The effect of linkers on the self-assembling and anti-tumor efficacy of disulfide-linked doxorubicin drug-drug conjugate nanoparticles

Accepted Manuscript The effect of linkers on the self-assembling and anti-tumor efficacy of disulfide-linked doxorubicin drug-drug conjugate nanoparticles

Yaoqi Wang, Xing Wang, Feiyang Deng, Nan Zheng, Yanqin Liang, Hua Zhang, Bing He, Wenbing Dai, Xueqing Wang, Qiang Zhang PII: DOI: Reference:

S0168-3659(18)30203-7 doi:10.1016/j.jconrel.2018.04.019 COREL 9246

To appear in:

Journal of Controlled Release

Received date: Revised date: Accepted date:

26 December 2017 26 March 2018 11 April 2018

Please cite this article as: Yaoqi Wang, Xing Wang, Feiyang Deng, Nan Zheng, Yanqin Liang, Hua Zhang, Bing He, Wenbing Dai, Xueqing Wang, Qiang Zhang , The effect of linkers on the self-assembling and anti-tumor efficacy of disulfide-linked doxorubicin drug-drug conjugate nanoparticles. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Corel(2018), doi:10.1016/ j.jconrel.2018.04.019

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT The effect of linkers on the self-assembling and anti-tumor efficacy of disulfide-linked doxorubicin drug-drug conjugate nanoparticles

Yaoqi Wang1# , Xing Wang1# , Feiyang Deng1 , Nan Zheng2 , Yanqin Liang1 , Hua Zhang1 ,

1

PT

Bing He1 , Wenbing Dai1 , Xueqing Wang1 *, Qiang Zhang1,3

Beijing Key Laboratory of Molecular Pharmaceutics and New Drug Delivery

RI

Systems, School of Pharmaceutical Sciences, Peking University, Beijing 100191,

2

SC

China

National Drug Clinical Trial Center, Key laboratory of Carcinogenesis and

NU

Translational Research (Ministry of Education), Peking University Cancer Hospital & Institute, Beijing 100142, China

State Key Laboratory of Natural and Biomimetic Drugs, Peking University, Beijing

MA

3

ED

100191, China

*Address correspondence to this author at the Beijing Key Laboratory of Molecular

Peking

EP T

Pharmaceutics and New Drug Delivery Systems, School of Pharmaceutical Sciences, University,

Beijing

100191,

China;

#

AC C

+86-10-82805935; E-mail: [email protected]

Both authors contributed equally to this manuscript.

Tel:

+86-10-82805935;

Fax:

ACCEPTED MANUSCRIPT Abstract: Drug-drug conjugate nanoparticles (DDC NPs) is a potential method for overcoming poor solubility and nonspecific action in cancer therapy, which is based on its high drug loading efficiency and passive tumor-target properties. Our laboratory has prepared DOX-SS-DOX NPs based on disulfide-linked doxorubicin (DOX) drug-drug

PT

conjugate, which showed well physical stability and similar anti-tumor efficacy as liposomes. However, how structures of DDCs influence the self-assembling and

RI

anti-tumor efficacy is still seldom clarified and needs further investigation. Here, we discussed the role of linker types, length and linkage site in the NPs self-assembling

SC

and anti-tumor efficacy. A series of DOX prodrugs were prepared and all the prodrugs could self-assemble into NPs except DOX-SS-DOX (2), indicating the linker length

NU

played an important role during self-assembling process. The linkage sites and types of linker exhibited great influence on in vitro cytotoxicity and in vivo anti-tumor

MA

efficacy, particularly, modification on C-14 hydroxyl was more efficient for DOX release than on amino group. Besides, disulfide-bond was not cleaved and DOX-SH

ED

release did not occur in the metabolism process. The function of disulfide-bond was to enhance the release of DOX in the hydrolysis process. These findings is meaningful

EP T

for effective prodrug NPs design for therapeutics.

Keywords: Doxorubicin conjugates, Self-assembling nanoparticles, Linker type,

AC C

Linker length, Linkage site, Anti-tumor efficiency

ACCEPTED MANUSCRIPT 1. Introduction Nano-drug delivery systems based on small drug conjugates (SDC), including peptide–drug conjugate nanoparticles (PDC NPs) [1, 2], fluorescence probe-drug conjugate NPs (FDC NPs) [3-5], drug-drug conjugate NPs (DDC NPs) [6], have attracted great research interests. Apart from the tumor passive targeting capacity via

PT

enhanced permeability and retention (EPR) effect as traditional NPs [7,8], SDC NPs demonstrated great advantages in (1) high drug loading, which nearly reaches 100%

RI

[9]; (2) avoiding the side-effects from carriers [6]; (3) precise chemical structure, whose metabolism process could be readily monitored. Compared with physical

SC

mixture, the conjugation of two anticancer drugs through a biodegradable bond exhibited synergetic cytotoxicity to the tumor cells [6]. For example: Wang et al.

NU

prepared DDC NPs from conjugates of methotrexate (MTX) and gemcitabine (GEM) with high drug loading capacity (100%), in which MTX and GEM could be

MA

self-delivered without any carriers and release synchronously in cancer cells [9]. There have been some achievements in SDC NPs study. For example, PDC NPs are

ED

utilized to improve the drug solubility by linkage with hydrophilic peptides and to provide additional targeting effect [1]. Our lab has reported four amphiphilic PDC

EP T

NPs (APDCs@NPs) with maytansinoid (DM1) as the cytotoxic agent and cRGDfK as the homing peptide, which significantly decreased the toxicity of free DM1 and improved its therapeutic effect via passive and RGD- mediated targeting strategies

AC C

[10]. Another example is FDC NPs, which combine therapeutic modalities and diagnostic imaging agents to develop theranostics. Xie and his co-workers reported a reduction-responsive fluorescence off–on theranostic prodrug NPs based on boron dipyrromethene (BODIPY) and camptothecin (CPT) with a long flexible disulfide linker, whose self-reporting drug release demonstrated theranostic drug uptake-related imaging that might allow more precise monitoring of dosage levels, as well as an improved understanding of the cellular uptake and release mechanisms [3]. For specific application of SDC, various smart linkers are designed, including hydrolyzable ester linkage [6], reductive-responsive disulfide bond [11], acid cleavable covalent links (pH-responsive) [12], enzyme-sensitive linkers [13], redox

ACCEPTED MANUSCRIPT dual-responsive linkers [14] and multiple stimuli-activated linkers [15]. Particularly, disulfide bond is an attractive linker for construction of stimuli-responsive delivery nano-system [3] because it is not only susceptible to the reductive environment but also vital for self-assembling nanostructures formation. It has been reported that the insertion of a single disulfide bond into hydrophobic molecules could promote and

PT

stabilize the self-assembling of nanomedicines by balancing the competition between intermolecular forces [16].

