Engineered pH-Responsive Mesoporous Carbon Nanoparticles for Drug Delivery

Abstract

In this work, two types of mesoporous carbon particles with different morphology, size, and pore structure have been functionalized with a self-immolative polymer sensitive to changes in pH and tested as drug nanocarriers. It is shown that their textural properties allow significantly higher loading capacity compared to typical mesoporous silica nanoparticles. In vial release experiments of a model Ru dye at pH 7.4 and 5 confirm the pH-responsiveness of the hybrid systems, showing that only small amounts of the cargo are released at physiological pH, whereas at slightly acidic pH (e.g., that of lysosomes), self-immolation takes place and a significant amount of the cargo is released. Cytotoxicity studies using human osteosarcoma cells show that the hybrid nanocarriers are not cytotoxic by themselves but induce significant cell growth inhibition when loaded with a chemotherapeutic drug such as doxorubicin. In preparation of an in vivo application, in vial responsiveness of the hybrid system to short-term pH-triggering is confirmed. The consecutive in vivo study shows no substantial cargo release over a period of 96 h under physiological pH conditions. Short-term exposure to acidic pH releases an experimental fluorescent cargo during and continuously after the triggering period over 72 h.

Introduction

Recent advances in nanotechnology have provided a new arsenal to modern medicine leading to the development of a new field, nanomedicine, which has inspired more specific and efficient treatments toward the treatment of complex diseases, such as cancer.^1–4 The benefits of using nanoparticles for drug delivery versus systemic treatments include enhancement of pharmacokinetic profiles, the possibility of releasing therapeutic molecules to specific tissues, thus reducing undesirable side effects, and the ability to bypass potential biological barriers. In this sense, nanoparticles acting as delivery vehicles of a variety of pharmaceutical agents currently represent ca. 75% of the market share of approved nanomedicines.⁵ A great variety of nanoparticles have been proposed as nanomedicines, spanning from lipid-based, protein, and polymeric nanoparticles as well as polymer-drug conjugates to different inorganic nano-particles.^6–8 Among them, mesoporous materials, and in particular, mesoporous silica nanoparticles (MSNs), have become very popular as the basis of smart drug nanocarriers because of their outstanding physicochemical properties, including tunable pore size, high pore volume, and large surface area, among others, that provide a great storage capacity within the porous network coupled with controlled release of the cargo due to nanoconfinement.^9–15 MSNs can easily be loaded using different techniques, such as electrospray or impregnation.¹⁶

In contrast, mesoporous carbon nanoparticles, presenting in general similar structural properties with traditional MSNs, have not been fully explored. Carbon nanoparticles have significant advantages over their silica counterparts with respect to their textural properties, which may translate into drastic benefits regarding their capacity to adsorb various molecules. For example, regular MSNs present specific surface areas and average pore volumes of ca. 1000 m²/g and ca. 1 cm³/g, respectively. On the other hand, mesoporous carbons offer specific surface areas of ca. 2000 m²/g and pore volumes of 1.5 cm³/g. In addition, mesoporous carbons present outstanding biocompatibility, great loading capacity of drugs showing reduced hydrophilicity, and provide superior loading capacity for aromatic drugs as a consequence of their additional supramolecular π- (or pi-) stacking interactions.^17–20

Although the open internal structure of mesoporous carbon nanoparticles makes them ideal candidates to introduce active compounds, the loaded biomolecules can easily diffuse out when the materials are placed in solution. Thus, closing the pore entrances is essential for the nanocarrier to control the cargo desorption and avoid premature release. One of the most successful approaches for blocking the pore entrances is coating the particle surface with a responsive polymer, so cargo leakage is impeded until some stimulus might change the conformation of the polymer, triggering the cargo release. Such hybrid carriers, known as stimuli-responsive systems, permit to tailor the release profiles of the cargo, enabling thus spatial, temporal, and dosage control. Various types of stimuli-responsive mesoporous carbon systems have been developed so far, showing sensitivity to for example pH,^21–23 redox,^24,25 enzymes,^26,27 ultrasound,²⁸ light,^29,30 temperature³¹ or magnetic fields.³² Among them, pH-responsive mesoporous carbon systems have been investigated the most because of the ample applications due to the natural pH gradients that exist between healthy and diseased tissues. In fact, in most tumors, the pH is lower than the physiologic value (7.4) because of the high rate of glycolysis of cancer cells, which leads to lactic acid accumulation. Additionally, other areas of the body present acidic pHs, such as the gastrointestinal track or some subcellular compartments, such as endosomes or lysosomes.³³ Some examples of pH-responsive gatekeepers employed to close the pore entrances of mesoporous carbons include biodegradable polymers,²⁶ small degradable nanoparticles,^34,35 or large polymers able to change their conformation and open the pores upon changes in the pH.^25,36 However, regardless of the number of scientific publications on biomedical applications of mesoporous carbon nanoparticles, this area is still in its infancy, with many significant challenges still to be solved before translation into clinical practice.³⁷

The primary aim of this work has been to investigate the loading capacity of mesoporous carbon particles and evaluate their potential use as pH-responsive delivery systems. For this purpose, we have developed responsive drug delivery systems based on two mesoporous carbon matrices (CMK-3 and spherical CMK-1) that are functionalized with a self-immolative polyurethane.³⁸ Because of their interesting chemical properties, several self-immolative structures have been used in nanomedicine.³⁹ In particular, we have used a polyurethane bearing a tert-butyloxycarbonyl (BOC) moiety as end-cap, which provides pH sensitivity (Scheme 1). At physiological pH, the pH-responsive trigger would preserve its integrity. However, after a drop in pH to more acidic values, the trigger unit would be cleaved, thus initiating the disassembly of the self-immolative backbone. As a consequence, the pores of the host matrix would be unlocked and the cargo would be released only in acidic environments. This coating approach has been successfully used by the authors for MSNs;⁴⁰ the advantage of using mesoporous carbons relies on the higher loading capacity that can be achieved (especially for the case of poorly soluble active pharmaceutical ingredients) coupled with the inert carbon surface that minimizes toxic effects. An additional benefit of this approach is the potential colloidal stabilization of the particles in aqueous media, which is a general problem related to unmodified carbon particles.

