Bioengineered yeast-derived vacuoles with enhanced tissue-penetrating ability for targeted cancer therapy

Significance

Tumor tissues have formidable physiological barriers, such as a high interstitial pressure and a densely entangled ECM. Most synthetic nanomaterials used for drug delivery fail to penetrate tumor tissues deeply and localize only in perivascular areas, thereby limiting their therapeutic efficacy. This report describes bioengineered yeast-derived natural nanocarriers for cancer-specific targeting and drug delivery. Budding yeast was genetically engineered to produce large numbers of nanosized compartments—vacuoles that display cancer—targeting ligands on their surface. The nanosized vacuoles significantly enhanced drug penetration in tumor xenografts, and consequently prevented tumor growth without eliciting immune responses. This result shows that the biological nanocarriers overcome the limitations associated with synthetic cancer-targeting nanomaterials, and thus can be used to treat various cancers.

Abstract

Despite the appreciable success of synthetic nanomaterials for targeted cancer therapy in preclinical studies, technical challenges involving their large-scale, cost-effective production and intrinsic toxicity associated with the materials, as well as their inability to penetrate tumor tissues deeply, limit their clinical translation. Here, we describe biologically derived nanocarriers developed from a bioengineered yeast strain that may overcome such impediments. The budding yeast Saccharomyces cerevisiae was genetically engineered to produce nanosized vacuoles displaying human epidermal growth factor receptor 2 (HER2)-specific affibody for active targeting. These nanosized vacuoles efficiently loaded the anticancer drug doxorubicin (Dox) and were effectively endocytosed by cultured cancer cells. Their cancer-targeting ability, along with their unique endomembrane compositions, significantly enhanced drug penetration in multicellular cultures and improved drug distribution in a tumor xenograft. Furthermore, Dox-loaded vacuoles successfully prevented tumor growth without eliciting any prolonged immune responses. The current study provides a platform technology for generating cancer-specific, tissue-penetrating, safe, and scalable biological nanoparticles for targeted cancer therapy.

Recent decades have witnessed the emergence of the nanomedicine era in cancer diagnosis and therapy, made possible by the advent of new synthetic nanomaterials suitable for imaging or drug delivery (1, 2). Such nanomaterial-based drug delivery systems have shown improved therapeutic efficacy with lower unwanted adverse effects compared with conventional chemotherapy, at least in animal models (1–5). Despite the appreciable success of cancer nanomedicines in preclinical studies, only a few such agents have entered clinical trials, and fewer still have shown promising therapeutic outcomes and advanced to subsequent trial stages (6–8). Among factors likely to hamper the clinical feasibility of synthetic cancer-targeting nanomaterials are technical challenges involving their pilot scale, cost-effective production, and intrinsic cytotoxicity associated with these materials. In addition to these limitations, tumor tissues have formidable physiological barriers, such as a high interstitial pressure and densely entangled ECM. Most synthetic nanomaterials fail to penetrate tumor tissues deeply and localize only in perivascular areas, thereby limiting their therapeutic efficacy (6–8). To overcome the limitations associated with synthetic nanomedicines, researchers have recently developed biologically derived, nanosized vesicles as novel drug delivery systems (9–12). Minicells and outer membrane vesicles derived from genetically engineered bacteria and mammalian cell-derived exosomes have shown potent antitumor efficacy (10–14). Moreover, such biological vectors can be engineered to achieve cancer cell-specific delivery of diverse cargos by introducing cancer-targeting ligands onto their surfaces (10–12). However, concerns remain about the immunogenicity of bacteria-derived vehicles. For exosome-based delivery systems, large-scale, cost-effective methods for producing cancer cell-specific exosomes are currently bottlenecks for future clinical applications. Thus, there is a need for a new class of biologically derived drug delivery systems that have an enhanced tissue-penetrating ability and cancer cell specificity, as well as low immunogenicity and facile drug loading, and that can be produced cost-effectively on a large scale.

To this end, we report here the development of a bio-inspired drug delivery system using vacuoles isolated from genetically engineered yeast cells that may fulfill these unmet needs. Budding yeast (Saccharomyces cerevisiae) is nonhazardous and nonpathogenic, and thus has been used in fermentation products for millennia (15). In addition, its simple genetic and biochemical manipulations, together with the fact that many yeast genes are conserved among higher eukaryotes, including humans, make it an ideal model for various biological studies (15, 16). Unlike bacteria, its lipid composition is similar to the composition of mammalian cell membranes, potentially increasing fusion efficiency with the plasma membrane or endolysosomes, and thereby facilitating the release of drugs into targeted cells or tissues (17–20). Exploiting these features, we genetically engineered S. cerevisiae to express human epidermal growth factor receptor 2 (HER2)-specific affibody on the vacuolar membrane, enabling the targeting of HER2 receptors that are expressed on various cancers (12, 21, 22). We further loaded the anti-HER2 affibody-expressing vacuoles (Affi_HER2Vacuole) with the chemotherapeutic drug doxorubicin (Dox), which is widely used in treating solid tumors. We then examined the anticancer effects of the drug-carrying, HER2-targeted vacuoles (Affi_HER2Vacuole^Dox) and compared their anticancer efficacy with the anticancer efficacy of the free drug (Free^Dox); drug-free vacuoles (Affi_HER2Vacuole); and affibody-free, Dox-carrying vacuoles (Vacuole^Dox) in HER2-overexpressing, in vitro-cultured, cell-based models and an in vivo mouse xenograft model.

The proposed bioengineered yeast vacuole-based platform partly imparts limitations associated with synthetic cancer-targeting nanomaterials and provides inspiration and future directions for designing of an effective anticancer drug delivery system that could be considered in clinical applications to treat various cancers.

Results and Discussion

Preparation and Characterization of Cancer-Targeting Vacuoles.

