Protein nanocages that penetrate airway mucus and tumor tissue

Abstract

Reports on drug delivery systems capable of overcoming multiple biological barriers are rare. We introduce a nanoparticle-based drug delivery technology capable of rapidly penetrating both lung tumor tissue and the mucus layer that protects airway tissues from nanoscale objects. Specifically, human ferritin heavy-chain nanocages (FTn) were functionalized with polyethylene glycol (PEG) in a unique manner that allows robust control over PEG location (nanoparticle surface only) and surface density. We varied PEG surface density and molecular weight to discover PEGylated FTn that rapidly penetrated both mucus barriers and tumor tissues in vitro and in vivo. Upon inhalation in mice, PEGylated FTn with optimized PEGylation rapidly penetrated the mucus gel layer and thus provided a uniform distribution throughout the airways. Subsequently, PEGylated FTn preferentially penetrated and distributed within orthotopic lung tumor tissue, and selectively entered cancer cells, in a transferrin receptor 1-dependent manner, which is up-regulated in most cancers. To test the potential therapeutic benefits, doxorubicin (DOX) was conjugated to PEGylated FTn via an acid-labile linker to facilitate intracellular release of DOX after cell entry. Inhalation of DOX-loaded PEGylated FTn led to 60% survival, compared with 10% survival in the group that inhaled DOX in solution at the maximally tolerated dose, in a murine model of malignant airway lung cancer. This approach may provide benefits as an adjuvant therapy combined with systemic chemo- or immunotherapy or as a stand-alone therapy for patients with tumors confined to the airways.

Keywords: PEG; biological barriers; human ferritin; lung cancer; nanoparticle.

Fig. 1.

Physicochemical characterization of FTn. (A) TEM images of nanostructures of FTn at pH 7.4 and pH 2.0 (stained with 1% uranyl acetate). (Scale bars, 50 nm.) (B) The diameters of FTn at pH 7.4 and pH 2.0 were analyzed by size chromatography. (C) Surface modification does not affect reassembly capacity of FTn. FTn with different modification were reassembled by controlling pH. (Top) Gel electrophoresis of the reassembled FTn formulated with varying ratios of FTn to FTn–PEG_2k (the ratio of AF488-labeled FTn used was fixed; 8 of 24 subunits). (Bottom) Gel electrophoresis analysis of the reassembled particles with AF488-labeled FTn and non-PEGylated/PEGylated FTn (PEG molecular mass = 2, 5, and 10 kDa) (the ratio of AF488-labeled FTn used was fixed).

Fig. S1.

Construction of standard curve between HD and retention time by using GFC equipped with Superose 6 column. (A) GFC analysis of protein standards (PBS, pH 7.4). Molecular mass markers M1 (thyroglobulin: 669 kDa, 18.8 nm HD), M2 (γ-globulin: 158 kDa, 11.9 nm HD), M3 (ovalbumin: 44 kDa, 6.13 nm HD), M4 (myoglobin: 17 kDa, 3.83 nm HD), and M5 (vitamin B12: 1.35 kDa, 1.48 nm HD) are shown by arrows. The retention time of protein standards at an absorbance of 280 nm was analyzed by reversed-phase HPLC. (B) GFC standard curve between HD and retention time. HD (nm) = –1.409 × time (min) + 43.106.

Fig. S2.

PEGylation does not affect reassembly of FTn. (A) TEM images of FTn and PEGylated FTn with different MWs of PEG. (Scale bars, 100 nm.) (B) TEM images of reassembled FTn with different modification. FTn–Cy5, Cy5-conjugated FTn; FTn–PEG_x, PEGylated FTn, x represents MW of PEG. (Scale bars, 100 nm.)

Fig. 2.

Distribution of FTn/FTn–PEG_x throughout the mucus-covered mouse lung airways. (A) Representative images of the in vivo distribution of Cy5-labeled FTn or FTn/FTn–PEG_x (x = 2, 5, and 10 kDa) (red) in the large airways of mice (n = 4 mice). Cell nuclei are stained with DAPI (blue). (Scale bars, 200 µm.) (B) Image-based quantification of the coverage area of various particles in the large airways. Tissues were harvested 10 min after intratracheal administration. Data represent mean percentage coverage ± SEM. **P < 0.01 compared with the non-PEGylated FTn. (C) Image-based quantification of the total fluorescence intensity distributed throughout mouse airways. *P < 0.05 compared with the non-PEGylated FTn.

Fig. S3.