Scheme 1. Schematic Representation of the pH-Responsive Mesoporous Carbons^a.

^a At neutral pH, the self-immolative coating remains collapsed on the surface. However, when placed in an acid environment, the polymers undergo self-immolation, leading to payload release.

The pH-responsive release behavior of the two new systems has been demonstrated in vial, in vitro, and in vivo. In addition, the selective cytotoxicity of the drug-loaded materials in comparison to the nonloaded ones has been validated in vitro. Finally, in order to lay the grounds for future in vivo applications, an in vivo proof-of-concept trial has been performed in a preclinical rodent model. To confirm the in vivo pH-responsiveness, the integrity of the materials under physiological and acidic conditions was investigated, whereas local reaction toward the particles was monitored to assess the materials’ biocompatibility.

Results and Discussion

Properties of the As-Produced Mesoporous Carbon Carriers

Two different mesoporous carbons (namely CMK-3 and spherical CMK-1, denoted as C3 and C1Sph, respectively) were developed. Each carbon carrier was synthesized through a nanocasting procedure, using silica templates (SBA-15 and spherical MCM-48, respectively) that were removed after the generation of the carbon replica. The as-produced mesoporous carbons were characterized by scanning and transmission electron microscopy (SEM, TEM), N₂ porosimetry, small angle X-ray scattering (SAXS), and X-ray photoelectron spectroscopy (XPS). (Details are given in the Supporting Information).

As presented in Figure 1A, the rod-like elongated macro-structure characteristic of CMK-3 type carbons was obtained for the C3 material,⁴¹ showing a diameter of 0.2–0.3 μm and length of 0.5–1 μm. The aligned mesochannels of the C3 material (Figure 1A, inset) are characteristic of their 2-D periodic hexagonal mesostructure. With regard to the C1Sph material, the nanoparticles presented an average size of ca. 150 nm and a spherical morphology (Figure 1B). In addition, the TEM microscopy revealed the expected ordered porous network of the herein synthesized carbon particles (Figure 1C for C3 and Figure 1D for C1Sph).

Figure 1.

SEM and TEM micrographs of the as-synthesized carbons. (A) SEM micrographs of C3 carbons with rod-like elongated macrostructure. Inset: C3 mesochannels. (B) SEM micrographs of C1Sph with well-defined spherical morphology. (C) TEM micrographs of C3. (D) TEM micrographs of C1Sph.

The pore network structure of the carbon materials was evaluated through SAXS measurements (Figure S4). The pattern obtained for C3 shows the (10), (11), and (20) reflections characteristic of a 2D and hexagonally ordered array of pores (P6mm space group).^42,43 On the other hand, the two well-defined diffraction peaks (110) and (211) in the C1Sph pattern are compatible with the 3-D cubic pore structure (I4₁32 space group) of this sample.⁴⁴

The porous structure of the carbon materials was analyzed by N₂ adsorption–desorption measurements at 77 K (Figure S5), revealing their excellent textural properties (see Table 1 and the Supporting Information for further details). In brief, C3 has a mean pore size of 4.5 nm, a Brunauer–Emmett–Teller (BET) area of 1340 m² g^–1, and a total (micro- and meso-) pore volume of 1.4 cm³ g^–1, whereas C1Sph (mean pore size 3.2 nm) displays a BET area of 1650 m² g^–1 and a total pore volume of 1.2 cm³ g^–1.

Table 1. Comparison of the Textural Properties and the Organic Matter Content of C3, C3-SIP, C1Sph, and C1Sph-SIP.

	S BET (m2/g)	TPV (cm3g)	V micro (cm3/g)	V meso (cm3/g)	pore width (nm)	% weight loss
C3	1340	1.4	0.1	1.3	4.5
C3-SIP	620	0.6	0.0	0.6	4.1	14.8a
C1Sph	1650	1.2	0.2	1.0	3.2
C1Sph-SIP	560	0.4	0.1	0.3	3.0	14.9a

^aWeight loss compared to the nonmodified carbon carriers.

Loading Capacity of the As-Produced Mesoporous Carbon Carriers

A model ruthenium complex was loaded in conventional MSNs and in both carbon carriers to evaluate whether their remarkable textural properties translated into a higher loading capacity. Indeed, thermogravimetric analyses (TGA) revealed that the C3 and C1Sph mesoporous carbons could accommodate significant quantities of Ru dye. For the case of C1Sph, the amount of Ru molecules is almost 3 times higher than that adsorbed by conventional MSNs, which were used as reference (Figure 2). The obtained data confirm that the storage capacity is related to the material surface area, as initially assumed. Furthermore, it can be concluded that because mesoporous carbons have in general increased areas compared to silicas, there are indeed significant advantages in using the former as drug carriers at least in terms of loading capacities.

