Ferucarbotran, a carboxydextran-coated superparamagnetic iron oxide nanoparticle, induces endosomal recycling, contributing to cellular and exosomal EGFR overexpression for cancer therapy

Abstract

Superparamagnetic iron oxide (SPIO) nanoparticles have shown many impacts on stem cell attributes when they are used as labels for cellular magnetic resonance imaging (MRI) in the application of stem cell-based therapy. Although it is plausible that iron ions from the lysosomal degradation of SPIO nanoparticles are one of the possible candidates, the mechanisms underlying SPIO-induced cellular responses remain unclear. Herein, the mechanism of ferucarbotran, an ionic SPIO, for the regulation of epidermal growth factor receptor (EGFR) expression in human mesenchymal stem cells (hMSCs) is explored. Ferucarbotran can be internalized into EGFR-localized endosomes, and the endosomal EGFRs in ferucarbotran-labeled hMSCs, compared to unlabeled cells, are mainly localized on the early endosomes and recycling endosomes, but not on late endosomes/lysosomes, and thus escape from lysosomal degradation. Afterward, the recycling endosomal EGFRs are transferred to the cellular membrane and extracellular exosomal vesicles (exosomes) through back fusion and a secretory pathway, respectively, resulting in EGFR-overexpressed hMSCs and EGFR-overexpressed exosomes. Moreover, as EGFR-overexpressed hMSCs, EGFR-overexpressed exosomes can more effectively capture tumorous EGF than native exosomes, which may contribute to the inhibition of tumor growth. This is the first report to find that the SPIO nanoparticles have an impact on stem cell attributes through inducing endosomal recycling instead of them undergoing lysosomal degradation.

---

Introduction

Stem cell-based therapy is a rapidly evolving research in the area of regenerative medicine and cancer. Being able to effectively and noninvasively track the fate and distribution of transplanted cells in vivo is crucial for the improvement of stem cell therapy. Recently, superparamagnetic iron oxide (SPIO) nanoparticles have appeared to be the most feasible probe to label stem cells for cellular magnetic resonance imaging (MRI), which has been proposed as the most attractive modality for repeated noninvasive monitoring of magnetically labeled transplanted stem cells. Although SPIO nanoparticles are generally believed to be biocompatible, the impacts of SPIO nanoparticles on stem cell attributes remain unresolved.^1,2 Furthermore, the increasing evidence in the literature has demonstrated that cellular responses to SPIO labeling are indeed observed,^3–5 suggesting the need for a more comprehensive survey of the biosafety implications of using SPIO nanoparticles to track stem cells in MRI. For instance, in our previous study, ferucarbotran (Resovist), an ionic SPIO nanoparticle with a carboxydextran coating, was shown to be able to stimulate the proliferation of human mesenchymal stem cells (hMSCs) from bone marrow,³ which prompted us to undertake a trilogy of studies.^6–8

First, ferucarbotran was shown to activate the migration and therefore to prohibit the osteogenesis of hMSCs,⁶ which raised some concerns about using ferucarbotran to label stem cells for osteogenic MRI. Second, ferucarbotran was also shown to be capable of promoting the migration of hMSCs toward glioma cells/gliomas in vitro and in vivo, which was attributed to the ferucarbotran-increased expression of the chemokine receptor CXCR4 in hMSCs.⁷ Moreover, in the third episode,⁸ when using ferucarbotran-labeled hMSCs, we found an overexpression of epidermal growth factor receptor (EGFR), which was shown to be associated with the tropism of MSCs for tumors,⁹ and also demonstrated that in a colon cancer model, ferucarbotran-labeled hMSCs exerted more effective antitumor activity than unlabeled hMSCs due to the stronger tumor tropism of EGFR-overexpressed hMSCs. These data demonstrate that the impacts of ferucarbotran on the expressions of tumor tropism receptors are not necessarily harmful, but can render stem cells more favorable for cancer therapy.

After their internalization into cells, SPIO nanoparticles can be transferred to lysosomes in which lysosomal degradation of the SPIO nanoparticles may occur, and then free iron (Fe) is believed to be capable of releasing itself into the cytoplasm.^2,6 Because iron is well known to play important roles in diverse cellular events, it is likely that Fe from the lysosomal degradation of ferucarbotran may be vitally important for ferucarbotran-mediated cellular impacts. Although Fe has been suggested to be involved in ferucarbotran-promoted migration and the related osteogenesis inhibition of hMSCs because of the antagonism effect of desferrioxamine (DFO), an iron chelator, in the abovementioned events,⁶ it was observed that the ferucarbotran up-regulated the EGFR protein level but not the gene transcript,⁸ in contrast to how DFO affects cellular events through expression of numerous genes,^10,11 implying the possibility that the alternatives more dominant than Fe may be responsible for the ferucarbotran-induced EGFR overexpression. To ensure that the possible benefits of the ferucarbotran-promoted tumor tropism of hMSCs in stem cell-based cancer therapy are maximized, it is essential to explore how ferucarbotran increases the expressions of tumor tropism receptors. Therefore, in this study we explored whether ferucarbotran could affect EGFR expression through a distinct mechanism(s) from the lysosomal degradation of ferucarbotran.

