Co-delivery of doxorubicin and conferone by novel pH-responsive β-cyclodextrin grafted micelles triggers apoptosis of metastatic human breast cancer cells

Adjuvant-aided combination chemotherapy is one of the most effective ways of cancer treatment by overcoming the multidrug resistance (MDR) and reducing the side-effects of anticancer drugs. In this study, Conferone (Conf) was used as an adjuvant in combination with Doxorubicin (Dox) for inducing apoptosis to MDA-MB-231 cells. Herein, the novel biodegradable amphiphilic β-cyclodextrin grafted poly maleate-co-PLGA was synthesized by thiol-ene addition and ring-opening process. Micelles obtained from the novel copolymer showed exceptional properties such as small size of around 34.5 nm, CMC of 0.1 μg/mL, and cell internalization of around 100% at 30 min. These novel engineered micelles were used for combination delivery of doxorubicin-conferone with high encapsulation efficiency of near 100% for both drugs. Our results show that combination delivery of Dox and Conf to MDA-MB-231 cells had synergistic effects (CI < 1). According to cell cycle and Annexin-V apoptosis analysis, Dox-Conf loaded micelle significantly induce tumor cell apoptosis (more than 98% of cells population showed apoptosis at IC50 = 0.259 μg/mL). RT-PCR and western-blot tests show that Dox-Conf loaded βCD-g-PMA-co-PLGA micelle induced apoptosis via intrinsic pathway. Therefore, the unique design of multi-functional pH-sensitive micelles open a new perspective for the development of nanomedicine for combination chemo-adjuvant therapy against malignant cancer.

Block-Copolymer Synthesis. Hydroxy terminated poly maleic anhydride synthesis. The synthesis method of poly maleic anhydride with hydroxy termination (PMA-OH), was reported in our previously published paper 6 . Briefly, after dissolving of 3.93 mg maleic anhydride (MA) in 60 mL toluene under refluxing and nitrogen purging, 3.5 mL 2-mercapto ethanol (ME) was poured into the solution by a syringe. After temperature reached 110 °C, 0.147 g of Azobis isobutyronitrile (AIBN) in dry toluene, was added to the flask via injection. Twenty hours was allowed for the completion of the reaction. The light-yellow product was then purified and precipitated by solvent/antisolvent system (respectively acetone/toluene). The prepared PMA-OH was then dried by freeze-dryer.
Preparation of beta cyclodextrin grafted PMA-OH. In order to activate beta-cyclodextrin (βCD), 0.98 g (equivalent to 0.00086 mol) of βCD was dissolved in 40 mL of dry dimethyl formamide (dry DMF) in a two-necked flask under a nitrogen atmosphere and stirring. After the complete dissolution of βCD, 0.17 g (equivalent to 0.007 mol) of NaH in the solid state, was added to the reaction solution. After completing of βCD activation at room temperature (24 h), the reaction flask was placed in an oil bath and the temperature was raised to 100 °C. Then, 0.58 g of PMA-OH (equivalent to approximately 0.003 mol MA) solution in dry DMF was added to the contents of the flask, dropwise, under the nitrogen purging and stirring. The reaction was continued for 24 h at 100 °C, under nitrogen atmosphere and stirring. After 24 h, the reaction mixture was poured into 150 mL of a mixture of acetone, acetic acid and water (100 mL acetone, 10 mL acetic acid and 50 mL distilled water) and stirred for 30 min to inactivate and wash the excess or unreacted NaH. Then, the product was precipitated again with pure acetone. The product of the second stage (βCD grafted hydroxy terminated poly maleate = βCD-g-PMA-OH), which was a creamy pale-yellow precipitate, was dried and stored.
Preparation of βCD-g-PMA-co-PLGA. βCD-g-PMA-OH (0.4 g, approximately equal to 0.0003 mol), lactide (1.5 g, 0.01 mol), and glycolide (0.5 g, 0.0043 mol) were poured into a two-necked flask. After complete melting of the material at 120 °C, under the nitrogen atmosphere and stirring, a certain amount of tin (II) octoate, Sn (Oct) 2 , (1-3% w/w of the total monomers) as the catalyst, was added to the contents of the flask. The mixture was stirred at 120 °C for 24 h. The prepared final copolymer (βCD-g-PMA-co-PLGA) was then purified by solvent/ antisolvent precipitation (Dichloromethane/Diethyl ether) and dried by freeze-dryer.
FTIR, 1 HNMR, 13 CNMR, CHNS and DSC analyses were used for investigating chemical structure and physicochemical properties of copolymer.
Degradation test of copolymer. For investigating in-vitro biodegradability of the copolymer, it was examined at two different pH environments. For each experiment, 5 mg of copolymer was dispensed in 2 mL of PBS at pH values of 7.4 and 5.5 and incubated at 37 °C in a shaker-incubator. For each of the pH values, and each specified time interval, two repetition were considered. In other words, after each specified time intervals (7, 11, 16, 21 and 30 day), four samples were centrifuged (12,000 rpm, 30 min) and the supernatants were separated from the copolymer precipitants. The supernatants pH, were measured, separately. After complete drying of copolymer precipitants, they were weighed and then analyzed by FTIR. The supernatants pH variation and weight loss percentage (WL %) of copolymer in different time intervals were calculated using Eq. (1) 21 .
where W i is the initial sample weight and W t is the sample weight at time t.
(1) www.nature.com/scientificreports/ Determination of critical micelle concentration (CMC). Spectro fluorometry method with pyrene probe was used to find critical micelle concentration (CMC) of the copolymer. One μL of pyrene solution (1 mg of pyrene in 10 mL of acetone) was added into dark flasks. After evaporation of acetone, the copolymer solution in dimethyl sulfoxide, DMSO, was poured into flasks. The final volume of flask was reached to 20 mL (5 mL copolymer solution and 15 mL deionized water) and copolymer final concentration was adjusted at 0.05, 0.1, 0.5, 1, 2.5, 5, 10, 25, 50, 100, 250, 500, 1000 μg/mL. The flasks were micellized by ultrasound probe and then incubated in a shaker incubator at 37 °C for 18 h, in order to balance pyrene partition between two phases. After cooling the samples to room temperature, the emission spectra of pyrene in each of the samples was studied by a spectrofluorometer. The excitation and emission wavelengths for pyrene spectra were 334 nm and 373 nm (I 1 ) and 393 nm (I 3 ), respectively.
