Optimizing Mannose “Click” Conjugation to Polymeric Nanoparticles for Targeted siRNA Delivery to Human and Murine Macrophages

Abstract

“Smart”, dual pH-responsive, and endosomolytic polymeric nanoparticles have demonstrated great potential for localized drug delivery, especially for siRNA delivery to the cytoplasm of cells. However, targeted delivery to a specific cell phenotype requires an additional level of functionality. Copper-catalyzed azide–alkyne cycloaddition (CuAAC) is a highly selective bioconjugation reaction that can be performed in conjunction with other polymerization techniques without adversely affecting reaction kinetics, but there exists some concern for residual copper causing cytotoxicity. To alleviate these concerns, we evaluated conjugation efficiency, residual copper content, and cell viability in relation to copper catalyst concentration. Our results demonstrated an optimal range for minimizing cytotoxicity while maintaining high levels of conjugation efficiency, and these conditions produced polymers with increased targeting to M2-polarized macrophages, as well as successful delivery of therapeutic siRNA that reprogrammed the macrophages to a proinflammatory phenotype.

Introduction

The development of bioorthogonal chemistry has allowed for transformative progress in using polymer chemistry in living systems. “Click” chemistry encompasses the most specific of these reactions and can be performed in aqueous conditions and in the presence of oxygen.¹ The discovery of copper-catalyzed azide–alkyne cycloaddition (CuAAC) in the early 2000s popularized the field since the addition of copper(I) salt significantly increased the reaction rate of 1,3-dipolar cycloaddition between azides and alkynes.^2,3 Further development of a Cu(II) “precatalyst” with a reducing agent, such as sodium ascorbate, enabled the use of solvents composed of water and alcohol, which is necessary for reagents not soluble in water alone.^2,4 The nonharsh reaction conditions of CuAAC allow for direct in vivo functionalization,^1,5−7 and the high selectivity allows for the use of click chemistry in conjunction with other polymerization techniques.⁸ However, one primary concern with this reaction is the potential for copper-associated toxicity that can result from residual copper catalyst.⁹⁻¹¹ Although copper-free click reactions, such as strain-promoted azide–alkyne cycloaddition, are available, the required cyclooctynes are difficult/expensive to synthesize, bulky and hydrophobic, and fail to produce regiospecificity associated with CuAAC.¹²⁻¹⁴ Therefore, it is crucial to establish a method that ensures the biological compatibility of CuAAC reaction products.

CuAAC is a powerful tool for functionalizing polymeric nanoparticles (PNPs) with a targeting moiety by conjugating each reagent with an azide or alkyne. Additionally, reverse addition–fragmentation chain-transfer (RAFT) polymerization can be combined with click chemistry to develop complex, “smart” PNPs that respond to physiological cues such as pH or temperature.^8,15−22 pH-responsive polymers can be used to induce endosomal escape and thus preserve the functionality of small interfering RNA (siRNA) for RNA interference (RNAi) therapies. These siRNA-condensing polymeric complexes (polyplexes) can then be decorated with small molecules using CuAAC. Our previous work has shown moderate success using a decorated triblock polymer,^23,24 but here, we describe click conjugation to an improved diblock copolymer. The triblock copolymer contained a 2-propylacrylic acid (PAA) block, but this reagent complicated the overall synthesis procedure, making the polymer less reproducible. For this reason, we developed a diblock copolymer, excluding PAA, which was simpler to fabricate, easier to reproduce, and maintained overall functionality.

The improved diblock copolymer comprises a poly(ethylene glycol) (PEG) corona and a diblock core consisting of dimethylaminoethyl methacrylate (DMAEMA) and butyl methacrylate (BMA).²⁵⁻²⁷ The polymer, termed PEGDB, can be fabricated with an azide group on the “outer” end of the PEG chain, which allows for click conjugation with an alkyne-functionalized targeting moiety. Here, we chose mannose to deliver RNAi therapies to M2-polarized macrophages, which overexpress the CD206 mannose receptor. Macrophages are the desired target since they are the most prevalent immune cell type in many tumors, and elevated levels of tumor-associated macrophages (TAMs) correlate with poor prognosis and reduced survival.²⁸⁻³² Recent studies have shown the therapeutic potential for targeting and “reprogramming” TAMs to function as M1 inflammatory macrophages, such as delivering siRNA against IκBα, an inhibitor of the classical nuclear factor-kappa B (NF-κB) pathway.^24,33−36 By conjugating a mannose-alkyne onto our azide-PEGDB polymer (MnPEGDB) via CuAAC, we develop a system capable of delivering therapeutic siRNA to the desired macrophage phenotype, while retaining the ability to induce endosomal escape. However, the residual copper in CuAAC reaction products and targeting efficacy of MnPEGDB have not been quantitatively characterized, creating uncertainties in the potential for toxicity or off-target effects of this polymer formulation.

When initially developing the diblock copolymer, we observed unexpected toxicity in the MnPEGDB group compared to that in a PEGDB control polymer (Supporting Information Figure S1). The presence of copper ions in the synthesis procedure of MnPEGDB led us to examine residual copper content postreaction. To this end, we evaluated conjugation efficiency and residual copper content using a range of copper catalyst concentrations. We also evaluated toxicity and uptake in several cell types. Once we determined the optimal reaction conditions, we used that polymer to deliver siRNA to polarized macrophages to quantify changes in gene and protein expressions related to macrophage phenotype.