RI

Despite the mechanism of disulfide bond for SDCs formation has been clarified as mentioned above, the research on how structures of SDCs influence the

SC

self-assembling and efficacy is still limited and needs further investigation. We have previously prepared doxorubicin drug-drug conjugate nanoparticles (DOX-SS-DOX

NU

NPs) with dithiobis (succinimidyl propionate) (DSP), a cleavable disulfide linker, conjugating with the DOX via its amino group, with three carbon atoms between

MA

disulfide bond and DOX. The NPs showed well physical stability and similar anti-tumor efficacy as liposomes, which set up a satisfactory model for further

ED

improvement [17]. In this article, we investigated the effect of SDC structure, such as the type and length of linker and the linkage site, on the self-assembling and the

EP T

anti-tumor efficacy of the NPs to provide guidance for design of DDC NPs. As shown in Scheme 1, a series of DOX DDCs were synthesized: (1) To investigate the effect of length of linkers on the self-assembling and the anti- tumor efficacy of the

AC C

NPs, we synthesized DOX-SS-DOX (2) (DSSD-2), DOX-SS-DOX (3) (DSSD-3) and DOX-SS-DOX (4) (DSSD-4), in which the linkers were composed of two, three and four carbon atoms between disulfide bond and DOX, conjugating with DOX via its amino group. (2) To investigate the effect of linkage sites of linkers on self-assembling and anti-tumor efficacy of the NPs, we synthesized DOX-SS-DOX (OH) (DSSD-OH), in which the linker molecule was connected to the position C14-OH of DOX by ester bond. The other part was identical with DSSD-3. (3) To investigate the importance of different types of linker, we synthesized DOX-CC-DOX (3) (DCCD-3), same as DSSD-3 except disuccinimidyl suberate (DSS) with non-reductive alkane bonds (“C-C”) instead of dithiobis (succinimidyl propionate)

ACCEPTED MANUSCRIPT (DSP) with disulfide bond (“S-S”). DOX-S-DOX (DSD-OH) was also synthesized with redox dual-responsive single thioether bond [14], which differs from DSSD-OH. We studied the self-assembling characteristics of these DDCs into nanoparticles, analyzed the in vitro and in vivo anti-tumor efficacy and the mechanism of differences by detecting the cellular uptake, co-localization with lysosomes and release free DOX

PT

from prodrug NPs. The results indicated that the length of linker could influence self-assembling process, and linker types and linkage site could affect anti-tumor

AC C

EP T

ED

MA

NU

SC

RI

efficacy.

Scheme 1. The structure of (A) DOX-SS-DOX (3) (DSSD-3), (B) DOX-CC-DOX (3) (DCCD-3), (C) DOX-SS-DOX (2) (DSSD-2), (D) DOX-SS-DOX (4) (DSSD-4), (E) DOX-SS-DOX (OH) (DSSD-OH) and (F) DOX-S-DOX (DSD-OH). 2. Materials and Methods 2.1 Synthesis and characterization of DDCs Six kinds of drug-drug conjugates (DDCs), DSSD-3, DCCD-3, DSSD-2, DSSD-4, DSSD-OH and

DSD-OH,

were

synthesized.

All

DDCs

preparation and

ACCEPTED MANUSCRIPT characterization are shown in the Supplementary Materials. 2.2 Preparation of DDC NPs DOX DDC NPs were prepared with nanoprecipitation method [16]. To prepare DSSD-3 and DCCD-3 NPs, DSSD-3 or DCCD-3 was dissolved with PEG2K-DSPE in DMSO and added to water dropwisely under sonication using supersonic cell disruptor (JY92-2D, China) for 10 min (the final concentration of PEG2K -DSPE was

PT

0.0086 μM). To prepare DSSD-4, DSSD-OH and DSD-OH NPs, DSSD-4, DSSD-OH

RI

or DSD-OH was dissolved in DMSO and then added to PEG2K -DSPE aqueous solution dropwise under sonication using supersonic cell disruptor (JY92-2D, China)

SC

for 10 min (the final concentration of PEG2K -DSPE was 0.0086 μM). 2.3 Characterization of DOX DDC NPs

NU

The particle size and zeta potential of different DOX DDC NPs were analyzed by dynamic light scattering analysis using Malvern Zetasizer Nano ZS (Malvern, United

MA

Kingdom). The morphology was observed with transmission electron microscopy (TEM, JEOL, JEM-200CX, Japan). The drug loading and encapsulation efficiency of

ED

DDC NPs were determined with ultrafiltration method (4 mL, 100kDa, Amicon Ultra).

EP T

2.4 The optical characterization of DDCs and their DDC NPs UV spectrum of equivalent of different DOX prodrug (50 μg/mL DOX) was acquired from 200 nm to 600 nm at a speed of 400 nm/min in ultraviolet spectrophotometer

AC C

(HITACHI, UH5300). Various DOX DDC NPs or DMSO solution of DOX DDC (10 μg/mL DOX) were scanned from 500 nm to 700 nm with excitation wavelength 471 nm at a speed of 240 nm/min (HITACHI, F7000). 2.5 In vitro cytotoxicity of DDC NPs The cytotoxicity of DOX DDC NPs, free DOX·HCl and DOX liposomes were evaluated by SRB assay on MCF-7 cells. Briefly, MCF-7 cells were seeded in a 96-well plate at a density of 5000 cells/well and incubated for 24 h. Then they were exposed to different DOX formulations diluted with RPMI 1640 medium (the concentration calculated as DOX). After 48 h incubation, the cells were fixed with 10% cold trichloroacetic acid at 4 °C for 1 h and washed with deionized water. The cells

ACCEPTED MANUSCRIPT were air-dried and then incubated with 0.4% SRB for 30 min, followed by washing with 1% acetic acid and air drying. The cell bound SRB was dissolved by 10 mM Tris and the UV absorbance at 540 nm was determined using a 96-well plate reader (Thermo Scientific, Multiskan FC, USA). 2.6 In vivo antitumor efficacy of DDC NPs Female nu/nu nude mice were subcutaneously injected with MCF-7 cells (5 × 106 ) in

PT

the right armpit. When the tumor volume reached 50−100 mm3 , mice were randomly

RI

divided into seven groups of PBS (control), DOX liposomes (the preparation method was described in Supplementary Materials), DSSD-3 NPs, DCCD-3 NPs, DSSD-4

SC

NPs, DSSD-OH NPs and DSD-OH NPs. The formulations were administrated at the dosage of 2 mg/kg DOX via tail vein injection every 2 days for 4 times and then every

NU

4 days for 3 times. Tumor volume and body weight were recorded every 2 days since the first administration. Tumor volumes were calculated according to the equation

MA

(width2 × length)/2. Mice were sacrificed at the thirtieth days and various organs were harvested. Heart, liver, spleen, lung, and kidney and tumor tissues were histologically

ED

evaluated with hematoxylin and eosin (H&E). The apoptosis of tumor cells was analyzed via the terminal deoxynucleotide transferase (TdT)- mediated dUTP-biotin

EP T

nickend labeling (TUNEL) assay.

