Research ArticleHEALTH AND MEDICINE

Trisulfide bond–mediated doxorubicin dimeric prodrug nanoassemblies with high drug loading, high self-assembly stability, and high tumor selectivity

See allHide authors and affiliations

Science Advances  04 Nov 2020:
Vol. 6, no. 45, eabc1725
DOI: 10.1126/sciadv.abc1725

Abstract

Rational design of nanoparticulate drug delivery systems (nano-DDS) for efficient cancer therapy is still a challenge, restricted by poor drug loading, poor stability, and poor tumor selectivity. Here, we report that simple insertion of a trisulfide bond can turn doxorubicin homodimeric prodrugs into self-assembled nanoparticles with three benefits: high drug loading (67.24%, w/w), high self-assembly stability, and high tumor selectivity. Compared with disulfide and thioether bonds, the trisulfide bond effectively promotes the self-assembly ability of doxorubicin homodimeric prodrugs, thereby improving the colloidal stability and in vivo fate of prodrug nanoassemblies. The trisulfide bond also shows higher glutathione sensitivity compared to the conventional disulfide bond, and this sensitivity enables efficient tumor-specific drug release. Therefore, trisulfide bond–bridged prodrug nanoassemblies exhibit high selective cytotoxicity on tumor cells compared with normal cells, notably reducing the systemic toxicity of doxorubicin. Our findings provide new insights into the design of advanced redox-sensitive nano-DDS for cancer therapy.

INTRODUCTION

Nanoparticulate drug delivery systems (nano-DDS) have been widely investigated for cancer therapy during the past decades (1). Unfortunately, only a few products have been successfully translated into clinical application (2, 3). Most antitumor drugs lack affinity with the nanocarriers, leading to low drug loading (usually less than 10%), poor colloidal stability, and premature drug leakage in the systemic circulation (2, 4, 5). In addition, how to achieve tumor-specific drug release is also a challenge for the design of advanced nano-DDS (6, 7). Doxorubicin (DOX) is a first-line chemotherapy drug with potent antitumor activity (8). However, the efficiency of DOX is limited by the serious adverse effects, including cardiotoxicity and myelosuppression (8). Doxil (liposomal DOX), the first nanomedicine approved by the U.S. Food and Drug Administration, significantly reduces the cardiotoxicity of DOX (9). However, the drug loading of Doxil is only 11% even when using remote loading technology, and the heavily used lipid materials cause a potential security risk (9). In addition, the antitumor efficacy of Doxil is hindered by the inefficient release of the active drug at the tumor sites (9, 10).

In response to these problems, prodrug-based self-assembled nanoparticles (NPs) have been developed, integrating the advantages of both prodrug strategies and nanocarriers (11, 12). Compared with the traditional nano-DDS, prodrug nanoassemblies exhibit distinct high drug loading and low excipient-related toxicity because the prodrugs act as both the carriers and the cargos (11). In particular, homodimeric prodrugs, formed by conjugating two drug molecules together, could further improve the drug loading of prodrug nanoassemblies to 60% or higher (1315). In previous studies, homodimeric prodrugs based on DOX, camptothecin (CPT), or SN-38 were developed to construct high–drug loading prodrug nanoassemblies (1416). However, anticancer agents such as DOX/CPT/SN-38 have strong intermolecular forces due to the π-π stacking interaction among the planar aromatic ring (14, 15). As a result, the corresponding homodimeric prodrugs exhibit poor self-assembly ability and tend to precipitate in water. To prevent the formation of large aggregates, the conventional approach introduces “structural defects” to the prodrugs by the insertion of bridging groups (e.g., rotatable σ bond-bearing benzene ring) (15, 17) or flexible linkers (e.g., sulfide bond) (18, 19), which could make the molecular structure less rigid, thereby preventing long range–ordered molecular packing. One of the most used flexible linkers is the disulfide bond (18, 20), with nearly vertical double bond angles and a single dihedral angle, which play an essential role in improving structural flexibility and balancing intermolecular forces during molecule self-assembly. However, even with the insertion of a disulfide bond, the self-assembly ability of the currently developed homodimeric prodrugs is still far from satisfactory. Most of them are unable to self-assemble into stable nanostructures without the help of surfactants. Compared with the disulfide bond, the trisulfide bond contains an extra sulfur atom in the bridge, which is the key linkage in biomolecules with functions ranging from subtle modulation of the conformation and assembly stability in proteins and peptides (21). The trisulfide bond has three sulfur atoms and two sulfur-containing dihedral angles, which perhaps further facilitate the self-assembly of homodimeric prodrugs.

Ideally, we hope that the prodrug nanoassemblies can remain intact in the systemic circulation and normal cells while intelligently releasing the active parent drugs in tumor cells (1, 5). It has been reported that the concentration of glutathione (GSH) in tumor cells is over 1000-fold higher than in blood and much higher than that in normal cells (22, 23). In response to the overexpressed GSH in tumor cells, the disulfide bond has been widely used as the “golden standard” to design redox-responsive DDS (24, 25). The disulfide bond can be reduced to the hydrophilic thiol group by GSH, facilitating the release of parent drugs (25). Compared with the disulfide bond, the trisulfide bond might be more sensitive to GSH because it has three redox reaction sites and higher redox potential. Therefore, we expect the trisulfide bond to be a reduction-supersensitive linkage with ultrahigh tumor selectivity.

On the basis of the above considerations, we attempted to explore the impact of the trisulfide bond on the homodimeric prodrug nanoassemblies. As shown in Fig. 1, three homodimeric DOX prodrugs were synthesized using a thioether bond, disulfide bond, or trisulfide bond as linkages. Simple insertion of a trisulfide bond could transform the DOX homodimeric prodrug into a uniform nanostructure with three highlights: high drug loading (67.24%, w/w), high self-assembly stability, and high tumor selectivity. Compared with the disulfide bond and thioether bond, the trisulfide bond significantly promoted the self-assembly of DOX homodimeric prodrugs, thereby improving the colloidal stability, pharmacokinetics, and tumor accumulation of prodrug nanoassemblies. Furthermore, the trisulfide bond showed higher GSH sensitivity than disulfide bond, and the triggering mechanism was also elucidated in detail. As a result, trisulfide bond–bridged prodrug nanoassemblies exhibited highly selective cytotoxicity on tumor cells as opposed to normal cells, significantly reducing the systemic toxicity of DOX.

Fig. 1 The trisulfide bond–bridged prodrug nanoassemblies for cancer therapy.

Simple insertion of a trisulfide bond–transformed DOX homodimeric prodrugs into self-assembled nanomedicines with three highlights: high drug loading, high self-assembly stability, and high tumor selectivity.

