Macrophage-Encapsulated Bioorthogonal Nanozymes for Targeting Cancer Cells.

Macrophages migrate to tumor sites by following chemoattractant gradients secreted by tumor cells, providing a truly active targeting strategy for cancer therapy. However, macrophage-based delivery faces challenges of cargo loading, control of release, and effects of the payload on the macrophage vehicle. We present a strategy that employs bioorthogonal "nanozymes" featuring transition metal catalysts (TMCs) to provide intracellular "factories" for the conversion of prodyes and prodrugs into imaging agents and chemotherapeutics. These nanozymes solubilize and stabilize the TMCs by embedding them into self-assembled monolayer coating gold nanoparticles. Nanozymes delivered into macrophages were intracellularly localized and retained activity even after prolonged (72 h) incubation. Significantly, nanozyme-loaded macrophages maintained their inherent migratory ability toward tumor cell chemoattractants, efficiently killing cancer cells in cocultures. This work establishes the potential of nanozyme-loaded macrophages for tumor site activation of prodrugs, providing readily tunable dosages and delivery rates while minimizing off-target toxicity of chemotherapeutics.

Drug targeting can reduce off-target toxicity, improving survival time and quality of life for cancer patients.[1−4] Recent metanalyses, however, have shown that the tumor targeting efficiency of most delivery systems is quite low[5−8] and that much of the delivered therapeutic payloads are taken up by tumor-associated macrophages rather than the targeted tumor cells.[9−11] Cell-based therapies provide a potential strategy for the enhanced targeting of tumors.[12−18] However, commonly used cellular vehicles face challenges for tumor targeting.[19,20] Macrophages are inherently attracted to specific tissue environments, including hypoxic, ischemic, and necrotic areas associated with tumors.[21−23] Concurrently, this homing ability is complemented by the secretion of macrophage chemoattractants by many cancer cell types to recruit macrophages to tumors[24,25] that can contribute up to 50% of tumor mass.[26] Taken together, these characteristics make macrophages particularly attractive for use as cell-based delivery vehicles for treating solid tumors that are difficult to reach using conventional targeting strategies.[13,15,27]

Previous therapeutic delivery studies using macrophages as cell-based carriers have loaded them with nanoparticle–drug conjugates or free chemotherapeutics.[14,16,17] These agents provided greater efficacy and reduced off-target toxicity when compared with free drugs.[17,21] The direct loading of therapeutics into macrophages is challenging, however, due to limitations in the amounts of therapeutics that can be loaded into the cells before compromising their viability and/or homing efficiency.[28] To a certain extent, these issues can be addressed by utilizing macrophages bearing stimuli-responsive nanoparticles that control the release of therapeutics.[29,30] In these approaches, external stimuli such as thermal energy, light, or ultrasound can be used to trigger the release of drug molecules from nanoparticles specifically at the tumor site. While this strategy is promising, limited loading and increased complexity of the release process remain as challenges.

Bioorthogonal catalysis[31−37] provides a strategy for creating drug “factories” for on-site activation of pro-imaging agents and prodrugs,[38−40] providing essentially unlimited quantities of active entities at desired cells, tissues, and organs.[41−44] We report here the integration of on-site manufacturing capability with the inherent homing properties of macrophages using cell-internalized bioorthogonal nanozymes (NZs).[45,46] These NZs use gold nanoparticles (AuNPs) to solubilize and stabilize transition metal catalysts (TMCs) through encapsulation in the AuNP monolayer. The TMCs can then generate imaging and therapeutic agents via bioorthogonal uncaging of inactive small molecule precursors in cells.[47−54] NZs were delivered into RAW 264.7 macrophages to provide NZ-loaded macrophages (RAW_NZ). The efficacy of RAW_NZ for pro-fluorophore and prodrug activation was demonstrated in a coculture model with HeLa (human cervical cancer) cells. Macrophages retain their migratory behavior toward colony stimulating factor-1 (CSF-1), a major chemoattractant secreted by cancer cells to recruit macrophages at tumor sites.[55] Significant cancer cell toxicity was achieved in the presence of NZ-bearing macrophages, even at the lowest concentration of the prodrug administered. Taken together, this strategy integrates the targeting ability of cell-based drug delivery with the on-site generation of therapeutics, providing a new approach for targeted drug delivery systems.[56]

