution of the green fluorescent DQ-OVA fragments from the MNBs, which instead remain confined within the intracellular acidic vesicles (i.e., endosomes/lysosomes). However, in some areas of the plate, cells displaying clusters of fluorescent MNBs could be found (Figure S6, white arrow in panel f). This likely occurred because in these cells the DQ-OVA fragments were still being digested, as denoted by the bright green fluorescence signal, but had not yet been released into the cytosol. Indeed, since the cells were administered with the MMNBs continuously for 24 h, the detection of different stages of the intracellular processing of DQ-OVA at 48 h can be ascribed to a random uptake of the MMNBs within the first 24 h and, consequently, to compartmentalization, digestion and release at different times. The cellular internalization of nanoparticles in general, here of MNBs, is a statistical process therefore, over time, MNBs that are being phagocytosed, or have already been internalized and processed by the cells, can be found in the same area under analysis.

Structural TEM characterization of MDA cells incubated with DQ-OVA loaded MMNBs shows their intracellular localization over the time (Figure ). At 2 h, the MNBs are engulfed by the membrane (Figure a and b), while after 24 and 48 h they are confined within vesicular structures, most likely endolysosomes (Figure c–f).

Figure 6

Ultrastructural characterization of cells administered with DQ-OVA loaded MMNBs and kept under incubation for (a) 2 h, (c) 24 h, and (e) 48 h. (b, d, f) Zoomed area of individual MMNBs. Solid arrows: vesicle’s membrane; dashed arrows: MMNBs. MNBs, multilayered magnetic nanobeads; V, intracellular vesicles.

These observations are in good agreement with our previous findings on the uptake and intracellular localization of MNBs through endocytic pathways.[24,60] Accordingly, we assume that once into the endolysosome compartments, the biodegradable shell and the DQ-OVA are cleaved by the enzymatic reservoir, and the resulting fluorescent fragments diffuse out into the cytoplasm. In literature, several mechanisms of the endosomal escape of nanocarriers and of their cargo have been described, from the membrane destabilization to the pore formation to the endosomal rupture.[61−63] These mechanisms are generally governed by the chemical composition and the charge of the nanomaterials, and the effects they produce can be temporary, such as in the case of the membrane destabilization, or permanent, as they are in the endosome rupture.[64] In this study, as observed in the ultrastructural characterization of the cells incubated with the DQ-OVA loaded MMNBs (Figure ), the whole structures were internalized by the cells and were later found into the endolysosome where the proteolytic digestion occurred. Further, the magnetic core remained confined into vesicles up to 120 h. On the other hand, once the PARG and PGA polyelectrolytes layers and the DQ-OVA were digested by the lysosomal proteases, the fluorescent signal (raising from the cleaved OVA peptide fragments) started to light on and spread into the cytosol (see Figure , after 48 h). Reasonably, this process involved an endolysosomal escape process that drives the diffusion of the fluorescent fragments in the cell cytosol. Upon proteolytic digestion, ovalbumin (that is a 45 kDa protein) produces peptide fragments with molecular weights ranging from 10 kDa to less than 1 kDa.[65,66] As already reported in literature,[67,68] it is likely that the presence of oppositely charged species triggers a temporary destabilization of the organelle membranes (the so-called “flip–flop” mechanism), that facilitates the cytosolic release of small molecules, like the digested fragments, but not of the magnetic nanobeads.[69] Therefore, the system here presented allows for the cytosolic delivery of small molecules, that is, peptides in this specific case, by means of a magnetic carrier coated by a stimulus-responsive biodegradable shell that can host biomolecules and trigger their release upon proteolytic degradation. Successful degradation of the polyelectrolyte layers and the consequent release of the hosted molecules has been proven by the fluorescent lighting of DQ-OVA, followed by the cytosolic diffusion of the fluorescent fragments.

