Charge-selective fractions of naturally occurring nanoparticles as bioactive nanocarriers for cancer therapy

Acta Biomaterialia 10 (2014) 4269–4284

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Charge-selective fractions of naturally occurring nanoparticles as bioactive nanocarriers for cancer therapy Yongzhong Wang a,b,1, Sijia Yi b,1, Leming Sun a,b, Yujian Huang a,b, Mingjun Zhang a,b,⇑ a b

Department of Biomedical Engineering, Dorothy M. Davis Heart & Lung Research Institute, The Ohio State University, Columbus, OH 43210, USA Department of Mechanical, Aerospace and Biomedical Engineering, The University of Tennessee, Knoxville, TN 37996, USA

a r t i c l e

i n f o

Article history: Received 5 February 2014 Received in revised form 28 May 2014 Accepted 12 June 2014 Available online 19 June 2014 Keywords: Nanoparticle fractionation Immunostimulation Drug delivery Cancer therapy Fungus

a b s t r a c t A carnivorous fungus, Arthrobotrys oligospora, has been shown to secrete nanoparticles. In the present work, the potential of two charge-selective fractions of fungal nanoparticles (FNPs) as bioactive nanocarriers in cancer therapy is explored by investigating their immunostimulatory activities, cytotoxic mechanisms and in vitro immunochemotherapeutic effects. A surface charge-selective fractionation procedure to purify crude FNPs has been established, and two FNP fractions (i.e. FNP1 and FNP2), with different surface charges and similarly reduced diameters of 100–200 nm, are obtained. Both FNP fractions enhance the secretion of multiple proinflammatory cytokines and chemokines from macrophages and splenocytes. However, FNP2 has stronger cytotoxicity than FNP1. It is FNP2 not FNP1 that could clearly inhibit cell proliferation by inducing apoptosis and arresting cells at the sub G0/G1 phase. Both the FNP fractions can form pH-responsive nanocomplexes with doxorubicin (DOX) via electrostatic interactions. For direct cytotoxicity, DOX–FNP2 complexes demonstrate higher activity than DOX against multiple tumor cells, while DOX–FNP1 complexes show weaker activity than DOX. Interestingly, in a co-culture experiment where splenocytes are co-cultured with tumor cells, both DOX–FNP complexes demonstrate higher cytotoxicity than DOX. In conclusion, this work proposes a combined therapeutics for cancer treatment using charge-selective fractions of FNPs as bioactive nanocarriers. Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Cancer is a leading cause of death worldwide, and it is estimated that 13.1 million will die of this disease in 2030 [1]. Chemotherapy is generally regarded as the first-line approach for the treatment of malignant cancer [2,3]. To avoid the emergence of systemic toxicity and therapy resistance, it is essential to develop new treatment modalities with multiple mechanisms of cell killing in tumors, i.e. combined therapy. A few combination therapies using engineered nanoparticle-based delivery systems, including nanoparticles [4,5], liposomes [6–8] and macromolecular conjugates [9] in conjunction with different chemical drugs and immune-stimulants, have been reported [10]. However, among these nanoparticle-enhanced combinatorial therapies, few engineered biomaterials play the role of immunostimulants or adjuvants. They are usually inert biomaterials, simply conjugated or ⇑ Corresponding author at: Department of Biomedical Engineering, Dorothy M. Davis Heart & Lung Research Institute, The Ohio State University, 270 Bevis Hall, 1080 Carmack Rd, Columbus, OH 43210, USA. Tel.: +1 614 292 1625. E-mail address: [email protected] (M. Zhang). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.actbio.2014.06.020 1742-7061/Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

encapsulated with an immunostimulatory agent and a chemodrug in the combined antitumor therapy. Naturally occurring nanoparticles are an alternative source for producing bioactive biopolymer-based nanoparticles with diverse chemophysical properties and biofunctions. The use of naturally occurring organic nanoparticles and biomimetic/bioinspired nanomaterials in medicine has drawn increasing interest in recent years. It is anticipated that study of naturally occurring nanoparticles will provide significant insight into the development of bioactive nanomaterials for cancer treatment. In 2012, our group first discovered that nanoparticles secreted from a carnivorous fungus, Arthrobotrys oligospora, had promising properties as immunostimulatory and antitumor agents for cancer treatment [11]. A. oligospora is a representative flesh eater in the fungal kingdom. It can develop into specialized 3-D adhesive traps for capturing, penetrating and digesting free-living nematodes in diverse environments [12]. A scalable and robust platform was developed to produce these fungal nanoparticles (FNPs) from a sitting drop culture system [11]. From this platform, the FNPs collected by a washing-dialysis procedure showed a size of 200–300 nm in diameter measured by scanning electron microscopy (SEM)/atomic force microscopy (AFM), and 300–400 nm in aqueous suspension

4270

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

measured by dynamic light scattering (DLS) [11]. From the perspective of mere passive tumor targeting in vivo, the upper bound size of the nanoparticles participating in the enhanced permeability and retention (EPR) effect is believed to be 400 nm [13], and an effective drug carrier for in vivo cancer treatment should have a diameter of <200 nm considering the multiple factors in vivo, such as limited fenestration size of the leaky vasculature in tumors, and rapid systemic clearance [14]. Thus, in order to effectively utilize these FNPs as drug carriers for chemical drug delivery into the tumor tissue in vivo, we need to further purify these naturally occurring FNPs to reduce their average particle size without compromising their bioactivities, such as immunostimulation and cytotoxicity. For such a purpose, we have established a surface charge-selective fractionation approach to purify the crude FNPs secreted from the sitting drop culture system. At the same time, the physicochemical properties of the newly isolated FNP fractions are characterized, and their potential use as bioactive nanocarriers in cancer therapy is finally explored by investigating their bioactivities against distinct immunocytes and tumor cells, as well as the combined immunochemotherapeutic effects in an in vitro co-culture system. 2. Materials and methods 2.1. Chemicals, fungus and cell lines A. oligospora (ATCC 24927), A549 human non-small-cell lung cancer cells (CCL-185) and RAW 264.7 murine macrophages (TIB71) were obtained from the American Type Culture Collection (Manassas, VA). B16BL6 murine melanoma cells, MCF-7 human breast tumor cell line, and multidrug resistant cell line MCF-7/ ADR were obtained from the National Cancer Institute-Central Repository (Frederick, MD). Splenocytes, derived from C57BL/6 mice, were purchased from the Allcells Company (Emeryville, CA). HEPES, 1,9-dimethyl-methylene blue (DMMB), chondroitin sulfate (CS), Sephadex G75, DEAE-cellulose, and phosphate-buffered saline (PBS) were purchased from Sigma-Aldrich (St Louis, MO). Doxorubicin hydrochloride (DOX) was purchased from Abcam (Cambridge, MA). LysoTracker Green DND-26 and Hoechst 33342 were purchased from Invitrogen Life Technologies (Grand Island, NY). Fetal bovine serum, DMEM medium and RPMI 1640 medium were purchased from Mediatech (Manassas, VA). Penicillin (10,000 units ml1)–streptomycin (10,000 lg ml1) solution was obtained from MP biomedicals (Solon, OH). 2.2. Arthrobotrys oligospora culture and FNPs fractionation A. oligospora was cultured in the sitting drop culture system proposed early [11] with some modifications to scale-up production and improve purification quality. Briefly, conidia suspension (about 1000–2000 conidia in 200 ll) was inoculated into the media droplet and incubated at 25 °C for 7 days. The isolation procedure was shown in Fig. 1A. First, the mycelia developed on the cover slip were washed over 10 times using distilled water. The collected water containing nanoparticles were then filtered through a 0.2 lm syringe filter (cellulose acetate, VWR, Radnor, PA). The FNPs were then desalted by a size exclusion chromatography (SEC, Sephadex G75) column [15]. The desalted FNPs is designated as FNP0, which is a crude sample. To further purify the FNP0, weak anion-exchange (WAX) chromatography on DEAE-cellulose was performed [16]. The DEAE-cellulose columns were then eluted in a stepwise fashion with 0.1, 0.2, 0.3, 0.5 and 1.0 M NaCl. As reported in a previous study [11], glycosaminoglycan (GAG) has been determined to be one of the main components in the FNPs. Thus, the colorimetric assay (k525nm) for GAG with

