Inhibition of Wnt signaling by Frizzled7 antibody-coated nanoshells sensitizes triple-negative breast cancer cells to the autophagy regulator chloroquine

Abstract

Despite improvements in our understanding of the biology behind triple-negative breast cancer (TNBC), it remains a devastating disease due to lack of an effective targeted therapy. Inhibiting Wnt signaling is a promising strategy to combat TNBC because Wnt signaling drives TNBC progression, chemoresistance, and stemness. However, Wnt inhibition can lead to upregulation of autophagy, which confers therapeutic resistance. This provides an opportunity for combination therapy, as autophagy inhibitors applied concurrently with Wnt inhibitors could increase treatment efficacy. Here, we applied the autophagy inhibitor chloroquine (CQ) to TNBC cells in combination with Frizzled7 antibody-coated nanoshells (FZD7-NS) that suppress Wnt signaling by blocking Wnt ligand/FZD7 receptor interactions, and evaluated this dual treatment in vitro. We found that FZD7-NS can inhibit Axin2 and CyclinD1, two targets of canonical Wnt signaling, and increase the expression of LC3, an autophagy marker. When FZD7-NS and CQ are applied together, they reduce the expression of several stemness genes in TNBC cells, leading to inhibition of TNBC cell migration and self-renewal. Notably, co-delivery of FZD7-NS and CQ is more effective than either therapy alone or the combination of CQ with free FZD7 antibodies. This demonstrates that the nanocarrier design is important to its therapeutic utility. Overall, these findings indicate that combined regulation of Wnt signaling and autophagy by FZD7-NS and CQ is a promising strategy to combat TNBC.

1. Introduction

Triple-negative breast cancer (TNBC), which does not express estrogen receptor, progesterone receptor, and human epidermal growth factor receptor-2, accounts for 15%–20% of all breast cancers yet it suffers from lack of available targeted therapies [1]. Since TNBC is not susceptible to conventional targeted or hormonal therapies, patients with this disease suffer from earlier recurrence and lower survival rates than patients with other subtypes of breast cancer [2]. It is imperative that researchers develop effective targeted therapies specifically for TNBC. Here, we introduce a nanomedicine-based therapeutic strategy that exploits TNBC cells’ dependence on Wnt signaling and the interplay between this pathway and autophagy signaling.

It is well established that the developmental Wnt signaling pathway is dysregulated in TNBC and supports tumor formation, progression, and metastasis by facilitating cell survival, proliferation, and stem-like behavior [3–5]. Hence, manipulating Wnt signaling represents a promising strategy for TNBC therapy. Wnt signaling is activated in TNBC cells when extracellular Wnt ligands bind Frizzled7 (FZD7) receptors that are overexpressed on the cells’ surface [6]. This leads to cytoplasmic accumulation of β-catenin, the key mediator of Wnt signaling, and β-catenin then translocates to the nucleus where it activates the transcription of several downstream oncogenes, including Axin2, cyclin D1, and others (Scheme 1(a)). Since Wnt signaling is initiated in TNBC cells through FZD7 receptors, locking these molecules in a Wnt ligand-unresponsive state is an attractive therapeutic strategy. Previously, we reported that FZD7 antibody-coated nanoshells (FZD7-NS) could bind TNBC cells to suppress Wnt signaling and reduce cellular metabolic activity in vitro [7]. Impressively, FZD7-NS were much more effective than freely delivered FZD7 antibodies, which we attributed to their ability to exploit multivalent binding to yield enhanced signal cascade interference [7]. Here, we expand on the potential of this system by exploring its use in combination with the autophagy regulator chloroquine (CQ).

Scheme 1.

Depiction of the Wnt and autophagy pathways in TNBC. (a) Wnt signaling is activated in TNBC cells when Wnt ligands bind FZD7 cell surface receptors. This leads to nuclear translocation of β-catenin, which activates Wnt target genes to promote cell stemness and migration. (b) FZD7-NS suppress Wnt signaling in TNBC cells by blocking ligand/receptor interactions. When Wnt signaling is suppressed, autophagy is activated, contributing to cellular resistance. Applying the autophagy inhibitor CQ to TNBC cells in combination with FZD7-NS can overcome this resistance to impair aggressive cell behaviors.

In TNBC and other forms of cancer, Wnt signaling exhibits crosstalk with diverse signaling pathways [8–10]. For example, increasing evidence indicates that Wnt signaling and autophagy are inversely regulated, as inhibition of Wnt signaling leads to activation of autophagy in colorectal cancer, breast cancer, multiple myeloma, and liver cancer [11–15]. Autophagy, or “self-eating”, is a natural, regulated cellular process to degrade and eliminate misfolded proteins and damaged organelles in adaptation to starvation, hypoxia and other stress. Deregulation of autophagy is linked with different pathological conditions, including neurodegeneration, aging, and cancer [16]. Autophagy plays a complex role in cancer. In the early stages of tumor development, autophagy is tumor suppressive [16]. However, in later stages of cancer, autophagy provides tolerance to stress and confers therapeutic resistance [11]. Indeed, anticancer therapies, including Wnt antagonists, often activate autophagy [17]. This provides an opportunity for combination therapy, as the concomitant application of Wnt inhibitors and autophagy inhibitors should increase cancer cell killing [17, 18]. In this study, we evaluated the use of Wnt inhibitory FZD7-NS in combination with the autophagy blocker CQ as a treatment for TNBC (Scheme 1(b)).

