Advancing Breast Cancer Therapeutics: Targeted Gene Delivery Systems Unveiling the Potential of Estrogen Receptor-Targeting Ligands

Although curcumin has been well known as a phytochemical drug that inhibits tumor promotion by modulating multiple molecular targets, its potential was not reported as a targeting ligand in the field of drug delivery system. Here, we aimed to assess the tumor-targeting efficiency of curcumin and its derivatives such as phenylalanine, cinnamic acid, coumaric acid, and ferulic acid. Curcumin exhibited a high affinity for estrogen receptors through a pull-down assay using the membrane proteins of MCF-7, a breast cancer cell line, followed by designation of a polymer-based gene therapy system. As a basic backbone for gene binding, dextran grafted with branched polyethylenimine was synthesized, and curcumin and its derivatives were linked to lysine dendrimers. In vitro and in vivo antitumor effects were evaluated using plasmid DNA expressing anti-bcl-2 short hairpin RNA. All synthesized gene carriers showed excellent DNA binding, protective effects against nuclease, and gene transfection efficiency in MCF-7 and SKBr3 breast cancer cells. Preincubation with curcumin or 17α-estradiol resulted in a marked dose-dependent decrease in gene transfer efficiency and suggested targeting specificity of curcumin. Our study indicates the potential of curcumin and its derivatives as novel targeting ligands for tumor cells and tissues.


Introduction
Cancer continues to be one of the most lethal diseases worldwide, despite the increasing number of interventions available.Advances in cancer diagnosis and treatment have not improved overall survival rates for various reasons, including drug resist ance, metastasis, and side effects [1].Among current antitumor treatments, chemotherapy plays an essential role, but most chemical agents have low specificity for cancer cells, resulting in toxicity and side effects in the body [2].To overcome these obstacles, many studies have focused on tumorspecific targeting systems, which are expected to play an important role in the next generation of chemotherapy treatments [3,4].
Drug delivery systems are classified into passive delivery systems, which transport drugs by taking advantage of ana tomical differences between normal and tumor tissues, and active delivery systems, which deliver drugs intensively to spe cific tissues through targeting ligands [2,[5][6][7].Unlike treatment via passive drug delivery, nanoparticles with a targeting moiety can effectively accumulate in cancer tissues and cells by binding to receptors expressed in specific tumor tissues and are effi ciently introduced into cells through receptormediated endo cytosis.Therefore, cancer treatments using targeting moieties can be applied to various cancer cells using 1 effective drug via the selection of targeting ligands that can bind to receptors overexpressed in specific tumor tissues [8,9].
Gene therapies are being developed for various diseases, with the development of effective gene delivery systems a primary area of research.Recently, nonviral vectors, including cationic polymers, lipids, dendrimers, and peptides, have been proposed as safer alternatives to gene therapy and offer several advantages, such as mild immune responses, ability to produce the vectors in large quantities with high reproducibility, acceptable cost, stability to storage in nonexotic conditions, and ease of admin istration to patients [10][11][12].Nonviral gene delivery systems are typically composed of genes condensed into nanoparticles pre pared with cationic polymers such as polylysine, polyethylenei mine (PEI), polyamidoamine, chitosan, and dendrimers [13,14].PEI is a positively charged cationic polymer with various struc tures such as branched or linear polymer chains that undergoes functionalization for conjugation [15].Lysine dendrimers can strongly bind to genes via electrostatic interactions.They are expected to exhibit high gene transfection efficiency because of their rapid endosomal escape via the proton sponge mechanism in a broad range of cell types when compared with the efficiency of other cationic polymers.
Dextran is a biodegradable, biocompatible, and nonimmu nogenic polysaccharide that can be used as a potential module for delivery systems.It inhibits nonspecific interactions with serum proteins and prolongs the biological halflife of its con jugates, thereby increasing their stability in vivo.Dextran has been shown to improve cell viability and serum stability of PEI based systems, with promising results in gene delivery applica tions [16][17][18].
Curcumin demonstrates a wide range of pharmacological activities, including antiinflammatory, anticancer, antioxidant, woundhealing, and antimicrobial effects [19][20][21].It suppresses the proliferation of various tumor cells, including those from colorectal cancer, pancreatic cancer, breast cancer, prostate cancer, multiple myeloma, lung cancer, and oral cancer [19].Curcumin inhibits cancer promotion by modulating multiple molecular targets such as transcription factors, enzymes, cell cycle proteins, cell surface adhesion proteins, components of cell survival pathways, and cytokines [22].Despite its therapeu tic efficacy, curcumin has some disadvantages, including poor aqueous solubility, relatively low bioavailability, and intense staining.To overcome these limitations, various curcumin formulations, including curcumin in nanoparticles, liposomes, micelles, and phospholipid complexes, have been designed [23,24].Curcumin has only been studied for use in nanoparticles as a therapeutic drug, but its potential for other uses has not been explored.
This study aimed to assess the tumortargeting efficiency of 5 targeting moieties: phenylalanine, cinnamic acid, coumaric acid, ferulic acid, and curcumin.Here, we hypothesized that curcumin could best be used as a tumortargeting ligand if a specific membrane receptor that transports it was identified.We selected MCF7, a breast cancer cell line, because curcumin has shown excellent antitumor activity in breast cancer cells, along with diverse other cancer cell types.The targeting capaci ties of polyplexes complexed with PEIgrafted dextran, lysine dendrimerlinked targeting ligands, and short hairpin RNA (shRNA)expressing plasmid DNA (pDNA) were investigated using transfection efficiency, confocal microscopy, and com petition assays.The in vitro and in vivo antitumor effects of these formulations were evaluated using RNAsilencing gene delivery.Our findings will help confirm the potential of cur cumin and its derivatives as novel targeting ligands for tumor cells and tissues.