2.7 The anti-tumor mechanism analysis of these DDC NPs 2.7.1 The reduction triggered cytotoxicity of DOX DDC NPs

AC C

To study the reduction responsive cytotoxicity of DDC NPs, MCF-7 cells was pre-incubated with GSH-OEt for 2 h and then DDC NPs (10 μg/mL calculated as DOX) was added [18]. The following procedures were the same as In vitro cytotoxicity of DDC NPs. 2.7.2 In vitro cellular uptake of DDC NPs For flow cytometry study, MCF-7 cells (2.5 × 105 cells/mL) were seeded on a 12-well plate and cultured at 37 °C for 24 h. Then 20 μg/mL of free DOX·HCl, DOX liposomes, and the DDC NPs were added to incubate with cells for 3 h. After incubation, cells were washed with cold PBS and collected by trypsinization. Then the cells were washed three times and analyzed by FACScan flow cytometer at the

ACCEPTED MANUSCRIPT wavelength of Ex /Em of 488 nm (Becton Dickinson, San Jose, CA, USA). Confocal microscopy analysis was also performed to investigate the internalization and intracellular distribution of different DOX DDC NPs. Briefly, MCF-7 cells seeded on cell culture dishes with glass bottoms were treated with 20 μg/mL of DOX·HCl, DOX liposomes, and the DDC NPs for 3 h and 5 μg/mL for 48 h, respectively. After

PT

washing with cold PBS, the cells were fixed with 4% paraformaldehyde and stained with Hoechst 33258. Then the samples were analyzed using a confocal laser scanning

RI

microscope (CLSM, LEICA, TCS SP5, Germany). 2.7.3 Endocytosis pathway investigation of DDC NPs

SC

The endocytosis inhibitors and their concentrations are shown in Table S1 whose cytotoxicity was determined by LDH assay as previously described [17]. For flow

NU

cytometry analysis, MCF-7 cells were seeded in a 12-well plate and incubated for 24 h. After pre-incubation with serum- free medium solutions of inhibitors for 0.5 h, the

MA

cells were treated with 20 μg/mL of free DOX-SH, DSSD-4 NPs, DSSD-OH NPs, and DSD-OH NPs at 37 °C with inhibitors of the same concentration for another 1 h. The

ED

uptake of various DDC NPs was also explored at 4 °C for 1 h. The following steps were the same as In vitro cellular uptake of DDC NPs.

EP T

2.7.4 Co-localization analysis between lysosomes and DDC NPs MCF-7 cells seeded on cell culture dishes and 20 μg/mL of DOX-SH, DSSD-4 NPs, DSSD-OH NPs, and DSD-OH NPs were added to incubate for 3 h. The NP solutions

AC C

were removed and the cells were incubated with LysoTracker® for 0.5 h. Co-localization between NPs and lysosomes was studied by CLSM. 2.7.5 DOX release in the cells MCF-7 cells were seeded on a 12-well plate at the density of 2.5 × 105 cells/well. After the confluency reached 80%, the cells were incubated with 5 μg/mL of DSSD-OH NPs, DSSD-4 NPs and DSD-OH NPs for 0.5, 1, 3, 6, 12, 24, 36 h, respectively. The cells were collected by trypsinization, washed twice with cold PBS and suspended in 250 μL ultrapure water. Then the cell samples were quickly frozen in liquid nitrogen and thawed at 37 °C for three cycles, and then ultrasonicated for 5 min to dissociate cells followed by immediate extraction [19]. Cellular concentrations

ACCEPTED MANUSCRIPT of prodrugs and their metabolites in MCF-7 cells were analyzed by LC–MS/MS (Agilent Technologies, USA). Protein concentrations of the samples were tested by Enhanced BCA Protein Assay Kit (Beyo time Biotechnology, China). Drug concentrations were calculated as: Drug concentration per unit protein = Drug Concentration/ Protein Concentration.

PT

2.8 Statistical Analysis. Data were expressed as mean ± SD and evaluated by one-way ANOVA analysis in

3. Results

SC

3.1 Preparation and characterization of DDC NPs

RI

SPSS. The differences were considered statistically significant if p < 0.05.

Different DOX DDCs were synthesized according to the synthesis route shown in

NU

Scheme S1. All these DDCs were achieved by one-step reaction. To obtain DSSD-OH and DSD-OH, the C14-OH esterification product, catalyst DMAP was used with

MA

anhydrous and oxygen-free condition and pH was adjusted to weak alkalescent. The results of ESI/MS and 1 H-NMR are shown in supplementary materials (Figure S1-7).

ED

Then DDC NPs were prepared with nanoprecipitation method. In the preparation process, DMSO was used and the amount of DMSO was controlled less than 0.5%,

EP T

which was practicable for intravenous injection [20]. As shown in Table S2, the drug loading (calculated according to DOX) were all above 60% and the encapsulation efficiency were all higher than 80%.

AC C

TEM results showed that all these NPs were in spherical shape (Figure 1A-E). Mean diameters, polydispersity index (PDI), size distribution curves and zeta potential of these NPs are shown in Figure 1F and Table 1. DSSD-3 NPs had the smallest size of 75 nm while DSD-OH NPs had the largest of 177 nm. All the DDC NPs were in negative charge except the neutral DSD-OH NPs. The absence of DSSD-2 was due to no self-assembling into nanoparticles despite a few methods were tried.

PT

ACCEPTED MANUSCRIPT

SC

RI

Figure 1. (A-E) The morphology of DSSD-3 NPs, DCCD-3 NPs, DSSD-4 NPs, DSSD-OH NPs and DSD-OH NPs observed by TEM. Scale bars=200 nm. (F) Particle size distributions of the DDC NPs determined by DLS.

Table 1. The particle size, PDI, zeta potential and fluorescence quench ratio

Size (nm) 75.21±1.44

DCCD-3 NPs

110.47±1.33

DSSD-4 NPs

106.83±0.51

DSSD-OH NPs

145.91±9.91

EP T

DSD-OH NPs

177.37±19.54

Quench

Zeta Potential (mV) Raito

0.23±0.024

-19.9

5.12

0.28±0.041

-17.8

9.59

0.21±0.004

-13.5

4.63

0.28±0.022

-9.04

2.18

0.42±0.058

0.00183

1.80

ED

DSSD-3 NPs

PDI

MA

Formulations

NU

(DDC/DDC NP) of DDC NPs

AC C

3.2 The optical characterization of DDCs and their DDC NPs

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

ED

Figure 2. (A) The fluorescence spectrum of different DDCs dissolved in DMSO at a DOX concentration of 10 μg/mL. (B) The fluorescence spectrum of different DDC NPs solution in water at a DOX concentration of 10 μg/mL.

EP T

The UV-spectrum analysis of DDCs were shown in Figure S8. Figure 2A showed the fluorescence spectra of DDCs in DMSO. Length and types of the linkers did not obviously alter the fluorescence spectrum, while peak wavelength shift could be

AC C

observed (from 550 nm to 575 nm) when DOX was modified by C-14 hydroxyl and amino group respectively. Nevertheless, the fluorescence intensity of NPs dropped apparently after preparation into nanoparticles. The quench ratio of NPs at 595 nm are shown in Table 1. DCCD-3 NPs demonstrated the strongest quench ratio, then amino group modification DDC NPs followed. While after C-14 OH modified, DSSD-OH NPs and DSD-OH NPs showed the weak quench. 3.3 In vitro cytotoxicity The cell viability of MCF-7 cells after treatment with DDC NPs, free DOX·HCl and DOX liposomes at a series of DOX concentrations is shown in Figure 3 and Figure S9. All the groups exhibited cytotoxicity in a dose-dependent manner and free DOX

ACCEPTED MANUSCRIPT showed the smallest IC 50 (0.87μg/mL). Among the NPs, the IC50 of DSSD-OH NPs was the lowest, while DCCD-3 NPs exhibited the weakest cytotoxicity (because the inhibition rate of DCCD-3 NPs against MCF-7 cells was less than 50%, the IC 50 was not calculated). No significant difference was observed although the cell viability was lower when treated with DSSD-4 NPs than DSSD-3 NPs. Despite both connected by

MA

NU

SC

RI

PT

ester bond, DSSD-OH NPs showed higher cytotoxicity than DSD-OH NPs.

