Nitrogen-Doped Carbon Nanotube Cups for Cancer Therapy

Abstract

Carbon nanomaterials have attracted significant attention for a variety of biomedical applications including sensing and detection, photothermal therapy, and delivery of therapeutic cargo. The ease of chemical functionalization, tunable length scales and morphologies, and ability to undergo complete enzymatic degradation make carbon nanomaterials an ideal drug delivery system. Much work has been done to synthesize carbon nanomaterials ranging from carbon dots, graphene, and carbon nanotubes to carbon nanocapsules, specifically carbon nanohorns or nitrogen-doped carbon nanocups. Here, we analyze specific properties of nitrogen-doped carbon nanotube cups which have been designed and utilized as drug delivery systems with the focus on the loading of these nanocapsules with specific therapeutic cargo and the targeted delivery for cancer therapy. We also summarize our targeted synthesis of gold nanoparticles on the open edge of nitrogen-doped carbon nanotube cups to create loaded and sealed nanocarriers for the delivery of chemotherapeutic agents to myeloid regulatory cells responsible for the immunosuppressive properties of the tumor microenvironment and thus tumor immune escape.

Graphical Abstract

Introduction

Nanomaterials, characterized by at least one dimension being at length scale between 1 nm and 100 nm, are able to effectively interact with different cell types and biological systems presenting an opportunity to be used as drug-delivery systems for improved delivery and efficacy compared to conventional carriers.^1–4 A multitude of drug delivery systems are being explored, including polymer-based nanocarriers,^{5, 6} carbon nanomaterial-based systems (e.g., carbon nanotubes and graphene oxide),⁷ inorganic nanoparticles (e.g., iron oxides, silica and gold nanoparticles),^{8, 9} metal-organic frameworks (MOF),¹⁰ covalent organic frameworks (COF),¹⁰ nucleic acid-based^{11, 12} and lipid-based systems.¹³ Particularly, the advances in the application of carbon nanomaterials have enabled the rational design of both their structure and properties to suit the targeted, versatile and efficient drug delivery requirements.^14–16 Carbon nanomaterials benefit from a myriad of modifications through covalent functionalization, the ability to not only interact with numerous cell lines but also penetrate cellular membranes, and to control their bio-distribution through the modification of their length scales.^{17, 18} Additionally, the presence of carbon particles has been shown to enhance the cellular uptake of micro-/nano-sized particles drug delivery systems.¹⁹ Carbon nanomaterials exhibit strong absorbance of infrared radiation, the ability to participate in Raman scattering, and can have additional photo-activatable properties which promise in vivo applications including light activated drug delivery, thermal ablation cell destruction, or localized biomolecule detection.^20–22

The graphitic lattice of carbon nanomaterials can be exploited for chemical functionalization through either strong covalent functionalization with therapeutic cargo,²³ or through noncovalent decoration utilizing weaker π-π stacking interactions with the aromatic domains of drug molecules, providing two separate loading mechanisms with different binding strength for delivery applications.²⁰ Carbon nanotubes can be further modified in order to make unique structures such as nano test tubes,²⁴ nanohorns,²⁵ or nanocups²⁶ where one side of the nanotube remains open while the other side of the nanotube is sealed off as shown in Figure 1A. These nanocups are a unique material that could be used for drug delivery applications through loading of the nanocup and sealing the open end. Loading the interior of the nanocapsules is theorized to occur through a nanoprecipitation mechanism.²⁷ During nanoprecipitation, the desired cargo has a stronger attraction to the interior of the capsule than the surrounding media which causes it to passively load into the capsule. This attraction can be controlled by managing the level of oxidation between the interior and exterior of the nanocapsule. Ajima, et al. illustrated that single-walled carbon nanohorns can be effectively loaded with cisplatin through nanoprecipitation, where an aqueous solution of nanohorns is mixed with cisplatin dissolved in DMF and the DMF slowly evaporated.²⁷ As the DMF evaporates, the cisplatin experiences lower solubility in the aqueous media and is driven into the interior of the hydrophobic nanohorn cavity producing cisplatin nanoparticles as shown in Figure 1B and 1C.

Figure 1.

