Doxorubicin and siRNA Co-Delivery System Based on Carbon Dots Inhibits Chemoresistance of Lung Cancer

The resistance to the anti-cancer agent limits the chemotherapy effect in the cancer therapy. Tumor easily develops resistance to anti-cancer drugs leading to decreased therapy eciency of chemotherapies. Targeting signaling molecules related with chemoresistance through strategy of co-delivery siRNA and chemotherapeutics may overcome the multidrug resistance to chemotherapy. A co-delivery nanosystem that could carry siRNA and DOX simultaneously has been studied in this work. The co-delivery is based on carbon dots was surface-modied with poly-ethylenimine (PEI), and loaded the siMRP1 and chemotherapeutics by electronstatic interactions on the surface with pH-triggered drug release. The CD-PEI was synthesized by one-step microwave assisted method; the PEI were raw materials and passivator during the reaction process that makes CD exhibit excellent optical property and the capability of loading siRNA. The CD-PEI was capable of loading and delivering siMRP1 and DOX to tumor and release synchronously in cell by pH-triggered manner using ow cytometry and confocal laser scanning microscopy analysis. MRP1 was successfully knocked down by siRNA. The silencing of MRP1 by co-delivery system could increase DOX accumulation and signicantly enhance the inhibitory effect of metastatic potential elicited by doxorubicin in A549 and A549/ADM cells. loaded released siRNA DOX agents targeted overcoming resistant By suppressing MRP1, CD-PEI-DOX-siMRP1 increase implying the intercellular uptake is increased and more drugs located at the perinuclear regions. And the co-delivery system incorporating suppression of MRP1 and anticancer agents could enhance anti-cancer ecacy against chemoresistance in the lung cancer. The real-time tracking of the delivery process is realized through the excellent optical property of co-delivery carrier. Results from our study would deepen the regulator network of MRP1 in modulating chemoresistance of malignant cancer. It would fasten the clinical potential of co-delivery system with a combination of chemotherapeutics and siRNA sequences.

delivery needs the synthesis of nanocarrier complex, and the selective release of co-loaded agents is fairly di cult.
Carbon dots (CD) is a novel kind of the quantum-sized nanomaterials, which have enormous application potential in bio-imaging, drug carrier, and nanosystems [7][8][9]. The application of semi-conductor quantum dots (QDs) is limited by the toxic of heavy metal. However, the carbon dots are composed by carbon atoms, which have excellent inherent biocompatibility, have the feasibility in the biotechnology application. CD have the capability of carry chemotherapeutic to the tumor and decrease cytotoxicity and enhance chemotherapy e cacy [10], and the CD with surface-modi ed groups can carry multiple drugs simultaneously [11,12].
Lung cancer was reported to be easily resistant to drugs after several cycles of chemotherapy [13].
Doxorubicin was widely used in treating lung cancer. A plethora of proteins were identi ed in resistance against chemotherapy drugs in treating lung cancer. Multidrug resistant protein 1 (MRP1) is a pivotal member of ABCC subfamily belonging to ATP-binding cassette transporter superfamily, which transport anti-cancer drugs out of cells thus reducing the treatment e cacy [14]. How MRP1 was involved in chemoresistance against doxorubicin was elusive [15]. Whether MRP1 was involved in uptake of doxorubicin and in uencing therapeutic e cacy of it warrants further study. Using small interfering RNA (siRNA) to downregulate proteins associated with multidrug resistance, such as MRP1, is a promising method for reversing drug resistance [16].
In this work, the CD are fabricated by microwave and passivized by polyethyleneimine (PEI). There are a lot of functional groups at the surface of CD due to the addition of PEI during the microwave synthesis process. The surface of CD is modi ed by PEI, as a passivation agent, in order to enhance the uorescence intensity of CD and possess the capability of loading and delivery siRNA. The siRNA and DOX anti-cancer agents are delivery simultaneously to the targeted tumor by the CD with surfacemodi ed PEI. By suppressing drug e ux pumps, the intercellular uptake is increased and more drugs located at the perinuclear regions. And the co-delivery system incorporating suppression of MRP1 and anticancer agents could enhance anti-cancer e cacy against chemoresistance in the lung cancer. The real-time tracking of the delivery process is realized through the excellent optical property of co-delivery carrier. Results from our study would deepen the regulator network of MRP1 in modulating chemoresistance of malignant cancer. It would fasten the clinical potential of co-delivery system with a combination of chemotherapeutics and siRNA sequences.