EP T

ED

Figure 3. (A) Cytotoxicity of different DOX DDC NPs against MCF-7 cells after 48 h incubation by SRB assay (n=6). (B) The IC 50 of different DDC NPs, ***P < 0.001, ****P < 0.0001. Because the inhibition rate of DCCD-3 NPs against MCF-7 cells was less than 50%, the IC 50 was not calculated.

AC C

3.4 In vivo anti-tumor efficacy

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC C

EP T

ED

Figure 4. In vivo antitumor efficacy of PBS, DOX liposomes and various DDC NPs of MCF-7 tumor bearing nu/nu mice. (A) Tumor growth curves of mice after treatment with different DOX formulations. (n=6) *P < 0.05 vs PBS. **P < 0.01 vs PBS. #P < 0.05 vs DCCD-3 NPs. ## P < 0.01 vs DCCD-3 NPs. (B) The tumor photoprints of different treatment groups after tumors were excised at the end of the test. (n=6) (C) Cell apoptosis detected by TUNEL assay. Although the brightness of DOX liposomes looks lower than that of other groups, its accumulated fluorescence intensity is the greatest because of its uniform intracellular distribution (Figure S10). Broken DNA strands were stained by FITC (green); Nuclei were labeled with Hoechst 33258 (blue); Apoptotic cells were indicated by the merging of these two labels.

In vivo anti-tumor efficacy of these NPs was evaluated using female nu/nu nude mice bearing MCF-7 tumor xenografts. As Figure 4A and 4B shows, tumor volume was slightly reduced when treated with DCCD-3 NPs, and higher anti- tumor efficacy was observed with DSSD-3 NPs, which was accordant with the cytotoxicity result. Obvious difference was not observed when treated with DSSD-3 and DSSD-4 NPs, both of which were modified with amino group. Nevertheless, the hydroxyl group modified DSD-OH NPs demonstrated better therapeutic effect than DSSD-OH NPs

ACCEPTED MANUSCRIPT with the most significant anti-tumor efficacy. According to TUNEL analysis (Figure 4C and Figure S10), all groups showed extensive region of TUNEL positive cells except DCCD-3, indicating that most NPs induced tumor apoptosis. From the H&E results in Figure S11, we could figure out that no obvious changes were observed in different DOX DDC NPs, which exhibited good biosafety.

PT

3.5 Anti-tumor mechanism analysis of DDC NPs According to the results mentioned above, the anti-tumor efficiency was closely

RI

dependent on the linker type and modification, while the effect of linker length was negligible. Further analysis of endocytosis process and intracellular fate was

SC

performed to clarify the reasons for the phenomenon.

3.5.1 The reduction triggered cytotoxicity analysis of DCC NPs

NU

GSH- mediated cytotoxicity assay was conducted to study the function of disulfide bond in cytotoxicity of DDC NPs. GSH-OEt was reported to be capable to penetrate

MA

cell membrane and rapidly increase the intracellular GSH concentration through ethyl ester hydrolysis process in cytoplasm [18]. As shown in Figure S12, the cell viability

ED

of these DDC NPs significantly decreased (by 7.14%-18.07%) after pre- incubated with GSH-OEt, indicating the reduction responsive cytotoxicity existed. However,

EP T

DCCD-3 NPs did not show reduction triggered cytotoxicity (p > 0.05). 3.5.2 In vitro cellular uptake of NPs Cellular uptake analysis of the NPs was performed in MCF-7 cells with flow

AC C

cytometry. As the fluorescence intensity of DDC NPs differed a lot from DOX at the same concentration, the original intracellular fluorescence intensity (Figure 5A) was further normalized according to the methods mentioned above. As shown in Figure 5B, DSSD-4 NPs showed the most cellular uptake while DSD-OH NPs revealed the least. It was understandable that the results of confocal microscopy were accordant with the original flow cytometry related to the fluorescence intensity of themselves (Figure 5C). Different from free DOX·HCl, the red fluorescence of nanoparticles showed dispersed distribution, mostly accumulated in cytoplasm and little into nucleus after incubation for 3 h, suggesting that the nanoparticles were internalized in intact form.

ACCEPTED MANUSCRIPT After incubation for 48 h, as shown in Figure 5D, the internalization into nucleus increased compared with that at 3 h. To accurately locate the position of fluorescence in the cells, DSSD-4 NPs, DSSD-OH NPs and DSD-OH NPs were selected and incubated with the cells treated for 3 h. Then the NPs removed and the cells were observed via CLSM after incubation with serum- free medium for another 24 h. The

PT

confocal images were shown in Figure S13. We found the fluorescence in nucleus were more obvious than before with the time prolonging, which indicated that more

AC C

EP T

ED

MA

NU

SC

RI

DOX was released and entered nucleus.

Figure 5. (A) Cellular uptake in MCF-7 cells after incubation with different DDC NPs and controls for 3 h by flow cytometry; (B) The normalized process by flow cytometry, **P < 0.01, ****P < 0.0001; (C) Confocal images of MCF-7 cells treated by various DDC NPs after incubation for 3 h or 48 h at 37 °C. Red represents the different NPs and blue represents the nucleus stained by Hoechst 33258. Scale bars are 25 µm.

ACCEPTED MANUSCRIPT

AC C

EP T

ED

MA

NU

SC

RI

PT

3.5.3 Investigation of Endocytosis Pathway.

Figure 6. Cellular uptake in MCF-7 cells treated by (A) DSSD-OH NPs , (B) DSD-OH NPs, (C) DSSD-4 NPs (D) DOX-SH solutions after pre- incubation with different inhibitors for 0.5 h at 37 °C by flow cytometry.,*P<0.05, **P < 0.01, ***P < 0.001. ****P < 0.0001 vs DSSD-3 NPs. (E) was the confocal laser scanning micrographs of MCF-7 cells after incubation with DSSD-OH NPs, DSD-OH NPs, DSSD-4 NPs and DOX-SH solutions for 3 h and treated with lysosome tracker for 0.5 h. Red represents different DOX NPs and green represents lysosomes. Scale bars are 25 µm. DSSD-4 NPs, DSSD-OH NPs, and DSD-OH NPs, were selected to further study as their higher anti-cancer efficacy and cellular uptake. As shown in Figure 6A-C, when incubation temperature was lowered to 4 °C, the uptake ratio of different NPs was

ACCEPTED MANUSCRIPT reduced by 50%, indicating that their uptake was an energy-dependent process, which differed from free DOX-SH whose uptake was not affected by low temperature. CPZ and hypertonic sucrose are clathrin- mediated endocytosis inhibitors. CPZ usually inhibits the clathrin- mediated endocytosis through disruption of clathrin and the AP2 complex on the cell surface [21]. Hypertonic sucrose inhibits the clathrin- mediated

PT

endocytosis by blocking the clathrin-coated pit formation. MβCD and filipin are known as the inhibitors of caveolae/lipid raft- mediated endocytosis [22]. As Figure

RI

6A-C showed, hypertonic sucrose and MβCD obviously decreased the uptake, which meant that both clathrin and caveolae participated in the endocytosis. CPZ

SC

participated little in the process, which meant that the endocytosis might be weakly related to the AP2 complex on the cell membrane. As for free DOX-SH, EIPA

NU

exhibited significant inhibition of endocytosis, with almost 40% inhibition ratio, while CPZ, MβCD and Sucrose demonstrated only 15% (Figure 6D). Therefore, the

MA

endocytosis of DOX-SH was mediated mostly by macropinocytosis. These results showed that the endocytosis pathway of DDC NPs was different from DOX-SH

ED

despite some similarity.

