Endosome-Mimicking Nanogels for Targeted Drug Delivery

Drug delivery systems inspired by natural particulates hold great promise for targeted cancer therapy. Endosome formed by internalization of plasma membrane has massive of membrane proteins and receptors on the surface, which is able to specifically target to the homotypic cells. Herein, we describe a simple method to fabricate an endosome membrane-coated nanogel (EM-NG) from source cancer cells. Following intracellular uptake of methacrylated hyaluronic acid (m-HA) adsorbed SiO 2 /Fe 3 O 4 nanoparticles encapsulating crosslinker and photoiniator, EM-NG was readily prepared through in situ crosslinking initiated under UV irradiation inside endosome. The resulting endosome mimetic nanogels loaded with Doxorubicin (DOX) displayed enhanced internalization efficiency to the source cells through a specific homotypic affinity in vitro. However, when treated the non-source cells, the EM-NGs exhibited insignificant difference in therapeutic efficiency compared to bare HA nanogel with DOX. This study illustrates the potential of utilizing endosome membrane-mimicking formulation for targeted cancer therapy, and offers guideline for developing natural particulates-inspired drug delivery system.


Introduction
Drug delivery systems based on natural particulates have emerged as one of the most promising strategies for cancer therapy 1 . Taking advantages of physical morphologies and biological functions of natural particulates, these biomimetic drug delivery carriers offer several significant advantages such as selective targeting, prolonged circulation time, and low immunogenicity 2-6 . Among them, natural biological membrane-derived nanoparticles (NPs) have received extensive attention as a simple and viable manner due to their capability of mimicking the natural membrane properties. For example, red blood cell membranecoated PLGA nanoparticles have shown superior circulation half-life in vivo [7][8][9][10] . In addition, it has been validated that cancer cell membrane-coated nanoparticles with full array of tumor antigens can promote a tumor-specific immune response and target to the source cancer cells via an inherent homotypic binding interaction 11 . Our previously reported work demonstrated that the platelet membrane coated core-shell nanovehicle with overexpressed P-Selectin on the membrane could specifically bind to CD44 receptors on the surface of cancer cells 12 .
Endosome plays an important role in regulating fundamental processes in eukaryotic cells, such as nutrient uptake, signaling, immunity and adhesion 13 The massive amount of membrane is internalized into endosome by several endocytic pathways, and the membrane lipids and proteins can be recycled back to the plasma membrane in an efficient manner [13][14][15] . Although precise mechanisms of the recycling and fusion remain to be fully elucidated, several previous studies revealed endosome-to-plasma membrane transport is mediated with several proteins including Eps15p homology (EH) domains containing proteins 16,17 . The proteins from the sorting nexin protein family and others can also promote the fusion of endosome with the plasma membrane [18][19][20][21] .
In this study, we describe an endosome membrane-coated nanogel (denoted as EM-NG) which is easily extracted from the source cancer cells for targeting and specific delivery of small molecular drug. As shown in Fig. 1, the EM-NG has an inner core composed of hyaluronic acid (HA) nanogel containing SiO 2 /Fe 3 O 4 nanoparticles, and an endosome membrane based outer shell. HA is chosen since it is highly biocompatible and it can target to the hyaluronnan receptors CD44, which are overexpressed in a variety of tumor types [22][23][24][25][26] . In order to allow HA nanogel form inside endosome directly, we firstly use coreshell mesoporous silica NPs with Fe 3 O 4 nanocrystals as the core (SiO 2 /Fe 3 O 4 NPs) to encapsulate crosslinker and photoinitiator, and coat with methacrylated HA (m-HA) on the surface through electrostatic interaction. Following the incubation of resulting NPs with source cells, an in situ formed nanogel in endosome is obtained by the photo-polymerization upon UV light irradiation [27][28][29][30][31][32] . The endosome containing HA nanogel is readily collected via the magnetic extraction due to the entrapped magnetic Fe 3 O 4 nanocrystals. After loading with anticancer drug Doxorubicin (DOX), these EM-NGs can actively target to source cancer cells by taking the advantage of the specific interaction with their source cells, and subsequently internalize to release DOX.