Results and discussion

Colocalizations of EGFRs with early endosomal markers in hMSCs

As shown in Fig. 1a, by incubating hMSCs with an equally relevant concentration of iron as free iron in the form of FeCl₃ or ferucarbotran nanoparticles, both FeCl₃ and ferucarbotran up-regulated EGFR protein expression; however, FeCl₃ increased EGFR gene-expression but ferucarbotran did not, suggesting that ferucarbotran would induce EGFR overexpression through other mechanism(s) distinct from gene up-regulation by Fe from the lysosomal degradation of ferucarbotran. Without regard to the contributing roles of several steps behind the transcription in the gene-expression process for the synthesis of proteins, the most likely pathway responsible for ferucarbotran-induced EGFR overexpression would be a decreased EGFR protein degradation. EGFR is a transmembrane glycoprotein that can be internalized through endocytosis starting with early endosomes (EEs), which are incorporated into the intraluminal vesicles (ILVs) within the multivesicular bodies (MVBs). Under most physiological conditions, MVBs (late endosomes) fuse with lysosomes, wherein liganded EGFR are destined for degradation. In contrast, for the constitutive internalization of the unliganded EGFR, MVBs move through the recycling pathway to the cell periphery and fuse with the plasma membrane, leading not only to the incorporation of both EGFR and the peripheral membrane of the MVBs back into the plasma membrane, but also to the release of the ILVs as exosomes.^12–14 Therefore, we assumed that ferucarbotran would be internalized into EGFR-loaded endosomes and would affect the trafficking and fate of endocytic EGFR.

Fig. 1 Ferucarbotran induces more localization of EGFRs to early endosomes in hMSCs. (a) The expression profiles of EGFR proteins and mRNAs in unlabeled hMSCs (control), ferucarbotran-labeled hMSCs (ferucarbotran), and iron(III) chloride-treated hMSCs (iron chloride). β-actin was the internal control. Data from the densitometry analysis for the relative levels of EGFR proteins and mRNAs are expressed as the mean ± standard error of three separate experiments (*p < 0.05 as compared with control). (b and c) The hMSCs were treated without (hMSCs) or with 300 μg mL⁻¹ ferucarbotran (*hMSCs) for 1 h and then fixed, permeabilized, and co-stained for indirect immunofluorescence using an EGFR mouse antibody and Rab5 rabbit antibody (b) or an EGFR rabbit antibody and EEA-1 mouse antibody (c). EGFR staining is shown in red; the marker proteins for the early endosomes (Rab5 and EEA-1) are shown in green. Colocalization masks of EGFR and vesicular markers (white dots as indicated with the white hollow arrow) were collected with an Olympus FV10i microscope. Shown are representative images from three independent experiments. Scale bar = 10 μm. For the large-scale image, please see the ESI, Fig. S2.†

To explore our assumption, we used indirect immunofluorescence to probe the colocalization between EGFR and markers of the vesicular trafficking pathway for describing the compartments in which the EGFR were distributed. Rab proteins, a large family of small molecular weight guanine nucleotide binding proteins (G-proteins), are localized to distinct intracellular organelles of both the endocytic and exocytic pathways and represent key regulators of the vesicular trafficking, and therefore implicated in the trafficking control of EGFR. Therefore, many Rab proteins have been widely used as markers for intracellular vesicular trafficking for the assignment of transport routes and the steps of EGFR. EGFR was visualized using Alexa Fluor 488-labelled antibodies (red), whereas Rab proteins were visualized with Alexa Fluor 594-labelled antibodies (green). Rab5 has a role in mediating the entry of EGFR from the cell surface into the early endosome.^12,15,16 Rab5 as well as early endosome autoantigen-1 (EEA1) were therefore used to identify the early endosomes. As seen in Fig. 1b and c, in the control (unlabeled) hMSCs, the EGFR exhibited basal colocalization with Rab5 and EEA1, indicating the internalization of EGFR into the early endosomal compartment. However, the internalized EGFR showed more colocalization with Rab5 (Fig. 1b) and EEA1 (Fig. 1c) in ferucarbotran-labeled hMSCs than in the control cells, suggesting that either ferucarbotran accelerated the internalization of EGFR into early endosomes or, more likely that ferucarbotran altered the routine trafficking of EGFR from the early endosomes into the next endocytic compartments.

Colocalizations of EGFRs with vesicular markers for the late endosomes/lysosomes and recycling endosomes