Results and Discussion

In this study, we directly examined the effects of altering the copper catalyst concentration for CuAAC reactions on both conjugation efficiency and residual copper content. Although CuAAC chemistry has been used previously to functionalize/decorate materials used in living systems,⁵⁻⁷ our polymers, functionalized using a 1 mM copper catalyst concentration, exhibited evident toxicity in contact with immortalized human macrophages (Supporting Information Figure S1); this toxicity was not observed in similar mannose-functionalized polymer formulations used previously by our lab.^23,24 Additionally, we have previously used undecorated PEGDB polymers to deliver siRNA with limited toxic effects observed both in vitro and in vivo;^25,26,37 so, we hypothesized the resulting toxicity stemmed from residual copper in this iteration of our polymer system. To our knowledge, no studies have been performed that directly quantify and compare CuAAC conjugation efficiency and residual copper content as a function of copper catalyst concentration. Moreover, the targeting specificity to M2-polarized macrophages for the resulting polyplexes has not been described in relation to mannose decoration and residual copper. Here, we show that an optimal range for copper catalyst concentration can be achieved that improves M2 macrophage-specific targeting while also reducing copper-associated toxicity.

Synthesis and Characterization of Mannose-Functionalized Diblock Copolymers

The reaction scheme for functionalizing azide-PEGDB (AzPEGDB) with mannose-alkyne is shown in Scheme 1. To quantify the conjugation of mannose onto the AzPEGDB polymers at each catalyst concentration, Fourier transform infrared (FTIR) spectroscopy was used to examine the change in height of the azide peak at 2100 cm^–1. By comparing peak height in the MnPEGDB polymers to that of the precursor AzPEGDB, we estimated conjugation efficiency as a function of copper catalyst concentration (Figure 1A). The lowest copper catalyst concentration (0.1 mM) resulted in an azide peak height, which was not different from the height of the unreacted AzPEGDB. All other copper catalyst concentrations (0.25 mM and above) resulted in a significantly decreased peak height at 2100 cm^–1, suggesting the conversion of the azide to a 1,2,3-triazole bond (Figure 1B).^38,39 The average reduction in azide peak for the higher catalyst concentrations (excluding 0.1 mM) was 64%, which is in the range of normal conjugation efficiency values for click reactions involving large molecules/polymers that cause steric hindrance.^40,41 This significant decrease in peak height suggests successful click conjugation of mannose onto the polymers when using a catalyst concentration of 0.25 mM and above. There was no significant improvement for conjugation efficiency when using catalyst concentrations above 0.25 mM, which is a 4-fold reduction relative to the commonly reported CuAAC catalyst concentration for polymer conjugation.^23,42

Scheme 1. Copper Catalyst Concentration Controls the Level of Residual Copper in the Final Polymer.

Figure 1.

Characterization of CuAAC reaction products. (A) FTIR spectra of polymers, including a detailed view of azide peaks. (B) Quantified height of the azide peaks confirmed that the CuAAC reaction depends on copper catalyst concentration. (C) Residual copper content, normalized to the amount of polymer, revealed that altering the copper catalyst concentration does affect the amount of copper ions associated with the final polymer.

After fabrication of the MnPEGDB and removal of residual copper using a Chelex resin following previously published methods,^23,43 the polymers were dissolved in ethanol and molecular-grade water. Residual copper was measured using a colorimetric assay that detects Cu²⁺ ions, and the assay results (copper ions in μg/dL) were normalized to provide mg of copper per mg of polymer. Residual copper in the final polymer product was significantly reduced for all catalyst concentrations below 1 mM (Figure 1C). Residual copper concentration for the lowest tested catalyst concentration (0.1 mM) was not significantly different from the PEGDB control polymer, which was not exposed to the CuAAC reaction conditions. Residual copper in the polymers produced using the three middle catalyst concentrations (0.25, 0.5, and 0.75 mM) was not significantly different from each other.

Other groups have previously reported the use of copper ligands as an alternative technique for employing CuAAC reactions to minimize the amount of residual copper.^44,45 However, this method is primarily employed for click conjugation performed in biological systems, such as directly functionalizing cell membranes, since postreaction treatment is not feasible. For our applications, we can treat the resulting polymers with a chelating reagent that functions similar to the ligands by binding and removing excess copper ions. To minimize the complexity of polymer fabrication, we chose to use a copper salt catalyst that could be removed post-treatment. This technique revealed that copper catalyst concentrations of 0.25, 0.5, and 0.75 mM supported robust click conjugation and resulted in significantly reduced residual copper concentrations, suggesting that these were the optimal candidates for the synthesis of MnPEGDB intended for future use in living systems.

Formation of Mannosylated Polyplexes

To assess the toxicity and delivery of the mannosylated carriers, we complexed the polymers with oligonucleotides to form polyplexes (Scheme 2). The MnPEGDB polymers, as well as a nonmannosylated PEGDB control polymer, were dissolved in a pH 4 buffer to protonate the amine groups on the DMAEMA monomers. Then, either Cy5-functionalized dsDNA (Cy5-dsDNA) or siRNA was added to the polymer solution and allowed to complex for 30 min. The solution was restored to physiological pH by adding 5× volume of a pH 8 buffer. Polyplex diameter (z-average) and surface charge were examined to ensure all polymer formulations were similar. The average diameter was approximately 150 nm with no differences among polymer groups (Figure 2A,B). The polymer synthesis conditions were selected to produce micelle diameters of approximately 150 nm for eventual efficiency in tumor accumulation after intravenous injection in vivo.⁴⁶⁻⁴⁸ The polydispersity index for each polymer was less than 0.3 (Supporting Information Table S2), which is below the generally accepted value for drug delivery applications.⁴⁹ ζ-Potentials for all polymers were not statistically different (p > 0.05) and were near neutral, which is appropriate for intravascular administration (Figure 2C,D).⁵⁰ Overall, the size and surface charge characteristics of the polyplexes were appropriate for delivery to the tumor microenvironment after intravascular injection.