Results and Discussion

Fabrication of Nanozymes

The NZ scaffold was provided by 2 nm cationic AuNPs functionalized with thioalkyl tetra(ethylene glycol) trimethylammonium (TTMA), previously shown to have both high cellular uptake and low toxicity.[57−61] The ligand monolayer of TTMA contains a crucial hydrophobic alkane interior for catalyst encapsulation and a tetra(ethylene glycol) spacer to provide biocompatibility and improve stability in aqueous intracellular environments (Figure ). TTMA particles were generated from pentanethiol-capped 2 nm core AuNPs using a place exchange reaction (synthesis and characterization of NPs are described in the Supporting Information and Supporting Information Figures S1–S3).[62,63]

Figure 1

Schematic representation of macrophage-mediated delivery of bioorthogonal nanozymes (TTMA-NZ) for prodrug (pro-5FU) and pro-fluorophore (pro-Rho) activation selectively at tumor cells.

Nanozyme TTMA-NZ was generated by immobilizing a hydrophobic palladium TMC (1,1′bis(diphenylphosphino)ferrocene)palladium(II)dichloride[64−66] in the hydrophobic portion of AuNP monolayers.[47] In practice, the catalyst was dissolved in acetone and added into an aqueous solution of AuNPs in a 1:1 ratio by volume. Acetone was evaporated slowly, and the excess catalyst was filtered away to provide the desired TTMA-NZ (details of NZ preparation and characterization are provided in the Supporting Information). There was no aggregation observed before or after encapsulation of TMCs, as confirmed by dynamic light scattering (Table S1 and Figure S4) and transmission electron microscopy (Figure S5). Quantification of Pd relative to Au by inductively coupled plasma mass spectrometry (ICP-MS) (Table S2) indicated that 37 ± 1 catalysts were present per AuNP.

Loading of Nanozymes into Macrophages

The toxicity of NZs to macrophages (RAW 264.7) was studied, with little effect on cell viability observed over a wide range of concentrations after 24 h incubation, as determined by Alamar blue assay (Figure a). We next performed hemolysis assays to further test the safety of NZs. As shown in Figure b, TTMA-NZ caused minimal hemolysis to human RBCs up to 300 nM. For the cellular uptake study, macrophages were incubated with TTMA-NZ for 24 h and then washed with phosphate-buffered saline (PBS) three times to remove noninternalized NZs. The use of additional washes has been shown to remove essentially all externally adsorbed NZs.[67] This process did not affect the effective internalization of TTMA-NZ, as determined by ICP-MS analyses of the levels of Au or Pd present in the macrophages (Figure S6). As expected, the resulting RAW_NZ macrophages feature Au/Pd ratios consistent with those of the precursor TTMA-NZ (Figures c,d and S6). The NZ content can be readily tuned through variation in TTMA-NZ concentrations during incubation, providing a ∼fourfold difference in the NZ level for RAW_NZ-50 nM versus RAW_NZ-250 nM after incubation. Significantly, retention of Au and Pd was observed over extended (48 and 72 h) periods (Figure c,d).

Figure 2

(a) Cytotoxicity of NZs on HeLa cells. (b) Hemolysis of red blood cells by different concentrations of TTMA-NZ. (c and d) Retention of NZs in macrophages after prolonged incubation time. Data from ICP-MS evaluation of (c) Au (ng/well) and (d) Pd (ng/well) nanozyme components in RAW 264.7 macrophages (20,000 cells/well) indicate retention after 24, 48, and 72 h incubations. The data shown are averages of experimental triplicates; error bars indicate standard deviations.