Preparation of blank and drug-loaded micelles. Blank βCD-g-PMA-co-PLGA micelles were prepared by adding of copolymer solution (200 mg of copolymer in 6 mL of DMSO) dropwise into polyvinyl alcohol (PVA) solution (20 mL, 1% w/v), under sonication in an ice bath. Then the blank βCD-g-PMA-co-PLGA micelle solution was centrifuged (4500 rpm, 10 min) by Amicon centrifugal filter (MWCO: 50 KDa). In order to remove the residue of DMSO, the blank βCD-g-PMA-co-PLGA micelles suspension in 2 mL deionized water was transferred into a dialysis membrane (CelluSep H1, MWCO: 2000 Da) and purified by dialysis method against the deionized water as external phase for 24 h. The old external solution was removed several times and replaced with fresh deionized water. The purified blank micelles inside of the dialysis membrane was freeze-dried and kept at − 24 °C. In order to prepare Doxorubicin (Dox) loaded βCD-g-PMA-co-PLGA micelles, the copolymer solution (200 mg copolymer in 6 mL DMSO) was added dropwise to PVA solution (20 mL, 1% w/v) containing 20 mg Dox, and then was sonicated by ultrasound probe, while the pH of micelle solution was adjusted at 7.4 by sodium hydroxide (NaOH) solution. After centrifuging by Amicon centrifugal filter (4500 rpm, 10 min), the Dox loaded βCD-g-PMA-co-PLGA micelles were collected, dried and stored at − 24 °C. After centrifuging, the supernatant was utilized to determine Dox loading percentage.
For loading conferone (Conf) in βCD-g-PMA-co-PLGA micelles, first, 200 mg of copolymer and 20 mg of conferone were dissolved in 6 mL of DMSO. Then, the prepared solution was added to PVA solution similar to Dox loaded micelle preparation. After ultrasonication and centrifuging micelles, the Conf loaded βCD-g-PMAco-PLGA micelles were dried and stored at − 24 °C. The supernatant solution was used to quantify drug loading percentage.
The co-drug loaded βCD-g-PMA-co-PLGA micelles, was prepared by gradually adding of copolymer and Conf (200 mg and 10 mg, respectively) solution in DMSO (6 mL), to PVA solution containing Dox (10 mg of Dox/20 mL PVA), with sonication in dark and ice bath. Acidity of solution was adjusted at 7.4. The following steps were done similar to Dox loading process.
Drug loading and release amounts were determined by UV-Vis spectrophotometer, for which λ max of Dox and Conf was 480 and 324 nm, respectively. Then, the drug encapsulation efficiency (DEE %) was obtained using Eq. (2) 22 : Characterization of copolymeric micelles. The size, morphology and zeta potential of blank micelles were investigated by SEM and DLS-Zeta analyses. Moreover, FTIR spectra and zeta potential of the blank and co-drug loaded micelles were studied in order to confirm the drug loading into micelles.
In-Vitro study of drug release. First, 1 mg of dried single-and co-drug loaded βCD-g-PMA-co-PLGA micelles were weighted in microtubes and then were suspended in 2 mL of sink solution with two pH values, separately. Due to lower solubility of conferone in PBS buffer solutions, the release study was conducted in sink solution contained 0.5% DMSO, 0.5% Tween 20 and 99% PBS (two pH values of 5.5 and 7.4) to improve conferone solubility. Then, the microtubes (contain the samples) were placed in a shaker-incubator at 37 °C. After different time intervals (1, 2, 3, 7, 9, 11, 14 and 16 day), the microtubes containing samples were centrifuged (12,000 rpm, 25 min). After supernatant collection in each time interval, 2 mL of fresh sink solution was added to precipitant and sample was re-incubated in a shaker-incubator at 37 °C. The drugs amount in collected supernatant were detected by UV-Vis spectrophotometer and then release percentage of drugs were measured using Eq. (3) 23 . All stages of this test were duplicated for each pH value.
Study of nano-formulations intracellular uptake. In order to preparation of rhodamine-B-labeled blank βCD-g-PMA-co-PLGA micelles, first, 10 mg of copolymer and 0.1 mg of rhodamine-B (RB) was dissolved in 1 mL of DMSO. Then the prepared solution was added dropwise to 4 mL of PVA solution (1% w/v), under sonication in an ice bath and dark condition. After centrifuging (8000 rpm, 15 min) of micelle solution, the supernatant was removed. The precipitated rhodamine B-labeled blank βCD-g-PMA-co-PLGA micelles were washed by distilled water and centrifuged several times for complete removal of unloaded rhodamine-B. The precipitated rhodamine B-labeled blank βCD-g-PMA-co-PLGA micelles were dispersed in deionized water (1 mL) and were kept at − 24 °C. For preparing of co-drug loaded βCD-g-PMA-co-PLGA micelles labeled by rhodamine-B, 10 mg of copolymer, 0.5 mg of Conf and 0.1 mg of RB, were dissolved in 1 mL DMSO and added to PVA 1% w/v solution containing 0.5 mg of Dox, under sonication. Then polymer/conf solution containing RB was added dropwise to PVA/Dox solution under sonication in an ice bath and dark condition. Next steps were done like rhodamine B-labeled blank βCD-g-PMA-co-PLGA micelles preparation procedure.
Subsequently, the MDA-MB-231 cells were cultured in the 6-well palates (with population of 2 × 10 5 cell per well) in complete RPMI medium containing 10% FBS. After incubation at 37 °C with 5% CO 2 for 48 h, the cells were treated with rhodamine B-labeled blank and co-drug loaded βCD-g-PMA-co-PLGA micelles for 0.5, 1.5 and 3 h. The un-treated cells were chosen as the control group. Next, the cells were washed with PBS and trypsinized. After centrifuging (1500 rpm, 5 min) the cells were washed with PBS again. In order to quantify the fluorescent intensity of internalized rhodamine-B-labeled blank and co-drug-loaded βCD-g-PMA-co-PLGA micelles, the washed cells were dispersed in PBS (about 300 μL) and analyzed with FACS Calibur flow cytometer. For qualitative analysis, fluorescent imaging by a fluorescence microscope was also utilized. The images of treated cells (with rhodamine B-labeled co-drug-loaded βCD-g-PMA-co-PLGA micelles), were prepared similar to our previously published paper 6 .
