Uptake and function of membrane‐destabilizing cationic nanogels for intracellular drug delivery

Abstract The design of intracellular drug delivery vehicles demands an in‐depth understanding of their internalization and function upon entering the cell to tailor the physicochemical characteristics of these platforms and achieve efficacious treatments. Polymeric cationic systems have been broadly accepted to be membrane disruptive thus being beneficial for drug delivery inside the cell. However, if excessive destabilization takes place, it can lead to adverse effects. One of the strategies used to modulate the cationic charge is the incorporation of hydrophobic moieties, thus increasing the hydrophobic content. We have demonstrated the successful synthesis of nanogels based on diethylaminoethyl methacrylate and poly(ethylene glycol) methyl ether methacrylate. Addition of the hydrophobic monomers tert‐butyl methacrylate or 2‐(tert‐butylamino)ethyl methacrylate shows improved polymer hydrophobicity and modulation of the critical swelling pH. Here, we evaluate the cytocompatibility, uptake, and function of these membrane‐destabilizing cationic methacrylated nanogels using in vitro models. The obtained results suggest that the incorporation of hydrophobic monomers decreases the cytotoxicity of the nanogels to epithelial colorectal adenocarcinoma cells. Furthermore, analysis of the internalization pathways of these vehicles using inhibitors and imaging flow cytometry showed a significant decrease in uptake when macropinocytosis/phagocytosis inhibitors were present. The membrane‐disruptive abilities of the cationic polymeric nanogels were confirmed using three different models. They demonstrated to cause hemolysis in sheep erythrocytes, lactate dehydrogenase leakage from a model cell line, and disrupt giant unilamellar vesicles. These findings provide new insights of the potential of polymeric nanoformulations for intracellular delivery.

understanding the mechanism of uptake is critical because the internalization pathway influences subcellular trafficking, sorting, and exposure to variable enzymatic and pH conditions. [10][11][12] Additionally, compositional considerations of the particle chemistry, such as the balance between cationic and nonionic, hydrophilic components, and ratio of hydrophobic monomers have significant impact on resultant drug delivery properties (i.e., transfection efficiency, complex stability, etc.). 13 These parameters must be carefully investigated and optimized in the development of polymer drug delivery systems.
It is generally understood that increasing cationic content leads to increased loading efficiency of negatively charged molecules. [14][15][16] Previous work on the interaction between poly(dimethylaminoethyl methacrylate; PDMAEMA) or poly(aminoethyl methacrylate; PAEMA) and negatively-charged molecules (i.e., DNA) showed that PAEMA interacts more strongly with DNA while PDMAEMA exhibited superior buffering capacity, 13 which could lead to increased endosomolytic activity. However, excess cationic content in polymeric delivery systems can have deleterious effects. High cationic charge density is frequently correlated with toxicity of conventional cationic polymers like poly(ethyleneimine) 17 and may host undesirable consequences in vivo. 18 The synthetic strategy developed by Peppas and coworkers [19][20][21] allowed the decrease of critical swelling pH to increase endosomolytic 22 and gene transfection 23  Previous studies demonstrated the pH-dependent aqueous solution behavior of P(DEAEMA-g-PEGMA; PDET) and P(DEAEMA-co-TBMA-g-PEGMA; PDETB) nanogels. 21 TBMA-modified networks exhibited a more tightly collapsed network at elevated pH, 19 which could theoretically provide improved protection of encapsulated payload. Moreover, a decrease in the pH required to induce a critical transition (as demonstrated by PDETB30) may minimize premature release of the cargo before the intended site of action. 21 The work presented herein analyzes the role of hydrophobicity in modulating the cytotoxicity, uptake, and membrane destabilization properties of polymeric cationic nanogels. The hydrophobic monomers TBMA and 2-(tert-butylamino)ethyl methacrylate (TBAEMA) were incorporated to the base formulation of PDET. The toxicity of the synthesized drug delivery vehicles was evaluated using a mammalian cell model. Mechanistic studies using imaging flow cytometry were then performed to identify the particle internalization pathways to elucidate the principal cellular mechanisms used for their uptake. 24 Finally, evaluation of the membrane disruption abilities of these drug delivery vehicles was performed using three model membrane systems.
Sheep erythrocytes were used to assess the pH-and concentrationdependent membrane destabilization of lipid bilayers, lactate dehydrogenase (LDH) leakage was measured from Caco-2 cells to evaluate the nonspecific membrane destabilization in live cells, and giant unilamellar vesicles (GUVs) were used to understand the mechanism of membrane disruption by cationic polymeric nanogels. 25 The insights obtained from these studies will help in the design of novel drug delivery platforms for intracellular delivery by understanding the role of charge and hydrophobicity in the uptake and function of cationic nanoparticle platforms.

| Polymer synthesis and purification
Polymer synthesis and purification proceeded as described previously. 21 kept at pH 8.5. The mixture was emulsified using a Misonix Ultrasonicator (Misonix, Inc., Newtown, CT). The emulsion was purged with nitrogen gas and exposed to a UV source for 2.5 hr with constant stirring. MyTAB, Brij 30, and unreacted monomers were then removed by repeatedly inducing polymer-ionomer collapse, separating particles by centrifugation, and resuspending in 0.5 N HCl. Polymer particles were dialyzed against ddH20 for 7 days with twice daily water changes. Following dialysis, polymeric particles were flash frozen in liquid N2 and lyophilized for 5 days. Transmission electron microscopy was used to determine the diameter of the dry nanogels and was conducted as previously described. 21

| Cytocompatibility studies
In vitro cytocompatibility of polycationic nanoscale hydrogels was evaluated using commercially available cytotoxicity assays. MTS assays were performed using the CellTiter 96 AQueous One Solu-

| Fluorescent polymer synthesis
PDETB30 was synthesized and purified as described above. To enable the covalent conjugation of a fluorescent probe, 2-aminoethyl methacrylate hydrochloride (AEMA) was included in the pre-polymer feed mixture at 5 mol% of DEAEMA. The resulting copolymer was named PDETB30f to signify the amine functionality. The primary amine of AEMA was verified with a fluorescamine assay after synthesis and purification. 19 Oregon Green 488 carboxylic acid, succinimidyl ester (OG488) was purchased from Molecular Probes (Eugene, OR). The solid dye was dissolved in DMSO to yield a 10 mg ml −1 solution. To form the fluorescent polymer conjugate, PDETB30f was suspended at 10 mg ml −1 in 150 mM sodium bicarbonate buffer, pH 8.30. OG488 was added to the PDETB30f suspension to give a 1:1 mol ratio between AEMA and OG488. The reaction was stirred in the dark for 6 hr. Following reaction completion, unreacted dye was separated from labeled PDETB30f through dialysis against DI water. Dialysis proceeded for 3 days using 12,000-14,000 MWCO dialysis tubing (Spectrum Labs, Rancho Dominguez, CA) for 3 days. Labeled nanogels, PDETB30-OG488, were lyophilized in the dark for 3 days.
Caco-2 cells were seeded at 1 x 10 5 cells/well in 6-well plates and allowed to grow to 80% confluence before performing the experiment. Immediately prior to exposure to inhibitors, cells were washed with 2 ml DPBS and media was replaced with 1.8 ml serum-free DMEM. Concentrated suspensions (20X) of inhibitors were added to wells in 100 μl increments and allowed to incubate with cells for 30 min in a 37 C, 5% CO 2 atmosphere. Cells inhibited by refrigeration were placed at 4 C for 30 min prior to nanogel exposure.
Following the 30 min equilibration period, 100 μl of PDETB30-OG488 at 500 μg ml −1 in PBS was added to test well to yield a final concentration of 25 μg ml −1 . Control wells received 100 μl PBS. Nanogel exposure occurred for 60 min at 37 C or 4 C.
Following the exposure period, cells were rinsed 3× DPBS (with calcium and magnesium) and the media was replaced with 2 ml serum-free DMEM. Hoechst 33342 was added to each well for nuclear staining at a final concentration of 2.5 μg ml −1 . The nuclear staining process was completed for 45 min at 37 C, 5% CO 2 . Following Hoechst incubation, cells were rinsed 3× with DPBS (without calcium and magnesium).
Caco-2 cells were isolated by replacing the final DPBS wash with 500 μl 0.25% trypsin-EDTA and incubating at 37 C, 5% CO 2 for 8 min. Trypsin was neutralized by adding 3 ml DMEM with 10% FBS and without phenol red. Cell suspensions were centrifuged for 5 min at 500g. The supernatant was discarded and cell pellet resuspended in 100 μl flow cytometry buffer. All cell suspensions were kept on ice until analysis with Image Stream Cytometry. Propidium iodide (PI) was used as a live/dead discriminator and was added to cell suspensions immediately before analysis at a final concentration of 1 μg ml −1 .

| Hemolysis
Sheep blood in sodium citrate was obtained from Hemostat Laboratories (Dixon, CA) and used for up to 2 weeks after receipt. Phosphate buffers (150 mM) from pH 5.0 to 8.0 were prepared. Dry nanogels were suspended in 150 mM phosphate buffer at the desired pH at a concentration of 2.5 mg ml −1 and allowed to equilibrate overnight.
Erythrocytes were isolated from whole sheep blood by three successive washes with freshly prepared 150 mM NaCl. Red blood cells (RBCs) were separated by centrifugation from 10 min at 2,000g. The supernatant was carefully aspirated and discarded. After removing the supernatant following the final wash, RBCs were suspended in a volume of 150 mM phosphate buffer identical to that of the original blood aliquot at the pH matching that of the suspended polymers.
This solution was diluted 10-fold in 150 mM phosphate buffer to yield an RBC suspension of approximately 5 x 10 8 cells/ml. In a typical experiment, 1 x 10 8 RBCs were exposed to nanogels at specified concentrations while shaking in a bead bath (LabArmor, Cornelius, OR) pre-equilibrated at 37 C. Following a 60 min incubation period, samples were centrifuged at 14,500 RPM for 5 min to separate cells and membrane fragments. An aliquot of each sample was transferred to a clear 96-well plate and hemoglobin absorbance was measured at 541 nm. Negative controls (0% lysis) consisted of 150 mM phosphate buffer at experimental pH and positive controls (100% lysis) consisted of RBCs incubated in ultrapure DI water.
The pH values tested in this analysis range from pH 5.0 to 8.0; experiments performed at pH 5.00, 5.50, 6.00, 6.50, 7.40, 7.60, 7.80, and 8.00. The concentrations tested range from 1 to 2,000 μg ml −1 ; with experiments performed with 2,000, 1,000, 500, 250, 100, 50, 25, 10, 5, 2.5, and 1 μg ml −1 nanogel suspended in 150 mM phosphate buffer at the specified pH.  GUVs were placed in 35 mm glass-bottom petri dishes for realtime confocal microscopy imaging. PDET and PDETB30 were prepared at 2 mg ml −1 in 100 mM phosphate buffer adjusted to pH 6.50. The osmolarity of the resulting suspensions was measured and adjusted with sucrose to~350 mOsm as needed. 1 ml of GUV suspension was transferred to the glass-bottom petri dish and was allowed to sediment for 5 min. About 25 μl of the nanogel suspension was carefully injected into the dish so as not to disturb the spatial distribution of focused GUVs. Images were collected every 5 s at a fixed focal plane.

| Statistical analysis
Statistical comparisons between experimental and control groups were made with two-tailed, unpaired, Student's t-test. Differences were accepted as statistically significant with p < .05.

| RESULTS AND DISCUSSION
Our previous work has demonstrated the ability to synthesize cationic polymeric nanogels based on a core of PDEAEMA using a photoemulsion polymerization. [19][20][21] Furthermore, we were able to tailor their hydrophobicity by the incorporation of the hydrophobic monomers TBMA and TBAEMA. 21,26 The physicochemical characteristics of the resulting particles displayed promising potential as drug delivery vehicles including their size, pH-responsiveness, and swelling volume. Based on these results, the most promising formulations were selected to evaluate their capabilities in vitro as membrane-destabilizing platforms for intracellular drug delivery. This work was focused on the assessment of nanogels prepared from copolymers with 0-30% mol of hydrophobic monomer (TBMA or TBAEMA).
Polycationic nanoscale hydrogels were successfully synthesized using a photoemulsion polymerization as previously described. 21 Copolymers with TBMA or TBAEMA (0, 10, 20, 30 mol%) were formulated to increase the core hydrophobicity. The polymer formulation FIGURE 1 Cytocompatibility of polycationic nanogels as a function of polymer concentration. Symbols represent PDET ( ), PDETB20 ( ), PDETB30 ( ), PDETBA20 ( ), or PDETBA30 ( ). Proliferation of Caco-2 cells was determined by MTS assay following 90 min nanogel exposure and is expressed as a fraction of the control (untreated) cells. Data are expressed as mean AE SEM, n = 8. Statistical significance determined via pairwise t-test between cells exposed to PDETB20 and PDET or PDETB30 and PDET (*p < .005) nomenclature includes the numerical suffix on the polymer name (e.g., PDETB30 or PDETBA20) which refers to the moles of hydrophobic monomer (TBMA or TBAEMA) per 100 mol of DEAEMA. The composition and morphology of the resulting polymeric particles were consistent with previously reported results. 21 Representative micrographs of the synthesized nanogels are presented in Supporting Information Figure S1.
Following the synthesis of cationic nanogels, the initial step in the evaluation of their biological properties was the assessment of their cytocompatibility. These materials will interact with a variety of cellular populations upon entering the body. Hence, a critical characteristic as drug delivery vehicles is their lack of cytotoxic effects.
The influence of polymer concentration and composition on cellular proliferation was assessed using MTS assays. These data are important to determine the nontoxic polymer doses for future drug delivery experiments. In this assay, the metabolic activity of an experimental population relative to control populations can be given by the ratio: where A s is the absorbance (λ = 490 nm) from sample wells, A bkg is the background absorbance from DMEM/MTS solution, and A PBS is the absorbance from wells in which cells were incubated only with DPBS.
As seen in Figure  can also enhance internalization of the drug carrier. 16,29 Previous studies have shown that surface hydrophobicity increases particle uptake in antigen presenting cells. 9,[30][31][32] This phenomenon has shown to be related to opsonization, as hydrophobicity has shown to increase the amount and variety of serum proteins adsorbed onto different particle platforms. 31 PDETB30f, was successfully synthesized through the inclusion of AEMA in the nanogel core. Following previous work, 34 solid AEMA was added to the pre-polymer mixture immediately before sonication.
While this monomer is stable in its hydrochloride salt form, it readily undergoes a cyclic rearrangement to 2-hydroxyethyl methacrylamide upon neutralization. 35 Prior to the conjugation reaction, the primary amine content of PDETB30f was determined to be 47.5 AE 0.6 μmol g −1 , which represents a 32% incorporation efficiency.
Oregon Green 488 (OG488), an amine reactive dye, was then successfully conjugated to the primary amines in the nanogel core.   In this study, several uptake inhibitors (Table 1) were applied to Caco-2 cells to elucidate the primary uptake pathways into enterocytes and phagocytes, respectively. As demonstrated previously, the membrane-disruptive activity of PDETB30 is highly dependent on environmental conditions (e.g., pH). Therefore, to exert their membrane-disruptive effect and enable cytoplasmic delivery of FIGURE 3 Representative fluorescent micrographs of Caco-2 cells exposed to endocytosis inhibitors and PDETB30-OG488. Images sampled from median intensity region of OG488 fluorescent histogram. Scale bar represents 7 μm encapsulate cargoes, these polybasic nanogels must be exposed to a   Wortmannin caused a 39% reduction in intracellular fluorescence and amiloride caused a 31% reduction.
Inhibition of energy-dependent processes by incubation at 4 C caused a 63% reduction in the intracellular fluorescence. Notably, an appreciable portion of PDETB30-OG488 uptake in Caco-2 cells occurs through an energy-independent process. Other reports have noted energy-independent transport of nanoparticles, specifically with respect to cationic lipids and breast cancer cells 42 and PLGA nanoparticles and Caco-2 cells. 43 This process is thought to be due to particle fusion with the cell membrane and has been reported in several types of cationic delivery vectors, including lipoplexes, 44 dendrimers, 45 and crosslinked poly(ethyleneimine) nanogels. 46 These results indicate that  Following the uptake analysis of our cationic nanogels that shows the internalization of these delivery vehicles, the evaluation of their membrane destabilizing capabilities was performed. This series of experiments was constructed to identify nanogels capable of selective membrane destabilization. An optimal nanogel would be relatively inert and nondisruptive under normal physiological conditions. Upon transition to endosomal conditions, this optimal nanogel would undergo a conformational transition to render it capable of potent membrane destabilization. Conversely, a nonoptimal nanogel would mediate membrane disruption under physiological conditions and/or be nondisruptive in endosomal conditions. This assessment was carried out using three different models: a hemolysis assay, LDH leakage, and GUV disruption.
First, hemolysis experiments were used to approximate the endosomolytic ability of these nanogels. The pH-and concentrationdependent hemolysis was determined according to the following equation: where A sample represents RBCs exposed to polymer at a given pH and concentration, A blank is the absorbance of the supernatant after RBC exposure to phosphate buffer at a given pH, and A max represents maximum lysis following RBC exposure to DI water. The relative lysis for nanoscale hydrogels containing varying amounts of TBMA or TBAEMA is shown in contour plot form in Figure 5, panel a. These data demonstrate that polymer composition has a clear impact on membrane-disruptive capabilities. As demonstrated previously with dynamic light scattering studies, 21 the presence of a t-butyl group alone in the copolymer is not the critical parameter for exerting control over resultant physicochemical properties. Rather, the increased network hydrophobicity of TBMA-containing nanogels seems to govern the interactions with biological membranes.
As seen in Figure 5, inclusion of TMBA in the nanogels markedly expands both the pH and concentration range at which these networks effectively disrupt erythrocyte membranes. Hemolysis of red blood cells occurs between 7.4 and 5.5, which corresponds to the pH from physiological to late endosomal conditions. The pH transition of the synthesized formulations (as shown in Figure 5, panel e) ranges from 6.78 (PDETB30) to 7.66 (PDETBA30). The optimal formulation for intracellular delivery would have a pH transition closer to endosomal pH levels, than physiological pH. For example, PDET demonstrates efficient hemolysis at high concentrations (>0.25 mg ml −1 ) and between pH 7.0 and pH 7.6. In contrast, PDETB30 demonstrates highly efficient hemolysis in the pH range of early endosomes (pH 5.5-6.5) at concentrations as low as 1 μg ml −1 . The enhanced hemolytic ability of PDETB30 at pH 6.0 is depicted in Figure 5, panel b, along with that of PDET and PDETBA30. Notably, PDETB30 is 10× more efficient (on a mass basis) than previously reported polycationic block copolymer systems with demonstrated efficacy in in vitro siRNA delivery 22 and 25× more efficient than phenylalanine-grafted pseudopeptides 47 with demonstrated utility in intracellular protein delivery. 48 These data indicate that the membrane-disruptive properties of these nanogels can be tuned by adjusting hydrophobic monomer incorporation, an observation in accordance with several previous studies. [49][50][51][52][53] To analyze the conformational transition of pH-responsive nanogels, the ratio of the first to third vibronic peak (I 1 /I 3 ) in the fluorescence emission spectra of pyrene was used as previously described. 21 In the fluorescence spectra of pyrene occurs a characteristic shift depending on the polarity of the pyrene microenvironment. If dissolved in a highly polar, aqueous solvent the I 1 /I 3 ratio in the emission spectra is approximately 1.59. This ratio decreases to 0.61 in nonpolar, aliphatic hydrocarbons such as n-hexane or dodecane. 54 Thus, a decrease in the emission I 1 /I 3 ratio denotes the preferential partition of pyrene into hydrophobic domains.
Therefore, the pH-responsive transition regime (from collapsed hydrophobe to swollen hydrophile) is a critical factor in determining the membrane-disruptive ability of these nanogels. In all cases, nanogels demonstrated maximum hemolysis at or near the pH app determined by pyrene fluorescence studies ( Figure 5, panels c and d). If this pH app is near physiological pH, this membrane-disruptive effect was  LDH leakage as a function of nanogel concentration and exposure time is shown in Figure 6 for The micrographs in Figure 8 suggest that transient nanopore formation is the predominant mechanism through which PDETB30 exerts a membrane-destabilizing effect. For these initial studies, pH 6.50 was selected to approximate the pH of an early endosomal environment. Based on the hemolysis studies presented in Figure 5, PDET should be nondisruptive and PDETB30 should be highlydisruptive at these conditions. Following an injection to bring the PDET to 50 μg ml −1 in the buffered GUV solution, no discernible change was detected in membrane integrity. The sucrose-Texas Red remains entrapped in the GUV for several minutes after injection, confirming the persistence of membrane integrity as shown in panels a and c.
In contrast, the micrographs in Figure

| CONCLUSIONS
Physicochemical properties of nanoscale hydrogel networks, including critical phase transition pH, membrane disruption, and cytocompatibility can be modulated by tuning polymer composition. The mechanisms of cellular internalization of fluorescent nanogels were studied using imaging flow cytometry in Caco-2 cells, which showed that despite the lack of any targeting moieties, these nanogels are readily taken up. After 60 min exposure, the intracellular PDETB30-OG488 fluorescence increased over 25× in treated cells. Additionally, this analysis also showed that macropinocytosis is the dominant mechanism of nanogel internalization in a model cell line. Membrane vesicles arising from clathrin-mediated endocytosis and macropinocytosis both undergo acidification. As PDETB30 requires a slightly acidic pH to exert its membrane-destabilizing effects, these internalization pathways are desirable for uptake and subsequent endosomal escape of PDETB30 and encapsulated therapeutics. Additionally, the breadth of the pH range for maximum membrane disruption is related to the pH range for hydrophobic-hydrophilic transition. In particular, PDETB30 is membrane-disruptive over a broader pH range than other nanogels that undergo a more rapid hydrophobic-hydrophilic phase transition (e.g., PDET and PDETBA30). For these reasons, we have shown that TBMA-containing nanogels exhibit favorable pH-responsive phase transition behavior for intracellular delivery and offer an excellent combination of cytocompatibility, hemolytic ability, and membrane-disruptive properties. These characteristics are crucial to their promise as intracellular drug delivery vehicles.