RESULTS

Design and synthesis of DOX homodimeric prodrugs

To explore the potential of the trisulfide bond on self-assembly, DOX was used as a model drug to design homodimeric prodrugs because of its rigid planar pentacyclic aromatic ring structure, which could provide strong intramolecular forces (15). These DOX homodimeric prodrugs were more challenging to self-assemble into stable NPs. The synthesis route of the designed DOX homodimeric prodrugs is shown in fig. S1A. The DOX homodimeric prodrugs were synthesized by conjugating two DOX molecules together via the amine group using 3,3′-thiodipropionic acid, 3,3′-dithiodipropionic acid, or 3,3′-trithiodipropionic acid as linkers. The corresponding prodrugs were named DOX-S-DOX (DSD), DOX-SS-DOX (DSSD), and DOX-SSS-DOX (DSSSD), respectively. The chemical structures of the prodrugs were confirmed by 1H nuclear magnetic resonance spectroscopy (1H NMR) and mass spectrometry (MS) (figs. S2 to S4).

Self-assembly ability and mechanism of prodrugs

The self-assembly ability of these prodrugs in the absence of surfactants was investigated using a simple nanoprecipitation method. DSD or DSSD alone failed to self-assemble into stable NPs in aqueous solution. As shown in fig. S5A, all three prodrugs initially formed a transparent red solution (0.2 mg/ml). However, both DSD NPs and DSSD NPs gradually precipitated with time. Floccules were first observed in DSD NPs, and larger amount of aggregates was formed within 6 hours (fig. S5A). In comparison, DSSSD could spontaneously assemble into spherical NPs with an average diameter of ~150 nm (fig. S5E), and no precipitation was observed within 48 hours (Fig. 2A). Furthermore, DSSSD NPs could still remain clear at a high concentration (0.6 mg/ml), but obvious precipitation was observed in DSD NPs and DSSD NPs even at low concentrations (0.1 mg/ml; fig. S5B). Therefore, DSSSD had much better self-assembly ability than DSD and DSSD. The three prodrugs had the same molecular composition, and the only difference existed in the thioether/disulfide/trisulfide linkages. These findings suggested that subtle changes in the linkages could significantly affect the self-assembly ability of prodrugs, and the trisulfide bond effectively facilitated the self-assembly of DOX homodimeric prodrugs.

Fig. 2 Preparation and characterization of DOX homodimeric prodrug nanoassemblies.

(A) Appearance of non-PEGylated prodrug nanoassemblies (0.2 mg/ml) stored at room temperature for 48 hours. (B) Prodrug nanoassemblies were prepared by self-assembly or co-assembly with DSPE-PEG2K (20%, w/w) using a simple nanoprecipitation method. (C) Appearance of PEGylated prodrug nanoassemblies (1 mg/ml) after storage at 4°C for 24 hours. (D) Particle size distribution and zeta potential of the PEGylated DSSD NPs. Polydispersity index (PDI). (E) Particle size distribution and zeta potential of the PEGylated DSSSD NPs. (F) TEM images of the PEGylated DSSD NPs. (G) TEM images of the PEGylated DSSSD NPs. Free energy changes during the dimerization of the (H) DSD, (I) DSSD, and (J) DSSSD prodrugs in water, respectively. (K) Appearance of PEGylated prodrug nanoassemblies (1 mg/ml) after dialysis and stored at 4°C for 48 hours. S, DSD NPs; SS, DSSD NPs; SSS, DSSSD NPs. Photo credit: Yinxian Yang, Department of Pharmaceutics, College of Pharmacy, Shenyang Pharmaceutical University, Shenyang 110016, China.

It has been reported that the bond angle/dihedral angle of the disulfide bond could improve the structure flexibility of prodrugs, thereby enhancing the stability of prodrug nanoassemblies (18). The most stable conformation would be constructed when the bond/dihedral angle approaches 90° (18, 19). Thus, the bond angles and dihedral angles of thioether/disulfide/trisulfide bonds in the optimized geometry of DOX homodimeric prodrugs were calculated using the quantum chemical method. The bond angles are shown in fig. S6A: ─C─S─C─ (97.534°), ─C─SS─C─ (96.897°/96.350°), and ─C─SSS─C─ (91.908°/92.848°/95.168°). Compared with the thioether/disulfide bond, the trisulfide bond had more sulfur-containing bond angles, and they were closest to 90°. Moreover, both dihedral angles of the trisulfide (83.59°/97.56°) were closer to 90° than disulfide (81.88°). The single thioether bond showed a poor effect on the self-assembly ability of DOX homodimeric prodrugs, as it has only one sulfur-containing bond angle. The disulfide bond could improve the self-assembly ability of the DOX homodimeric prodrug to some extent, but it was still insufficient to balance the strong intermolecular forces to form stable NPs. Therefore, DSSD NPs were dynamically unstable and slowly precipitated during storage, but the precipitation rate was slower than DSD NPs. In comparison, the trisulfide bond had three sulfur-containing bond angles and two dihedral angles, and their angles were closer to 90°. As a result, the trisulfide bond could provide more sufficient structural flexibility for DSSSD to balance intermolecular forces, thereby establishing a favorable conformation during self-assembly.

Furthermore, the free energy of prodrug nanoassemblies was calculated using the quantum chemical method and molecular dynamics simulations (26). From the thermodynamic perspective, the lower the free energy (ΔG < 0), the better is the system stability (27). As shown in Fig. 2 (H to J), the calculated free energy values of the prodrug nanoassemblies were DSSSD NPs (−59.7 kcal/mol) < DSSD NPs (−45.7 kcal/mol) < DSD NPs (−33.8 kcal/mol), agreeing well with the self-assembly stability. These results confirmed that the trisulfide bond significantly improved the structural flexibility of DOX homodimeric prodrugs and helped to establish a stable conformation with low free energies during self-assembly (Fig. 2B).