Results And Discussion
The characteristic of CD-PEI and nanodrug system based on CD-PEI The TEM images showed that the CD-PEI was uniform and spherical, with an average size of 4.25 nm which obtained by measuring the sizes of hundreds CD nanoparticals in TEM images. It is indicated that the average diameter of CD-PEI is 5.503 nm measured by dynamic light scattering (DLS). CD-PEI was dark yellow under white light and well-distributed in aqueous solution. Under the UV light, the CD-PEI solution had obvious green emission indicating that CD-PEI had excellent uorescence property, as shown in Fig. S1. The photoluminescence (PL) spectra of CD-PEI and CD-PEI-DOX-siMRP1 had been characteristic by various excitation wavelengths with 20 nm increments. The emissions of CD are redshifted with the increase of the excitation wavelength. The emission in CD could result from the surface effects and nanometer quantum con nement effect [17,18]. The PL spectra of CD-PEI and CD-PEI-DOX-siMRP1 were obviously different. Due to the loading of DOX agents, the PL and UV-vis spectrum has changed. The emission intensity of CD-PEI-DOX-siMRP1 is lower than that of CD-PEI in spectra, and the peaks have red-shift. The increase of emission peak con rmed the DOX has been successfully loaded on the surface of CD-PEI [19]. The CD-PEI exhibited excellent and stable PL property that is bene cial for the tracking of drug delivery in vivo. In addition, the CD-PEI possess excitation-dependent emission behavior [20,21], that when the excitation wavelength increase from 330 nm to 510 nm, the emission peak shift from 450 to 600 nm. When the excitation wavelength is 350 nm, the emission intensity is highest. In this work, we studied the PL behavior of CD-PEI-DOX-siMRP1 through increasing excitation wavelength from 270 to 510 nm. The CD-PEI-DOX-siMRP1 had no obvious emission peak when the excitation wavelength is lower than 330 nm. When excitation wavelengths increase from 350 nm to 510 nm, the emission intensity of CD-PEI-DOX-siMRP1 is increasing, and reaches maximum at 470 nm. The CD-PEI-DOX-siMRP1 has two emission peak centers at 560 and 690 nm.
There is obvious absorption band correspond to the O-H stretching in the FTIR spectrum of the CD. The absorption peaks centered at 1096 cm -1 and 1655 cm -1 are corresponded to the C-O and C=O stretching, respectively. The absorption bands at 1453 cm -1 and 2949 cm -1 are corresponded to the C-H and CH 2 stretching, respectively. The UV-Vis spectrum of CD-PEI had two bands at 304 nm and 346 nm, corresponding to the π-π * transition and n-π * transition [22][23][24]. The CD-PEI-DOX, CD-PEI-DOX-siNC, and CD-PEI-DOX-siMRP1 had a broad peak band at 480-600 nm among which was the DOX characteristic absorbance peak. These results con rmed that the DOX was loaded on the surface of CD-PEI, and the drug loading e ciency of CD-PEI was 16.7%.
The drug release behavior of CD-PEI-DOX-siMRP1 was revealed by in vitro release test. The free DOX release totally during 50 hours, while CD-PEI-DOX-siMRP1 prolong the drug release time up to 72 hours when pH was 5.2. CD-PEI-DOX-siMRP1 released fairly minimal DOX under pH 7.4 and illustrated that the nanocarrier would avoid the detrimental release of DOX in normal tissue and cell. When the pH changed from 7.4 to 5.2, the release amount of DOX from CD-PEI-DOX-siMRP1 increased about 5-times indicating that CD-PEI-DOX-siMRP1 was pH-sensitivity and triggered the release of DOX by acid environment because the water solubility of DOX was increased at lower pH. The advantage of selective drug release behavior of nanosystem could enhance the therapeutic effect of tumor and decrease the toxic to normal tissue.
The information of the chemical composition and functional groups of nanodrug were characterized by X-ray photoelectron spectroscopy (XPS) in Fig. 1I. As shown in Fig. 1, the XPS spectra of CD-PEI-DOX-siMRP1 exibit three peaks corresponding to C1s peak at 284.5 eV, N1s peak at 399.5 eV, and O1s peak at 531.5 eV, respectively. This result indicates that CD-PEI-DOX-siMRP1 are mainly composed of C, N, and O, with some small inorganic elements due to the synthesis while attaining a solution system at pH 7.4 with PBS [25]. The C1s spectrum of CD-PEI-DOX-siMRP1 (Fig. 1J) shows four peaks at 284.8 eV, 285.8 eV, 287.7 eV, and 288.3 eV, which are attributed to C-C/C=C (78.4%), C-N/C-OH (16.4%), C=O (2.7%), and O-C=O (2.5%), respectively. The N1s spectrum of CD-PEI-DOX-siMRP1 (Fig. 1k) exhibits two components located at 398.9 eV and 400.2 eV, assigned to C=C-N (80.5%) and N-(C) 3 (19.5%) groups. The O1s spectrum of CD-PEI-DOX-siMRP1 (Fig. 1L) shows two peaks at 531.0 eV and 531.9 eV, which are corresponded to the C=O (47.7%) and C-OH/C-O-C (52.3%) groups. As shown in Table S1, the change in the chemical groups and the percentage of them in the spectra gives further evidence of successful complex conjugation.
MRP1 was involved in chemoresistance of A549 against CD-PEI-DOX.
Failure of the chemotherapy to malignant tumor was mainly attributable to insensitivity to drugs.
Elucidating the mechanisms how tumor cells modulate the chemoresistance was critical for improving the chemotherapeutic effect of various drugs. We previously con rmed that CD-PEI-DOX treatment in hepatocellular carcinoma inhibit tumor growth through actively targeting tumors [10]. This led us to postulate: whether CD-PEI-DOX would elicit side effects on malignant tumors such as increasing chemoresistance in lung cancer. We rstly detected the expression of molecules associated with chemoresistance including p-glycoprotein (P-gp), MRP1, and ABCG2 in adherent, sphere and chemoresistant lung cancer cell line A549. From Fig. 3A, we vividly observed that molecules related with chemoresistance were elevated in spheres formed by A549 cells, which were considered to bear more traits of stemness compared with adherent cells. Consistently, the expression of these molecules elevated most in A549/ADM cells. These results show that A549/ADM cells expressed a panel of molecules involved in chemoresistance. Subsequently, we analyzed the cell viability of A549 and chemoresistant A549/ADM cells after doxorubicin treatment (Fig.3B). We discovered that A549/ADM cells were resistant to doxorubicin treatment. To investigate the involvement of MRP1 in chemoresistance of A549 against doxorubicin, we tentatively knockdown the expression of MRP1 using siRNA transfection. We determined the best knockdown e ciency of siRNA targeting MRP1 using qPCR methods. From Fig. 3C and 3D, we determined siRNA #2 was the most e cient sequence targeting MRP1. CD-PEI-DOX were conjugated with this sequence with best e ciency and CCK8 assays were performed to analyze the cell viability after treatment by various drugs. The IC 50 of drugs in both A549 and A549/ADM cells were calculated from three independent cell toxicity experiments. As shown in Fig. 4A-4B and Table 1, free doxorubicin in A549/ADM cells was approximately 5 fold higher compared with A549 cells. Accordingly, the IC 50 of CD-PEI-DOX in A549/ADM cells were markedly higher (151.7μg/mL) than that (6.72μg/mL) in the A549 cells.
When CD-PEI-DOX were loaded with MRP1 siRNA sequences, the cells became more vulnerable to CD-PEI-DOX, IC 50 was only 17.4μg/mL in A549/ADM cells. The CD-PEI has been proved to be non-toxic to cancer cells and normal cells, as shown in Fig. S3.
Based on these ndings, we conclude that the free DOX exhibited different toxic behavior to A549 and A549/ADM cells. While CD-PEI-DOX-siMRP1 treatment caused more toxicity to A549/ADM cells compared with free DOX and CD-PEI-DOX groups at all-time points. It indicated that CD-PEI-DOX-siMRP1 possess better antitumor effect than DOX and CD-PEI-DOX. CD-PEI-DOX-siMRP1 possesses high toxic to A549/ADM due to the co-delivery and synergistic effect of siRNA and DOX, indicating that the MRP1 was critically involved in chemoresistance of A549/ADM cells against doxorubicin. The CD-PEI have the capability of transfection siRNA to the cells and interfere the expression of MRP1. Fig. 4D shows that the expression of MRP1 in A549/ADM cell could be suppressed by CD-PEI-DOX-siMRP1.
These results led us to postulate: whether MRP1 was induced by CD-PEI-DOX and suppressing MRP1 would reverse this effect? To test our hypothesis, we treated both A549 and A549/ADM cells with the indicated drugs and analyzed the expression of MRP1 after treatment. As demonstrated in Fig. 4F-4F, the expression of MRP1 increased markedly in case of CD-PEI-DOX treatment. In contrast, MRP1 expression decreased after siRNA targeting MRP1 only in A549/ADM cells (Fig. 4F) was loaded on CD-PEI-DOX particles. Combined with the cell viability results, knockdown the expression of MRP1 augment the killing effect of CD-PEI-DOX. The internalization of DOX and nanodrug system based on CD-PEI The cellular uptake e ciency of free DOX, and CD-PEI-DOX-siMRP1 has been studied by laser scanning confocal microscopy and ow cytometry. The A549 and A549/ADM cells were incubated for 24 hours with PBS, free DOX, CD-PEI-DOX, CD-PEI-DOX-siNC, and CD-PEI-DOX-siMRP1, respectively. The results revealed that more CD-PEI-DOX-siMRP1 is uptaken by the cell and located in the perinuclear regions and the nuclei. The phenomenon revealed that the CD-PEI-DOX-siMRP1 could enhance the permeability of DOX in to the nuclei, which is bene cial to increase the therapeutic e ciency of DOX. In the A549/ADM, the CD-PEI-DOX-siMRP1 may increase the uptake e ciency of the DOX into the cell, and the DOX wrapped by nanocarrier could enter into the A549/ADM more than the free DOX group, for the reason that the nano carrier may increase the permeability of free drug into the cells. The permeability of free DOX, and DOX wrapped by CD-PEI has been investigated using A549 and A549/ADM cell spheres by the Zstack mode of laser scanning confocal microscopy. The Z-stack mode could be used to observe the penetrate situation of different treatment by gaining the different layers of the cell mammosphere. The results revealed that the group treated by free DOX has the weakest permeability among all the groups. And the CD-PEI-DOX-siMRP1 could penetrate the inner region of mammosphere, as shown in Fig. 6. This result con rmed that the CD-PEI-DOX-siMRP1 could have the permeability to enter into the tumor tissue, which can maximize the therapeutic effect.
The uptake of DOX and nanodrug system has been measured by ow cytometry. Fig. 7 is the uptake in A549 and A549/ADM cells during the different intervals. The group incubated with CD-PEI-DOX-siMRP1 have higher mean uorescence intensity than other group in both A549 and A549/ADM, indicated that CD-PEI-DOX-siMRP1 could delivery more DOX into cell. The cellular uptake of free DOX is lower than other treated groups at every interval in both A549 and A549/ADM cells. The nanodrug could decrease the drug e ux and gather high intracellular concentration of the drug, because the nanodrugs bypass the e ux pumps. And in the A549/ADM, the uptake amount of CD-PEI-DOX-siMRP1 is higher than other groups, results from the disrupting of siRNA to cellular pathway. This result indicated that the disturbing MRP1 could enhance the uptake of nanodrug effectively. The co-delivery DOX and siRNA would more effective in overcoming resistance of cancer cells.
CD-PEI and CD-PEI-siMRP1 enhanced the inhibitory effect on stemness and metastatic potential elicited by doxorubicin The ability of metastasis is the critical factor that in uence the therapeutic e ciency of tumor. We used the Transwell assay to investigate the migration and invasion in A549 and A549/ADM cells affected by respective drugs. Both A549 and A549/ADM treated with DOX, CD-PEI-DOX and CD-PEI-DOX-siNC and CD-PEI-DOX-siMRP1 exhibited decreased migration and invasion compared to the control, with migration counts shown in Fig. 8, respectively. The results indicated that CD-PEI-DOX and CD-PEI-DOX-siNC and CD-PEI-DOX-siMRP1 generated stronger inhibitory effects than free DOX. We found that CD-PEI-DOX distinctly inhibited the migration and invasion of both A549/ADM and ADM cells compared with the group treated by free DOX. After silencing MRP1, the decline in migration and invasion was reversed elicited by CD-PEI-DOX treatment in chemoresistant A549 cells. These results collectively showed that MRP1 is critical for mediating the chemoresistant effect of A549 against doxorubicin.
To elucidate the regulation of stemness in lung cancer by MRP1, we conducted sphere-forming assays. As revealed in Fig. 9, CD-PEI-DOX-siMRP1 treatment speci cally suppressed the size and number of spheres in A549/ADM cells. A conclusion could be drawn from the above ndings that suppressing MRP1 expression attenuated the resistance to regulation of stemness, migration and invasion induced by CD-PEI-DOX in DOX-resistant A549 cells.