3.5.4 Co-localization analysis of DDC NPs and lysosomes

EP T

As Figure 6E & S14 shows, at 3 h, DSSD-OH NPs, DSD-OH NPs and DSSD-4 NPs were mostly accumulated in lysosomes with co- localization coefficient of 0.70, 0.74 and 0.66 respectively. This might be owing to their similar intracellular destiny.

AC C

Nevertheless, the accumulation of free DOX-SH was significantly less in lysosomes with co-localization coefficient 0.38. 3.5.5 DOX release in the cells Before the formal metabolism experiment, we first evaluated the methodology of LC– MS/MS for determining the prodrug DSSD-4, DSSD-OH, DSD-OH and their possible metabolite DOX and DOX-SH. Mass spectra and proposed fragmentation patterns for DSSD-4, DSD-OH, DOX, DOX-SH are shown in Figure 7A. The linear concentration range could satisfy the analysis in the following cellular metabolism studies. While DSSD-OH, because of its poor solubility in all our gettable solvent, it was not determined.

ACCEPTED MANUSCRIPT As shown in Figure 7B and C, all the three nanoparticles could metabolize into DOX. DOX was continuously released in 12 h and gradually reached a plateau in 36 h. In the three NPs, the amount of DOX produced at 12, 24, 36 h sequenced as: DSSD-OH> DSSD-4> DSD-OH, which was accordant with results of cell cytotoxicity (Figure 3). Intracellular DSSD-4 NPs was at a higher ratio than DSD-OH NPs, which was

PT

consistent with the results of Flow Cytometry Analysis. Compared with the amide bond conjugated DSSD-4 NPs, the ester bond conjugated DSD-OH NPs showed

RI

higher release ratio (Figure S15). DOX-SH was not detected in cells after DSSD-4

AC C

EP T

ED

MA

NU

DOX-SH were all below limit of detection.

SC

NPs incubation (Figure 7D). It was considered that the released concentrations of

Figure 7. Metabolism of DDC NPs in MCF-7 cells. Cells were harvested at 0.5, 1, 3, 6, 12, 24 and 36 h after drug administration. (A) Representative mass spectra and proposed fragmentation patterns for DSSD-4, DSD-OH, DOX, and DOX-SH. (B) DOX concentrations per unit protein in MCF-7 cells treated with DSSD-OH NPs, DSSD-4 NPs and DSD-OH NPs (at a DOX concentration of 5 μg/mL). (C) DSSD-4 and DSD-OH concentrations per unit protein in MCF-7 cells treated with DSSD-4 NPs and DSD-OH NPs (at a DOX concentration of 5 μg/mL). (D) Chromatogram of DSSD-4 NPs and its metabolite in MCF-7 cells.

4. Discussion Six types of DOX DDCs were synthesized and five of which could successfully self-assemble into

nanoparticles except DSSD-2. In the NPs preparation,

ACCEPTED MANUSCRIPT PEG2K-DSPE was added as a stabilizer to improve the physical stability of the nanoparticles and could achieve long circulation of the NPs after injected into veins. π−π interaction between the huge amount of benzene rings in DOX molecules might account for the formation of NPs. The fluorescence quench phenomenon (Figure 2A and B) further proved our hypothesis. As it is reported, when molecules aggregated,

PT

fluorescence quench usually happened because of the π-π stacking interactions of the molecules in H2 O [15, 23, 24].

RI

During self-assembling process, linker length played an important role. Compared with DSSD-3 and DSSD-4, the shorter linker of DSSD-2 was less flexible, which

SC

hindered its self-assembling. In addition, the stability of NPs was greatly influenced by linker types. As shown in Figure S16, the size of DSSD-3 NPs, DSSD-4 NPs and

NU

DSSD-OH NPs was almost constant in 30 days, while DCCD-3 NPs and DSD-OH NPs demonstrated significant precipitation. In plasma, the disulfide bond linked

MA

nanoparticles also showed better stability than the other two groups (Figure S17). That is to say the disulfide bond were beneficial to keep the stability of the NPs. It

ED

might be explained that S−S bonds showed a distinct preference for dihedral angles approaching 90°, which might be essential for balancing intermolecular forces and

EP T

establishing a favorable conformation [16], while the alkane bonds (“C-C”) and single thioether bond lacked flexibility. Both ester bond and amide bond modification were favorable to self- assembling, but

AC C

amide bond linked particles showed higher stability (Figure S16), less particle size (Table 1) and higher fluorescence quenching ratio (Table 1). It was proposed that the formation of loose aggregates (weak aggregates with large intermolecular distances and insignificant intermolecular energy transfer) limited quenching efficiency because the available surface area of the molecule was reduced upon weak aggregation [25]. In this way, the ester bond connected DSSD-OH NPs and DSD-OH NPs showed less stability than amide bond connected DDC NPs. Distinction of physicochemical properties of different linker types led to the difference in therapeutic effect. DCCD-3 NPs showed poorer cytotoxicity and anti-tumor efficacy than DSSD-3 as carbon-carbon bond was non-reducible and DOX

ACCEPTED MANUSCRIPT release was hindered (Figure S12). The result of TUNEL analysis was consistent as DCCD-3 NPs exhibited almost no cellular apoptosis (Figure 4C). However, DSSD-4 NPs and DSSD-3 NPs showed similar anti-tumor efficacy in vitro and in vivo (no significant difference), even the length of linkers differed. It was reasonable when the higher endocytosis amount of DSSD-4 was taken into consideration. Despite larger

PT

intracellular accumulation, the long alkyl chain of DSSD-4 NPs was more hydrophobic (as shown in Figure S18, the retention time of DSSD-4 was longer than

RI

that of the DSSD-3 under the same elution conditions), which led to difficulty in attracting enough water molecules to enhance the cleavage of amide bond. As a result,

SC

DOX released from both DSSD-4 NPs and DSSD-3 NPs group were similar, supporting our hypothesis (Figure S19).

NU

Compared with DSSD-3 NPs and DSSD-4 NPs, DSSD-OH NPs showed higher cytotoxicity and anti-tumor efficacy. It has been reported that the anti-tumor efficacy

MA

could be maintained when the aminoglycoside of doxorubicin was exposed, accounting for the high activity [26] of DSSD-OH. Besides, ester bond is more likely

ED

to hydrolyze than amide bond for DOX release [27], which was also confirmed via CLSM and LC-MS/MS analysis that more free DOX were produced when incubated

EP T

with DSSD-OH NPs (Figure 7B and Figure S13,15). Besides, DSSD-OH NPs exhibited higher cytotoxicity than DSD-OH NPs (P <0.001). This was reasonable since the cellular uptake of DSSD-OH NPs was 1.6 fold that of

AC C

DSD-OH NPs (Figure 5B), whereas the free DOX released from DSSD-OH NPs was 4.6, 4.4, 4.0, 3.2 and 3.1 fold that from DSD-OH NPs in 3 h, 6 h, 12 h, 24 h, 36 h (Figure 7B), which indicated that the ester bond in DSSD-OH NPs was more susceptible to be hydrolyzed than that in DSD-OH NPs. As disulfide bond was more hydrophilic than thioether bond, DSSD-OH was more readily for attacking by water molecules [28], which led to quicker release of DOX. On the contrary, single thioether was more hydrophobic, and water was not attracted unless oxidized into hydrophilic sulfone [14]. However, DSSD-OH and DSD-OH demonstrated similar anti-tumor efficiency in vivo. We speculated that the complexity of the internal environment might enhance the

ACCEPTED MANUSCRIPT oxidation process of DSD-OH NPs and accelerate the alteration from hydrophobicity to hydrophilicity, which was further supported by the sensitivity of DSD-OH to ROS (Figure S20). In this way, more DOX could be released from DSD-OH NPs in vivo and developed efficacy, which resulted in the similar anti-tumor efficacy with DSSD-OH NPs.