Catalytic Activity of Nanozymes in Living Cells

Effective application of NZ-based cell therapy requires the intracellular activation of substrates. The cell-internalized catalysts of RAW_NZ were used to uncage nonfluorescent di(propargyloxycarbonyl)-caged rhodamine 110 (pro-Rho)[65] as a pro-fluorophore. RAW_NZ was cultured in serum-containing media for 24, 48, or 72 h, followed by addition of pro-Rho and incubation for a further 24 h. Confocal microscopy imaging indicated that RAW_NZ successfully activated pro-Rho to rhodamine 110, independent of RAW_NZ generation time (Figures and S7). Efficient NZ activity was observed even after 72 h of cell internalization, consistent with the presence of NZ components observed via ICP-MS (Figure ). Taken together, NZs in RAW_NZ both remain inside of macrophages and retain their activity for prolonged periods.

Figure 3

Confocal images of pro-rhodamine 110 fluorophore (pro-Rho) activation by RAW_NZ. RAW_NZ was generated for 24, 48, and 72 h before the addition of pro-Rho. In each case, images were acquired 24 h following exposure of nanozyme to pro-fluorophore. Pro-Rho was used as the negative control, and Rho was used as the positive control. The scale bars are 15 μm.

Efficient Chemotactic Migration of RAW_NZ

We next investigated the effects of NZ internalization on macrophage response to chemotactic signals. CSF-1 was used as the chemoattractant, and migration was evaluated by transwell membrane (Boyden Chamber) assay[23] (Figure ). We compared the abilities of nonmodified macrophages (RAW 264.7) without an NZ versus those loaded with an NZ (RAW_NZ) to traffic through a membrane in response to CSF-1 presence. Macrophages were stained with crystal violet at the conclusion of the experiment to visualize and quantify the migrated cells (the detailed procedure is described in the Supporting Information). No significant differences were observed in the behaviors of macrophages (Figure b), in the presence or absence of CSF-1, indicating that the NZs do not affect the migratory behavior and chemotaxis capabilities of the macrophages. These in vitro Boyden chamber chemotaxis assays demonstrate that the migratory ability of the NZ-loaded macrophage (RAW_NZ) is consistent with that of untreated cells (RAW 264.7). As seen in the transwell images and the corresponding plots, the cell migration of the NZ-loaded macrophages and untreated cells toward chemoattractants was found to be essentially identical.

Figure 4

Chemotaxis capabilities are retained by RAW_NZ as determined by transwell membrane assay. (a) Confocal imaging of migrated macrophages with NZs (RAW_NZ) and without NZs (RAW 264.7) in the presence and absence of chemoattractant CSF-1. All cells were stained with crystal violet to facilitate detection. Scale bar = 100 μm. (b) Quantification of migrated RAW 264.7 cells and RAW_NZ in the presence and absence of CSF-1. Nine panels of cells were counted per treatment (n = 9, from three biological replicates). The box constitutes the interquartile range (25th to 75th percentile), the intersecting line designates the median, the small square in the center represents the mean, and the bottom and top whiskers specify the 5th and 95th percentiles, respectively. n.s. = not significant, ***p < 0.0001.