Although WT yeast cells have two to five large vacuoles that are around 1 μm in diameter, their size and number vary depending upon the cell cycle and environmental conditions, reflecting vacuole-associated fission and fusion events (17, 19, 20, 23). The processes of fission and fusion are tightly regulated by multiple factors, including Rab GTPases that mediate membrane tethering and docking, which are prerequisites for homotypic vacuole fusion (24–29). The major Rab GTPase for yeast vacuole fusion is yeast protein transport 7 (YPT7); thus, its deletion results in a large number of fragmented, nanosized vacuoles (25–27, 30–32). To obtain a high yield of nanosized vacuoles, we generated a yeast strain with a YPT7 deletion (YPT7Δ) (Fig. 1A and Fig. S1A). Additionally, the yeast cells were genetically engineered to express the HER2-specific affibody on the vacuolar membrane for cancer targeting. Such modified YPT7Δ yeast cells contained a large number of fragmented vacuoles with enhanced colloidal stability. To express the Affibody_HER2 on the vacuolar membrane, we genetically fused the gene encoding myc-tagged Affibody_HER2 to the 5′ end of the PHO8 gene, which encodes vacuolar transmembrane alkaline phosphatase. The Affibody_HER2-expressing vacuoles (Affi_HER2Vacuoles) were then purified as described in SI Materials and Methods, Yeast Strains, Plasmid Construction, and Vacuole Isolation. Specific expression of the PHO8-Affibody_HER2 chimeric protein in the Affi_HER2Vacuole was shown by Western blot analysis using antibodies against PHO8 and the myc tag (Fig. 1B). These Affi_HER2Vacuoles were characterized in terms of their size and morphology using electrophoretic light scattering (ELS) and transmission electron microscopy (TEM) (Fig. 1C and Fig. S1 B and C). These analyses of the Affi_HER2Vacuoles revealed a bilayered, circular morphology with a hydrodynamic diameter of ∼200 nm. The HER2 specificity of the Affi_HER2Vacuole was verified by ELISA. The Affi_HER2Vacuole showed an ∼14-fold higher affinity for HER2 compared with the controls (Fig. 1D). Importantly, after proteinase kinase (PK) treatment, the Affi_HER2Vacuole showed an almost complete loss of HER2 affinity, indicating that the functional form of the Affibody_HER2 was exposed on the exterior vacuole surface. Next, by confocal imaging, we showed that the Affi_HER2Vacuoles were specifically bound to HER2 receptors on the cell surface and internalized through receptor-mediated endocytosis (Fig. 1E). A large proportion of the Affi_HER2Vacuoles were endocytosed by HER2-overexpressing SKOV3 cells, whereas a relatively low nonspecific uptake was evident in HER2-negative MDA-MB-231 cells. Collectively, these results clearly indicate that the Affi_HER2Vacuole can be used for cancer-specific targeting and drug delivery. Because of their subnanometer size, the engineered, drug-carrying vacuoles could rapidly extravasate and accumulate in the periphery of tumor tissue, where, after cancer receptor engagement, they could undergo endocytosis before releasing their payload and producing cytotoxic effects. In addition, with the available rapid protocols for vacuole preparation and methods for large-scale fermentation of yeasts, it is possible to obtain high yields of vacuoles in a cost-effective manner (26, 33).

Fig. 1.

Generation and characterization of Affi_HER2Vacuoles. (A) Schematic illustration of fragmented vacuole generation and drug loading into anti-HER2 affibody-expressing vacuoles (Affi_HER2Vacuole^Dox). (B) Western blot analysis of PHO8-affibody and myc-tagged affibody in Affi_HER2Vacuoles. The analyses confirmed the localization of the affibody in vacuoles as a fusion partner. (C) Representative TEM image of the Affi_HER2Vacuoles. (Scale bar: 50 nm.) (D) ELISA results showing the specificity of the Affi_HER2Vacuoles for HER2 protein. Treatment with PK caused almost a complete loss of HER2p affinity. Data represent the mean ± SD of three replicates. (E) Confocal images showing cell binding (0 h) and uptake (3 h and 6 h) of Affi_HER2Vacuoles in HER2-overexpressing SKOV3 cells and HER2-negative MDA-MB-231 cells. The nuclei were stained with DRAQ5 (blue), and Affi_HER2Vacuoles were stained with anti-affibody antibody (green). Affi_HER2Vacuoles exhibited receptor-specific binding and internalization in HER2-overexpressing SKOV3 cells, whereas very low and nonspecific uptake was evident in HER2-negative MDA-MB-231 cells. (Scale bars: 20 μm.)

Fig. S1.

Characterization of the vacuoles. (A) Bright-field images showing normal or fragmented vacuoles inside the WT or modified BJ3505 yeast strain, respectively. Vacuolar membrane staining was done with FM4/64 lipophilic dye. (Magnification: 100×.) (B) Hydrodynamic size distribution of the Affi_HER2Vacuoles analyzed before drug loading. (C) Representative TEM image of the Affi_HER2Vacuoles. (Scale bars: 100 nm; Inset, 50 nm.)

Drug Loading, Quantification, and Release Kinetics.

We expected that Dox, a membrane-permeable drug, could be loaded into the membrane and lumen of the Affi_HER2Vacuole through physical adsorption because the lipid composition of the vacuoles is relatively similar to the lipid composition of mammalian cell membranes (10, 17–19). As shown, the amount of drug loaded in the vacuoles increased by increasing the initial Dox concentration, suggesting that drug loading was mediated by a concentration gradient (Fig. S2A). Characterization of the size and shape of the Dox-loaded vacuoles (Affi_HER2Vacuole^Dox) by ELS and TEM analyses showed that these vesicles retained an intact circular shape (Fig. S2 B and C), confirming that the drug loading did not affect the physical properties of the Affi_HER2Vacuole. Next, we investigated drug release from Affi_HER2Vacuole^Dox. As shown, less than 10% of the Dox was released during 24 h of incubation at pH 7.4, whereas drug release was accelerated in an acidic (pH 5.1) environment (Fig. S2D). These results suggest that the vacuole structure becomes destabilized or disrupted in an acidic pH environment, like that in the lysosome of a cell, enabling the rapid release of the payload, but remains intact and relatively stable under physiological conditions, minimizing leaking of the loaded drug into the circulatory system. This drug release profile is crucial for reducing nonspecific cytotoxicity caused by the free drug and for improving the terminal half-life (t_1/2) of the drug in vivo by preventing its rapid metabolism or excretion.

Fig. S2.

Characterization of the Dox-loaded vacuoles and drug efflux analysis. (A) Quantification of the Dox loading of the Affi_HER2Vacuoles (100 μg) following incubation with different amounts (100 μg, 200 μg, or 300 μg) of the drug (n = 3 replicates per sample). Drug loading was evaluated by measuring the fluorescence intensity of each sample, as described in SI Materials and Methods. (B) Hydrodynamic size distribution of Affi_HER2Vacuole^Dox analyzed after drug loading. (C) Representative TEM image of Affi_HER2Vacuole^Dox. (Scale bars: 100 nm; Inset, 50 nm.) (D) Dox-release profile in physiological (pH 7.4) and acidic (pH 5.1) environments. Rapid efflux was detected under acidic conditions. The fluorescence intensity of each sample was evaluated as described in SI Materials and Methods.