To further prove that the release and diffusion of DQ-OVA fragments were exclusively triggered by intracellular proteases, we prepared MNBs coated with nondegradabale polyelectrolytes, the anionic PSS and the cationic PAH, into which the cationic BSA-RITC was loaded as fluorescent probe. Indeed, the fluorescence emission of BSA-RITC does not depend on the presence of proteases (as instead occurs in the case of DQ-OVA), and the use of nondegradable polyelectrolytes should keep the fluorescent protein entrapped within the shell.

Incubation of cells with MNBs@(PAH/PSS)(BSA-RITC/PAH)(PSS/PAH) particles was followed by internalization of the beads and their localization within intracellular vesicles at 24 h (Figure S8a). As expected, after 120 h of incubation, both the beads and the red emitting BSA-RITC were still visible in the vesicles, whereas no fluorescence signal of BSA-RITC could be detected in the cell cytosol (Figure S8b). These data provide additional support to our hypothesis that the release of DQ-OVA was exclusively triggered by the endolysosomal proteases that digest the biodegradable multilayer shell of MNBs.

Finally, we analyzed quantitatively the cellular uptake of the magnetic nanobeads with and without polyelectrolytes shell via elemental analysis to investigate the role of the surface coating in the interaction with the cellular membrane and the endocytosis process.

To this aim, cells were incubated with either MNBs or MMNBs for 24, 48, and 96 h. As shown in Table 1 in the Supporting Information, at each time point, being the administered bead amount at the same iron concentration, cells treated with MMNBs exhibited a higher intracellular amount of Fe compared to cells incubated with MNBs. This in turn corresponds to a greater internalization efficiency (I.E.): while at 24 h, cells incubated with MNBs displayed an I.E. equal to 4.3%, the cellular samples administered with MMNBs showed 30.9% I.E., thus 7 times greater.

In addition, it looks that in the case of cells incubated with MMNBs the uptake occurred rapidly within the first 24 h because the I.E. was almost the same at each time point. On the other hand, in the case of cells incubated with MNBs, the amount of particles internalized increases with time, being the I.E. 4.3, 5.1, and 7.4%, respectively, at 24, 48, and 96 h. Reasonably, the MNBs adhered to the cell membrane and were then internalized at a slower speed than MMNBs.

This behavior would depend on the net positive charge of the MMNBs considering that the last layer is made of PARG (approximately +40 mV), while the starting MNBs have a negative surface (−40 mV) as shown in Figure a.

The charge dependent uptake has been largely demonstrated in literature for various types of nanomaterials.[50−52]

Conclusions

We have described an efficient method of preparing multifunctional stimuli-responsive nanostructures based on iron oxide nanobeads functionalized with a protease-sensitive multilayer polymer shell that acts as a degradable compartment for the temporary loading and hosting of bioactive cargos. The magnetic response of the polymeric-inorganic core allowed for the fast preparation of the MMNBs and provided a magnetic guidance for the hosted molecules placed within the multilayer shell. Special care was given to control the size of the magnetic nanobeads that should fall in the optimal range required for in vivo delivery, and the stability of MMNBs by setting the optimal conditions for layer adsorption and payload encapsulation without affecting the colloidal stability of the final structure in physiological conditions.

The degradation of the LbL shell and the consequent release of the cargo were demonstrated both in cuvette tests and in cellular studies following the activation of the fluorescent DQ-OVA upon protease activity.

It is worth mentioning that ovalbumin (OVA) is as an allergen model. Immunization with ovalbumin-loaded capsules has been reported by De Geest and collaborators who explored the capacity of ovalbumin-loaded polymeric multilayer capsules to elicit immune response in murine models of cancer (B16 melanoma) and infection (influenza).[70] However, in our study, the DQ-OVA was merely used as drug model to easily track the intracellular drug release.