Characterization of PEGylated prodrug nanoassemblies

A PEGylation strategy could reduce the clearance of NPs by the reticuloendothelial system and improved the colloidal stability of prodrug nanoassemblies (18, 28). As shown in fig. S5F, the non-PEGylated prodrug nanoassemblies showed poor stability in phosphate-buffered saline (PBS) with significantly increased aggregates within 2 hours. Therefore, PEGylated prodrug nanoassemblies were prepared for subsequent experiments by coassembling DOX homodimeric prodrugs with DSPE-PEG2K [1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N [methoxy (polyethyleneglycol)-2000; 20%, w/w; Fig. 2B]. As shown in Fig. 2C and fig. S5C, DSD could still not self-assemble into stable NPs, and obvious precipitates were observed within 24 hours. Therefore, DSD was excluded from subsequent studies. In comparison, DSSD and DSSSD could self-assemble into transparent NP solutions with no precipitates within 24 hours. The residual organic solvents in DSSD NPs and DSSSD NPs were removed by dialysis. The final NPs showed hydrodynamic diameters of ~80 nm and a negative surface charge of ~18 mV (Fig. 2, D and E). Transmission electron microscopy (TEM) images also revealed the successful fabrication of spheroidal NPs (Fig. 2, F and G). Notably, the PEGylated homodimeric prodrug nanoassemblies had an impressive ultrahigh drug-loading capacity (68.94% for DSSD NPs and 67.24% for DSSSD NPs), which could help to reduce the excipient-associated side effects.

However, after dialysis, DSSD NPs were relatively unstable during storage. Large precipitates were observed after stored at 4°C for 48 hours (Fig. 2K), and DSSD NPs stored within 12 hours were used for further studies (fig. S5C). In comparison, DSSSD NPs were extremely stable under the same conditions for 7 days. The stability of DSSD NPs and DSSSD NPs was further evaluated by dilution in PBS (pH 7.4) supplemented with 10% fetal bovine serum (FBS). As shown in fig. S5D, DSSSD NPs showed good colloidal stability, while DSSD NPs were destroyed and precipitated during dilution. These results were highly consistent with the self-assembly ability of prodrugs, supporting the assertion that simple insertion of a trisulfide bond could promote the self-assembly process and enhance the stability of prodrug nanoassemblies (Fig. 2B).

In vitro GSH-triggered drug release and mechanism

Next, we investigated the in vitro drug release of DSSD NPs and DSSSD NPs. As shown in fig. S6B, the prodrugs were extremely stable in blank medium without GSH, which could be attributed to the stable amide bond, and this would help to reduce undesired exposure of the parent drug in the blood circulation and normal cells and subsequently alleviate the systemic toxicity of DOX. We then investigated the GSH-triggered drug release of prodrug nanoassemblies, and the reaction was monitored by high-performance liquid chromatography (HPLC) and validated using ultraperformance liquid chromatography (UPLC)–MS/MS measurement. As shown in fig. S6 (F to H), the same intermediate compound occurred at 5.16 min for GSH-responsive drug release from both DSSD (5.98 min) and DSSSD (7.98 min), which was thiol-modified DOX (DOX-SH) confirmed by HPLC-MS (fig. S6, D and E). Because of the good stability of the amide bond in PBS, no free DOX was released from DOX-SH (fig. S6F). As shown in Fig. 3 (C to E), DSSSD was more quickly converted to DOX-SH than DSSD in various concentrations of GSH. For instance, more than 25% of DSSSD was converted to DOX-SH within 30 min in the presence of 0.5 mM GSH but less than 10% was converted for DSSD (Fig. 3D). Furthermore, 68% of DOX-SH was released from DSSSD NPs in 60 min at 1 mM GSH, compared with only 35% for DSSD NPs under the same conditions (Fig. 3E). In response to the overexpressed GSH in tumor cells, the disulfide bond has been widely used as the golden standard to design GSH-responsive DDS. However, our results showed that the trisulfide bond was more sensitive to GSH than the disulfide bond, thus highlighting the potential of trisulfide bond to be used as a GSH-responsive linkage. In addition, we investigated the drug release of prodrug nanoassemblies in PBS (pH 7.0) (cytosol of tumor cells) (29). As shown in fig. S6C, there is no significant decrease on the drug release of prodrug nanoassemblies in pH 7.0 compared to pH 7.4 (Fig. 3E), suggesting that the reductive cleavage of the disulfide/trisulfide linker would not be affected in the cytoplasm of tumor cells.

Fig. 3 GSH-triggered drug release and the responsive mechanism of prodrug nanoassemblies.

(A) GSH-responsive mechanism of disulfide bond. (B) GSH-responsive mechanism of trisulfide bond. (C) In vitro drug release in 0.1 mM GSH. (D) In vitro drug release in 0.5 mM GSH. (E) In vitro drug release in 1 mM GSH. Data are presented as means ± SD (n = 3).

The GSH-responsive mechanism of prodrug nanoassemblies is illustrated in Fig. 3 (A and B). In the presence of GSH, the disulfide bond was degraded into DOX-SH and the mechanism of this reaction has been investigated in detail (25). Trisulfide has been commonly used as H2S donator, which could release H2S upon reaction with thiols (30, 31). Compared with disulfide bond, the reaction of trisulfide bond with GSH is more complex because GSH can attack at either the α- or β-sulfur (30). For example, nucleophilic attack by GSH at the β-sulfur resulted in the formation of DOX-SH and DOX-GSH trisulfide (DOX-SSSG) (30). Alternatively, nucleophilic attack by GSH at the α-sulfur led to the formation of a hydrophilic thiol group (DOX-SSH) and DOX-GSH disulfide (DOX-SSG) (30). DOX-SSH, DOX-SSG, and DOX-SSSG would further react with GSH, leading to the generation of oxidized GSH (GSSG) and DOX-SH (fig. S6E). In general, redox reactions were characterized by the transfer of electrons, with one reactant undergoing oxidation (losing electrons), while another undergoes reduction (gaining electrons) (32). The trisulfide bond is more reduction sensitive than the disulfide bond, possibly owing to the following three points. (i) The trisulfide bond has more GSH-triggering sites than the disulfide bond (three versus two) (32). (ii) The reduction potential (electron-gain ability) of polysulfide can be calculated as (2n – 2) e−1 (n ≥ 2) for R-Sn-R (32). Therefore, trisulfide bond (4e−1) has a higher reduction potential than the disulfide bond (2e−1). (iii) The β-sulfur in the trisulfide bond has a higher chemical valence (0) than the sulfur atoms of the disulfide (−1), thereby undergoing reduction more easily (33).