Self-lmmolative Polymers

Having confirmed that greater surface areas led to greater loading capacities, the next step was to synthesize the pH-responsive coating. The monomer and the trigger (compounds 1 and 2, respectively) were produced from commercially available 4-aminobenzyl alcohol, using phenyl chloroformate for the former and BOC anhydride for the latter, which provided pH-sensitivity to the polymeric chain. The self-immolative polymer (compound 3) was produced from compounds 1 and 2 in the presence of a tin catalyst, yielding polymers composed of ca. 20 units (3300 g/mol).

Properties of Polymer-Coated Mesoporous Carbons

The surface properties of the as-produced carbon carriers were evaluated through XPS analyses to unravel the chemical groups available on the surface and develop an adequate grafting protocol. The spectra (Figure S6) indicated the presence of hydroxyl, carboxyl, and epoxy groups. In consequence, the synthesis of C3-SIP and C1Sph-SIP was accomplished by addition of the self-immolative polymer (SIP) to the carbon nanoparticles to form stable ether groups (Scheme S2).

The SIP-coated mesoporous carbon particles were characterized using TEM, N₂ adsorption–desorption at 77 K, TGA, dynamic light scattering (DLS), and XPS. TEM micrographs (Figure S7) confirmed the successful particle coating, whereas Table 1 shows the modifications in the textural properties and the organic content of the carbon carriers before and after the functionalization (see also Figures S8–S11). DLS measurements on pristine and coated particles demonstrated that the polymeric coating prevents agglomeration of the particles and thus improves their colloidal stability, making them more suitable for biological applications (Figure S12). In addition, the SIP-coated particles were subjected to harsh conditions (75 °C, 4 days) and then analyzed by DLS measurements, demonstrating that the coated particles remained unaffected and assuring their integrity under physiological conditions (Figure S13).

TGA showed a difference in weight loss of ca. 15% within the range 150–450 °C for both SIP-coated carbons when compared with the bare particles due to the presence of organic matter on the coated materials (Figure S8 for C1Sph-SIP and Figure S9 for C3-SIP). The successful SIP coating was also confirmed by comparison of the N₂ adsorption–desorption isotherms at 77 K measured for C1Sph-SIP (Figure S10) and C3-SIP (Figure S11) against those of the pristine carbons, revealing a great decrease of the BET area (ca. 65% for the C1Sph-SIP and ca. 55% for C3-SIP) and pore volume after SIP coating, as shown in Table 1. Furthermore, based on the N₂ adsorption isotherms (external surface) and the TGA results (SIP loading), grafting densities of ca. 0.4–0.5 μmol/m² were deduced for both samples (see the Supporting Information).

The hybrid materials were also subjected to XPS analysis. The obtained spectra (Figure 3) show some differences when compared to those of the bare particles (Figure S6), indicating their successful modification (Table S1 for detailed analysis). The amount of C–C/C–H peaks for the SIP-coated carbons appears reduced, while the signals that may pertain to C-N bonds (amines, 286.0 eV;^45–51 amides, 288.7 eV;⁴⁹ carba-mates, 290.1 eV^45,49), have in general increased. Nevertheless, the most significant evidence of the interaction of the polymer with the carbon surface is the presence of the N 1s photoelectron peak (400.5 eV)^45–47,50 for both SIP-coated carbon materials (Figure 3A,B, insets), which is ascribed to the carbamate groups present throughout the self-immolative polyurethane, indicating the successful grafting of the polymer on the surface of both C3 and C1Sph carbon materials.

Figure 3.

XPS analyses of the SIP-coated carbons. (A) C1Sph-SIP. (B) C3-SIP. The plots, that show the expected signals for a carbon framework, exhibit signals associated to the carbamates present in the polymers, confirming the interaction of the polymer with the surface.

Release Experiments from SIP-Coated Mesoporous Carbon Particles

Several release experiments were performed to evaluate the pH-responsiveness of the hybrid mesoporous carbons, using a model fluorescent ruthenium complex as cargo. The use of nontoxic dyes as the model payload to study the smart behavior of mesoporous nano-matrices is well-stablished and allows the evaluation of their stimuli-responsiveness without using cytotoxic compounds, thereby reducing potential experiment-associated toxicity.^52–56 To mimic the physiological conditions, the Ru-loaded and SIP-coated nanocarriers were subjected to orbital stirring at 37 °C, using two pH values (5 and 7.4), corresponding to the lysosomal environment and the physiological body fluids, respectively (Figure 4).

Figure 4.

Release experiments. (A) C1Sph-SIP. (B) C3-SIP. At neutral pH, the polymer remains on the surface, inhibiting drug release. Nevertheless, the polymer self-immolates at acid pH, leading the opening of the pores and subsequent release. Error bars indicate standard deviation (n = 3 per condition).

Both release kinetics clearly show that the polymer is capable of hampering the dye release at pH 7.4, which would assure the biosafety of the carriers in vivo. However, at pH 5, the polymer disassembles and the dye can be easily released from both mesoporous carbon particles in a sustained fashion. In particular, for C1Sph-SIP (Figure 4A) even after 50 h of experiment, only 20% of the dye was released at pH 7.4. Unlike the samples at physiological pH, a 5-fold release was measured for those at pH 5. With regard to C3-SIP (Figure 4B), 28% of dye was released after 50 h at pH 7.4, whereas a ca. 4-fold release was observed at pH 5. Compared to C1Sph-SIP, more dye was released at physiological pH for C3-SIP, which can be connected with the larger pore size of C3 carbon but also with its more open pore architecture compared to C1Sph (in CMK-3 the connectivity is infinite because all pores are directly connected with the external space; on the contrary, in CMK-1, there is a more confined worm-like pore system).