1,9-dimethyl-methylene blue was used to monitor the FNP fractions in the eluates from the SEC or DEAE-cellulose column. The elution profiles of the FNPs, reflected from GAG concentration, were plotted vs. elution volumes. The collected peaks containing FNPs from WAX column were subjected to the Sephadex G-75 column for desalting. The desalted FNP fractions were concentrated to final volume of 150 ll using a centrifugal filter tube (Amicon Ultra15 100K, Merck Millipore, Ireland). 2.3. Characterization of FNP fractions To characterize nanomorphology and particle size of the FNP fractions, the samples were analyzed using AFM (MFP-3D, Asylum Research, Santa Barbara, CA) with an IGOR Pro control system. Briefly, 10 ll of the particle suspension was air-dried on a glass cover slip, and imaged in AC mode at room temperature using a silicon probe PPP-NCHR-20 (Nanosensors™, Neuchatel, Switzerland) with a cantilever spring constant of 42 N m1 and a resonance frequency of 330 kHz. The nanoparticle samples were also analyzed by DLS and electrophoretic light scattering (ELS) to determine the size distribution and zeta potential in aqueous suspension using a Zetasizer Nano (Malvern Instruments Ltd., Worcestershine, UK) with a He–Ne laser (wavelength of 633 nm) and a detector angle of 173°. All samples were measured in triplicate. To qualitatively determine the chemical components in the nanoparticles, SDS–PAGE was used, and then the GAG, neutral polysaccharides and proteins in the nanoparticles were stained using Alcian blue, PAS reagents and silver staining reagents, respectively. To quantitatively determine the chemical components in the nanoparticles, total amounts of polysaccharides were measured using anthronesulfuric acid assay [17]. The amount of GAG in each sample was determined by a proteoglycan detection kit (1,9-dimethylmethylene blue, Astarte Biologics, Redmond, WA) [11], and the uronic acid in the nanoparticles was determine using carbozole assay [18]. Meanwhile, the concentration of proteins in the samples was quantitatively determined by the BCA protein assay (Pierce, Rockford, IL) following the manufacturer’s instructions. 2.4. Immunostimulatory activity The mouse macrophage RAW 264.7 cells (ATCC TIB-71) and splenocytes derived from C57BL/6 mice were cultured in DMEM and RPMI 1640 culture media, respectively. Both media were supplemented with 10% FBS and 1% penicillin–streptomycin at 37 °C in 5% CO2. The cells were plated in 12 well plates at a density of 5  106 cells ml1, treated with the FNPs at the GAG concentration of 5 lg ml1. After a 24 h incubation, the supernatants were collected for ELISArray. Mouse common cytokines and chemokines multi-analyte ELISArray kits (SABiosciences Corporation, Frederick, MD) were used to determine 12 cytokines (IL-1A, IL-1B, IL-2, IL-4, IL-6, IL-10, IL-12, IL-17A, IFNc, TNFa, G-CSF and GM-CSF) and 12 chemokines (RANTES, MCP-1, MIP-1a, MIP-1b, SDF-1, IP-10, MIG, Eotaxin, TARC, MDC, KC and 6Ckine) in the supernatants following the manufacturer’s instructions. The concentration of nitric oxide (NO) in the supernatants of both cells treated with the FNP samples were also determined using Griess assay, as described elsewhere [19]. 2.5. MTT assay The cytotoxicity of the purified FNP fractions and the DOX–FNP complexes against four cancer cell lines (A549, B16BL6, MCF-7 and MCF-7/ADR cells) was evaluated by MTT assay [15,20]. Biocompatibility of the purified FNP fractions toward mouse fibroblast NIH3T3 cell was also measured through MTT assay. Briefly, 8000–10,000 cells were plated in 96-well plates in 100 ll culture

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

4271

Fig. 1. Schematic of the newly developed charge-selective fractionation procedure, size exclusion chromatography (SEC)–weak anion exchange (WAX)–SEC procedure, for FNP purification (A), and the elution profiles of the FNP-containing media collected by washing mycelia in the sitting drop culture system using an SEC (Sephadex G75, 15 mm  70 mm) column (B), and a WAX (DEAE–cellulose, 10 mm  70 mm) column (C). The FNPs isolated from the SEC column were designated as FNP0, and are crude nanoparticles. After loading the FNP0 into the WAX column, two FNP fractions eluted from 0.5 and 1.0 M NaCl were designated as FNP1 and FNP2, respectively. The FNPs were qualitatively determined by measuring the absorbance (k525) of glycosaminoglycan (GAG) colorimetrically in the FNPs.

media per well and incubated at 37 °C in 5% CO2 for 24 h to allow the cells to attach. Specifically, DMEM medium was used for A549 and B16BL6 cells, RPMI 1640 medium for MCF-7 and MCF-7/ADR cells, and DMEM-a medium for MIH3T3 cells. The media was supplemented with 10% fetal bovine serum (for tumor cells) or calf serum (only for NIH3T3), and 1.0% penicillin–streptomycin. The cells were then treated with different concentrations of the FNP fractions or the DOX–FNP complexes. After a 48 h treatment, 10 ll of MTT solution (5 mg ml1 in PBS, pH 7.4) was then added to each well, and the plates were incubated for another 4 h. The cell culture media was removed and replaced with 100 ll DMSO. The absorbance was measured by a microplate reader (Bio-Tek lQuant) at 570 nm. For the DOX–FNP complexes, the average IC50 value was determined by cell survival plots using the ‘‘DoseResp’’ function in OriginPro 8.0. 2.6. Apoptotic assay The apoptosis assay was conducted by evaluating DNA ladder formation [21]. Briefly, A549 cells and B16BL6 cells were treated with the FNP fractions at the GAG concentration of 10 lg ml1, and then incubated at 37 °C in 5% CO2 for 48 h. Apoptotic cells were then identified by TdT-mediated dUTP nick and labeling (TUNEL) assay using APO-BrdU™ TUNEL Assay Kit (Invitrogen, Eugene, OR) following the manufacturer’s instructions. The cells were analyzed using flow cytometry (Epics XL Analyzer, Beckman Coulter Inc., Brea, CA) by collecting 20,000 events for each sample and measuring the cell-associated fluorescence.

37 °C in 5% CO2 for 24 h. The cells were then trypsinized, washed with PBS, and fixed in 75% ethanol at 4 °C for 2 h. The fixed cells were stained with propidium iodide/RNase A staining buffer (Invitrogen, Eugene, OR) at 37 °C for 30 min in the dark. The cell cycle analysis was conducted using flow cytometry (Epics XL Analyzer, Beckman Coulter Inc., Brea, CA) by collecting 20,000 events for each sample and measuring the cell-associated fluorescence. 2.8. Formation of DOX–FNP complexes To prepare the complexes formed by the DOX and the FNP fractions, 60 ll of DOX (3 mM) was mixed with 100–200 ll of FNP samples (containing 15 lg GAG in each sample) in 20 mM HEPES buffer at pH 7.0, and then incubated at room temperature overnight. The precipitates formed by the DOX and the FNP fractions were then centrifuged at 10,000 rpm for 10 min. The precipitates were dispersed in 500 ll PBS buffer, and then sonicated for 10 min in a bath sonicator (Aquasonic 7500, VWR). The amount of DOX in the dispersed precipitates was quantified by measuring UV absorbance at 480 nm and the entrapment ratios of DOX in the complexes were calculated as previously reported [15]. To determine the stability of DOX in the complexes in the PBS buffer, the dispersed complexes were applied to a Sephadex G75 column. The first peak (standing for stable complexes) and the second peak (standing for free DOX) were collected for quantification. The nanomorphology, particle size and zeta potential of the DOX–FNP complexes were characterized using AFM, DLS and ELS analysis. 2.9. In vitro release study

2.7. Cell cycle analysis To determine cell cycle distribution, A549 and B16BL6 cells were seeded into 24-well plates and treated with the FNP fractions at the GAG concentration of 10 lg ml1, and then incubated at

DOX release from the DOX–FNP complexes was measured at different pH values. 150 ll of the DOX–FNP1 complexes (168 lM for DOX), 200 ll of DOX–FNP2 complexes (126 lM for DOX), or 84 ll of free DOX (300 lM) were placed in a dialysis tube (MWCO

4272

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

300K, Spectrum Labs, CA), and then immersed in tubes containing 6 ml of release buffers at different pH values (1 PBS, pH 7.4; 0.1 M acetic acid buffer, pH 5.5). All tubes were incubated at 37 °C under mild agitation. The dialysate sample (0.5 ml) was collected at different time points and replenished immediately with the same volume of the fresh medium. The concentration of DOX in the dialysate was determined fluorometrically at kex480 nm and kem590 nm, and the cumulative release profiles were plotted vs. release times. 2.10. Cellular uptake and confocal microscopy study Quantification of intracellular DOX uptake in cancer cells was evaluated by flow cytometry. A549 and B16BL6 cells were cultured in 6-well plates at densities of 1  106 cells ml1, and incubated at 37 °C in 5% CO2. The DOX–FNP complexes at the DOX concentration of 10 lM were added into each well, and free DOX was used as a control. After a 4 h incubation, the media were aspirated. The cell monolayer was rinsed with PBS for three times, and then trypsinized. Flow cytometry analysis was carried out on a FACSCalibur (BD Biosciences) by collecting 20,000 events for each sample and the data were analyzed by FlowJo software (Tree Star, Ashland, OR). Confocal laser scanning microscopy (FluoView FV1000, Olympus, Japan) was used to investigate intracellular DOX distribution in tumor cells treated with the DOX–FNP complexes. Briefly, the cells were seeded on cover slips with a density of 106 cells ml1 on 6-well plates and cultured at 37 °C in 5% CO2 for 24 h. The cells were then treated with the DOX–FNP complexes at the DOX concentration of 10 lM for 4 h. Free DOX was used as a control. To observe the intracellular distribution, endolysosome- and nuclear-specific markers, LysoTrackerÒ green (100 nm) and Hoechst 33342 (4 lM), were incubated with the cells for 30 min prior to the confocal imaging. After that, the cover slip was washed with PBS three times, set on a microscope slide, and then examined by confocal microscopy. To observe the internalization of the FNPs, the FNP fractions were first labeled with FITC. Briefly, 3.5 mg ml1 of FITC in DMSO was diluted to 0.7 mg ml1 in 100 mM carbonate buffer (pH 9.3), and then 150 ll of the FNP1 (138.56 lg ml1) and FNP2 (56.83 lg ml1) were added into 400 ll of the above carbonate buffer. After a 24 h incubation at the room temperature, resulting solution was dialyzed (MWCO 300K, Spectrum Labs, CA) against PBS buffer for 3 days. The cells were further treated with the FITC-labeled FNP fractions at the FITC concentration of 2 ng ml1, and the internalization of the FNP fractions was then imaged by confocal microscopy. 2.11. Co-culture system to evaluate immunochemotherapeutic activity A co-culture system using B16BL6 tumor cells and the splenocytes derived from C57BL/6 mice was used to evaluate the immunotherapeutic effect of the DOX–FNP complexes in vitro [4,9]. Briefly, 2  105 tumor cells, labeled with 5 lM CFSE, were co-cultured with 5  106 splenocytes, and then the co-cultures were treated with free DOX and both DOX–FNP complexes at the DOX concentration of 1 lM. After a 24 h incubation, the death of tumor cells was determined by the PI uptake method using flow cytometry after gating on the CFSE-labeled cancer cells. 2.12. Statistical analysis All the values were presented as mean ± standard deviation (SD) of at least three independent measurements. Statistical significance was tested by one-way ANOVA followed by a Student’s