Study of nano-formulations effect on cell cycle. The MDA-MB-231 cells (3 × 10 5 cell per well) were seeded in 6-well plates and incubated (48 h). Then, all formulations with IC 50 dosage, were applied for treatment of the cells. The un-treated cells were selected as the control group. Then, the plates were incubated for another 48 h. After moving the medium of each treated cells into separate tubes, the cells were washed with PBS, trypsinized and transferred back to corresponding tubes. As soon as centrifuging of tubes and removing of their supernatant were completed, the cells were dispersed in PBS (700 µL) and centrifuged again. Then, the supernatants were discarded and the cells were dispersed in 300 µL cold PBS. For fixing of cells, 700 µL of cold ethanol 70%, was poured to each of the tubes and mixed. The tubes were located at 4 °C, in dark condition for 3 days. Then, the samples were centrifuged and after removing of supernatants, the cells were dispersed in 300 µL of PBS. Next, after adding 10 µL of Ribonuclease-A (10 mg/mL) and 45 min incubating, 10 µL propidium iodide (1 mg/mL) was added to each of the samples and vortexed. After 10 min incubation at the room temperature and dark condition, the cells were examined by FACS Calibur flow cytometer for estimation of cell cycle phases.
Apoptosis study induced by nano-formulations. The effect of formulations on MDA-MB-231 cells were studied by Exbio apoptosis kit of Annexin V-FITC/PI. As soon as reaching 60% confluency, the cells were cultured (1 × 10 5 cell per well) in 6-well plates and were incubated (48 h). The nano-formulations (PB, B2D, BD and BC) with IC 50 doses were used for treatment of the cells. After 48 h, the medium of wells were transferred to separate tubes. Then, the cells were washed with PBS, collected, and transferred back into the corresponding tubes. The tubes were centrifuged and the cells were washed with PBS two times, after removing of supernatants. After washing by annexin binding buffer (BB), the cells were dispersed in of binding buffer (100 μL). Then, Annexin V-FITC (5 μL) and of propidium Iodide (PI, 5 μL), were added to cell dispersions and vortexed gently. After incubation at room temperature in dark condition (15 min), the samples were centrifuged and the supernatants were discarded. Finally, the cells were suspended in binding buffer (100 μL) and were evaluated by a FACS Calibur flow cytometer. The un-treated unstained cells were selected as the auto-fluorescence control group.
Real-time PCR analysis. The MDA-MB-231 cells were seeded and treated with all formulations, like the protocol of the previous section. After 48 h, the cells were washed with PBS twice and then trypsinized and centrifuged. The supernatants were removed and the cells were dispersed in PBS (250 μL). According to TRIzol method for RNA isolation, RiboEx (750 μL) and then chloroform (200 μL) were poured into samples to lysis of cells and extract RNA. Following a short incubation (2 min, at room temperature), the samples were centrifuged (12,000 g, 20 min, 4 °C) and the upper aqueous layer (RNA phase), were separated. Then, isopropanol (500 μL) was added to separate RNA solution and the samples were centrifuged ( Finally, in order to perform quantitative PCR (qPCR) and investigate apoptotic pathway of treated cells, the samples were prepared as a mixture of SYBR Green Master Mix (5 μL, 2x), cDNA (2 μL), primer pair mix (5 pmol/μL) and deionized water (3 μL). This mixture was prepared for each of the formulations separately. In the PCR program, initial denaturation of samples was done for 15 min at 95 °C. Then, the run was proceeded at 95 °C for 15 s, which was repeated for 45 cycles. The annealing/extension stage was completed for 50 s at 60 °C. The sequences of used primers are presented in Table S1. The GAPDH was considered as the references gene. Lastly, the fold changes of genes expression were calculated by − ∆∆C t method.

Study of protein expression by western blot method.
Like the previous section, the cells were cultured in 6-well plates and treated by co-drug loaded βCD-g-PMA-co-PLGA micelles (B2D, with IC 50 dosage) and un-treated cells were considered as the control group. After 48 h incubation, the cells were washed with PBS and harvested. Then, radioimmunoprecipitation assay buffer (RIPA buffer) at 4 °C, was used for cell lysing. The RIPA buffer was composed of protease inhibitor cocktail (1 tablet), Tris-HCL (pH = 8, 500 µL), NaCl (0.08 g), EDTA (0.003 g), Sodium deoxycholate (0.025 g), Triton NP40 (10 µl, 1%) and SDS (0.01 g). Subsequently, the samples were centrifuged (12,000 rpm, 10 min, 4 °C) and the protein content of supernatant, was determined by a spectrophotometer, according to protocols of Bradford assay (Bio-Rad Laboratories, USA). The target fragments of proteins that were separated from the SDS-PAGE gel electrophoresis, were moved to the PVDF membrane (polyvinylidene difluoride membrane) and were blocked with 5% w/v of skim milk and 0.1% v/v of Tween 20 in tris buffered saline (TBS) for masking of unspecific bands. Specific primary antibodies were added to the blocked PVDF membranes that contained the target proteins and were incubated (overnight at 4 °C). After washing with TBS-T, the membranes were incubated with secondary antibodies, for 1 h at room temperature. The bands related to the target proteins were visualized using enhanced chemiluminescence detection kit (Thermo Fisher Scientific, the Netherlands) and were measured with Amersham Imager. Lastly, after normalizing of the outcomes of western blot using GAPDH expression as the control, the blots were calculated using Image J software, version 1.52n. The used primary and secondary antibodies were presented in our previously published paper 6 .
Statistical analyses. The duplicate or triplicated outcomes of analyses, were presented as ± standard deviation (± SD) using Graph pad prism software, version-8 or Microsoft Excel (2019). The student's t-test and ANOVA were used as statistical analyses for two-way and multiples comparisons, respectively. The statistically significant results had the P value lesser or equal to 0.05.