High-resolution transmission electron microscopy (TEM) images of A) a representative nano test tube, nanohorn, or nanocup structure. Reproduced from Hu et al., J. Phys. Chem., 2007, 111, 5830-5834.²⁴ Copyright 2007 American Chemical Society. B) A single-walled carbon nanohorn prior to loading with cisplatin; and C) a single-walled carbon nanohorn after incubation with cisplatin in which the black spot is a cisplatin cluster, insets show illustration of before and after loading; scale bars in panels B and C are both 2 nm. Reproduced from Ajima et al.,Mol. Pharm., 2005, 2, 475-480.²⁷ Copyright 2005 American Chemical Society.

Our group has achieved the incorporation of nitrogen atoms into the graphitic lattice of carbon nanotubes at the point of chemical vapor deposition (CVD) synthesis resulting in a unique stacked cups morphology referred to as nitrogen-doped carbon nanotube cups (NCNCs).^{26, 28} Individual and/or short-stacked NCNCs were obtained through oxidation of the pristine fibers in a mixture of nitric and sulfuric acid followed by high energy direct sonication of the oxidized fibers, as shown in Figure 2. These nitrogen-doped carbon nanocups have also been synthesized through pyrolysis,^{29, 30} or post synthetic doping by either chemical³¹ or electrochemical³² functionalization. The incorporation of sp³ nitrogen atoms into the conjugated sp² graphitic structure of carbon nanotubes results in added strain to the carbon structure which bends the sidewalls of the nanotubes resulting in shorter segmented nanotubes. The disruption of the conjugated sp² structure of carbon nanotubes also results in changes to the local electronic environment resulting in distinct new interactions with the local chemical environment.³³

Figure 2.

(a) Schematic illustration of NCNC as synthesized by chemical vapor deposition and further separated by oxidation and sonication. Adapted with permission from ref.²⁸ Copyright 2013 MyJoVE Corporation. (b) TEM image pristine NCNC fibers and (c) high-resolution TEM image illustrating the stacked cup structure of adjacent segments, the white dashed lines indicate that the graphitic lattice of the carbon nanotube runs parallel with the inner cavity of the nanocup (d) TEM image of a separated NCNC after oxidation and sonication and (e) high-resolution TEM image of a separated nanocup showing the now accessible interior. Reproduced from Zhao et al., ACS Nano, 2012, 6, 6912-6921.³⁴ Copyright 2012 American Chemical Society.

The successful translation of carbon nanomaterials for applications in drug delivery and cancer therapy is predicated on the demonstration of their enhanced performance over that of existing therapies, in addition to documenting their nontoxic behavior with biological systems. Initial studies indicated that sp² hybridized carbon nanomaterials would be resistant to biodegradation resulting in prolonged circulation times, bio-persistence, and toxic profiles.³⁵ The origin of the perceived resistance of carbon nanomaterials to biodegradation are required to be well-understood for successful implemental in biomedical applications, and to minimize unintended exposures in both occupational and environmental settings. Carbon nanotubes in particular have been almost eliminated from the discussion for applications in drug delivery applications based upon a fiber-like structure which draws similarities to previous reports of the bio-persistence and health concerns associated with asbestos.³⁶ While carbon nanomaterials were initially theorized to result in oxidative stress, by functionalizing their surface, full biological degradation is achieved with minimal side effects.^{37, 38} More specifically, single-walled carbon nanotubes can be characterized by rapid excretion through the renal system resulting in a biological half-life of several hours.³⁹ Bio-persistence of multiwalled carbon nanotubes has been correlated to their level of surface functionalization, resulting in a method to control their circulations times with more functionalized materials experiencing shorter circulation times and less functionalized materials accumulating in the liver and spleen.²³ The incorporation of nitrogen into multi-walled carbon nanotubes further increases their biocompatibility resulting in nontoxic behavior completely distinguishable from the reported behavior of pristine carbon nanomaterials and asbestos alike.^{40, 41} The origin of this behavior is due to the nitrogen dopants serving as anchoring and initiation sites for enzymatic degradation resulting in nitrogen doped carbon nanocups representing a unique material for drug delivery applications.⁴²