Cytotoxicity Studies

The in vitro biocompatibility and cytotoxicity (measured by Alamar Blue) of the polymer-grafted carriers was evaluated using human osteosarcoma (HOS) cells at 24 and 48 h for three particle concentrations (25, 50, and 100 μg/mL). For this purpose, three different setups were used for each nanocarrier. The biocompatibility was investigated using unloaded polymer-grafted carbons (denoted as X-SIP, X indicating either C3 or C1Sph). To evaluate their cytotoxic effect, both carbon materials were loaded with a widely used chemotherapeutic drug, namely doxorubicin, and then coated with the polymeric layer (denoted as X-DOX-SIP). In addition, to confirm that the polyurethane actually undergoes self-immolation upon exposure to acid pH and, consequently, can effectively act as the gatekeeper, an additional group of materials was prepared. For this purpose, both mesoporous carriers were loaded with DOX and then coated with an analogous polyurethane without the triggering moiety responsive to acid pH (denoted as X-NOBOC) (Figure 5).

Figure 5.

Cytotoxicity assay measured by Alamar Blue in HOS cells. (A) 24 and (B) 48 h experiments of C1Sph-SIP, C1Sph-DOX-SIP, and C1Sph-NOBOC. (C) 24 and (D) 48 h experiments of C3-SIP, C3-DOX-SIP, and C3-NOBOC. Data are mean ± SEM of three independent experiments performed in duplicate. *p < 0.05 vs X-SIP and control.

In both cases, the drug-free nanocarriers showed negligible cytotoxicity on the cells for all studied concentrations and times. Interestingly, a significant increase of cell proliferation was observed for C1Sph-SIP when a concentration of 50 μg/mL was employed (Figure 5A). This behavior has also been observed for carbon nanotubes on different cell lines.^57,58 When using C1Sph-DOX-SIP, the hybrid nanocarrier was capable of exerting cytotoxicity at 24 and 48 h (Figure 5A,B) for all concentrations, being higher at 24 h. Nonetheless, for C3-SIP, only the highest concentration was capable of inducing significant reduction of the viability at 24 h (Figure 5C). A plausible explanation for that would be that doxorubicin release from the particle is mediated by the disruption of π–π interactions between the carbon matrix and the host molecules, being faster at the beginning.³³ On this basis, the majority of cells would die within the first 24 h. Finally, in both cases, the control groups loaded with the drug and functionalized with a polymer without the pH-responsive trigger (in orange) did not show inhibition of the cell viability. In other words, the presence of the triggering moiety responsive to acid pH is required for the self-immolation of the polyurethane and subsequent opening of the pore entrances and drug release to take place, thereby verifying in vitro the pH-responsiveness of the polymer-coated carbons.

In general, the cytotoxic effect of C3-SIP is much lower than that observed for C1Sph-SIP. A plausible hypothesis would be that, generally speaking, small nanoparticles are internalized by cells more easily than those which are bigger and as such, a smaller amount of C3-SIP particles is expected to be present inside the cells. In consequence, less amount of drug would be released inside the cells, leading to a reduced effect on their viability. In this sense, only the 100 μg/mL C3-SIP group was found to exert cell inhibition at 24 h, whereas the particles at a concentration of 50 μg/mL induced the first cytotoxic effect at 48 h. Apart from the size, the higher loading capacity of C1sph compared to that of C3 might also be an explanation for the greater cytotoxic effect of the SIP-coated C1Sph carbons.

In Vivo Proof-of-Concept Study: Preliminary in Vial Short-Term pH Triggering

These promising results, both in vitro and in vial, called for further in vivo proof-of-concept validation of the systems designed here. C1Sph-SIP was selected for the in vivo experiments due to their outstanding textural properties, release profile, and cytotoxic effects on cells. The experiments were carried out in mice. Fluorescent Ru was chosen as a model drug in order to allow fluorescence-based detection of payload release from the carrier, both in vial and in vivo.

Before performing the in vivo experiments, the short-term pH-responsiveness of C1Sph-SIP loaded with Ru was analyzed in vial (Figure 6), showing that only those materials incubated at acid pH, for either 30 min or 24 h, led to significant payload release. This result indicates that 0.5 h is enough time to trigger the self-immolation of the gatekeeper and initiate the payload release. In consequence, the SIP-coated mesoporous carbon nanocarrier should be unaffected by any potential in vivo homeostasis.

Figure 6.

In vial release experiment of C1Sph-SIP loaded with Ru. The materials were immersed in acid (pH 4, red) or physiological solutions (pH 7.4, black), and the supernatants were analyzed by the IVIS apparatus. After that, all groups were soaked in a solution at physiological pH and analyzed again 24 h later. This test confirms that just the groups that had been previously treated with the acid environment are capable of inducing significant and continuous payload release. Top images constitute representative IVIS images for each condition. Data are mean ± standard deviation (n = 3; n = 2 for 24 h trigger). Statistics: unpaired, two-tailed t-test, p* ≤ 0.05, conditions always referred to control per time point.

After this test, those materials were immersed in a solution at physiological pH (pH 7.4) for further 24 h, observing that only the materials that had been previously treated with the acid solution kept releasing the payload. On the other hand, the particles initially soaked at physiological pH kept showing residual release, unquestionably demonstrating the pH-responsiveness of the nanocarrier.

In Vivo Proof-of-Concept Study: Integrity of C1Sph-SIP and Biocompatibility

Having confirmed in vial that a short exposure to acid pH can trigger the payload release, the next objective was to find out in vivo if (1) the pores remain closed at physiological pH and (2) the materials induce in vivo reactions toward the particles.