Results and discussion
The general procedure is briefly presented as a schematic illustration in Fig. 1. Fig. 2, the synthesis of pH-responsive βCD-g-poly maleate-co-PLGA was done in three stages. In the first stage, 2-mercaptoethanol (ME) and Maleic anhydride (MA) were polymerized with radical thiol-ene addition in the presence of AIBN as the initiator. The product of this stage is the hydroxy terminated poly maleic anhydride (PMA-OH). As a result of forming of ME radicals with initiator, the C=C band of MA was reacted radically with • S end of thiol radical and then polymerized without any ring opening. The end -OH group of PMA was required for the last stage of synthesis.

Designing and synthesizing of copolymer. As demonstrated in
In the second stage, the rings of PMA were opened with activated βCD (which was transferred to epoxy form by NaH) and then were esterificated. Esterification may be accomplished by one or more locations of hydroxyl www.nature.com/scientificreports/ groups of βCD and anhydride rings of PMA. Therefore, the βCD-g-PMA-OH was formed with one or more branches of PMA per one molecule of βCD. This βCD grafted polymer has a carboxylic acid in every unit of polymer that was required for pH-sensitivity of delivery system, formation of hydrophilic section of copolymer as the shell of micelles, and enhancing of water solubility of copolymer. About 1 g of product was obtained from this step (efficiency 64.1%).
In the final stage, the -OH end group of βCD-g-PMA-OH, with the catalyzing effect of Sn(Oct) 2 , caused ring openings of lactide and glycolide and their esterification to PLGA form (as the tail of copolymer and core of micelles). About 1.4 g of βCD-g-PMA-co-PLGA was obtained from this step (efficiency 58.3%).
Characterization of copolymer. FTIR results. The FTIR spectra of βCD and all stages of synthesis are presented in Fig. S1 in the supplementary file and is enlarged for better visualization of details in Fig. S2.
The detailed explanation of FTIR spectra of PMA-OH, βCD-g-PMA-OH and βCD-g-PMA-co-PLGA were presented in supplementary file.
NMR results. Results and detailed discussion of 1 HNMR and 13 CNMR spectra of PMA-OH, were presented in our previously published paper in detail 6 . The 1 HNMR and 13 CNMR spectra of βCD-g-PMA-OH, are shown in Fig. S3-A and S3-B, and the 1 HNMR and 13 CNMR spectra of βCD-g-PMA-co-PLGA, are presented in Fig. 3-A  and 3-B, respectively. The detailed explanation of 1 HNMR and 13 CNMR spectra of βCD-g-PMA-OH and βCD-g-PMA-co-PLGA were presented in supplementary file. www.nature.com/scientificreports/ The molecular weight of βCD-g-PMA-co-PLGA could not be investigated with gel permeation chromatography (GPC), due to the insufficient solubility of the copolymer in DMF solvent. Furthermore, because of the vigorous interaction of hydroxyls of the βCD with the GPC column, other researchers have also reported problems with calculating the molecular weight of their polymer by GPC 13,24 . Therefore, Eqs. (4) and (5) were used for determining molar mass (M n ) of βCD-g-PMA-co-PLGA, with the aid of integrating of the peaks in 1 HNMR spectrum 25 .
In the Eq. (4), "m" is the number of used signs of copolymer, and "p i " and "I i " are the number and integration of protons that pertained to ith peak of copolymer. The calculation of molar mass of copolymer are presented in supplementary file. The results of related calculations are reported in Table 1.

Results of CHNS elemental analysis.
For elemental analysis of copolymer by CHNS analyzer, 4.568 mg of copolymer was used. Results of test are presented in Table S2. According to data, the presence of sulfur (4.44% w), confirmed the existence of -S-linkage in copolymer structure. In the same way, the negligible amount of "N" (0.85% w) is probably related to impurities such as solvents residues (such as DMF). The CHNS time versus voltage plot are shown in Fig. S4.
Results of DSC study. Figure S5, presents the DSC or temperature against heat flow plot of copolymer. The T g or glass transition temperature of copolymer was determined with an endothermic peak in DSC plot, about 38.69 °C. Absence of T m or melting point of copolymer in the plot was a sign of amorphous structure without any crystallinity. This result is confirmed with DSC results of other studies related to PLGA-based polymers and copolymers. According to these reports, T m was not observed because of amorphous structure of PLGA, and also T g was reported to be between 35-65 °C, related to LA: GL ratio in PLGA structure (50-10% of LA in PLGA) that showed the T g was decreased with increasing of GL content of PLGA 26 .
Results of in-vitro degradation test of copolymer. Biodegradable copolymers encounter gradual degradation in contact of aqueous solution. In most polyesters, such as PLGA-based copolymers, hydrolysis of esteric-band and cleavage of copolymer is the main reason of degradation 27 . The produced soluble cleaved-copolymers and monomers such as lactic and glycolic acid are produced, that decreases pH of the solution. Therefore, timewise investigation of structure and weight of residual copolymer and pH of solution (that was in contact with copolymer during degradation test) are the suitable ways to determine degradation time and process. Results of degradation test are shown in Figs. S6-A, S6-B, S7 and S8. Fragmentation of the copolymer by degradation cause a gradual decrease in copolymer weight. Diagram of weight loss (WL %) of copolymer versus time is shown in Fig. S6-A. As shown in Fig. S6-A, copolymer initial weight decreased with time gradually. However, percentage of weight loss at pH = 7.4 was more than at pH = 5.5, that is probably due to the more hydrolysis of ester and carboxylic acid groups and subsequent more dissolution in PBS (with pH = 7.4). After 30 days, the WL % was reached to 19 and 20%, at pH = 7.4 and 5.5, respectively. As could be seen at day 30, the WL percentage at pH 5.5 excelled over the pH = 7.4, due to the initiation of major degradation process of copolymer (due to cleavage) rather than slight degradation (due to dissolution).
Variation of the pH of degradation medium are plotted versus time in Fig. S6-B. Hydrolysis of carboxylic acid groups related to maleate block, caused the initial sharp decrease in pH. After 7 days, pH-decrease slowed down, as a result of slight copolymer degradation. Finally, after 30 days, pH value reached to 6.2 and 3.8 for initial pH = 7.4 and 5.5, respectively. These pH values are similar to other reports about PLGA or PLGA-based copolymers that were about pH ≈ 5.48-7.4 with initial neutral pH 28,29 . However, compared to PLGA-based copolymers degradation results of our previously published article (pH 3.1 in 16 days) 6 , the upper pH value in the similar time interval is because of the lesser carboxylic acid groups in new copolymer maleate block.