In vitro cytotoxicity

The cytotoxicity of DOX and prodrug nanoassemblies on tumor cells and normal cells (B16-F10 cells, 4T1 cells, KB cells, and L02 cells) is shown in fig. S7 (A to D), and the half-maximal inhibitory concentration (IC50) values are summarized in Fig. 4A. DOX solution showed obvious dose-dependent cytotoxicity on both cancer cells and normal cells. Compared with DOX solution, DSSD NPs and DSSSD NPs had lower cytotoxicity on cancer cells, which might be attributed to the delayed release of the active drug from prodrug nanoassemblies. Furthermore, prodrug nanoassemblies, especially DSSSD NPs, showed much lower cytotoxicity on L02 cells compared with cancer cells (Fig. 4A and fig. S7D). The selectivity index (SI) of these formulations between normal and tumor cells is calculated in fig. S7E. A high SI value correlates with high tumor selectivity. DSSSD NPs showed higher SI values than DSSD NPs and DOX solution, indicating that DSSSD NPs could selectively kill tumor cells while showing good safety for normal cells. This phenomenon could be attributed to the high selective drug release of DSSSD NPs between tumor cells and normal cells (Fig. 4B). The higher GSH-sensitive trisulfide bond enabled rapid drug release of DSSSD NPs in tumor cells. In addition, the trisulfide bond improved the stability of prodrug nanoassemblies, which could help to maintain the intact nanostructure of DSSSD NPs and subsequently reduce the active drug release in normal cells.

Fig. 4 Cytotoxicity, pharmacokinetics, and in vivo biodistribution of prodrug nanoassemblies.

(A) IC50 values (nM) of free DOX and prodrug nanoassemblies against three tumor cell lines and one normal cell line for 48 hours (n = 3). (B) Selective bioactivation of DSSSD NPs in tumor cells and normal cells. (C) Intracellular GSH concentrations of the L02, 4T1, B16-F10, and KB cells (n = 3). (D) Fluorescence spectrum of DOX or DOX derivative (in methanol) and prodrug nanoassembles (in water) at a DOX concentration of 2 μg/ml. (E) Flow cytometric analyses of B16-F10 cells after incubation with free DOX or prodrug nanoassemblies for 2 and 6 hours (n = 3). (F and G) CLSM images of B16-F10 cells incubated with free DOX or prodrug nanoassemblies for 2 and 6 hours. Scale bar, 20 μm. (H) Molar concentration–time curves of the prodrugs (n = 5). (I) Molar concentration–time curves of the released DOX (n = 5). (J) Tumor accumulation of DOX and prodrug nanoassemblies (n = 3). *P < 0.05, **P < 0.01, and ***P < 0.001 by two-tailed Student’s t test. n.s., not significant.

The DSSSD NPs exhibited much higher cytotoxicity than DSSD NPs on B16-F10 and KB cells, while the cytotoxicity of DSSSD NPs was mostly close to that of the DSSD NPs in 4T1 cells. It has been found that faster drug release from prodrug nanoassemblies results in higher cytotoxicity (19, 28). As shown in fig. S7 (F to J), DOX-SH showed comparable cytotoxicity with the free DOX in both the cancer cells and normal cells. The similar cytotoxicity of both DOX and DOX-SH indicated that the sulfhydryl modification on DOX did not affect DOX’s cytotoxicity, which was in good agreement with other studies (34, 35). Therefore, cleavage of the disulfide/trisulfide bond by intracellular GSH would be the critical rate-limiting step to elicit cytotoxicity. Then, we evaluated the levels of GSH in different tumor cells and normal cells (KB, B16-F10, 4T1, and L02 cells). As shown in Fig. 4C, a significantly higher level of GSH was detected in the three tumor cells compared to L02 cells. These findings supported that the prodrug nanoassemblies could selectively kill tumor cells rather than L02 cells due to the GSH-triggered intracellular drug release. Moreover, DSSSD NPs exhibited stronger cytotoxicity on GSH-overproduced B16-F10 and KB cells than DSSD NPs, which was consistent with the in vitro drug release. In addition, 4T1 cells exhibited a lower GSH level compared to B16-F10 and KB cells, which may account for the similar cytotoxicity between DSSSD NPs and DSSD NPs.

Intracellular distribution of prodrug nanoassemblies

We first investigated the fluorescence spectrum of DOX, prodrugs, and prodrug nanoassemblies. As shown in Fig. 4D, the fluorescence intensity of DSSD and DSSSD was significantly decreased compared with DOX, while the released DOX-SH can restore the fluorescence of DOX. Notably, DSSD NPs and DSSSD NPs suffered from aggregation-caused quenching effect. Then, the endocytosis and intracellular distribution of prodrug nanoassemblies were investigated in B16-F10 cells by confocal laser scanning microscope (CLSM). It was reported that DOX and DOX-SH could transport to the nucleus by forming a DOX-proteasome complex (34). As shown in Fig. 4F, only a small amount of drug was released from prodrug nanoassemblies at the first 2 hours, with inefficient internalization into cell nucleus (Fig. 4F). With the time increasing to 6 hours, the cleavage of the disulfide/trisulfide linkage allows efficient DOX-SH release and fluorescence recovery, and a large amount of DOX-SH/DOX moved from the cytoplasm to the nucleus (Fig. 4G). Therefore, a time-dependent increase in fluorescence intensity was observed from the cell cytoplasm (Fig. 4F) to the cell nucleus (Fig. 4G) due to intracellular GSH-induced activation of the prodrug and migration of the DOX-SH/DOX. Furthermore, the intracellular fluorescence intensity was measured by flow cytometry. In general, DSSD NPs and DSSSD NPs should show similar cellular uptake due to their similar surface properties and particle size (19). After internalization into tumor cells, the release of DOX-SH would be the critical step for fluorescence recovery. As shown in Fig. 4E, the DSSSD NP–treated group exhibited a higher fluorescence intensity than DSSD NPs at 6 hours, which could be attributed to the faster drug release from DSSSD NPs than DSSD NPs. These results were in good agreement with the in vitro release and cytotoxicity.

In vivo pharmacokinetics and biodistribution

Next, the pharmacokinetics of prodrug nanoassemblies and DOX solution in Sprague-Dawley (SD) rats was investigated. For the prodrug nanoassembly–treated groups, no DOX-SH was detected, possibly due to the low concentration (2 to 10 μM) of GSH in blood (29). Therefore, most prodrugs remained intact in the blood circulation, which was in good agreement with the in vitro drug release in PBS (pH 7.4) without GSH. Furthermore, only a small amount of DOX was released because of the presence of amidase in blood. The results are summarized in Fig. 4 (H to I) and fig. S8A. DOX solution showed rapid systemic clearance and a short half-life (t1/2). In contrast, prodrug nanoassemblies significantly prolonged the circulation time of DOX. The area under the blood concentration curve (AUC) of DSSSD NPs and DSSD NPs was 8.31- and 2.61-fold greater than that of the DOX solution, respectively. Furthermore, DSSSD NPs exhibited a much higher AUC and t1/2 than DSSD NPs as it had better colloidal stability. In comparison, the nanostructure of DSSD NPs should be destroyed in vivo and then quickly cleared from the body. These results indicated that the stability of prodrug nanoassemblies had a significant impact on the pharmacokinetics and highlighted the key role of the trisulfide bond in improving the in vivo fate of prodrug nanoassemblies.