Scheme 2. MnPEGDB Is Complexed with Small Oligonucleotides (dsDNA or siRNA) for 30 min in a pH 4 Buffer.

The protonated amine group on the DMAEMA electrostatically interacts with the negative phosphates on the oligonucleotide, spontaneously forming a stable micelle with mannose presented on the corona of the polyplex.

Figure 2.

Polyplexes were suspended in deionized water, and Z-average diameter (A) distribution and (B) averages were measured with dynamic light scattering. Additionally, surface charge (C) distribution and (D) averages were measured by ζ-potential measurement.

In Vitro Polyplex Toxicity Examination

The toxicity associated with residual copper in polyplexes was examined in multiple cell types. All cells were treated with a polyplex concentration corresponding to a 50 nM dose of siRNA, as done previously.²³⁻²⁵ Cells were incubated with polyplexes for 24 h before evaluating viability with a luminescent ATP-reporting assay. ThP-1 immortalized human macrophages demonstrated a significant decrease in viability when treated with polymers prepared with 0.1 or 1 mM copper catalyst compared to untreated controls (Figure 3A). The polymer fabricated with 1 mM catalyst had the highest amount of residual copper (Figure 1C), which is presumed to lead to the observed toxicity. Notably, the 0.1 mM catalyst group, which had the least amount of residual copper, also produced significant toxicity. The FTIR results (Figure 1B), however, suggest that this polymer had the largest residual azide peak, indicating the least amount of mannose conjugation and, presumably, the largest number of residual azides on the polymer. Azides are also known to be cytotoxic, and we interpret the toxicity of the 0.1 mM polymer to be due to unreacted azides on the AzPEGDB.⁵¹ This interesting observation of azide toxicity has been neglected in previous considerations of cytotoxicity from CuAAC reaction products, which have focused exclusively on the role of residual copper. Clearly, the optimal copper catalyst concentration is a balance between excess residual copper and relatively few unreacted azides and low residual copper but a greater concentration of remaining cytotoxic azides. Therefore, the optimal catalyst concentration range to limit cell toxicity in human macrophages for our polyplexes is between 0.25 and 0.75 mM, which corroborates the results of click conjugation and copper content.

Figure 3.

Macrophage viability assessed with polyplex treatment and Cu²⁺ treatment. (A) Viability of ThP-1 human macrophages was significantly decreased when treated with polyplexes fabricated with 0.1 and 1 mM copper catalysts. (B) M0 and M2 bone-marrow-derived macrophages (BMDMs) displayed no toxicity in any treatment groups, but all M1 BMDMs had slight toxicity. (C) All macrophages displayed a trend of decreasing viability as the concentration of copper salt treatment increased.

To explore the impact of polyplexes on the viability of primary macrophages with varying levels of CD206 expression, bone-marrow-derived macrophages (BMDMs) were cultured with cytokines to polarize them to either an M1 (CD206_low) or M2 (CD206_high) phenotype. Interferon-γ (IFN-γ) and lipopolysaccharide (LPS) were added to polarize BMDMs toward M1, and interleukin- (IL-) 4 and IL-13 were used to induce M2 polarization.⁵² An additional group of unpolarized BMDMs (M0) was cultured in media with no cytokine treatment. BMDM skewing was characterized via flow cytometry by examining the expression of CD11b and F4/80 (general macrophage markers), CD86 (M1 marker), and CD206 (M2 marker) (Supporting Information Figure S5). Following cytokine treatment, we observed significantly higher levels of CD86 expression in M1 macrophages compared to those in both M0 and M2 phenotypes. M2-polarized BMDMs had significantly higher expression of CD206 compared to both M1 and M0, and the M1 macrophages had a decreased level, though not statistically different, compared to unpolarized BMDMs. These results are consistent with the establishment of a CD206_low population of M1s and a CD206_high population of M2s. All three groups of BMDMs were treated with polyplexes as performed with ThP-1 cells. Unpolarized (M0) and CD206_high (M2) BMDMs demonstrated no significant changes in viability with all treatments remaining above 90% viable compared to an untreated control (Figure 3B). Interestingly, the viability of CD206_low M1 macrophages, but not M0 or M2 macrophages, was significantly reduced following exposure to polyplexes prepared with a 0.25–1 mM copper catalyst concentration (Figure 3B). This broad-ranging effect suggests a mechanism independent of the residual copper content. Classically activated macrophages initially respond to infection by producing inflammatory cytokines, chemokines, and reactive oxygen species but the resolution of the inflammatory reaction has been reported to involve mitochondrial-dependent macrophage cell death.^53,54 Therefore, we hypothesize that the loss of viability observed in the M1-polarized macrophages is induced by their additional activation in response to particle endocytosis. The lack of toxicity in M2 macrophages is encouraging for our applications. It is also important to note that although the same copper-associated toxicity found in ThP-1 macrophages was not observed in BMDMs, there is still a concern for the toxic effects of copper in human macrophages for potential future applications. Additionally, we show that the polymer alone (PEGDB control) does not induce a significant change in macrophage viability. This diblock copolymer has been previously shown to be nontoxic to cells, and we confirm those results here.^26,27,37