Accordingly, we compared the activation status and the protein levels of EGFR between EGF-stimulated hMSCs and ferucarbotran-treated hMSCs. As shown in Fig. 2a, EGF stimulation caused EGFR degradation as well as EGFR phosphorylation compared to the control; however, ferucarbotran could induce the phosphorylation but not the degradation of EGFR. Instead of the degradation of EGFR, ferucarbotran induced the accumulation of EGFR (also shown in Fig. 1a). These results suggest that ferucarbotran may be internalized through EGFR and colocalized with EGFR inside the identical endosomes. Furthermore, we observed the colocalization of internalized EGFR with Rab7, the molecular motor protein that is localized to late endosomes, is responsible for the transport of EGFR-loaded early endosomes to lysosomes for degradation^17–19 and with lysosomal-associated marker protein (LAMP1) to examine whether ferucarbotran could affect the trafficking of EGFR from the early endosomes into late endosomes/lysosomes. Interestingly, ferucarbotran labeling resulted in a significantly diminished entry of internalized EGFR from the early endosomes into late endosomes and lysosomes, because the internalized EGFR was less colocalized with Rab7 (Fig. 2b) and LAMP1 (Fig. 2c) in ferucarbotran-labeled hMSCs than in the control cells, suggesting that there was less EGFR degradation in ferucarbotran-labeled hMSCs than in unlabeled hMSCs. Based on that the internalized EGFRs are either degraded through late endosomes/lysosomes or recycled back to the plasma membrane through the recycling pathway; therefore, we next examined the colocalization of internalized EGFR with Rab4 and Rab11, two distinct markers for recycling pathways, also known as fast (Rab4) and slow (Rab11) recycling.²⁰ Rab4 mediates fast endocytic recycling directly from early endosomes, whereas Rab11 mediates slow endocytic recycling through recycling endosomes and has been shown to be able to facilitate EGFR recycling to the plasma membrane.^21,22 As shown in Fig. 2d, there was a mild but significant difference in the colocalization of internalized EGFR with Rab4 in unlabeled hMSCs and ferucarbotran-labeled hMSCs, suggesting a contributing role of Rab4 in the ferucarbotran-mediated endocytic recycling of EGFR, whereas a dramatic increase in the colocalization of internalized EGFR with Rab11 was observed in ferucarbotran-labeled hMSCs (Fig. 2e), indicative of a facilitated transport of internalized EGFR to Rab11-postive recycling endosomes by ferucarbotran. We also quantified the percentages of the colocalizations of internalized EGFR with vesicular trafficking markers (Fig. 2f). Taken together, these data suggest that ferucarbotran can orientate EGFR-loaded early endosomes to Rab4 and Rab11-postive recycling endosomes, thus protecting against the transfer of internalized EGFR to late endosomes/lysosomes for degradation, resulting in the overexpression of EGFR protein without gene transcript.

Fig. 2 Ferucarbotran orientates EGFR-loaded early endosomes to recycling endosomes and protects against the transfer of internalized EGFR to late endosomes/lysosomes for degradation. (a) The expression profiles of phosphorylated EGFR and total EGFR proteins in untreated hMSCs (control; white bar), EGF-treated hMSCs (EGF; grey bar), and ferucarbotran-labeled hMSCs (ferucarbotran; black bar). β-actin was the internal control. Results of the western blot analysis are the representative of the three separate experiments. Data of densitometry analysis for the relative levels of p-EGFR and EGFR proteins are expressed as the mean ± standard error of six separate experiments (*p < 0.05 as compared with control). (b–e) The hMSCs were treated without (hMSCs) or with 300 μg mL⁻¹ ferucarbotran (*hMSCs) for 1 h and then fixed, permeabilized, and co-stained for indirect immunofluorescence using an EGFR mouse antibody and a Rab7 rabbit antibody (b), an EGFR rabbit antibody and a LAMP1 mouse antibody (c), an EGFR mouse antibody and a Rab4 rabbit antibody (d), or an EGFR mouse antibody and a Rab11 rabbit antibody (e). EGFR staining is shown in red; the marker proteins for the late endosomes/lysosomes (Rab7 and LAMP1) and for the recycling pathway (Rab4 and Rab11) are shown in green. Colocalization masks of EGFR and vesicular markers (white dots as indicated with the white hollow arrows) were collected with an Olympus FV10i microscope. Herein shown are the representative images from three independent experiments. Scale bar = 10 μm. (f) The percentages of colocalizations of internalized EGFR with vesicular trafficking markers were measured using Olympus FV10i software version 4.1. *p < 0.05 as compared *hMSCs (black bar) with hMSCs (grey bar). For the large-scale image, please see the ESI, Fig. S3.†

The contribution of ferucarbotran-induced EGFR recycling to the cellular membrane and the TEM observation of ferucarbotran trafficking

In our previous report, we attributed the more potent antitumor activity of ferucarbotran-labeled hMSCs to the fact that hMSCs with ferucarbotran-induced EGFR overexpression could more effectively migrate toward tumors and capture tumorous EGF; however, we only examined the total cellular form and not the membranous form of the ferucarbotran-induced EGFR. Although the recycling endosomal EGFR inside cells can contribute to total cellular EGFR expression, they cannot interact with extracellular tumorous EGF, which prompts the question as to whether the recycling endosomal EGFR could contribute to exerting a tumor tropism and capturing tumorous EGF. As mentioned above, the recycling endosomal EGFR can be either incorporated into the plasma membrane or released into secretory exosomes; therefore, we investigated the contribution patterns of the recycling endosomal EGFR in ferucarbotran-induced EGFR overexpression. Using flow cytometry, we observed an increased expression of membranous EGFR in ferucarbotran-labeled hMSCs (Fig. 3a); in addition, using transmission electron microscopy (TEM), we observed ferucarbotran-enclosed vesicles in the cell periphery (Fig. 3b; green arrow) as well as extracellular and free (unenclosed) ferucarbotran adherent to the plasma membrane (Fig. 3b; red arrow). These results suggest that, through back fusion, ferucarbotran enclosed in EGFR-loaded vesicles could be one candidate for contributing the recycling endosomal EGFR into the plasma membrane and that ferucarbotran is then extruded. Interestingly, we also observed the existence of extracellular but enclosed ferucarbotran (Fig. 3c; green arrow), which suggests that ferucarbotran can be extruded through being enclosed in the secretory exosomes. Therefore, we wondered whether the recycling endosomal EGFR could be released into ferucarbotran-enclosed exosomes.