To evaluate dual Wnt and autophagy regulation as a treatment for TNBC, we exposed TNBC cells cultured in serum-containing media first to FZD7-NS followed by CQ two days later, and performed several gene expression and functional cell assays. In initial studies, we examined the cellular trafficking of FZD7-NS, and found that they localize primarily with mitochondria. Analysis of Axin2 and CyclinD1 messenger RNA (mRNA) expression by quantitative real-time polymerase chain reaction (qRT-PCR) confirmed that FZD7-NS suppressed Wnt signaling in TNBC cells, and they also increased expression of the autophagy marker LC3. This effect was not observed with freely delivered FZD7 antibodies nor with the delivery of nanoshells coated with non-specific IgG antibodies (IgG-NS). Further analysis of several stemness genes revealed these were suppressed to the greatest extent when cells were exposed to FZD7-NS and CQ. Finally, transwell migration and spheroid formation assays revealed that the combined application of FZD7-NS and CQ reduced TNBC cell migration and self-renewal capabilities. These results confirm that targeting the interplay of Wnt signaling and autophagy is a promising strategy to combat TNBC and other cancers characterized by dependence on Wnt signaling.

2. Experimental

2.1. Nanoshell synthesis and antibody functionalization

Nanoshells with ~ 120 nm diameter spherical silica cores and ~ 15 nm thick gold shells were synthesized by the method of Oldenburg et al. as previously described [19]. To conjugate either rabbit anti-human FZD7 (Fisher Scientific, or LifeSpan Biosciences, Seattle, WA, USA) or rabbit anti-human IgG control (Genscript, Piscataway, NJ, USA) antibodies to NS, anti-FZD7 or anti-IgG were linked to 5 kDa orthopyridyl disulfidepoly(ethylene glycol)-succinimidyl valerate (OPSS-PEG_5k-SVA) (Laysan Bio, Arab, AL, USA) linkers in sodium bicarbonate. One part OPSS-PEG_5k-SVA was reacted with nine parts antibody at a 2:1 PEG-to-antibody molar ratio overnight at 4 °C to form OPSS-PEG_5k-FZD7 or OPSS-PEG_5k-IgG. Then, the PEGylated antibodies were purified with Amicon Ultra-4 centrifugal filters with a 10 kDa molecular weight cutoff (Millipore Sigma, Burlington, MA) to remove any unconjugated linkers. Finally, OPSS-PEG_5k-FZD7 or OPSS-PEG_5k-IgG was added to NS suspended in purified water at a concentration of approximately 4 × 10⁹ NS/mL at a ratio of 500–1,000 antibodies per NS and the samples reacted for 4 h at 4 °C. To passivate the NS, 5 kDa methoxy-poly(ethylene glycol)-thiol (mPEG_5k-SH) (Laysan Bio, Arab, AL, USA) was added to the samples to a final concentration of 20 μM and reacted overnight at 4 °C. To produce Cy5-labeled FZD7-NS or IgG-NS, Cy5-PEG_5k-SH (Biochempeg Scientific, Watertown, MA, USA) was used in place of mPEG_5k-SH. After antibody and PEG functionalization, FZD7-NS and IgG-NS were purified by centrifugation (800g, 10 min) using LoBind Protein microcentrifuge tubes (Eppendorf, Hauppauge, NY, USA). The antibody-conjugated NS formed a pellet at the bottom of the tubes and unbound molecules were removed with the supernatant. Purified FZD7-NS or IgG-NS were diluted in water or medium as indicated for each experiment, and freshly prepared samples were used in all studies.

2.2. Characterization of nanoshell conjugates

The extinction profile of bare NS, IgG-NS, and FZD7-NS suspended in water was characterized by UV–visible spectroscopy (Cary60 spectrometer, Agilent, Santa Clara, CA, USA). The hydrodynamic diameter and zeta potential of bare NS, IgG-NS, and FZD7-NS (suspended in water to an optical density (OD) of 1 at 800 nm wavelength, corresponding to 2.7 × 10⁹ NS/mL) were measured using a Litesizer500 instrument (Anton Paar, Graz, Austria), and the reported hydrodynamic diameter (using z-average mean) and zeta potential are the mean of three sample measurements. FZD7-NS samples for scanning electron microscopy (SEM) were diluted to 2.7 × 10⁹ NS/mL (OD 1) in 200 proof ethanol and dried directly onto a polished carbon stub prior to imaging (S4700, Hitachi, Chiyoda, Tokyo, Japan).

2.3. Antibody loading quantification

The loading of FZD7 antibodies or IgG control antibodies on NS was quantified by a solution-based ELISA described previously [7]. Briefly, FZD7-NS, IgG-NS, or PEG-NS were incubated with 10 μg/mL horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (HRP-AR, KPL) for 1 h at room temperature. Samples were pelleted by centrifugation (500g, 5 min, thrice) to remove unbound secondary antibodies in the supernatant and then suspended in 3% bovine serum albumin in PBS (PBSA). Samples were developed in 3,3′,5,5′-tetramethylbenzidine (TMB core) (Bio-Rad, Hercules, CA) for 5 min and then sulfuric acid was added to stop the reaction. Absorbance at 450 nm was measured on a Hybrid Synergy H1M plate reader and compared to a standard curve of known HRP concentration to calculate the number of FZD7 or IgG antibodies per NS.

2.4. Binding avidity of FZD7-NS

The binding avidity of FZD7-NS to TNBC cells was analyzed as described previously with minor modification [7]. The effective dissociation constant $K_{\text{D}}^{\text{eff}}$ was determined by fitting data to a model in JMP software using the equation $\text{OD}=\frac{{\text{Ab}}_{\text{conc}}}{K_{\text{D}}^{\text{eff}}+{\text{Ab}}_{\text{conc}}}$. In this equation, Ab_conc is the concentration of antibody added to cells and $K_{\text{D}}^{\text{eff}}$ represents the binding avidity. Further, OD is the optical density at each antibody concentration normalized according to the equation $\text{OD}=\frac{{\text{OD}}_{\text{raw}}-{\text{OD}}_{\text{back}}}{{\text{OD}}_{\text{high}}-{\text{OD}}_{\text{back}}}$, where OD_raw is the initial optical density reading prior to any calculations, OD_back is the optical density of background signal (cells treated with no primary antibody and treated with secondary antibody), and OD_high is the optical density of the highest (saturated) signal.