All viability studies were repeated in human and murine triple-negative breast cancer (TNBC) cell lines and a nontumorigenic human mammary epithelial cell line. MDA-MB-231 cells were chosen since they are well-established models with well-characterized properties and display a basal-like triple-negative phenotype.⁵⁵ E0771 cells were chosen as the murine equivalent since they are also basal-like triple-negative breast cancer cells, but are less aggressive than the often-used 4T1 model and therefore a better model for our applications.⁵⁶ MCF10a cells provide a nontumorigenic, mammary epithelial control, but a murine analog could not be identified among many commercial sources; so, only human mammary cells were included in this study. Treatment with polyplexes for 24 h generated minimal toxicity in both TNBC cell lines. MDA-MB-231 cells treated with the 0.25, 0.5, and 0.75 mM copper-catalyzed polymers maintained viability of over 90% but were still statistically decreased compared to the untreated control (Supporting Information Figure S6A). The E0771 cells exhibited a significant decrease in viability in all polyplex treatments, but none of the treated groups were different from each other (Supporting Information Figure S6B). Similarly, the MCF10a human mammary epithelial cells had baseline toxicity from the polyplex treatments compared to the untreated control (Supporting Information Figure S6C). Interestingly, when compared to the PEGDB control polymer, MCF10a cells only had significantly decreased viability when treated with polymers made with 0.1 and 1 mM copper catalysts. These results are consistent with the ThP-1 human macrophages as these groups are presumed to have the highest number of azides (0.1 mM) and highest residual copper content (1 mM). In both E0771 and MCF10a cells, there does appear to be a slight decrease in viability in all polymer groups, indicating a potential cytotoxic effect from just the polymer system. The lack of toxicity in macrophages is encouraging for our applications, but we will take the off-target toxicity into consideration as we continue forward with our experiments. Overall, the range of a 0.25–0.75 mM copper catalyst appears optimal for fabricating polyplexes that have minimal toxicity to both macrophages and mammary cells.

Effects of Copper Salt on Cell Viability

To assess the contributions of copper(II) alone to cell viability, each cell type was treated with a range of known CuCl₂ concentrations. In general, copper is known to be toxic to cells, but only above a threshold concentration.^10,57 This study allowed us to characterize the lower limits of copper cytotoxicity and inform the selection of CuAAC reaction conditions to enable the use of the click reactions for biological applications. In human immortalized macrophages, as well as primary BMDMs, we observed a trend of decreasing viability as Cu²⁺ concentration increased, but there was no significant change in viability up to 0.009 mg/mL, and only the ThP-1 cells had a significant decrease in viability at 0.01 mg/mL (Figure 3C). These findings are important for our applications since the highest amount of residual copper associated with the 1 mM MnPEGDB was approximately 6-fold lower at 0.0015 mg/mL, as determined with the copper assay. Although these results indicate that our CuAAC products are well below the cutoff for copper-induced cytotoxicity, our polymer system, in particular, appears to cause cell death at an elevated level when the residual copper concentration is above 0.0006 mg/mL. This relationship between residual copper content and cell viability demonstrates that functionalization via CuAAC can produce biocompatible biomaterials by altering the amount of copper catalyst used.

Copper salt toxicity was also examined in TNBC and epithelial cell lines. MDA-MB-231 human TNBC cells and MCF10a human epithelial cells showed no significant change in viability at any Cu²⁺ concentration, although both did exhibit a trend in decreasing viability with increasing copper concentration. E0771 murine TNBC cells produced a trend of decreasing viability, though not statistically significant until the maximum dose (Supporting Information Figure S6D). Notably, the human cell lines (MDA-MB-231 and MCF10a) were not as susceptible to copper-associated toxicity at 0.1 mg/mL compared to all other examined cells. This result indicates that the concern for copper-associated toxicity in the human breast tumor microenvironment is minimal, providing support for the use of CuAAC in potential human therapeutics. Based on these results, the residual Cu²⁺ alone is not a predictor of cytotoxicity. Rather, the context of polymer, cargo, cell type, and the interaction with residual copper and azide control overall cytotoxicity.

Targeting Efficacy of Mannose-Functionalized Micelles

To evaluate the ability of mannose conjugation to increase uptake in CD206_high cells, polyplexes were fabricated with Cy5-dsDNA and treated with cells at an optimized time point of 2 h. Cell-associated fluorescence intensity was measured as an index of polyplex uptake, and results were normalized to the PEGDB control polymer to demonstrate mannose-specific uptake of the decorated polyplexes. In ThP-1 human macrophages, the MnPEGDB produced with a 0.75 mM copper catalyst significantly increased uptake efficiency compared to all other polymer groups (Figure 4A). The polymers catalyzed with 0.25 and 0.5 mM copper show slight, but nonsignificant, increases in uptake. Unchanged uptake in the 0.1 and 1 mM groups is presumed to be due to increased toxicity associated with those polymers. Our measure of cytotoxicity was an ATP reporter, and since endocytosis is an ATP-dependent process, early disruption in ATP production may significantly suppress subsequent endocytosis of polyplexes. Polarized BMDMs were then used to examine the effects of CD206 expression on polyplex uptake. There was no significant difference in uptake regardless of mannose decoration for M0 or M1 BMDMs. In the M2-polarized macrophages, however, the polymer produced with a 0.75 mM copper catalyst led to significantly increased uptake compared to the 0.1 and 1 mM groups, but not the 0.25 and 0.5 mM catalyst groups (Figure 4B). These results are consistent with our observations in human macrophages (Figure 4A), as well as the FTIR mannose conjugation results (Figure 1B), which indicated that CuAAC efficiency was not changed when increasing beyond a 0.25 mM copper catalyst concentration. Even more importantly, the 0.75 mM group was the only micelle treatment that produced significantly increased uptake in CD206_high macrophages compared to both M0 and CD206_low, M1 macrophages. To verify that uptake was due to CD206-specific uptake, we repeated uptake experiment in M2-polarized BMDMs using the optimized polymer groups (0.25–0.75 mM) and cotreated with free mannose sugar to block the CD206 receptor. By averaging across the three optimized groups, we determined that mannosylated polyplex binding was decreased by almost 60% (Supporting Information Figure S7). These results indicate that polyplex uptake is CD206-dependent. Based on these results from both human and murine macrophages, using a copper catalyst concentration of 0.75 mM for the CuAAC led to significantly increased macrophage uptake and was shown to increase uptake in CD206_high macrophages.

Figure 4.

Polyplex uptake evaluated in human and murine macrophages. (A) ThP-1 human macrophages displayed preferential uptake of polyplexes formed with the 0.75 mM copper catalyst. (B) 0.75 mM catalyzed polymer targeted CD206_high BMDMs over CD206_low, as well as unpolarized macrophages (^#p < 0.05), indicating mannose specificity. The 0.75 mM group also outperformed the PEGDB control, 0.1 mM, and 1 mM groups in just M2-polarized BMDMs (**p < 0.01).

When examining uptake in mammary epithelial cells, both human cell lines (MDA-MB-231 and MCF10a) revealed no preferential uptake among the polymer groups (Supporting Information Figure S8). The E0771 murine TNBC cells demonstrated a significant increase in uptake for the 0.75 mM group compared to that for the 1 mM group, but this change was not statistically different compared to all other polymer groups. Overall, our results indicate an optimal concentration of a 0.75 mM copper catalyst for minimizing copper-induced toxicity, as well as promoting mannose-associated targeting to CD206-expressing macrophages.

Evaluation of siRNA-Induced Macrophage Repolarization

We used the optimal MnPEGDB polymer prepared with a 0.75 mM copper catalyst concentration to treat M1- and M2-polarized BMDMs with either scrambled or IκBα siRNA. We treated the macrophages with siRNA-loaded polyplexes for 24 h before collecting the cells and isolating RNAs for use in reverse transcription polymerase chain reaction (RT-PCR). To examine a shift in phenotype, we examined mRNA expression of six genes: three M1 genes (CD86, TNF-α, iNOS), one M2 gene (CD206), and one general macrophage marker (F4/80).^24,58,59 The final mRNA we evaluated was IκBα, which we used to assess the ability of our siRNA sequence to knock down the target gene. Relative gene expression was normalized to an untreated control group of the corresponding polarization.

Treatment with IκBα siRNA resulted in a significant increase in CD86 and TNF-α compared to that with untreated control and scrambled siRNA treatments in both M1 and M2 BMDMs, and we observed a significant increase in iNOS expression, but only in the M2-polarized macrophages (Figure 5A–C). Furthermore, CD206 significantly decreased after IκBα siRNA treatment compared to the untreated control in both phenotypes, but, interestingly, there was no significant change in CD206 expression between the scrambled siRNA control and the IκBα siRNA in M2 BMDMs. Most importantly, treatment with the therapeutic siRNA significantly decreased IκBα expression in both phenotypes, which is necessary for activating the classical pathway of NF-κB (Figure 5D,E). Taken together, these results indicate that not only is the IκBα siRNA pushing M2 macrophages toward an M1 phenotype but the treatment does not significantly impact overall M1 macrophage gene expression. Finally, the expression of F4/80, which is known to be associated with macrophage activation,^60,61 increases in M2 BMDMs treated with IκBα siRNA (Figure 5F). This result indicates that not only are the alternatively activated macrophages being skewed toward an M1 phenotype but they are also in a more activated state. The delivery of IκBα siRNA with our optimized polyplex system successfully skews M2 macrophages back to a classical phenotype, which is necessary for eventual application in inducing an inflammatory macrophage response in tumors.

Figure 5.

RT-PCR was used to examine relative mRNA expression after treatment with scrambled or IκBα siRNA. Expression was normalized to an untreated control group of the corresponding polarization (M1 vs M2). To assess a change in macrophage phenotype, we examined the expression of the following: (A) CD86, a surface marker for inflammatory immune cells, (B) TNF-α, a proinflammatory cytokine, (C) iNOS, an IFN-γ induced isoform of nitric oxide synthase, (D) CD206, the macrophage mannose receptor overexpressed on M2 macrophages, (E) IκBα, an inhibitor of the classical NF-κB pathway, and (F) F4/80, a general macrophage marker.

To confirm macrophage repolarization after the delivery of IκBα siRNA, we also evaluated protein expression. BMDMs expressing an NF-κB-dependent green fluorescent protein (GFP)/luciferase reporter (BMDM-NGL) were polarized to an M2 phenotype and then treated with either phosphate-buffered saline (PBS) (negative control), M1 cytokines (IFN-γ and LPS, positive control), or mannosylated polyplexes loaded with scrambled or IκBα siRNA. After 24 h of treatment, the cells were harvested in a reporter lysis buffer, which allowed for direct quantification of luminescence. The IκBα siRNA demonstrated elevated NF-κB activity as indicated by an increase in luminescence. This activation was on the level of M1 polarization, which provides evidence for successful phenotype repolarization (Figure 6A). In addition, the same treatments were repeated, and the cells were collected for Western blot to directly examine protein expression levels. By delivering the IκBα siRNA in our optimized MnPEGDB polymer, we demonstrate a significant knockdown in IκBα protein expression, as well as a significant decrease in arginase-1 (Arg-1) expression, which is a protein overexpressed in M2 macrophages (Figure 6B,C). In conjunction with the results of RT-PCR, these results indicate that our polyplex system loaded with IκBα siRNA is able to repolarize macrophages from an M2 toward an M1 phenotype.

Figure 6.

BMDM-NGL cells were polarized toward M2, followed by treatment with PBS, IFN-γ and LPS, or mannose-decorated polyplexes containing scrambled or IκBα-targeting siRNA. Effects on (A) luciferase activity of the NF-κB-dependent reporter and (B) protein levels of IκBα, and the M2 macrophage marker, arginase-1 (Arg-1) were evaluated. (C) Densitometry analysis of the expression of IκBα and Arg-1 relative to β-actin loading control.

Conclusions

We have optimized the reaction conditions for functionalizing polymeric micelles with a targeting moiety using CuAAC. By altering the concentration of a copper catalyst, we demonstrated an optimal range for efficient conjugation while minimizing residual copper and azide contents. Our results indicated that cytotoxicity is a balance between excess Cu²⁺ and residual azides. Our observation of azide toxicity due to incomplete CuAAC reaction is an underappreciated and avoidable potential cause of cell injury. Moreover, we showed that copper-associated toxicity can be eliminated from our polymer system and that residual copper alone is not the main predictor of cytotoxicity. The toxicity observed in our polymer system appears to be synergistically associated with residual copper and the material system as a whole, both of which may exhibit exacerbated toxicity due to the intracellular delivery of the polyplexes. To our knowledge, this is the first study to examine the cytotoxic effects of copper salt alone compared to those of copper ions associated with our polyplex system. These results indicate that concerns for negative biological impacts of CuAAC reaction products can be mitigated by altering the reaction conditions. Overall, we show that CuAAC should not be discouraged for in vivo applications, but we recognize that some optimization of copper catalyst concentration and confirmation of low toxicity for each formulation/system may be necessary.

Mannosylated micelles prepared from our simplified polymer system enable increased delivery to human macrophages, as well as CD206_high primary murine macrophages, which is an improvement over the previous PEGDB polymer that can deliver siRNA, but does not confer cell-specific delivery. This optimized polymer system also successfully delivered therapeutic siRNA and induced a shift in gene expression indicative of macrophage “repolarization” from an M2 phenotype toward an inflammatory M1 phenotype. We confirmed this repolarization by demonstrating a shift in mRNA expression for several M1/M2 markers, as well as increased activation of NF-κB protein, indicative of classical macrophage activation. This improved polyplex formulation will be used in future studies to repolarize TAMs in cocultures with tumor cells to quantify the cancer-killing properties of the macrophages.

Materials and Methods

Materials

All materials were purchased from Sigma-Aldrich unless otherwise noted. Inhibitors were removed from dimethylaminoethyl methacrylate (DMAEMA) and butyl methacrylate (BMA) using an activated basic aluminum oxide column.^25,26 All DNA oligonucleotides were purchased from Integrated DNA Technologies (Coralville, IA). dsDNA was designed to be length-matched to the therapeutic IκBα siRNA and functionalized with a cyanine-5 (Cy5) fluorophore on the 5′ end of the antisense strand. IκBα siRNA was designed based on previous studies done in our lab.²⁴

Polymer Synthesis

Mannose-alkyne was fabricated as previously described.²³ All diblock copolymers were fabricated using 4-cyano-4-(ethylsulfanylthiocarbonyl)-sulfanylpentanoic acid (ECT) as a chain-transfer agent (CTA) conjugated to either azide-PEG (AzPEG) or PEG. ECT was synthesized as previously described.^62,63 The AzPEG and PEG macro-CTAs were then RAFT-polymerized with DMAEMA and BMA at a 50:50 molar ratio as previously described.²³⁻²⁵ AzPEGDB was then conjugated with mannose-alkyne via CuAAC chemistry to produce MnPEGDB. All polymers were characterized using ¹H nuclear magnetic resonance (¹H NMR) spectroscopy (Bruker, 400 MHz), Fourier transform infrared spectroscopy (FTIR, Bruker Tensor 27), and a copper assay kit (Sigma-Aldrich). All NMR spectra are shown in Supporting Information Figures S2–S4. Details for the polymer synthesis appear in Supporting Information Materials and Methods.

Polyplex Formation

All polymeric complexes (polyplexes) were formed as previously described.²³⁻²⁶ Initially, polymers were complexed with Cy5-labeled dsDNA, scrambled siRNA, or IκBα siRNA for 30 min in a 10 mM citrate buffer (pH = 4). The solution was restored to pH = 7.4 by adding a 10 mM phosphate buffer (pH = 8) at 5× volume of the pH 4 solution. Polyplex N⁺/P^– ratio was determined by the mole ratio of protonated amines in DMAEMA polymer (assuming 50% protonation at physiological pH) to the number of phosphates on dsDNA/siRNA.^25,26 All polyplex treatments were performed at a dose of 50 nM dsDNA/siRNA with N⁺/P^– 10:1. Particle size and ζ-potential were characterized using a Malvern Zetasizer located in the Vanderbilt Institute of Nanoscale Science and Engineering (VINSE) core facility.

Cell Culture

ThP-1

Immortalized human macrophage cells (ThP-1) were chosen to evaluate toxicity in a human equivalent to inform potential future translatability of our polymer system. ThP-1 human monocytes were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher Scientific; 11879-020) supplemented with 10% fetal bovine serum (FBS), 25 mM HEPES, 1% penicillin–streptomycin (P/S), 1% minimum essential medium vitamins, and 4.4 μL β-mercaptoethanol at 37 °C in a 5% CO₂ humidified atmosphere. To differentiate monocytes into mature macrophages, ThP-1 cells were plated in the aforementioned media supplemented with 0.1% (v/v) phorbol 12-myristate 13-acetate (Thermo Fisher Scientific) and incubated for 4 days to allow for differentiation into mature macrophages.^64,65 These cells were primarily chosen as a human line to use for viability studies, but the expression of a mannose receptor also allows us to examine targeted uptake.⁶⁶ Cells were plated in 96-well plates at 1 × 10⁵ cells/well in 100 μL media. Polyplex or copper salt treatments were added on day 4.

L929

L929 murine fibroblasts were used to produce supplemental media for culturing bone-marrow-derived macrophages (BMDMs) taken from mice. L929 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Corning, Inc., Corning, NY; 15-013-CV) supplemented with 10% FBS, 1% l-glutamine, and 1% P/S. L929 cells were grown to confluency in T-175 flasks, and 55 mL fresh media was added. On day 7, the media was collected, labeled L929 Week 1 media, and stored at −20 °C. A fresh 55 mL media was added, cultured for an additional 7 days, collected, and labeled L929 Week 2 media before being stored at −20 °C.

BMDM

All animal work was approved by the Vanderbilt University Institutional Animal Care and Use Committee. Healthy female FVB mice were sacrificed at 4–8 weeks of age, and the femurs and tibias were extracted. The bone marrow was flushed out with DMEM using a 5 mL syringe and collected in DMEM (Corning; 15-018-CV). The cell suspension was centrifuged (Thermo Scientific, Sorvall ST 8 Centrifuge) at 1000g for 5 min. Media was aspirated before resuspending the cell pellet in a 2 mL ACK (ammonium, chloride, potassium) lysing buffer (KD Medical, Columbia, MD) and incubating on ice for 2 min to lyse red blood cells. The lysis solution was diluted in 20 mL DMEM and again centrifuged at 1000g for 5 min. The media was aspirated, and the resulting BMDMs were resuspended in a 10 mL BMDM media: DMEM (15-018-CV) with 10% FBS, 1% P/S, 1% l-glutamine, and 14% 1:1 (v/v) L929 week 1 and week 2 media. The cells were counted by mixing a 10 μL cell suspension with 10 μL Trypan Blue stain (Thermo Fisher Scientific) and pipetting 10 μL of the resulting mixture into a cell counter slide (Bio-Rad) and running on an automated cell counter (Bio-Rad, TC20). BMDMs were seeded in 12-well plates at 1 × 10⁶ cells/well in 1 mL media or in 96-well plates at 1 × 10⁵ cells/well in 100 μL media. To induce M1 and M2 polarizations, BMDMs were incubated with M1- or M2-inducing cytokines.⁵² Briefly, all cells were plated on day 0. On day 2, cells were washed with sterile phosphate-buffered saline (PBS) and then cultured with fresh media. On day 4, fresh media was added to M0 and M1 BMDMs, while M2 macrophages were treated with media supplemented with 0.01 μg/mL IL-4 and 0.02 μg/mL IL-13. On day 7, M0 and M2 macrophages were not changed, while M1 BMDMs received media supplemented with 0.1 μg/mL IFN-γ and 0.1 ng/mL LPS. On day 8, cells were treated for the appropriate experiment.

E0771

Murine breast cancer cells were used as a comparison to the BMDMs described previously. E0771 cells were cultured in RPMI 1640 (Thermo Fisher Scientific; 11875-093) supplemented with 10% FBS, 1% P/S, and 25 mM HEPES.

MDA-MB-231

Human breast cancer cells were used as a comparison to human ThP-1 macrophages. MDA-MB-231 cells were cultured in DMEM (Thermo Fisher Scientific; 11960-044) supplemented with 10% FBS, 1% l-glutamine, and 1% P/S.

MCF10a

Human epithelial cells were used as a healthy tissue control for all treatments. MCF10a cells were cultured in bronchial epithelial cell growth medium (BEBM) supplemented with a BEBM Bulletkit (Lonza, Morristown, NJ).

Flow Cytometry

BMDMs were polarized, washed with 0.5 mL PBS, and then incubated with 0.5 mL of 0.25% trypsin–EDTA (Thermo Fisher Scientific) for 5 min. Media of 1 mL was added to each well, and cells were repeatedly aspirated with a disposable pipette to dislodge them from the surface and collected in 15 mL conical tubes. Tubes were centrifuged at 1500 rpm for 5 min, and the supernatant was aspirated. The cell pellet was resuspended in 2 mL of fresh BMDM media, and the cells were counted as previously described. Cells were then placed in a 96-well round-bottom weight flask at 1 × 10⁶ cells/well in 300 μL/well. The plate was spun down at 1500 rpm for 5 min and inverted and lightly tapped to remove the supernatant without losing cell pellets. An Fc block consisting of 1 μL Fc block (Biolegend, San Diego, CA) and 50 μL of flow cytometry (FACS) staining buffer (PBS with 2% FBS) was added to each well. Plates were stored at 4 °C for 10 min. A macrophage panel of antibodies consisting of the following was added: CD11b (1:400), F4/80 (1:200), CD86 (1:200), and CD206 (1:200) (Invitrogen, Carlsbad, CA). Each antibody was added in 50 μL of FAC buffer and so the total volume when calculating the concentrations was in 100 μL per sample. Plates were stored at 4 °C in the dark to allow for staining. Plates were spun down at 1500 rpm for 5 min and inverted and tapped to remove the supernatant. Each well was resuspended in a 200 μL FACS buffer and analyzed by flow cytometry.

Viability Assays

Polyplex Toxicity

Each cell type was cultured in 96-well plates for viability assays. BMDMs and ThP-1s were plated at 1 × 10⁵ cells/well, while all other cells were plated at 25 000 cells/well. BMDMs and ThP-1s followed the plating protocols listed above. The other cell types were plated, incubated overnight to allow cells to adhere, and treated with polyplexes. All cells were treated with 50 nM of Cy5-dsDNA loaded into the various polymer formulations. The cells were incubated for 24 h before conducting a CellTiter-Glo Luminescence Assay (Promega, Madison, WI). All luminescent results were normalized to the average of the control well luminescence.

Copper Salt Toxicity

All cell types were plated as performed to assess polyplex toxicity. CuCl₂ was dissolved at 1 mg/mL in 10% (v/v) 200-proof ethanol and media (specific for cell type). This solution was then diluted to a range of 50–10 000 μg/dL, and 100 μL was added to each well. All cells were incubated for 24 h before running the CellTiter-Glo luminescence assay.

Polyplex Uptake

Cells were plated in 96-well plates and treated with 50 nM of Cy5-dsDNA-loaded polyplexes. The cells were incubated with the polyplexes for 2 h and then washed 3× with 100 μL PBS. A final volume of 100 μL PBS was added, and the fluorescence intensity was measured (Tecan Infinite M1000 Pro). All fluorescence results were normalized to the fluorescence of the PEGDB control polyplexes to determine mannose-associated uptake. To examine CD206-specific uptake of mannosylated polyplexes, we also performed a receptor-blocking experiment where polyplexes were added to media containing 100 mg/mL molecular-grade D-mannose, as done previously.²³ The free mannose binds CD206 and prevents receptor-mediated uptake of fluorescent polyplexes. To evaluate the decrease in uptake, the optimized copper catalyst concentration (0.25–0.75 mM) groups were averaged together since they demonstrated similar azide reduction.

Quantitative Reverse Transcriptase Polymerase Chain Reaction (RT-PCR)

BMDMs were plated in 12-well plates as described previously. After polarization to M1 and M2, the MnPEGDB fabricated with a 0.75 mM copper catalyst was used to form polyplexes with either scrambled siRNA or IκBα siRNA. BMDMs were treated for 24 h before mRNA was isolated from the cultured cells using the RNeasy Mini kit (Qiagen, Valencia, CA). Residual DNA was removed using the RNase-Free DNase set (Qiagen). cDNA synthesis was performed using SuperScript IV reverse transcriptase kit (Invitrogen). The RT-PCR reaction was performed using an SYBR Green PCR super mix (Bio-Rad) and CFX real-time PCR instrument and software (Bio-Rad). Normalized levels of mRNA expression were calculated using the ΔΔCt method with GAPDH as an internal control. Each mRNA was normalized to an untreated control group of the corresponding polarization (M1 vs M2). All primer sequences used are summarized in Supporting Information Table S1.

Luciferase Assays

Luciferase activity was measured in immortalized bone-marrow-derived macrophages derived from transgenic mice, which carry an NF-κB-dependent GFP/luciferase reporter (BMDM-NGL) on the FVB strain background. BMDM-NGL cells were polarized toward M2 by IL-4 treatment for 24 h, followed by a further 24 h treatment with PBS (control), IFN-γ and LPS (M1-polarized control), or mannosylated polyplexes loaded with either scrambled siRNA or IκBα-targeting siRNA. Cells were harvested in reporter lysis buffer and luminescence measured by the Promega Luciferase Assay system (Cat #4030) using a GloMax Luminometer (Promega, Madison, WI). Results were expressed as relative light units normalized for protein content, as measured by the Bradford assay (Bio-Rad, Cat #500-0002).

Western Blotting

BMDM-NGL cells were polarized toward M2 by IL-4 treatment for 24 h, followed by a further 24 h treatment with PBS (control), IFN-γ and LPS (M1-polarized control), or mannosylated polyplexes loaded with either scrambled siRNA or IκBα-targeting siRNA. Whole cell protein isolation, Western blotting, and signal detection were performed as described (PMID: 26215403). Primary antibodies used were mouse monoclonal anti-IκBα (Cell Signaling Technology; Cat #4814; 1:1000 dilution), rabbit polyclonal antiarginase-1 (Gene Tex; Cat #GTX109242; 1:200 dilution), and mouse monoclonal anti-β-actin (Sigma Chemical Co., Cat #A5441 1:10 000 dilution) as loading control.

Statistical Analysis

All data are presented as mean ± standard error of the mean (SEM) except RT-PCR, which is shown as mean ± upper/lower limit. For all FTIR, copper assay, and ThP-1 polyplex toxicity studies, a one-way analysis of variance (ANOVA) with Tukey’s post hoc test was used to compare all groups to all other groups. For copper salt viability assays, a two-way ANOVA with Dunnett’s post hoc test was used to compare all concentrations to the control within each cell type. For the BMDM polyplex toxicity studies and all polyplex uptake studies, a two-way ANOVA with Tukey’s post hoc test was used to compare all groups to all other groups. For all RT-PCR results, ΔΔCt values were compared using a one-ANOVA with Tukey’s post hoc test. Luciferase assay and Western blotting signal detection were analyzed using Student’s t test. To establish statistical significance, we used the following: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsomega.9b01465.

• Experimental methods for polymer synthesis; preliminary cell toxicity data; NMR characterization of each polymer fabrication step; RT-PCR primer sequences; polyplex diameter and ζ-potential data; flow cytometry characterization of BMDM polarization; toxicity of polyplexes in nonmacrophage cells; mannose-blocking of CD206-specific polyplex uptake; and uptake of polyplexes in nonmacrophage cells (PDF)

Author Contributions

^⊥ F.E.Y. and T.D.G. contributed equally to this work.

The authors declare no competing financial interest.