Peptide synthesis
Peptides were prepared using solidphase methods with Fmoc protected amino acids on a Liberty Microwave Peptide Synthesizer (CEM Co.).FmocLys(Fmoc)OH was used as the backbone material for dendrimer peptide synthesis.Fully protected K 8 den drimers (K 8 [K 2 K] 2 KC) (dK) were synthesized on 2chlorotrityl chloride resin (ChemPep, Inc., Wellington, FL, USA) with Oxyma/N,N′diisopropylcarbodiimide (coupling agent) and 20% piperidine (v/v, deprotection agent) in dimethylformamide (DMF) under 90 °C heating.For each coupling step, Fmoc protected amino acids and coupling agents were added in 5fold molar excess to the resin concentration.After cleavage from the resin in the presence of a cocktail solution (trifluoroacetic acid [TFA]/triisopropylsilane/diH 2 O (95: 2.5: 2.5, v/v/v) for 50 min at 40 °C, the fully protected peptide was precipitated and freezedried.

Preparation of the copolymer and targeting ligands
To prepare dextran copolymer conjugated with bPEI (DP), carboxyl group of FmocGlu (OtBu)OH (E) was activated with 1ethyl3(3dimethylaminopropyl) carbodiimide hydrochloride in a cosolvent of 50 mM 2(Nmorpholino)ethanesulfonic acid (MES) buffer (pH 4.95) and dimethyl sulfoxide (DMSO) at 25 °C for 30 min.The bPEI was added dropwise to activated E (molar ratio, 1:1) and incubated at 25 °C for 6 h (EP).After acetone pre cipitation, OtBu protecting groups on EP were cleaved with 95% TFA, and the EP was reprecipitated using ethyl ether.The product was dialyzed to remove unreacted Fmocamino acids and cleaved OtBu with DMSO for 1 d and further dialyzed in deionized water (DW) for 2 d, followed by freeze drying.Next, EP was conjugated chemically to dextran in a 1% molar ratio with DCC/4dimethylaminopyridine in DMSO at 25 °C for 12 h, resulting in DP.Then, DP was precipitated with acetone, and Fmoc groups were deprotected with 20% piperidine/DMF at 50 °C for 2 h.DP was dialyzed with DW for 48 h and freezedried (Fig. S1).

Pull-down assay and western blot analysis
MCF7 cells were grown at 37 °C in a 5% CO 2 incubator.Cytosolic and membrane proteins were isolated using a MemPER plus membrane protein extraction kit (Thermo Fisher Scientific Inc.) according to the manufacturer's protocol.Cells were collected by centrifugation at 300 × g for 5 min, and the cell pellets were washed twice with a cell wash and centrifuged again.The super natant was carefully discarded and permeabilization buffer with a protease and phosphate inhibitor cocktail (Thermo Fisher Scientific Inc.) added.Then, the cell pellet was incubated at 4 °C for 10 min with constant mixing, followed by centrifugation for 15 min at 16,000 × g.The supernatant containing cytosolic proteins was transferred to a new tube, and the residual cell pellet was resuspended in a solubilization buffer with a pro tease and phosphate inhibitor cocktail and incubated at 4 °C for 30 min with constant mixing.The tubes were then centrifuged at 16,000 × g for 15 min at 4 °C to collect the supernatant containing solubilized membrane proteins [25,26].These membrane pro teins were stored at −80 °C for future use.
For pulldown assay, membrane proteins were dialyzed with phosphatebuffered saline (PBS) (10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , 2.7 mM KCl, and 137 mM NaCl, pH 7.4) for 3 h at 4 °C [27].The curcumin Sepharose 4B resin was washed 3 times with cold PBS.The mixture was centrifuged at 2,000 × g for 5 min, the supernatant was discarded, and then 600 μl of PBS was added to the Sepharose resins to prepare a 50% slurry.Following this, 3 mg of membrane proteins was combined with the curcumin Sepharose resins and incubated overnight at 4 °C on a rotary shaker.These resins were washed thrice with a 10bed volume of cooled PBS and centrifuged at 2,000 × g for 5 min at 4 °C.

Molecular docking
The structures of ERα (7qvj and 8duk) and Erβ (4J26 and 5toa) were obtained from RCSB Protein Data Bank (RCSB PDB).The structure of curcumin (CC9) was obtained from National Center for Biotechnology Information structure.These struc tures were prepared for simulating docking in a Patch Dock server (https://bioinfo3d.cs.tau.ac.il/PatchDock/) and visual ized using PyMOL software.

Electrophoretic mobility shift assay and deoxyribonuclease stability
The pEGFPN1 amplified in Escherichia coli DH5α cells were purified using a pDNA midiprep kit (GeneJET Plasmid Midiprep Kit, Thermo Fisher Scientific Inc.) according to the manufac turer's protocol.The purity of pDNA was assessed by measuring the absorbance ratio at optical density at 260 nm (OD 260 )/OD 280 .DP polymer was mixed with dKPh (DPKPh), dKCi (DPKCi), dKCo (DPKCo), dKFe (DPKFe), or dKCu (DPKCu) at 9.5:0.5, 9:1, or 8.2 weight ratio in PBS solution, after which the polymers were complexed with pDNA at various weight ratios ranging from 0.01 to 64 in PBS and incubated for 30 min at 25 °C.DNA protec tion capacity of the polymers was assessed by adding deoxyribo nuclease (DNase) after polyplex formation.Briefly, 1.5 μl of RNaseFree DNase I (1 U/μl in buffer containing 100 mM tris, 25 mM MgCl 2 , and 5 mM CaCl 2 ) was incubated with polyplexes for 20 min at 37 °C, followed by inactivation of DNase by incuba tion at 75 °C for 10 min.DNA was released from polyplexes by treatment with heparin (60 units) for 10 min at 25 °C.All com plexes were electrophoresed on a 0.8% (w/v) agarose gel at 100 V, stained with SafeView classic (Applied Biological Materials Inc., Richmond, BC, Canada), and illuminated using a blue light emitting diode illuminator.

Size distribution and zeta potential analysis of polyplexes
The particle sizes and zeta potentials of DP, DPK, DPKPh, DPKCi, DPKCo, DPKFe, and DPKCu complexed with pDNA were investigated using an ELS8000 electrophoretic lightscattering spectrophotometer (Otsuka Electronics, Osaka, Japan) equipped with a HeNe laser operating at 25 °C and a fixed scattering angle of 90°.