2.5. Cell culture and stable gene expression

Human MDA-MB-231 TNBC cells (American Type Culture Collection, Manassas, VA, USA) and 293TN cells (System Biosciences, Palo Alto, CA, USA) were cultured in Dulbecco”s Modified Eagle Medium (DMEM) (VWR, Radnor, PA, USA) supplemented with 10% fetal bovine serum (FBS) (Gemini Bio-Products, West Sacramento, CA, USA) and 1% penicillin/streptomycin (Thermo Fisher Scientific, Waltham, MA, USA), and maintained in a humidified incubator at 37 °C, 5% CO₂. For spheroid formation experiments, MDA-MB-231 were cultured in MammoCult (STEMCELL Technologies, Vancouver, BC, Canada) medium supplemented with proliferation supplement, heparin (4 μg/mL) and hydrocortisone (0.48 μg/mL). To study the intracellular trafficking of Ab-NS, MDA-MB-231 cells were stably transduced with GFP-tagged endosomal markers using lentiviral transduction procedures as described previously [20]. Briefly, Rab5-Clover (Addgene # 56530), Rab8a-EGFP (Addgene # 86075), Rab11-GFP (Addgene # 12674), or LAMP1-mGFP (Addgene # 34831) were cloned into a lentiviral transfer vector (System Biosciences, Palo Alto, CA, USA) by restriction cloning. Lentiviral particles were produced by triple-transfecting (TransIT-Lenti transfection reagent; Mirus Bio, Madison, WI, USA) 293TN cells (System Biosciences, Palo Alto, CA, USA) with either transfer vector and lentiviral packaging and envelope plasmids (Addgene #12260,12259). Lentivirus was harvested, filtered, and diluted in cell culture medium to transduce MDA-MB-231 cells. Cells stably expressing the desired fusion protein were selected with 1 mg/mL puromycin (VWR, Radnor, PA, USA).

2.6. Analysis of nanoparticle colocalization with intracellular compartments

Wild type MDA-MB-231 cells or MDA-MB-231 cells stably expressing Rab5-Clover, Rab8a-EGFP, Rab11-GFP, or LAMP1-mGFP were seeded in 8-well Lab-Tek™ chambered coverglasses (VWR, Radnor, PA) to 60%–70% confluence. Cells were incubated with Cy5-labeled FZD7-NS or IgG-NS at a concentration of 2.7 × 10¹⁰ NS/mL for 24 h. GFP-expressing cells were counterstained with 1 μM Hoechst 33342 for 15 min to label cell nuclei. For mitochondria staining, wild type cells were treated with 1 μM Hoechst 33342 and 1 μM Rhodamine123 after the nanoparticles had been applied. Cells were then washed with 1× PBS and imaged with a Zeiss Axioobserver Z1 Inverted Fluorescent Microscope equipped with an apotome using the GFP (intracellular compartment) (excitation, 495 nm; emission, 525/50 nm), Cy5 (NS) (excitation, 615 nm; emission, 675/100 nm), and DAPI (nuclei) (excitation, 395 nm; emission, 445/50 nm) fluorescence channels. For colocalization analysis of antibody-conjugated NS and different subcellular compartments, Manders’ colocalization coefficients [21] were calculated for each compartment visualized under the EGFP fluorescence channel using the ImageJ plugin EzColocalization [22].

2.7. Western blot analysis for LC3 protein expression

To assess LC3 protein expression, MDA-MB-231 cells were seeded at 50,000 cells per well in 24 well plates, and treated with complete medium containing 2.7 × 10¹⁰ NS/mL IgG-NS or FZD7-NS, or 4.5 nM free FZD7 antibodies (equivalent to antibody loading on NS) or 10 μM rapamycin for 24 h. Cells were washed 1× with phosphate-buffered saline (PBS) and lysed in RIPA buffer (Amresco) supplemented with Halt Protease Inhibitor Cocktail (Life Technologies) on ice. Protein concentration was determined using a DC Protein Assay (Bio-Rad, Hercules, CA, USA) relative to a BSA standard. A 30 μg aliquot of protein was loaded into each well of 4%–12% Bis-tris gels (ThermoFisher Scientific, Waltham, MA, USA) and separated at 135 V for 60 min. The protein was then transferred to a 0.45 μm nitrocellulose membrane for 10 min using the Pierce Power System (Thermo Scientific). Membranes were blocked for 60 min in 5% nonfat milk in tris-buffered saline containing 0.1% Tween-20 (TBST). Membranes were probed with rabbit anti-human LC3 (1:100) or mouse anti-human β-actin (1:2,000) primary antibodies (Cell Signaling Technology, Danvers, MA, USA) at 4 °C overnight, followed by incubation with goat anti-rabbit (1:12,500) or anti-mouse (1:25,000) horseradish peroxidase-conjugated secondary antibodies (Kirkegaard & Perry Laboratories, Inc., Gaithersburg, MD, USA) for 1 h at room temperature. Protein bands were detected by chemiluminescence using VisiGlo^™ Select Chemiluminescent Substrate (Amresco, Solon, OH, USA) and imaged on a ChemiDoc-It2 Imager (UVP, Upland, CA, USA). The blot shown represents the average band density across experiments.

2.8. Assessment of cellular metabolic activity in response to CQ

To identify a dose of CQ for use in subsequent studies, MDA-MB-231 cells were seeded into 96-well plates at 20,000 cells/well, then treated with media containing 0‒500 μM CQ. After 24 h incubation, the medium was removed and an Alamar blue viability reagent (ThermoFisher, Waltham, MA, USA; diluted 1:10 in complete cell culture medium) was added for further incubation per manufacturer recommendations. Sample fluorescence was measured on a Synergy H1M plate reader with excitation and emission wavelengths of 560 and 590 nm, respectively. To analyze the data, background (Alamar blue reagent without cells) was subtracted from the fluorescence reading in each well.