To detect potential Ru release from the material (1), the mice (n = 3) were administered the hybrid particles in a physiological solution with pH 7.4 via subcutaneous (sc) injection. After that, the mice were visualized along the longitudinal axis using the in vivo imaging system (IVIS) before and after administering the particles. The animals were also imaged at different time points (24, 48, and 96 h) after the administration of the SIP-coated mesoporous carbon (Figure 7A). Of note, the Ru-loaded C1Sph-SIP carbons emitted imperceptible fluorescence signal, meaning that the SIP remained intact, therefore, closing the pores. The injected material itself cannot be visualized in the IVIS, only the release of the fluorophore can be tracked. Overall, only little fluorophore release was detected, indicating that there was no substantial degradation of the SIP polymer in vivo under physiological conditions and in the absence of a pH-trigger.

Figure 7.

Behavior of Ru-loaded C1Sph-SIP particles in the absence of an exogenous pH trigger after subcutaneous injection into mice. (A) Imaging of the dorsal site of the animals along the longitudinal axis using the IVIS before and directly post injection of the Ru-loaded C1Sph-SIP particles, as well as 24, 48, and 96 h post injection (images from left to right). The following filter set for excitation/emission was used: 465 nm/Cy5.5. Images of one animal are shown as examples. No exogenous pH-trigger was applied. (B) H&E-stained sections of paraffin-embedded skin tissue with (B1) and without (B2) injected C1sph-SIP (black) 96 h post injection. Nuclei stained in purple, cytoplasm and connective tissue in light pink, muscle tissue in dark pink. A and B comprise overview mosaic images, magnification 10×; the numbered and colored arrows point to the sites that are presented in higher magnification (40×) in a separate image. B1.I (green): moderate capsule formed near the administered mesoporous carbons and only low presence of immune cells can be observed; B1.II (yellow): substantial immune reaction against the wounded skin area due to tissue puncture during injection; B2.III (grey): control healthy skin tissue with dermis, subcutaneous adipose, muscle, as well as loose connective tissue. (Representative images are shown).

To study the general biocompatibility (2), the mice were visually analyzed each day for irritations of the site of the injection, general aspect, and behavior. Moreover, the weight was monitored at all imaging time points. No adverse reactions due to the presence of the materials were observed for any of the mice. The mice were imaged after 96 h and subsequently sacrificed. The injection site was explanted and further prepared for histological analysis. Figure 7B entails representative hematoxylin and eosin (H&E) stained images of skin and subcutaneous tissue with injected particles (Figure 7B1) as well as noninjected skin tissue as control (Figure 7B2). The organism reacted to the injected materials by forming a moderate capsule around the mesoporous nanocarriers and by low presence of immune cells in the injection site (Figure 7B1.I), compared to the injection-free skin (Figure 7B2.III). Figure 7B1.II shows the injection site where the injury from the needle results in a more pronounced immune reaction. The comparison of the immune reaction toward the skin wound (Figure 7B1.II) with the reaction toward the particles (Figure 7B1.I) shows only a mild foreign-body response to the material.

In Vivo Proof-of-Concept Study: Fluorophore Release in Response to Exogenous pH-Triggering

Having confirmed the responsiveness of the material to short-term pH-triggering in vial, the only minor reaction of the organism to the particles, and the integrity of the hybrid system over 96 h in vivo, it was investigated whether the self-immolation of the polyurethane could be initiated in an in vivo setup upon application of an exogenous acid pH trigger. A solution at pH 4 was used to slow down pH homeostasis that might interfere with the pH-sensitive particles. As a control, a physiological pH trigger was included (pH 7.4). For this, pH-responsiveness was monitored with the IVIS; the trigger was applied via injections of pH solution (multiple cycles of injection of pH solution followed by 10 min incubation time, adding up to at least 30 min of triggering) into the injection site. The animals were imaged at different time points (before as well as during and after applying the trigger). The mice were also imaged at 24 and 72 h after the triggering period. In a proof-of-concept approach, fluorophore release in response to triggering with low pH (pH 4, n = 4) was detected during the trigger period, whereas no fluorophore release could be observed after administration of the control solution (pH 7.4, n = 2) (Figure 8).

Figure 8.

Monitoring of Ru release from C1Sph-SIP particles in the presence of an exogenous pH trigger after subcutaneous injection into mice. Longitudinal IVIS imaging over 72 h, shown for two animals (1, 2) that received particles and pH trigger [1 = pH 4 (top); 2 = pH 7.4 (bottom)]. pH triggering was performed via multiple injections using a dwelling cannula over 30 min and imaged during the trigger period and over 20 min, 24 and 72 h post triggering. The animals were sacrificed, and the areas where the injections took place extracted, inverted, and visualized again. The following filter set for excitation/emission was used: 465 nm/Cy5.5. Representative images for each condition are shown.

After the initial triggering, the fluorescence of the Ru payload could be continuously detected during the whole experiment (72 h) for the group that received the solution at pH 4, whereas just a small amount of fluorescence was seen for the mice that received the solution at pH 7.4. The mice were euthanized after 72 h, and the place where the injections took place was collected, inverted, and visualized, showing strong fluorescent signal for pH 4-triggered hybrid systems and only minor signals for the pH 7.4-treated animals. This indicates a successful SIP degradation in response to administration of exogenous acidic pH-triggering over a short time period and, thus, validated the pH-responsiveness of the hybrid system in an in vivo and more complex setting.