Myeloid-derived suppressor cells (MDSC), a heterogeneous population of immature myeloid cells, have been shown to be integral in inhibiting antitumor immune response and supporting tumor growth and progression.⁴³ There are two different types of MDSC - polymorphonuclear MDSC (PMN-MDSC) which are morphologically and phenotypically similar to neutrophils, and monocytic MDSC (M-MDSC) which are similar to monocytes. In the tumor microenvironment, most M-MDSC differentiate into immune suppressive tumor-associated macrophages (TAMs) hence PMN-MDSC remain the major immunesuppressive myeloid cells.⁴⁴ While MDSC have routine biological functions, they are only found to circulate and accumulate under pathological conditions, such as sepsis, infection, and cancer. A direct comparison of MDSC from spleens and tumors of mice demonstrated that tumor MDSC have a more potent suppressive activity.^45–48 Splenic MDSC suppress antigen-specific immune response of T cells, while tumor-infiltrating MDSC from the same mice demonstrated more potent antigen-specific suppressive activity and also acquired ability to inhibit non-specific anti-CD3/28-stimulated response.^49–51 As a result, a highly suppressive environment is created in tumors and prevents rejection of tumors via immune mediated mechanisms. Recently, three populations of PMN-MDSC have been identified in tumor-bearing mice based on the expression of CD14 marker: classical PMN (CD14⁻) with no detectable suppressive activity; weakly suppressive CD14^int PMN-MDSC, and highly suppressive CD14^hi PMN-MDSC.⁵² In tumor-free mice, more than 95% of spleen PMN were CD14⁻. CD14^int PMN represented about 40% of all PMN in spleens of tumor-bearing mice. In tumor tissues, CD14^int and CD14^hi represented more than 75% of all PMN. More than 30% of all PMN were CD14^hi cells. This was associated with markedly higher expression of Arg1, Nos2, and Ptgs2. Using cytometry by time-of-flight (CyTOF) two distinct PMN populations were also identified in patients tumor tissues: population of classical PMN with phenotype CD11b⁺CD15⁺CD16⁺CD66b^hiCD10⁻Lox1⁻Arg1^+/−pSTAT3⁻S100A9^lo and PMN-MDSC with phenotype CD11b^hiCD15^hiCD66b^hiCD16^hiCD10^hiLox1^hiArg1^hipSTAT3^hiS100A9^hip38⁺. PMN-MDSC represented the vast majority of PMN in tumors (70%).⁵² PMN-MDSC have been shown to promote the spread of tumors and metastasis, intratumoral angiogenesis, and may maintain tumor resistance to current immunotherapies, antiangiogenic drugs and chemotherapeutic techniques.⁵³ A positive correlation of MDSC in peripheral blood with cancer stage and tumor burden has been shown in many types of cancer.^54–59 In meta-analysis, elevated MDSC in the circulation were found to be an independent indicator of poor outcomes in patients with solid tumors.⁶⁰ The circulating MDSC were negatively correlated with objective clinical response to check point inhibitors.^61–64

Tumor-associated PMN-MDSC have also been found to overexpress various oxidative machineries, including myeloperoxidase (MPO), which suppress the activity and lifetime of immune effector cells in the tumor microenvironment.⁶⁵ Interestingly, the oxidative machinery expressed by MDSC can also be used to initiate the effective biodegradation of carbonaceous nanoparticles with complete degradation of nitrogen-doped carbon nanomaterials reported.^{37, 38} The critical role of PMN-MDSC within the tumor microenvironment has made them an attractive target for advanced anticancer immune therapies.⁶⁶ We have recently shown that nitrogen-doped carbon nanotube cups (NCNC) can be effectively separated into individual cups and sealed with gold nanoparticles through a citrate reduction of chloroauric acid.⁶⁷ Upon interaction with MPO, gold corked NCNC (Au-NCNC) are able to be effectively uncorked representing an attractive drug delivery vehicle for specific targeting to PMN-MDSC-rich environments. Furthermore, Au-NCNC loaded with paclitaxel, a widely used chemotherapeutic, and injected into tumor-bearing mice are uncorked by MDSC in vivo altering the tumor microenvironment and decreasing tumor growth rates.⁶⁸ Because carbon nanomaterials can also be degraded by reactive intermediates generated by macrophages , particularly by peroxynitrite (ONOO⁻) produced by the combined action of iNOS (generating NO•) and NADPH oxidase (generating ${\text{O}}_2{}^{\underset{-}{•}}$),^{69, 70} new selective approaches targeting immune-suppressive macrophages may be developed within the general optimized anti-cancer strategies.

NCNC Cytotoxicity.

Pristine carbon nanotubes (CNT) have shown inherent toxicity to cells, therefore, understanding the complete cytotoxic profiles of these uniquely shaped NCNC is essential before application as a novel drug nanocarrier. NCNC are synthesized through a floating catalyst chemical vapor deposition (CVD) method resulting in segmented growth and the formation of individual cup segments held together through easily disrupted van der Waals forces. Unlike the synthesis of nano test tubes which relies on templating and etching²⁴ or the synthesis of carbon nanohorns which relies on laser ablation,²⁷ CVD synthesis of NCNC can be easily scaled up showing its ability for mass production and advantage over the synthesis of analogous materials. These adjacent cups are then able to be isolated into individual cups and short stacked segments through bath sonication in alkali salts⁷¹ or oxidation with mineral acids and high energy probe-tip sonication.⁶⁷ In comparing these different separation techniques, separation by sonication in alkali salts maintains the graphitic lattice of the carbon nanotube and minimizes incorporation of external dopants. However, oxidation with mineral acids and high energy sonication result in the incorporation of oxygen containing defect sites which have been shown to enhance the binding of carbon nanotubes to the active site of peroxidase enzymes thus enhancing their biodegradation.³⁵

As NCNC exhibit differing chemical and physical structure compared to pristine CNT, their interactions with various cell types are also theorized to be drastically different. The interaction of NCNC with HeLa cells and RPE-1 cells was therefore evaluated in order to identify any cytotoxic effects of NCNC and their ability to penetrate various cells.⁷² As shown in Figure 3A and 3B, the metabolic activity of HeLa cells was tested in response to incubation with varying concentrations of NCNC through water-soluble tetrazolium (WST-1) assay and cell proliferation activity as determined through cell counts. NCNC exhibited no toxicity either metabolically or proliferatively to HeLa cells despite varying concentration and prolonged exposure times (up to 84 h). This nontoxic behavior was also observed in RPE-1 cells which are an immortalized epithelial cell line. Additionally, exposure of HeLa cells to NCNC showed no change in the mitotic frequency of the cells as shown by measures of the number of cells in mitosis and the number of multinucleated cells. Unlike that of pristine CNT, incubation of HeLa cells with NCNC showed no increase in reactive oxygen species cellular levels, as shown in Figure 3C, which has been one of the largest concerns of graphitic carbon materials interacting with biological systems. Finally, NCNC that were imaged inside of cells were largely (50 – 60%) found to reside within double membrane vesicles as shown in Figure 3D. This indicates that NCNC are taken into cells through endocytosis, with the remaining NCNC seeming to have entered the cell through a passive “needle-like” penetration of the cell membrane which has been previously reported for pristine CNT.¹⁷ Another report has indicated that the presence of oxidized carbon particles can enhance the uptake of micro- and nano-sized particles into cells through nonendocytotic pathways.¹⁹ These data illustrates that NCNC are not only able to be rapidly taken-up by cells but that their interaction with cells does not result in cell damage, cell death, or alteration of normal cell cycle. The biocompatibility of NCNC enables them an ideal nanocarrier for drug delivery applications. It should be noted that conflicting results on the effect of nitrogen-doped carbon nanotubes on cell proliferation have been previously published.^{73, 74} We believe that the interaction of carbon nanomaterials with various cell types is unique to each individual material depending upon its synthetic conditions and chemical/physical composition. Therefore, instead of viewing all nitrogen-doped carbon nanomaterials as inherently toxic or nontoxic, each unique material should be properly characterized for its desired biological application.

Figure 3.

Metabolic and proliferation assays on HeLa cells treated with NCNC. (A) HeLa cells were treated with increasing concentrations of NCNC for a duration of 24 h. After incubation, cells were assayed for metabolic activity using a WST-1 assay, n = 3 readings for each concentration. (B) Cell number was calculated on HeLa cells treated with 2 μg/mL NCNC over a span of 84 h using a hemocytometer, n = 2 counts for each time point. (C) quantification of mean pixel intensity from a dihydroethidium (DHE) fluorescence assay to quantify the presence of reactive oxygen species. (D) TEM image of double membrane autophagic vesicles containing NCNC in HeLa cells after 48 hours, scale bar is 100 nm. Adapted with permission from ref. ⁷² (open access). Copyright 2019 MDPI.