Cellular Drug Delivery and Cytotoxicity of Affi_HER2Vacuole^Dox.

We next investigated the cellular uptake and cytotoxic effects of Affi_HER2Vacuole^Dox in various cancer cell lines. As shown in Fig. 2A, compared with the nontargeting Vacuole^Dox, Affi_HER2Vacuole^Dox was largely endocytosed by HER2-overexpressing NIH3T6.7 cells. At 1 h and 12 h, the red fluorescence of Dox was detected largely in the cytoplasm (Fig. S3A), indicating that Dox remained associated with the vacuoles. However, after 12 h, most of the Dox had been released and was readily internalized in the nucleus. Although a clear difference in uptake efficiency was observed between both types of Dox-loaded vacuoles, an almost similar pattern of delayed drug release and nuclear uptake was evident with Affi_HER2Vacuole^Dox and Vacuole^Dox. In contrast, due to its intrinsic property, the Free^Dox could readily penetrate through the cell membrane and internalize into the nuclei, which was observed within 1 h of treatment. The results clearly show the differences in intracellular kinetics of Free^Dox and Dox-loaded vacuoles. Furthermore, compared with HER2-overexpressing SKOV3 and NIH3T6.7 cells, HER2-negative MDA-MB-231 cells exhibited a very low uptake of Affi_HER2Vacuole^Dox (Fig. S3B). These results show that the receptor-specific binding of Affi_HER2Vacuole^Dox is crucial for cellular uptake, whereas drug release from Affi_HER2Vacuole^Dox takes place presumably due to deformation of the Affi_HER2Vacuole upon lysosomal acidification. Next, the cytotoxic effect of Affi_HER2Vacuole^Dox was quantitatively evaluated with colorimetric 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide assays. In contrast to treatment with PBS, vacuole or Vacuole^Dox and treatment with Free^Dox or Affi_HER2Vacuole^Dox caused significantly greater cytotoxicity after 96 h of incubation (Fig. 2B), reducing cell viability in both the HER2-overexpressing NIH3T6.7 and SKOV3 cells by 75% and 72%, respectively. The cytotoxicity of Free^Dox in HER2-negative MDA-MB-231 cells was significantly higher than the cytotoxicity of Affi_HER2Vacuole^Dox, indicating that the HER2-specific targeting of Affi_HER2Vacuole^Dox is necessary for cell-selective cytotoxicity.

Fig. 2.

In vitro drug delivery and cytotoxicity. (A) Confocal images showing in vitro uptake of HER2-targeted Affi_HER2Vacuole^Dox and nontargeting Vacuole^Dox by NIH3T6.7 cells. Higher uptake was evident in the Affi_HER2Vacuole^Dox-treated cells. (Scale bars: 20 μm.) (B) In vitro cytotoxic efficacy (analyzed by 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide assay) of the Affi_HER2Vacuole^Dox was compared with the in vitro cytotoxic efficacy of various controls in different cell lines (NIH3T6.7, SKOV3, and MD-AMB-231). Cell viability was evaluated after 96 h of incubation with Affi_HER2Vacuole^Dox (Dox: 1 μM) and compared with empty vacuole (20 μg), Vacuole^Dox (Dox: 1 μM), and two sets of Free^Dox (Dox: 100 nM and 500 nM, respectively). Data represent the mean ± SD of three replicates, expressed relative to the PBS-treated control (**P < 0.01). (C) Confocal images showing the time-dependent uptake and penetration of the Affi_HER2Vacuole^Dox in NIH3T6.7 spheroids. (Scale bar: 200 μm.) (D) Bright-field images of the NIH3T6.7 spheroids after 96 h of treatment with (i) Free^Dox, (ii) Vacuole^Dox, (iii) Affi_HER2Vacuole^Dox, and (iv) PBS. The cytotoxicity and total volume control were the highest in the Affi_HER2Vacuole^Dox-treated samples compared with the controls. (Scale bar: 200 μm.)

Fig. S3.

Intracellular kinetics and cell-specific uptake analysis. (A) Confocal images showing the time-dependent uptake of Affi_HER2Vacuoles^Dox compared with the controls; within 1 h of treatment, Affi_HER2Vacuoles^Dox were largely internalized by NIH3T6.7 cells and remained in the cytosol until 12 h. At 24 h and 48 h, nuclear translocation of Dox was evident, indicating drug release from the vacuoles after 12 h. (Scale bars: 10 μm.) (B) Confocal microscope images showing HER2 receptor-specific uptake of the Affi_HER2Vacuoles^Dox. Bright red fluorescence from the cytosol of HER2-overexpressing cells (SKOV3 and NIH3T6.7) indicates greater uptake of Affi_HER2Vacuoles^Dox compared with the HER2-negative MDA-MB-231 cells. (Scale bars: 10 μm.)

Drug Penetration and Cytotoxicity in Multicellular Cultures.

In addition to endocytosis, we monitored the transcytosis or penetration of Affi_HER2Vacuole^Dox with NIH3T6.7 cells cultured in multiple layers on a Transwell filter system (Costar). Cells in the Transwell filter system were treated with Free^Dox, nontargeting Vacuole^Dox, or Affi_HER2Vacuole^Dox, and penetration was assessed as a function of time by monitoring the fluorescence of the cells cultured below the transwell. Drug uptake and accumulation were increased compared with the controls after incubation with Affi_HER2Vacuole^Dox (Fig. S4), suggesting that, following receptor-mediated endocytosis, Affi_HER2Vacuole^Dox might undergo transcytosis through layers of cells. In contrast, Free^Dox could not penetrate the cellular layers due to rapid nuclear entrapment and Vacuole^Dox exhibited very low transcytosis because of reduced uptake by cells in the transwell and was largely removed by washing.

Fig. S4.

Transwell assay for drug penetration and uptake analysis. Confocal images show enhanced transcytosis and uptake of the Affi_HER2Vacuoles^Dox compared with the controls by NIH3T6.7 cells cultured in six-well plates below the Transwell system. Confocal imaging was done at different time points. (Scale bars: 10 μm.)