Conclusions

In this work, two types of mesoporous carbon particles with different morphology and pore structure have been successfully synthesized and functionalized with a pH-responsive polymer. The initial hypothesis of the higher loading capacity of the carbon materials, compared to the well-studied mesoporous silica nanoparticles, has been verified through TGA, showing up to 3 times the loading capacity of conventional mesoporous silica. The acid-responsive nature of the carriers has been in vial, in vitro, and in vivo at physiological and lysosomal pH, demonstrating that the drug release would only take place inside the target cell. Moreover, the cellular experiments on HOS cells have demonstrated that only the doxorubicin-loaded hybrid materials are able to induce significant decrease in the cell viability, especially when using the C1Sph-SIP materials. Furthermore, the integrity of C1Sph-SIP, which presented the highest loading capacity, the best release profile, and the highest cytotoxicity effect on tumoral cells, has been confirmed over 96 h in vivo and no adverse reactions were observed during the investigated time period, showing the high biocompatibility of the material. Additionally, the rapid pH-responsiveness of the hybrid system to short-term triggering with acidic pH was successfully demonstrated in vivo with a continuous payload release over a 72 h period. In consequence, the obtained results demonstrate the suitability of the studied carbon materials to be considered as highly efficient and biocompatible smart drug delivery nanocarriers of, for example, anticancer active agents.

Materials and Methods

Materials

Tetraethyl orthosilicate (TEOS); Pluronic P123 and F127; cetyltrimethylammonium bromide (CTAB); phenyl chlorofor-mate; 4-aminobenzyl alcohol; dibutyltin dilaurate (DBTL); N,N-diisopropylethylamine (DIPEA); tris(2,2′-bipyridyl)-dichlororuthenium(II) hexahydrate (Ru), and di-tert-butyl dicarbon-ate were purchased from Sigma-Aldrich Inc. Solvents (dimethyl sulfoxide (DMSO); tetrahydrofuran (THF); dichloromethane (DCM), N,N-dimethylformamide (DMF); ethanol; heptane; ethyl acetate; methanol) were also purchased from Sigma-Aldrich Inc.

Synthesis of CMK-3 Type Ordered Mesoporous Carbon Nanoparticles

CMK-3 type ordered mesoporous carbons (denoted as C3) were produced through a nanocasting procedure,^44,59 employing SBA-15 mesoporous silica as the template. Sucrose was employed as the carbon precursor. The silica template was synthesized following previously reported methods,^60–62 using TEOS as the silica source and the triblock copolymer Pluronic P123 as the surfactant templating agent. The following composition was employed: TEOS (2 g): P123 (1 g) in 38 mL of HCl (1.6 M). The mixture was placed in an autoclave at 35 °C for 20 h and then aged at 90 °C for 24 h, dried and finally calcined at 550 °C in air to remove the P123 soft template. The calcined SBA-15 was impregnated twice with 5 mL of aqueous solution containing 1.25 g (first impregnation) and 0.8 g (second impregnation) of sucrose per gram of silica and minute amounts of sulfuric acid (95–97%) as the catalyst. After each impregnation step, the mixture was thermopoly-merized at 100 °C (6 h) and 160 °C (6 h). The final carbon replica was obtained by carbonization of the composite material at 900 °C (ramp rate 10 °C/min) for 2 h under N₂ flow (80 cm³/min), followed by cooling to room temperature and dissolution of the silica framework using 48% hydrofluoric acid solution at room temperature.

Synthesis of CMK-1 Type Ordered Mesoporous Carbon Nanospheres

A similar procedure was also adopted for the preparation of CMK-1 type mesoporous carbon with spherical morphology (denoted as C1Sph). In this case, MCM-48 ordered mesoporous silica spheres were employed as the starting hard template and were synthesized according to the Stöber method with some modifications,⁶³ employing a mixture of surfactants in a solution containing H₂O, NH₃, and EtOH. The reaction was carried out at room temperature and under static conditions. In brief, TEOS was used as the silica source, CTAB (cationic surfactant) was employed as the templating agent, and Pluronic F127 (nonionic surfactant) as suppressant of the silica particle grain growth.¹⁸ The carbon replica was obtained following the same procedure as for C3, that is, infiltration (twice) of the calcined MCM-48 silica “mold” with acidic sucrose solution, thermopolymerization, carbonization at 900 °C under N₂ flow, and silica removal with HF.

Synthesis of Mesoporous Silica Nanoparticles

Apart from the C3 and C1Sph carbons, mesoporous silica nanoparticles (denoted hereafter as MSNs) were also synthesized (to serve as control) following a modified Stöber method reported elsewhere.⁶⁴ In brief, 1 g (2.74 mmol) of CTAB was dissolved in a flask containing 480 mL of H₂O and 3.5 mL of NaOH at 80 °C. After that, 5 mL (22.39 mmol) of TEOS was added dropwise (0.25 mL/min) and the whole reaction mixture was kept for 2 h at 80 °C with stirring. Afterward, the nanoparticles were centrifuged and washed (water, ethanol). Finally, the template was removed refluxing the as-produced nanoparticles 3 times in 500 mL of an ethanolic solution (95%) of NH₄NO₃ (10 mg/mL).

Analytical Methods

Details on ¹H NMR, SAXS, N₂ adsorption–desorption (77 K), XPS, TGA, DLS, SEM, and TEM measurements are given as Supporting Information.

Determination of the Loading Capacity

Portions of 15 mg of each of the three nanoparticles (MSNs, C3, and C1Sph) were soaked in 2 mL of ethanol containing 34 mg (0.04 mmol) of Ru overnight. The impregnated materials were filtered, washed with ethanol (40 mL), and vacuum-dried. The loading capacity was evaluated through the TGA of all Ru-loaded materials, in comparison with each unloaded counterpart. The percentage of weight loss due to the cargo was calculated between 250 and 350 °C.