Cargo Loading, Au Corking and Enzymatic Uncorking of NCNC.

Nitrogen dopants in separated NCNC have been found to be heavily distributed at the opening of the cup structure providing a modification site for facile functionalization. Gold nanoparticles can be preferentially coordinated to the open end of the cup through reaction of the NCNC with chloroauric acid and sodium citrate resulting in completely sealed nanocapsules as shown in Schemes 1 and 2 and Figure 4b.^{67, 68} This same incubation of pristine CNT with chloroauric acid and sodium citrate was found to have no preference in binding to the nanotube surface,⁷⁵ illustrating the importance of the nitrogen dopants and cup shape morphology for controlled modification with inorganic nanoparticles. The binding energies of gold clusters to the oxygen and nitrogen functionalities present in NCNC were determined through DFT calculations, which illustrated that gold clusters have a binding energy to aliphatic primary amines which is 20-30% larger than the binding energies to any other oxygen functionalities, specifically carboxylic acids and hydroxyl groups.⁶⁷

Scheme 1.

Review article outline for the application of nitrogen-doped carbon nanocups as targeted drug delivery vehicles to myeloid-derived suppressor cells for cancer therapy.

Scheme 2.

Illustration for the loading of NCNC with cargo and corking with gold nanoparticles, followed by release of cargo through incubation with enzymatic oxidation.

Due to the order of synthetic conditions, oxidation followed by separation, separated NCNC are characterized by an oxidized exterior providing aqueous stability and a non-oxidized, hydrophobic interior which acts as a driving force for loading of cargo with poor aqueous solubility. We believe this loading mechanism is related to nanoprecipitation as the poorly soluble cargo becomes more stable within the non-oxidized interior of the NCNC cavity.^{27, 68} In addition to TEM results shown in Figure 4, the cargo loading, Au corking and enzymatic uncorking process have been further confirmed by Raman spectroscopy (Figure 5). The presence of GNP on NCNC causes a strong surface-enhanced Raman scattering (SERS) effect, allowing sensitive detection of this hybrid material by Raman spectroscopy in biological samples. This SERS effect originates due to charge transfer between the GNP cork and the NCNC container as a result of direct contact between the two which allows for easier electronic interactions.⁷⁶ Figure 5a illustrates this point as Raman spectra of NCNC decorated with GNP (Au-NCNC) show significantly enhanced Raman signals as compared to unfunctionalized NCNC. Enhancements of about 15- and 18-fold were noticed for the intensities of D (~1350 cm⁻¹) and G (~1582 cm⁻¹) bands, respectively. The physical mixing of NCNC with commercial citrate-coated GNP does not result in this same Raman enhancement, which confirms the importance of direct contact between the GNP cork and the NCNC capsule for these SERS effects to be observed. This SERS effect can also be exploited as a means of detecting loaded cargo within the interior of the sealed Au-NCNC. Au-NCNC were synthesized through the normal citrate reduction of chloroauric acid in the presence of Rh123 which is a fluorescent dye that is also frequently used as Raman probes for high-sensitivity SERS analysis. We expect Rh123, given its structure and poor solubility in water, to have a stronger affinity for the interior hydrophobic portion of the NCNC as opposed to the oxidized exterior of the NCNC, providing a driving force for the loading of the Raman probe to the interior of the NCNC. Figure 5b confirms the successful cargo loading of Rh123, as the inherent signals of the molecule are also observed in the loaded and corked sample of Au-NCNC. However, if the Rh123 molecule is simply mixed with already synthesized Au-NCNC, this same Raman enhancement is not observed. These results suggest that desired cargo can be loaded into the interior of NCNC through a nanoprecipitation method and effectively trapped upon the completion of the GNP cork. Additionally, the SERS enhancement of the characteristic Raman peaks of Rh123, and presumably other Raman active molecules, provides an avenue for the qualitative confirmation of successful loading of Au-NCNC.

Figure 4.

TEM images of (a) separated NCNCs, (b) Au-NCNC (c) Au-NCNCs after incubation with MPO, H₂O₂, and NaCl for 5 days, and (d) Au-NCNC upon incubation with MDSC. (a, c) Adapted with permission from ref.²⁶. Copyright 2015 John Wiley & Sons, Inc. (d) Adapted with permission from ref.⁶⁸. Copyright 2018 Royal Society of Chemistry.