We next tested the uptake, penetration, and cytotoxic effects of Affi_HER2Vacuole^Dox in 3D spheroid cultures generated from NIH3T6.7 cells. Nonvascularized cancer cell spheroids are known to share several features in common with tumors studied in vivo (34–36). As shown in Fig. 2C, internalization of Free^Dox was limited to the outer region of rapidly proliferating cells; such uptake could be attributable to the intrinsic high membrane permeability of the drug. Vacuole^Dox, lacking the HER2-binding ability, as well as Free^Dox showed a low level of uptake in the spheroids. In contrast, Affi_HER2Vacuole^Dox not only showed a relatively higher uptake in the outer proliferating regions but also penetrated much more deeply into the inner parts of the spheroids, possibly reflecting the advantages of the membrane composition and subnanometer size (mostly below 100 nm) of the vacuoles, which could penetrate more upon saturation of the outer region receptors. The greater penetration and distribution of Affi_HER2Vacuole^Dox in both the transwell and spheroid cultures compared with Free^Dox were dependent upon incubation time. Moreover, growth inhibition and volume control of the cancer cell spheroids after 96 h of treatment were much greater with Affi_HER2Vacuole^Dox than with the other controls (Fig. 2D). These results are in accordance with previous in vitro findings, showing that cellular targeting and enhanced penetration improved the efficacy of Dox (35). Taken together, these results show that Affi_HER2Vacuole^Dox can selectively deliver a drug payload to targeted cancer cells and exert cytotoxic effects. Importantly, Affi_HER2Vacuole alone is not cytotoxic in cells; therefore, the targeted Affi_HER2Vacuole^Dox system could be used to safely treat cancers in vivo.

In Vivo Cancer Targeting, Biodistribution, and Antitumor Efficacy.

Researchers have developed a variety of nanocarriers to improve the therapeutic index of drugs, mainly by increasing their efficacy and reducing their toxic side effects. However, curtailing vector-associated toxicity is technically challenging. Therefore, biomaterials that are biocompatible, bioresponsive, and biodegradable could be advantageous over synthetic vectors and continue to be actively developed (8, 9). To investigate the tumor targeting and therapeutic efficacy of Affi_HER2Vacuole^Dox in vivo, we used a HER2-overexpressing NIH3T6.7 cell-based mouse xenograft model (37). Mice bearing NIH3T6.7 xenografts were systemically injected with Affi_HER2Vacuole^Dox, Vacuole^Dox, or Free^Dox via the tail vein, and tumor tissue sections were analyzed for cellular drug distribution at fixed time points (6 h and 12 h) by monitoring the red fluorescence of the accumulated Dox. Images obtained from the various tissue sections (from the top to the center) of each tumor showed strong Dox fluorescence in the Affi_HER2Vacuole^Dox-injected animals compared with the control animals (Fig. 3A and Fig. S5A). The results show a greater accumulation and penetration of the Affi_HER2Vacuole^Dox compared with the controls. Although Vacuole^Dox showed a relatively low cellular uptake due to passive targeting, its enhanced distribution, like the enhanced distribution of the Affi_HER2Vacuole^Dox, ensures the effectiveness of biological carriers in penetrating tissue deeply. By comparison, the least accumulation was observed with Free^Dox, indicating that passive targeting prevents effective accumulation in tumors and that the drug is cleared through the circulation. We extended this analysis by assessing the biodistribution of Affi_HER2Vacuole^Dox compared with the biodistribution of nontargeting Vacuole^Dox and Free^Dox in NIH3T6.7 xenograft mice with each i.v.-administered formulation at a Dox dose of 1 mg/kg. Tumors and vital organs were collected 6 h postinjection to quantify the drug. As shown in Fig. 3B, animals injected with Affi_HER2Vacuole^Dox or Vacuole^Dox showed a relatively greater tumor accumulation of Dox compared with animals injected with Free^Dox, suggesting the extended t_1/2 of Dox-carrying vacuoles resulted in enhanced distribution to tumors. Furthermore, a comparison of targeted and nontargeted delivery showed a greater tumor-specific retention and accumulation of the HER2-targeted Affi_HER2Vacuole^Dox (∼11% injected dose per gram of tissue), suggesting that the HER2-specific affibody is responsible for the improved tumor-specific accumulation of the drug. To evaluate the behavior of Affi_HER2Vacuole^Dox and validate if the vacuole-based formulation could prolong the systemic circulation of Dox, the pharmacokinetic parameters were evaluated in tumor-free mice. The Affi_HER2Vacuole^Dox and Free^Dox were administered i.v., and the plasma drug concentrations were measured as a function of time. The Affi_HER2Vacuole^Dox showed the highest level of Dox retention in the circulation compared with Free^Dox (Fig. S5B). The pharmacokinetic parameters obtained after fitting the data in a noncompartmental pharmacokinetic model are also given (Fig. S5B). The terminal t_1/2 and area under the curve (AUC) of Affi_HER2Vacuole^Dox were 6.16-fold and 15.37-fold higher, respectively, than the terminal t_1/2 and AUC of Free^Dox. Taken together, these results suggest that the HER2-targeted Affi_HER2Vacuole^Dox performed well as a targeted drug delivery platform. Importantly, Affi_HER2Vacuole^Dox was able to circulate for a significantly longer period. An extended t_1/2, which ensures enhanced extravasation and receptor-specific endocytosis, is an essential parameter that achieves effective delivery of the cargo and long-term accumulation of the drug at the target tumor site.

Fig. 3.

In vivo tumor targeting, biodistribution, and antitumor effects of Affi_HER2Vacuole^Dox in NIH3T6.7 xenograft mice. (A) Confocal images of tumor sections show the drug distribution in cancer cells after 6 h of treatment with Affi_HER2Vacuole^Dox (1 mg/kg) compared with the controls. (Scale bars: 10 μm.) (B) Tumor-specific distribution of Affi_HER2Vacuole^Dox (1 mg/kg) analyzed 6 h after i.v. administration in NIH3T6.7 xenograft mice. Values are reported as the mean ± SD of triplicate samples. (C) Tumor growth was monitored throughout the course of the treatment (Dox at a dose of 1 mg/kg administered every other day) in each treatment group. The percentage of tumor growth inhibition (TGI) caused by Affi_HER2Vacuole^Dox was evaluated on the final day of treatment by comparison with PBS-treated controls; each value represents the mean ± SD (**P < 0.01 vs. control; n = 5 mice per group). (D) Apoptosis was determined by TUNEL assay measured as BrdU-FITC–positive cells (green) in tumor tissue sections from animals treated with Affi_HER2Vacuole^Dox compared with PBS controls. (Scale bars: 10 μm.) PI, propidium iodide.