t-test for multiple comparison tests. A P value of <0.05 was considered statistically significant. 3. Results 3.1. Purification and characterization of FNP Fractions As shown in Fig. 1A, after inoculation of at least 1000 conidia per cover slip on which 500 ll of media was added, the fungal mycelia were grown for 7 days and the mycelia that thrived on the cover slip were washed using distilled water. The collected wash media were filtered to remove any debris, and then applied to the Sephadex G75 column. The first peak collected from the SEC column is crude FNPs, designated as FNP0 (Fig. 1B), which was further characterized by AFM. As shown in Fig. 2A,B, spheroidal nanoparticles with a diameter of 100–300 nm were observed. DLS and ELS anlyses showed an average size of 300 nm (Fig. 2C) and a negative zeta potential of 30 mV (Table 1). These crude FNP0 nanoparticles were further applied to the DEAE-cellulose column, resulting in production of two FNP fractions that were eluted using 0.5 and 1.0 M NaCl (Fig. 1C), respectively. Both the FNP fractions, designated as FNP1 and FNP2, were further characterized for their nanomorphology, hydrodynamic size and zeta potential using AFM, DLS and ELS. As shown in Fig. 2D–F, the first peak (FNP1), eluted from 0.5 M NaCl, was spheroidal nanoparticles with a hydrodynamic diameter of 150 nm. The second peak, FNP2, had similar hydrodynamic size with spheroidal morphology (Fig. 2G,I). However, as shown in Table 1, the FNP1 had a zeta potential of 27 mV, which is lower than that of FNP2 (32 mV). Apart from the differences in morphology, sizes and zeta potential, FNP1 and FNP2 both had chemical components similar to that of the crude FNP0. As shown in Fig. S1A,B, GAG and neutral polysaccharides were demonstrated to be the major components for both FNP1 and FNP2 fractions, which are similar to the chemical components of the crude FNP0. The quantitative data for GAG, uronic acid, total sugars and proteins in the three FNPs are listed in Table 1. 3.2. FNP fractions stimulate secretion of multiple proinflammatory cytokines and chemokines from macrophage and splenocytes Cytokine secretion profiles for macrophages were first analyzed after treatment with the FNPs. As shown in Fig. 3A, of the panel of 12 cytokines assayed, the levels of IL-6, TNF-a, and G-CSF from the macrophage treated by the three FNP samples, FNP0, FNP1 and FNP2, were significantly elevated as compared to the untreated cells after a 24 h incubation. In addition, low but statistically significant increases in IL-1a and IL-17A were also detected for FNP1 and FNP2 fractions in the supernatant of the treated macrophage. Similar to macrophages, significantly higher amounts of IL-6 and TNF-a were detected in the culture supernatant of the splenocytes treated with the three FNP samples (Fig. 3B) in comparison with the untreated control. Other stimulatory cytokines, including IL1a, IL1b and IL-2, were also secreted in a low, but significantly higher amount from the treated groups with the three FNP samples. Additionally, in the treated splenocytes, only the crude fungal nanoparticle, FNP0, induced significantly higher amount of IL-10, IL-17A, IFN-c, and G-CSF as compared to the untreated control, while the purified FNP fractions did not show these activities. A panel of 12 chemokines was further assayed in both macrophages and splenocytes. As shown in Fig. 3C, TRANTES, MCP-1 and IP-10 were highly induced in the treated macrophages by the three FNPs, and low but statistically significant increases in MDC was also observed in the FNP2-treated macrophages. For the activation of the splenocytes, except for RENTES, MCP-1,

4273

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

Fig. 2. AFM images (A,B, D,E and G,H) and size distributions (C, F and I) of the FNP fractions obtained from the SEC–WAX–SEC procedure. (A–C) The FNP0 nanoparticles, obtained by following only the SEC procedure (Sephadex G75 column). (D–F) The FNP1 fraction, obtained by following the SEC–WAX–SEC procedure, and elution with 0.5 M NaCl. (G,H) The FNP2 fraction, obtained by following the SEC–WAX–SEC procedure, and elution with 1.0 M NaCl.

Table 1 Physicochemical characterization of the FNP fractions prepared following the SEC-WAX-SEC procedure.

FNP0 FNP1 FNP2

Size (nm)

Polydispersity index

Zeta potential (mV)

Protein (lg/ml)a

GAG (lg/ml)a

Uronic acid (lg/ml)a

Total Sugar (lg/ml)a

294.2 ± 152.3 147.5 ± 78.4 148.5 ± 67.4

0.267 0.202 0.195

30.7 ± 9.1 26.9 ± 6.9 32.1 ± 7.6

661.1 ± 10.7 86.8 ± 6.3 3.7 ± 0.7

187.6 ± 10.7 296.5 ± 38.1 98.7 ± 7.4

162.6 ± 23.1 153.9 ± 10.8 40.4 ± 7.7

410.1 ± 6.4 506.2 ± 25.2 162.7 ± 8.5

a Concentrations of protein, glycosaminoglycan (GAG), uronic acid and sugar in the FNP fractions were determined after the FNPs were concentrated to 150 ll using a spin filter. Therefore, the unit here is designated as lg/ml nanoparticle suspension instead of lg/mg freeze-dried nanoparticles. The FNP0 was prepared using one batch of fungal culture (40 small disks), and the purified FNP1 and FNP2 were prepared using three batches of fungal culture (120 small disks).

IP-10 and MDC, elevated levels of the MIP-1a, MIP-1b, TARC and KC were also observed after treatment with the three FNPs as compared to untreated control (Fig. 3D). Apart from multiple cytokines and chemokines, the bactericidal mediator nitric oxide (NO) stimulated by the FNP fractions was further evaluated on both macrophages and splenocytes. As shown in Fig. 3E,F, compared to the untreated cells, significantly elevated levels of NO were observed in both macrophages and splenocytes treated with the FNPs.

3.3. FNP fractions induce cytotoxicity in tumor cells via apoptosis and cell cycle arrest The in vitro cytotoxicities of the two FNP fractions against four different tumor cell lines, A549, B16BL6, MCF-7 and MCF-7/ADR, were further investigated. As shown in Fig. S2A–C, the three FNP samples showed a dose-dependent cytotoxicity against the four tumor cell lines. For the crude FNP0, 11–39% inhibition of cell

4274

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

Fig. 3. Effects of the FNP fractions on the secretion of cytokines (A,B), chemokines (C,D) and nitric oxide (E,F) from RAW 264.7 macrophage cells (A,C and E) and splenocytes (B,D and F). The macrophage RAW264.7 cells and splenocytes were treated with the FNP fractions at the GAG concentration of 5 lg ml1 for 24 h, and then the culture media were collected. A panel of 12 cytokines and 12 chemokines in the culture media were measured using ELISArray kits, and the nitric oxide in the culture media was detected with Griess assay. The results are expressed as mean ± SD. ⁄P < 0.05,  P < 0.01, significantly different from the controls.

proliferation were obtained in A549, B16BL6 and MCF-7 cells at concentrations ranging from 1 to 10 lg ml1. For the purified FNP2 fraction, inhibition rates of 26–37% for B16BL6 cells, 9–33% for A549 cells, and 3–30% for MCF-7 cells were observed at the

same concentration range. However, using a higher concentration (2–25 lg ml1), the purified FNP1 fraction showed similar inhibition rates in the three tumor cells, i.e. 15–32% for B16BL6 cells, 11–28% for A549 cells, and 3–28% for MCF-7 cells. In addition, it

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

was observed that all the three FNP samples showed lower inhibition rates (8–18% for the FNP0, 5–15% for the FNP1, and 9–17% for the FNP2) in the multidrug-resistant cell line MCF-7/ADR. Apart from tumor cells, the cytotoxicity of the FNPs was also examined in mouse fibroblast cell line NIH3T3. Less than 20% inhibition rates were seen in NIH3T3 cells treated with the three FNPs at the respective concentrations ranges (Fig. S2D–F). Further apoptosis analysis showed that the purified FNP2 induced strong apoptosis in the A549 cells and B16BL6 cells after a 48 h incubation, and the crude FNP0 had similar but weaker apoptosis induction in both tumor cells; however, the purified FNP1 could not induce significantly apoptotic effect in A549 tumor

4275

cells and induced only slight apoptotic effect in B16BL6 cells (Fig. 4A,B). The cell cycle arrest analysis is consistent with that of the apoptosis analysis. As shown in Fig. 4C,D, only the crude FNP0 and the purified FNP2 fraction arrested the cell cycle at sub G0/G1 phase in both tumor cells after a 24 h incubation, and the FNP1 fraction did not show this activity in both tumor cells. 3.4. Formation of pH-responsive complexes by efficiently binding DOX to FNP fractions Due to negative surface charges of the FNP1 and FNP2 (Table 1), DOX, which carries positive charges from protonation of the amino

Fig. 4. Apoptosis (A,B) and cell cycle arrest (C,D) in human non-small-cell lung cancer A549 cells (A,C) and mouse melanoma B16BL6 cells (B,D) induced by the FNP fractions. The cells were treated with the FNP fractions at the GAG concentration of 10 lg ml1 for 48 h (apoptosis assay) or 24 h (cell cycle analysis). For the apoptosis assay, the fragmented DNA was stained with TUNEL method (Section 2.6) and then measured by flow cytometry. For cell cycle analysis, the cells were stained with PI, and then measured with flow cytometry.

4276

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

group at pH 7.0 in HEPES buffer, could easily be bound to the FNPs via electrostatic interactions. As shown in Fig. 5A, when DOX is freshly dissolved in a neutral pH (20 mM HEPES buffer at pH 7.0) or acidic pH, no obvious precipitates were observed in the DOX solution after centrifugation at 10,000 rpm for 10 min (Fig. 5A). However, DOX could be efficiently bound to both FNP1 and FNP2 fractions when mixing DOX and the FNPs in 20 mM HEPES buffer at pH 7.0. The binding between DOX and the FNPs was highly efficient, leading to formation of precipitates overnight after centrifugation at 10,000 rpm for 10 min (Fig. 5A). The collected precipitates

were then dispersed in PBS buffer (pH 7.4), and nanosized DOX– FNP complexes were formed. As shown in Table 2, for both DOX– FNP1 and DOX–FNP2 complexes, the hydrodynamic diameters, measured by DLS analysis, were <200 nm, which was slightly increased compared to the blank FNP fractions (Table 1). The morphology of the dispersed DOX–FNP complexes was also imaged using AFM. Both DOX–FNP complexes were spheroidal nanoparticles with diameters of <200 nm (Fig. 5C,D), similar to the blank FNP fractions (Fig. 2). More importantly, the entrapment ratio of DOX in the FNP fractions could be as high as 72–77%, and the

Fig. 5. Characterization of the DOX–FNP complexes and pH-responsive release of DOX from the complexes. (A) The DOX–FNP complexes were precipitated in 20 mM HEPES buffer (pH 7.0) after centrifugation at 10,000 rpm for 10 min. (B) Purification of the precipitated DOX–FNP complexes from free DOX using the SEC column after being well dispersed in PBS buffer (pH 7.4). (C) AFM images of the DOX–FNP1 complexes. (D) AFM images of the DOX–FNP2 complexes. (E) pH-responsive release of DOX from the complexes; non-linear curve fitting with an exponential function (y ¼ A1  ex=t1 þ A2  ex=t2 þ y0 ) in OriginPro 8.0 was applied to the release data. R2 for the fitted curves are greater than 0.89 for all samples.

4277

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284 Table 2 Physical characteristics and cytotoxicity of the DOX–FNP complexes.

DOX-FNP1 DOX-FNP2 Free DOX

Size (nm)

Polydispersity index

Zeta potential (mV)

Entrapment ratio (%)

194.5 ± 79.5 186.9 ± 89.7 –

0.241 0.279 –

22.20 ± 7.48 24.24 ± 5.95 –

77.4% ± 2.4% 72.2% ± 0.72% –

precipitated DOX–FNP complexes with such a high drug loading were demonstrated to be stable after nanosized DOX–FNP complexes had formed in the PBS buffer. As shown in Fig. 5B, the amount of DOX dissociated from the freshly prepared complexes in the PBS buffer was as low as 20%. The DOX-loaded nanoparticles dispersed in PBS buffer also showed good storage stability. After the nanocomplex dispersion in PBS buffer had been stored at 4 °C for 1 week, the entrapment ratio of DOX in FNP fractions still maintained 70%. Hence, the DOX–FNPs complexes dispersed in the PBS buffer were directly used as a nanosized antitumor agent without further purification to remove free DOX for the following studies, including uptake, cytotoxicity and in vitro immunochemotherapeutic effects. The release profiles of DOX from both complexes at different pH were further evaluated by immersing the dialysis tubes in large volume centrifuge tubes containing 6 ml of release buffers. As shown in Fig. 5E, the free DOX control confirmed that dialysis membrane tubing with 300K MWCO in this study could not restrict diffusion of the released drugs into the bulk release media, indicating that the sink condition was established. The free DOX release was able to reach 100% release after 5 h; however, the release of DOX from both complexes at different pH values could not reach a plateau until at least 9–10 h. The total released drug from both DOX–FNP complexes was significantly different under different pH conditions (Fig. 5E). Before 5 h, the release rates of DOX from free DOX solution were higher than those from both complexes. As for both complexes, the release rates of DOX increased with decreasing in the pH of release medium during 5 h. After 24 h, up to 55% and 65% of total drug were released at the physiological pH 7.4 for the DOX–FNP1 and DOX–FNP2 complexes, respectively; however, 80% of total drug released at pH 5.5 were observed for both complexes. 3.5. Cytotoxicity, uptake and intracellular distribution of DOX–FNP complexes in tumor cells For the cytotoxicity of nanocomplexes, the DOX–FNP2 complexes showed significantly higher cytotoxicity against four tumor cell lines than free DOX after a 48 h incubation (Fig. S2 G–J). The IC50 of the DOX–FNP2 complexes and free DOX are listed in Table 2. As for the A549, B16BL6 and MCF-7 cell lines, the IC50 for the DOX– FNP2 complexes was 1.5- to 1.8-fold lower than the free DOX; even for the multidrug resistant cell line MCF-7/ADR, the IC50 for the DOX–FNP2 was still 1.2-fold lower than the free DOX. However, the DOX–FNP1 complexes showed weaker cytotoxicity compared to free DOX. As shown in Table 2, the IC50 values for the DOX–FNP1 complexes were 1.1-to 2.8-fold higher than the free DOX against four tumor cell lines. DOX uptake and intracellular distribution were further investigated after the tumor cells had been treated with the respective nanocomplexes. As shown in Fig. 6, for both A549 and B16BL6 tumor cells, there was no significant difference in DOX fluorescence for both the complexes and free DOX at the DOX concentration of 10 lM after a 4 h incubation. Before investigating intracellular distribution of both the DOX–FNP complexes, we first tested if the FNPs could be efficiently internalized in the tumor

IC50 (nM) A549

B16BL6

MCF-7

MCF-7/ADR

1170.6 ± 92.33 599.34 ± 15.85 1052.5 ± 67.58

494.87 ± 38.00 209.63 ± 23.72 308.82 ± 26.55

1830.5 ± 270.5 355.71 ± 23.06 648.39 ± 75.77

5464.6 ± 16.87 3522.6 ± 110.0 4177.4 ± 116.1

cells. For this purpose, the FNPs were conjugated with FITC (Table S1), and then incubated with the A549 tumor cells for 4 h. As shown in Fig. S3, the FNP1 and FNP2 fractions were confirmed to be efficiently taken up by both tumor cells after a 4 h incubation. We further analyzed the intracellular distribution of the nanocomplexes. As shown in Fig. 7, the confocal imaging showed that a different intracellular distribution of both DOX–FNP complexes and free DOX was observed in both tumor cells after a 4 h treatment. The majority of DOX in both cells incubated with both the complexes were predominantly distributed in the endolysosomal compartment, while most of the free DOX was located outside the organelle (Fig. 7). 3.6. Immunochemotherapeutic activity of DOX–FNP complexes in an in vitro co-culture system The co-cultures were incubated with either DOX–FNP complexes or free DOX at the DOX concentration of 1 lM for 24 h. As shown in Fig. 8, significantly higher death rates of the cancer cells were observed with the treatment of both complexes as compared to free DOX. The percentage of dead tumor cells for both DOX– FNP2 and DOX–FNP1 complexes are 31.5% and 29.1%, respectively, significantly higher than the 25.5% of dead tumor cells for the free DOX. 4. Discussion The sitting drop culture method for the fungus A. oligospora has been previously established in our laboratory [11]. Using a dialysis procedure, the secreted FNPs were isolated. These crude FNPs have been determined to be spheroidal in shape with diameter of 100– 500 nm [11]. In this study, we established a surface charge-selective fractionation procedure by combining SEC and WAX to purify the crude FNPs. Two fractions, FNP1 and FNP2, having a similar particle size of 150 nm and different surface charges (27 mV for FNP1 vs. 32 mV for FNP2), were obtained. Both FNP fractions showed narrow size distribution with a polydispersity index (PDI) of 0.2. The difference in zeta potentials for both FNP fractions is consistent with their elution profiles through the WAX column. Due to the relatively lower surface charge of the FNP1 fraction, it was eluted more easily at low concentrations of NaCl than the FNP2 fraction (Fig. 1C) [16]. Therefore, at low salt concentration, i.e. 0.5 M NaCl, the FNP1 was first eluted, and then the FNP2 was selectively eluted at higher concentration, i.e. 1.0 M NaCl. For the protein component in the three FNP samples, the ratio of protein to total sugar in the FNP0 was much higher than those in the FNP1 and FNP2 (Table 1), indicating that most proteins in the crude FNP0 were free and unbound. These unbound proteins were completely washed away through the fractionation process with lower salt concentration (<0.5 M NaCl). Compared to the amount of total sugar, the purified FNP1 and FNP2 fractions had much lower ratios of proteins to total sugars. It is likely that most GAGs are covalently attached to core proteins to form proteoglycan [22]. As such, we presume that some proteins probably were bound to the GAGs in the nanoparticles. Two protein bands were observed in the FNP samples prepared using a dialysis method in

4278

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

Fig. 6. Quantitative analyses of DOX uptake by human non-small-cell lung cancer A549 cells (A,B) and mouse melanoma B16BL6 cells (C,D). Both cells were treated with the DOX–FNP complexes or free DOX at the DOX concentration of 10 lM for 4 h. The mean DOX fluorescence associated with the cells was then measured by collecting 20,000 events for each sample.

the previous report [11]. The protein bands with similar MWs have also been observed in the FNP0 samples in this study. For the purified FNP1 and FNP2 fraction, similar protein bands with relatively weaker intensities also appeared in the SDS–PAGE analysis (data not shown). Independent of the chemically bound or physically associated proteins, polysaccharides, including acidic GAG and neutral polysaccharides, are the main chemical components in the purified FNP fractions. Our previous study has shown that FNPs could induce the secretion of TNF-a from a macrophage cell line RAW264.7 in a dosedependent manner [11], indicating the potential antitumor immunity of FNPs. In this work, the macrophage stimulatory activity of the purified FNP fractions (FNP1 and FNP2) was first studied using the cultured mouse macrophage RAW 264.7. A panel of 12 cytokines and 12 chemokines in the culture media were measured using ELISArry analysis. The three FNP samples (FNP0, FNP1 and FNP2) could significantly enhance the secretion of IL-6, TNF-a and G-CSF from the macrophage (Fig. 3A). As is well known, TNFa is a Th1-biased cytokine and has been regarded as a potential anticancer agent for many years [23]. It plays a key role in apoptosis, cell survival, inflammation and immunity [24], and has been shown to be critical for antitumor T cell immunity in mice [25]. It could act synergistically with other drugs at the molecular level to trigger the apoptosis and dissociation of tumor vascular endothelial cells in cancer treatment [26,27]. As a Th2-biased cytokine, IL6 plays key roles in T-cell-mediated immune responses, acting as a cofactor for T-cell proliferation [28]. As a growth-inhibiting factor, the antitumor effect of IL6 on multiple murine tumor in vivo has been reported [29–31]. G-CSF could enhance the differentiation of stem cells in bone marrow, facilitate the mobilization of hematopoietic precursor cells into the bloodstream [32], and accelerate recovery from chemotherapy-induced myelosuppression

[33]. The synergistic antitumor effect of TNF-a and G-CSF has been established, and the antitumor effect of TNF was enhanced by combining with G-CSF in multiple tumors in vivo [34]. As such, we presume that the elevated levels of TNF-a, IL6 and G-CSF secreted from the macrophages will be beneficial for adjuvant anticancer therapy using the FNPs. To confirm these results, immunostimulation of the FNPs was further evaluated using primary splenocytes isolated from C57BL/ 6 mice, which have all types of immune cells, and the cross-talk between immunocytes, including macrophage and T cells [4,9]. As expected, significantly higher amounts of IL-6 and TNF-a were detected in the culture supernatant of the splenocytes treated with the three FNP samples. In addition, the secretion of IL1a, IL1b and IL-2 were also significantly enhanced in the treated groups (Fig. 3B). IL1, including IL-1a and b, has a number of properties potentially useful in the treatment of cancer, including direct antiproliferative activity against tumor cells, the activation of effector cells in vitro, and the inhibition of tumor angiogenesis [35]. In addition, IL1 has the capacity to protect and restore the bone marrow from radiation- or chemotherapy-induced injury [35]. The cytokine IL2 is known to be a T-cell growth factor, inducing clonal expansion of T cells following antigen stimulation, and is also important for the differentiation of CD4+ T cells into Th1 and Th2 effector subsets [28]. IL2 has been used for the treatment of melanoma and renal cell carcinoma in clinic [36], and recombinant human IL2 is a potent cytokine and a FDA-approved anticancer drug [37]. As such, we believe that the three cytokines, IL1a, IL1b and IL-2, with elevated levels in the supernatant of splenocytes after the treatment with the FNP fractions also favor the antitumor immunity in the cancer treatment. Comparing the crude nanoparticles with both purified nanoparticles, there is bias in the cytokine secretion profiles. Only the

4279

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

(A)

DOX

Lysotracker Green

Hoechst

Merge

Free DOX DOX-FNP1 DOX-FNP2

(B) Free DOX DOX-FNP1 DOX-FNP2 Fig. 7. Confocal images of intracellular distribution of the DOX–FNP complexes at the DOX concentration of 10 lM in human non-small-cell lung cancer A549 cells (A) and mouse melanoma B16BL6 cells (B). The cells were incubated with the samples at 37 °C in 5% CO2 for 4 h, and then 100 nM Lysotracer Green DND-26 and 4 lM Hoechst 33342 were added for 30 min incubation prior to imaging by the confocal microscopy. Scale bars represent 10 lm.

crude fungal nanoparticle, FNP0, induced significantly higher amount of IL-10, IL-17A, IFN-c, and G-CSF as compared to the untreated control, while the purified FNP fractions, FNP1 and FNP2, did not show these activities (Fig. 3B). As is well known,

IFN-c is a functionally pleiotropic cytokine, and has direct antiproliferative effects on some tumor cell lines [38]. Hence, similar to G-CSF, IFN-c may also be beneficial for antitumor immunity. Unfortunately, as for the purified fractions, FNP1 and FNP2, they

4280

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

Fig. 8. Immunochemotherapeutic activities of the DOX–FNP complexes in an in vitro co-culture system where B16BL6 tumor cells were first labeled with CFSE and then cocultured with the splenocytes derived from C57BL/6 mice. The co-cultures were treated with free DOX (B) and both the DOX–FNP complexes (C,D) at the DOX concentration of 1 lM, incubated for 24 h, and then the death of the tumor cells was determined by the PI uptake method using flow cytometry after gating on the CFSE-labeled cancer cells. The untreated co-cultured cells were used as a negative control (A). Statistical analysis of the mean DOX fluorescence associated with the cells was performed, and significant differences (P < 0.05) between different treatments were observed (E).

did not show the activity on either macrophage or splenocytes. However, the bias in immunostimulatory activity between the purified FNP fractions (FNP1 and FNP2) and the crude FNP0 may be helpful to establish the antitumor immunity in vivo, because the elevated level of IL10, an anti-inflammatory and immunosuppressive cytokine that favors tumor escape from immune surveillance [39], was observed in the supernatant of splenocytes treated with the crude fungal nanoparticles, FNP0. In addition, a higher amount of cytokine, IL17A, was also observed from the treated splenocytes by the crude FNP0 (Fig. 3B), and a low but significant increase in this cytokine was observed in the macrophage treated with FNP2 (Fig. 3A). The role of IL17A in antitumor immunity is controversial and remains elusive, and both proand antitumor activities of IL17A have been reported [40–41]. Overall, these results suggest that the FNPs, especially the purified fractions FNP1 and FNP2, can potentially modulate the immune cells to an activated state to induce an efficient antitumor response in a potentially combined therapy.

Subsequently, we investigated the production of chemokines by the FNP fractions in both macrophages and splenocytes. Chemokines are small chemotactic cytokines, which can induce migration of leukocytes, activate inflammatory responses, and are implicated in the regulation of tumor development and growth [42]. The three FNPs significantly enhanced the secretion of TRANTES, MCP-1 and IP-10 from the treated macrophages. A significant increase in MDC was also observed in the FNP2-treated macrophages (Fig. 3C). It has been reported that RANTES enhances antitumor immunity in a mouse model in part through direct T cell effector recruitment [43,44]. MCP-1 has been reported to augment the antitumor effects by promoting lymphocyte infiltration into the tumor and subsequent cytokine production [45]. IP-10 has been demonstrated to elicit strong antitumor and antimetastatic properties, and its immunological properties appear to be dependent on the attraction of monocytes and T lymphocytes [46]. MDC is chemotactic for a variety of leukocytes, and has been shown to be involved in Th-2-mediated cellular immunity [47]. In addition, the three FNPs

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

also significantly enhanced splenocytes to secrete RENTES, MCP-1, IP-10, MDC, MIP-1a, MIP-1b, TARC and KC (Fig. 3D). As is well known, MIP-1a shows a potent antitumor effect after intravenous administration along with intratumor injection of certain adjuvants [48], while MIP-1b is a chemokine which can chemoattract T cells and NK cells, inducing efficient antitumor responses in a pre-established tumor model [42]. In conjunction with RANTES, TARC, which mediates the chemoattraction of both antitumorspecific effector T cells, has been demonstrated to enhance the antitumor immune effects of GM-CSF [49,50]. The production of chemokine KC may amplify filtration of inflammatory cells, creating a more sustained antivascular action [51]. Given the elevated cytokines and chemokines from both macrophage and splenocytes, the results here provide evidence that favorable antitumor immunity in vivo could be established by stimulating different immunocytes with the FNPs, especially the purified FNP fractions, FNP1 and FNP2. Apart from the elevated levels of cytokines and chemokines as discussed above, an increase in NO could also benefit antitumor therapy (Fig. 3E,F). NO is an important regulator and mediator of macrophage-directed cytotoxicity against tumor cells and microbes [19]. Significantly higher amounts of NO production from the treated macrophages and splenocytes thus substantiate the potential of anticancer immunity in the cancer therapy using polysaccharide-based nanoparticles secreted from fungi as bioactive nanocarriers. Apart from the immunostimulatory activity, we further confirmed that the purified fractions, FNP1 and FNP2, both had mild to moderate cytotoxic activity against multiple tumor cell lines. In comparison of the FNP1with the FNP2 at the same GAG concentration, the latter showed around a 2-fold increase in cytotoxicity against the four tumor cell lines (Fig. S2A–C), indicating that the purified FNP2 fraction has stronger cytotoxic activity than the purified FNP1 fraction. In addition, all the three FNP samples showed lower inhibition rates in the multidrug-resistant cell line MCF-7/ ADR compared to its sensitive tumor cell line MCF-7. The lower inhibition rates here further suggest that the FNPs only possess mild to moderate cytotoxic activity, and the maximal concentration tested in this study is still not enough to effectively inhibit the proliferation of the resistant cells. Even for the sensitive tumor cells, the highest inhibition rates were still <40% at the maximal concentration of 10 lg ml1 for the FNP2 fraction or 25 lg ml1 for the FNP1 fraction. The mild to moderate cytotoxic activity of the FNPs was substantially confirmed by an in vitro biocompatibility test using NIH3T3 cell line. NIH3T3 is a mouse embryo fibroblast, which is commonly used in biocompatibility evaluation of nanomaterials [52]. Less than 20% inhibition rates in this cell line suggests that the FNPs did not have a strong cytotoxic effect against normal cells, but had a slightly higher cytotoxicity against tumor cells, especially towards the sensitive tumor cells. Even though the FNPs only possess mild to moderate cytotoxic activity, given that the FNPs induced the secretion of multiple proinflammatory cytokines and chemokines from immunocytes (Fig. 3), we believe that they are a potential immunomodulator of biological responses in the adjuvant antitumor therapy in which the synergistic effect could be reached between the mild cytotoxic activity and the immunostimulatory activity. For the purified FNP fractions, we have demonstrated a similar immunostimulatory activity between the FNP1 and FNP2, inducing almost the same levels of cytokines and chemokines from immunocytes (Fig. 3). However, as far as the cytotoxic activity concerns, the FNP2 fraction showed around 2-fold stronger activity in the tumor cells tested in this study (Fig. S2A–C). To better understand the difference in the cytotoxicity between the purified FNP fractions, we further investigated the apoptotic effect and cell cycle arrest in the tumor cells treated by both purified FNP fractions. As expected, it is the purified FNP2, not FNP1, that induced strong apoptosis in

4281

the A549 cells and B16BL6 cells (Fig. 4), which may explain the weaker cytotoxicity of the FNP1 as compared to the FNP2. The cell cycle arrest analysis using A549 and B16Bl6 cells treated with both purified FNPs substantially agrees with the apoptosis assay. Only the FNP2 fraction and crude FNP0 arrested the cell cycle at sub G0/G1 phase in both tumor cells (Fig. 4). A significant increase in the sub G0/G1 peak, which corresponds to apoptotic cells, indicates that the tumor has undergone apoptosis [53,54]. However, the purified FNP1 fraction could not significantly induce the cell cycle arrested at the sub G0/G1 phage, indicating that there should be different mechanisms for the cytotoxicity of the purified FNP1 and FNP2 fractions, although they showed similar immunostimulatory activity. To use the purified FNP fractions as nanocarriers for chemodrug delivery by exploiting the synergistic effect between the immunostimulation from FNPs and cytotoxicity from both the FNPs and chemodrugs, DOX was chosen as a model chemodrug to form a DOX–FNP nanocomplex [55]. Through electrostatic interaction between FNPs and DOX, the nanocomplexes, DOX–FNP1 and DOX–FNP2, with diameters of <200 nm were obtained (Fig. 5C,D). Both nanocomplexes showed a narrow size distribution with a PDI of <0.3. As compared to both the blank FNP fractions, there is significant decrease in zeta potentials for the DOX–FNP complexes (Table 2), indicating a direct association of DOX with the FNP fractions via electrostatic interactions [56]. In principle, pH-responsive release of DOX was expected from the nanocomplexes formed via an electrostatic driving force between DOX and nanoparticles. This could provide a stimulus-responsive release mechanism after internalization by tumor cells or penetration into the tumor tissue in vivo [57]. In this study, lower pH (pH 5.5) could lead to more DOX release as compared to pH 7.4 (Fig. 5E). The release rates of DOX from both complexes increased with decreasing pH of the release medium, indicating a pH-sensitive release behavior with an accelerated release of DOX in an acidic environment from both complexes. Since DOX is a weak amphipathic base with pKa 8.3–8.5 [56], lower pH results in stronger protonation of the amine group of DOX, leading to higher positive charge. On the contrary, the fungal nanoparticle fractions had negative charges at neutral pH (Table 1), which are likely from carboxyl groups or sulfate groups of glycosaminoglycan or core proteins in the nanoparticles. Thus, lower pH also results in stronger protonation of the carboxyl or sulfate group on the surface of the nanoparticles, leading to lower negative charges. Therefore, lower pH could lead to weaker electrostatic interaction and faster dissociation of the complexes as compared with neutral pH values. This favorable property is believed to facilitate passive tumor targeting since the interstitial space of solid tumors have a lower pH value [15]. Our previous study has demonstrated that there was synergistic cytotoxicity exerted by covalently conjugating DOX with FNPs via amide bonds [11]. A synergistic cytotoxic effect between DOX and the FNPs when forming the physical complexes via electrostatic interactions is also evidenced in this study. The DOX–FNP2 complexes showed significantly higher cytotoxicity against four tumor cells than free DOX after a 48 h incubation, demonstrating a synergistic cytotoxic effect between DOX and the FNP2 fraction. As shown in Fig. S2G–J, at the IC50 values of free DOX (Table 2), free DOX inhibited cell proliferation by 50%, whereas the DOX–FNP2 complexes showed 62–75% inhibition against the four tumor cells. Based on a 72% entrapment ratio for the DOX–FNP2 complexes (Table 2), the concentrations of the FNP2 in the complexes at the respective IC50 values for the A549, B16Bl6 and MCF-7 cells were calculated to be <0.25 lg ml1 of the GAG concentration, and for the resistant MCF-7/ADR cells <1.0 lg ml1 of the GAG concentration. Fig. S2A–C shows almost no significant inhibition effect on the four tumor cells at that concentration ranges for the FNP2, indicating that DOX and FNP2 in their physical complexes exerted

