Lipid raft-mediated and upregulated coordination pathways assist transport of glycocholic acid-modified nanoparticle in a human breast cancer cell line of SK-BR-3

Abstract

Bile acid transporter-targeting has been proven to be an effective strategy to improve drug delivery to hepatocytes and enterocytes. With increasing discoveries of bile acid transporter expression on tumor cells, bile acid-modified anticancer drugs are gradually attaining interests. In our previous study, we confirmed the efficacy of glycocholic acid-conjugated polystyrene nanoparticles (GCPN) on apical sodium bile acid transporter (ASBT)-expressed SK-BR-3 cells. However, the transport mechanisms remain unknown, due to the nanosized carriers are unlikely to be pumped through the narrow cavities of ASBT. To clarify their transport pathways, in this article, pharmacological inhibition and gene knocking-down studies were performed, which revealed that GCPN were primarily internalized via non-caveolar lipid raft-mediated endocytosis. Proteomics was analyzed to explore the in-depth mechanisms. In total 561 proteins were identified and statistical overrepresentation test was used to analyze the gene ontology (GO) upregulated pathways based on the highly expressed proteins. It was found that multiple pathways were upregulated and might coordinate to assist the location of the GCPN-ASBT complex and the recycling of ASBT. Among the highly expressed proteins, myelin and lymphocyte protein 2 (MAL2) was selected and confirmed to colocalize with GCPN, which further supported the lipid raft-mediated process. These findings will help set up a platform for design the bile acid-modified nanomedicines and regulate their transport to improve their anticancer efficacy.

Graphical Abstract

1. INTRODUCTION

Bile acids are digestive surfactants that derives from cholesterol in the hepatocytes (Staels and Fonseca, 2009). In the human body, the bile acid pool is around 3 g, which circulates 4–12 times a day between liver and intestine (Chiang, 2013). In this “enterohepatic circulation” process, only ~5% of which are lost in feces each day and replenished by newly synthesized ones (Mertens et al., 2017). The bile acid transporters account for this high transport and recovery efficiency. In the liver, the bile acids are taken up by sodium-dependent taurocholic co-transporting polypeptide (NTCP) or organic-anion-transporting polypeptides (OATP) from the basolateral side (Suga et al., 2017), shuttled to the apical side with the assistance of fatty acid binding proteins (FABP)(Meier and Stieger, 2002), and exported by bile salts export pump (BSEP) or Multidrug Resistance Protein (MRP2) (Alrefai and Gill, 2007; Kosters and Karpen, 2008). In the ileum, the bile acids enter the enterocytes from lumen via apical sodium-dependent bile acid transporter (ASBT), delivered to the basolateral side by ileal bile acid binding protein (IBABP) and released by organic solute transporter (OST α/β) into circulation (Dekaney et al., 2008).

In the last two decades, bile acid transporter-mediated drug delivery system has been attracting research interests in enhancing drug delivery. These studies mostly focus on targeting the bile acid transporters in hepatocytes and ileal enterocytes, which utilize the natural circulation properties of bile acids. With deeper understanding of bile acids transporters in recent years, however, cancer-targeting drugs with bile acid modification are surging due to the increasing discoveries of bile acid transporter expression on cancer cells. For examples, Xiao et al. conjugated deoxycholic acid to camptothecin and found that the conjugates significantly improved the uptake of camptothecin in HepG2 3D cell models (Xiao et al., 2019). Lozano and co-workers designed cisplatin-ursodeoxycholate conjugate as a model ASBT-targeting drug, which demonstrated higher efficacy in inhibition of human colon adenocarcinoma after i.v. injection in mice (Lozano et al., 2015). Grosser and collogues reported that daidzein-7-β-D-glucuronide-4′-sulfate, a cholic acid analogue, could be transported by Organic anion transporting polypeptide 2B1, a transporter whose expression increased in breast cancer cells than normal breast tissue (Grosser et al., 2015). These studies suggest the promising prospect of bile acid transporter-targeting anticancer strategy in clinical applications.

Despite rising of bile acid transporter-mediated anticancer therapeutics, whether this strategy is applicable in nanomedicine remains largely unknown. Based on the design principle of bile acid-conjugated drugs, the modified drugs are expected to be internalized into the tumor cells through the bile acid transporters. For instance, ASBT exploits the Na+ gradient to pump bile acids into cytoplasm, with the stoichiometry of two Na+ for one bile acid molecule (Alrefai and Gill, 2007); however, it is unlikely that the bile acid-modified nanomedicines enter the tumor cells simply by transporter pumping effect due to the small cavity size of ASBT (6Å×12Å×14Å) (Hu et al., 2011). This indicates that the complexity of bile acid-modified nanomedicines for anticancer treatment: whether they could be internalized into tumor cells by bile acid transporters and whether they utilize an unknown pathway different from free bile acids.

In our previous study, we synthesized glycocholic acid-conjugated carboxylated polystyrene nanoparticles (GCPN) and used them as the probe to investigate their transport efficiency in SK-BR-3 cells (a breast cancer line with high ASBT expression) (Kim et al., 2018b). It was found that GCPN demonstrated higher uptake than CPN, while the treatment of free glycocholic acid could significantly inhibit the uptake of GCPN. Meanwhile, GCPN were found to be internalized together with ASBT and then separated, and the localization of GCPN in endoplasmic reticulum (ER) and Golgi apparatus was observed. It was also found that that GCPN exhibited 2.9-fold reduced exocytosis rate compared with CPN in SK-BR-3 cells, indicating that the conjugation of glycocholic acid altered the intracellular transport of CPN (Kim et al., 2020). These findings illustrated the potential of breast cancer cell-targeting properties of GCPN, and depicted their preliminary transport behaviors. Here, in order to clarify the specific transport mechanisms of bile acid-modified nanomedicines in tumors, GCPN were further studied in SK-BR-3 cells. The endocytosis was investigated with pharmacological inhibition and gene knockdown analysis. Proteomics studies were performed to explore the in-depth transport pathway in the cells.

2. MATERIALS AND METHODS

2.1. Materials

Glycocholic acid and CFL488 labeled secondary antibody were ordered from Santa Cruz (Dallas, TX, USA). CPN (100 nm, carboxylate, fluorescence E_x = 552 nm, E_m = 580 nm) were purchased from Micromod Partikeltechnologie GmbH (Rostock, Germany). Dynasore and ZCL278 were provided by Abmole Bioscience (Houston, TX, USA). Genistein, cytochalasin D (CytD), 5-(N-Ethyl-N-isopropyl) amiloride (EIPA), tannic acid, chlorpromazine (CPZ), methyl-β-cyclodextrin (MβCD), and siRNAs of Ras homolog family member A (RhoA), ADP-ribosylation factor (Arf6) and Flotillin-1 (Flot-1) and MAL2 were purchased from Sigma Aldrich (St Louis, Mo, USA). The primers for quantitative polymerase chain reaction (qPCR) analysis were synthesized by Eurofins Genomics (Louisville, KY, USA). The antibody of MAL2 was obtained from Bioss Antibodies (Woburn, MA, USA). The antibody of actin beta (ACTB) was purchased from Abcam (Cambridge, UK). SK-BR-3 and Caco-2 cell lines were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA).

2.2. Synthesis and characterization of GCPN

GCPN is synthesized according to the previously described method (Kim et al., 2018b). In brief, GCA (300 mg) was dissolved in 3 mL dimethyl formamide. Then 173 mg dicyclohexylcarbodiimide (DCC) and 96 mg N-hydroxysuccinimide (NHS) were added to active the carboxyl group, and 2.1 mL ethylenediamine (EDA) was added and stirred at room temperature overnight. After the reaction, unreacted EDA was removed by vacuum evaporation at 80 °C. The mixture was filtered and precipitated in ethyl acetate. The pellet (GCA-EDA) was dissolved in 10 mL water and lyophilized. CPN (0.1 mL) was dispersed in 2 mL 2-(N-morpholino)ethanesulfonic acid buffer (0.1M, pH= 6.0). Then 5 mg 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide hydrochloride and 5 mg NHS were added and stirred for 0.5 h. The synthesized GCA-EDA (10 mg) was added and stirred overnight in dark. After the reaction, GCPN was collected by ultracentrifugation (M_w = 10 kD) at 3500 rpm for 10 min. The synthesis was confirmed by Nuclear magnetic resonance (NMR, Mercury 400, Varian, Palo Alto, CA, USA). The size and Zeta potential were measured by dynamic laser scattering (DLS) using Malvern Zetasizer Nano-ZS (Malvern, Worcestershire, UK). The morphology of CPN and GCPN was observed by JEOL JEM 2800 Scanning Transmission Electron Microscopes (JEOL, MA, USA).

2.3. ASBT expression in SK-BR-3 cells

SK-BR-3 was seeded in 6-well plates. After reaching 80% confluency, the cells were washed with PBS and the RNA was extracted by Trizol kit according to the manufacturer’s instruction. The cDNA was then synthesized using RNA (1 μg), dNTP, oligo (dT) 20-mer, and reverse transcriptase. Then qPCR analysis was performed by StepOnePlus real-time PCR system (Applied Biosystem) with human β-actin for gene expression normalization. The oligonucleotide primer sequences were shown in Table S1. For Western Blot analysis, the cells were collected and the protein was extracted by radioimmunoprecipitation assay buffer (RIPA) buffer after NPs treatment. The protein was separated by electrophoresis, transferred to PVDF film and blocked with 5% BSA. The films were incubated with primary antibody against ASBT at 4 °C overnight and HRP-labeled secondary antibody at 37°C for 2 h. The human α-tubulin was used as the reference. The sample was illuminated and observed by ImageQuant system (GE Healthcare, LAS500, USA).

2.4. Temperature effect on the uptake of GCPN in SK-BR-3 cells

SK-BR-3 cells were seeded in 12-well plates. After reaching 80% confluency, the cells were preincubated at the determined temperatures (37 °C, 20 °C, 4 °C) for 1 h. Then, the cells were treated with 100 μg/mL CPN or GCPN at 37 °C for 30 min, respectively. After the incubation, the cells were washed with cold PBS and collected by trypsin. The cells were further rinsed and the intracellular fluorescence intensity was measured by flow cytometry (FACS Canto II, BD Biosciences, San Jose, CA, USA).

2.5. Pharmacological inhibition study of endocytosis mechanism of GCPN in SK-BR-3 cells

The cytotoxicity of various pharmacological inhibitors (Table 1) was tested by 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Briefly, 5 × 103/well of SK-BR-3 cells were seeded in 96-well plate and treated with the inhibitors for 1 h at 37 °C. Then the cells were washed and treated with 100 μL of 0.5 mg/mL MTT at 37 °C for 4 h. The media was carefully removed and 100 μL DMSO was added. The absorption was measured at UV 570 nm by the microplate reader (CYTATION3, BioTek, USA) after 30 min incubation at 37 °C. To study the effect of the pharmacological inhibitors, 1 × 105/well cells were seeded in 12-well plates and incubated with the inhibitors at 37 °C for 0.5 h. Then the cells were treated with CPN or GCPN with the inhibitors for another 0.5 h. The intracellular fluorescence intensity was measured with flow cytometry.

Table 1.

Pharmacological inhibitors used in the endocytosis study of GCPN in SK-BR-3 cells.

Inhibitors	Concentration	Function
CPZ	30 μM	Blocking clathrin-mediated endocytosis by inhibiting the adaptor AP2 and thus interfere the coated pit assembly (Wang et al., 2016).
MβCD	5 mM	Blocking lipid raft-mediated endocytosis by depleting cholesterol (Valapala and Vishwanatha, 2011).
Genistein	100 μM	Blocking caveolae-mediated endocytosis by inhibiting the Src kinase-dependent phosphorylation of caveolin-1 (Zhang et al., 2018).
EIPA	50 μM	Blocking micropinocytosis by inhibiting Na+/H exchanger (Jayashankar and Edinger, 2020).
Dynasore	80 μg/mL	A small GTPase inhibitor which inhibits dynamin and blocks the formation of pinched off vesicles (Macia et al., 2006).
CytD	0.5 μM	Inducing actin depolymerization (Shoji et al., 2012).
Tannicacid	2 μM	Affecting the fluidity and permeability of lipid bilayer membranes (Nagesh et al., 2020).
ZCL 278	50 μM	Inhibiting Cdc42 by targeting Cdc42–intersectin interaction (Friesland et al., 2013).

2.6. Clathrin/Caveolae-independent endocytosis pathway study

2.6.1. Knocking down Flot-1, Arf6 and RhoA

SK-BR-3 cells were seeded in 6-well plate at a density of 1.5 × 105/well. After the cells reached 50% confluency, siRNA of RhoA, Arf6 and Flot-1 were transfected to the cells with jetPRIME® according to the manufacturer’s instruction. After 48 h, the cells were washed with PBS and the RNA was extracted by Trizol kit according to the manufacturer’s instruction. The qPCR analysis was performed as described in the previous part. The oligonucleotide primer sequences were shown in Table S1.

2.6.2. Effect of Clathrin/Caveolae-independent pathway on the endocytosis of GCPN on SK-BR-3 cells

SK-BR-3 cells were seeded in 12-well plate at 1 × 105/well. After the cells reached 50% confluency, siRNA of RhoA, Arf6 and Flot-1 were transfected to the cells with jetPRIME® according to the manufacturer’s instruction. After 48 h, the cells were washed with PBS and treated with 100 μg/mL of GCPN and incubated for 0.5 h. For Cdc42 effect exploration, the cells were treated with 50 μM ZCL278 followed by ZCL278 with the NPs. The intracellular fluorescence intensity was measured with flow cytometry.

2.7. Proteomics analysis of SK-BR-3 cells with GCPN uptake

2.7.1. Protein separation

SK-BR-3 cells were seeded in 100 mm culture dish. At 80% confluency, 100 μg/mL GCPN was added to the cells and incubated at 37°C for 24 h. The cells were washed with PBS, scrapped, collected by centrifugation and lysed followed by the separation of proteins by electrophoresis. Separated proteins were observed with Coomassie Brilliant Blue staining.

2.7.2. Digestion of in-gel proteins

Gel spots were destained with 50 mM ammonium bicarbonate in water/methanol (50:50, v/v). Proteins were reduced with dithiothreitol (DTT) for 45 min at 60 °C and then alkylated with iodoacetamide (IAA) for 30 min at room temperature in the dark. Gel spots were washed three times in 50 mM ammonium bicarbonate in water for 45 min/wash cycle. Gel spots were cut into small pieces and dehydrated using 100% acetonitrile. Proteins were digested overnight at 38 °C with Trypsin/LysC mixture at 1 μg trypsin per sample. The digestion was quenched by acidification with 1% formic acid to a pH of 2–3. Peptides were extracted from the gel using 50% acetonitrile with 1% formic acid and then concentrated in vacuo to a final volume of 5 μL.

2.7.3. LC/MS/MS Analysis

Reversed-phase nano-Liquid Chromatography with tandem mass spectrometry (LC-MS-MS) was performed on an Eksigent Ekspert nanoLC 425 system (SciEx) coupled to a Bruker MAXIS ETD II QToF mass spectrometer equipped with a nanoelectrospray source. Concentrated samples were diluted with sample: 0.1% formic acid (1:1) in water. Then 5 μL of the samples were injected onto the liquid chromatograph. A gradient of reversed-phase buffers (Buffer A: 0.2% formic acid in water; Buffer B: 0.2% formic acid in acetonitrile) at a flow rate of 150 μL/min at 60 °C was set up. The LC run lasted for 83 min with a starting concentration of 5% buffer B increasing to 55% over the initial 53 min and a further increase in concentration to 95% over 63 min. A 15 cm long/100 μm inner diameter nanocolumn was employed for chromatographic separation. The column was packed, in-house, with reverse-phase BEH C18 3.5 μm resin (Xbridge). MS/MS data was acquired using an auto-MS/MS method selecting the most abundant precursor ions for fragmentation. The mass-to-charge range was set to 350–1800.

2.7.4. Analysis of Data

Mascot generic format (MGF) files were generated from the raw MS/MS data. Mascot (version 2.6) uses the MGF file for database searching and protein identification. The SwissProt database was searched with homo sapiens taxonomy selected. The parameters used for the Mascot searches were: trypsin digest, two missed cleavages, carbamidomethylation of cysteine as fixed modification, oxidation of methionine as variable modifications, and the maximum allowed mass deviation at 11 ppm. For the final Scaffold data file, the Scaffold viewer program (http://www.proteomesoftware.com/products/scaffold/) was used to perform proteomics quantification. The expression degree was determined by label free quantification (LFQ). Based on the quantification result, statistical overrepresentation test of the upregulated proteins was performed via PANTHER (http://www.pantherdb.org/). The information on proteins were referred from UniProt database (https://www.uniprot.org/).

2.8. Validation of effect of MAL2 on GCPN transport

2.8.1. Colocalization between GCPN and MAL2 in SK-BR-3 cells

SK-BR-3 cells were seeded on glass slides in 24-well plate and treated with 200 μg/mL of GCPN for 1 h. Then MAL2 was stained with primary antibody at 4 °C overnight and CFL488 labeled secondary antibody at 37 °C for 2 h after the cells were fixed with 4% paraformaldehyde, permeabilization with 0.1% TritonX-100 and blocking with 1% BSA. The samples were observed by confocal laser scanning microscopy. Fifteen areas were selected randomly to calculate the average Pearson’s coefficient. Cells with no NPs added were used as negative control. In the meantime, ACTB was stained as the control protein.

2.8.2. Effect of MAL2 on the endocytosis of GCPN on SK-BR-3 cells

To study the effect of MAL2 on the uptake of GCPN, MAL2 was knocked down in SK-BR-3 cells with the same methods described above. The expression of MAL2 was evaluated by qPCR. The oligonucleotide primer sequences were shown in Table S1. For endocytosis study, the cells were treated with GCPN for 0.5 h and harvested. The intracellular fluorescence intensity was measured by flow cytometry.

2.9. Statistical analysis

The results are expressed as mean ± SD unless otherwise stated. All results were analyzed by analysis of variance (ANOVA) or t test unless specifically clarified. A p-value less than 0.05 was considered statistically significant, while a p-value less than 0.01 was considered to be highly significant.

3. RESULT AND DISSCUSSION

Currently, bile acid-mediated oral drug delivery mainly stays in targeting ASBT in ileal enterocytes, which has been confirmed an effective strategy. By contrast, studies on bile acid-mediated tumor-targeting are still rare. Some enterocyte-mimicking models, including Caco-2 cells or ASBT-transfected MDCK cells, demonstrated excellent properties as the targets for bile acid-modified drugs (Alam et al., 2014; Kim et al., 2018a). However, we discovered that the abundance of ASBT in SK-BR-3 cells were much higher than that in Caco-2 cells (Figure S1). This implies that bile acid-modified nanomedicines might be better as candidates for breast cancer treatment.

To investigate the transport mechanism of bile acid-modified nanomedicines in SK-BR-3 cells, commercial CPN (100 nm in nominal diameter and labeled with red fluorescence) were selected as an inert probe to study the transport pathways due to their stability under biological conditions. The synthesis route was displayed in Fig.1a. As shown in Fig.1b, the peaks of methyl group in GCA was detected in the NMR of GCPN, indicating the successful conjugation of GCA to the CPN surface. Compared with CPN, GCPN was in slightly larger size (127.3 ± 1.3 vs 109.0 ± 1.1 nm), and the spherical morphology was maintained after conjugation (Fig. 1c and 1d). The Zeta potential of GCPN (−29.2 ± 0.4 mV) was less negative compared with that of CPN (−49.1 ± 1.3 mV), which was due to the reduction of carboxyl groups on the particle surface. In our previous study, CPN demonstrated uniform intensity at different pH, biomimetic and physiological solvents, suggesting it could suitable as a probe and the quantification based on the fluorescence intensity is reliable (Kim et al., 2020).

Fig.1.

Synthesis and characterization of GCPN. a) Synthesis scheme of GCPN. b) NMR spectrum of GCPN. The peaks a, b and c represent the methyl groups from GCA. c) DLS analysis of particle size. GCPN showed slightly larger particle size than CPN. d) TEM images of CPN and GCPN. Both particles showed spherical shape.

To determine the endocytic behaviors of GCPN in SK-BR-3 cells, the temperature effect on the uptake was tested. As shown in Fig.2a, GCPN exhibited decreasing uptake at 20 °C, and the least uptake at 4 °C. As the endocytosis is an active process and energy-dependent, it can be blocked at low temperature (4 °C) (Saraste et al., 1986). At ~20 °C, cell membranes are known to experience many abnormal changes, like an irregularly extended network of labyrinthine channels, tubular elements and alpha vacuoles (Kosaka and Ikeda, 1983). Besides, at the inflection point (20 – 25 °C), subtle changes in the membrane have been reported including microdomain separation, liquid phase transition and protein conformational changes (Chanaday and Kavalali, 2018). These results suggest that GCPN uptake is an energy-dependent endocytic pathway and relies on the normal membrane morphology.

Fig.2.

Effect of temperature and pharmacological inhibitor on GCPN uptake. a) Temperature-dependent effect on the endocytosis of GCPN. At 20 °C and 4 °C, the uptake significantly decreased. b) Pharmacological inhibition study on the endocytosis of CPN and GCPN. CPZ, genistein and EIPA demonstrated statistical difference but the discrepancy was lower than 20% in GCPN uptake; tannic acid, MβCD, CytD and Dynasore showed significant inhibition effect on GCPN endocytosis. CPN uptake was inhibited by genistein, tannic acid, CytD and MβCD. n=3. **: p < 0.01; ***: p < 0.001; ****: p < 0.0001.

To study the endocytosis behaviors of GCPN, the uptake of GCPN was tested with pharmacological inhibitors in SK-BR-3 cells. The cytotoxicity of the inhibitors was studied by MTT assay (Figure S2). At determined concentrations, no inhibitor showed apparent effect on the cell viability, indicating they did not affect other critical biological functions apart from the particular endocytosis route. The working mechanisms of the inhibitors are described in Table 1.

As shown in Fig.2b, the most significant inhibition effect was observed with tannic acid treatment, which illustrated the essential role of membrane fluidity. Dynasore and CytD showed significant blocking effect, suggesting the key role of dynamin and actin filaments. Although genistein, CPZ, and EIPA showed statistical inhibition or enhancement effect, the discrepancy was lower than 20%, indicating the limited involvement of clathrin- or caveolae-mediated endocytosis or macropinocytosis in the GCPN uptake. It is noted that MβCD displayed about 40% inhibition in the uptake of GCPN in SK-BR-3 cells, which illustrates that lipid raft is crucial for the GCPN internalization. Caveolae are subdomains of lipid raft of the membrane with abundance of cholesterol, sphingolipids and caveolin (Mulcahy et al., 2014). Thus, caveolae-mediated endocytosis depends on the normal function of lipid raft. Conversely, the lipid raft-mediated endocytosis could be independent from caveolae, because the lipid and protein compositions of the different lipid raft domains are also different (Lajoie and Nabi, 2007; Yao et al., 2009). Owing to the lack of effect of genistein, GCPN were likely to be internalized by lipid raft-mediated but non-caveolar pathway. On the contrary, CPN uptake shared the mechanism of lipid raft and actin filaments regulation, with significant inhibition by MβCD (~40%) and CytD, while genistein also showed more than 30% inhibition, which indicates that CPN were largely internalized via caveolae-mediated pathway. This result reveals that GCA modification alter the transport from caveolar to non-caveolar route.

To ensure the survival condition of cells, the concentration of MβCD was adjusted to an effective but might not 100% inhibition range. This caused that there were still undergoing endocytosis. Besides, MβCD is known to activate some other pathways, such as increasing the staurosporine-sensitive kinase activity or ligand-independent activation of the epidermal growth factor receptor (Chen and Resh, 2002; Smith et al., 2010), which might cause some complementary uptake of GCPN. Therefore, more evidence is required for validation of the lipid raft-mediated pathway of GCPN. Some other pathways, called “clathrin/caveolae independent pathway”, have also been gradually regarded as important endocytosis routes, which include some key proteins such as Flot-1, Cdc42, Arf6, and RhoA (Canton and Battaglia, 2012; Iversen et al., 2011). The basic description is listed in Table 2. To study whether these pathways are involved in the endocytosis route of GCPN, Flot-1, Arf6, and RhoA were knocked down with siRNA, and Cdc42 was blocked by a known inhibitor, ZCL278. The knocking-down efficacy was over 70% (Fig.3a–c), indicating the successful block of Flot-1, Arf6, and RhoA. As shown in Fig.3d and 3e, silencing of RhoA and Arf6 or inhibition of Cdc42 showed no effect on the GCPN uptake. On the contrary, Flot-1, the lipid-associated protein, demonstrated significant involvement. It is also known that Flot-1-mediated endocytosis relies on the dynamin for vesicle scission (Fekri et al., 2019). This result is consistent with the pharmacological inhibition study and illustrates that lipid raft-mediated endocytosis is the main route of GCPN with the assistance of Flot-1. It has been reported that ASBT-V5 fusion protein is present in the membrane lipid raft layer after an Optiprep density gradient, indicating ASBT is directly associated with lipid rafts; meanwhile, depletion of plasma membrane cholesterol decreases ASBT activity, suggesting the content of cholesterol in lipid rafts is essential for the optimal activity of ASBT (Annaba et al., 2008). These findings further confirm the crucial role of lipid rafts in the GCPN endocytosis.

Table 2.

Clathrin/caveolae-independent pathways.

Pathways	Description
Arf6	A small GTPase that can regulate actin polymerization and induce endocytosis via Arf6-enriched vesicles (Van Acker et al., 2019).
Cdc42	Belonging to Rho family GTPase, regulating endocytic and secretary pathway (Chi et al., 2013).
Flot-1	Associated with lipid rafts and mediating endocytosis (Mulcahy et al., 2014).
RhoA	Belonging to Rho family GTPase, regulating the endocytosis by controlling Rho kinase (Chi et al., 2013).

Fig.3.

Clathrin/caveolae-independent pathway analysis of GCPN uptake. a-c) QPCR analysis of expression of Flot-1, Arf6 and RhoA after knocking down by siRNA in SK-BR-3 cells. d-e) Endocytosis of GCPN in Arf6(−), RhoA(−) and Flot-1(−) and cdc42(−) SK-BR-3 cells. Flot-1(−) group demonstrated significant reduction in GCPN uptake. N=3. **, p < 0.01.

For deeper understanding of the GCPN transport mechanism in SK-BR-3 cells, a proteomics study was performed to identify the specific factors and pathways in the uptake. The protein abundance was measured by label-free quantification (LFQ) based on LC/MS/MS. The minimum number of peptides to determine a protein was set 2. The protein expression was regarded as upregulated when the LFQ intensity of GCPN/Control was larger than 1.4, and as downregulated when lower than 0.7. Fig.4a and 4b showed the Venn diagrams of proteomics. The identified proteins in GCPN group were nearly the same with the control, with only 4 unique proteins in GCPN group and 1 in the control group. Among the 561 identified proteins, 118 were upregulated and 166 downregulated in GCPN group. The scattered plot and heat map of the LFQ intensities were shown in Fig.4c and 4d.

Fig.4.

Proteomics study of GCPN transport in SK-BR-3 cell lines. a) Venn Diagrams of distribution of kinds of identified proteins in control and GCPN groups. b) Venn Diagrams of distribution of up- and downregulated proteins in GCPN treated groups, with >1.4 fold to the control group as higher expression, and <0.7 as lower. c-d) Scattered plot and heat map of LFQ intensities of identified proteins. e) The statistical overrepresentation test of the top 5 pathways with the highest number of upregulated proteins based on the GO term cellular component, molecular function and biological process. f) The statistical overrepresentation test of the transport-related pathways based on the upregulated proteins. The p-values of the listed pathways were all <0.05.

The upregulated proteins were further studied by gene ontology (GO). The statistical overrepresentation test was performed in different GO terms: biological process, cellular component and molecular function. The top 5 pathways with the highest numbers of upregulated proteins are shown in Fig.4e, which are mainly related to cellular structure, protein bindings or metabolic process, indicating that the transport of GCPN required the coordination of various cellular components. To comprehensively elucidate the regulatory factors of the GCPN transport, the upregulated transport-related pathways were selected and shown in Fig.4f. The higher -log P value means the higher upregulation significance. It could be found that some pathways of cytoskeleton, including actin binding and actin cytoskeleton, are upregulated, which reveals that the actin filaments play an important role in GCPN transport and is consistent with the pharmacological study. It could also be found that the pathways of organelles, especially the membrane-bounded structures like ER, suggesting the uptake and intracellular delivery of GCPN among the different compartments were assisted by vesicles, which requires the packing and sorting in ER. According to our previous study, ASBT was internalized together with GCPN and then recycled to the cytoplasmic membrane (Kim et al., 2018b). The enhanced pathways, like protein-containing complex binding, cellular protein localization and cellular macromolecule localization, might account for the intracellular locating of GCPN-ASBT complex and the anchoring of ASBT back to the membrane.

Meanwhile, to identify the role of the specific proteins in the transport, some transport-relevant proteins with higher expression was shown in Figure S3, and the predicted transport routes in SK-BR-3 cells were depicted based on these results. Among the highly expressed proteins, MAL2 was selected for further study, as MAL2 is relevant to the raft-dependent transport (Information from Uniprot Database: https://www.uniprot.org/). By immunofluorescence staining, it was found that GCPN were colocalized with MAL2, revealing MAL2 were involved in the transport of GCPN (Fig.5a and b). On the contrary, the control protein, ACTB, whose expression was not significantly affected by GCPN (88% compared with the non-treated group), did not show colocalization to GCPN. Generally, proteins adsorbed on the NPs surface could be retained until degraded in lysosomes, indicating its rationale for determining the internalization and intracellular destiny (Bertoli et al., 2016). Therefore, this result could primarily confirm the participation of MAL2 in GCPN uptake. However, the complex intracellular milieu might also affect the adsorption of MAL2: the different expression level of MAL2 led by GCPN could conversely alter the affinity between GCPN and MAL2, and their interaction might also be interfered by other proteins (O’Connell et al., 2015; Vilanova et al., 2016). Besides, the formation of protein corona could also in turn influence the subsequent transport of NPs (Peng et al., 2019; Zhang et al., 2021).

Fig.5.

Colocalization of GCPN and MAL2 in SK-BR-3 cells. a) Confocal images of the colocalization (red: GCPN, green: MAL2, ACTB). The white arrow indicates the colocalization site. b) Pearson’s coefficient calculation of colocalization. GCPN group demonstrated significantly higher Pearson’s coefficient compared with the control group. c) QPCR analysis of expression of MAL2 after knocking down by siRNA in SK-BR-3 cells. d) Endocytosis of GCPN in MAL2(−) SK-BR-3 cells. ****, p< 0.0001; *, p<0.05.

To further validate the effect of MAL2 on the uptake of GCPN, MAL2 was knocked down (Fig. 1c) and the uptake was analyzed. It was found that the uptake of GCPN significantly decreased (Fig.5d), which confirmed the dependence the MAL2 and lipid raft in the endocytosis of GCPN in SK-BR-3 cells. Therefore, the lipid raft-mediated pathway is the crucial mechanism for GCPN internalization, which might inspire the rational design and regulation of the bile acid-modified nanomedicine in the future, for instance, by boosting the lipid raft affinity of the formulations or co-delivery the biodrugs to improve the lipid raft expression.

4. CONCLUSION

The transport pathways of GCPN in SK-BR-3 cells are summarized (Fig.6). GCPN are internalized together with ASBT via active endocytosis. The main pathway is non-caveolar lipid raft-mediated uptake, with the assistance of actin filaments, dynamin, Flot-1 and MAL2. The transport of GCPN induces the synergetic response of membrane-bounded organelles and cellular component bindings, which regulates the location of GCPN-ASBT complexes and the recycling of ASBT. Clarification of the transport mechanism will set up a stepping stone for design of ASBT-targeting nanomedicines and regulations of their antitumor behaviors in clinical therapeutics in the future.

Fig.6.

Schematic illustration of GCPN transport pathway in SK-BR-3 cells. GCPN entered the cells by associating with ASBT and lipid raft-mediated endocytosis, with the assistance of Flot-1, MAL2, actin filaments and dynamin. Pathways that are relevant to the membrane-bounded organelle and protein bindings were regulated and controlled the locating and separation of GCPN-ASBT complexes as well as the recycling of ASBT to the plasma membrane.