The results here reported clearly show the potential of these multilayer magnetic structures for multi-stimuli-triggered delivery of biomolecules. Future experiments will follow regarding the co-delivery and sequential release of two or multiple drugs to strengthen the therapeutic efficacy in chemoresistant tumors. Also, the implementation of such an LbL approach to magnetic beads made of iron oxide nanocubes[71] would enable the controlled delivery of an internal cellular stimulus and a magnetic hyperthermia by applying an external alternating magnetic field. Alternatively, the use as core template for LbL protocol of magnetic–fluorescent beads[24] consisting of iron oxide NPs as MRI contrast agent and quantum dots as fluorescent probes, would enable to develop a drug delivery system with multimodal imaging features.

Methods

Materials

Poly-l-arginine hydrochloride (PARG, Mw > 70 kDa, no. P3892), poly-l-glutamic acid sodium salt (PGA, Mw = 50–100 kDa, no. P4886), protease (no. P5147) were obtained from Sigma-Aldrich. DQ-ovalbumin (Mw = 45 kDa, no. D12053) was purchased from Invitrogen.

Preparation, Proteolytic Degradation, and Characterization of LbL-Coated MNB

Iron oxide nanoparticles (6 nm diameter) were prepared via thermal decomposition method.[72] Then, they were embedded with poly(maleic anhydride-alt-1-octadecene) to form colloidal and stable MNBs with an average size of 100 nm, as reported by Di Corato et al.[18,24] The negatively charged MNBs (1 mL, 1 mM Fe concentration, in ultrapure water) were incubated with 1 mL of PARG (2 mg/mL, 50 mM NaCl, pH 6.5). Following 1 h of incubation, MNBs were washed 2 times with 3 mL of ultrapure water. MNBs were collected by the application of a constant magnetic field (0.3 T neodynium/boron magnet). Then, a layer of PGA was deposited by adding 1 mL of PGA (2 mg/mL, 50 mM NaCl, pH 6.5). After the washing steps, a layer of DQ-OVA (20 μg/mL, in PBS, pH 6.5) was added. Preliminary experimental sets were performed to fix the amount of DQ-OVA to be added: increasing amounts of ovalbumin (from 200 to 2 μg/mL) were incubated with the MNBs and, after magnetic collection, the supernatants were recovered, digested with protease in order to evaluate the amount of DQ-OVA non loaded.

On the other hand, the amount of loaded DQ-OVA was calculated according to the following equation:Then, three additional layers were added to obtain nanobeads with the following multilayer shell: (PARG/PGLUT)(DQ-OVA/PARG)(PGLUT/PARG).

The proteolytic degradation of the polyelectrolyte shell was performed through the incubation of the LbL-coated MNBs with protease (protease type XIV from Streptomyces griseus) (1 mL, 5 mg mL–1 in Tris-HCl buffer, pH 7.4) at 37 °C O.N. under vigorous stirring. Then, the MNBs were collected by means of a magnet and the supernatant was analyzed via a fluorescence spectrometer (Cary Eclipse). PL spectra of the cleaved fragments were recorded under 488 nm excitation wavelength.

Low-magnification TEM images of LbL-coated MNBs were recorded on a JEOL Jem1011 microscope operating at an accelerating voltage of 100 kV. ζ-potential measurements were performed on a Zetasizer Nano ZS90 (Malvern, USA) equipped with a 4.0 mW He–Ne laser operating at 633 nm and with an avalanche photodiode detector. Measurements were made in water at 25 °C.

Fe concentration was assessed by means of inductively coupled plasma atomic emission spectrometry (ICP-AES, Varian 720-ES) through the preparation of a Fe calibration curve.

Cellular Studies

MDA-MB-231 cells were routinely maintained at a RPMI medium supplemented with l-glutamine (2 mM), penicillin (100 units mL-1), streptomycin (100 mg mL-1), and 10% heat-inactivated fetal bovine serum (FBS). Cells were maintained at 37 °C in a water-saturated atmosphere of 5% CO2 in air.