Fig. S5.

In vivo tumor-specific accumulation and distribution and plasma pharmacokinetics. (A) Confocal analysis showing the intracellular distribution of Affi_HER2Vacuole^Dox (Dox at a dose of 1 mg/kg) at the 12-h time point, observed in different sections of the tumor (from the top to the center of the tissue) and compared with the Free^Dox- and Vacuole^Dox-treated control groups. The greater fluorescence from the Affi_HER2Vacuole^Dox-treated tissue sections indicates enhanced cellular accumulation and uptake due to HER2 targeting. (Scale bars: 10 μm.) (B) Plot of the plasma concentration vs. time of Dox in animals treated with either Affi_HER2Vacuole^Dox (1 mg/kg) or Free^Dox (1 mg/kg) and the respective pharmacokinetic parameters (n = 5). The pharmacokinetic parameters were obtained after fitting the data in a noncompartmental pharmacokinetic model.

We next compared the antitumor efficacy of Affi_HER2Vacuole^Dox with the antitumor efficacy of the nontargeting Vacuole^Dox and Free^Dox in NIH3T6.7 xenografts. After the tumors had become established, mice were divided into five groups: (i) vehicle control (PBS-treated), (ii) empty Affi_HER2Vacuole, (iii) Free^Dox (1 mg/kg), (iv) nontargeting Vacuole^Dox (1 mg/kg), and (v) HER2-targeted Affi_HER2Vacuole^Dox (1 mg/kg). Each regimen was administered i.v. every other day for a total of eight injections. Tumor growth rates and animal body weights were recorded during the course of each treatment. Tumor growth was significantly delayed in mice injected with Affi_HER2Vacuole^Dox compared with mice injected with PBS, empty Affi_HER2Vacuole, nontargeting Vacuole^Dox, or Free^Dox (Fig. 3C and Fig. S6A). The percentage of tumor growth inhibition on the last day of the Affi_HER2Vacuole^Dox treatment was 77.8%. Next, the harvested tumors were weighed and analyzed to assess the mechanism of cell killing. Consistent with the observed regression of tumor growth, the average weight of the excised tumors was significantly less in the Affi_HER2Vacuole^Dox-treated group than in the control groups (Fig. S6B). Throughout the treatment, the animals in all of the groups appeared normal, showing no overt signs of toxicity and a consistent body weight (Fig. S6C). Finally, the apoptosis of cells in the tumor tissue was assessed with TUNEL staining. This assay revealed abundant BrdU-FITC–positive apoptotic cells in the tumors excised from the mice treated with Affi_HER2Vacuole^Dox (Fig. 3D and Fig. S6D). Consistent with the drug distribution results, tumors from mice treated with Free^Dox or nontargeting Vacuole^Dox showed minimal apoptosis; no signs of apoptosis were detected in mice treated with PBS or empty Affi_HER2Vacuole. Moreover, the HER2-specific targeting of Affi_HER2Vacuole^Dox enabled a preferential concentration of the drug in the tumor tissue through the enhanced permeability and retention effect, as well as the promotion of the subsequent receptor-mediated endocytosis and intratumoral distribution. Collectively, these results indicate that nanocarrier-based drug delivery and active targeting substantially enhance therapeutic efficacy and increase the therapeutic index of the drug.

Fig. S6.

In vivo antitumor effects of the Affi_HER2Vacuole^Dox. (A) Sizes of the tumors dissected from each group were compared. (B) Actual weight of the dissected tumors from each group was recorded on the last day of treatment, and each value represents the mean ± SD (n = 5 mice per group; **P < 0.01). (C) Body weight of all of the animals was recorded during each treatment. All animals appeared healthy throughout the study. (D) Mechanism of cell death was confirmed by monitoring apoptotic cells with TUNEL staining in each tumor tissue section: (i) control (PBS), (ii) Affi_HER2Vacuole, (iii) Free^Dox (1 mg/kg), (iv) Vacuole^Dox (1 mg/kg), and (v) Affi_HER2Vacuole^Dox (1 mg/kg), respectively. The apoptotic cells were stained green with BrdU-FITC. (Scale bars: 10 μm.)

Cytotoxicity and Immune Responses.

To test the cytotoxic effects of the vacuoles in vitro, we treated human umbilical vein endothelial cells (HUVECs) with either wild type (WT) or Affi_HER2-modified vacuoles for 24 h, and evaluated the apoptotic effects by immunostaining. Compared with the vehicle control (methanol)-treated cells, vacuole-treated cells appeared normal; no signs of cytotoxicity were evident, and overall cell viability remained high (Fig. S7A).

Fig. S7.

In vitro biocompatibility and raw cell stimulation study. (A) HUVECs were treated with Affi_HER2Vacuole for 24 h and monitored by confocal microscopy. The vacuole-treated cells appeared normal with no signs of cytotoxicity, whereas methanol treatment caused massive cytotoxic effects (stained red with EthD-1). (Scale bars: 20 μm.) (B) ELISA was done to quantify TNF-α production from raw cells using media harvested 24 h after treatment. Levels of secreted TNF-α were significantly higher in raw cells treated with LPS (100 ng/mL) than in cells treated with either form of the vacuole (10 μg and 50 μg).

Next, we evaluated vacuole-induced immune responses both in vitro and in vivo. For the in vitro analyses, we stimulated murine macrophage-like RAW 264.7 cells with either form of the vacuole; LPS was used as a positive control. The TNF-α levels in the vacuole-treated cells were indistinguishable from the TNF-α levels in the PBS-treated cells (Fig. S7B), whereas LPS caused a significant and substantial cytokine response. These results suggest that the vacuoles used herein are very weakly immunogenic and could be well tolerated upon systemic administration. To confirm the in vivo safety of the vacuoles, we investigated whether repeated systemic injection of vacuoles (WT or Affibody_HER2-modified) overstimulated the immune system. Immune responses were recorded in C57BL/6 mice after systemic administration of 100 or 200 μg of vacuoles for 4 consecutive days by measuring the serum levels of TNF-α, IL-6, and IFN-γ with ELISA. Cytokine levels were evaluated at two time points, 2 h and 24 h, to monitor early and delayed immune responses. Both forms of vacuoles slightly increased the serum levels of TNF-α and IL-6 at the early (2-h) time point but had no effect on IFN-γ levels under any condition (Fig. 4). Importantly, however, cytokine levels returned to baseline by 24 h in all treatment groups. Moreover, there was no significant difference in immune stimulation at the two different concentrations of vacuoles, and neither treatment resulted in body weight loss or lethality. These results suggest that the vacuoles are nontoxic and well tolerated upon systemic administration. Therefore, the vacuoles themselves do not likely contribute to the antitumor effects by overstimulating immune pathways and, as such, are safe for in vivo systemic administration. Although these preliminary results in mice are encouraging, further blood chemistry and safety profile analyses in higher animals will be required.