The biodistribution and intratumoral accumulation of prodrug nanoassemblies were examined in 4T1 tumor–bearing BALB/C mice. It was difficult to detect the DOX-SH in tissue samples because the active thiol could react with the massive bioactive compounds in vivo. As shown in fig. S8B, DOX-SH was only detected in liver, possibly due to the high local concentrations of GSH in hepatocytes and rapid accumulation of NPs in liver (36). As shown in Fig. 4J, DSSSD NPs exhibited higher accumulation in tumors than DOX solution and DSSD NPs, which could be attributed to the good colloidal stability of DSSSD NPs, contributing to the favorable pharmacokinetic and enhanced permeability and retention effect.

Antitumor efficiency

The in vivo antitumor activities of the prodrug nanoassemblies were evaluated using a subcutaneous 4T1 breast xenograft model. As shown in Fig. 5 (A to C), there were no significant differences between DSSD NPs and DOX solution in terms of tumor volume and tumor burden, which could be ascribed to the poor colloidal stability, rapid systemic clearance, and limited tumor accumulation of DSSD NPs. In contrast, DSSSD NPs exhibited multiple therapeutic advantages including high drug loading, good colloidal stability, a long circulation time, high tumor accumulation, and efficient drug release in tumor cells. As a result, DSSSD NPs displayed potent antitumor activity with almost no growth in tumor volume. These results suggested the significant improvement of therapeutic efficacy of the trisulfide bond–based systems over the conventional disulfide bond–based systems. In addition, the body weights of DOX solution–treated mice (2 mg/kg) dropped by 16.43% after the treatment (Fig. 5G), indicating that DOX induced serious adverse side effects. In comparison, the body weights (Fig. 5G) and hematological parameters (fig. S9, A and B) of the prodrug nanoassembly–treated groups showed no difference from the saline group. In addition, no obvious histological damage was observed in hematoxylin and eosin (H&E)–stained tissue sections of major organs (heart, liver, spleen, lung, and kidney; fig. S9C). These results suggested that DSSSD NPs were “highly selective”: DSSSD NPs could effectively inhibit tumor growth but were well tolerated and had negligible nonspecific toxicity in major organs and tissues.

Fig. 5 Antitumor efficacy of prodrug nanoassemblies.

(A) Tumor volume growth profiles of BALB/C mice bearing 4T1 xenografts. (B) Tumor burden after different treatments in BALB/C mice bearing 4T1 xenografts. (C) Digital photos of excised tumors after different treatments in BALB/C mice bearing 4T1 xenografts. (D) Tumor volume growth profiles of C57BL/6 mice bearing B16-F10 melanoma. (E) Tumor burden after different treatments in C57BL/6 mice bearing B16-F10 melanoma. (F) Digital photos of excised tumors after different treatments in C57BL/6 mice bearing B16-F10 melanoma. (G) Body weight changes of BALB/C mice bearing 4T1 xenografts. (H) Body weight changes of C57BL/6 mice bearing B16-F10 melanoma. Data are presented as means ± SD (n = 5). *P < 0.05, **P < 0.01, and ***P < 0.001 by two-tailed Student’s t test. Photo credit: Yinxian Yang, Department of Pharmaceutics, College of Pharmacy, Shenyang Pharmaceutical University, Shenyang 110016, China.

Inspired by the effect of DSSSD NPs, the in vivo antitumor efficacy of prodrug nanoassemblies was further investigated in B16-F10 melanoma–bearing C57BL/6 mice with two dose schedules (3 and 9 mg/kg). As shown in Fig. 5 (D to F), the therapeutic benefit of DOX solution is much obvious in B16-F10 melanoma tumors compared to 4T1 tumors, resulting in a significant tumor inhibition on B16-F10 melanoma–bearing C57BL/6 mice after treatment with DSSD NPs or DSSSD NPs. In comparison, DSSSD NPs still showed stronger tumor inhibition on tumor sizes than DSSD NPs at the low doses and significantly inhibited tumor growth at either low or high doses. These results were further confirmed by the statistically significant difference that existed between free DOX and DSSSD NPs at either low or high doses but only for free DOX and DSSD NP at high doses. Together, the above results further confirmed that DSSSD NPs exhibited improved therapeutic efficacy over the DSSD NPs. Although DOX solution was only administered at a low dose (3 mg/kg), the body weights of the mice were significantly decreased by 19.36% (Fig. 5H). Furthermore, DOX solution caused severe blood toxicity because the numbers of red blood cell (RBC) and blood platelet (PLT) were abnormal (Fig. 6, A and B). In comparison, there were no significant differences between the prodrug nanoassemblies and saline group in terms of body weights (Fig. 5H), blood index values (Fig. 6, A to C), hematological parameters (fig. S10, A and B), and H&E-stained tissue sections (fig. S10C). Therefore, DSSSD NPs showed the potential to be administered at a higher dose and then achieved more potent antitumor activity.

Fig. 6 In vivo safety of prodrug nanoassemblies.

The blood index values after different treatments in C57BL/6 mice bearing B16-F10 melanoma. (A) Red blood cell (RBC; 1012/liter). (B) Blood platelet (PLT; 109/liter). (C) White blood cell (WBC; 109/liter). Data are presented as means ± SD (n = 3). (D) Body weight changes during treatment with different formulations in healthy BALB/C mice (n = 3). (E) Body weight changes during treatment with different formulations in healthy C57BL/6 mice (n = 4). Hepatorenal toxicity and cardiotoxicity parameters in healthy BALB/C mice. (F) Aspartate aminotransferase (AST; U/liter). (G) Alanine aminotransferase (ALT; U/liter). (H) Creatine kinase (CK; U/liter). (I) Lactate dehydrogenase (LDH; U/liter). (J) Urea nitrogen (UREA; mM). (K) Creatinine (CREA; μM). Data are presented as means ± SD (n = 5). *P < 0.05, **P < 0.01, and ***P < 0.001 by two-tailed Student’s t test.

Acute toxicity studies

The acute toxicities of prodrug nanoassemblies and DOX were assessed in normal BALB/C and C57BL/6 mice. After two injections, the body weights of the DOX-treated mice (10 mg/kg) decreased by over 30%, and all the mice died on day 6 or 7 (Fig. 6, D and E). In contrast, mice treated with prodrug nanoassemblies at an elevated dose of 30 mg/kg (DOX equivalent) for three injections showed no changes in body weights (Fig. 6, D and E), and no mice died during the study. In addition, the hepatorenal toxicity and cardiotoxicity markers were also evaluated based on levels of serum aspartate transaminase (AST), alanine transaminase (ALT), urea nitrogen (UREA), creatinine (CREA), creatine kinase (CK), and lactate dehydrogenase (LDH). The hematological parameters of mice treated with prodrug nanoassemblies (DOX equivalent, 30 mg/kg, three injections) showed no differences from the saline group (Fig. 6, F to K). These results further confirmed that the prodrug nanoassemblies were well tolerated, significantly reducing the systemic toxicity of DOX.