RAW_NZ Kills Cancer Cells in Coculture Models

Having established the stability and retention of inherent chemotactic capabilities of macrophages of RAW_NZs, we next investigated their therapeutic potential in a coculture model with HeLa cells (Figure ). For this study, propargyl-protected 5-fluorouracil (pro-5FU)[66] was chosen as a model prodrug due to the broad application of its active counterpart (5FU; Figure a).[68,69] The caged propargyl group blocks the active site of 5FU, reducing cytotoxicity ∼500-fold (Figure S8). The activation of pro-5FU was first studied in PBS solution and demonstrated by checking the thin layer chromatography (TLC, Figure S9). The reaction mixture was further analyzed using electrospray ionization mass spectrometry. The presence of the 5FU peak (m/z 129) indicates the successful activation of pro-5FU by NZs (Figure S10). For the coculture experiment, RAW_NZ or RAW 264.7 cells were seeded on glass slides that were then immersed into wells seeded with HeLa cells (Figure b). Macrophage-free slides and HeLa-free wells were used for additional control conditions. Pro-5FU (0 to 1 mM) was added to the coculture and control wells and incubated for 24 h. Slides with RAW_NZ and controls were removed before performing Alamar blue assays separately on HeLa cells and macrophages to differentiate viabilities by the cell type (details provided in the Supporting Information). In the presence of RAW_NZ, the viabilities of both HeLa (Figure c) and RAW 264.7 carrier (Figure d) cells were substantially reduced with increasing concentrations of the prodrug. This dose-dependent cytotoxicity indicates the successful conversion of pro-5FU into the active therapeutic by NZs. The activity against the HeLa cells demonstrates that the uncaged drug was able to diffuse from the macrophages to the target cells. HeLa (and macrophage) cells that received increasing concentrations of prodrug but were not cultured with RAW_NZ (or TTMA-NZ) did not show any reduction in cell viability, indicating successful caging of 5FU (Figure c,d). RAW_NZ was also cocultured with green fluorescent protein (GFP) expressing HeLa cells (GFP-HeLa) and separately, U2OS cells (GFP-U2OS). In these experiments, the toxicity of the activated prodrug toward target cancer cells was assessed via monitoring of fluorescent signals. The green fluorescence of both GFP-HeLa cells and GFP-U2OS cells was greatly reduced with the increasing concentration of prodrug (Figure S11), indicating cell death and effective therapeutic activity against both cell lines.

Figure 5

5FU activation in a coculture system with RAW_NZ. (a) Pro-5FU activation by TTMA catalysis. (b) Graphical scheme of the coculture experiment to evaluate therapeutic efficacy of RAW_NZ in HeLa cells. Viability of (c) HeLa cells and (d) RAW 264.7 cells following pro-5FU (pro-drug) activation by RAW_NZ in the coculture experiment and control conditions. The data are average of triplicates, and the error bars indicate standard deviations. n.s. = not significant, *p < 0.05, ***p < 0.001.

Carrier cell death is not an issue in the therapeutic setting, so long as it occurs after the target tissue has been reached. Diminished viability of RAW 264.7 cells was observed only in treatments where the NZ and prodrug were both present. Likewise, free NZs and AuNPs were nontoxic to cancer cells in the absence of the prodrug. NZs showed toxicity only in the presence of the prodrug, whereas AuNPs remained nontoxic at all prodrug concentrations (Figure S11). In related experiments, RAW_NZ was compared to free TTMA-NZ to evaluate whether cellular loading affects the catalytic activity of the agents with similar activity observed for the NZ and RAW_NZ (Figure S12). Based on our previous research,[23] we are optimistic about the targeting efficiency of macrophages toward the tumor; however, liver accumulation should be considered as a potential challenge. Dosing of the prodrug is expected to not be an issue since this 5FU prodrug has been used at the levels used for our studies recently in vivo, showing no toxicity to mice.[70]

Conclusions

In summary, we have demonstrated a macrophage-based bioorthogonal strategy for tumor imaging and therapy. Macrophages bearing internalized bioorthogonal NZs featured catalytic activity extending over days. The carrier macrophages retained chemotactic behavior and exhibited efficient generation of therapeutics in a coculture model, illustrating the potential of this macrophage-based therapy for therapeutic uses. Coupling the chemotactic ability of macrophages and the ability to generate therapeutic and imaging agents at tumor sites presents a new strategy for reducing off-target effects and extending on-demand delivery, providing the potential to bring drug-activating factories into tumors. Future studies are underway exploring the use of this platform for specific drug activation in in vivo tumor models.