Representative images of mucus layer in the airway. (A) Fluorescence image showing distribution of Cy5-labeled FTn/FTn–PEG_2k throughout the mucus-covered large airway (i.e., trachea) in the mouse lung. (B) Enlarged images of the dashed box in A. The air–liquid interface layer of mouse appears to be around 50 µm in thickness.

Fig. 3.

Specific cell uptake and effect of PEGylation on the intrinsic cancer cell targeting capacity of FTn. (A) Confocal images of 3LL cells following incubation with the FTn. Blue, nucleus; green, membrane; red, Cy5-labeled FTn. (Scale bar, 20 µm.) (B) Flow cytometry analysis of cell uptake of Cy5-labeled FTn at different concentrations 2 h after the treatment. (C) Flow cytometry analysis of TfR 1-dependent cell binding of Cy5-labeled FTn to 3LL or A549 cells in the presence of anti-TfR 1 Ab. **P < 0.01. (D) Effect of PEG MW on the cell uptake of PEGylated FTn. *P < 0.05, **P < 0.01 compared with non-PEGylated FTn. (E) Competition binding of Cy5-labeled FTn to 3LL cells at different concentrations of FTn or FTn/FTn–PEG_2k. (F) Western blot analysis of TfR 1 expression in human NSCLC and SCLC cell lines as well as C2C12 normal cell line. (G) Flow cytometry analysis of the cell uptake of Cy5-labeled FTn/FTn–PEG_2k in various lung cancer cell lines.

Fig. 4.

Deep penetration of PEGylated FTn throughout tumor tissues. (A) Effect of PEGylation on the penetration of FTn through a 3D tumor spheroid model established with 3LL cells. (Left) Representative confocal images of tumor penetration of the various particles in 3D spheroids. (Scale bar, 200 µm.) (Right) Image-based quantification of mean Cy5 fluorescence signal intensity in whole 3D-constructed cell spheroids. Data represent the average of n ≥ 8 cell spheroids (±SEM). **P < 0.01 compared with non-PEGylated FTn. (B) Quantification analysis of tumor penetration of Cy5-labeled FTn in reconstructed 3D cell spheroids in the absence or presence of excess amounts of unlabeled FTn. **P < 0.01. (C) The penetration of the FTn and FTn/FTn–PEG_2k through cell spheroids in the absence or presence of a 10-fold molar excess of anti-TfR 1 Ab. (Scale bar, 200 µm.) (D) Distribution of the Cy5 fluorescence signal in the middle section images of C. (Left) Representative middle section image. (Right) Image-based quantification of the mean intensity at different penetration distance (Top, FTn; Bottom, FTn/FTn–PEG_2k). Radius, 0 and 1 indicate the center and edge of the tumor spheroid, respectively. **P < 0.01 compared with FTn or FTn/FTn–PEG_2k. (Scale bar, 200 µm.) (E) Western blot analysis of TfR 1 expression in proximal and distal lung as well as 3LL tumor tissues. The tissues were collected from n = 3 mice for each group. **P < 0.01. (F) Colocalization (yellow) of intratracheally administered FTn/FTn–PEG_2k (red) with an orthotopic 3LL lung cancer (mCherry; green). Cell nuclei are stained with DAPI (blue). (Scale bar, 100 µm.)

Fig. S4.

H&E-stained tissue sections show that untreated and FTn/FTn–PEG_2k-treated lung tissues are virtually identical with no significant histopathological sign of acute inflammation or cytotoxicity in both proximal and distal lung tissues. (Scale bars, 100 μm.)

Fig. 5.

Intracellular drug release and cell uptake of FTn/FTn–PEG_2k/DOX conjugates. (A) Schematic illustration of FTn/FTn–PEG_2k conjugated with DOX via an acid-sensitive linker. (B) Representative TEM image of FTn/FTn–PEG_2k/DOX. (Scale bar, 100 nm.) (C) DOX release kinetics from FTn/FTn–PEG_2k/DOX incubated at pH 7.4 and pH 5.0. (D) Representative confocal images of the colocalization (yellow) of the Cy5-labeled FTn/FTn–PEG_2k (red) and lysosome (green) in 3LL cells. (Scale bar, 10 µm.) (E) Representative confocal images of the cellular localization of free DOX (red) and FTn/FTn–PEG_2k/DOX (red) in 3LL cells. Cell membrane and nuclei are stained with Vybrant DiO (green) and DAPI (blue), respectively. (Scale bar, 10 µm.) (F) Relative amounts of free DOX or FTn/FTn–PEG_2k/DOX in cytoplasm over nuclei at different time points. ns, nonsignificant; **P < 0.01. (G) Tumor penetration of free DOX and FTn/FTn–PEG_2k/DOX in 3LL-based spheroids. (Left) Representative 3D reconstructed confocal images. (Right) Image-based quantification of the mean fluorescence intensity throughout the entire 3D reconstructed spheroids. **P < 0.01 compared with the free DOX. (Scale bars, 100 µm.) (H) Distribution of the DOX fluorescence signal in the middle section of spheroids. (Left) Representative middle section image. (Right) Image-based quantification of the mean intensity at different penetration distance. *P < 0.05, **P < 0.01 compared with the free DOX. (Scale bars, 100 µm.)