Fig. 4.

Investigation of immune responses upon treatment with Affi_HER2Vacuoles in C57BL/6 mice. The serum levels of cytokines TNF-α, IL-6, and IFN-γ were quantified by ELISA. Analyses were performed at early (2 h) and late (24 h) time points after treatment with PBS or vacuoles (WT or Affi_HER2-modified: 100 μg and 200 μg each) for 4 consecutive days. Each value represents the mean ± SD (n = 5 mice per group) and was analyzed as described in SI Materials and Methods, Serum Cytokine Analysis.

SI Materials and Methods

Yeast Strains, Plasmid Construction, and Vacuole Isolation.

Vacuoles were purified from S. cerevisiae strains BJ3505, BJ3505-YPT7Δ, and BJ3505-YPT7Δ-Affi_HER2 as previously described (27, 38). The Affibody_HER2-PHO8 chimera construct was made by PCR amplification of myc-tagged Affibody_HER2 from a pGEX4T1-ClyA-affibody construct (12), digested with EcoRI and BamHI, and subcloned into pYJ406 (39), yielding pYJ406-Myc-Affibody_HER2. The PHO8 gene was then PCR-amplified from yeast genomic DNA, digested with BamHI and SacII, and inserted into pYJ406-Myc-Affibody_HER2, generating pYJ406-Myc-Affibody_HER2-PHO8. This construct contained a linker (GGGGGGGS) between Myc-Affibody_HER2 and PHO8.

Cell Culture and Spheroid Preparation.

NIH3T6.7 mouse fibroblast cells stably transfected with HER2 (kindly provided by Yong-Min Huh, Yonsei University, Seoul, Korea) (37) were cultured in DMEM supplemented with 10% (vol/vol) FBS (GIBCO) and antibiotics. SKOV3 and MDA-MB-231 cells were cultured in RPMI-1640 medium containing 10% (vol/vol) FBS and antibiotics. All cells were incubated at 37 °C in 5% CO₂. Cell lines were purchased from the Korean Cell Line Bank (KCLB).

For the NIH3T6.7 cell-based spheroid preparation, we used a 96-well plate-based system as previously described (40). Spheroids with a uniform circular morphology and diameter of ∼500 μm were used for the analyses.

Characterization of the Vacuoles.

Protein immunoblotting.

The total protein concentration of the vacuoles was quantified with the Bradford assay. Affi_HER2-protein expression was assessed by Western blotting. Proteins in the lysates were resolved by SDS/PAGE on 12% (vol/vol) gels, followed by immobilization on a PVDF membrane by electrotransfer. Blots were blocked by incubating with 5% (wt/vol) nonfat dry milk for 1 h, probed with primary antibodies against PHO8 and myc tag, and subsequently incubated with HRP-conjugated secondary antibodies. Immunoreactive proteins were visualized with Western blotting detection reagents (GE Healthcare) following the manufacturer’s instructions.

Particle size analysis.

Vacuoles were characterized with respect to size and morphology using an ELS apparatus (Malvern Instruments) and an electron microscope (in situ TEM; JEOL), respectively. For TEM analysis, vacuole samples were dropped onto a copper grid (200 mesh); after drying, the samples were stained with 2% (wt/vol) uranyl acetate (Polysciences, Inc.) for 5–10 min and washed five times with distilled water. Grids were dried in a desiccator and examined under a transmission electron microscope. Vacuoles within yeast cells were visualized by bright-field imaging after staining with FM4/64 lipophilic styryl dye (Invitrogen).

Monitoring HER2 protein specificity with ELISA.

The specificity of Affi_HER2Vacuoles and nontargeting vacuoles for HER2 protein was determined by ELISA after pretreating the vacuoles with PK (0.1 mg/mL), following a previously described protocol with some modifications (12). Briefly, the vacuoles were added to each well and immobilized by incubation for 12 h at 4 °C. The next day, after removal of the coating solution, 200 μL of blocking buffer [2% (wt/vol) skim milk in PBS] was added and the plates were incubated for 2 h at 25 °C. The blocking buffer was removed, and each well was washed six times with PBS containing 0.1% Tween-20 (PBST). HER2 protein (10 μg/mL in PBS; Bender Medsystems, Inc.) was added to each well and incubated for 2 h at 25 °C. Plates were washed six times with PBST, and bound HER2 protein was detected by adding 100 μL of anti-HER2 antibody (Santa Cruz Biotechnology) in PBST containing 2% (wt/vol) skim milk. Plates were washed as above and incubated with HRP-conjugated secondary antibody. Immunoreactivity was visualized by incubating with the chromogenic HRP substrate, 3,3′,5,5′-tetramethybenzidine (BD OptEIA), and the absorbance was measured at 450 nm.

Cell Binding and Uptake Assay.

Cell binding and uptake were monitored according to a previously described protocol (12), with slight modifications. HER2-overexpressing SKOV3 cells and HER2-negative MDA-MB-231 cells were grown on chamber slides in their respective media for 24 h to reach 60% confluence. For the binding assays, cells were fixed with 4% (wt/vol) paraformaldehyde (PFA) (SIGMA ALDRICH) for 15 min, washed three times with PBS, and blocked by incubating with 2% (wt/vol) BSA for 1 h at room temperature. After washing, the cells were coincubated with Affi_HER2Vacuoles (25 μg) for 1 h at 4 °C. Free, unbound Affi_HER2Vacuoles were removed by washing with PBS. Affi_HER2Vacuoles were visualized by incubating samples sequentially with anti-affibody primary antibody (goat polyclonal; Abcam) and FITC-labeled anti-goat secondary antibody (Santa Cruz Biotechnology). All samples were mounted with fluorescence mounting medium (Dako) and examined under a confocal microscope. For cellular uptake assays, cells were grown to 60% confluence in chamber slides and then treated with Affi_HER2Vacuoles (25 μg) in the OPTIMEM media (GIBCO) for 1 h under standard culture conditions. After washing with PBS, the cells were further incubated under standard culture conditions. Nuclei were stained with DRAQ5 (Cell Signaling Technology) following the manufacturer’s instructions. Samples were then washed twice with PBS and fixed with 4% (wt/vol) PFA at 3 h and 6 h posttreatment. Cells were permeabilized by incubation with 0.25% Triton X-100 for 5 min After washing cells with PBS, slides were incubated with anti-affibody primary antibody and FITC-labeled anti-goat secondary antibody, and then coverslip-mounted and examined under a confocal microscope.