2.9. Analysis of mRNA by qRT-PCR

qRT-PCR was employed to evaluate the mRNA expression of several genes in MDA-MB-231 cells after treatment with IgG-NS, free FZD7, or FZD7-NS alone or in combination with CQ. MDA-MB-231 cells were cultured in 6-well plates in complete medium at a density of 200,000 cells per well, and then treated with 4.5 nM free FZD7 antibodies, or with FZD7-NS or IgG-NS at a concentration of 2.7 × 10¹⁰ NS/mL on day 1. On day 3, samples received either 0 or 50 μM CQ. Finally, on day 4, total RNA was isolated from treated cells using an Isolate II RNA Mini Kit (Bioline, Taunton, MA, USA), and qRT-PCR was performed using SensiFAST SYBR One-Step Master Mix on a LightCycler 96 instrument (Roche Diagnostics Corporation, Indianapolis, IN, USA). Gene expression was normalized to that of GAPDH. Primer sequences are listed in Table S1 in the Electronic Supplementary Material (ESM).

2.10. Transwell migration assay

Transwell assays were employed to evaluate the migration ability of MDA-MB-231 cells after treatment. MDA-MB-231 cells stably expressing GFP were cultured in 96-well plates in complete medium at a density of 10,000 cells per well. On day 1, cells were treated with complete medium containing IgG-NS or FZD7-NS at 2.7 × 10¹⁰ NS/mL, or with 4.5 nM free FZD7 antibodies (equivalent to antibody loading on NS). On day 3, only cells subjected to combination therapy were supplemented with CQ to a final concentration of 50 μM. On day 4, cells were trypsinized, resuspended in serum-free DMEM at a density of 3,000 cells per 200 μL media and loaded into a transwell insert of 8 μm pore size (Corning, Corning, NY, USA) placed in a 24-well transwell apparatus (Cellvis, Mountain View, CA, USA). The bottom chamber was loaded with 800 μL complete DMEM media and then the samples were cultured at 37 °C, 5% CO₂. After 1 and 5 d, the entirety of each well was imaged by fluorescence microscopy on a Zeiss AxioObserver Z1 microscope to visualize cells that had passed through the membrane using the GFP channel (excitation, 495 nm; emission, 525/50 nm). The number of cells that passed through the filter membrane was counted using ImageJ. Quantitative data were reported as the average number of cells per treatment group across three independent replicates.

2.11. Spheroid formation assay

A spheroid formation assay was employed to evaluate the self-renewal capacity of MDA-MB-231 cells after treatment. MDA-MB-231 cells were cultured in 96-well plates in complete medium at a density of 10,000 cells per well. Cells were treated with free FZD7 antibodies, FZD7-NS, IgG-NS, either with or without CQ in an identical manner to transwell experiments. On day 4, cells were seeded as a single-cell suspension in 24-well Costar® low-adhesion plates (Corning, Corning, NY, USA) at a density of 250 cells per well in MammoCult (STEMCELL Technologies, Vancouver, BC, Canada) medium supplemented with proliferation supplement, heparin (4 μg/mL) and hydrocortisone (0.48 μg/mL). After incubation for 7 d at 37 °C, 5% CO₂, the entirety of each well was imaged by bright-field microscopy using a Zeiss AxioObserver Z1 microscope. The images shown represent only a small region of interest within the well, but are representative of the spheroid growth in each treatment group. To quantify spheroid formation in each sample, the number of spheroids with 4 or more cells in each treatment group was measured across the entirety of each well. Spheroid growth data are reported as the average number of spheroids per treatment group across three independent replicates.

2.12. Statistical analysis

All experiments were performed in triplicate, and data represent means ± standard deviations from three independent replicates unless otherwise indicated. Groups with significant differences were identified using one-way ANOVA with a post-hoc Turkey test, and differences were considered significant at p < 0.05. Statistical tests were performed in Minitab software (Minitab, State College, PA), and flow cytometry data were analyzed using NovoExpress software (ACEA Biosciences, San Diego, CA, USA).

3. Results and discussion

3.1. Synthesis and characterization of antibody-nanoshell conjugates

Nanoshells (NS) composed of ~ 120 nm diameter silica cores and ~ 15 nm thick gold shells (total diameter ~ 150 nm) were synthesized per established methods [19], and coated with antibodies and methoxy-poly(ethylene glycol)-thiol (mPEG-SH) as shown in Fig. 1(a). We developed two types of NS formulations for these studies: NS coated with FZD7 antibodies and mPEG-SH (FZD7-NS) and NS coated with non-specific IgG antibodies and mPEG-SH (IgG-NS). Compared to our prior study [7], which used 2 kDa heterobifunctional PEG linkers to attach antibodies to NS, here we utilized 5 kDa PEG linkers in order to increase the flexibility and presentation of antibodies on the NS and increase the loading density. UV–visible spectrophotometry revealed that bare NS had a peak plasmon resonance near 800 nm (Fig. 1(b)), which is consistent with the optical properties of 150 nm diameter NS [23]. A slight bathochromic-shift was observed when the NS were coated with IgG or FZD7 antibodies, providing evidence of successful surface functionalization. The functionalization was further confirmed by dynamic light scattering (DLS), which showed an increased hydrodynamic diameter from bare NS (151.9 ± 1.6 nm) to IgG-NS (185.1 ± 3.4 nm) and FZD7-NS (177.4 ± 1.6 nm) (Fig. 1(c)). The zeta potential of the nanoparticles also increased from −35 mV for bare NS to −27.3 mV for IgG-NS and −26.6 mV for FZD7-NS (Fig. 1(c)). Scanning electron microscopy (SEM) images indicated that FZD7-NS were highly monodisperse with respect to their size and shape (Fig. 1(d)). Antibody loading on NS was quantified using a previously reported ELISA assay [7, 24]. IgG-NS and FZD7-NS had similar antibody loading of 122 ± 4 and 115 ± 8 antibodies per nanoshell, respectively (Fig. 1(e)). This loading is higher than that of the antibody-NS conjugates we created previously using 2 kDa PEG linkers [7].