PT

It was once considered that the disulfide bond would be cleaved in the reductive environment to generate DOX-SH [29] and then metabolized into DOX (Figure S21).

RI

However, when incubated with cells, DOX-SH exhibited poor cytotoxicity (Figure S22). According to our previous study, DOX-SH and their prodrug (pHLIP-SS-DOX)

SC

were mostly internalized via passive diffusion or cell membrane insertion to cytoplasm [30], while DSSD-3 NPs entered cells through endocytosis and

NU

demonstrated significantly improved therapeutic efficacy. Thus, we hypothesized the uptake pathway might play the role and affect the metabolism mechanism.

MA

To clarify the anti-tumor mechanism, we further explored the endocytosis and intracellular transport of DDCs NPs and DOX-SH. The result showed that all the

ED

DDC NPs entered cells by clathrin and caveolae medicated endocytosis (Figure 6 A-C). Furthermore, highly co- localization between NPs and lysosomes was observed

EP T

(Figure 6E). On the contrary, DOX-SH were internalized mainly by micropinocytosis (Figure 6D) and rarely entered lysosomes (Figure 6E), which might be owing to the aggregation of poorly soluble DOX-SH into crude particles under our experiment

AC C

condition (Figure S23). As lysosomes were low pH (5.0-6.0) environment with abundant enzymes [27], ester or amide bond in DDC NPs might be cleaved by acid environment or various kinds of enzymes, such as amidases [31], which led to release of free DOX. Then DOX could be dispersed into nucleus, the therapeutic site. The DOX-SH, by contrast, were mostly accumulated in cytoplasm and not able to be metabolized into DOX, which could explain the lower cytotoxicity of DOX-SH compared with DDC NPs. Therefore, the lysosome environment was essential for anti-tumor efficacy of DDC NPs. To further confirm this, we investigated the metabolism process of DDC NPs in cells. When DSSD-4 was incubated in cell lysate, where amidase and esterase did not

ACCEPTED MANUSCRIPT work efficiently but reductive substances such as GSH were more active, DOX-SH rather than DOX were detected. By contrast, DSD-OH prodrug showed good stability, nearly 100% remained after 6 h incubation with cell lysate (Figure S24). Thus, the detected DOX-SH was produced by the S-S bond cleavage capacity in cytoplasm. However, according to Figure 7D, intracellular DOX existed when DDC NPs were with

MCF-7

cells,

while

no

DOX-SH

and

its

metabolite,

PT

incubated

2-oxotetrahydrothiophene (Figure S21), were detected, indicating the disulfide bond

RI

was not cleaved [32-34]. This result illustrates that the DDC NPs were not exposed to cytoplasm environment but directly released DOX in lysosomes, which accounted for

SC

the anti-tumor efficacy and was consistent with the transport pathway study. Besides, cleavage of amide or ester bonds was the metabolism mechanism and S-S bond

NU

attracted enough water molecules to enhance the hydrolysis rather than be fractured

AC C

EP T

ED

MA

directly.

Scheme 2. Schematic illustrations of cellular pathways of DDC NPs and hypothetical release mechanism of DOX.

We concluded the mechanisms of DOX release from DDC NPs as shown in Scheme 2:

ACCEPTED MANUSCRIPT The DDC nanoparticles were internalized by MCF-7 cells through clathrin and caveolae medicated endocytosis and further transported to lysosomes. In lysosomes, different kinds of DDC NPs released DOX via different mechanisms: (1) In the amide bond and disulfide-bond connected DDC NPs, such as DSSD-3 and DSSD-4 NPs, the amide bond was cleaved by amidases and other related enzymes [31] to release DOX.

PT

(2) In the ester bond and disulfide-bond connected DDC NPs, DSSD-OH NPs, the ester bond was easily to be hydrolyzed because of its low pH environment. The

RI

function of disulfide-bond in both the amide bond and ester bond hydrolysis was to attract enough water molecules to form the intermediates [28] that enhanced the

SC

release of DOX. (3) DSD-OH NPs might show the DOX release in response to ROS, which could take place in three steps: (i) oxidation of the thioether to hydrophilic

NU

sulfone [14]; (ii) hydrolysis of ester bond; (iii) release of active DOX molecule.

MA

5. Conclusion

In this study, six DOX DDC prodrugs, DSSD-3, DCCD-3, DSSD-2, DSSD-4,

ED

DSSD-OH, DSD-OH, were synthesized with different linker types, length or linkage sites. Except for DSSD-2, all these DOX DDC prodrugs could self-assemble into

EP T

nanoparticles via the nanoprecipitation method. The fluorescence quench of nanoparticles indicated the self-assembly mechanism was π−π interaction between the benzene ring of the two DOX molecules. The length of linker was essential for

AC C

self-assembling of NPs, while it showed little influence on the cytotoxicity. Disulfide bond was not only vital for self-assembling and stability of NPs but also beneficial to the cleavage of connected bond in low pH environment by water attacking for anti-tumor efficacy, and C-14 hydroxyl modification was more efficient than the amino group for DDC NPs in therapy. In a word, to design a stimuli-sensitive DOX prodrug nanoparticle, the modified DOX DDC was recommended. The finding in study is significant in clarifying the important role of disulfide bond and C-14 hydroxyl group in DDC NPs, and is meaningful for effective SDC design for therapeutics.

AC C

EP T

ED

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Acknowledgment: The study was supported by Beijing Natural Science Foundation ( Grants NO.

AC C

EP T

ED

MA

NU

SC

RI

PT

7162108).

ACCEPTED MANUSCRIPT Competing Interests:

AC C

EP T

ED

MA

NU

SC

RI

PT

The authors have declared that no competing interest exists.

ACCEPTED MANUSCRIPT References： [1] H. Cui, M.J. Webber, S.I. Stupp, Self-assembly of peptide amphiphiles: from molecules to nanostructures to biomaterials, Biopolymers 94(1) (2010) 1-18. [2] Y. Hu, R. Lin, K. Patel, A.G. Cheetham, C. Kan, H. Cui, Spatiotemporal control of the creation and immolation of peptide assemblies, Coordination Chemistry Reviews

PT

320-321 (2016) 2-17. [3] Y. Liu, Q. Pei, L. Chen, Z. Li, Z. Xie, Reduction-responsive fluorescence off–on

RI

BODIPY–camptothecin conjugates for self-reporting drug release, J. Mater. Chem. B 4(13) (2016) 2332-2337.