Methods

The catalyst was dissolved in acetone (2 mg/mL) and mixed with AuNPs in a 1:1 ratio by volume. The organic layer was slowly evaporated. The resulting mixture was filtered using a syringe-driven filter unit (pore size = 0.2 μm). The excess catalyst was removed via multiple filtrations through a molecular weight cut-off filter (10 K), followed by dialysis.[49]

Preparation of Engineered Macrophages (RAW_NZ)

Macrophages (RAW 264.7) were seeded at a concentration of 20,000 cells/well in a 24 well plate and allowed to attach overnight (at 37 °C in a humidified atmosphere of 5% CO2). After 24 h, cells were washed three times with PBS to remove any dead cells. NZ solution (100 nM) in macrophage growth media (described above) was added to the cells and incubated together for 24 h (at 37 °C in a humidified atmosphere of 5% CO2), except where otherwise noted for specific experiments. After this, the cells were washed with PBS three times to remove any excess NZ, to provide RAW_NZ. RAW_NZ was detached via treatment with trypsin, depending on the nature of the experiment to follow.

Chemotaxis/Boyden Chamber Assay

Cell migration toward CSF-1 was investigated by a Boyden chamber assay following previously described protocols.[23,71] Briefly, a transwell membrane with 8 mm pore size was coated with 10 mg/mL fibronectin. After 4 h, the excess fibronectin was rinsed with PBS and left to dry overnight. The next day, designated wells of a 24-well plate received 650 mL of serum-free growth media supplemented with 40 ng/mL rCSF-1; control wells received serum-free growth media only. The fibronectin-coated inserts were then placed onto the wells. Next, 100 mL of RAW_NZ (100,000 cells/well) solution was added into each insert and incubated for 12 h at 37 °C and 5% CO2. Nonmigratory cells were removed with a Q-tip, and migratory cells at the bottom of the inserts were fixed in 4% formaldehyde and stained with a 0.1% crystal violet solution in 25% methanol. Membranes were removed precisely, mounted onto cover-glass, and visualized under a Zeiss Axio Observer Z1 with an Axio Cam 506 Color attachment under a 20× objective lens. The cells were counted from three random, nonoverlapping fields of view per membrane, with three membranes per condition (n = 3 × 3 = 9). Box and whisker plots were generated using OriginPro 2017.

Coculture of RAW_NZ with HeLa Cells and Viability Evaluation

Glass slides were coated with poly-lysine solution, washed with PBS, and dried overnight. Each of the dry glass slides was placed onto each well of a 6-well plate. Macrophages (RAW 264.7) (100,000/well) in standard growth media were seeded in designated wells with glass slides; 3 mL of media was used in each well to ensure that the glass slides remained fully immersed in solution. The cells were treated with 100 nM NZs for wells designated as RAW_NZ. In parallel, HeLa cells (100,000/well) were seeded in separate 6-well plates. All plates were stored under 5% CO2 at 37 °C for 24 h. HeLa cells were washed three times with PBS followed by adding fresh, standard growth media. The glass slides coated with RAW 264.7/RAW_NZ were thoroughly washed with PBS to remove any nonadhered cells and/or excess NZ. The slides were then removed carefully with a tweezer from their designated wells and placed within wells containing HeLa cells (atop the cells) for coculture. As a result, both HeLa cells and RAW 264.7/RAW_NZ were in the same wells and solution. For control experiments with only HeLa cells, blank glass slides were placed on top of the cells. For control experiments with only RAW 264.7 cells, the slides coated with RAW 264.7 cells were placed in wells containing only media. The cells were incubated with the prodrug/drug for 24 h, followed by washing three times with PBS. The glass slides were removed and placed in separate 6-well plates. All wells (with cells now separated) were thoroughly washed with PBS three times. Finally, 10% Alamar blue assay in serum-containing media was performed separately for HeLa and RAW 264.7 cells.