Characterization of Drug-Loaded Vacuoles.

Drug loading.

Dox hydrochloride (100 μg, 200 μg, and 300 μg) was mixed with a volume of Affi_HER2Vacuoles containing a fixed amount (100 μg) of total protein, and the final volume was adjusted to 1,000 μL using PS buffer [10 mM Pipes⋅KOH (pH 6.8), 200 mM sorbitol]. Loading was achieved by simple mixing with a magnetic stirrer overnight at 4 °C. Free drug (unloaded) was removed by ultrafiltration with 100-kDa Millipore membranes. Samples were washed six to eight times with cold PS buffer and concentrated to 1 mL. Vacuoles were then washed twice with cold PS buffer by centrifugation (for 10 min at 10,000 rpm) (Hanil Science Co., Ltd; Rotor No. 1). Before the fluorescence intensity measurements, vacuoles were lysed with a probe sonicator (10 times for 5 s each time at an interval of 10 s) at an amplitude setting of 36. After centrifuging the lysed samples at 10,000 rpm (Hanil Science Co., Ltd; Rotor No. 1) for 15 min, the supernatant was collected to estimate the fluorescence. The loading efficiency was determined by measuring the fluorescence intensity of Dox at 480 nm. For all further experiments, after a brief centrifugation, the concentrated vacuoles were resuspended in an appropriate buffer or medium before use for in vitro or in vivo treatment.

Drug release kinetics.

The drug efflux was analyzed according to a previously described protocol (41), with slight changes. Briefly, Dox-loaded vacuoles were transferred to a dialysis tube [Float-A-Lyzer (Spectrum Laboratories, Inc.); molecular mass cutoff of 100 kDa] and immersed in 10 mL of PBS (pH 7.4) or acetate saline buffer (pH 5.1) at 37 °C with gentle stirring. At predefined time points, buffer was collected for fluorescence analysis and replaced with fresh buffer solution. The amount of Dox released was determined by fluorescence analysis with excitation at 480 nm.

In Vitro Drug Delivery and Cytotoxicity.

Intracellular trafficking studies of drug-loaded Affi_HER2Vacuole^Dox were performed with NIH3T6.7 cells cultured in DMEM containing 10% (vol/vol) FBS and antibiotics on gelatin-coated coverslips in 12-well dishes. After the cells reached 70% confluence, the medium was replaced with OPTIMEM medium, and cells were incubated for 30 min. Affi_HER2Vacuole^Dox was then added, and the cells were incubated for 1 h. The cells were then washed twice with PBS, and fresh DMEM was added containing 10% (vol/vol) FBS and antibiotics. At specific coincubation time points (1 h, 12 h, 24 h, and 48 h), cells on the glass slides were washed with PBS, fixed with 4% (wt/vol) PFA, and coverslip-mounted using fluorescence mounting medium (Dako). All samples were examined with a confocal microscope equipped with 40× and 60× lenses. Cytotoxicity assays using 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide were performed as previously described (42).

For the transcytosis and penetration analyses, NIH3T6.7 cells were cultured for 1 wk in six-well plate Transwell supports (0.4-μm polyester membrane, 24 mm insert; Costar); under these conditions, the cells grew beyond confluence, yielding multiple layers of cells. Cells in the Transwell supports were treated for 2 h with Affi_HER2Vacuole^Dox; after washing, they were further incubated in a six-well plate containing NIH3T6.7 cell monolayers. The penetration and uptake of Affi_HER2Vacuole^Dox into the NIH3T6.7 cell monolayers were monitored at different time points (12 h, 24 h, and 48 h) by confocal microscopy and compared with the controls. The NIH3T6.7 spheroid-based penetration and cytotoxicity studies were done by treating the spheroids with Free^Dox, Affi_HER2Vacuole^Dox, or Vacuole^Dox (25 μM Dox) in OPTIMEM media for 2 h under standard culture conditions. After washing with PBS, the spheroids were further incubated in standard medium and analyzed for drug uptake, penetration, and cytotoxic effects at fixed time points (6 h, 12 h, 24 h, 48 h, and 96 h) with confocal (10×) and bright-field microscopes.

Xenograft Model, In Vivo Targeting, Biodistribution, and Pharmacokinetic Analysis.

In vivo targeting and biodistribution studies were done with tumor-bearing, 6-wk-old, female BALB/c nude mice, and pharmacokinetic analysis was done in tumor-free, 6-wk-old, female ICR mice (Orient Bio, Inc.). All animal experiments were done according to established guidelines and with the approval of the KAIST-Institutional Animal Care Committee (KA2013-25). Sample sizes were chosen based on guidance from the literature. Investigators were not blinded to the identity of the groups. HER2-overexpressing NIH3T6.7 cells (1 × 10⁷) in 100 μL of 1:1 PBS/Matrigel (BD Biosciences) were implanted s.c. into the flanks of nude mice. Tumor volume, calculated as (length × width²)/2, was monitored at regular intervals and expressed as the group mean ± SD. For in vivo cancer targeting, mice were given a single injection of Dox (1 mg/kg) or an equivalent amount of Affi_HER2Vacuole^Dox via the tail vein (n = 3). Tissue homing and retention were monitored at 6 h and 12 h postinjection with a confocal imaging system (ZEISS LSM 510). In the biodistribution studies, drug treatment was initiated when tumor volumes reached ∼300 mm³; mice were injected with Dox at a dose of 1 mg/kg or an equivalent amount of Affi_HER2Vacuole^Dox via the tail vein (n = 3). Six hours after the blood was collected (retroorbitally), animals were killed and the tumors and vital organs (liver, kidney, spleen, heart, and lung) were harvested. Tissues were homogenized in 1 mL of tissue lysis buffer with an IKA T10 basic homogenizer. Drug extraction was done by incubating samples in methanol and 12 mM phosphoric acid with vigorous vortexing at 15 °C. Samples were centrifuged at 10,000 × g for 8 min at 4 °C, and the supernatant was filtered through 0.2-μm syringe filters (Whatman). The concentration of Dox was determined by HPLC analysis with a fluorescence detector (Agilent Technologies). The percentage of injected dose and percentage of injected dose per gram of tissue were calculated. The pharmacokinetic analysis was done as previously described, with slight changes to the protocol (43). Briefly, mice were injected with PBS, Dox at a dose of 1 mg/kg, or an equivalent amount of Affi_HER2Vacuole^Dox via the tail vein (n = 5). At fixed time points (10 min and 1 h, 3 h, 6 h, 12 h, and 24 h), blood was collected (retroorbitally) and plasma was separated and processed as mentioned above in the biodistribution study to extract the Dox. Each sample was loaded onto a 96-well plate for fluorescence determination (excitation at 485 nm, emission at 590 nm). The concentration of Dox at various time points was obtained by comparing the fluorescence with a calibration curve generated from known amounts of Dox. Pharmacokinetic parameters, such as t_1/2 and AUC, were calculated by fitting the plasma Dox concentrations to a noncompartment model with the WinNonlin 5.2.1 pharmacokinetic software package.