DISCUSSION

The development of advanced nanomedicines for efficient cancer therapy is always restricted by the poor drug loading, poor colloidal stability, and poor tumor selectivity (4, 5). Here, we introduced the trisulfide bond into DOX homodimeric prodrugs to engineer stable NPs with ultrahigh drug loading (67.24%, w/w). It was evident that the increase in drug loading of NPs not only could reduce the adverse effects caused by excessive excipient agents but also had a potential to reduce the cost of the nanomaterials. In previous studies, some DOX homodimeric prodrugs were developed to construct high–drug loading prodrug nanoassemblies (14, 15). However, most of them exhibited poor self-assembly ability and could not self-assemble into stable nanostructures in the absence of surfactant because of the strong intermolecular forces (14). We have previously demonstrated that the disulfide bond can help to balance the intermolecular forces during self-assembly (18). The disulfide bond–bridged paclitaxel-tocopherol prodrug has been observed to self-assemble into NPs, while precipitation was observed for thioether bond–bridged prodrugs (18). In this study, both thioether and disulfide bond–bridged DOX homodimeric prodrugs alone failed to self-assemble into stable NPs. Compared with the thioether bond and disulfide bond, the trisulfide bond provided more sufficient structural flexibility for the prodrugs to balance intermolecular forces and establish a favorable conformation during self-assembly. As a result, the trisulfide bond significantly promoted the self-assembly ability of DOX homodimeric prodrugs, thereby improving the colloidal stability and in vivo fate of prodrug nanoassemblies. The trisulfide bond–induced nanomedicines strategy expand the functional space of conventional self-assembling prodrugs, providing opportunities for the development of advanced nanocarriers with improved loading and stability.

Anticancer drug delivery achieving high therapeutic efficacy and low side effects requires a nanocarrier to remain intact in the systemic circulation and normal cells but has the ability to intelligently release the active parent drugs at tumor sites (5, 29). Here, we found that DSSSD NPs had higher tumor selectivity than DSSD NPs and DOX solution, significantly reducing the systemic toxicity of DOX. Our results showed that the trisulfide bond was more sensitive to GSH than the disulfide bond, thereby enabling more rapid release of active drugs in tumor cells. Thus, DSSSD NPs exhibited stronger cytotoxicity than DSSD NPs on tumor cells. In addition, insertion of a trisulfide bond effectively improved the stability of prodrug nanoassemblies, which could help maintain the intact nanostructure and subsequently reduce the release of active drug in the blood circulation and normal cells. As a result, DSSSD NPs effectively inhibited tumor growth but were well tolerated. We did not find any adverse effects of the prodrug nanoassemblies even when they were administered at an elevated dose in the toxicity studies, including body weight measurements, blood cell counts, heart and hepatorenal enzyme function measurements in blood samples, and histological analysis of heart, liver, spleen, lung, and kidney tissue.

The introduction of trisulfide bridges into synthetic chemistry opens up a range of possibilities, in particular for drug delivery and imaging applications. For instance, linker-based conjugates, such as antibody drug conjugates (ADCs) and small-molecule theranostic prodrugs, have become promising in drug discovery (37). Because of the embedded cavity and chiral environment in cellular redox systems, the stability of the disulfide linker mainly relies on the adjacent steric and stereoscopic effects (38, 39). However, increasing the stability of the disulfide causes a corresponding decrease in the cleavage and payload release (38). Compared with disulfide bridges, trisulfide bridges with longer sulfur chains provide higher steric flexibility and higher redox potential for the molecular system. Thus, the construction of stereoscopic trisulfides may be promising to improve the stability of ADCs and maintains rapid release in target sites. For the imaging of intracellular thiols, a number of disulfide-based fluorescent probes have been investigated (24). Like disulfides, the trisulfide bond can also be cleaved by the nucleophilic thiolate of the thiol, which provides the possibility of efficient imaging tools or chemosensors owing to its greater sensitivity. In addition, polymeric carriers with a trisulfide bridge insertion could undergo thiol-mediated cleavage, and release H2S from long sulfur chains (31). Notably, this trisulfide bond–bridged carrier exhibited more rapid thiol-mediated cleavage and disassembly than the disulfide bond, which was consistent with the findings of our study. Disulfide-based strategies have demonstrated high performance due to their good stability in the bloodstream and efficient cleavage by cellular thiols (24, 40). In comparison to disulfide bonds, trisulfide bonds exhibited comparable stability at the low level of GSH and blood pools but more efficient cleavage at the elevated GSH level. Considering these findings, we believe that the trisulfide linker has potential as a promising candidate in the development of drug delivery, imaging, and materials.

In summary, we found that simple insertion of a single trisulfide bond could turn DOX homodimeric prodrugs into self-assembled nanomedicines with three benefits: high drug loading, high self-assembly stability, and high tumor selectivity. Compared with the disulfide bond and thioether bond, the trisulfide bond effectively promoted the self-assembly of DOX homodimeric prodrugs, thereby improving the colloidal stability, pharmacokinetics, and tumor accumulations of prodrug nanoassemblies. Furthermore, the trisulfide bond showed ultrahigh sensitivity to the overproduced GSH in tumor cells, enabling efficient drug release in tumor sites. As a result, trisulfide bond–bridged prodrug nanoassemblies exhibited highly selective cytotoxicity over tumor cells than normal cells, significantly reducing the systemic toxicity of DOX. Our results provide novel insights into the rational design of prodrug nanoassemblies and highlight the potential of the trisulfide bond for the development of advanced redox-sensitive nanomedicines.

MATERIALS AND METHODS

Materials

DOX was purchased from Dalian Meilun Biotech Co. Ltd. (Dalian, China). O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU), N,N-diisopropylethylamine (DIEA), 3,3′-dithiodipropionic acid, 3,3′-thiodipropionic acid, 3-bromopropionic acid, dithiothreitol (DTT), and H2O2 were purchased from Aladdin Biochemical Technology Co. Ltd. (Shanghai, China). Sodium thiosulfate and sodium sulfide were obtained from Yuwang Pharm Co. Ltd. (Shangdong, China). Cell culture reagents were purchased from GIBCO, Invitrogen Corp. (Carlsbad, CA, USA). DSPE-PEG2k was purchased from Shanghai Advanced Vehicle Technology Co. Ltd. (Shanghai, China). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) was obtained from Dalian Meilun Biotech Co. Ltd. (Dalian, China). Hoechst 33342 was purchased from Beijing Solarbio Science and Technology Co. Ltd. (China). The GSH Assay Kit was obtained from Beyotime Biotechnology (Shanghai, China). The other reagents were of analytical or HPLC grade.