Fig. 3 Cellular membranous EGFR expression and the TEM observation of ferucarbotran trafficking. (a) Cell membranous EGFR expression was measured by flow cytometry. The hMSCs were treated without (hMSCs) or with 300 μg mL⁻¹ ferucarbotran (*hMSCs) for 1 h and then immediately processed for the flow cytometry study. The semi-quantified membranous EGFR expression was represented as the relative mean fluorescence intensity of the indirect immunofluorescent staining of membranous EGFR of *hMSCs compared with that of hMSCs as the control. Data are expressed as the mean ± standard error of three independent experiments (*p < 0.05 as compared with hMSCs). (b and c) TEM images of ferucarbotran distribution after cellular internalization. The hMSCs were treated with 300 μg mL⁻¹ ferucarbotran for 1 h and then immediately processed for TEM observation. (b) Ferucarbotran-enclosed vesicles (green arrow) were observed in the cell periphery; extracellular and free (unenclosed) ferucarbotran (red arrow) adherent to the plasma membrane was visible. (c) Extracellular but enclosed ferucarbotran (green arrow) was observed.

Examination of the exosomes and the contribution of ferucarbotran-induced EGFR recycling to the exosomes

As shown in Fig. 4, exosomes isolated from control hMSCs (Fig. 4a) and ferucarbotran-labeled hMSCs (Fig. 4b) were observed by TEM. The mean diameter of the exosomal nanoparticles obtained from the control hMSCs was 116.5 ± 9.7 nm, whereas it was 98.1 ± 1.5 nm in the case of ferucarbotran-labeled hMSCs. The zeta potentials of the exosomes from the control hMSCs and ferucarbotran-labeled hMSCs were about −21.9 ± 0.3 and −25.1 ± 5.4 mV, respectively. No significant difference in either particle size or surface charge between the control hMSC-generated exsosomes and ferucarbotran-labeled hMSC-generated exosomes were observed. To verify the presence of ferucarbotran enclosed in exosomes, as observed in Fig. 3c, energy-dispersive X-ray (EDX) spectroscopy was used, where the selected area EDX elemental analysis showed that Fe was significantly detected in the exosomes from ferucarbotran-labeled hMSCs (Fig. 4d) but not in those from the control hMSCs (Fig. 4c), which confirmed that ferucarbotran was enclosed in the exosome. Therefore, we then compared the EGFR expression of control hMSC-generated exosomes with that of ferucarbotran-labeled hMSC-generated exosomes. Exosomal protein markers, such as CD81 and flotillin-1,^23,24 were identified in both types of exosomes. We demonstrated that the expression of EGFR was increased in the exosomes from ferucarbotran-labeled hMSCs compared with that from the control hMSCs (Fig. 4e). These results suggest that ferucarbotran also facilitates the contribution of recycling endosomal EGFR to the ferucarbotran-enclosed exosomes, resulting in EGFR-overexpressed exosomes.

Fig. 4 Examination of exosomal EGFR and ferucarbotran and the captured binding of EGF by exosomal EGFR. (a and b) TEM images of exosomes. The hMSCs were treated without or with 300 μg mL⁻¹ ferucarbotran for 1 h and then grown in a regular cultured medium for 48 h. Exosomes in the cell culture supernatant from unlabeled hMSCs (hMSCs-exosomes) (a) or ferucarbotran-labeled hMSCs (*hMSCs-exosomes) (b) were collected and processed for TEM observation. (c) EDX measurement of selected area of (a). (d) EDX measurement of selected area of (b). (e) The protein expression profiles of EGFR and exosomal markers (CD81 and flotillin-1) in unlabeled hMSC-derived exosomes (exosome from hMSCs) and ferucarbotran-labeled hMSCs-derived exosomes (exosome from *hMSCs). β-actin was the internal control. The densitometry analysis for the relative level of EGFR of three separate experiments is indicated under each protein band (*p < 0.05 as compared with exosome from hMSCs). (f) The capture binding of tumorous EGF by unlabeled hMSCs-derived exosomes (exosome from hMSCs) and ferucarbotran-labeled hMSCs-derived exosomes (exosome from *hMSCs) at 16.7 and 50 μg, respectively (* p < 0.05 as compared with control; # p < 0.05 between two indicated groups).

Binding capture of EGF by ferucarbotran-induced exosomal EGFR and its effect on tumor growth

The abovementioned findings—that the ferucarbotran-facilitated endosomal recycling of EGFR resulted in EGFR-overexpressed hMSCs and EGFR-overexpressed exosomes, and that EGFR-overexpressed hMSCs (ferucarbotran-labeled hMSCs) performed better antitumor activity than control hMSCs (unlabeled hMSCs) through more potent capture binding of tumorous EGF⁸—prompted us to investigate whether EGFR-overexpressed exosomes from ferucarbotran-labeled hMSCs could exert stronger antitumor activity than control exosomes from unlabeled hMSCs. We first examined the binding capture activity of EGFR-expressed exosomes. As shown in Fig. 4f, EGF secreted by 10⁶ HT-29 colon cancer cells into the culture supernatant (6 mL) for 48 h was measured as the control. After the incubation of the tumor cell culture supernatant with exosomes derived from unlabeled hMSCs (exosome from hMSCs in Fig. 4f) or ferucarbotran-labeled hMSCs (exosome from *hMSCs in Fig. 4f), the amount of EGF in the supernatant was significantly decreased (Fig. 4f). Irrespective of whether the exosomes were derived from unlabeled hMSCs or ferucarbotran-labeled hMSCs, the supernatant treated with a higher dose of exosomes showed a greater decrease of EGF content, suggesting an exosome (or EGFR) dose-dependent capture binding of EGF (p < 0.0001). Moreover, at the same dose of exosome treatment, the amount of EGF decreased in the supernatant treated with the exosomes derived from ferucarbotran-labeled hMSCs than in the supernatant treated with exosomes derived from unlabeled hMSCs (p < 0.0001, adjusted R squared = 0.76), which strongly suggests that EGFR-overexpressed exosomes from ferucarbotran-labeled hMSCs can capture tumor-secreted EGF more potently than EGFR-expressed exosomes from unlabeled hMSCs. To further assess the role of the EGF-capture binding activity of EGFR-expressed or EGFR-overexpressed exosomes in tumor growth in vivo, we established tumor models in nude mice by subcutaneously injecting HT-29 cells alone or HT-29 cells mixed with unlabeled hMSC-derived or ferucarbotran-labeled hMSC-derived exosomes. Although the statistical significance of the antitumor activity of unlabeled hMSC-derived exosomes or ferucarbotran-labeled hMSC-derived exosomes was not clearly shown in our small-scale animal experiment (ESI Fig. S1†), a tendency of stronger tumor growth inhibition was indeed observed in the treatment of ferucarbotran-labeled hMSC-derived exosomes than in unlabeled hMSC-derived exosomes, suggesting a positive correlation between the EGFR expression level and the antitumor activity of the exosomes. Our in vivo results suggest that EGFR-overexpressed exosomes derived from ferucarbotran-labeled hMSCs may exert a superior contribution toward antitumor impact than native exosomes derived from unlabeled hMSCs. Furthermore, in vivo and in vitro studies are needed to confirm this hypothesis.

Experimental section

Materials

The mouse anti-EGFR (610016) antibody was obtained from BD BioSciences. Mouse anti-EGFR (sc-120), rabbit anti-EGFR (sc-03), rabbit anti-Rab4A (sc-312), rabbit anti-Rab7 (sc-10767), rabbit anti-Rab11 (sc-25510), mouse anti-EEA-1 (sc-137130) and mouse anti-LAMP-1 (sc-17768) antibodies were from Santa Cruz, CA. Rabbit anti-EGFR (100448), rabbit anti-phosphorylated EGFR (phosphor Tyr1092; 61353), rabbit anti-Rab5 (13253) and anti-β-actin (109639) antibodies were obtained from Genetex. Alexa Fluor 488-goat-anti-rabbit (A11008), Alexa Fluor 594-goat-anti-rabbit (A11012), Alexa Fluor 488-goat-anti-mouse (A11001), Alexa Fluor 594-goat-anti-mouse (A11005) antibodies and 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, D1306) were from Molecular Probes. Horseradish peroxidase (HRP)-conjugated anti-rabbit (#7074S) and anti-mouse (A#7076S) antibodies were obtained from Cell Signaling Technologies. Recombinant human EGF (rhEGF) was procured from R&D. Iron(III) chloride hexahydrate (FeCl₃·6H₂O) was purchased from Sigma-Aldrich. Superparamagnetic iron oxide nanoparticles: ferucarbotran (Resovist, Bayer Schering Pharma AG, Berlin, Germany) consists of SPIO nanoparticles coated with carboxydextran, which offers the complex a net negative charge and ensures stable dispersion of the nanoparticles within aqueous.

Cell culture

Human mesenchymal stem cells (hMSCs) were isolated from the bone marrow of normal donors, with informed consent approved according to the procedures of the institutional review board (EC1021003-E, NHRI, Taiwan). The bone marrow aspirate was added to low-glucose Dulbecco's modified Eagle's medium (DMEM; Gibco) containing 25 U mL⁻¹ heparin in a 1 : 1 ratio, fractionated by Ficoll-Paque density gradient centrifugation. The hMSCs-enriched low-density fraction was collected, rinsed with DMEM, and plated in T25 flasks at 5 × 10⁷ nucleated cells per flask in 5 mL regular growth medium, which consisted of low-glucose DMEM supplemented with 10% fetal bovine serum (FBS; HyClone, Logan, UT), 4 mM L-glutamine, 100 U mL⁻¹ penicillin, and 100 μg mL⁻¹ streptomycin (Sigma-Aldrich). During the first two-week incubation for cell adherence and initial expansion, 5 mL of fresh growth medium was added twice weekly for the first week. Then, the medium changes were carried out twice weekly. When the adherent cells reached ∼60%–70% confluence, they were detached with 0.25% trypsin–EDTA (ethylenediaminetetraacetic acid; Gibco) and replated at 1 : 3 in regular growth medium to allow for continued passaging. All the cultures were kept in an atmosphere of 5% CO₂, 95% air at 37 °C.

RNA extraction and RT-PCR analysis

hMSC cells were seeded at 0.6 × 10⁶ cells per mL in 4 mL regular cultured medium in 60 mm dishes. The hMSC cells were treated with ferucarbotran (300 μg mL⁻¹) or iron chloride (300 μg mL⁻¹) in serum-free medium at 37 °C for 1 h. Total RNA was isolated using an RNeasy micro kit (QIAGEN) according to the manufacturers' instructions. RNA (1 μg) was reversely transcribed with Super Script III (18080-051, Invitrogen) in the presence of oligo-dT primer (2.5 μM), dNTP (0.5 mM), and DTT (0.01 M). PCR was performed using primers as described in a previous report:⁸ hEGFR forward (+): 5′-CGGCGTCCGCAAGTGTAAG and hEGFR reverse (−): 5′-CGGCTGACATTCCGGCAAG. The reaction was performed using a GeneAmp PCR System 9700 (Applied Biosystem) after an initial heating at 95 °C for 5 min, followed by 35 cycles of denaturation at 95 °C for 40 s, annealing at 53 °C for 40 s, and elongation at 72 °C for 1 min, with an additional 7 min incubation at 72 °C after completing the last cycle. The amplified DNA pre-stained with Ezvision three DNA dye buffer (N313-1M, AMRESCO) was loaded onto a 1.5% agarose gel. After electrophoresis, the DNA bands were photographed under UV light. The gel was photographed using an image analyzer (AlphaEase FC 2200, Alpha Innotech). The signal intensity for the EGFR gene product was normalized to their respective β-actin expression.

Western blot analysis

hMSC cells were seeded at 0.6 × 10⁶ cells per mL in 10 mL regular cultured medium in 100 mm dishes. The hMSCs were treated with ferucarbotran (300 μg mL⁻¹) or iron chloride (300 μg mL⁻¹) in serum-free medium at 37 °C for 1 h (Fig. 1a) or treated with ferucarbotran (300 μg mL⁻¹) or EGF (300 μg mL⁻¹) for 5 min (Fig. 2a). After treatment, the cells were rinsed with ice-cold 1 × PBS and were then lysed by the addition of lysis buffer (25 mM HEPES, pH 7.5, 150 mM NaCl, 1% Igepal CA-630, 10 mM MgCl₂, 1 mM EDTA and 2% glycerol, 1 μM phenylmethylsulfonyl fluoride, 1 μg mL⁻¹ leupeptin, and 10 μg mL⁻¹ aprotinin) for 1 h at 4 °C. The suspensions were centrifuged at 15 700g for 20 min at 4 °C. The protein concentration of the supernatant was assessed by the Bio-Rad protein assay kit.

Proteins were separated by electrophoresis in a 10% polyacrylamide gel and transferred to a polyvinylidene difluoride membrane. After incubation at room temperature in 0.1% Tween 20 with TBS (TBST) plus 5% bovine serum albumin (BSA) for 1 h, primary antibodies (mouse anti-EGFR, rabbit anti-pEGFR or rabbit anti-β-actin antibodies) were added to TBST containing 1% BSA and incubated with the membranes at 4 °C overnight. The membranes were then washed three times in TBST, for 5 min each time. After washing, horseradish peroxidase (HRP)-conjugated anti-rabbit (#7074S; 1 : 5000) or anti-mouse (A#7076S; 1 : 5000) antibodies were incubated with the membranes for 1 h at room temperature. After washing, the membranes were developed using the Luminata Cresendo Western HRP Substrate kit (WBLUR0500, Millipore).

Immunofluorescence staining of EGFR with vesicular trafficking markers

hMSC cells were seeded at 0.6 × 10⁶ cells per mL in 400 μL regular cultured medium per well in a 4-well Millicell EZ Slide (Millipore). The hMSCs were treated with ferucarbotran (300 μg mL⁻¹) in serum-free medium at 37 °C for 1 h. After treatment, the cells were rinsed with PBS twice and fixed with 3.5% paraformaldehyde for 20 min, permeabilized with 0.05% saponin/PBS for 5 min, and blocked with 2% BSA/PBS for 2 h. After washing, the cells were incubated at 4 °C with mouse anti-EGFR (sc-120, 1 : 50) antibodies or rabbit anti-EGFR (sc-03, 1 : 50) antibodies with rabbit anti-Rab4A (sc-312, 1 : 25) antibodies, rabbit anti-Rab5 (13253, 1 : 25) antibodies, rabbit anti-Rab7 (sc-10767, 1 : 25) antibodies, rabbit anti-Rab11 (sc-25510, 1 : 100) antibodies, mouse anti-EEA-1 (sc-137130, 1 : 25) antibodies, or mouse anti-LAMP1 (sc-17768, 1 : 25) antibodies overnight. The cells were then washed with 2% BSA/PBS three times and visualized with the secondary antibodies: Alexa Fluor 488-goat-anti-rabbit (A11008, 1 : 100) antibodies, Alexa Fluor 594-goat-anti-rabbit (A11012, 1 : 100) antibodies, Alexa Fluor 488-goat-anti-mouse (A11001, 1 : 100) antibodies, or Alexa Fluor 594-goat-anti-mouse (A11005, 1 : 100) antibodies for 1 h. After washing, the samples were counterstained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) (D1306, Molecular Probes, 1 : 1000) before mounting them onto slides (H-1000, Vector). A colocalization analysis for the dual stained samples (green for vesicular markers and red for EGFR) was carried out using a function of Olympus FV10i software, version 4.1. A scatterplot diagram of the result was set from each control sample as the threshold level of the signal intensity of each channel. Each white dot (as typically indicated with a white hollow arrow) shows the colocalization of both channels. Each condition was performed in duplicate, and six cells were processed to acquire the percentages of the colocalizations of the internalized EGFR with the vesicular trafficking markers. A representative image of each condition was shown.