In Vivo Antitumor Effects in NIH3T6.7 Xenografts.

For the in vivo studies, NIH3T6.7 xenografts were prepared. All in vivo experiments were done according to established guidelines and with the approval of the KAIST-Institutional Animal Care Committee (KA2013-25). Sample sizes were chosen based on guidance from the literature. Investigators were not blinded to the identity of the groups. Ten to 12 d after cancer implantation, animals were divided into five groups based on body weight and tumor volume (n = 5 mice per group): (i) vehicle control (PBS), (ii) empty Affi_HER2Vacuole, (iii) Free^Dox, (iv) nontargeted Vacuole^Dox, and (v) Affi_HER2Vacuole^Dox (Dox at a dose of 1 mg/kg). Mice were injected via the tail vein every other day for a total of eight injections. Tumor volumes and body weights were monitored throughout the course of the treatment, and tumor growth inhibition (TGI) was determined on the final day as TGI = 100% × (T_vol^Control − T_vol^Treated)/T_vol^Control, where T_vol = final tumor volume − initial tumor volume (44). Differences with a P value <0.05 were considered significant. Tumors were collected and sectioned for analysis of apoptotic cells by TUNEL assay (BioVision) according to the manufacturer’s instructions.

HUVEC Cytotoxicity Study.

Cytotoxicity was monitored in HUVECs by analyzing apoptosis with immunostaining. HUVECs were cultured in a six-well dish in M199 medium containing 15% (vol/vol) serum (GIBCO). Both forms of the vacuoles (WT and Affi_HER2-modified) were added to the cells at a concentration of 100 μg/mL and incubated for 4 h. After washing with PBS, cells were incubated for 24 h under standard culture conditions. Cells were further stained with the molecular probes calcein AM and ethidium homodimer (Invitrogen) following the manufacturer’s instructions, and were then examined under a confocal microscope to determine the live and dead cells. The protocol was validated by treating one group with 70% (vol/vol) methanol for 30 min before staining and analysis as above.

RAW 264.7 Cell Stimulation Assay.

The RAW 264.7 cell line, obtained from the KCLB, was stimulated and assayed as previously described (45), with minor changes. Briefly, cells grown in DMEM containing 10% (vol/vol) FBS (GIBCO) were activated by incubation with 100 ng/mL LPS or were treated with vacuoles (10 μg/mL and 50 μg/mL, WT or Affi_HER2-modified) for 2 h. After replacing media with fresh media, cells were further incubated under standard culture conditions for 24 h and TNF-α was quantified by ELISA (Thermo Scientific) following the manufacturer’s instructions.

Serum Cytokine Analysis.

Animal studies were done following the approval and guidelines of the KAIST-Institutional Animal Care Committee, using 6-wk-old female C57BL/6 mice (Orient Bio, Inc.) as previously described (12), with minor changes. Animals were divided into three groups (n = 5 mice per group): (i) vehicle control (PBS), (ii) vacuoles, and (iii) Affi_HER2Vacuoles (100 μg and 200 μg each). Vacuoles were administered i.v. for 4 consecutive days. Thereafter, blood was collected retroorbitally at 2 h and 24 h to monitor early and delayed responses, respectively. The serum cytokines TNF-α, IL-6, and IFN-γ were quantified by analyzing samples at 450 nm with the respective ELISA kits (R&D Systems) following the manufacturer’s instructions. Investigators were not blinded to the identity of the groups.

Statistics.

Statistical analyses of the data were done with the SPSS 18.0 program by applying the nonparametric Kruskal–Wallis and Mann–Whitney U tests. Results are expressed as the mean ± SD, and differences are considered statistically significant for P values <0.05.

Conclusion

In conclusion, the proposed approach suggests a platform technology for generating cancer-specific, tissue-penetrating, safe, and scalable biological nanoparticles from a nonhazardous yeast strain as a potential targeted cancer therapy. The developed nanocarriers can be generated on a large scale in a cost-effective manner by optimizing the fermentation and purification methods. The results showed significant tumor growth inhibition due to active drug delivery, enhanced biodistribution within tumors, and minimal toxic side effects using Affi_HER2Vacuole^Dox, suggesting that bioengineered vacuoles have potential for use as drug delivery carriers in treating cancers. In the future, we plan to investigate the possibility of loading diverse therapeutic agents, including toxins, within vacuoles and testing their fate, both in vitro and in vivo.

Materials and Methods

Detailed methods on the following subjects are available in SI Materials and Methods: yeast strains, plasmid construction, and vacuole isolation; cell culture and spheroid preparation; characterization of the vacuoles; cell binding and uptake assay; characterization of drug-loaded vacuoles; in vitro drug delivery and cytotoxicity; xenograft model, in vivo targeting, biodistribution and pharmacokinetic analysis; in vivo antitumor effects in NIH3T6.7 xenografts; HUVEC cytotoxicity study; RAW 264.7 cell stimulation assay; serum cytokine analysis; and statistical analyses. All animal experiments were done according to established guidelines and with the approval of the KAIST-Institutional Animal Care Committee (KA2013-25).