Figure 5.

(a) Raman spectra of separated NCNC (black), NCNC mixed with commercial GNPs (blue), and NCNC corked with GNPs by in situ reduction process (red). The dotted line indicates the baseline. (b) Raman spectra of free Rh123 drop-casted on a glass slide at the concentration of 15 μM (black), (1) the precipitate of NCNC functionalized with GNP in the presence of 0.15 μM Rh123, after repetitive wash, and (2) the precipitate of 0.15 μM Rh123 mixed with as-functionalized NCNC/GNP conjugates, after repetitive wash; the spectrum was taken at 10% laser intensity to weaken the NCNC background. (c) Raman spectroscopy illustrating the uncorking of Au-NCNC upon incubation with MPO, H₂O₂, and NaCl in comparison to pristine NCNC. Reproduced from Zhao et al., J. Am. Chem. Soc., 2015, 137, 675-684.⁶⁷ Copyright 2015 American Chemistry Society. (d) Fluorescence intensity of Rh123 loaded Au-NCNC in response to enzymatic oxidation over 30 hours. Fluorescence was measured by an Infinite M1000 Pro microplate reader with an excitation of 500 nm and emission of 525 nm.

Peroxidase enzymes, containing a ferriprotoporphyrin IX heme structure, have been shown to effectively degrade carbon nanomaterials which have been functionalized with oxygen or nitrogen defects,³⁷ however pristine carbon nanomaterials have been found to be resistant to such enzymatic degradation.⁷⁷ The rate of enzymatic degradation has also been shown to be dependent upon the level of oxidation or the number of defect sites present on the CNT surface,⁴² providing a method of control over the release rates of loaded cargo which can be tailored to the application of interest. Nitrogen defect sites have been shown to have a similar effect and represent areas of initiation for enzymatic degradation. As the nitrogen sites can be the start of degradation for the NCNC capsules, it is reasonable to suspect that the GNP cork would be removed first, followed by complete degradation of the remaining NCNC capsule. Therefore, exploration into the interaction of Au-NCNC with peroxidase enzymes, such as myeloperoxidase (MPO) or eosinophil peroxidase,⁷⁸ represent an attractive uncorking mechanism for the delivery of loaded cargo. In response to incubation with MPO/H₂O₂/Cl⁻ enzymatic oxidation environment or MDSC, the gold can be uncorked from Au-NCNC (Scheme 1 and Figures 4c and 4d) thus triggering a targeted drug release.⁶⁷ In Figure 5c, Au-NCNC were incubated with MPO, H₂O₂, and NaCl for 0, 24, and 48 hours before being collected and Raman spectra measured. As MPO incubation proceeds the gold nanoparticles are removed from the NCNC, as evidenced by TEM images (Figure 4c), thus lowering the Raman enhancement due to the association of the plasmonic particles with the graphitic structure. For comparison a Raman spectrum of separated NCNC without any attached gold nanoparticles is also presented.

The release of loaded cargo from Au-NCNC upon exposure to oxidative environments is illustrated through the incubation of Rh123 loaded Au-NCNC with peroxidase enzymes for 30 hours and the fluorescence of the Rh123 cargo detected upon exiting the cup interior. When excited at 500 nm, Rh123 exhibits fluorescence at 525 nm. Rh123 loaded Au-NCNC were incubated in MPO, NaCl, and H₂O₂ for 30 hours with fluorescence measurements taken every hour, the detection of fluorescence illustrates the release of Rh123 from the inner cavity of the Au-NCNC as a result of the separation of the gold nanoparticle cork and the open end of the NCNC. As can be seen in Figure 5d, fluorescence increases over the first 12 hours and slowly plateaus until a maximum intensity is reached after 24 hours. Fluorescence is seen within 1 hour of enzymatic oxidation showing an immediate delivery of cargo upon interaction with MPO. Additionally, the fluorescence intensity remains consistent between 24 and 30 hours illustrating that the degradation rate of released cargo is longer than 6 hours.