Cells (2 × 105) were seeded on coverslips placed in 6-well plates with 2 mL of culture medium. After 24 h incubation, the medium was replaced with 2 mL of fresh medium containing the LbL-coated MNBs at a Fe concentration equal to 50 μM. The cells were kept under incubation with the MNBs for either 2 or 24 h. Then, in the case of the “2 h” time point, the cells were washed with PBS, fixed with 4% paraformaldehyde, and mounted in glycerol. On the other hand, the cellular samples incubated continuously with the MNBs for 24 h were divided into four different sets. In the first set, the cells were fixed at 24 h; in the others, the medium was replaced with fresh one and the cells were kept in incubation for an additional 24, 72, and 96 h, respectively, prior to being fixed and imaged. In this way, we could monitor the intracellular localization of DQ-OVA in a time frame of 2 to 120 h.

Confocal Laser Scanning Microscope (CLSM)

Fluorescent microscopy images were collected using a Leica confocal laser scanning microscope (CLSM) (TCS SP5; Leica, Microsystem GmbH, Mannheim, Germany). The fluorescence of DQ-OVA was excited at 488 and 543 nm and the emission was collected between 505–540 (green channel) and 560–670 nm (red channel). Cells were observed with a 63 × 1.40 na oil-immersion objective (sequential mode acquisition, scan speed 400 Hz, 512 × 512 pixels, pinhole aperture set to 1 Airy). To avoid fluorescent artifacts via CLSM observation, the imaging acquisition parameters were adjusted to remove any trace of green autofluorescence in cells or autofluorescence of polymers used for beads synthesis and for LbL coating. The images reported in Figure S7 show MDA-MB-231 cells incubated with MNBs@(PARG/PGA)2(PARG) for 24 h and then kept in incubation in an MNB-free medium up to 120 h. Any green or red fluorescent signal was detected under the aforementioned imaging setting conditions.

Structural Analysis of Cells

MDA-MB231 cells (5 × 105 suspended in 2 mL of medium) were seeded in a culture dish. After 24 h, the medium was replaced with a fresh medium containing LbL-coated MNBs at a Fe concentration equal to 50 μM, and the cells were kept under incubation at 37 °C for either 2 or 24 h. After 24 h, the cells were either processed or placed in fresh medium for an additional 24 h. Thus, three time points were considered: 2, 24, and 48 h.

For TEM analysis, the cells were washed with PBS and fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer at 4 °C for 30 min. The fixed specimens were washed three times with the same buffer and then a solution of 1% osmium tetroxide in cacodylate buffer was added and left in contact with the sample for 1 h. Thereafter, the cells were washed again and dehydrated with 25%, 50%, 75%, and 100% acetone. Two steps of infiltration in a mixture of resin/acetone (1/1 and 2/1 ratios) followed and finally the specimens were embedded in 100% resin at 60 °C for 48 h. Ultrathin sections (70 nm thick) were cut on an Ultramicrotome, stained with lead citrate, and observed under the electron microscope.

Quantification of Cellular Uptake

MDA-MB-231 cells were seeded in each well of a 6 well-plate in 2 mL of culture medium. After 24 h the medium was replaced with 2 mL of fresh medium containing either MNBs or MMNBs (Fe concentration equal to 91 μM). Three different time points were chosen, 24, 48, and 96 h. In detail, cells were incubated continuously with the nanobeads for 24 h. Then, the samples were either processed for elemental analysis, or kept in the incubator with fresh medium for additional 24 or 48 h. The cells were washed three times with PBS, trypsinized and counted. Then, 250 μL of a concentrated H2O2/HNO3 (1/2) solution was added and the cells digested for 3 h in a hot bath at 55 °C (in order to ensure the complete digestion of the cells components), under sonication. After this step, concentrated HCl was added (3/1 volume ratio respect to HNO3) to reach a final volume of 500 μL. The digestion proceeded overnight at room temperature. The solution was then diluted to 5 mL with Milli-Q water and filtered with a PVDF filter (0.45 μm) before the analysis. The intracellular Fe concentration was measured by inductively coupled plasma atomic emission spectrometry (ICP-AES, Thermofisher ICAP 6300 duo) and with the preparation of a Fe calibration curve (0.01–10 ppm).