Synthesis of Phenyl ((4-Hydroxymethyl)phenyl)carbamate (Compound 1)

For the synthesis of compound 1, a slightly modified reported method was followed.⁶⁵ In brief, 1.5 g (12 mmol) of 4-aminobenzyl alcohol was dissolved in 60 mL of a mixture of THF/saturated aqueous sodium bicarbonate/water (2:2:1). After that, 1.7 mL (12 mmol) of phenyl chloroformate was added dropwise and the reaction was stirred at room temperature until it was completed (controlled by TLC chromatography). Then, compound 1 was extracted in ethyl acetate and washed with sodium bicarbonate. Afterward, the organic phase was dried using sodium sulfate as the desiccant agent. Finally, the solvent was removed and the crude was recrystallized in chloroform to yield compound 1, which was characterized by proton nuclear magnetic resonance (¹H NMR).

Synthesis of tert-Butyl ((4-Hydroxymethyl)phenyl)-carbamate (Compound 2)

For the synthesis of compound 2, a slightly modified reported method was employed.⁶⁶ In brief, 1 g (8.12 mmol) of 4-aminobenzyl alcohol was dissolved in 14 mL of DCM. After that, 1.94 g (8.75 mmol) of di-tert-butyl dicarbonate in 3 mL of DCM was added, and the mixture was stirred overnight at room temperature. Then, the solvent was removed and the crude was purified on a silica column (ethyl acetate/heptane, 1:1), yielding compound 2, which was characterized by ¹H NMR.

Synthesis of Poly(phenyl (4-hydroxymethyl)phenyl)-carbamate (Compound 3)

Compound 1 (1 g, 4.12 mmol) was dissolved in 2.1 mL of dry DMSO. Then, 73 μL (3% mol) of DBTL were added and the reaction was stirred for 2 h at 85 °C. After that, the reaction was cooled down to 40 °C and 223 mg (1 mmol) of compound 2 in 0.5 mL of dry DMSO were added slowly. Then, the mixture was heated again at 85 °C and stirred for 2 h. Finally, compound 3, which was isolated by precipitation of the reaction in cold MeOH and subsequent centrifugation, was dried and characterized using ¹H NMR.

Synthesis of Self-Immolative Polymer-Coated Carbon Nanoparticles

Thirty mg of either C1Sph or C3 nanoparticles was dispersed in 3 mL of dry DMF. In a different vial, 300 mg (0.09 mmol) of compound 3 and 31 μL (0.17 mmol) of dry DIPEA were dissolved in 2 mL of dry DMF. The mixture was stirred for 30 min at room temperature. Then, the latter solution was added to the nanoparticles dispersion and the temperature was set at 80 °C for 24 h. Finally, the newly formed hybrid materials (denoted as C1Sph-SIP and C3-SIP, respectively) were centrifuged, washed (DMF, water, and ethanol), and dried. The hybrids were characterized by means of TGA, Z-potential, N₂ adsorption, XPS, and TEM.

Release Experiments from Ru-Loaded Polymer-Coated Carbon Nanoparticles

The carbons were loaded with a Ru dye before the SIP coating to evaluate the pH-responsiveness of the hybrid materials. For that purpose, 30 mg of either C1Sph or C3 were dispersed in 3 mL of dry DMF containing 40 mg (0.05 mmol) of Ru, and the solutions were stirred for 24 h at room temperature. Then, the protocol described above for the grafting of the polymers was followed again.

The evaluation of the pH-responsiveness of each hybrid nano-carrier was carried out via in vial release experiments. Two pH values were employed, namely pH 7.4 (0.01 M phosphate buffer) and pH 5 (0.01 M acetate buffer). For this purpose, two batches of 10 mg of the corresponding nanocarrier were prepared and dispersed in 1.8 mL of the corresponding buffer solution. Then, 0.5 mL of each suspension were placed on a Transwell permeable support with a 0.4 μm polycarbonate membrane (n = 3 for each condition). The corresponding medium (1.5 mL) was placed in the external well, and the suspension was kept under continuous orbital shaking (100 rpm, 37 °C). The solution from the external well was taken at every time point, and fresh medium was added. The dye released was determined by fluorescence spectrometry.

Cytotoxicity Studies

Mesoporous carbons were loaded with a cytotoxic drug to evaluate their effect on tumoral cells. For that purpose, 30 mg of either C3 or C1Sph was dispersed in 3 mL of dry DMF containing 42 mg (0.07 mmol) of doxorubicin hydrochloride. Then, the mixture was soaked at room temperature for 24 h. After that, the previously mentioned protocol for the grafting of the polymer was carried out again.

Cellular studies were carried out using HOS cells derived from a HOS (CRL-1543; ATCC, Manassas, VA). HOS cells (20,000 cm^–2) were seeded into each well of 24-well plates (Corning, CULTEK, Madrid, Spain) at 37 °C in a humidified atmosphere (5% CO₂) using Dulbecco’s modified Eagle’s medium (Sigma-Aldrich) containing 10% of heat-inactivated fetal bovine serum (Thermo Fisher Scientific) and 1% penicillin–streptomycin (Thermo Fisher Scientific). Then, the corresponding group of particles was placed in the corresponding wells, and the whole system was incubated for 24 and 48 h. Particle-free wells were also introduced and employed as controls.

Cell proliferation was determined by addition of Alamar Blue solution (Thermo Fisher Scientific) AbD at 10% (v/v) to the cell culture at each time point (24 and 48 h) of growth, following the manufacturer’s instructions. Fluorescence intensity was measured using excitation emission wavelengths of 570 and 600 nm, respectively, in a Unicam UV-500 UV–visible spectrophotometer.