Since the enzymatic machinery of MPO can effectively degrade the graphitic carbon capsule, we must also consider if peroxidase enzymes are able to effectively degrade loaded cargo once it is released. As will be discussed in more detail below, paclitaxel is a common chemotherapeutic which on its own suffers from nonspecific targeting and short circulation time. However, we have shown that paclitaxel can be selectively delivered to the tumor microenvironment using the Au-NCNC delivery system. Previous reports have indicated that free chemotherapeutics degrade more rapidly than if the desired payload has been coordinated to carbon nanotubes.⁷⁰ Therefore, we explored the rate at which paclitaxel undergoes enzymatic degradation as a potential mechanism for the observed chemoresistance in tumor bearing animals by degrading chemotherapeutic agents in vivo. Liquid chromatography mass spectrometry (LC-MS) was used in order to determine the extent of paclitaxel degradation by the oxidizing environment required to uncork Au-NCNC. Paclitaxel was incubated with MPO and H₂O₂ in PBS buffer for 48 hours. As shown in Table 1, after incubation in the MPO oxidizing environment for 24 hours, about 25% of the paclitaxel remained, and after 48 hours, about 15% remained which suggests that the released drug will suffer from degradation under enzymatic oxidation conditions. However, in our previous report we showed that paclitaxel that is released from Au-NCNC does not appear to undergo the same degradation as free paclitaxel.⁶⁸ These results illustrate that the carbon nanocapsule delivery system is able to improve the efficacy of loaded therapeutic cargo by scavenging the oxidative potential of the enzymatic machinery. As phenolic derivatives are a good substrate to reduce Compound I (one of the strong oxidants generated from the MPO cycle),⁷⁹ hydroxyl groups on the outer surface of NCNCs make the carbon nanocapsules more susceptible to enzymatic oxidation than the therapeutic drug. In the absence of the Au-NCNC delivery vehicle, paclitaxel is rapidly degraded minimizing its interaction time with the tumor microenvironment. Incorporation of oxidative scavengers could improve therapeutic efficacy by increasing the lifetime of cargo under the harsh enzymatic conditions of MPO. As will be discussed below, the delivery of paclitaxel using the Au-NCNC container showed prolonged antitumor response over the course of a 3-week experiment, which is significantly improved as compared to treatment with free paclitaxel which has a short-lived antitumor effect.⁸⁰ This illustrates another advantage of using carbon nanocapsules for drug delivery as their degradation helps to protect the loaded cargo resulting in longer interaction times with the biological environment.

Table 1.

Quantified amount of paclitaxel upon MPO/H₂O₂/Cl⁻ enzymatic oxidation of Au-NCNC.

Sample	Calculated Paclitaxel Concentration (μg/mL)
0 hour	7.245 ± 0.696
24 hours	2.011 ± 0.177
48 hours	1.191 ± 0.186
0.02 mg/mL Paclitaxel in PBS buffer	18.004 ± 1.0148

In vitro Drug Release and Efficacy.

MDSC are immature myeloid cells which have not yet undergone differentiation into mature myeloid cells such as dendritic cells (DC), macrophages, and neutrophils.⁸¹ The production of MDSC occurs within bone marrow and where subsequent release into the bloodstream results in accumulation within lymphoid tissue and tumor masses. MDSC have been shown to accumulate in tumor-bearing hosts, resulting in the inhibition of effector T cells which prevents an effective antitumor immune response. Tumor-associated MDSC overexpress MPO, providing a naturally occurring route for the complete degradation of NCNC and can be exploited for the targeted delivery of loaded cargo directly to the tumor micro-environment. Targeting MDSC with high concentrations of chemotherapeutic agents to induce cell death has been previously reported, however, it has also been shown that low doses of paclitaxel can result in differentiation of MDSC into immune stimulatory DC.^{80, 82} The overexpression of MPO in MDSC and the ability to differentiate MDSC into DC upon delivery of paclitaxel represents an attractive situation to apply Au-NCNC for a novel immunotherapeutic approach.