Synthesis of Fe 3 O 4 nanocrystals
A solution of 1 M iron (III) chloride hexahydrate and 0.5 M iron (II) chloride tetrahydrate in 25 mL DI water was added dropwise to a 0.5 M NaOH solution in 250 mL of DI water at 40°C. The mixture was stirred for 1 h and the nanocrystals subsequently washed with DI water until pH neutral. The resulting nanocrystals were dialyzed against DI water for 3 days. Finally, the Fe 3 O 4 nanocrystals were stabilized with oleic acid and dispersed in chloroform with a concentration of 6.7 mg Fe/mL. The zeta-potential and size distribution were measured on the Zetasizer (Nano ZS; Malvern). The TEM images were obtained on a JEOL 2000FX TEM instrument.

Synthesis of SiO 2 /Fe 3 O 4 NPs functionalized with quaternary ammonium groups
1 mL of the Fe 3 O 4 nanocrystals in chloroform was poured into 10 mL of 0.55M aqueous cethyltrimethylammonium bromide (CTAB) solution and emulsified by sonication for 10 min. The resulting turbid brown solution was stirred and heated up to 65 °C for 10 min to evaporate the chloroform, resulting in a transparent black Fe 3 O 4 /CTAB solution. Then, the Fe 3 O 4 /CTAB solution was added to a mixture of 90 mL of water and 0.6 mL of 2N NaOH solution, and the mixture was heated up to 70 °C under stirring. After adding 1 mL of tetraethylorthosilicate (TEOS) and 6 mL of ethylacetate, the solution was stirred at 70 °C for 3 h. Thereafter, the obtained SiO 2 /Fe 3 O 4 NPs were washed 3 times with ethanol to remove the unreacted species and dispersed in 30 mL of ethanol. To extract CTAB, 80 µL of HCl was added to the solution and stirred for 3 h at 60 °C. After washing them with ethanol three times, the NPs were dispersed in 10 mL of chloroform, and 120 µL of 3trimethoxysilylpropyl-N,N,N-trimethylammonium chloride (TMAPS) 50% in methanol was added. The reaction solution was stirred at room temperature for 24 h. The resulting SiO 2 / Fe 3 O 4 NPs functionalized with quaternary ammonium groups were wash 3 times with ethanol and dried in vacuum.

Preparation of HA/SiO 2 /Fe 3 O 4 NPs
3 mg SiO 2 /Fe 3 O 4 NPs was mixed with 0.5 mg N,N'-methylenebisacrylamide (MBA) and 0.1 mg photoinitiator (Irgacure 2959) in PBS buffer for 24 h. Unloaded crosslinker and photoinitiator was removed by filtering using a centrifugal filter (100,000 Da molecular mass cutoff, Millipore). Then, 1 mg m-HA or rhodamine-HA derivative was added and stirred for another 4 h to obtain HA/SiO 2 /Fe 3 O 4 NPs.

Cell culture
HeLa and A549 cells were obtained from Tissue Culture Facility of UNC Lineberger Comprehensive Cancer Center and cultured in Dulbecco's Modified Eagle's Medium supplemented with 10% (v/v) fetal bovine serum (FBS), penicillin (100 U/mL) and streptomycin (100 µg/mL) in a 37°C incubator (Thermal Scientific) under 5% CO 2 and 90% humidity. The cells were regularly sub-cultured with trypsin-EDTA (0.25%, w/w) and cell density was determined with hemocytometer before each experiment.

Preparation of EM-NG and HA NG
HeLa cells (1×10 5 cells per well) were seeded in 6-well plates. The cells were allowed to culture for 24 h before exposure to the HA/SiO 2 /Fe 3 O 4 NPs dispersions. The NPs dispersions were prepared by diluting the concentrated NPs solution into the FBS free medium. The cells were incubated with NPs for 4 h, then the NPs containing medium was discarded. After washing the cells by PBS twice, the cells were harvested with trypsin and centrifuged at 1000 rpm for 4 min. Then, the cells were resuspended in 1 mL PBS solution and exposed to UV irradiation (wavelength: 365 nm) for 1 min to form solid HA nanogel by crosslinking polymerization. The cells were lysed with Pierce IP lysis buffer (Thermo Scientific). Then, EM-NGs were collected using a magnet from the cells lysate. In order to obtain the bare HA nanogel without endosome membrane, the EM-NG were further lysed with lysis buffer for a second time, and collected using a magnet.