Membranous EGFR analysis

The semi-quantification of the EGFR expression on the cell membrane was performed using flow cytometry. hMSC cells were seeded at 0.6 × 10⁶ cells per mL in 10 mL regular cultured medium in 100 mm dishes. The hMSCs were treated with ferucarbotran (300 μg mL⁻¹) in serum-free medium at 37 °C for 1 h and then harvested by trypsinization. The cells were fixed with 3.5% paraformaldehyde in PBS (methanol-free) for 10 min at room temperature, followed by 2% BSA/PBS blocking for 20 min at 4 °C. The cells were incubated with monoclonal antibodies against human EGFR (100448, 1 : 100) for 45 min at 4 °C. After washing with PBS, the cells were incubated with the Alexa Fluor 488-conjugated goat anti-rabbit IgG antibodies for 30 min at 4 °C. The fluorescence was analyzed using a FACSCalibur flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA, USA).

Exosome extraction and identification

hMSC cells were seeded at 0.6 × 10⁶ cells per mL in 35 mL regular cultured medium in 150 mm dishes. The hMSCs were treated with ferucarbotran (300 μg mL⁻¹) in serum-free medium at 37 °C for 1 h. After washing with PBS once and following incubation with 35 mL regular cultured medium for 48 h, the cell culture medium was centrifuged for 20 min at 2000g to separate the cellular debris. The clarified supernatant was then concentrated 10-fold by 100 kDa hollow fiber membrane, followed by ultracentrifugation at 100 000g at 4 °C for 1 h. After washing with PBS, the exosomal pellets were either passed through 0.22 μm microcentrifuge filters and stored at −80 °C for the other experiments or the protein contents were determined using a BCA protein assay kit. The expressions of CD81 (sc-31234, 1 : 500), flotillin-1 (sc-74566, 1 : 500), EGFR and β-actin (109639, 1 : 1000) were measured by western blotting analysis.

Transmission electron microscopy and energy-dispersive X-ray (EDX) spectroscopy

Cells. hMSC cells were seeded at 0.6 × 10⁶ cells per mL in 400 μL regular cultured medium on ACLAR® Embedding Film (#50425-25, Electron Microscope Science) per well in 24-well plates. The hMSCs were treated with ferucarbotran (300 μg mL⁻¹) in serum-free medium at 37 °C for 1 h. The harvested cells were fixed in a 2.5% glutaraldehyde solution in PBS and then postfixed in 2% osmium tetroxide (OsO₄) in PBS for 1 h. The cells were washed twice with PBS and then dehydrated in a graded series of ethanol solutions (50%, 70%, 80%, 90%, 95%, and 100%); they were then soaked in a 1 : 1 ratio of 100% acetone and spurr embedding resin (Electron Microscope Science) for 1 h. The cells soaked in spurr resin were then placed at 68 °C for 15 h for polymerization of the resin. The polymerized blocks were sectioned and then visualized under a Hitachi H-7650 transmission electron microscope (Hitachi).

Exosomes. Purified and stored exosomes (see exosome extraction and identification) were fixed with 1% glutaraldehyde in PBS (pH 7.4). A 10 μL drop of the suspension was loaded onto a formvar/carbon-coated grid, negatively stained with 3% (w/v) aqueous phosphotungstic acid for 1 min, and then observed by a Hitachi H-7650 transmission electron microscope (Hitachi). The compositional element analysis of the cell section was measured simultaneously by using energy-dispersive X-ray spectroscopy (EDX, METAK).

EGF ELISA

To examine whether unlabeled hMSCs-derived and ferucarbotran-labeled hMSCs-derived exosome can capture tumorous EGF, HT-29 cells were seeded at 1 × 10⁶ cells per dish in 100 mm dishes with 10 mL of growth medium overnight and allowed the replacement of medium with 6 mL of growth medium to grow for 48 h; then, they were centrifuged at 2000g for 20 min at 4 °C. The HT-29 supernatants (1 mL each sample) were incubated with exosomes isolated from unlabeled hMSCs or ferucarbotran-labeled hMSCs (at 16.7 μg or 50 μg protein of exosomes) for 1 h at 4 °C. Then, the cell supernatants were centrifuged at 16 000g for 30 min at 4 °C and examined with a commercially available enzyme-linked immunosorbent EGF assay kit (ELISA; R&D Systems).

Statistical analysis

The data (except Fig. 4f) are presented as the mean ± standard error of the mean for the indicated numbers of the separate experiments. The results were compared using Student's t test in the case of two groups for comparison. Statistical significance was assigned if the probability value (p) was less than 0.05. In Fig. 4, we used a linear regression analysis for evaluating the efficacy of both the exosomes unlabeled hMSCs and the exosomes for ferucarbotran-labeled hMSCs toward the capture binding of tumorous EGF under the assistance of R 3.1.3 (GNU software for statistical analysis).