4282

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

synergistic cytotoxic effects and led to the IC50 values 1.2- to 1.8fold lower against the four types of tumor cells (Table 2). Unexpectedly, the cytotoxicity of the DOX–FNP1 complexes was similar or even lower than that of free DOX against the four different cell lines, which is completely different from the DOX–FNP2 complexes. The weaker cytotoxicity indicates that there is no obvious synergistic effect between DOX and the FNP1 upon forming the physical complexes. We postulate that the difference in the cytotoxicity between the two DOX–FNP complexes was due to the different cytotoxicity of the FNPs, because the FNP2 showed at least 2-fold higher cytotoxicity against different tumor cells than the FNP1 at the same GAG concentration (Fig. S2A–C). Although similar chemical components were characterized for both FNP fractions in Table 1 and Fig. S1, specific chemical structures for both FNPs still remained elusive. Presumably, the differences in the cytotoxicity of the FNPs reflect the different chemical structures of the FNP1 and FNP2 fractions, including polysaccharide chain, monosaccharide composites/linkages, uronic acid content, sulfation degree and possible core proteins. The different chemical structures in both purified FNPs eventually lead to different cytotoxicity of the FNPs and the DOX–FNP complexes, as well as the different physical properties including zeta potential and morphology (Table 1, Figs. 2 and 5). In order to further elucidate whether the different cytotoxicity between the DOX–FNP1 and DOX–FNP2 complexes is related to DOX uptake and intracellular distribution after forming the complexes, we quantitatively analyzed the cell-associated DOX fluorescence intensity using flow cytometry after treatment of the tumor cell lines, A549 and B16BL6, with the two complexes. No significant difference in the DOX fluorescence for both the complexes (Fig. 6) indicates that DOX uptake was not impeded upon the formation of the DOX–FNP complexes via the electrostatic interactions. The intracellular distribution of the two DOX–FNP complexes was subsequently examined in two tumor cells. As is well known, the majority of nanoparticles internalized via endocytosis in mammalian cells were mainly found within endosomes or lysosomes [58]. Thus, the higher DOX distribution in the endolysosomal compartment after treatment with the nanocomplexes (Fig. 7) indicates that the DOX–FNP complexes might be internalized by the endocytic pathway in both the tumor cell lines. The internalization of FNPs by A549 tumor cells (Fig. S3) indicates that the FNPs could mediate the uptake and distribution of DOX in tumor cells via the DOX–FNP complexes, instead of via free DOX released from the complexes. In addition, the DOX–FNP complexes were demonstrated to have a pH-sensitive release behavior with accelerated release of DOX in an acidic environment (Fig. 5B), which may facilitate DOX release from the nanocomplexes in endosomes or lysosomes after internalization. Overall, the experimental data here indicate that the formation of the DOX–FNP complexes did not decrease DOX uptake by both tumor cell lines, even though there was different DOX subcellular distribution as compared with free DOX. These results support the hypothesis that the different cytotoxicity of both complexes arises from the variations of cytotoxicity of the FNPs rather than the increase in the DOX uptake enhanced by the FNPs in the tumor cells. After demonstrating differences in cytotoxicity for both DOX– FNP complexes against tumor cells, the idea to combined cancer therapy using the complexes was further confirmed by the co-culture analysis. The co-culture study is an in vitro model system by mimicking in vivo situation [4,9]. For such a purpose, B16BL6 tumor cells were first labeled with CFSE, and co-incubated with splenocytes isolated from C57BL/6 mice. As expected, a significantly higher death rate of the cancer cells was observed with the treatment of the DOX–FNP2 complexes as compared to free DOX or the DOX–FNP1 complexes (Fig. 8). This result is consistent with the data in the direct cytotoxicity experiment using MTT assay (Fig. S2G–J and Table 2). As the DOX–FNP2 complexes had both

cytotoxic and immunostimulating activities, they might have been cooperating with each other to produce a synergistic effect, resulting in a higher death rate in the co-culture cells treated with the complexes. Interestingly, compared to the direct cytotoxicity experiment where the DOX–FNP1 complexes had lower cytotoxicity than free DOX (Fig. S2G–J and Table 2), in the co-culture experiment where a mixed culture of cancer cells and splenocytes were treated, the DOX–FNP1 complexes unexpectedly enhanced tumor cell death as compared to free DOX alone. In this co-culture experiment, a higher death rate of the cancer cells exerted by both complexes, especially the DOX–FNP1 complexes, could be attributed to the immune stimulatory activity of the FNPs. Both the FNP fractions have been shown to induce the secretion of multiple pro-antitumor cytokines and chemokines from splenocytes, such as TNF-a and MIP-1a, which has direct cytotoxic activity. Overall, the enhanced tumor cell death by both complexes confirmed the potential to use the DOX–FNP complexes in vivo for combined cancer therapy.

5. Conclusion The goal of this study was to explore the potential of the different fractions of FNPs for cancer therapy as bioactive nanocarriers by investigating their immunostimulatory activities, cytotoxic mechanisms and in vitro immunochemotherapeutic effects. We first established a surface charge-selective fractionation procedure to purify the FNPs through combining SEC and WAX, by which two highly purified FNP fractions, FNP1 and FNP2, were obtained. AFM imaging and DLS analysis showed that both the purified FNP fractions, with different surface charges (27 mV of zeta potentials for FNP1 and 32 mV for FNP2), had a reduced diameter of 100–200 nm compared to the crude FNP0. SDS–PAGE and chemical assays showed that polysaccharides, including GAG and neutral polysaccharides, are the main components in the purified FNP fractions. Both the TNP fractions were demonstrated to enhance the secretion of multiple proinflammatory cytokines and chemokines from macrophages and splenocytes, measured by ELISArray, suggesting the efficacy of the purified FNP fractions as a potential immunomodulator of biological responses in adjuvant antitumor therapy. MTT assay showed that both the purified FNP fractions had mild to moderate cytotoxicity against multiple tumor cells, but the FNP2 had stronger cytotoxic activity than the FNP1. The apoptotic assay and cell cycle analysis further demonstrated that FNP2, but not FNP1, could inhibit cell proliferation via inducing apoptosis and arresting tumor cells at sub G0/G1 phase. To test the combined cancer therapeutic effects, both FNP fractions formed pH-responsive nanocomplexes with the chemodrug DOX via electrostatic interactions. Upon binding of DOX to the FNP fractions, it was found that the DOX–FNP2 complexes had higher cytotoxic activity than free DOX against multiple tumor cells, while the cytotoxic activity of the DOX–FNP1 complexes was weaker than that of free DOX. Interestingly, in a co-culture experiment where splenocytes were co-cultured with tumor cells, both DOX–FNP complexes demonstrated higher antitumor activity than free DOX, suggesting a synergistic effect between the immunostimulation of the FNPs and cytotoxicity of the nanocomplexes in vitro. To summarize, we have developed a combined anticancer therapeutics using the charge-selective fractions of FNPs as bioactive nanocarriers, which may open a new avenue for combined cancer therapy.

Acknowledgements Y.W. and S.Y. contributed equally to this work. We appreciate early discussion related to the research topic with Dr. Fan Yuan at Duke University.

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 1–8 are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/j.actbio.2014. 06.020). Appendix B. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2014. 06.020. References [1] , [20.04.14]. [2] Morgan G, Ward R, Barton M. The contribution of cytotoxic chemotherapy to 5year survival in adult malignancies. Clin Oncol (R Coll Radiol) 2004;16:549–60. [3] Nowak AK, Lake RA, Robinson BW. Combined chemoimmunotherapy of solid tumours: improving vaccines? Adv Drug Deliv Rev 2006;58:975–90. [4] Roy A, Singh MS, Upadhyay P, Bhaskar S. Combined chemo-immunotherapy as a prospective strategy to combat cancer: a nanoparticle based approach. Mol Pharm 2010;7:1778–88. [5] Lee IH, An S, Yu MK, Kwon HK, Im SH, Jon S. Targeted chemoimmunotherapy using drug-loaded aptamer–dendrimer bioconjugates. J Control Release 2011;155:435–41. [6] Besch R et al. Proapoptotic signaling induced by RIG-I and MDA-5 results in type I interferon-independent apoptosis in human melanoma cells. J Clin Invest 2009;119:2399–411. [7] Poeck H et al. 50 -Triphosphate-siRNA: turning gene silencing and Rig-I activation against melanoma. Nat Med 2008;14:1256–63. [8] Karve S, Alaouie A, Zhou YP, Rotolo J, Sofou S. The use of pH-triggered leaky heterogeneities on rigid lipid bilayers to improve intracellular trafficking and therapeutic potential of targeted liposomal immunochemotherapy. Biomaterials 2009;30:6055–64. [9] Roy A, Chandra S, Mamilapally S, Upadhyay P, Bhaskar S. Anticancer and immunostimulatory activity by conjugate of paclitaxel and non-toxic derivative of LPS for combined chemo-immunotherapy. Pharm Res 2012;29:2294–309. [10] Praetorius NP, Mandal TK. Engineered nanoparticles in cancer therapy. Recent Pat Drug Deliv Formul 2007;1:37–51. [11] Wang Y, Sun L, Yi S, Huang Y, Lenaghan SC, Zhang M. Naturally occurring nanoparticles from Arthrobotrys oligospora as a potential immunostimulatory and antitumor agent. Adv Funct Mater 2013;23:2175–84. [12] Yang Y, Yang E, An ZQ, Liu XZ. Evolution of nematode-trapping cells of predatory fungi of the Orbiliaceae based on evidence from rRNA-encoding DNA and multiprotein sequences. Proc Natl Acad Sci USA 2007;104:8379–84. [13] Alexis F, Pridgen E, Molnar LK, Farokhzad OC. Factors affecting the clearance and biodistribution of polymeric nanoparticles. Mol Pharm 2008;5:505–15. [14] Steichen SD, Caldorera-Moore M, Peppas NA. A review of current nanoparticle and targeting moieties for the delivery of cancer therapeutics. Eur J Pharm Sci 2013;48:416–27. [15] Wang Y, Chen L, Ding Y, Yan W. Oxidized phospholipid based pH sensitive micelles for delivery of anthracyclines to resistant leukemia cells in vitro. Int J Pharm 2012;422:409–17. [16] Williams A, Frasca V. Ion-exchange chromatography. Curr Protoc Mol Biol 1998 [10.10.1-30]. [17] Nie S, Xie M, Fu Z, Wan Y, Yan A. Study on the purification and chemical compositions of tea glycoprotein. Carbohydr Polym 2008;71:626–33. [18] Fedarko NS. Purification of proteoglycans from mineralized tissues. Methods Mol Biol 2001;171:19–25. [19] Tincer G, Yerlikaya S, Yagci FC, Kahraman T, Atanur OM, Erbatur O, et al. Immunostimulatory activity of polysaccharide-poly(I:C) nanoparticles. Biomaterials 2011;32:4275–82. [20] Lv Z, Ou Y, Li Q, Zhang W, Ye B, Wu W. Expression, purification and bioactivities analysis of recombinant active peptide from shark liver. Mar Drugs 2009;7(2):258–67. [21] Pumphrey CY, Theus AM, Li S, Parrish RS, Sanderson RD. Neoglycans, carbodiimide-modified glycosaminoglycans: a new class of anticancer agents that inhibit cancer cell proliferation and induce apoptosis. Cancer Res 2002;62:3722–8. [22] Yip GW, Smollich M, Gotte M. Therapeutic value of glycosaminoglycans in cancer. Mol Cancer Ther 2006;5:2139–48. [23] Roberts NJ, Zhou SB, Diaz LA, Holdhoff M. Systemic use of tumor necrosis factor alpha as an anticancer agent. Oncotarget 2011;2:739–51. [24] Schioppa T et al. B regulatory cells and the tumor-promoting actions of TNF-a during squamous carcinogenesis. Proc Natl Acad Sci USA 2011;108:10662–7. [25] Calzascia T et al. TNF-a is critical for antitumor but not antiviral T cell immunity in mice. J Clin Invest 2007;117:3833–45.