Intracellular distribution of HA/SiO 2 /Fe 3 O 4 NPs
HeLa cells (1×10 5 cells per dish) were seeded in confocal dishes and cultured for 24 h. Then, the cells were incubated with rhodamine-HA/SiO 2 /Fe 3 O 4 NPs for 1h and 4 h. Afterward, the cells were washed with PBS twice and stained by LysoTracker green (50 nM) (Life Technologies) at 37°C for 30 min. Then, the cells were washed with PBS twice and stained with Hoechst 33342 (1µg/mL) for 10 min. After washing with PBS twice, the cells were immediately observed using CLSM (LCM 710, Zeiss).

Preparation of DOX-EM-NG
To prepare DOX-EM-NG, 0.03 mL of DOX solution (0.1 mg/mL) was added into 0.27 mL of EM-NG solution, and mixed at room temperature for 24 h. Then, the excess DOX was remove by filtering using a centrifugal filter (10,000 Da molecular mass cutoff, Millipore). The loading capacity (LC) of DOX-EM-NG were determined by measuring the amount of encapsulated DOX by analyzing fluorescence intensity of DOX at 590 nm with an excitation wavelength of 480 nm using microplate reader (Infinite M200 Pro, Tecan). LC was calculated as LC=encapsulated amount of DOX/total weight of DOX-EM-NGs.

Endosome membrane protein characterization
SDS-PAGE was used to separate the proteins contained in the endosome only, EM-NGs and NGs. Samples were diluted in protein loading buffer and incubated for 10 min at 100°C. 10 µL of sample was loaded into each well in a 10% polyacrylamide gel. After separation by SDS-PAGE, the proteins were transferred to a nitro cellulose membrane at 300 mA for 3 h using a wet transfer method. The membrane was incubated with a 3% BSA blocking solution in the PBS TWEEN solution (0.01%) for 30 min at room temperature, and then incubated with early endosome antigen 1 (EEA1) in blocking solution overnight at 4°C. Afterward, the membrane was washed 3 times with PBS TWEEN 0.01% and incubated with the anti-rabbit HRP in blocking solution for 1 h. The membrane was then washed with PBS TWEEN 0.01% for 3 times and visualized using a 1-StepTM TMB-Blotting solution (Pierce, USA).

In vitro DOX release study
The release profile of DOX was measured using dialysis method in PBS buffer. 400 µL of DOX-EM-NG solution was added into a dialysis tube (10,000 Da molecular mass cutoff) (Slide-A-Lyzer, Thermo Scientific) against 1 mL of PBS buffer solution with different pH (7.4 and 5.0). The dialysis tube was incubated at 37°C. At predetermined time intervals, the total buffer solution was withdrawn, followed by replacing with fresh buffer solution with the same pH. The amount of release DOX was measured through the same method mentioned above.

Determination of endocytosis pathways
HeLa cells (1×10 5 cells per well) were seeded in 6-well plates and cultured for 48 h. Afterwards, the cells were pre-incubated with different specific inhibitors for different endocytosis pathways, including chlorpromazine (CPZ, 10 µM) for the clathrin-mediated endocytosis, nystatin (NYS, 25 µg/mL) for the caveolin-mediated endocytosis inhibition, amiloride (AMI, 1 mM) for the macropinocytosis inhibition, and methyl-β-cyclodextrin (MCD, 3 mM) for the lipid raft inhibition. Then, the cells were incubated with DOX-EM-NG at DOX concentration of 1 µM in the presence of the inhibitors for another 2 h. After washing the cells by 4°C PBS twice, the fluorescence intensity of DOX in the cells were measured by the flow cytometry.

Evaluation of DOX-EM-NG and DOX-NG uptake of HeLa cells
HeLa cells (1×10 5 cells per well) were seeded in 6-well plates and cultured for 48 h. Afterwards, the cells was added with DOX-EM-NGs, DOX-NGs, and free DOX with same concentration, and incubated for 2 h. After washing the cells by 4 °C PBS twice, the fluorescence intensity of DOX in the cells were measured by the flow cytometry.