Figure 1.

Characterization of antibody-nanoshell (Ab-NS) conjugates. (a) Scheme depicting the synthesis of Ab-NS. (b) Extinction spectra of bare nanoshells versus nanoshells coated with IgG or FZD7 antibodies. (c) Hydrodynamic diameter (black) and zeta potential (blue) of bare NS, IgG-NS, and FZD7-NS. (d) SEM image of FZD7-NS. (e) Quantification of antibody loading on IgG-NS and FZD7-NS.

To evaluate the binding avidity of FZD7-NS to TNBC cells, we used a modified Langmuir isotherm model, as introduced by De Puig et al. [25]. and described in our prior study [7]. The averaged effective dissociation constant for FZD7-NS determined by this model among all three experiments was 6.03 × 10⁻¹⁰ ± 1.14 × 10⁻¹⁰ M (Fig. S1 in the ESM), which is similar to the value of 4.9 × 10⁻¹⁰ M we previously determined for FZD7-NS that were made using a 2 kDa PEG linker. Further studies would be needed to understand the relationship between PEG length, antibody loading, and strength of antibody-NS binding to cancer cells with varying levels of FZD7 expression.

3.2. FZD7-NS localize primarily with mitochondria in TNBC cells

To determine if FZD7-NS are internalized by TNBC cells or remain bound to the cell surface following binding, fluorescence microscopy was used to monitor the intracellular trafficking of Cy5-labeled FZD7-NS and IgG-NS after their addition to the cell culture media. MDA-MB-231 TNBC cells were engineered to stably express fluorescent markers of different intracellular compartments, including Rab5-Clover, Rab8a-EGFP, Rab11-GFP, and LAMP1-mGFP, which label early endosomes, slow recycling endosomes, endocytic recycling compartments (ERC), and lysosomes, respectively. The fluorescent dye Rhodamine123 was used to label mitochondria. Cells were incubated with Cy5-FZD7-NS or Cy5-IgG-NS for 24 h and imaged by fluorescence microscopy. To quantitatively assess colocalization between the antibody-coated NS and subcellular compartments, Manders’ colocalization coefficients (MCCs) were calculated for each subcellular compartment using the ImageJ plugin EzColocalization [22]. MCCs were calculated as the fractional overlap of red (NS) and green (subcellular compartment) fluorescent signals and range from 0 to 1, with MCC = 0 indicating no colocalization and MCC = 1 indicating perfect colocalization. We calculated both MCC₁ (red/green overlap) to reveal the fraction of FZD7-NS and IgG-NS within specific subcellular compartments, and MCC₂ (green/red overlap) to evaluate the fraction of each subcellular compartment that is loaded with NS.

We found that Cy5-FZD7-NS colocalize to the greatest extent with mitochondria (MCC₁ = 0.79), which was significantly greater than colocalization with early endosomes (MCC₁ = 0.5), slow recycling endosomes (MCC₁ = 0.62), endocytic recycling compartments (MCC₁ = 0.31), or lysosomes (MCC₁ = 0.48) (Figs. 2(a) and 2(b)). Additionally, a higher fraction of mitochondria contained Cy5-FZD7-NS (MCC₂ = 0.92) than other endosomal compartments (MCC₂ = 0.21–0.47) (Figs. 2(a) and 2(b)). Images acquired at 1 h post-incubation showed the same trend (data not shown). Therefore, we conclude that FZD7-NS are internalized by TNBC cells and primarily localize to mitochondria.

Figure 2.

Analysis of the intracellular trafficking of Cy5-labeled FZD7-NS. (a) Fluorescence microscopy images showing overlap of Cy5-labeled FZD7-NS (red) with various subcellular compartments (green) 24 h after the nanoparticles were added to the cell culture medium. Scale bars = 20 μm. (b) Results from quantitative colocalization analysis to calculate the fractional overlap of red-fluorescent FZD7-NS with green-fluorescent subcellular compartment and vice versa. MCC = Manders’ colocalization coefficient; letters a–d indicate grouping by one-way ANOVA with post hoc Turkey. Groups with different letters are significantly different at the 99% confidence level, except for the comparison of Rab8a and Rab 11 for MCC₂, which has p = 0.026.

We were curious whether the mitochondrial accumulation observed for FZD7-NS would also extend to IgG-NS, so the same imaging studies were performed for this nanoparticle formulation. Intriguingly, we found a similar distribution for IgG-NS, with most of these nanoparticles colocalizing with mitochondria (Figs. S2(a) and S2(b) in the ESM). The observation that FZD7-NS and IgG-NS colocalize with mitochondria is interesting because most nanoformulations become entrapped in the endo-lysosomal system [26–29]. More studies are needed to understand the uptake pathways of these nanoparticles and the mechanism by which they accumulate in mitochrondria, although similar mitochondrial targeting has been observed for serum protein-coated gold nanorods [30]. Given that our studies were performed in serum-containing medium, it would be intriguing to determine whether the serum content of the media influences the nanoparticles’ ultimate fate. Nevertheless, these results confirm that FZD7-NS and IgG-NS are internalized by TNBC cells and are exciting given that mitochondria perform essential cell functions (including regulation of autophagy) and massive efforts are currently being put towards developing mitochondria-targeted therapies [31–37]. We next aimed to evaluate the ability of FZD7-NS, free FZD7 antibodies, and IgG-NS to suppress Wnt signaling and activate autophagy in TNBC cells.