FTIR analysis was used for investigation of the variation in structure of copolymer during the degradation process (refer to Figs. S7 and S8 and detailed explanation in the supplementary file). The degradation results showed that until 30 days, the degradation of copolymer was not evident, but after that the copolymer started the main process of degradation, due to the higher LA / GL ratio in PLGA section. This result is in agreement with other reports about PLGA based copolymers 27,30 .  Table 1. M n value for copolymer calculated using 1 HNMR, with theoretical and spectrum-based calculated (by 1 HNMR) molar ratio of βCD-g-PMA-co-PLGA copolymer sections.  www.nature.com/scientificreports/ CMC results, characterization, encapsulation and loading efficiency of micelles. Critical micelle concentration (CMC) of copolymer was determined using a plot of concentration of pyrene loaded micellar solution versus the ratio of I 1 /I 3 (Fig. 4A). With increasing of micelle formation, the pyrene loading in core of micelles increased and as a result the pyrene intensity decreased. After formation of micelles, the final value of a sharp decrease in the ratio of intensities is considered as CMC. As could be seen in Fig. 4A, the plot is "µ-shaped" with two minimum points that are selected as CMC 1 and CMC 2 for copolymer micellar solution. According to Fig. 4A, the first CMC point is located at 0.1 µg/mL and the second CMC point is observed at 2.5 µg/mL. Such type of CMC diagrams (µ-shaped) appears in copolymers micellization and is related to selfassembly process and polydispersity of copolymers (due to variation in chain length of polymeric blocks) 31 . It is also mentioned that with increasing of copolymer concentration, cylindrical-shaped micelles are formed as a result of aggregation of spherical-shaped micelles, which causes the second CMC 32 . The low value of CMC is an important and favorite property for dynamic stability of micelles, particularly at very low concentration in physiological environments such as blood circulation 33 39 . Therefore, our newly developed micelles (βCD-g-PMA-co-PLGA) showed a CMC value of 9-520-fold smaller compared to CMC values for βCD-based micelles in the previously published reports. Doxorubicin (Dox), Conferone (Conf) and combination of them (2D) were loaded to micelles as single and co-drug loaded βCD-g-PMA-co-PLGA micelles (BD, BC and B2D, respectively), with copolymer/drug ratio of 10:1. For confirming of Dox and Conf loading into micelles, FTIR spectrum of co-drug loaded βCD-g-PMAco-PLGA micelles (B2D) was evaluated. According to Fig. S1-B2D, the presence of strong and broad peak at 1400-1500 cm −1 and 1650 cm −1 shows the stretching of C=C of Dox and Conf aromatic and alkene rings, respectively. Presence of =C-H in Dox and Conf proved by appearance of peak at 3100 cm −1 . Peaks are observed at: 700 cm −1 (out of plane bending of C-H of aromatic ring) and at 1423 cm −1 (stretching of C-C band of aromatic ring) was indicators for presence of Dox-Conf in nano-formulation.
Drug loading results of nano-formulations are presented in Table 2 as drug encapsulation efficiency (DEE %). The high values of DEE % (up to 98%) in Table 2, show that the copolymeric micelles have very great loading efficiency, due to presence of various drug trapping positions (binding electrostatically to -COO-groups of PMA section, βCD cavity and core of micelle). Our obtained DEE % shows a very higher efficiency compared to similar studies on βCD-based star micelles with a range of 21.44-86.4% [34][35][36]40 .
The blank and co-drug loaded βCD-g-PMA-co-PLGA micelles were analyzed with DLS-zeta test and results are presented in Figs. S9-a, S9-b and S10. According to Fig. S9, zeta-potential of blank and co-drug loaded βCDg-PMA-co-PLGA micelles are equal to − 19.7 and − 2.39 mV, respectively. This difference between zeta-potential of blank and co-drug loaded βCD-g-PMA-co-PLGA micelles is due to the electrostatic interactions between carboxylic acid groups of micelle surfaces (pK a = 6.6) and Dox amine groups (pK a = 8.3) at pH = 7.4. Decreasing of zeta-potential (from − 19.7 to − 2.39 mV) after drug loading, confirms loading of Dox on surfaces of micelles. However, Dox could be loaded into core of micelles, too. In the case of Conf, due to the high hydrophobicity, loading happens into the core of micelles completely. According to the published researches, the optimum range of zeta-potential for electrostatically stability and extended circulation time for nano-particles in blood, is ± 20 mV 41 . Therefore, the obtained zeta-potential values for blank and co-drug loaded βCD-g-PMA-co-PLGA micelles are located in the suitable range that complies with other related reports 42 .
Based on DLS results in Fig. S10, the blank βCD-g-PMA-co-PLGA micelles had an average hydrodynamic diameter of about 96.51 nm (with polydispersity index, PDI = 1). The obtained PDI value shows lower homogeneity of nano micelles 43,44 , that may be due to variation in amount of grafted-βCD, length of PLGA or PMA chains in copolymer. The size and morphology of blank βCD-g-PMA-co-PLGA micelles are analyzed with SEM and the prepared image is shown in Fig. S11. According to SEM results, the blank micelles had an average diameter of about 34.5 nm and a spherical-like shape. The DLS reported size is higher compared to what SEM reported, that is probably due to the swelling of micelles by water in DLS test versus the dry condition in SEM analysis 45 .
Altogether, based on the obtained desirable diameter and zeta-potential of the micelles, it can be claimed that the prepared micelles are capable of penetrating into cancer tissues and cells through passive targeting. On top of that, the prepared micelles had a smaller size in comparison with other βCD-based micelles in the published works up to now which improves the efficiency of diffusion into cells 34,35,46 . Investigation of in-vitro release test. Dox and/or Conf release from single-and co-drug loaded βCDg-PMA-co-PLGA micelles, are shown in Fig. 4B. As can be seen, co-drug loaded and Dox loaded βCD-g-PMAco-PLGA micelles showed pH-responsive release with more dominant release at pH = 5.5 compared to pH = 7.4. But conferone release from conferone loaded βCD-g-PMA-co-PLGA micelles did not follow the pH-responsive pattern. The reason for lower Dox release from Dox-loaded nano-formulations in physiological pH (pH = 7.4) was the presence of electrostatic interaction between protonated amine group of Dox (with positive charge) (pH < pKa = 8.3) and carboxylate groups of copolymer (with negative charge) (pH > pKa micelles = 6.6). While, in acidic pH (pH = 5.4), the carboxylate groups of copolymer are protonated (pH < pKa = 6.6) and transferred to -COOH group without any charge. Therefore, the electrostatic interaction between Dox and carboxylate part of micelle was removed that led to higher amount of Dox release. More importantly, Dox release from co-drug-  www.nature.com/scientificreports/ and Dox loaded βCD-g-PMA-co-PLGA micelles has two steps (from day-1 to day-7, and from day-7 to day-16), that is probably due to the presence of more than one loading mechanisms. Since Dox could be loaded either in core of micelle or interact with carboxylate groups on the surface of micelles and finally interact with βCD as inclusion-complex, the release profile could be different depending on the loading mechanism.
In the case of Conf, it could just be loaded in the core of micelles or trapped in βCD cavity, due to the high hydrophobicity. Therefore, its release depends on the micelle's deformation (with no pH-sensitivity) which increases with copolymer degradation or micelles swelling. The dual drug release in our study is clearly sustained compared with Dox release from βCD based micelles published previously 34,35,47,48 . For example, in a study by Xu et al., βCD-PLA-POEGMA/Dox micelles had shown a sustained Dox release of about 20% and 50%, at pH = 7.4 and 5.0, respectively after 24 h 40 . In the same time, in our study, Dox and Conf release from drug loaded βCDg-PMA-co-PLGA micelles was below 10% and 30%, at pH = 7.4 and 5.0, respectively. Sustained release of Dox in our work, may be due to the stability and rigidity of micelle structure and dominant loading of drugs in the core of micelles that cause a resistance against dilution and drug release.

Cell internalization ability of micelles. Internalization of rhodamine B-labeled blank βCD-g-PMA-co-
PLGA micelles (PB) and rhodamine B-labeled co-drug loaded βCD-g-PMA-co-PLGA micelles (B2D) into MDA-MB-231 cell line were investigated with flowcytometry and fluorescent microscope and the obtained results are presented in Figs. 5A-C. As shown in Fig. 5A, 100% of the cells have taken the rhodamine B-labelled blank βCDg-PMA-co-PLGA micelles and rhodamine B-labeled co-drug loaded βCD-g-PMA-co-PLGA micelles. According to Fig. 5B, the mean fluorescence intensity (%) was increased with time (0.5, 1.5 and 3 h). Similarly, rhodamine B-labeled co-drug loaded βCD-g-PMA-co-PLGA micelles showed higher mean fluorescence intensity compared to rhodamine B-labeled blank βCD-g-PMA-co-PLGA micelles in all mentioned time intervals (p value < 0.001). The higher cellular uptake of rhodamine B-labeled co-drug loaded co-drug loaded βCD-g-PMA-co-PLGA micelles compares to rhodamine B-labeled blank βCD-g-PMA-co-PLGA micelles is due to the decrease in the negative surface charge of co-drug loaded βCD-g-PMA-co-PLGA (− 2.39 mV) micelles compared to blank βCD-g-PMAco-PLGA (− 19.7 mV) micelles as determined by the zeta potential, because the lower negative charge has less electrostatic repulsion forces with the negative cell membrane, and therefore higher uptake into the cells 49 .
Based on the rapid and great uptake percentage of our novel developed βCD-g-PMA-co-PLGA micelles (100% at 0.5 h), we can claim that the prepared βCD-g-PMA-co-PLGA micelles had favorable structure, charge, and size for cell internalization. Superiority of this formulation is clear when comparing its cell internalization with reports of other researchers. For example, our previous work showed a lower internalization of functionalized PLGA-based blank micelles into MDA-MB-231 cells (33%, 60 and 81%, at 0.5, 1.5 and 3 h) which is very slower 6 . This phenomenon is because of lesser negative charge of blank βCD-g-PMA-co-PLGA micelles in the present study (− 19.7 mV) compared to the blank micelle charge in previous study (− 29.7 mV). This leads to inferior electrostatic repulsion forces between negative charges of blank βCD-g-PMA-co-PLGA micelles and cell membrane and consequently higher internalization into cells 49 Figure 4C shows that the single-and co-drug loaded βCD-g-PMA-co-PLGA micelles caused a higher level of cytotoxicity in comparison with the corresponding free drugs (Dox, Conf and Dox-Conf). This difference between result of nano-formulation and free drugs, was due to the higher intracellular uptake that overcome drug resistance as well as increasing of Conf solubility in micelle forms. GraphPad prism software (V. 8.0.1) was used to calculate IC 50 dosages of all formulations; and the results are presented in Fig. 4E and Table S3. The results of cell viability assay showed that the cells treated with dual drug loaded (Dox-Conf loaded βCD-g-PMA-co-PLGA) micelles caused in significantly lower viability than those treated with either single drug loaded micelles, indicating that combination of Dox and Conf demonstrated superior anticancer activity. This result suggesting that efficient delivery of Dox and Conf by βCD-g-PMA-co-PLGA micelles Table 2. Results of drug encapsulation efficiency (DEE %) for nano-formulations (Abbreviations: B2D: co-drug loaded βCD-g-PMA-co-PLGA micelles, BD: Dox loaded βCD-g-PMA-co-PLGA micelles, BC: Conf loaded βCD-g-PMA-co-PLGA micelles). www.nature.com/scientificreports/ contributes substantially to enhance combinational antitumor Effects (Fig. 4C,E). According to IC 50 results, the lowest IC 50 (0.259 μg/mL) belongs to co-drug loaded βCD-g-PMA-co-PLGA micelles. The effective dosage of Dox in co-drug loaded βCD-g-PMA-co-PLGA micelles (0.1295 μg/mL) was lower compared to free Dox and Dox loaded βCD-g-PMA-co-PLGA micelles (refer to Table S3) because conferone in combination with Dox, can overcome the P-gp-mediated drug resistance and lead to Dox accumulation in cells 52 . This caused a decrease in the required Dox therapeutic dosage and therefore a decrease in its side effects. Based on our literature review our novel developed co-drug loaded micelle showed superior anticancer efficacy compared to previously pub- The CompuSyn software (V. 1) was used for calculation of combination index (CI), and results are shown in Fig. S12 and Table S4. The combination of free Dox-Conf and Dox-Conf in co-drug loaded βCD-g-PMA-co-PLGA micelles showed synergistic effects in IC 50 dosage (CI < 1). The CI value of co-drug loaded βCD-g-PMA-co-PLGA micelles (0.5) is lower than free Dox-Conf (0.8), that shows a more synergistic effect of nano-formulated combination form (co-drug loaded βCD-g-PMA-co-PLGA micelles). As can be seen in Fig. 4C, among the nanoformulations, the co-drug loaded βCD-g-PMA-co-PLGA micelles showed higher cytotoxicity in comparison with single-drug loaded βCD-g-PMA-co-PLGA micelles that could be explained by lower IC 50 dose and synergistic effect. Drug efflux due to increasing in P-glycoprotein (P-gp) expression, is an important problem in progressive cancers which causes a decrease in drug accumulation in cells and hence a decrease in drug efficiency. Conferone in combination with Dox, can overcome the P-gp-mediated drug resistance and lead to Dox accumulation in cells 52 . As a result, co-drug loaded βCD-g-PMA-co-PLGA micelles acted as the most efficient nano-formulation www.nature.com/scientificreports/ because of higher accumulation level of Dox, increasing of Conf solubility, and synergistic effect. The statistical analysis showed that the results of comparison among groups was significant.

Evaluation of cell cycle arrest induced by drug loaded micelles. The cell cycle analysis investigates
the various stages of cell cycle and DNA duplication, containing: G1, S, G2 and M 53 . The obtained results are presented in Fig. 6 and Table S5. According to Fig. 6 and Table S5, the blank βCD-g-PMA-co-PLGA micelles did not show noticeable changes in cell cycle pattern in comparison with the control group, which shows almost no toxicity to MDA-MB-231 cells. The co-drug loaded βCD-g-PMA-co-PLGA micelles (G2/M: 86.4%), free Dox-Conf combination (G2/M: 95%), and free Dox caused (G2/M: 87%), G2/M arrests in treated cells (0.259 μg/mL) while, Dox loaded βCDg-PMA-co-PLGA micelles showed S (19.5%) and G2/M (67%) arrest compared to control group (S: 10.80%, G2/M = 15.30%) ( Table S5). The presence of Conf in formulations lead to S arrest in Conf loaded βCD-g-PMAco-PLGA micelles (S: 51.3%) and free Conf (S: 15.8%) (Table S5) Evaluation of apoptosis induction. Annexin-V is a fluorescent agent that stained the apoptotic cells and propidium iodide (PI) was used for staining nucleus of late apoptotic and necrotic cells 58 . To demonstrate that Dox-Conf loaded βCD-g-PMA-co-PLGA micelles produces greater level of cancer cell apoptosis compared to single drug loaded formulations and free drugs, MDA-MB-231 cells after treatment were analyzed by Annexin-V/PI double staining flowcytometry. The effect of free drugs (Dox, Conf and Dox-Conf) on apoptosis of MDA-MB-231 cells were presented in our previously published paper 6 . Figure 7 and Table S6, show the outcomes of apoptosis analysis. As could be seen in Fig. 7, the blank βCD-g-PMA-co-PLGA micelles did not show noticeable toxic effect (83.4% cell viability) to MDA-MB-231 cells. The Conf loaded βCD-g-PMA-co-PLGA micelles (BC) was not shown noticeable apoptosis and its result was similar to the result of blank βCD-g-PMAco-PLGA micelles (PB). Because IC 50 dosage (0.259 μg/mL) of co-drug loaded βCD-g-PMA-co-PLGA micelles was selected for all drug-loaded micelles in this test which is much lower than IC 50 dosage of Conf loaded βCDg-PMA-co-PLGA micelles (3.567 μg/mL). According to the results the co-drug loaded βCD-g-PMA-co-PLGA micelles showed synergistic effect with highest apoptosis (98.7%) and lowest necrosis (1.33%) compared to single-drug (Dox or Conf) loaded βCD-g-PMA-co-PLGA micelles. These results confirm that Dox-Conf loaded micelle acts as an effective intracellular co-delivery system that enhances combinational apoptosis-inducing effect. As expected from the results of cell cycle and MTT tests, the highest anticancer effect was observed in co-drug loaded βCD-g-PMA-co-PLGA micelles. This is the consequence of synergistic effect of drugs in co-drug loaded βCD-g-PMA-co-PLGA micelles as well as promotive effect of Conf on Dox intracellular accumulation.   www.nature.com/scientificreports/ (the caspases) are the essential factors for apoptosis and are divided into initiator (Caspase-8 and -9) and effector caspases (caspase-3 and -7) 56 . The caspase-8 and 12 upregulation are the signs of extrinsic pathway of apoptosis, but the caspase-9, caspase-3 and -7 upregulation show the intrinsic or mitochondria mediated pathway of cell apoptosis. Therefore, all the mentioned factors regulation changes after treating of cells with each formulation, which were investigated in real-time PCR analysis. The results of real-time PCR test were presented, as the heat map, of change in gene expressions related to the control group (gene expression = 1, Fig. 8A). In the heat map the light-yellow represented the lack of gene expression and red presented higher expression of genes. According to Fig. 8A, gene expression and regulation in blank βCD-g-PMA-co-PLGA micelles, compared to control group, did not show significant changes that confirmed its non-toxicity on MDA-MB-231 cells. Except for blank βCD-g-PMA-co-PLGA micelles, the rest of formulations showed concurrent Bax upregulation and Bcl-2 downregulation with the following order: co-drug loaded βCD-g-PMA-co-PLGA micelles > Dox-loaded βCD-g-PMAco-PLGA micelles > free Dox ≈ Conf-loaded βCD-g-PMA-co-PLGA micelles ≈ free Dox-Conf > free Conf. The mentioned concurrent regulations cause upregulation of caspase-9 expression with the same order. As a result of caspase-9 upregulation, the caspase-3 and -7 were activated and caused cell apoptosis. Therefore, the Bax, caspase-9, caspase-3, and caspase-7 were upregulated dominantly while the Bcl-2, was downregulated significantly in our nano formulations (Co-drug loaded, Dox-loaded and Conf-loaded βCD-g-PMA-co-PLGA micelles). Based on the results, it can be concluded that the nano-formulations specially co-drug loaded βCD-g-PMA-co-PLGA micelles caused the higher level of cell apoptosis via caspase-dependent and intrinsic pathway of apoptosis. This superiority conforms with results of previous research about the effect of Dox-Conf loaded micelles on apoptosis pathway in MDA-MB-231 cell line 6    www.nature.com/scientificreports/ Investigation of apoptosis pathway by western blotting. Since real-time PCR results showed that the highest level of caspase-dependent intrinsic pathway of apoptosis (at gene level) was induced by co-drug loaded βCD-g-PMA-co-PLGA micelles, the effect of co-drug loaded βCD-g-PMA-co-PLGA micelles on Bax, Bcl-2, pro-caspase-9, cleaved-caspase-9, pro-caspase-3, cleaved-caspase-3, pro-caspase-7, cleaved-caspase-7, p27 and p53 were evaluated using western blotting (at protein level). Cyclin dependent kinase inhibitor or KIP1 (p27) and the other tumor-suppressor proteins such as p53, are the cell cycle inhibitors. Upregulation of p27 and p53 cause Bax upregulation and Bcl-2 downregulation which lead to cell apoptotic death [65][66][67] . In the case of malignant tumors, the downregulated p53 prevents from apoptotic death 68 . Figure 8B,C and Table S7, present the western blotting results and fold-changes of protein expression in MDA-MB-231 cells treated by co-drug loaded βCD-g-PMA-co-PLGA micelles. These results show noticeable increase in expression of Bax (1.75-fold), cleaved-caspase-9 (5.41-fold), cleaved-caspase-3 (14-fold), cleaved-caspase-7 (22.55-fold), p27 (3.2-fold), p53 (2.87-fold), and decrease in expression of Bcl-2 (0.67-fold), pro-caspase-9 (0.37-fold), pro-caspase-3 (0.56fold) and pro-caspase-7 (0.34-fold), with respect to the control group. The upregulated p27 and p53, induced upregulation of Bax and reduction of Bcl-2 which caused a severe disturbance to cell cycle, and subsequently cell apoptosis. Increasing in Bax and decrease in Bcl-2 expression led to cytochrome-c release from mitochondria, and hence creation of apoptosome which led to pro-caspase-9 expression. Upregulation of pro-caspase-9, caused cleavage of caspase-9 in parallel with pro-caspase-9 downregulation. Cleaved-caspase-9 upregulation caused pro-caspase-3 and pro-caspase-7 upregulation and their cleavage (upregulation of cleaved-caspase-3 and cleaved-caspase-7). Finally, cleavage of death substrate increased and hence fragmentation of DNA was occurred as a result of upregulation of cleaved-caspase-3 and cleaved-caspase-7. Because of marked increase in expression of cleaved-caspase-9, -3 and -7, it was proved that the co-drug loaded βCD-g-PMA-co-PLGA micelles induced apoptosis to MDA-MB-231 cells via intrinsic mitochondrial pathway (p27, p53, Bcl-2/Bax, cleaved-caspspase-9, cleaved-caspase-7 and cleaved-caspase-3 axis) which confirmed the real-time PCR outcomes. Similarly showed that Dox-conferone loaded micelles induced apoptosis via intrinsic Bcl-2/Bax, cleaved-caspase-9, cleaved-caspase-7, cleaved-caspase-3 and p27 pathway 6 . Therefore, it can be stated that Dox combination therapy on MDA-MB-231 cells induces apoptosis via activation of the intrinsic pathway. Our novel developed co-drug loaded βCD-g-PMA-co-PLGA micelles acted with similar intrinsic apoptosis pathway.

Conclusion
The new pH-sensitive and biodegradable βCD-grafted poly maleate-block-PLGA micelles was developed for codelivery of Doxorubicin (Dox) and Conferone (Conf) into MDA-MB-231 cell line. Micelles with very low CMC (0.1 μg/mL), small size (34.5 nm) and negative zetapotential were obtained. The co-drug loaded βCD-g-PMA-co-PLGA micelles and Dox loaded βCD-g-PMA-co-PLGA micelles had a pH-sensitive and sustained drug release. The blank βCD-g-PMA-co-PLGA micelles and co-drug loaded βCD-g-PMA-co-PLGA micelles were internalized quickly (0.5 h) and completely (100%) into MDA-MB-231, because of their favorable size and zetapotential. The lowest IC 50 (0.259 μg/mL) was obtained in B2D nano-formulation because of: synergistic effect of Conf on Dox (CI = 0.529), inhibition of P-gp expression and Dox efflux by Conf in MDA-MB-231 cells. Furthermore, co-drug loaded βCD-g-PMA-co-PLGA micelles with G2/M arrest, caused a severe disturbance to cell cycle and therefore induced exceptional apoptosis (up to 98%, according to cell cycle and apoptosis tests). The induced apoptosis of MDA-MB-231 cells by co-drug loaded βCD-g-PMA-co-PLGA micelles was confirmed with real-time PCR (at gene level) and western blotting (at protein level) that proved the p27, p53, Bax/Bcl-2; caspase-9; caspase-7 and caspase-3, intrinsic mitochondrial apoptosis pathway. The new Dox-Conf loaded βCD-g-PMA-co-PLGA micelles improved Dox therapeutic function by minimizing Dox therapeutic dosage. Thus, based on the excellent capabilities for apoptosis induction, βCD-g-PMA-co-PLGA micelles loaded with Dox in combination with Conf as adjuvant are suggested for in-vivo application in the future animal studies. We also aim to draw the attention of the scientific community to more consider the mechanisms involved in the synergism effect of combination therapy of anticancer drug and adjuvants with reduced side effects, and conduct clinical studies, for the development of alternative therapeutic way to benefit cancer patients worldwide.