4283

[26] van Horssen R, ten Hagen TLM, Eggermont AMM. TNF-a in cancer treatment: molecular insights, antitumor effects, and clinical utility. Oncologist 2006;11:397–408. [27] Lokshin A, Raskovalova T, Huang X, Zacharia LC, Jackson EK, Gorelik E. Adenosine-mediated inhibition of the cytotoxic activity and cytokine production by activated natural killer cells. Cancer Res 2006;66:7758–65. [28] de oliveira Silva F et al. Immunostimulatory activity of ConBr: a focus on splenocyte proliferation and proliferative cytokine secretion. Cell Tissue Res 2011;346:237–44. [29] Kang HS, Cho DH, Kim SS, Pyun KH, Choi I. Antitumor effects of IL-6 on murine liver tumor cells in vivo. J Biomed Sci 1999;6:142–4. [30] Mule JJ, Custer MC, Travis WD, Rosenberg SA. Cellular mechanisms of the antitumor activity of recombinant IL-6 in mice. J Immunol 1992;148:2622–9. [31] Ishiguro H, Kishimoto T, Furuya M, Nagai Y, Watanabe T, Ishikura H. Tumorderived interleukin (IL)-6 induced anti-tumor effect in immune-compromised hosts. Cancer Immunol Immunother 2005;54:1191–9. [32] Thomas J, Liu F, Link DC. Mechanisms of mobilization of hematopoietic progenitors with granulocyte colony-stimulating factor. Curr Opin Hematol 2002;9:183–9. [33] Voloshin T et al. G-CSF supplementation with chemotherapy can promote revascularization and subsequent tumor regrowth: prevention by a CXCR4 antagonist. Blood 2011;118:3426–35. [34] Maeda M et al. Enhanced antitumor effect of recombinant human tumor necrosis factor in combination with recombinant human granulocyte colonystimulating factor in BALB/c mice. Jpn J Cancer Res 1993;84:921–7. [35] Veltri S, Smith 2nd JW. Interleukin 1 trials in cancer patients: a review of the toxicity, antitumor and hematopoietic effects. Oncologist 1996;1:190–200. [36] Carmenate T, Pacios A, Enamorado M, Moreno E, Garcia-Martínez K, Fuente D, et al. Human IL-2 mutein with higher antitumor efficacy than wild type IL-2. J Immunol 2013;190:6230–8. [37] Chou SH et al. Palmitate-derivatized human IL-2: a potential anticancer immunotherapeutic of low systemic toxicity. Cancer Immunol Immun 2013;62:597–603. [38] Tannenbaum CS, Hamilton TA. Immune-inflammatory mechanisms in IFNgamma-mediated anti-tumor activity. Semin Cancer Biol 2000; 10:113–23. [39] Jarnicki AG, Lysaght J, Todryk S, Mills KHG. Suppression of antitumor immunity by IL-10 and TGF-beta-producing T cells infiltrating the growing tumor: influence of tumor environment on the induction of CD4(+) and CD8(+) regulatory T cells. J Immunol 2006;177:896–904. [40] Wei C, Sirikanjanapong S, Lieberman S, Delacure M, Martiniuk F, Levis W, et al. Primary mucosal melanoma arising from the eustachian tube with CTLA-4, IL17A, IL-17C, and IL-17E upregulation. Ent-Ear Nose Throat 2013;92:36–40. [41] Oshiro K et al. Interleukin-17A is involved in enhancement of tumor progression in murine intestine. Immunobiology 2012;217:54–60. [42] Yu YZ, Luo XL, Liu SX, Xie Y, Cao XT. Intratumoral expression of MIP-1b induces antitumor responses in a pre-established tumor model through chemoattracting T cells and NK cells. Blood 2004;104 [402b-b]. [43] Mulé JJ et al. RANTES secretion by gene-modified tumor cells results in loss of tumorigenicity in vivo: role of immune cell subpopulations. Hum Gene Ther 1996;7:1545–53. [44] Bacon KB, Premack BA, Gardner P, Schall TJ. Activation of dual T cell signaling pathways by the chemokine RANTES. Science 1995;269:1727–30. [45] Nakasone Y et al. Host-derived MCP-1 and MIP-1alpha regulate protective anti-tumor immunity to localized and metastatic B16 melanoma. Am J Pathol 2012;180:365–74. [46] Keyser J, Schultz J, Ladell K, Elzaouk L, Heinzerling L, Pavlovic J, et al. IP-10encoding plasmid DNA therapy exhibits anti-tumor and anti-metastatic efficiency. Exp Dermatol 2004;13:380–90. [47] Lee JM, Merritt RE, Mahtabifard A, Yamada R, Kikuchi T, Crystal TG, et al. Intratumoral expression of macrophage-derived chemokine induces CD4(+) T cell-independent antitumor immunity in mice. J Immunother 2003;26:117–29. [48] Nakano A et al. Intravenous administration of MIP-1a with intra-tumor injection of P. acnes shows potent anti-tumor effect. Int Immunopharmacol 2007;7:845–57. [49] Inoue H et al. TARC and RANTES enhance antitumor immunity induced by the GM-CSF-transduced tumor vaccine in a mouse tumor model. Cancer Immunol Immun 2008;57:1399–411. [50] Ghia P et al. Chemoattractants MDC and TARC are secreted by malignant B-cell precursors following CD40 ligation and support the migration of leukemiaspecific T cells. Blood 2001;98:533–40. [51] Wang LC et al. Neutrophil influx and chemokine production during the early phases of the antitumor response to the vascular disrupting agent DMXAA (ASA404). Neoplasia 2009;11:793–803. [52] Meenach SA, Anderson AA, Suthar M, Anderson KW, Hilt JZ. Biocompatibility analysis of magnetic hydrogel nanocomposites based on poly(Nisopropylacrylamide) and iron oxide. J Biomed Mater Res A 2009; 91A:903–9. [53] Chuang JY, Tsai YY, Chen SC, Hsieh TJ, Chung JG. Induction of G0/G1 arrest and apoptosis by 3-hydroxycinnamic acid in human cervix epithelial carcinoma (HeLa) cells. In Vivo 2005;19:683–8. [54] Williams O. Flow cytometry-based methods for apoptosis detection in lymphoid cells. Methods Mol Biol 2004;282:31–42.

4284

Y. Wang et al. / Acta Biomaterialia 10 (2014) 4269–4284

[55] Wang Y, Ding Y, Liu Z, Liu X, Chen L, Yan W. Bioactive lipids-based pH sensitive micelles for co-delivery of doxorubicin and ceramide to overcome multidrug resistance in leukemia. Pharm Res 2013;30:2902–16. [56] Wang Y, Yi S, Sun L, Huang Y, Lenaghan SC, Zhang M. Doxorubicin-loaded cyclic peptide nanotube bundles overcome multidrug resistance in breast cancer cells. J Biomed Nanotechnol 2014;10:445–54.

[57] Chen D, Song P, Jiang F, Meng X, Sui W, Shu C, et al. PH-responsive mechanism of a deoxycholic acid and folate commodified chitosan micelle under cancerous environment. J Phys Chem B 2013;117:1261–8. [58] Iversen T-G, Skotland T, Sandvig K. Endocytosis and intracellular transport of nanoparticles: Present knowledge and need for future studies. Nano Today 2011;6:176–85.