Conclusions

In summary, our data suggest that ferucarbotran was internalized into EGFR-localized endosomes and affected the endosomal recycling and the destination of EGFR. Although the mechanisms underlying why ferucarbotran induced the endosomal recycling of EGFR still need to be fully elucidated, it is clear that ferucarbotran-affected endosomal recycling can switch the intracellular traffic of EGFR from late endosomes/lysosomes to recycling endosomes, thereby protecting against the lysosomal degradation of EGFR, and resulting in cellular (membranous) and exosomal EGFR overexpression. Furthermore, the EGF-capture binding activities of ferucarbotran-engineered EGFR-overexpressed hMSCs and exosomes appear to be beneficial for cancer therapy. This study highlights our new thoughts on developing a novel strategy to use the impacts of nanomaterials for the application of nanotechnology to biomedicine.

Acknowledgements

This study was supported by grants from the National Health Research Institutes, Taiwan (grants NM-102-PP-02, NM-103-PP-02 and BN-104-PP-22), and from the Ministry of Science and Technology, Taiwan (102-2314-B-400-008-MY3 and 102-2628-B-303-001-MY3).

References

1. H. Kostura, D. L. Kraitchman, A. M. Mackay, M. F. Pittenger and J. W. Bulte, NMR Biomed., 2004, 17, 513–517.
2. A. S. Arbab, G. T. Yocum, A. M. Rad, A. Y. Khakoo, V. Fellowes, E. J. Read and J. A. Frank, NMR Biomed., 2005, 18, 553–559.
3. D. M. Huang, J. K. Hsiao, Y. C. Chen, L. Y. Chien, M. Yao, Y. K. Chen, B. S. Ko, S. C. Hsu, L. A. Tai, H. Y. Cheng, S. W. Wang, C. S. Yang and Y. C. Chen, Biomaterials, 2009, 30, 3645–3651.
4. H. S. Kim, S. Y. Oh, H. J. Joo, K. R. Son, I. C. Song and W. K. Moon, NMR Biomed., 2010, 23, 514–522.
5. C. Y. Yang, J. K. Hsiao, M. F. Tai, S. T. Chen, H. Y. Cheng, J. L. Wang and H. M. Liu, Mol. Imag. Biol., 2011, 13, 443–451.
6. Y. C. Chen, J. K. Hsiao, H. M. Liu, I. Y. Lai, M. Yao, S. C. Hsu, B. S. Ko, Y. C. Chen, C. S. Yang and D. M. Huang, Toxicol. Appl. Pharmacol., 2010, 245, 272–279.
7. L. Y. Chien, J. K. Hsiao, S. C. Hsu, M. Yao, C. W. Lu, H. M. Liu, Y. C. Chen, C. S. Yang and D. M. Huang, Biomaterials, 2011, 32, 3275–3284.
8. T. H. Chung, J. K. Hsiao, S. C. Hsu, M. Yao, Y. C. Chen, S. W. Wang, Y. P. Kuo, C. S. Yang and D. M. Huang, ACS Nano, 2011, 5, 9807–9816.
9. H. Sato, N. Kuwashima, T. Sakaida, M. Hatano, J. E. Dusak, W. K. Fellows-Mayle, G. D. Papworth, S. C. Watkins, A. Gambotto, I. F. Pollack and H. Okada, Cancer Gene Ther., 2005, 12, 757–768.
10. F. Saletta, Y. S. Rahmanto, E. Noulsri and D. R. Richardson, Mol. Pharmacol., 2010, 77, 443–458.
11. Y. Yu and D. R. Richardson, J. Biol. Chem., 2011, 286, 15413–15427.
12. J. L. Dinneen and B. P. Ceresa, Traffic, 2004, 5, 606–615.
13. J. S. Rush and B. P. Ceresa, Mol. Cell. Endocrinol., 2013, 381, 188–197.
14. P. D. Robbins and A. E. Morelli, Nat. Rev. Immunol., 2014, 14, 195–208.
15. M. A. Barbieri, R. L. Roberts, A. Gumusboga, H. Highfield, C. Alvarez-Dominguez, A. Wells and P. D. Stahl, J. Cell Biol., 2000, 151, 3539–3550.
16. F. Huang, J. Biol. Chem., 2004, 279, 16657–16661.
17. P. Chavrier, R. G. Parton, H. P. Hauri, K. Simons and M. Zerial, Cell, 1990, 62, 317–329.
18. B. P. Ceresa and S. J. Bahr, J. Biol. Chem., 2006, 281, 1099–1106.
19. N. Taub, D. Teis, H. L. Ebner, M. W. Hess and L. A. Huber, Mol. Biol. Cell, 2007, 18, 4698–4710.
20. K. Hofer-Ballmer, A. E. Andersson, L. E. Ratcliffe and P. Berger, Blood, 2011, 118, 816–826.
21. H. Stenmark, Nat. Rev. Mol. Cell Biol., 2009, 10, 513–525.
22. D. N. Cullis, B. Philip, J. D. Baleja and L. A. Feig, J. Biol. Chem., 2002, 277, 49158–49166.
23. J. Conde-Vancell, E. Rodriguez-Suarez, N. Emade, D. Gil, R. Matthiesen, M. Valle, F. Elortza, S. C. Lu, J. M. Mato and J. M. Falcon-Perez, J. Proteome Res., 2008, 7, 5157–5166.
24. A. V. Vlassov, S. Magdaleno, R. Setterquist and R. Conrad, Biochim. Biophys. Acta, 2012, 1820, 940–948.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1 of animal study. Fig. S2 for Fig. 1 and Fig. S3 for Fig. 2 in large scale. See DOI: 10.1039/c5ra18810e

---