In Vial Pretesting of the pH-Responsive Nature of Ru-Loaded C1Sph-SIP upon Short-Term pH-Triggering in Preparation of in Vivo Application

The pH-responsiveness of the hybrid system (C1Sph-SIP loaded with the fluorophore Ru as a model drug) to short-term pH-triggering was investigated in vial by suspending the material in physiological solution (Sterofundin, B. Braun Melsungen, Germany) adjusted to a defined pH of 4 and 7.4 with HCl and NaOH, respectively. Each group of particles was incubated for 30 min or 24 h under either pH 4 or pH 7.4 conditions at 37 °C. After the trigger period, the particles were collected by centrifugation. The supernatant was collected, recentrifuged (2000g, 5 min) to remove the remaining particles, and imaged with the aid of the IVIS Lumina (Caliper LifeSciences, MA; ex/em filter: 465 nm/Cy5.5), as well as quantified in a plate reader (Tecan Infinite Pro 200, ex/em 450/620 nm). The just-centrifuged mesoporous carbons were dispersed again in solution at pH 7.4 and incubated at 37 °C for further 24 h. Afterward, the protocol (centrifugation, imaging of the supernatant and quantification) was carried as described above.

In Vivo Proof-of-Concept Study

The study in mice was carried out in agreement with the German Animal Welfare Act and received the approval of the local animal protection authorities (LaGeSo; permit numbers: G 0293/17). The animals were kept under obligatory hygiene standards as monitored according to the FELASA standards. The animals had access to water and food ad libitum, were kept in gangs, and randomly assigned to groups. The temperature was set to 20 ± 2 °C and a light/dark period of 12 h was utilized.

The mice were anesthetized by inhalation of isoflurane (2%) mixed with oxygen during the procedures. Before the intervention, mice were administered 0.03 mg/kg of buprenorphine solution (Temgesic, Schering-Plough, NJ) as an analgesic via subcutaneous (sc) injection distant from the site of the material injection. Eyes were protected by eye ointment and animals were kept on a heating pad throughout the surgical intervention.

The C1Sph-SIP hybrid systems loaded with Ru were administered to the C57BL/6 mice at a concentration of 3 mg/200 μL sc dorsally close to the shoulder blades via injection through a 20 G needle (Sterican, B. Braun-Melsungen, Germany) or a dwelling cannula (Vasofix, B. Braun-Melsungen, Germany), respectively. Prior to the injection, the fur in the dorsal region was clipped to avoid potential interferences of the mice hair with the IVIS. To test the behavior of the Ru-loaded C1Sph-SIP hybrid systems in the absence of an exogenous pH-trigger, the animals (n = 3) were imaged longitudinally before and after, as well as 24, 48, and 96 h post injection of the material. For the investigation of the in vivo responsiveness of the hybrid system to exogenous pH-triggering, the animals (n = 3) received repeated sc injection (three injections of 100 μL in a 10 min interval) of physiological solution with a defined pH 4 or 7.4 solution (Sterofundin, B. Braun Melsungen, Germany, adjusted with NaOH or HCl) in the material injection site using a dwelling cannula. The mice were again imaged longitudinally before and after injection of the hybrid system, repeatedly during and shortly after the pH-triggering, as well as 24 and 72 h post intervention. The behavior, body weight, and the skin area of the injection site were monitored to detect any adverse reaction during the investigation period.

The imaging was carried out employing the IVIS apparatus and the Living Image 3.1 software. Images were obtained using excitation/emission filters of 465 nm and Cy5.5, respectively. The exposure time was set to 0.25 min. The images shown in the article are the result of the overlay of bright field and fluorescent images. The final visualization was carried out in conditions of deep anesthesia, which was achieved by intraperitoneal (ip) injection of medetomidine [1 mg/kg BW (Cepetor, CP-Pharma, Germany)] and ketamine [75 mg/kg BW (Inresa Arzneimittel, Germany)]. Animals were terminated afterward by cervical dislocation, and the injection site was harvested, inverted, and imaged using the IVIS as well as prepared for histological analysis.

Histological Analysis of Injection Sites from Mice

The explanted injection sites were extended in histological embedding cassettes (Tissue-Tec, Sakura Finetek USA, CA), and tissue was fixated in 4% PFA for 24 h. Consecutively, tissues were dehydrated and embedded in paraffin in 5 μm sections and were stained with H&E to analyze the corresponding tissue and the injected hybrid systems. Bright-field images are shown at given magnifications (10× and 40×).

Statistical Analyses

The results shown throughout the article are displayed as mean ± SEM, unless otherwise stated. Statistical evaluation was carried out using nonparametric Kruskal–Wallis test and post hoc Dunn’s test or two-tailed student’s t-test, when applicable, p < 0.05 was considered to be significant.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.0c01786.

Description of the analytical methods; schematic synthesis of self-immolative polymers and their ¹H NMR characterization; SAXS of pristine C3 and C1Sph; N₂ adsorption–desorption isotherms of pristine C3 and C1Sph at 77 K; XPS analysis of pristine C3 and C1Sph; Grafting protocol; TEM images of pristine and SIP-coated C3 and C1Sph; TGA and N₂ adsorption–desorption measurements/pore size distributions of the SIP-coated versus pristine C3 and C1Sph; colloidal stability of pristine and SIP-coated C3 and C1Sph; thermal stability of SIP-coated C3 and C1Sph; grafting density; XPS analysis of hybrid materials; and XPS survey of plain and SIP-coated C3 and C1Sph (PDF)