Au-NCNC were loaded with paclitaxel and used to investigate their ability to delivery ultra-low doses of paclitaxel directly to MDSC. Similar to our previous results, the SERS effect of the GNP cork allowed for enhancement of the Raman peaks of the loaded paclitaxel in order to confirm the effective loading of the therapeutic cargo. Upon 48 h of incubation, NCNC are readily uncorked as shown by TEM images (Figure 4d) and the decreased Raman intensity (Figure 6c). The much lower Raman peaks than before the incubation indicates that most GNP are detached from NCNC. Both Au-NCNC loaded with paclitaxel and empty Au-NCNC were incubated with tumor-induced MDSC in order to determine the therapeutic efficacy of the material. As shown in Figure 6a, Au-NCNC were not enough to restore immune function in tumor-associated MDSC, however Au-NCNC loaded with paclitaxel were able to restore T cell proliferation to normal levels. This restoration of normal T cell activity was further associated with a decrease in the concertation of TGF-β, which is responsible for inhibiting T cells, upon incubation of MDSC with paclitaxel loaded Au-NCNC as shown in Figure 6b. Additionally, these results were also correlated with a 3-fold increase in the number of DC and an overall decrease in the number of MDSC in the tumor milieu. These results illustrate that not only are MDSC able to effectively degrade Au-NCNC but that this degradation results in the local delivery of paclitaxel which allows for MDSC differentiation into DC and a potentially enhanced antitumor immune response. Additionally, the SERS enhancement of the D and G peak of NCNC can also be used in order to track the material within cells through Raman spectroscopy. As shown in Figure 6c, cells imaged after incubation with Au-NCNC can be detected through Raman spectroscopy as a way of confirming interaction in specific tissues.

Figure 6.

(a,b) NCNC-delivered paclitaxel blocks immunosuppressive activity of tumor-associated MDSC. (c) Raman spectra of GNP-NCNCs after incubation with tumor-associated MDSC. The inset shows the optical microscopic image of NCNCs contained in the MDSC residues. Reproduced from Zhao et al., J. Am. Chem. Soc., 2015, 137, 675-684.⁶⁷ Copyright 2015 American Chemistry Society. (d) B16 melanoma cells were inoculated s.c. in syngeneic mice on day 0 and on day 6 animals were treated with saline (control), empty Au-NCNC, paclitaxel loaded Au-NCNC, and free paclitaxel. Adapted with permission from ref.⁶⁸. Copyright 2018 Royal Society of Chemistry.

Our in vitro results prove the ability for paclitaxel loaded Au-NCNC to be explored as a novel immunotherapy, this activity was explored in tumor-bearing mice. Previous work had illustrated that systemic delivery of paclitaxel with dosages from 1-5 mg/kg could restore immune function but required multiple injections or combination with other therapies to achieve these results.⁸⁰ The administration of a single dose of Au-NCNC with a loaded paclitaxel concentration of 1.8 mg/kg was investigated over the course of 3 weeks in comparison to direct injection of free paclitaxel as shown in Figure 6d. As can be observed, treatment with free paclitaxel results in a short-lived inhibition of tumor growth between days 5 and 9 in comparison to the control. This effect is quickly negated as the rate of tumor growth between days 9 and 14 are similar between the direct injection of paclitaxel and the control group with no therapy. Treatment with paclitaxel loaded Au-NCNC results in a pronounced effect over the entire course of the experiment where tumor growth is consistently suppressed as compared to empty Au-NCNC, free paclitaxel injection, and the control group. After day 15, 25-30% of the mice injected with paclitaxel loaded Au-NCNC were found to be tumor-free. We believe that this enhanced delivery of paclitaxel and resulting antitumor effect can be related to the slow release of paclitaxel from the Au-NCNC and the ability of the graphitic carbon capsule to scavenge the ROS produced in the harsh tumor microenvironment. All of these results illustrate that carbon nanocapsules make for a unique drug delivery vehicle which shows promising application in anticancer and immunotherapies.

Conclusions

We have summarized the recent synthesis and application of NCNCs towards applications in drug delivery and anticancer immunotherapies. NCNCs benefit from the facile scale-up potential afforded by their CVD based synthesis, easy functionalization through a vast array of organic reactions, ability for enzymatic degradation through biologically relevant peroxidase enzymes, and modification of physical structure through nitrogen doping and coordination with inorganic nanoparticles. We have also shown that unlike pristine carbon nanomaterials, nitrogen-doped carbon nanocapsules do not result in harmful interactions with different cell types. In the future, the level of oxidation and nitrogen doping could be altered in order to achieve a more effective control over the degradation rate of these carbon nanocapsules. Carbon nanocapsules could also be applied to other diseases and disorders associated with the recruitment and activation of innate immune cells and massive production of reactive intermediates.