Transfection efficiency of polyplexes in various cell lines
The transfection efficiencies of DPK, DPKPh, DPKCi, DPKCo, DPKFe, and DPKCu polyplexes with pEGFPN1 were investi gated in PC3, SKBr3, MDAMB231, MCF7, SKOV3, DLD1, and HaCaT cells.Cells were seeded at 2 × 10 4 cells/well in 24well plates and incubated for 24 h.The medium was aspirated, and polyplexes were added to wells at indicated weight/weight ratios in antibioticfree medium supplemented 10% FBS.TurboFect was used as a control transfection reagent.After incubation for 48 h, green fluorescent protein (GFP) expression in the cells was observed under a fluorescence microscope (IX71, Olympus, Japan).In addition, quantitative analysis of transfected cells was performed using an Attune NxT acoustic focusing cytometer (Thermo Fisher Scientific Co.).

Cellular uptake and localization of nanoparticles
Cellular uptake of the polyplexes was determined by confocal laser scanning microscopy.The DP polymers were labeled with NHSactivated Flamma 675 dye at a molar ratio of 1:1 in PBS.Flamma 675labeled DP/DP/DPK, DPKPh, DPKCi, DPKCo, DPKFe, and DPKCu polymers were complexed with pET28(a) at a 9:1 weight ratio and added to wells, where MCF7 cells were seeded at a density of 2×10 5 cells/ml in slide glass culture wells (SPL Life Science, Pocheonsi, Gyeonggido, Korea) and incu bated for 12 h.After incubation for 2 h, the glass slides were rinsed with PBS, stained with ActinGreen 488, and mounted with an antifade mountant with 4′,6diamidino2phenylindole (DAPI) according to the manufacturer's protocols.To investi gate cellular localization and endosomal escape of polyplexes in MCF7 cells, the polyplexes were prepared by the same method of cellular uptake.After incubating the polyplexes for 6 h, the glass slides were washed with PBS.Then, the cells were stained with LysoTracker and mounted using an antifade mountant with DAPI according to the manufacturer's proto cols.The intracellular polyplexes were observed using a Zeiss LSM510 confocal microscope (ZEISS, Jena, Germany) under identical conditions.The images were recorded digitally in a 512 × 512pixel format.

Cytotoxicity and antitumor activity
Hemolytic and cytotoxic effects of the polyplexes were evaluated using rat erythrocytes and HaCaT cells, respectively.Fresh blood obtained from a healthy rat tail was transferred to a heparin tube, and rat red blood cells (rRBCs) were collected by washing with PBS until the supernatant was clear.Polymers mixed with dKs at 9:1 (w/w) were complexed with psiRNAhBCL2 at 2:1 (w/w) and then added to 8% rRBCs (v/v) in PBS, followed by incubation with mild agitation for 1 h at 37 °C.The cells were centrifuged for 10 min at 800 × g, and the absorbance of the supernatants was measured at 414 nm.The controls for zero (blank) and 100% hemolysis were PBS alone and 0.3% (v/v) Triton X100, respec tively.Each measurement was conducted in triplicate.Cytotoxic effects and antitumor activities of the polyplexes were evaluated by CCK8 assay.A total of 5 × 10 3 cells/well were seeded in a 96well plate and incubated for 24 h.Subsequently, the polymers mixed with dKs at 9:1 (w/w) were complexed with psiRNA hBCL2 at 2:1 (w/w) in a minimal volume of Dulbecco's PBS.The polyplex solutions added to the culture medium (8× the complex volume) were placed in cell plates, after which plates were incu bated for 24 or 48 h at 37 °C.A 10μl CCK8 solution was added to each well, and the cells were further incubated for 4 h at 37 °C.Absorbances were measured at 430 nm (optical density) and 490 nm (with background subtracted) using a microplate reader (SpectraMax M5 Microplate Reader, Molecular Devices, Sunnyvale, CA, USA).Then, 100% cytotoxicity was determined following treatment with 0.3% (v/v) Triton X100.

Titration assay in gene transfection
Titration of polyplexes complexed with pEGFPN1 (1 μg) was performed in MCF7 and SKBR3 cells.Before polyplexes were added, the cells were treated with 10, 100, or 1,000 ng/ml at 2 × 10 4 cells/well.After incubation for 1 h, polyplexes were added to wells in an antibioticfree medium supplemented 10% FBS, followed by additional incubation for 36 h.GFP expres sion in cells was observed under a fluorescence microscope (IX71, Olympus, Japan).

Animal experiments
To xenograft with MCF7 cells, 5weekold female BALB/cnu mice (16 ± 1 g) obtained from Orient Bio Inc. (Seongnam, Gyeonggido, South Korea) were divided into 5 groups (n = 4, PBS, DPK, DPKCi, DPKCo, and DPKCu groups).Mice were housed separately with 4 animals in each cage under controlled temperature (22 ± 2 °C), humidity (55 ± 15%), and lighting (illuminance of 150 to 300 lx) conditions.MCF7 cells (2 × 10 7 cells/ml) were subcutaneously injected into the right flanks of the mice, which were anesthetized by inhalation of 5% (induc tion) and 2% (maintenance) isoflurane in pure oxygen.To com pare the in vivo biodistributions of Flamma 675labeled DPKCu/ psiRNAhBCL2 polyplexes, MDAMB231 and MCF7 cells were injected into the left and right flanks in the same mouse.The samples were injected intravenously into tail veins when tumor sizes reached approximately 20 mm 3  .After 2 h, tumor xenografted nude mice were imaged using a fluorescence imag ing system (FOBI, Cellgentek, Cheongju, Chungcheongbukdo, South Korea) in nearinfrared (NIR) light with filtration.Hearts, kidneys, livers, lungs, and spleens were collected from sacrificed mice after 24 h, and organ distributions were determined by imaging.Fluorescence intensity of each organ harvested was calculated and converted to rainbow images using NEOimage software [7].
To exam antitumor effects of polyplexes, the copolymers were complexed with 10 μg of psiRNAhBCL2 at 2:1 (w/w), and then the polyplexes were intravenously injected into tail veins when tumor diameters reached 4.4 to 5.2 mm 3. Tumor sizes were moni tored for 21 d, and the mice were euthanized by CO 2 inhalation.The excised tumor tissues were fixed in 4% paraformaldehyde, embedded in paraffin, stained by hematoxylin and eosin and BCL2 immunohistochemistry, followed by observation under a fluorescent microscope.

Statistical analysis
Data were analyzed using Excel software and presented as the mean ± SD.Comparisons among groups were performed using oneway analysis of variance Tukey's test of variance.**P < 0.05 was considered statistically significant.

Identification of curcumin-binding membrane protein in MCF-7 cells
The cell membrane fraction was isolated from cultured MCF7 cells, and membrane proteins were extracted.Subsequently, these proteins were applied to a Sepharose 4B resin conjugated with curcumin, and the binding proteins were eluted.The eluted pro teins were separated by 2D electrophoresis, and matrixassisted laser desorption/ionization timeofflight analysis was per formed on strongly expressed spots (Fig. 1A).Among receptors on the MCF7 cells, an ER was found to have the strongest affinity for curcumin (Fig. 1B).To clarify this result, fractions from the pulldown assay were subjected to SDSPAGE (Fig. 1C), and western blotting was performed using human ERα antibody (Fig. 1C).A strong band of ERα appeared in the total proteins, and it bound to the curcuminconjugated resin but did not bind to the curcuminfree resin (red arrow in Fig. 1D).To confirm whether curcumin can bind to ERs, we attempted molecular docking of curcumin in 2 ERα (Fig. 1E and F) and ERβ (Fig. 1G  and H) structures.Interestingly, unlike tamoxifen and peptides known to bind to ERs, which have existing binding sites inside the protein, curcumin was found to have binding pocket outside the protein.

Designation and characterization of synthesized copolymers
The bPEI2K was grafted onto dextran (DP) at 1% molar ratio to bind electrostatically with pDNA because uncharged dextran cannot be complexed with pDNA (Fig. 2A).To graft bPEI2K onto dextran, we first conjugated FmocGlu(OtBu)OH to bPEI2K through a chemical reaction (Egrafted bPEI), OtBu groups were removed, and the hydroxyl groups of dextran were conjugated with the carboxyl groups of Egrafted bPEI via esteri fication.Ester bonds are biocompatible linkages easily cleaved by cellular esterase.The DP copolymer was obtained by Fmoc deprotection (Fig. 2A).
PEGylated curcumin and its analogs (phenylalanine, cinnamic acid, coumaric acid, and ferulic acid) was conjugated to lysine dendrimer peptides with cysteine residue (K 8 [K 2 K] 2 KC, dk) through a maleimidethiol "click" reaction (Fig. 2B).Although PEGylated targeting ligands can be directly grafted onto DP, they were conjugated to dK peptides because these peptides can also bind with nucleic acids.Unexpected reactions due to multicon jugations can be minimized, and the percentage of targeting ligands employed is easily controlled by mixing DP and dKs.Furthermore, the present study aimed to investigate a favorable structure for specific targeting of tumor cells, including struc tures with phenylalanine, cinnamic acid, coumaric acid, ferulic acid, and curcumin.Successful synthesis was confirmed by 1 HNMR structural analysis of the products at each synthesis step (Figs.S1 to S5).
Figure 2C shows predicted mechanisms underlying effects of the targeted gene therapies developed in this study.Nanoparticles were formed by ternary complexing of DP, a targeted ligand grafted lysine dendrimer peptide, and pDNA expressing shRNA for the bcl2 gene, which targets the membrane ER (mER) in breast cancer cells.These particles were predicted to be internal ized by receptormediated endocytosis, followed by endosome formation and endosomal escape by the proton sponge effect.After the nanoparticles are dissociated and pDNA is released, pDNA is localized to the nucleus, and antibcl2 shRNA is expressed.Cytosolic small interfering RNA (siRNA) processing is achieved by Dicer, which silences bcl2 mRNA.

Nanoparticle formation of pDNA/copolymer complexes
To investigate the binding capacity of the copolymers to pDNA, we performed an electrophoretic mobility shift assay after complexing at various weight ratios.Dextran did not bind to pDNA even at a weight ratio of 64:1 because it has only hydroxyl groups (Fig. 3A, a), whereas bPEI2K bound strongly to DNA at a weight ratio of 0.2:1, because it has many amine groups (Fig. 3A, b).The bands of pDNA/DP polyplexes displayed complete condensation at a weight ratio of 0.2:1, although only 1% bPEI2K was grafted onto dextran (Fig. 3A, c).The DP and dK peptides with targeting ligands containing phenylala nine, cinnamic acid, coumaric acid, ferulic acid, and curcumin were mixed at a weight ratio of 9:1, followed by complexation with pDNA.dK, a lysine dendrimer peptide, and DPK, a mix ture of dK/DP (weight ratio of 1:9), were completely com plexed at weight ratios of 0.4:1 and 0.2:1, respectively (Fig. 3A,  d and e, respectively).DNA migration was completely retarded in the presence of DPKPh, DPKCo, DPKFe, and DPKCu car riers at a weight/weight ratio of 0.2:1 or 0.4:1 (Fig. 3A, f to j, respectively).
For gene delivery in vivo, an effective carrier must protect nucleic acids from degradation by nucleases in the serum and extracellular matrix.Therefore, the DNaseprotecting ability of the copolymers was investigated by adding DNase I, inactivat ing DNase I, and treatment with heparin.A negative control treated with only DNase I was completely digested, but pDNA complexed with DP, DPK, DPKPh, DPKCi, DPKCo, DPKFe, and DPKCu at a weight/weight ratio of 2:1 showed clear DNA bands, as did the positive control, indicating protective ability against nuclease digestion (Fig. 3B).

Transfection efficiency of polyplexes with pDNA
To compare the gene delivery capacity of copolymers, we com plexed pDNA with DPKs, which were formulated with a dK of 5%, 10%, or 20% (w/w), at 2 weight ratios of 1:1 and 2:1, in the 2 breast cancer cell lines MCF7 and SKBr3 (Fig. 4  fluorescence microscopy (Fig. 4A) and cytometry analysis (Fig. 4B) showed that enhanced GFP (EGFP) expression of DPKCi, DPKCo, and DPKCu polyplexes was higher than that of DPK, DPKPh, and DPKFe at a weight/weight ratio of 2:1.All polyplexes in the presence of targeting ligands exhibited potent transfection efficacy at a weight/weight ratio of 2:1 when compared to that of DPK.In addition, fluorescence intensities of the microtiter wells were measured to normalize gene transfection efficiency (Fig. S6).Polyplexes with 10% (w/w) targeting ligands showed the strongest EGFP expression, and those complexed with DPKCu emitted the highest green fluorescence.Although EGFP expression in SKBr3 cells was higher than that in MCF7 cells, transfection efficiency in SKBr3 cells was similar to that in MCF7 cells, except for the potent green fluorescence inten sity of DPK polyplexes in all tested formulations in SKBr3 cells (Fig. S7).

and Figs. S6 and S7). The pEGFPN1tranfected MCF7 cells observed by
In an investigation of gene transfection ability in various cancer cells, MDAMB231 (breast cancer), SKOV3 (ovarian cancer), PC3 (prostate cancer), and DLD1 (colon cancer) cells transfected with polyplexes in the presence of 10% (w/w) tar geting ligands showed no marked targeting capacity in any of the cell lines (Fig. S8). Figure S9 shows the gene transfection efficiency of the polyplexes in HaCaT cells, which are a normal cell line.

Cellular uptake and cellular distribution
To investigate the cellular uptake and distribution of pDNA/ copolymer complexes, we conjugated Flamma 675 fluorescence dye to the amine groups of DP, and unlabeled DP and labeled DP were mixed at a weight/weight ratio of 9:1 to minimize artificial effects of the conjugated fluorescent dye.The mixture was complexed with dKs containing targeting ligands (9:1, w/w) and pET28 DNA (a nonGFP expression vector), and the polyplexes were transfected into MCF7 cells.Sections of the cytosol and nucleus were stained with specific actin (ActinGreen 488 probe, green fluorescence) and nuclear (DAPI probe, blue fluorescence) antibodies.Figure 5A shows that strong red fluorescence was notable after transfection with DPKCi, DPKCo, and DPKCu polyplexes but that PEI2K polyplexes were not internalized.Intracellular fluorescence after DPK, DPKCo, and DPKFe trans fection was low.These results suggest that the targeting ligands DPKCi, DPKCo, and DPKCu have a specific binding affinity for ERs in the plasma membrane of MCF7 cells.
Furthermore, LysoTracker (green fluorescence) and DAPI were used to detect endosomes and nuclei, respectively.As shown in Fig. 5B, the red fluorescence of DPK, DPKPh, and DPKFe poly plexes exactly overlapped with the green fluorescence, indicating limited localization of polyplexes within endosomes, although low intracellular localization was observed in MCF7 cells compared to that of DPKCi, DPKCo, and DPKCu.In contrast, red fluores cence of the DPKCi, DPKCo, and DPKCu polyplexes showed that cellular was inconsistent, suggesting early endosomal escape in some cases.

Dose-dependent inhibition of gene transfection after preincubation of curcumin or estradiol
The gene transfection efficacy of polyplexes was analyzed after preincubation with curcumin or 17αestradiol, a ligand for ERs.As shown in Fig. 6, preincubation with curcumin inhibited gene transfection via DPKPh, DPKCi, DPKCo, DPKFe, and DPKCu in a dosedependent manner in both MCF7 (Fig. 6A) and SKBr3 (Fig. 6B) cells.However, the transfection efficiency of DPK was not altered by curcumin in either cancer cell line.Figures S10 and S11 show that the transfection of DPKPh, DPKCi, DPKCo, DPKFe, and DPKCu polyplexes were inhibited in the presence of 17αestradiol, an inhibitor of ER, in MCF7 cells.These results suggested that MCF7 and SKBr3 cells have one targetable receptor, at least in the plasma membrane.Therefore, this study focused on mERs as the target for curcumin.

Cytotoxic and antitumor effects of polyplexes
To evaluate safety of the polyplexes in normal cells, we performed hemolysis and cytotoxicity assays using rRBCs and HaCaT cells (human keratinocytes).As shown in Fig. 7A, the hemolysis rates of all copolymers except for PEI and DPK were less than 9% at 1,000 μg/ml, but that of DPK was 29.87% at 1,000 μg/ml.The pDNA/PEI2K complex induced a remarkable hemoglobin release from rRBC, even at the 10 μg/ml concentration, which was the experimental minimum (Fig. 7A).We assessed cytotoxicity of the polyplexes in HaCaT cells using the CCK8 assay.PEI poly plexes with 100 μg/ml pDNA showed 42.1% cell viability, indi cating highly cytotoxic effects (Fig. 7B).The percentage cell survival with pDNA/DPK polyplexes was 65.11% at 1,000 μg/ ml, whereas that with other copolymers and pDNA was more than 90% at 1,000 μg/ml.A psiRNAhBCL2 pDNA expressing shRNA for silencing the bcl-2 gene was used to demonstrate antitumor effects after tar geted gene delivery.These effects were evaluated by CCK8 assay in MCF7, MDAMB231, SKBr3, SKOV3, PC3, and DLD1 cancer cells.As shown in Fig. 7C, cell viabilities following delivery of psiRNAhBCL2 was remarkably reduced in MCF7 and SKBr3 cells.This result suggests that curcumin and its derivatives pos sess specific binding or targeting abilities for mER.To compare effects of targetingligands on antitumor effects, we performed dosedependent gene transfections in MCF7 (Fig. 7D) and SKBr3 (Fig. 7E) cells.The results showed that cell viability was dosedependently reduced in the presence of DKPCi and DPKCu, which is consistent with the results of transfection efficiencies.
To ascertain the induced apoptosis resulting from psiRNA hBCL2 delivery, we measured overexpression of caspase9 and caspase3 (Fig. 7F and G), which are molecules in the apoptotic cascade, by enzymelinked immunosorbent assay in MCF7 cells.Caspase9 was upregulated by transfection with the pDNA/DPKCi, pDNA/DPKCo, and pDNA/DPKCu polyplexes, and caspase3 was remarkably overexpressed following delivery of pDNA/DPKCi and pDNA/DPKCu.These results were con sistent with those of transfection efficiency.

In vivo targeted gene delivery and antitumor effects of DPKCu polyplexes
To prove the tumorhoming capacity of curcumin, we intrave nously injected Flamma 675labeled DPKCu/psiRNABCL2 (w/w, 2:1) polyplexes into MDAMB231 (left flank) and MCF7 (right flank) doublexenografted mice (Fig. 8A) and observed them using a NIR optical imaging system.After 2 h, a remarkable NIR signal in tumorbearing mice was observed on the MCF7xenografted flank but was not detected on the MDAMB231xenograted flank (Fig. 8B).This NIR signal weakened slightly over time but was maintained for 96 h (data not shown).Polyplexes were injected into the mouse tail vein; 24 h later, major organs (heart, kidney, liver,  and E) Polyplexes with psiRNA-hBCL2 (10, 100, 500, or 1,000 ng) (w/w, 2:1) were subjected to MCF-7 (D) or SKBr3 (E), and the cells were incubated for 48 h, followed by CCK-8 assay.(F and G) Upregulated caspase-9 (F) and caspase-3 (G) levels in the presence of polyplexes with psiRNA-hBCL2 in MCF-7 cells.After incubation of polyplexes with pDNA (3 or 6 μg) at weight/weight ratio of 2:1, intracellular caspase-9 and caspase-3 were quantitatively assayed using an enzyme-linked immunosorbent assay method.lung, and spleen) were removed, and biodistributions of the poly plexes were confirmed by observing NIR fluorescence.The NIR signal appeared in the liver but not in other organs (Fig. 8C).
To confirm antitumor effects in vivo, polyplexes were formed with 10 g of psiRNAhBCL2 using DPK, DPKci, DPKCo, or DPKCu (w/w, 9:1) at a weight ratio of 2:1 and injected into MCF 7xenografted mice.As shown in Fig. 8D, statistically significant tumor growth inhibition was observed in the groups treated with DPKCi and DPKCu polyplexes.In comparison to the PBS group, although cancer masses were inhibited by administration of DPK and DPKCo polyplexes, cancer formation was progressive.At 21 d after sample injection, tumor masses of the PBSadministered mice increased to 488 mm 2 , whereas those from mice adminis tered DPK polyplexes showed a mean mass size of 263 mm 2 .Mice treated with DPKCi, DPKCo, or DPKCu polyplexes had average tumor masses of 20, 130, or 14 mm 2 , respectively (Fig. 8E). Figure 8F shows histological analysis with hematoxylin and eosin and immunohisochemical analysis with antiBcl2 protein.The administration of DPKCi and DPKCu polyplexes remarkably reduced the cancer cells in the tumor tissue and inhibited Bcl2 protein expression, compared to the PBStreated group.

Discussion
Breast cancer is the second most common cancer in the world wide, with approximately 2.26 million incidences in 2020, and one of the main cancerrelated mortality in women [29].Several risk factors, such as family history, genetic mutations, aging, estro gen, sex, and lifestyle, increase the incidence of breast cancer.It is a metastatic cancer that can spread to organs such as the liver, lungs, bones, and brain after its onset [29].It is mainly classified into 3 main types by the presence of human epidermal growth factor 2, estrogen or progesterone receptor, and triplenegative breast cancer, lacking all 3 standard molecular markers [30].Early diagnosis by glycoproteins, blood angiogenic factors, hormone receptors, and other potential biomarkers increases survival rates.Approximately 50% to 60% of primary breast cancer lesions are diagnosed with ER overexpression (ERpositive).Two classes of ER exist: intracellular ERs (ERα and ERβ) regulate transcription driving growth, proliferation, and differentiation by estrogen sig naling; mERs (GPR30, ERX, and G q mER) are mostly G protein coupled receptors [31,32].Present study focused to verify the potential of curcumin as a specific targeting ligand in drug deliv ery via screening ER as a curcuminbinding protein among mem brane proteins of MCF7.
A variety of studies are being conducted to deliver antitumor drugs such as genes, chemical drugs, and peptides to tumor tissues and cells using nanoparticles.Although the size of the endothelial cell gap and the crossendothelial channels affect the exudation strength of nanoparticles [33][34][35], early studies have explored sys tems for delivering drugs to tumor tissue by the enhanced perme ability and retention (EPR) effect.However, this passive targeting does not have a marked therapeutic effect because it is difficult to obtain a sufficiently high drug concentration to manage tumor tis sue [36,37].The active targeting strategy of nanoparticles is mainly focused on the overexpression of many receptors such as epidermal growth factor, folate, transferrin, and integrin receptor present on the surface of vascular endothelial cells of cancer cells or tumors compared to normal cells [38,39].In general, active tumor targets are membrane proteins, which are overexpressed on tumor, angio genic endothelial, malignant, or inflammatory cells.There are approximately 7,000 known transmembrane proteins in cells, and ~150 of them are overexpressed in tumor cells or tumorassociated cells.They are potential drug or imaging diagnostic targets.Overexpression of special receptors can increase specific recogni tion between the nanoparticle and cancer cells by displaying various types of ligands, such as antibodies, aptamer, peptides, and polysac charides, on the nanoparticle surface, being increasingly used for active targeting.Therefore, finding various ligands that can specifi cally recognize tumor tissues or cells is very important for effective cancer treatment through active targeting.
Curcumin, a natural polyphenolic phytoalexin isolated from Curcuma longa, has a broad spectrum of biological effects [40].Recently, its anticancer effect was discovered after ingestion as a food or injection of its nanoformulation into the body [24].Elucidating the molecular mechanism by interfering multiple cell signaling pathways such as regulating cell cycle (cyclin D1 and cyclin E), apoptosis (caspase and downsignaling gene products), proliferation (human epidermal growth factor 2, epidermal growth factor receptor, and activating protein1), survival (phosphoinositide 3kinase/protein kinase B), invasion (matrix metallopeptidase9), angiogenesis (vascular endothelial growth factor), metastasis (CXCchemokine receptor4), and inflam mation (tumor necrosis factor, interleukin 1, interleukin8, interleukin12, nuclear factorkappa B, cyclooxygenase2, and 5lipoxygenase) holds potential of the development of novel cancer treatments [40].In addition, cinnamic acid inhibited human ovarian, colon, and breast cancer cell proliferation [41,42].Ferulic acid can inhibit cell proliferation and invasion in HeLa and Caski cells by induction of apoptosis and cell cycle arrest and have antiangiogenic action [43].p-Coumaric acid can inhibit the proliferation of a variety of tumor cells by causing cell cycle arrest and inducing apoptosis [44,45].
Although curcumin is a promising phytochemical with therapeutic benefits, studies elucidating its mechanism of action have been limited to intracellular targets such as reactive oxygen species, apoptosis signaling cascades, and DNA damage.We believe that curcumin must first enter cells to exert its diverse antitumor effects.If it randomly penetrates a cell membrane, it will not only exhibit greater anticancer activity but also result in serious cytotoxicity.Our research explores curcumin's poten tial as a tumortargeting ligand for gene delivery in breast can cer.We firstly identified ERs on membrane of MCF7 cells with a high affinity for curcumin via pulldown assay.This study sought to determine whether curcumin can target cancer cells via ERs and to evaluate the antitumor efficiency of a gene deliv ery system with curcumin in vitro and in vivo.Therefore, to eliminate the artificial effects of antitumor drugs or drug deliv ery carriers, we designed a gene delivery system using dextran as a vehicle to form nanoparticles.Although charged biopoly mers such as chitosan, hyaluronic acid, and alginate can be used, uncharged dextran, which is a biodegradable, biocompatible, and nonimmunogenic polysaccharide, was selected to mini mize nonspecific cellular delivery via electrostatic interactions.The bPEI2K was grafted onto dextran to form a polyplex by combining with genes via electrostatic attraction, which can improve endosomal escape via the proton sponge effect when the gene carrier is endocytosed.Target ligands can be directly conjugated to dextran, but it is difficult to conjugate each of the 5 ligands in the same ratio.Therefore, we conjugated PEGylated ligands to a cationic lysine dendrimer that can electrostatically bind to genes and completed polyplexes in which they were added in equal proportions.Additionally, lysine dendrimers can also induce the proton sponge effect.
As shown in Fig. 3, when bPEI2K was bound to the surface of pDNA, a large structure and other amine groups that do not participate in formation of this complex can prevent binding of another bPEI2K complex via cationic repulsion.Because most DP amine groups were bound to pDNA and dextran was exposed on the surface of the polyplex, no other interactions occurred.Therefore, the DP copolymer was considered a suitable cationic polymer for gene carriage in this study.In addition, curcumin and 4 analogs were conjugated to lysine dendrimer peptides and used as tumortargeting ligands of nanoparticles.The electro phoretic mobility shift assay results suggest that the targeting ligands do not interfere with pDNA binding.In addition, the protection capacity for DNase I showed that mixtures of copo lymers and dK derivatives strongly bind to DNA by electrostatic interactions and maintain the stable polyplexes in blood and other body fluids.Polyplex size is an important factor in the intercellular delivery of polyplexes and in vivo intravenous injec tions.The size of polyplexes ranged 80 to 140 nm suggested that all nanoparticles were of a suitable size for gene delivery.Surface charge is one of the key factors in endocytosis process of gene transfection as well as the cytotoxicity.Because the membrane of cancer cells is negatively charged, cationic surfacecharged nanoparticles may have a very high potential for cellular uptake.
The results of transfection efficiency in vitro are important for determining whether the targeted gene carriers developed in this study can be used in vivo.For example, if the DPKCi, DPKCo, and DPKCu complexes, which show high transfection efficiency in 2 breast cancer cell lines (MCF7 and SKBr3), are efficiently translocated into the cytosol of normal cells, they cannot be used as targeted gene carriers in tumor cells.However, their gene trans fection capacity is remarkably decreased in normal cells.These results suggest that cinnamic acid, coumaric acid, and curcumin specifically bind to or target the ERs of MCF7 and SkBr3 cells.Furthermore, these good transfection capacity via fast binding with tumor cells through targeting ligands and early endosomal escape via polycationicity.
A transfection inhibition assay using curcumin and 17α estradiol showed that curcumin displayed on nanoparticles specifically recognized ERs.Especially, 17αestradiol is an endog enous stereoisomer of the hormone 17βestradiol (E2), which targets ERs, but their affinities are different [46].17αestradiol is a shortacting estrogen that forms a complex with the estrogen ligand receptor in human breast cancer cells for a short period of time.As shown in Figs.S10 and S11, almost all gene delivery into MCF7 cells was inhibited by preincubation with 17αestradiol, with the exception of DPK polyplex delivery.SKBr3, a recognized ERnegative breast cancer cell line, can express both ERβ1 and 2 variants, which are also expressed in most of ERpositive cancer cells [47].A previous study proposed that G proteincoupled receptor 30 (GPR30) is related to estrogeninduced transactivation of epidermal growth factor receptor and adenylyl cyclase activity in SKBr3, which lacks nuclear ERs, and is a novel mER [31,32,46,47].Although the relationship between curcumin and G proteincou pled ERs is not clearly understood, direct evidence to define their roles is expected from future research.
As shown in Fig. 7A and B, all polyplexes did not show remark able toxicity or hemolysis in HaCaT cells and erythrocytes, resepctively.These results suggest that polyplexes without targeting ligands may be easily transferred into normal cells, and targeting ligands displayed on the surfaces of nanoparticles may prevent internalization across the cell membrane, which is consistent with results of the transfection in HaCaT cells.Cytotoxic and hemolytic evaluations suggested that DPK mixtures with targeting ligands are biocompatible and blood compatible.We used pDNA express ing bcl-2 shRNA as an antitumor gene.When pDNA was delivered into the cells, an shRNA, an artificial RNA fragment with a tight hairpin turn, was continuously expressed in the nucleus of trans fected cells.This shRNA had the advantages of relatively low deg radation and turnover compared with siRNA.Because the bcl-2 gene is an antiapoptotic gene that can suppress apoptosis of abnor mal cells, silencing of bcl-2 mRNA results in caspasedependent apoptosis and autophagic cell death.The bcl2 oncogene is over expressed in 50% to 70% of all human cancers, as well as breast cancers [48].Therefore, inhibition of bcl-2 mRNA can induce sensitivity to antitumor therapies, and delivery of this shRNA into cancer cells is a promising gene therapeutic approach.Caspase activity has been implicated in the induction of mitochondrial damage during apoptosis.Activation of caspase9 in the apopto some through dimerization leads to loss of mitochondrial mem brane potential, as well as cleavage of Bcl2, BclxL, and Mcl1.This activation can be regulated by downstream effector caspases, including caspase3, caspase6, and caspase7; procaspase3 can be activated by an increase in caspase9, followed by apoptosis [49].In Fig. 7F and G, gene transfection with DPKCi and DPKCu showed greater activation of caspase3 than caspase9, suggesting that cinnamic acid and curcumin, which are released from the dissociated polyplexes after the endosomal escape of internalized pDNA/DPKCi and pDNA/DPKCu polyplexes, may activate caspase3.This indicates synergistic induction of apoptosis following delivery of the bcl2 shRNA and target materials (cinnamic acid and curcumin).
In vivo targeting and antitumor effects are very important results in determining whether curcumin can actually be used clinically.It is a very interesting result that the NIR signal of nanoparticles injected to double tumorxenografted mice was strong only in the MCF7 region.These results indicate that curcumin displayed on the surface of the polyplexes allows spe cific homing to tumor tissues or cells in vivo.The absence of NIR signal in other organs except the liver suggests that most of the polyplexes accumulated in the tumor mass though some polyplexes remained in the liver as a "firstpass effect".Although more precise biotoxicity analyses need to be conducted in the future, we predict that in vivo toxicity will be minimal based on these results.Administration of polyplexes with pDNA signifi cantly inhibited the increase in tumor masses, and a decrease in Bcl2 protein was confirmed in immunestained tumor tissues.These findings suggest that curcumin and cinnamic acid hold potential as homing or targeting ligands and that silencing Bcl2 mRNA is very effective when applying gene therapies for cancer management.
In summary, this study examined whether curcumin can specifically target or home to cancerous tissues and cells in vivo, rather than having antitumor effects.In addition, the capacity of curcumin and its derivatives were compared as targeting ligands for tumor cells and tissues.Curcumin and cinnamic acid specifically target mERs in MCF7 and SKBr3 breast cancer cells.Further, these results suggest that DPKCi and DPKCu have great potential as gene delivery systems for targeted therapy of breast cancer cells.