Targeting and biocompatibility analysis of CD-PEI-DOX-siMRP1
Based on the ex vivo results, we conducted in vivo experiments to analyze the targeting of CD-PEI-DOX-siMRP1. As shown in Fig. 10A, speci c binding of CD-PEI-DOX-siMRP1 particles were observed, while no marked binding were seen in free doxorubucin treated group. CD-PEI-DOX-siMRP1 located on tumors spec cally at 12h. Tumors and organs were harvested, the uorescence intensity were stastistically analyzed (n=3). CD-PEI-DOX-siMRP1 bear strong binding on tumors compared with dox group (Fig. 10B-10C). Hemolysis assays were conducted to analyze the toxic effect of our nanodrugs. The results revealed that CD-PEI-DOX-siMRP1 have no obvious toxic effect on blood among all groups, even at high concentration. Triton X-100 was included as a positive control (Fig. 10D-10E). Strong biocompatibility of CD-PEI-DOX-siMRP1 could be concluded since no obvious changes were seen in routine blood test (Fig.   10F). Suseuqently, slices from organs of each group were obtanied and subjected to HE staining. No destructive structures of drugs on different organs were seen (Fig. 10G). These results collectively show that CD-PEI-DOX-siMRP1 consistently target tumor and bear good biocompatibility and no obvious toxic effect in vivo.

Conclusion
In this work, we offered a new approach to overcome the chemoresistance through a strategy of combining siRNA targeted to the MRP1 with chemotherapeutic by the co-delivery system based on CD-PEI. The surface of carbon dots was decorated with PEI groups, having the capability of carrying siRNA, and the DOX agent was loaded through electrostatic interactions. The co-delivery systems effectively loaded and released siRNA and DOX agents to the targeted tumor, overcoming the resistance to chemotherapy. By suppressing MRP1, CD-PEI-DOX-siMRP1 can obviously increase the drug intercellular accumulation and inhibit the cancer cell proliferation. The migration and invasion of A549/ADM cells were inhibited by CD-PEI-DOX-siMRP1. These results con rmed that the co-delivery systems enhance the therapy e ciency and overcome the chemoresistance. CD-PEI-DOX-siMRP1 co-delivery systems could inhibit the proliferation, migration, and invasion of the cancer cell because of the synergistic treatment from siRNA and anti-cancer agent. These works pave the way for exploring the application of co-delivery system underlying chemoresistance of anti-cancer drugs in treatment of lung cancer.

Materials And Methods
The synthesis of CD-PEI CD-PEI was fabricated by the microwave-assisted method with PEI-passivated at the surface. The glycerol and PBS solution (pH 7.4) have been mixed togather, and PEI (MWCO, 25 kDa) was added into the homogeneous solution. The mixed solution has been transform into a beaker, and placed into a microwave oven (80P) for heating 10 minutes. Then, the reaction product was diluted by 10 mL ultrapure water and dialysised for 2 days to remove the unreacted agents. The solution after dialysis was lyophilized and stored at 4 ℃.

The characterization of CD-PEI and a nanodrug system based on CD-PEI
The TEM image of CD-PEI was obtained by high-resolution transmission electron microscopy (JEM-2100) and the X-ray photoelectron spectroscopy of CD-PEI and CD-PEI-DOX were measured by a Thermo 250Xi Thermo K-Alpha. The UV-Vis absorbance spectra of the CD-PEI, CD-PEI-DOX, CD-PEI-DOX-siNC, and CD-PEI-DOX-siMRP1 were measured by a Shimadzu UV3600. The PL spectra were characterized by a ourescence spectro uorometer (Edinburgh, FLS 980-STM).

The loading and releasing of DOX
The DOX was loaded onto CD-PEI by electrostatic interactions, which has been studied in our previous work [10]. Brifely, the CD-PEI solution mixed with DOX was shaked for 24 h at 4 ℃ and then dialysised for 3 days to remove the excess DOX agent. The 40 nmol/L MRP1 siRNA was added into the CD-PEI-DOX solution and shaked for 24 h at 4 ℃ in order to combine them through electrostatic interactions.
The drug loading e ciency of DOX was calculated by the absobance at 480 nm. The DOX releasing by CD-PEI-DOX-siMRP1 was measured by the drug release experiment as following: 2 mL CD-PEI-DOX-siMRP1 solution was added into a dialysis bag (MWCO, 3  penicillin(100 unit mL -1 ) and streptomycin (100 µg mL -1 ) (Gibco, Carlsbad, CA, USA) were added into the DMEM solution. The cells were cultured in a incubator with 5% CO 2 at 37 ℃.

Cell cytotoxicity
The Cell Counting Kit-8 (CCK8) was used to measure the cell viability.

Western blot assays
Western blot assays were carried out as previously described [26]. Breie y, total cell lysates were prepared in RIPA buffer (Beyotime, Shanghai, China). and proteins were separated by the sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. After blocking with 5% skim milk for 2 h, the membranes were incubated with the primary antibodies against Growth inhibition of sphere forming.

Animal Study in vivo
All animal studies were approved by Animal ethics committee of the Fifth A liated Hospital Sun Yat-sen University. Brie y, A549 cells were subcutaneously inoculated into 4 weeks old male Balb/c nude mice to construct xenograft models. When tumors grew to palpable size, DOX and CD-PEI-DOX-siMRP1 were injected through tail vein. The bio-distribution of drugs were visualized by in vivo image system (IVIS) equipment. Hemolysis experiment were performed to assess the biocompatibility of drugs. The targeting of drugs on tumor were viusualized and photographed as the indicated time points. Penetration of the drugs into tumor were obtained using confocal microscopy LSM880 (Zeiss).

Immunohistochemical assays
Tumor and organs were obtained after mouse were scari ced humanely. Slices were prepared from frozen tissues. Immunohistochemical staining were conducted as prveiously described. Hematoxylin and eosin (H&E) staining (ZSBG-BIO, Beijing, China) were carried out to investigate the toxic effect of drugs.
Representative pictures were shown and the distribution of different drugs were visulized undrer optical microscopes (BX53 System Microscope, Olympus, Japan).

Statistical analysis
The data between groups was statistical analyzed by Student's t-test. P<0.05 was used as the criterion for statistical signi cance.

Consent for publication
Not applicable.

Declaration of interest
The authors declare that they have no known competing nancial interests or personal relationships that could have appeared to in uence the work reported in this paper.

Availability of data and materials
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study. Funding: The