Synthesis of DOX homodimeric prodrugs

First, 3,3′-trithiodipropionic acid was synthesized. Briefly, 3-bromopropionic acid (5 mmol) was dissolved in 20 ml of water, added to the aqueous sodium thiosulfate pentahydrate (6.5 mmol), and stirred at 60°C for 5 hours. When the solution turned clear, the reaction mixture was cooled to room temperature. Then, the sodium sulfide hydrates solution (1.5 mmol) was added, with stirring at room temperature overnight. The pH of the reaction mixture was adjusted to 4.0 to 5.0 using HCl solution. The supernatant was then extracted three times with ethyl acetate to obtain the targeted crude products and purified by preparative liquid chromatography, yielding 30 to 40% of the products. To obtain DOX homodimeric prodrugs, diacid (3,3′-thiodipropionic acid, 3,3′-dithiodipropionic acid, or 3,3′-trithiodipropionic acid, 0.12 mmol) was dissolved in 4 ml of dimethyl sulfoxide (DMSO) in a 50-ml round-bottom flask. Then, HBTU (0.48 mmol) and DIEA (0.24 mmol) were added dropwise sequentially and stirred in an ice bath for 1 hour under a nitrogen atmosphere. Next, DOX (0.25 mmol) was added, and the mixture was further stirred at room temperature for 24 hours under a nitrogen atmosphere. The completion of the reaction was monitored by thin-layer chromatography. The target products were purified to give a red powder. The products were confirmed by high-resolution mass spectra (Agilent 1100 Series LC/MSD Trap) and NMR spectral analyses (400 MHz for 1H, Bruker AV-400) using DMSO-d6 as the solvent.

Preparation and characterization of prodrug nanoassemblies

Self-assembly of DOX homodimeric prodrugs was investigated in aqueous solution via the nanoprecipitation method. Briefly, the DOX prodrugs were dissolved in DMSO and added dropwise into deionized water with vigorous stirring. The colloidal stability of the non-PEGylated DSSSD NPs was investigated in PBS. PEGylated prodrug nanoassemblies were prepared following the same procedure with the addition of DSPE-PEG2k (20%, w/w). The formulations were further dialyzed against ultrapure water overnight to remove the organic solvent. The hydrodynamic diameter and zeta potential of the prodrug nanoassemblies were measured by dynamic light scattering using the Zetasizer (Nano ZS, Malvern Co., UK). The morphology of these prodrug nanoassemblies was observed by TEM (JEOL, Japan). The colloidal stability of the prodrug nanoassemblies was investigated. Briefly, prodrug nanoassemblies with final concentrations of 0.05, 0.1, and 0.25 mg/ml were incubated in PBS (pH 7.4) supplemented with 10% of FBS for 24 hours at 37°C.

Self-assembly mechanism and molecular dynamics simulations

The self-assembly mechanism of hydrophobic homodimeric prodrug molecules was investigated to expound the possible theory. The quantum chemical method, called Geometry, Frequency, Noncovalent, eXtended TB (GFN-xTB), was used for calculations of the structures, vibrational frequencies, and noncovalent interactions of molecular systems. Briefly, geometry optimization and frequency calculations were performed with the xTB program, with the generalized born and surface area solvation (GBSA) implicit solvation model, taking water as the solvent. At least three to seven conformations generated by periodic annealing were optimized for each compound, and the one with the lowest free energy was accepted as the final structure. The dihedral angles of the disulfide/trisulfide linkages of the prodrugs were measured during the calculation of structures. The dimerization of DSD, DSSD, and DSSSD for molecular dynamics simulations, including the energy minimization, was also performed according to the literature methods (26).

GSH-triggered drug release

To investigate the GSH-induced drug release profile of DOX prodrug nanoassemblies, prodrug nanoassemblies (50 nM, 1 ml) were incubated with 30 ml of release medium (ethanol/PBS, v/v = 3/7, pH 7.4) with or without GSH at 37°C. In addition, we investigated the drug release of prodrug nanoassemblies in PBS (pH 7.0) with 1 mM GSH. At designated time intervals, the mixture (200 μl) was withdrawn. The released drug from prodrug nanoassemblies was monitored by HPLC. The detailed GSH-responsive drug release process of prodrugs was monitored by UPLC-MS/MS (Waters Co. Ltd., Milford, MA, USA).

Cell culture

KB cells were obtained from COBIOER Biotechnology Co. Ltd. (Nanjing, China). Other cell lines were obtained from the Cell Bank of Type Culture Collection of Chinese Academy of Sciences. Human normal liver cells (L02), human epidermoid carcinoma cells (KB), murine melanoma cells (B16-F10), and mouse mammary carcinoma cells (4T1) were maintained in accordance with the protocols of the American Type Culture Collection. Briefly, the KB and B16-F10 cells were grown in a 10% FBS–containing Dulbecco’s modified Eagle’s medium supplemented with penicillin (100 U/ml) and streptomycin (100 μg/ml). The L02 and 4T1 cells were propagated in a 10% FBS–containing Gibco 1640 medium containing penicillin (100 U/ml) and streptomycin (100 μg/ml). The cells were grown in a humidified incubator at 37°C with 5% CO2 atmosphere.

Cytotoxicity assay

The MTT assay was applied to evaluate the in vitro cytotoxicity of the DOX solution and prodrug nanoassemblies against different cells. Corresponding cells (5000 cells/well) were seeded in 96-well plates and incubated with the corresponding serial concentrations of drug/prodrug nanoassemblies at 37°C for 48 hours. At the end of the incubation period, each well was treated with 20 μl of MTT (5 μg/μl) for 4 hours. The resulting formazan crystals were dissolved in DMSO, and the absorbance was recorded at 570 nm using a microplate reader (Thermo Fisher Scientific, USA). The experiments were performed in triplicates, and cells without any treatment were used as the control. The IC50 values were calculated by GraphPad Prism 8, using molar concentration and cell viability ratio as parameters.

Intracellular distribution of prodrug nanoassemblies

The fluorescence spectrum of DOX, prodrugs, and prodrug nanoassemblies (DOX, 2 μg/ml) was scanned from 500 to 700 nm with excitation wavelength at 470 nm. B16-F10 cells (1 × 105 cells) were seeded on coverslips overnight. The medium was replaced with fresh medium including free DOX, DSSD NPs, and DSSSD NPs containing equivalent concentration of DOX (20 μg/ml). After incubation for 2 or 6 hours, the cells were washed and fixed with paraformaldehyde. Then, the cells were stained by Hoechst 33342. The prepared covered slips were observed by CLSM (TCS SP2/AOBS, LEICA, Germany). For quantitative analysis, the cells were washed, collected, and resuspended in PBS after incubation with DOX or prodrug nanoassemblies. Cellular uptake was analyzed by flow cytometry on a FACSCalibur instrument (Becton Dickinson).

Animal studies

All experimental animals were supplied by the Animal Centre of Shenyang Pharmaceutical University (Shenyang, China). All the animal experiments were carried out according to the Guide for the Care and Use of Laboratory Animals of Shenyang Pharmaceutical University.

Pharmacokinetic studies

The SD rats weighing 200 to 250 g were used to investigate the pharmacokinetic profiles of the prodrug nanoassemblies. The animals were intravenously administered DOX solution, DSSD NPs, and DSSSD NPs at a dose equivalent to DOX (2 mg/kg). Blood samples were collected at predetermined time points and centrifuged to obtain plasma. The plasma samples were stored at −80°C until analysis. The plasma concentration of DOX and prodrugs was measured by UPLC-MS/MS on an ACQUITY UPLC system (Waters Co. Ltd., Milford, MA, USA).

In vivo biodistribution

The in vivo biodistribution of DOX solution and different prodrug nanoassemblies was assessed in 4T1 tumor–bearing BALB/C mice. When the tumor volume reached approximately 500 mm3, DOX solution, DSSD NPs, and DSSSD NPs were administered via the tail vein at a dose equivalent to DOX (2 mg/kg). At designed time intervals, the mice were sacrificed and their major organs (heart, liver, spleen, lung, and kidney) and tumor were harvested, weighed, and stored at −80°C until analysis. The drugs were extracted by incubating 100 mg of tissue in 1 ml of methanol overnight. After that, the tissues were homogenized in methanol by a high-speed shear under ice bath. Then, the samples were vortexed and centrifuged at 13,000 rpm/min, and the supernatant was collected for analysis. The tissue concentration of prodrugs and DOX was measured by UPLC-MS/MS on an ACQUITY UPLC system (Waters Co. Ltd., Milford, MA, USA).

In vivo antitumor efficacy

To evaluate antitumor effects, 4T1 cells (5 × 106 cells) were inoculated subcutaneously into the right flank of female BALB/C mice. When the tumor reached approximately 120 to 150 mm3, the mice were treated with saline (control), DOX solution, DSSD NPs, and DSSSD NPs at a dose equivalent to DOX (2 mg/kg) five times every other day via the tail vein. Body weight and tumor volume were measured and recorded daily. Mice were sacrificed on the last day of observation, and blood was collected and centrifuged at 4000 rpm for 10 min to obtain the serum for hepatic and renal function analysis. The major organs (heart, liver, spleen, lung, and kidney) and tumor tissues were harvested for further study. The tumor volume (V) and tumor burden were calculated as follows: V (mm3) = (a2 × b)/2 (where a and b represented the width and length of the tumors). Tumor burden (%) = (Wtumor/Wmice) × 100% (“Wtumor” is the weight of the tumor and “Wmice” is the weight of the mouse).

In addition, another tumor model (B16-F10 melanoma–bearing C57BL/6 mice) was used to further evaluate the antitumor efficacy of prodrug nanoassemblies. Each mouse was injected subcutaneously with B16-F10 cells (1 × 106) in the right flank. When the tumor volume reached approximately 120 mm3 after 9 days, the mice were randomly divided into six groups: (i) saline (control), (ii) DOX solution (3 mg/kg), (iii) DSSD NPs (DOX equivalent, 3 mg/kg), (iv) DSSSD NPs (DOX equivalent, 3 mg/kg), (v) DSSD NPs (DOX equivalent, 9 mg/kg), and (vi) DSSD NPs (DOX equivalent, 9 mg/kg). The treatments were carried out using the abovementioned experimental scheme, and tumor volumes and body weights were measured every 2 days. Blood and plasma were collected for blood cell counts and hepatorenal function analysis. The major organs and tumor tissues were harvested for further study.

Acute toxicity studies

The systemic toxicity of the prodrug nanoassemblies was evaluated in healthy male BALB/C and C57BL/6 mice. The (i) saline (control), (ii) DOX solution (2 × 10 mg/kg), (iii) DSSD NP (DOX equivalent, 3 × 30 mg/kg), and (iv) DSSSD NP (DOX equivalent, 3 × 30 mg/kg) groups were intravenously injected every other day. The body weights of the mice were measured every 2 days, and the blood samples were harvested. After centrifugation at 4000 rpm for 10 min, the serum samples were collected for serological analysis. Organs including the heart, liver, spleen, lung, and kidney were also harvested and analyzed by H&E staining.

Statistical analysis

All data were presented as the mean value ± SD. Statistical analysis was performed using Student’s t test (two-tailed) and one-way analysis of variance (ANOVA). Statistical significance was established at P < 0.05, where *P < 0.05, **P < 0.01, and ***P < 0.001.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/45/eabc1725/DC1

https://creativecommons.org/licenses/by-nc/4.0/

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: Funding: This work was financially supported by the National Natural Science Foundation of China (no. 81872816), the Liaoning Revitalization Talents Program (no. XLYC180801), the China Postdoctoral Innovative Talents Support Program (no. BX20190219), and the China Postdoctoral Science Foundation (no. 2019 M661134). Author contributions: Y.Y., B.S., and J.S. conceived the project. Y.Y. and B.S. designed the experiments. Y.Y. and B.S. synthesized and characterized the prodrugs. Y.Y., B.S., S.Zu., X.L., and S.Zh. carried out the animal study. Y.Y., L.L., C.L., Y.W., and Z.H. carried out the drug release and pharmacokinetic studies. S.W. and J.S. contributed to the cytotoxicity study. Y.Y., B.S., and H.L. contributed to the molecular dynamics simulations. Y.Y. and B.S. analyzed the data and wrote the manuscript. M.C. and J.S. participated in the manuscript revision. All authors discussed the results and reviewed the manuscript. Competing interests: J.S. and Z.H. are inventors on a patent application related to this work filed by Shenyang Pharmaceutical University (filed under number 202010390944.X). The other authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Stay Connected to Science Advances

Navigate This Article