Ethical Approval
All the animal experiments were performed in accordance with the protocols of the Institutional Animal Care and Use Committee of Sunchon National University, Republic of Korea (approval no.SCNU IACUC202219).

Fig. 1 .
Fig. 1. Specific binding of curcumin to ER. Pull-down assay of MCF-7 cell membrane extraction on curcumin-displayed Sepharose 4B resin.(A and B) The curcumin-binding protein was identified as ERα by matrix-assisted laser desorption/ionization time-of-flight analysis.(C and D) Curcumin-binding proteins were analyzed on SDS-PAGE (C) and western blotting (D).(E to H) Molecular docking of curcumin to ERβ (E and F) and ERα (G and H).The structures of ERβ (PDB: 4J26 [A]; 5Toa [B]) and ERα (PDB: 7qvj [C]; 8duk [D]) were docked with curcumin (PDB: CC9) in a Patch Dock server.

Fig. 2 .
Fig.2.Synthesis and scheme of the targeted gene delivery system proposed in this study.(A) Synthetic procedures of dextran with bPEI (DP) copolymer by chemical reactions.After conjugation between Fmoc-Glu(OtBu)-OH amino acids and bPEI (Mw 2 kDa) (EP), OtBu protecting groups were removed by 95% TFA.Free carboxyl groups of EP were grafted to hydroxyl groups of dextran via esterification.DP copolymers were achieved by Fmoc elimination.(B) Synthetic procedures of targeting ligand-grafted lysine dendrimer peptide.(A) After PEGylation of L -phenylanine (Ph), trans-cinnamic acid (Ci), p-coumaric acid (Co), or trans-ferulic acid (Fe), it was grafted to lysine dendrimer peptide (K 8 [K 2 K] 2 KC, dk) via a facile maleimide-thiol "click" reaction.For curcumin-conjugation, CDI-activated curcumin was easily conjugated with Maleimide-PEG-NH 2 , followed by maleimidethiol "click" reaction with dK peptide.(C) Nanoparticles were formed by ternary complexing of DP, targeted ligand-grafted lysine dendrimer peptide, and pDNA expressing shRNA for bcl-2 gene.Nanoparticles were specifically targeted to ER-positive tumor cells, followed by receptor-mediated endocytosis, endosomal escape, and dissociation of nanoparticles.Plasmid DNA containing anti-bcl-2 shRNA was localized into nucleus and expressed anti-bcl-2 shRNA, after which siRNA can silence bcl-2 mRNA in tumor cells.