3.3. FZD7-NS inhibit Wnt signaling and promote autophagic flux in TNBC cells to a greater extent than free FZD7 antibodies or IgG-NS

As noted above, our previous studies demonstrated that FZD7-NS coated with antibodies using 2 kDa PEG linkers could effectively block Wnt signaling in TNBC cells that were pretreated with serum starvation [7]. Here, we aimed to confirm that FZD7-NS coated with antibodies using 5 kDa PEG linkers could suppress Wnt signaling in TNBC cells cultured in normal serum-containing media. We also sought to determine the impact of CQ co-treatment on gene expression. To investigate this, qRT-PCR was employed to evaluate the mRNA expression of Axin2 and Cyclin D1, which are both downstream targets of the canonical Wnt pathway. MDA-MB-231 cells were treated with IgG-NS, free FZD7 antibodies, or FZD7-NS at equivalent antibody doses, and two days later the cells were treated either with or without CQ. The dose of CQ to be applied was determined by first evaluating the impact of CQ on TNBC cell viability after the drug was administered to cells for 24 h at concentrations ranging from 0 to 500 μM. MDA-MB-231 cell viability began to decrease at doses above 100 μM CQ (Fig. S3 in the ESM), so 50 μM was selected at the CQ dose for PCR experiments and subsequent studies.

The qRT-PCR studies revealed that both Axin2 and Cyclin D1 mRNA expression significantly decreased in MDA-MB-231 cells treated with FZD7-NS alone or in combination with CQ (Figs. 3(a) and 3(b), and Table S2 in the ESM). In contrast, cells treated with IgG-NS or free FZD7 antibodies alone or with CQ did not exhibit reduced gene expression. This is consistent with our previous observation that FZD7-NS are much more effective than free FZD7 antibodies at suppressing Wnt signaling in TNBC cells [7]. We did note a slight decrease in Cyclin D1 expression for cells exposed to IgG-NS, which may be due to the presence of nanoparticles impacting cell-cell communication. Interestingly, Axin2 mRNA expression was slightly higher in cells treated with IgG-NS + CQ than those treated with only IgG-NS (Figs. 3(a) and 3(b)). This observation is consistent with previous reports that autophagy negatively regulates Wnt signaling [38]. However, this trend was not observed when comparing cells exposed to free FZD7 antibodies ± CQ or when comparing cells exposed to FZD7-NS ± CQ. This indicates that free FZD7 antibodies and FZD7-NS may overcome any CQ-mediated activation of Wnt signaling, and highlights the complexity of the crosstalk between Wnt signaling and autophagy.

Figure 3.

Examination of Wnt signaling and autophagy markers in TNBC cells exposed to various treatments. (a) and (b) qRT-PCR analysis of Axin2 and Cyclin D1 mRNA expression in MDA-MB-231 cells following treatment with media only (non-treated, NT), IgG-NS ± CQ, free FZD7 antibodies ± CQ, or FZD7-NS ± CQ. * p < 0.05 versus NT by one-way ANOVA with post hoc Turkey. # p < 0.05 by one-way ANOVA with post hoc Turkey. Precise p-values for all significant comparisons are provided in Table S2 in the ESM. (c) Western blot analysis of LC3 protein expression in MDA-MB-231 cells exposed to the autophagy promoter rapamycin, no treatment (NT), FZD7-NS, IgG-NS, or free FZD7 antibodies.

To confirm our hypothesis that Wnt inhibition mediated by FZD7-NS might upregulate autophagy in MDA-MB-231 cells, we employed Western blot to probe the protein expression of LC3, a marker of autophagy signaling, in cells exposed to various treatments. Specifically, cells were treated with IgG-NS, FZD7-NS, or free FZD7 antibodies at equivalent antibody doses for 24 h, or they were treated with rapamycin as a positive control that should activate autophagy. As expected, MDA-MB-231 cells treated with rapamycin exhibited a dramatic increase in LC3 protein expression compared to cells that received no treatment (NT) (Fig. 3(c)). By comparison, cells treated with IgG-NS or free FZD7 antibodies showed minimal increases in LC3 protein expression, which is probably attributable to the role of autophagy as a master regulator of cell response to stresses to maintain homeostasis. Importantly, cells treated with FZD7-NS exhibited a pronounced increase in LC3 protein level. These data support our hypothesis that Wnt signaling inhibition mediated by FZD7-NS can upregulate autophagy in MDA-MB-231 cells. This result agrees with prior work by Fu et al., who showed that Resveratrol can induce autophagy in MCF-7 and SUM159 breast cancer cells via Wnt signaling inhibition [14].

3.4. Dual inhibition of Wnt signaling and autophagy mediated by FZD7-NS and CQ impairs TNBC cell migration

Several studies have shown that active Wnt signaling promotes TNBC cell migration [3, 4, 39]. Indeed, Wnt signaling inhibition by pharmacological agents, shRNA, or siRNA therapeutics has been shown to reduce TNBC cell migration in vitro [4, 39]. Further, a recent study showed that silencing FSIP1 (fibrous sheath interacting protein 1) in TNBC cells reduces β-catenin nuclear localization, increases LC3 expression, elevates autophagosome production, and decreases cell migration [40]. These observations prompted us to investigate whether applying both FZD7-NS and CQ to TNBC cells could impair cell migration using transwell assays.

In these experiments, GFP-expressing MDA-MB-231 cells were pretreated with nothing, IgG-NS, free FZD7, or FZD7-NS for 48 h, followed by CQ (or no drug) for 24 h. The cells were then detached from their culture plate, suspended in serum-free media, and transferred to the top insert of a transwell apparatus (Fig. 4(a)). Over the subsequent 5 d, the samples were imaged by fluorescence microscopy to observe the number of cells that had migrated through the insert towards the serum-containing medium on the other side. After 1 d, cells treated with FZD7-NS + CQ showed reduced migration versus the non-treated group (Figs. S4(a) and S4(b) in the ESM). Further, FZD7-NS + CQ was significantly more effective at suppressing TNBC cell migration than free FZD7 antibodies + CQ (Figs. S4(a) and S4(b) in the ESM). The migration ability was evaluated again after 5 d, and this excitingly revealed the response was durable, as the migration of cells exposed to FZD7-NS + CQ was reduced over 80% compared to that of untreated cells (Figs. 4(b) and 4(c)). At this time point, FZD7-NS + CQ combination therapy was also still significantly more effective than free FZD7 + CQ. Although IgG-NS + CQ exhibited some ability to reduce cell migration, the difference was not statistically significant compared to the non-treated group (Figs. 4(b) and 4(c)). Together, these results demonstrate that dual Wnt and autophagy inhibition via FZD7-NS + CQ can impair TNBC cell migration to a greater extent than Wnt inhibition alone.

Figure 4.

Evaluation of TNBC cell migration in response to Wnt and autophagy inhibition. (a) Scheme depicting the transwell migration assay. (b) Representative fluorescence images of migrated GFP-expressing MDA-MB-231 cells 5 d after they were seeded in the transwell apparatus following treatment with NT, IgG-NS ± CQ, free FZD7 ± CQ, or FZD7-NS ± CQ. Scale = 1,000 μm. (c) Quantitative analysis of the number of migrated cells in each treatment group. # p < 0.01 versus FZD7-NS + CQ by one way ANOVA with post hoc Turkey. * p < 0.01 versus NT by one-way ANOVA with post hoc Turkey.

3.5. Combined application of FZD7-NS and CQ reduces the expression of several genes associated with aggressive TNBC behavior and limits TNBC cell self-renewal

We next sought to investigate how dual inhibition of Wnt and autophagy affects the stem-like and aggressive behavior of TNBC cells. Cancer stem cells (CSCs) are a small subpopulation of tumor cells that have been implicated in tumor initiation, treatment resistance, relapse, angiogenesis, and metastasis [41]. Most conventional cancer therapies fail because they do not eliminate CSCs. Both Wnt and autophagy signaling pathways are associated with cancer stemness [42–44], making their inhibition a promising approach to eliminate CSCs [45, 46]. To evaluate whether FZD7-NS + CQ treatment could reduce the stem-like behavior of TNBC cells, we used qRT-PCR to examine the expression of several genes associated with stemness and aggressive phenotypes (KLF4 [47, 48], Nanog [48, 49], Oct4 [50, 51], Epcam [52, 53], CD90 [54, 55], Survivin [56] and Bcl-2 [46]) in response to treatment. We also evaluated the impact of treatment on self-renewal by performing spheroid formation assays.

The qRT-PCR results showed that several genes (KLF4, Nanog, Oct4, and Bcl-2) were elevated in cells exposed to IgG-NS + CQ and/or to free FZD7 + CQ (Fig. 5, and Table S3 in the ESM). We postulate that the increased expression of these genes is a compensation mechanism for cell survival when autophagy is inhibited. This would agree with the findings of Cho et al., who showed that Oct4, Nanog, KLF4, and several other stemness genes were elevated in human embryonic stem cells treated with the autophagy inhibitor Bafilomycin A [57]. Indeed, KLF4, Oct4, and Nanog have all been shown to have interplay with autophagy [57–61]. Likewise, Bcl-2 is an anti-apoptotic gene [46], so its activation is likely a cellular compensation mechanism. Notably, in our studies the application of FZD7-NS alone or in combination with CQ counteracted this mechanism, as the expression of all four genes (KLF4, Nanog, Oct4, and Bcl-2) was decreased under these conditions (Fig. 5). The most profound impact was observed for KLF4, Nanog, and Oct4, whose expression decreased by 57%, 60%, and 18%, respectively, in FZD7-NS + CQ treated cells relative to controls.

Figure 5.

Analysis of mRNA expression of several genes in MDA-MB-231 cells treated with FZD7-NS + CQ or controls. GAPDH was used as a housekeeping gene, and expression in each treatment group was normalized to that in the NT group. Data are means ± standard deviations. * indicates p < 0.05 versus NT and # indicates p < 0.05 versus FZD7-NS + CQ by one-way ANOVA with post hoc Turkey. Other significant differences are not shown for simplicity. Precise p-values for all significant comparisons are provided in Table S3 in the ESM.

We also found that the mRNA expression of Epcam, CD90, and Survivin was most affected by FZD7-NS + CQ, as cells exposed to this combination therapy exhibited reductions in mRNA expression of 46% (Epcam), 58% (CD90), and 40% (Survivin), relative to control non-treated cells (Fig. 5). Epcam and CD90 are surface markers of CSCs that are associated with cancer cell proliferation, metastasis, and angiogenesis [62, 63]. Survivin, the smallest member of the inhibitor of apoptosis (IAP) family, is a cancer biomarker that is associated with apoptosis inhibition, chemoresistance, and tumor aggressiveness [64]. Studies have also shown that survivin is regulated by Wnt signaling and is a marker for quiescent breast cancer stem cells [65], consistent with our finding that survivin is decreased by FZD7-NS + CQ in TNBC cells. Overall, our qRT-PCR data provide evidence that combined application of FZD7-NS + CQ can reduce the expression of stemness-associated genes in TNBC cells.