SC

[4] X. Zheng, Z. Li, L. Chen, Z. Xie, X. Jing, Self- Assembly of Porphyrin-Paclitaxel

Chem Asian J 11(12) (2016) 1780-4.

NU

Conjugates Into Nanomedicines: Enhanced Cytotoxicity due to Endosomal Escape,

[5] W. Lin, T. Sun, Z. Xie, J. Gu, X. Jing, A dual-responsive nanocapsule via

MA

disulfide- induced self-assembly for therapeutic agent delivery, Chem. Sci. 7(3) (2016) 1846-1852.

ED

[6] P. Huang, D. Wang, Y. Su, W. Huang, Y. Zhou, D. Cui, X. Zhu, D. Yan, Combination of small molecule prodrug and nanodrug delivery: amphiphilic

EP T

drug-drug conjugate for cancer therapy, J Am Chem Soc 136(33) (2014) 11748-56. [7] J. Kopecek, Polymer-drug conjugates: origins, progress to date and future directions, Adv Drug Deliv Rev 65(1) (2013) 49-59.

AC C

[8] J.W. Nichols, Y.H. Bae, EPR: Evidence and fallacy, J Control Release 190 (2014) 451-64.

[9] Y. Wang, P. Huang, M. Hu, W. Huang, X. Zhu, D. Yan, Self- Delivery Nanoparticles of Amphiphilic Methotrexate-Gemcitabine Prodrug for Synergistic Combination Chemotherapy via Effect of Deoxyribonucleotide Poo ls, Bioconjug Chem 27(11) (2016) 2722-2733. [10] Y. Liang, S. Li, X. Wang, B. He, B. He, W. Dai, H. Zhang, X. Wang, Y. Wang, D. Zhou, Q. Zhang, A Nanosystem of Amphiphilic Oligopeptide-Drug Conjugate Actualizing Both alphavbeta3 Targeting and Reduction- Triggered Release for Maytansinoid, Theranostics 7(13) (2017) 3306-3318.

ACCEPTED MANUSCRIPT [11] J. Liu, W. Huang, Y. Pang, P. Huang, X. Zhu, Y. Zhou, D. Yan, Molecular self-assembly of a homopolymer: an alternative to fabricate drug-delivery platforms for cancer therapy, Angew Chem Int Ed Engl 50(39) (2011) 9162-6. [12] H. Wei, R.-X. Zhuo, X.-Z. Zhang, Design and development of polymeric micelles with cleavable links for intracellular drug delivery, Progress in Polymer

PT

Science 38(3-4) (2013) 503-535. [13] Z. Duan, Y. Zhang, H. Zhu, L. Sun, H. Cai, B. Li, Q. Gong, Z. Gu, K. Luo,

RI

Stimuli-Sensitive Biodegradable and Amphiphilic Block Copolymer-Gemcitabine Conjugates Self-Assemble into a Nanoscale Vehicle for Cancer Therapy, ACS Appl

SC

Mater Interfaces 9(4) (2017) 3474-3486.

[14] C. Luo, J. Sun, D. Liu, B. Sun, L. Miao, S. Musetti, J. Li, X. Han, Y. Du, L. Li, L.

by

Single

Thioether-Bridged

NU

Huang, Z. He, Self- Assembled Redox Dual-Responsive Prodrug-Nanosystem Formed Paclitaxel-Fatty

Acid

Conjugate

for

Cancer

MA

Chemotherapy, Nano Lett 16(9) (2016) 5401-8.

[15] Y. Li, X. Xu, X. Zhang, Y. Li, Z. Zhang, Z. Gu, Tumor-Specific Multiple

ED

Stimuli- Activated Dendrimeric Nanoassemblies with Metabolic Blockade Surmount Chemotherapy Resistance, ACS Nano 11(1) (2017) 416-429.

EP T

[16] Y. Wang, D. Liu, Q. Zheng, Q. Zhao, H. Zhang, Y. Ma, J.K. Fallon, Q. Fu, M.T. Haynes, G. Lin, R. Zhang, D. Wang, X. Yang, L. Zhao, Z. He, F. Liu, Disulfide bond bridge insertion turns hydrophobic anticancer prodrugs into self-assembled

AC C

nanomedicines, Nano Lett 14(10) (2014) 5577-83. [17] Q. Song, X. Wang, Y. Wang, Y. Liang, Y. Zhou, X. Song, B. He, H. Zhang, W. Dai, X. Wang, Q. Zhang, Reduction Responsive Self- Assembled Nanoparticles Based on Disulfide-Linked Drug-Drug Conjugate with High Drug Loading and Antitumor Efficacy, Mol Pharm 13(1) (2016) 190-201. [18] A.N. Koo, H.J. Lee, S.E. Kim, J.H. Chang, C. Park, C. Kim, J.H. Park, S.C. Lee, Disulfide-cross- linked

PEG-poly(amino

acid)s

copolymer

micelles

for

glutathione- mediated intracellular drug delivery, Chem Commun (Camb) (48) (2008) 6570-2. [19] N. Zheng, X. Wang, Y. Wang, G. Xu, H. Zhang, W. Dai, B. He, Q. Zhang, J. Ji, X.

ACCEPTED MANUSCRIPT Wang, A sensitive liquid chromatography/electrospray tandem mass spectroscopy method for simultaneous quantification of a disulfide bond doxorubicin conjugation prodrug and activated doxorubicin: Application to cellular pharmacokinetic and catabolism studies, J Chromatogr B Analyt Technol Biomed Life Sci 1065-1066 (2017) 96-103.

PT

[20] Thackaberry EA, Wang X, Schweiger M, Messick K, Valle N, Dean B, Sambrone A, Bowman T, Xie M, Solvent-based formulations for intravenous mouse

RI

pharmacokinetic studies: tolerability and recommended solvent dose limits. Xenobiotica; the fate of foreign compounds in biological systems, 2014,

SC

44(3):235-41.

[21] D. Yu, P. Peng, S.S. Dharap, Y. Wang, M. Mehlig, P. Chandna, H. Zhao, D.

NU

Filpula, K. Yang, V. Borowski, G. Borchard, Z. Zhang, T. Minko, Antitumor activity of poly(ethylene glycol)-camptothecin conjugate: the inhibition of tumor growth in

MA

vivo, J Control Release 110(1) (2005) 90-102.

[22] B. He, Z. Jia, W. Du, C. Yu, Y. Fan, W. Dai, L. Yuan, H. Zhang, X. Wang, J.

ED

Wang, X. Zhang, Q. Zhang, The transport pathways of polymer nanoparticles in MDCK epithelial cells, Biomaterials 34(17) (2013) 4309-26.

EP T

[23] C. Ren, H. Wang, D. Mao, X. Zhang, Q. Fengzhao, Y. Shi, D. Ding, D. Kong, L. Wang, Z. Yang, When molecular probes meet self-assembly: an enhanced quenching effect, Angew Chem Int Ed Engl 54(16) (2015) 4823-7.