In vitro cytotoxicity
HeLa or A549 cells (6×10 3 cells per well) were seeded in the 96-well plates. After 24 h culture, the cells were exposed to EM-NGs, free DOX solution, DOX-EM-NGs and DOX-NGs with different concentrations of DOX in FBS free medium for 24 h, respectively. Then, 20 µL per well of MTT solution (5 mg/mL) was added and incubated for another 4 h. After removing the medium, 150 µL DMSO was added to each well. The absorbance was measured at a test wavelength of 570 nm with a reference wavelength of 630 nm by a microplate reader (Infinite M200 PRO, Tecan).

Statistical analysis
All results presented are Mean ± SEM. Statistical analysis was performed using Student's ttest or ANOVA test. With a p value < 0.05, the differences between experimental groups and control groups were considered statistically significant.

Results and discussion Preparation and characterization of HA/SiO 2 /Fe 3 O 4 NPs
Magnetic Fe 3 O 4 nanocrystals was synthesized by a traditional aqueous co-precipitation technique 33 . The obtained Fe 3 O 4 nanocrystals was characterized by dynamic light scattering (DLS) and transmission electron microscope (TEM). As shown in Fig. 2a and b, the Fe 3 O 4 nanocrystals were monodispersed with an average diameter of 22.5 nm. These nanocrystals were then coated with mesoporous silica using a sol-gel method 34 . DLS results revealed that the size of SiO 2 /Fe 3 O 4 NPs increased to 115.3 nm, and the wormhole-like mesopores with a size of about 2-3 nm were clearly observed in the silica shells in TEM image (Fig. 2c and d).
The zeta potential of silica nanoparticles was determined as -30.1±3.3 mV due to surface hydroxyl groups.

Preparation and characterization of EM-NGs
To achieve the endosome membrane coating HA nanogels, HA/SiO 2 /Fe 3 O 4 NPs were first required to interact with endosomes. Human cervical carcinoma epithelial (HeLa) cell was chosen as a model cell line, and HA/SiO 2 /Fe 3 O 4 NPs were incubated in the cell culture medium with cells. HA/SiO 2 /Fe 3 O 4 NPs was allowed to internalize the endosome via the endo-lysosomal pathway. The successful uptake of HA/SiO 2 /Fe 3 O 4 NPs in endosome was confirmed using the confocal laser scanning microscopy (CLSM). The fluorescence signal of the rhodamine tagged HA/SiO 2 /Fe 3 O 4 NPs was clearly observed in cells after 4 h of coincubation, and the fluorescence signals of rhodamine and LysoTracker Green showed high colocalization (Fig. 3a), suggesting most of HA/SiO 2 /Fe 3 O 4 NPs existed in the endosomes. After removing the free NPs, the cells were exposed to UV light to initiate the in situ photo-polymerization to form the crosslinked HA nanogel in the endosome, namely EM-NG. The cells were then lysed, and EM-NG containing magnetic nanoparticle were collected via magnetic extraction (Fig. 3d). As determined by DLS, the EM-NG was 262.3 nm in hydrodynamic diameter (Fig. 3c). The TEM picture showed the EM-NGs were roundoval in shape, and the membrane with a thickness of 10 nm and the multiple of inclusions of silica NPs were clearly observed in each EM-NG (Fig. 3b). To further confirm the successful coating of endosome membrane, the western blotting analysis was performed against the early endosome antigen 1 (EEA1), the endosome membrane-specific marker. The result showed a significant enrichment of EEA1 was present on the EM-NGs rather than the bare HA NGs (Fig. 3e).

Drug loading and release from EM-NGs
DOX, as a model hydrophilic anticancer drug, was loaded into EM-NG via dispersion 35,36 . The loading capacity was determined as 7.3%. The in vitro drug release behavior was investigated under different pH conditions over time. The DOX-EM-NG exhibited near a zero-order release kinetics at pH 7.4 (Fig. 3f). In contrast, DOX released from nanogel was much faster at an acidic condition than physiological condition. The accelerated drug release at acidic pH is mainly due to the weakening of the binding between the EM-NG and drug, and improved solubility of DOX at low pH 37,38 . This pH-dependent drug release behavior can play a crucial role in tumor-targeted drug delivery via endocytosis pathway.

In vitro delivery of DOX by EM-NG
In order to determine the endocytosis pathway of DOX-EM-NG, HeLa cells were preincubated with several specific inhibitors of different kinds of endocytosis, including chlorpromazine (CPZ) for clathrin-mediated endocytosis, nystatin (NYS) for the caveolinmediated endocytosis inhibition, amiloride (AMI) for the macropinocytosis inhibition, and methyl-β-cyclodextrin (MCD) for the lipid raft inhibition. As shown in Fig. 4b, CPZ, NYS and AMI all reduced the uptake of DOX-EM-NG significantly, suggesting that DOX-EM-NG were taken up by HeLa cells through clathrin-mediated endocytosis, caveolin-mediated endocytosis, and lipid raft. In contrast, there was insignificant inhibition of uptake efficiency in the cells pretreated with AMI for the macropinocytosis inhibition. These results indicated that several pathways were involved in the internalization of cell membrane with DOX-EM-NG, due to a large number of receptors and proteins on the surface of endosome membrane shell of DOX-EM-NG 13,15 . The fluorescence of DOX was clearly observed in cells after 1 h of incubation with DOX-EM-NG, visualized by CLSM (Fig. 4a and Fig. S1 in the ESI †), which validated the cellular internalization of DOX-EM-NG. When prolonged the incubation time to 4 h, DOX was remarkably released and delivered into the nuclei of cells. To explore the targeting capability of DOX-EM-NG, EM-NG extracted from HeLa cells was further lysed to obtain the bare HA nanogel (NG) by removing the endosome membrane 39 . By quantitative analysis using the flow cytometry, it was demonstrated that incubation of DOX-EM-NG with HeLa cells in vitro leaded to significantly increased uptake as compared to the DOX loaded bare HA nanogel (DOX-NG) and free DOX (Fig. 4c).

In vitro cytotoxicity of DOX-EM-NGs
Next, the in vitro cytotoxicity of nanogel against HeLa cells was evaluated by using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Both DOX-EM-NG and DOX-NG showed significantly enhanced cytotoxicity toward HeLa cells compared to free DOX solution upon 24-hour incubation (Fig. 5a). Notably, the cell viability of cells treated with DOX-EM-NG was much lower than that of those treated with DOX-NG, suggesting the better cell targeting ability of endosome membrane than the pure HA. The results were consistent with the cellular uptake studies measured above. Bare EM-NG did not exhibit significant cytotoxicity within the studied range of concentrations. In order to assess the capability of DOX-EM-NG to homotypically target cancer cells, the human lung adenocarcinoma epithelial (A549) cells, as a heterotypic cell line, were incubated with DOX-EM-NG or DOX-NG for 24 h. The results showed there was insignificant difference in cell viability for DOX-EM-NG and DOX-NG (Fig. 5b), which further indicated that the enhanced binding effect was specifically associated with the membrane coating.

Conclusions
We have developed an innovative strategy utilizing endosome membrane-coated nanogels for enhanced delivery of anticancer drug. HA/SiO 2 /Fe 3 O 4 NPs encapsulating crosslinker and photo-initiator were confirmed to efficiently interact with endosomes by CLSM, and a subsequent in situ polymerization happened inside endosome, resulting in the endosome membrane coating HA nanogels. These EM-NGs containing magnetic nanocrystals were easily collected using a magnet after cell lysis. Through the western blotting analysis against endosome membrane specific protein (EEA1), the successful coating of endosome membrane was substantiated, and the endosome membrane shell could be removed via further lysis for long time. After loading with DOX, the resulting nanogel was demonstrated to efficiently target the source cancer cells through a specific homotypic affinity, which increased 1.5-fold in uptake efficiency compared to the bare HA NGs. Moreover, DOX-EM-NGs were able to deliver anticancer drug to source cancer cells with enhanced efficacy, while there were insignificant differences compared to DOX-NGs when treated toward nonsource cancer cells.
Furthermore, stimuli-responsive moieties can be integrated with these EM-NGs to achieve controllable drug release 40 . We will also evaluate in vivo targeting capability, antitumor efficacy and systemic toxicity of nanogel. This strategy provides guideline to develop endosome membrane-coated nanomedicine from primary tumors of patients for personalized anticancer treatments.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.