Based on our data indicating that FZD7-NS + CQ can suppress several genes associated with stem-like and aggressive behavior in TNBC cells, we next sought to evaluate whether dual Wnt and autophagy inhibition can functionally impair TNBC self-renewal using spheroid formation assays. In these experiments, MDA-MB-231 cells cultured in adherent conditions were pretreated with IgG-NS, free FZD7, or FZD7-NS in the presence or absence of CQ for 72 h and the cells were then dissociated and transferred to new plates in a single cell suspension in MammoCult medium to form spheroids over a period of 7 d (Fig. 6(a)). Bright-field images (Fig. 6(b)) and quantitative analysis of the number of spheroids with > 4 cells (Fig. 6(c)) show that the total number of spheroids did not decrease in cells treated with IgG-NS compared to non-treated (NT) cells, but cells exposed to free FZD7 antibodies produced ~ 20% fewer spheroids and cells exposed to FZD7-NS formed ~ 50% less spheroids. When CQ was added to each of these treatment regimens, it significantly reduced the number of spheroids formed compared to each monotherapy, indicating the importance of dual Wnt and autophagy inhibition. Specifically, the amount of spheroids formed in IgG-NS + CQ treated cells was ~ 24% that of NT cells, and the amount formed in free FZD7 + CQ treated cells was ~ 35% that of NT cells. Impressively, no spheroids were formed from cells treated with FZD7-NS + CQ (Figs. 6(b) and 6(c)). Although this reduction was not statistically significant versus the IgG-NS + CQ treatment group, it was significant versus all other treatment groups (Fig. 6(c)). These results, combined with the qRT-PCR data, suggest that dual inhibition of Wnt and autophagy can impair the self-renewal capacity of MDA-MB-231 cells, which may be attributed to impacts on the aggressive CSC subpopulation.

Figure 6.

Self-renewal capacity of MDA-MB-231 cells subjected to Wnt and autophagy inhibition. (a) Scheme depicting the spheroid culture model; red cells illustrate breast cancer stem cells. (b) Representative bright-field images of spheroids cultured from MDA-MB-231 cells after exposure to NT, IgG-NS ± CQ, free FZD7 ± CQ, or FZD7-NS ± CQ. Scale bars = 50 μm. (c) Quantitative analysis of the number of spheroids with > 4 cells in each treatment group. *p < 0.01 compared to NT, while ^# p < 0.05 and ^## p < 0.01 compared to FZD7-NS + CQ by one-way ANOVA with post hoc Turkey. Other comparisons are not depicted for simplicity.

4. Conclusions

Wnt signaling and autophagy play a complex role in TNBC, and each of these pathways have been identified as promising targets for therapeutic intervention [5, 17, 66–68]. Since inhibiting Wnt signaling in TNBC and other cancers has been shown to upregulate autophagy signaling [11–15], we wanted to take advantage of this crosstalk to develop a more effective treatment for TNBC. Recently, Nager et al. showed that inhibiting Wnt signaling in glioblastoma cells upregulates autophagy through mTOR inhibition, and this renders the cells more sensitive to autophagy blockers [11]. This encouraged us investigate whether blocking autophagy with CQ after inhibiting Wnt signaling with FZD7-NS might be an effective treatment strategy for TNBC. Currently, CQ and its derivative hydroxychloroquine (HCQ) are the only FDA-approved autophagy inhibitors for cancer therapy [69]. Moreover, CQ has been found to improve drug delivery and efficacy by reducing immunological clearance of nanoparticles, normalizing tumor vasculature, and suppressing several oncogenic and stress tolerance pathways, such as autophagy [70]. Therefore, we chose CQ as the autophagy inhibitor for our study.

In initial experiments, we determined that FZD7-NS are internalized by TNBC cells and colocalize primarily with mitochondria (Fig. 2). This is an intriguing result deserving of further investigation in future studies. In agreement with our prior report [7], we found that FZD7-NS inhibited Wnt signaling in MDA-MB-231 cells more effectively than free FZD7 antibodies, and they also upregulated the autophagy marker LC3 to a greater extent than free FZD7 antibodies (Fig. 3). Moreover, we showed that several genes (KLF4, Oct4, Nanog, Bcl-2, Survivin, Epcam, and CD90) that are associated with CSCs and implicated in cancer cell migration, self-renewal, and other aggressive behaviors [47, 49, 51, 52, 62, 64, 71], are suppressed in TNBC cells exposed to FZD7-NS + CQ (Fig. 5). This reduced gene expression corresponded with decreases in cell migration (Fig. 4) and impaired self-renewal ability (Fig. 6). These findings agree with prior studies that show loss of these genes hinders TNBC growth. For example, Zhou et al. demonstrated that KLF4 inhibition reduces the spheroid formation capacity of TNBC cells [72], consistent with our results. Likewise, Dey et al. and Bilir et al. showed that blockade of Wnt signaling suppresses TNBC cell migration and reduces spheroid formation [4, 39]. Our results also agree with several studies that demonstrate CQ can impair the spheroid formation and migration of TNBC cells [73–75]. Taken together, our results confirm that dual Wnt and autophagy inhibition mediated by FZD7-NS and CQ can impair the aggressive behavior of TNBC cells.

Moving forward, in vivo studies will need to be performed to validate dual Wnt and autophagy inhibition as a strategy to reduce TNBC tumor growth, recurrence, and metastasis. Before testing FZD7-NS extensively in vivo, however, it will be imperative to determine how antibody loading density, PEG length, and other features of the nanoparticle design impact their efficacy. Dosing optimization should also be performed to determine what concentrations of FZD7-NS and CQ are synergistic, and to reveal the best dosing schedule. Given the positive results observed in this study, it is likely that continued development of this treatment strategy would yield an effective regimen to eliminate TNBC. In conclusion, the combined application of FZD7-NS to inhibit Wnt signaling and CQ to suppress autophagy is an intriguing and promising approach to combat TNBC that is deserving of further investigation.