Fig. S5.

Conjugation of DOX to FTn/FTn–PEG_2k. (A) Schematic diagram showing conjugation of the FTn/FTn–PEG_2k and DOX via acid-activated cross-linker. (B) HPLC analysis of acid-activated DOX.

Fig. S6.

Representative confocal images demonstrating colocalization (yellow) of the FTn/FTn–PEG_2k/DOX (red) and lysosomes (green; AF488-conjugated anti-LAMP 1 Ab) in 3LL cells at 2 h postadministration of the nanocages. (Scale bar, 10 µm.)

Fig. S7.

FTn/FTn–PEG_2k improve tumor penetration of DOX into tumor spheroid established with different lung cancer cells, including (A) A549, (B) H1975, and (C) H460 cells. Confocal representative images (Left) and quantification analysis (Right) of tumor penetration of free DOX or FTn/FTn–PEG_2k/DOX throughout the entire 3D-reconstructed cell spheroids. (Scale bars, 100 µm.) **P < 0.01 compared with the free DOX.

Fig. S8.

In vitro antitumor activity of DOX and FTn/FTn–PEG_2k/DOX in 3LL-Luc cells. In vitro cytotoxicity was investigated by firefly luciferase bioluminescence imaging after 24 h of treatment at different concentrations of DOX and FTn/FTn–PEG_2k/DOX (equivalent DOX quantity). (A) Luciferase activity of 3LL-Luc was acquired and (B) analyzed using a Xenogen IVIS imaging system. (C) Representative bright-field images of 3LL-based spheroids untreated and treated with either DOX or FTn/FTn–PEG_2k/DOX. (Scale bars, 200 µm.) (D) Imaged-based quantification of the relative average diameter of untreated and treated cell spheroids. Data represent the average of n ≥ 20 cell spheroids (±SEM). *P < 0.05 compared with free DOX.

Fig. 6.

Improved therapeutic efficacy provided by locally administered FTn/FTn–PEG_2k/DOX in an orthotopic mouse model of proximal lung cancer. (A) Colocalization of FTn/FTn–PEG_2k/DOX (red) and an orthotopically established 3LL lung tumor tissue (green) at 2 h postadministration of the nanocages. Cell nuclei are stained with DAPI (blue). (Scale bar, 100 µm.) (B) Bioluminescence signal over time postinoculation through IVIS imaging. Mice were untreated or treated with intratracheally administered free DOX or FTn/FTn–PEG_2k/DOX, 3 d following the tumor inoculation (n = 10 mice per group). (C) Quantitative analysis of the bioluminescence signal over time. *P < 0.05 compared with untreated or free DOX. (D) Kaplan–Meier survival curves for the mice with and without treatment. **P < 0.01 compared with untreated or free DOX.

Fig. S9.

Construction of orthotopic mouse models of proximal lung cancer and bioluminescence imaging for monitoring the tumor growth. (A) Distribution of cancer cells in lung following intratracheal intubation of 3LL-Luc cells (Left). The tumor growth was visualized on an IVIS bioluminescence system 3 d after the inoculation of 3LL-Luc cells (Right). (B) Timeline of model establishment, treatment, and imaging for the proximal model shown in A.

Fig. S10.

Blank FTn/FTn–PEG_2k devoid of DOX do not affect the tumor growth and survival of an orthotopic mouse model of proximal lung cancer. (A) Bioluminescence signal over time postinoculation thorough IVIS imaging. Mice were untreated (n = 7) or treated with intratracheally administered FTn/FTn–PEG_2k (n = 8), 3 d following the tumor inoculation. (B) Quantitative analysis of the bioluminescence signal over time. (C) Kaplan–Meier survival curves for the mice with and without treatment.