AC C

[24] B. Fu, J. Huang, D. Bai, Y. Xie, Y. Wang, S. Wang, X. Zhou, Label- free detection of pH based on the i- motif using an aggregation-caused quenching strategy, Chem Commun (Camb) 51(95) (2015) 16960-3. [25] S.W.I. Thomas, G.D. Joly, T.M. Swager, Chemical Sensors Based on Amplifying Fluorescent Conjugated Polymers, Cheminform 38(31) (2007) 1339. [26] V.M. Garsky, P.K. Lumma, D.M. Feng, J. Wai, H.G. Ramjit, M.K. Sardana, A. Oliff, R.E. Jones, D. Defeo-Jones, R.M. Freidinger, The synthesis of a prodrug of doxorubicin designed to provide reduced systemic toxicity and greater target efficacy, Journal of Medicinal Chemistry 44(24) (2011) 4216-4224. [27] R. Soudy, C. Chen, K. Kaur, Novel peptide-doxorubucin conjugates for targeting

ACCEPTED MANUSCRIPT breast cancer cells including the multidrug resistant cells, J Med Chem 56(19) (2013) 7564-73. [28] S. Lv, Z. Tang, D. Zhang, W. Song, M. Li, J. Lin, H. Liu, X. Chen, Well-defined polymer-drug conjugate engineered with redox and pH-sensitive release mechanism for efficient delivery of paclitaxel, J Control Release 194 (2014) 220-7.

PT

[29] Z. Xu, S. Liu, Y. Kang, M. Wang, Glutathione- and pH-responsive nonporous silica prodrug nanoparticles for controlled release and cancer therapy, Nanoscale 7(13)

RI

(2015) 5859-68.

[30] Q. Song, X. Chuan, B. Chen, B. He, H. Zhang, W. Dai, X. Wang, Q. Zhang, A

SC

smart tumor targeting peptide-drug conjugate, pHLIP-SS-DOX: synthesis and cellular uptake on MCF-7 and MCF-7/Adr cells, Drug Deliv 23(5) (2016) 1734-46.

NU

[31] E. Romeo, S. Ponzano, A. Armirotti, M. Summa, F. Bertozzi, G. Garau, T. Bandiera, D. Piomelli, Activity- Based Probe for N-Acylethanolamine Acid Amidase,

MA

ACS Chem Biol 10(9) (2015) 2057-2064.

[32] A.M. Sauer, A. Schlossbauer, N. Ruthardt, V. Cauda, T. Bein, C. Brauchle, Role

ED

of endosomal escape for disulfide-based drug delivery from colloidal mesoporous silica evaluated by live-cell imaging, Nano Lett 10(9) (2010) 3684-91.

EP T

[33] R. Carden, J.L. Dellifraine, Real- time fluorescence detection of protein transduction into live cells, J Am Chem Soc 130(8) (2008) 2398-9. [34] L. Brulisauer, M.A. Gauthier, J.C. Leroux, Disulfide-containing parenteral

AC C

delivery systems and their redox-biological fate, J Control Release 195 (2014) 147-54.

ACCEPTED MANUSCRIPT Scheme legends: Scheme 1. The structure of (A) DOX-SS-DOX (3) (DSSD-3), (B) DOX-CC-DOX (3) (DCCD-3), (C) DOX-SS-DOX (2) (DSSD-2), (D) DOX-SS-DOX (4) (DSSD-4), (E) DOX-SS-DOX (OH) (DSSD-OH) and (F) DOX-S-DOX (DSD-OH).

AC C

EP T

ED

MA

NU

SC

RI

PT

Scheme 2. Schematic illustrations of cellular pathways of DDC NPs and hypothetical release mechanism of DOX.

ACCEPTED MANUSCRIPT Figure legends: Figure 1. (A-E) The morphology of DSSD-3 NPs, DCCD-3 NPs, DSSD-4 NPs, DSSD-OH NPs and DSD-OH NPs observed by TEM. Scale bars=200 nm. (F) Particle size distributions of the DDC NPs determined by DLS.

PT

Figure 2. (A) The fluorescence spectrum of different DDCs dissolved in DMSO at a DOX concentration of 10 μg/mL. (B) The fluorescence spectrum of different DDC NPs solution in water at a DOX concentration of 10 μg/mL.

SC

RI

Figure 3. (A) Cytotoxicity of different DOX DDC NPs against MCF-7 cells after 48 h incubation by SRB assay (n=6). (B) The IC 50 of different DDC NPs, ***P < 0.001, ****P < 0.0001. Because the inhibition rate of DCCD-3 NPs against MCF-7 cells was less than 50%, the IC50 was not calculated.

ED

MA

NU

Figure 4. In vivo antitumor efficacy of PBS, DOX liposomes and various DDC NPs of MCF-7 tumor bearing nu/nu mice. (A) Tumor growth curves of mice after treatment with different DOX formulations. (n=6) *P < 0.05 vs PBS. **P < 0.01 vs PBS. #P < 0.05 vs DCCD-3 NPs. ## P < 0.01 vs DCCD-3 NPs. (B) The tumor photoprints of different treatment groups after tumors were excised at the end of the test. (n=6) (C) Cell apoptosis detected by TUNEL assay. Although the brightness of DOX liposomes looks lower than that of other groups, its accumulated fluorescence intensity is the greatest because of its uniform intracellular distribution (Figure S10). Broken DNA strands were stained by FITC (green); Nuclei were labeled with Hoechst 33258 (blue); Apoptotic cells were indicated by the merging of these two labels.

AC C

EP T

Figure 5. (A) Cellular uptake in MCF-7 cells after incubation with different DDC NPs and controls for 3 h by flow cytometry; (B) The normalized process by flow cytometry, **P < 0.01, ****P < 0.0001; (C) Confocal images of MCF-7 cells treated by various DDC NPs after incubation for 3 h or 48 h at 37 °C. Red represents the different NPs and blue represents the nucleus stained by Hoechst 33258. Scale bars are 25 µm. Figure 6. Cellular uptake in MCF-7 cells treated by (A) DSSD-OH NPs , (B) DSD-OH NPs, (C) DSSD-4 NPs (D) DOX-SH solutions after pre- incubation with different inhibitors for 0.5 h at 37 °C by flow cytometry.,*P<0.05, **P < 0.01, ***P < 0.001. ****P < 0.0001 vs DSSD-3 NPs. (E) was the confocal laser scanning micrographs of MCF-7 cells after incubation with DSSD-OH NPs, DSD-OH NPs, DSSD-4 NPs and DOX-SH solutions for 3 h and treated with lysosome tracker for 0.5 h. Red represents different DOX NPs and green represents lysosomes. Scale bars are 25 µm. Figure 7. Metabolism of DDC NPs in MCF-7 cells. Cells were harvested at 0.5, 1, 3, 6, 12, 24 and 36 h after drug administration. (A) Representative mass spectra and

ACCEPTED MANUSCRIPT

AC C

EP T

ED

MA

NU

SC

RI

PT

proposed fragmentation patterns for DSSD-4, DSD-OH, DOX, and DOX-SH. (B) DOX concentrations per unit protein in MCF-7 cells treated with DSSD-OH NPs, DSSD-4 NPs and DSD-OH NPs (at a DOX concentration of 5 μg/mL). (C) DSSD-4 and DSD-OH concentrations per unit protein in MCF-7 cells treated with DSSD-4 NPs and DSD-OH NPs (at a DOX concentration of 5 μg/mL). (D) Chromatogram of DSSD-4 NPs and its metabolite in MCF-7 cells.

ACCEPTED MANUSCRIPT

AC C

EP T

ED

MA

NU

SC

RI

PT

Graphical Abstract: