Doxorubicin hydrochloride and L-arginine co-loaded nanovesicle for drug resistance reversal stimulated by near-infrared light

Abstract

Drug resistance is accountable for the inadequate outcome of chemotherapy in clinics. The newly emerging role of nitric oxide (NO) to conquer drug resistance has been recognized as a potential strategy. However, it remains a great challenge to realize targeted delivery as well as accurate release of NO at desired sites. Herein, we developed a PEGylated indocyanine green (mPEG-ICG) integrated nanovesicle system (PIDA) to simultaneously load doxorubicin hydrochloride (DOX⋅HCl) and the NO donor L-arginine (L-Arg), which can produce NO triggered by NIR light irradiation and exert multimodal therapy to sensitize drug-resistant cancers. Upon 808 nm irradiation, the NO released from PIDA led to a decrease in mitochondrial membrane potential, an increase in ROS and significant ATP depletion in K562/ADR cells, thus inhibiting cell growth and resolving the problem of drug resistance. Consequently, the in vivo experiment on K562/ADR-bearing nude mice indicated that PIDA nanovesicles achieved significant anticancer efficacy with a tumor inhibition rate of 80.8%. Above all, PIDA nanovesicles offer guidance for designing nanoplatforms for drug-resistant cancer treatment.

Graphical abstract

L-Arginine and doxorubicin hydrochloride co-loaded PIDA nanovesicles integrated with mPEG-ICG were constructed to achieve drug resistance reversal when stimulated by near-infrared light.

1. Introduction

As the first gaseous signaling molecule to be recognized, nitric oxide (NO) displays some medical influence on the nervous, cardiovascular, reproduction, and immune systems [1,2]. Over the years, NO has also been found to exert antitumorigenic effects by causing mitochondria/DNA damage, suppressing cellular respiration and inhibiting DNA repair. Moreover, it has been demonstrated recently that NO can reverse drug resistance [3], [4], [5], which is known as a serious barrier to chemotherapy in clinics. The mechanisms are currently under intense study and could be roughly concluded as reducing and inhibiting P-glycoprotein (P-gp) expression to increase drug concentration in cancer cells [6], [7], [8], [9], inducing the nitration of proteins participating in DNA repair and inducing DNA damage [10,11], promoting glutathione (GSH) depletion [12], [13], [14] or inhibiting NF-κb-associated pathways [15], [16], [17]. Scheme 1.

Scheme 1.

Fabrication schematics of PIDA nanovesicles and proposed mechanism of PIDA nanovesicles to overcome drug resistance.

Nonetheless, the unmediated use of NO in the gaseous state suffers from its short half-life and susceptibility to react with various biological molecules, such as glutathione, hemoglobin, oxygen or reactive oxygen intermediates [18,19]. Therefore, realizing the spatiotemporal release of NO at target sites remains a great challenge. To overcome this problem, the present research has focused on the strategy of incorporating or conjugating NO donors to nanomaterials to produce NO on demand [20], [21], [22]. To that aim, a series of active NO donors have been developed. Among them, N-diazeniumdiolate (NONOate) and N,N′-di‑sec‑butyl‑N,N′-dinitroso-1,4-phenylenediamine (BNN6) have been the most extensively studied. NONOate could decompose quickly in solution at physiological pH and temperature, which then provided two molar equivalents of NO [23], [24], [25]. BNN6, the cage small molecule NO donor, was designed to release two molar equivalents of NO upon ultraviolet and visible light irradiation [26]. However, the toxicity of NONOate and its metabolite (carcinogenic secondary nitrosamines) hinders its clinical application [27]. Additionally, the toxicity of BNN6, which contains a nitrosamine functional group, must be mitigated [28]. In contrast, the NO donor L-arginine (L-Arg) is considerably biocompatible and stable and can be transferred to NO and citrulline via NO synthases (NOSs) or reacted with high oxidative reactive oxygen species (ROS) [28], [29], [30]. In light of these resources, how to accurately deliver NO donors to tumor sites is a critically predominant issue. Several kinds of inorganic nanomaterials have been applied as carriers to load NO donors via chemical conjugation or physical adsorption. For example, core shell-structured Cu_2-xSe@SiO₂ was constructed to graft NONOates by a nucleophilic addition reaction of secondary amines of silane coupling agent (N-[3-(tri-methoxysilyl)-propyl]ethylenediamine, AEAPTS) with NO gas [31]. PEG-MoS₂-BNN6 was developed by grafting PEG to MoS₂ nanoflowers, followed by BNN6 loaded via hydrophobic interactions [32]. Studies have also attempted to use inorganic carriers to load L-Arg via physical adsorption or attach L-Arg to polymers through chemical bonds [33,34]. For example, GMOF-LA (Au NPs and L-Arg-coloaded Zr-TCPP MOF nanosheets) taking advantage of the ultralarge surface area in 2D materials made the successful loading of L-Arg possible [35]. A-MPDI (L-Arg-modified mesoporous polydopamine) was prepared by the covalent connection between the nucleophilic amine groups of L-Arg and the 5,6-dihydroxyindole (DHI) units of PDA [36]. However, the restricted biodegradability, unpredictable systemic toxicity and side effects of inorganic-based carriers have made clinical applications rather difficult [37,38]. The complicated preparation procedure, limited loading content of NO donors, and burst release would also cause huge obstacles. Consequently, it is still difficult to simultaneously fulfill the convenient preparation, high loading and good biocompatibility of the efficient delivery of L-Arg.

In this study, we constructed doxorubicin hydrochloride (DOX⋅HCl) and L-Arg-coloaded nanovesicles (PIDA), which could release NO triggered by NIR light and accomplish multimodal combined precision therapy with the aim of improving the inhibitory effect against drug-resistant cancer. First, polymer carriers formed by poly[(PEG)(ethyl-p-aminobenzoate)phosphazene] (PEP) with an optimized chemical structure were synthesized. Using the conventional dialysis method, self-assembled PEP nanovesicles were constructed, which could encapsulate DOX⋅HCl and L-Arg in the center aqueous space of the nanovesicles and integrate PEGylated indocyanine green (mPEG-ICG) into the hydrophobic vesicle membrane at the same time. The hydrophobic interaction between mPEG-ICG and PEP enhanced the stability of the nanovesicles, which prevented drug leakage during blood circulation and ensured that as much L-Arg and DOX⋅HCl reached the tumor as possible via the EPR effect of the nanovesicles. Moreover, our previous work confirmed that mPEG-ICG could act as a reliable photosensitizer and produce heat and ROS under 808 nm irradiation [39]; namely, mPEG-ICG could act as an agent for both mild photothermal therapy (MPTT) and photodynamic therapy (PDT). Then, based on the reaction between ROS and L-Arg, the newborn NO would cause further sensitization of cancer cells to DOX⋅HCl via mitochondrial dysfunction [40]. Simultaneously, the cell apoptosis or necrosis caused by NO and DOX⋅HCl could compensate for the uneven heat distribution of the MPTT treatment and achieve better treatment efficiency with a lower light power density and dosage. As described above, NIR-triggered multimodal synergistic cancer therapy, including NO, MPTT, PDT and chemotherapy, is integrated into novel nanovesicles with convenient preparation to realize better therapeutic efficiency. Additionally, with the help of mPEG-ICG, the nanosystem could act as a fluorescence emitter to image tumors in mice. To validate this assumption, in vitro and in vivo investigations, including ROS and NO generation, cytotoxicity and apoptosis, mitochondrial dysfunction, and antitumor activity, were performed.

2. Materials and methods

2.1. Materials

Doxorobicin hydrochloride (DOX·HCl) was obtained from HaiKou Manfangyuan Chemical Company (Haikou, China). L-Arginine (L-Arg) was obtained from Innochem Co., Ltd. (Beijing, China). α-Naphthol was obtained from Infinity Scientific (Beijing) Co., Ltd. (China). 2,3-Butanedione was purchased from Shanghai Macklin Biochemical Co., Ltd. (China). 1,3-Diphenylisobenzofuran (DPBF) was obtained from Shanghai Aladdin Biochemical Technology Co., Ltd. (China). DAF-FM DA NO fluorescence probe, griess reagent kit, reactive oxygen species (ROS) assay kit and enhanced ATP assay kit were all obtained from Beyotime Institute of Biotechnology (China). Annexin V-FITC/PI apoptosis kit was obtained from Multi Sciences (Lianke) Biotech, Co., Ltd. (China). Image-iT™ TMRM reagent was obtained from ThermoFisher Scientific. All other chemical reagents could be achieved commercially and used as received.

2.2. Synthesis and characterization of PEP and mPEG-ICG

As previously reported, PEP was synthesized by sequent substitution reaction and PEGylated indocyanine (mPEG-ICG) was synthesized via the condensation reaction [39,41]. The details of syntheses were depicted in the supporting information.

2.3. The preparation of nanovesicles

The assembly of mPEG-ICG incorporated nanovesicle (PI), mPEG-ICG incorporated/ DOX⋅HCl loaded nanovesicle (PID), and mPEG-ICG incorporated/ DOX⋅HCl & l-Arg-coloaded nanovesicle (PIDA) was achieved via the simple dialysis method. For the preparation of PI, 3 mg mPEG-ICG and 10 mg PEP were dissolved in 500 µl DMF and then the same volume of deionized water was added dropwise under stirring, followed by 6 h dialysis against water. The preparation of PID nanovesicle was almost the same as PI nanovesicle, with the exception of containing 2 mg DOX⋅HCl in the deionized water. For PIDA, 2 mg DOX⋅HCl was dissolved in 250 µl deionized water and 20 mg L-Arg was dissolved in PBS 5.5. Two kinds of solution were mixed and then drippled into DMF solution as stated above, followed by 6 h dialysis against water.

2.4. Characterization

Various sample solutions (1 ml) were pipetted into the plastic cuvette and inserted into sample chamber of dynamic light scattering (DLS, Malvern Nano-ZS 90, U.K.) for the detection of both particle sizes and ζ-potentials of nanovesicles. To study the morphologies, samples were carefully dripped on the support films and transferred under Transmission electron microscope (TEM, JEM-1010, Japan) for observation. The UV‒vis spectra of substances in deionized water were obtained by UV‒vis spectrometry (TU-1800 PC, Beijing Persee General Instrument Co., Ltd., China). The NIR light was provided by 808 nm laser (LSR808NL-3.5W-FC, China).

The concentration of mPEG-ICG and DOX⋅HCl was determined by UV‒vis spectrophotometry using TU-1800 PC spectrophotometer (Beijing Persee General Instrument Co., Ltd., China) at 780 nm and 480 nm, respectively. First, the standard curves of mPEG-ICG and DOX·HCl were determined in DMF. Then, the demulsification process was conducted by adding 9900 µl of DMF to 100 µl of various formulations. The absorbances at 780 nm and 480 nm were measured and converted to the concentration according to the standard curves. The loading content (LC) of nanovesicles was determined with the following formula:

$$\begin{array}{ccc}\hfill \text{LC}(\%)& =& \text{weight}\text{of}\text{drug}(\text{mPEG}\text{-}\text{ICG}\text{or}\text{DOX}·\text{HCl})\text{loaded}\text{in}\hfill \\ & & \text{nanovesicles}/\text{total}\text{weight}\text{of}\text{nanovesicles}\times 100\%.\hfill \end{array}$$

The concentration of L-Arg was determined by the classical diacetyl-α-naphthol assay [42]. In detail, the standard coloration liquid was prepared by dissolving 1 g of α-naphthol and 5 µl diacetyl in 20 ml n-propanol. Then, adding 0.17 ml 14wt% NaOH, 0.5 ml standard coloration liquid and 0.33 ml n-propanol to 1 ml of the test sample and maintained in a 35 °C water bath for 10 min. A cooling process in water was followed. Besides, in order to prevent the absorbance interference of DOX·HCl at 530 nm, DOX·HCl solution with the same concentration as PIDA sample was prepared as the background sample and the L-Arg-concentration was detected via the same process as above. Finally, the absorbance at 530 nm (Abs_sample - Abs_DOX·HCl) was used to convert the content of L-Arg through the standard curve. LC was calculated according to the equation: LC (%) = amount of loaded L-Arg/amount of L-Arg-loaded nanovesicles× 100%.

2.5. In vitro drug release

Various samples with DOX⋅HCl mass of 500 µg were added into the dialysis bag (MWCO: 8–14 kDa) and sealed. Every dialysis bag was put inside a disposable centrifuge tube consisted of 10 ml PBS solution (pH = 5.5, 6.5 and 7.4) and put into the shaker with the temperature fixed at 37 °C and shaking speed at 110 rpm. At regular intervals, 1 ml of solution in the tube for every sample was taken out for testing and same volume of fresh PBS at 37 °C with correspondent pH was then added. The concentration of DOX⋅HCl of various samples was measured by UV‒vis spectrometry and calculated through the standard curves. All experiments had three duplicated samples.

ROS generation of mPEG-ICG, PI, PID and PIDA was measured by DPBF. Briefly speaking, DPBF ethanol solution was first prepared as detection reagent in dark. Then, PI, PID and PIDA solution was mixed evenly with the detection reagent in equal volumes. DPBF absorbance spectra were recorded at different times (0, 15, 30 s and 1, 2, 3, 4, 5 min) after exposure to 808 nm laser (0.5 W/cm²).

2.7. In vitro no release

The NO generation in various formulations was quantitatively measured by typical griess reagent kit. PIDA solution (1 ml) was pipetted to a quartz cuvette and exposed to 808 nm laser for 5 min at a power intensity of 0.5 W/cm². After irradiation, 50 µl PIDA was taken out and gently pipetted into a 96-well plate. Then, 50 µl Griess Reagent I was pipetted into the well and shaken for 5 min to mix evenly. After shaken, another 50 µl Griess Reagent II was pipetted to the well and shaken for another 5 min. To avoid the signal interference of DOX·HCl, DOX·HCl solution with the same concentration as PIDA sample was prepared as the background sample and the NO generation was detected via the same process as above. The absorbance values could then be obtained using microplate reader (Multiskan MK3, ThermoFisher Scientific, America). Standard curve of NO was made as the protocol suggested. The absorbance value difference (Abs_sample - Abs_DOX·HCl) was used to calculate the concentration of NO by standard curve.

2.8. In vitro photothermal performance

To investigate the photothermal performance, we irradiated mPEG-ICG, PI, PID and PIDA with the same concentration of mPEG-ICG by NIR light (0.5 W/cm²). The changes in temperature were monitored by Fotric 235 (Fotric, China) and recorded every 10 s. The photothermal stability was obtained from three cycles of heating and cooling via laser ON/OFF. The photothermal conversion efficiency was calculated through the heating-cooling curve.

2.9. Cell

The drug-resistant human chronic myelogenous leukemia K562/ADR cells were obtained from Key Gen Biotechnology Co., Ltd. (Nanjing, China). Cells were maintained in RMPI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (Sijiqing Biologic, Hangzhou, China) under the condition of 37 °C and 5% CO₂.

2.10. Intracellular generation of NO

DAF-FM DA NO probe was used to investigate NO production within cells through CLSM (LSM-880, Zeiss, Germany). Typically, 5 × 10⁵ K562/ADR cells were gently pipetted in the well of 12-well plates and transferred to incubator for 24 h. Afterward, 1 ml detection solution (5 µM DAF-FM DA per well) was gently pipetted into each well and co-incubated with cells for 20 min at 37 °C. The cells were incubated with PBS, DOX⋅HCl, PI, PID and PIDA for 8 h after being washed PBS. For PI NIR, PID NIR and PIDA NIR groups, cells were exposed to 808 nm laser (0.5 W/cm², 5 min). After staining nuclei by Hoechst 33342, the cells were washed by PBS to remove the staining reagent and fixed with 4% paraformaldehyde. Eventually, the cells were transferred and observed under CLSM.

2.11. Intracellular mitochondria membrane potential assay

The change of mitochondria membrane potential was explored through TMRM probe. K562/ADR cells (2 × 10⁵) were gently pipetted in the well of 12-well plates and transferred to incubator for 24 h. Then, the cells were cocultured with PBS, DOX⋅HCl, PI, PID and PIDA for 4 h. NIR groups were irradiated by 808 nm laser for 5 min at a density of 0.5 W/cm². Nuclei were stained after being washed with PBS for three times. Finally, the cells were observed under CLSM (LSM-880, Zeiss, Germany) after being incubated with 1 µl TMRM stock solution (30 min, 37 °C).

2.12. ATP level measurement and ROS production

The enhanced ATP assay kit was conducted for ATP level measurement in K562/ADR. 5 × 10⁴ K562/ADR cells were gently pipetted in the well of 24-well plates and transferred to incubator for 24 h. Afterward, cells were co-incubated with different samples at the same dosage of intracellular mitochondria membrane potential assay. NIR groups were irradiated by 808 nm laser for 5 min at a density of 0.5 W/cm² after 6 h. Under 24 h incubation, cells were obtained and treated as recommended in the protocol. The luminescent signals of ATP were obtained by microplate spectrophotometer (Molecular Devices, SpectraMax M2e, USA).

Reactive oxygen species assay kit was used to detect ROS production in K562/ADR cells. Briefly, cells were seeded and coincubated with different samples according to the protocol of ATP detection. Cells were possessed in accordance with the protocol and analyzed via flow cytometry (Cytoflex S, Beckman) after 24 h incubation.

2.13. Cellular uptake of PID and PIDA

In detail, 2 × 10⁵ K562/ADR cells were pipetted gently in the well and transferred to incubator for 24 h. Afterward, DOX⋅HCl, PID and PIDA were added and incubated for 2, 4, 8 and 24 h. All NIR groups were exposed to the laser immediately after drug feeding. To be noticed, the total mass of DOX⋅HCl in each well was fixed at 10 µg. At predetermined time points, the cells were gently digested by trypsin and fixed by 4% paraformaldehyde. After resuspending by PBS, cells were analyzed by flow cytometer (Cytoflex S, Beckman). We prepared three duplicated samples for each timepoint and each group.

2.14. Cell cytotoxicity and apoptosis assay

CCK-8 assay (ApexBio, MA, USA) was utilized to analyze the cell cytotoxicity of DOX⋅HCl, PEP, PI, PID and PIDA against K562/ADR cells. Briefly, 1 × 10⁵ K562/ADR cells were gently pipetted in the well of 96-well plates and transferred to incubator for 24 h. After that, samples at different concentration (25 µl) were added to each well and coincubated with cells for another 24 h. For NIR irradiation groups, cells were irradiated under 0.5 W/cm² for 5 min. Then, CCK-8 solution (10 µl) was pipetted into each well and another 2 h was needed for coincubation. Finally, the OD value of each well was obtained via a microplate reader (Multiskan MK3, ThermoFisher Scientific, America) at 450 nm. Moreover, the reversal index (RI) was calculated as following formula: RI (%) = IC₅₀ against DOX⋅HCl treated cells/IC₅₀ against nanoparticles treated cells × 100%

The apoptosis of cells was conducted via Annexin V-FITC/PI apoptosis kit. In general, 2 × 10⁵ K562/ADR cells were gently pipetted in the well of 12-well plates and transferred to incubator for 24 h. Afterward, the cells were cocultured with PBS, DOX⋅HCl, PI, PID and PIDA (containing 0.5 µg/ml DOX⋅HCl and 1.5 µg/ml mPEG-ICG) for 24 h in total. PI NIR, PID NIR and PIDA NIR groups were exposed to 808 nm laser (0.5 W/cm², 5 min) 6 h after coincubation with different samples. The cells were stained with 5 µl Annexin V-FITC solution and 10 µl PI solution just before analyzation via flow cytometry (Cytoflex S, Beckman).

2.15. Animals

The BALB/C athymic nude mice (female, 4–6 weeks old) were obtained from Shanghai SLAC Laboratory Animal Co., Ltd (Shanghai, China). The tumor model could be successfully established through subcutaneously injection 2 × 10⁷ K562/ADR cells into the right armpit of nude mice after approximately two weeks. And all animal experiments were approved by Animal Ethics Committee of Zhejiang University (Hangzhou, China).

2.16. Fluorescence imaging

PI, PID and PIDA were injected intravenously when the tumor volume reached ∼400 mm³. Next, in vivo imaging of mice was performed. Specifically, the mice for different times after injection were anesthetized by isoflurane and then gently placed on the platform inside in vivo imaging system (CLS136341/F, PerkinElmer, Germany). After making the above preparations, images were captured and obtained through the ICG fluorescence channel. Ex vivo imaging was performed under the same in vivo imaging system by collecting organs and tumors from mice 24 h after drug injection. Two channels of fluorescence were used for the detection of DOX⋅HCl and mPEG-ICG, respectively.

2.17. In vivo anti-tumor therapy

The total mice were divided into eight groups randomly as tumor grew to ∼80 mm³, namely (1) PBS, (2) PI, (3) PI NIR, (4) DOX⋅HCl, (5) PID, (6) PID NIR, (7) PIDA and (8) PIDA NIR. Mice were given intravenously with all the group mentioned above on Day 0, 2, 4 (4 mg/kg DOX⋅HCl and 12 mg/kg ICG). For irradiation groups, mice were exposed to NIR irradiation using 808 nm laser (5 min, 0.5 W/cm²) on Day 1, 3, 5. The size of tumor and body weight of mice were monitored every other day. Tumor volume was determined by the formula as follows: Tumor volume (mm³) = length × width × width /2. Mice were sacrificed after 16 d. At the same time, tumors and organs were collected and analyzed by H&E staining.

2.18. Statistical analysis

Descriptive statistics were shown as means ± SD. Statistical significance between data was calculated by the Student's t-test via GraphPad Prism 9 (GraphPad Software, CA, USA). The level of statistical significance was set as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

3. Results and discussion

3.1. Synthesis and characterization of PEP and mPEG-ICG

Based on the experience of regulating the PEG weight ratio to guarantee the self-assembly of amphiphilic graft polyphosphazenes into nanovesicles [31,41,43], we synthesized PEP at the optimal feed ratio of mPEG2k to EAB via a sequential substitution reaction. The structure of the PEP copolymer was characterized using ¹H NMR (Fig. S1). According to the integrated area of the methyl proton of EAB (1.3 ppm) and the methyl ether proton peak of mPEG2k (3.3 ppm), the weight ratio of mPEG2k was calculated as 0.67, which was in accordance with our previous reports for nanovesicle formation.

PEGylated ICG (mPEG-ICG) was synthesized under optimal conditions [39]. In short, the amidation reaction between ICG-COOH and the amine groups of mPEG2k-NH₂ was carried out to synthesize mPEG-ICG. The successful synthesis of mPEG-ICG was confirmed by ¹H NMR (Fig. S2). The conjugation ratio was determined to be 75% according to the integrated area ratio of two peaks (3.3 ppm for - OCH₃ and 1.9 ppm for -CH₃).

3.2. Drug loading and characterization of nanoparticles

PEP nanoparticles were simply prepared via the dialysis method. Under TEM observation, PEP nanoparticles exhibited the typical structure of nanovesicles, which were equipped with a light hollow chamber and a deep distinct periphery (Fig. 1A). After introducing 23.1 wt% mPEG-ICG (PI nanovesicle), the typical vesicle morphology was maintained, which indicated that the insertion of mPEG-ICG had no influence on vesicle formation (Fig. 1B). Compared with water-soluble ICG sold on the market, mPEG-ICG is an amphiphilic macromolecule comprising the mPEG segment and the hydrophobic part of ICG. The good compatibility of PEP and mPEG-ICG was further verified through DSC detection, which showed that the curve of PI exhibited only one melting peak at 44.26 °C but not the individual melting peaks of PEP and mPEG-ICG (Fig. S3). Therefore, mPEG-ICG could be inserted into the membranes of PEP nanovesicles through hydrophobic interactions between the polycyclic benzoindotricarbocyanine part of mPEG-ICG and EAB in PEP. The DOX⋅HCl-loaded PI (PID) and DOX⋅HCl/L-Arg-coloaded PI (PIDA) were also obtained via the dialysis method. According to Fig. 1C, the center of the nanovesicles darkened after DOX⋅HCl loading, indicating that DOX⋅HCl was distributed in the hydrophilic hollow cavities, which was also in accordance with our previous work [41,[43], [44], [45]]. Moreover, the NO donor L-Arg could also be integrated together into the aqueous lumen, as shown in Fig. 1D. The average hydrodynamic sizes of PEP, PI, PID and PIDA were 217.4 ± 7.5 nm, 193.1 ± 5.0 nm, 206.0 ± 9.4 nm and 154.6 ± 8.5 nm, respectively (Fig. S4). The ζ-potentials of various nanoparticles were almost in the range of −10 to −20 mV (Fig. 1E). To confirm the integration of mPEG-ICG and coencapsulation of DOX⋅HCl and L-Arg, the absorbances of mPEG-ICG in PI, PID and PIDA were first investigated in the UV‒vis spectra and all showed a maximum absorbance at approximately 780 nm (Fig. 1F). With the aid of a standard curve at 780 nm (Fig. S5A), the LCs of mPEG-ICG were calculated to be 22.56%, 20.78% and 18.19%, respectively. For DOX⋅HCl and L-Arg, the standard curves were measured at 480 and 530 nm. (Fig. S5B, C) The LCs of DOX⋅HCl of PID and PIDA were 8.66% and 8.62%, respectively, and the LC of L-Arg of PIDA was 10.21%. It is noteworthy that the LCs of both DOX⋅HCl and L-Arg were rather high compared with many other works. As reported, the optimal DOX concentration in Arg-Ag@Cu-DOX was ∼0.4236 mg/ml [46], while the DOX·HCl concentration in PIDA could reach as high as ∼0.5974 g/ml. In another reference [47], the LC of DOX in UCNPs(DOX)@CS-RBS was only 4.10% ± 0.29% with a theoretical loading content of 10%. L-Arg-encapsulation was reported to have ∼7% loading capacities in _AAu_1–4@CuS YSNPs and L-Arg@Ce6@P [8,48]. Hence, this PEP-based nanovesicle platform exhibited superior coloading capacities for both DOX·HCl and L-Arg.

Fig. 1.

TEM images of (A) PEP, (B) PI, (C) PID and (D) PIDA nanoparticles (scale bar = 200 nm). (E) Zeta potentials of PEP, PI, PID and PIDA nanoparticles. (F) UV‒vis spectra of PEP, DOX⋅HCl, mPEG-ICG, PI, PID and PIDA.

The in vitro release profile of DOX⋅HCl from PIDA was investigated under different pH values (7.4, 6.5 and 5.5) and calculated with standard curves (Fig. S6). As a comparison (Fig. 2A), a burst release of free DOX⋅HCl occurred at pH 7.4 (44.8% within 5 min and 86.4% after 3 h), which was accelerated at pH 5.5 because of the enhanced solubility of DOX⋅HCl at lower pH. After encapsulation in PEP to form a DOX·HCl-loaded PEP nanovesicle (PD), DOX·HCl showed a controlled release behavior, indicating the superiority of the nanocarrier PEP (Fig. S7A). Upon the integration of mPEG-ICG, the DOX·HCl release from PID (Fig. S7B) and PIDA (Fig. 2B) at pH 7.4 was further inhibited. After 24 h, 12.8% DOX·HCl was released from PIDA at pH 7.4, while 74.7% was released at pH 5.5. For mPEG-ICG, less than 5% mPEG-ICG was released from PI, PID and PIDA after 48 h (Fig. S8). The in vitro release of L-Arg in PIDA could hardly be detected by the detection reagent. From the results above, we speculated that mPEG-ICG, which was integrated into the membrane of PEP, acted as a plug to inhibit the leakage of the content within the nanovesicles.

Fig. 2.

The release curves of (A) free DOX⋅HCl and (B) DOX⋅HCl from PIDA at pH 5.5, 6.5 and 7.4. (C) UV‒vis spectra of the PIDA and DPBF mixed solutions after exposure to NIR irradiation (0.5 W/cm²) for different times. (D) Amounts of NO detected in the mixed solution of mPEG-ICG and l-Arg after irradiation for 1, 2, 3, 4 and 5 min. (E) NO release curves for PIDA under different power densities (0, 0.3, 0.5 and 1.0 W/cm²) for 5 min. (F) Photothermal curves and (G) infrared thermographic images for various groups under NIR irradiation (0.5 W/cm², 5 min).

3.3. ROS generation

ICG, a typical type II photosensitizer, produces ROS mainly in the form of singlet oxygen (¹O₂) through the interaction between excited ICG and molecular oxygen [49,50]. To determine the mPEG-ICG-generating ¹O₂, DPBF was used as the proper indicator [51,52]. The furan ring in DPBF could be irreversibly destroyed by ¹O₂; therefore, ¹O₂ generation could be monitored by the decay behavior of DPBF at 410 nm in the spectra. First, the ¹O₂ generation ability of mPEG-ICG was investigated. The DPBF absorbance at 410 nm showed a gradual decrease as the time exposed to NIR light was extended (Fig. S9A), which revealed the excellent ¹O₂ generation ability of mPEG-ICG. Then, the ¹O₂ generation capacities of the various nanovesicles integrated with mPEG-ICG (PI, PID and PIDA) were explored. As shown in Fig. S9B and S9C, PI and PID with the same concentration of mPEG-ICG exhibited similar characteristics of ¹O₂ generation. Fig. 2C shows the PIDA spectra at different time points after exposure to NIR light (0.5 W/cm²). An attenuation in the DPBF absorbance at 410 nm could be detected, which indicated the ¹O₂ generation ability of PIDA nanoparticles and provided powerful ROS resources for NO generation.

3.4. NO release in vitro

The NO release behavior of PIDA nanoparticles was investigated via the Griess reagent and calculated based on a standard curve (Fig. S10). It was documented that ROS with high oxidation ability could oxidize L-Arg-molecules to generate NO [53,54]. Based on the results of time-dependent ROS generation, we first investigated the amount of NO in the mixed solution of L-Arg and mPEG-ICG with different irradiation times. As shown in Fig. 2D, the amount of NO increased when prolonging the time of exposure to NIR, which coincided with the gradual generation of ROS at different time points. Afterward, the NO release from PIDA was studied. Fig. 2E shows that NO could not be produced without NIR irradiation. As the power density of NIR light increased from 0.3 W/cm² to 0.5 W/cm², the ultimate amount of NO increased from 16.1 to 19.6 µM. The density-dependent release behavior could be largely ascribed to the amount and speed of ROS generation under different power densities (Fig. S11). Furthermore, as the release of mPEG-ICG and L-Arg from PIDA was low at pH 7.4 or pH 5.5 (Fig. S8), the NO generation at pH 5.5 under different laser power densities was almost the same as that in H₂O (Fig. S12). Considering the amount of NO production and the effect of laser power on skin in biological studies, an optimal power density of 0.5 W/cm² was designated for the next experiments.

3.5. In vitro photothermal effect

The in vitro photothermal effect was evaluated under 808 nm laser irradiation at 0.5 W/cm² for 5 min. As shown in Fig. 2F, mPEG-ICG reached ∼37 °C under NIR irradiation. The temperatures of PI, PID and PIDA reached 40 °C in 5 min, demonstrating that the integration of mPEG-ICG and the encapsulation of both DOX⋅HCl and L-Arg by PEP had no influence on the photothermal properties. As shown in Fig. 2G, we could directly observe a steady increase in temperature in mPEG-ICG-containing formulations. The photostabilities of PI, PID and PIDA were further studied by recording the temperature of the solutions for three successive cycles of heating/cooling processes. As illustrated in Fig. S13- S16A, the variations in mPEG-ICG, PI, PID and PIDA at the maximum temperature during each cycle were less than 5%; thus, they all showed excellent photostability within three cycles. Moreover, the photothermal transduction efficiencies of mPEG-ICG, PI, PID and PIDA were calculated by the cooling period as 3.26%, 7.04%, 9.66% and 13.80%, respectively. The enhanced photothermal performances of PI, PID and PIDA could be largely explained by the morphology of the nanovesicles to retain heat [55,56].

3.6. Intracellular generation of NO

The NO generation of PIDA in K562/ADR cells was monitored using a commercial NO-sensitive DAF-FM DA probe, which could trap NO and then produce triazole compounds with strong green fluorescence in cells. After staining with DAF-FM DA probes, the production of NO in cells was observed by CLSM. In detail, K562/ADR cells incubated with PI, PID and PIDA with or without NIR irradiation were set as experimental groups, with PBS- and DOX⋅HCl-treated cells as the control groups. As observed in Fig. 3A, there was no NO generation in the control, PI, PID or PIDA groups without irradiation. Only the cells treated with PIDA exposed to NIR irradiation exhibited obvious and strong green fluorescence. It was reported that L-Arg is capable of being transferred to NO via NOSs, but the amount of the produced NO was quite limited since NOSs are scarcely expressed in cancer cells. Therefore, NIR irradiation was necessary for the oxidation of L-Arg by the new-born ROS of mPEG-ICG as well as the efficient generation of NO in cells.

Fig. 3.

(A) CLSM images of Hoechst 33342 and DAF-FM DA-stained K562/ADR cells. (B) CLSM images of TMRM-stained K562/ADR cells. (C) ATP and (D) ROS levels of K562/ADR cells coincubated with different groups. (E) Cellular DOX·HCl contents analyzed by flow cytometry for K562/ADR cells treated with different groups at different time points. The scale bars represent 50 µm.

3.7. Mitochondria-dependent pathway

It has been reported that NO can promote cell apoptosis by blocking mitochondrial respiration, inducing nitration of important proteins, triggering the transformation of mitochondrial permeability and stimulating the mitochondria-associated pathway of apoptosis. The sensitization role of NO in chemotherapy has also been validated by a large number of studies on mitochondrial dysfunction caused by NO [40,57,58]. In this regard, we investigated the effect of NO on mitochondrial activity.

It is widely recognized that moderate NO levels can acutely induce ROS production from mitochondria; moreover, ROS accumulation can cause cell damage and physiological dysfunction. First, we used the TMRM probe to measure the mitochondrial membrane potential and evaluate the integrity of mitochondria. The strong red fluorescence of TMRM was observed in the normal-state hyperpolarized mitochondria, while weak fluorescence could be detected if the mitochondrial membrane was depolarized. As visualized by TMRM in Fig. 3B, red fluorescence was barely noted in PIDA NIR-treated cells, clearly demonstrating not only the precipitous decrease in mitochondrial membrane potential but also the destruction of membrane integrity.

The ATP level in K562/ADR cells was then characterized because ATP acted as the direct energy source in cell metabolism, the amount of which acted as direct evidence for mitochondrial dysfunction. When cells were incubated with PI, the ATP level remained as high as that in untreated K562/ADR cells. Nevertheless, samples underwent significant decreases in the level of ATP among the DOX⋅HCl-containing groups, and the PIDA NIR-treated group realized the most remarkable decrease in ATP levels (Fig. 3C).

Next, the intracellular production of ROS was investigated using a DCFH-DA assay. As shown in Fig. 3D, DOX⋅HCl and PI made little difference in the level of ROS compared with the blank. For the PI with NIR irradiation group, the ROS level in cells was enhanced due to the photodynamic characteristics of mPEG-ICG and the probable oxidative stress caused by PTT and PDT damage, also indicating the efficient cellular uptake of the nanoparticles. The PID NIR group that integrated chemotherapy, PDT and PTT together showed a significant increase in ROS compared with the DOX⋅HCl group. However, PIDA NIR with NO-releasing ability proved even more capable of increasing ROS levels than PID NIR. The differences between the PID NIR group and the PIDA NIR group could be largely ascribed to the fact that NO could efficiently increase ROS generation.

All the data confirmed that NO served as a powerful regulator of mitochondrial functions and efficiently induced depletion of mitochondrial membrane potential and increases in ROS and ATP reduction.

3.8. Cellular uptake

P-gp is an important cell surface protein whose main function is to reduce cytotoxicity and maintain normal cell function by using energy generated by ATP hydrolysis to transfer various toxic molecules out of the cell [59]. Once the ATP level is reduced, P-gp function is impaired, and the intracellular DOX·HCl level increases, inhibiting drug efflux and ultimately reversing drug resistance in tumors. Herein, DOX·HCl accumulation within K562/ADR cells was investigated. First, as shown in Fig. S17, the amount of mPEG-ICG internalized by cells increased slightly within 24 h, and there was no obvious difference among PI, PID and PIDA, indicating the efficient uptake of the three nanovesicles. However, after treatment for 8 h, the intracellular DOX⋅HCl accumulation in the PIDA NIR group was significantly higher than that in the PIDA group (Fig. 3E). After ATP depletion by NO, there was no efficient energy available for P-gp, thus reducing drug efflux and reversing drug resistance.

3.9. Cytotoxicity and cell apoptosis

In vitro proliferation inhibition of K562/ADR cells treated with different groups was carried out via CCK-8 assay. As depicted in Fig. S18, mPEG-ICG barely affected cell proliferation with or without NIR irradiation, nor did PI nanovesicles, indicating that mPEG-ICG-derived MPTT and PDT resulted in little effect on cell growth inhibition under this laser irradiation condition (0.5 W/cm², 5 min). Meanwhile, the viability of K562/ADR cells declined slowly after incubation with DOX⋅HCl, indicating the resistance of K562/ADR cells to DOX⋅HCl. However, the PID group showed enhanced cytotoxicity compared with the free DOX⋅HCl group, which could be interpreted by the excellent cellular uptake efficiency provided by the polymeric nanocarriers. The cytotoxicity of the PID NIR group was much better than that of the PID group due to the combination of chemotherapy, PTT and PDT. Interestingly, PIDA with 808 nm laser irradiation exhibited obvious toxicity compared with other groups, which could be mainly ascribed to the combination of NO with the multimodal treatments. To investigate the difference in cytotoxicity between various formulations, the IC₅₀ values were calculated in K562/ADR cells and are shown in Fig. 4C. Among all the groups, the PIDA NIR group had a significantly enhanced killing effect with an IC₅₀ value of 5.54 µg/ml compared with the values of DOX⋅HCl (20.64 µg/ml) and PID NIR (8.38 µg/ml), indicating the sensitization role of NO to chemotherapy. Moreover, the reversal indices (RIs) for PID NIR and PIDA NIR were 2.46 and 3.73, respectively, confirming the crucial effect of NO in overcoming drug resistance. Fig. 4D also revealed the proapoptotic effect of PIDA on K562/ADR cells. The PIDA NIR group effectively induced 28.71% apoptosis compared with 6.84% in the DOX⋅HCl group, 16.02% in the PIDA group and 17.11% in the PID NIR group, confirming the crucial roles of NO and synergistic therapy in the induction of cell apoptosis.

Fig. 4.

Viability of K562/ADR cells treated with (A) DOX⋅HCl, PIDA and PIDA NIR and (B) DOX⋅HCl, PIDA and PIDA NIR (0.5 W/cm², 5 min). (C) IC₅₀ values calculated by DOX⋅HCl against K562/ADR cells treated with different groups. (D) K562/ADR cell apoptosis after coincubation with different groups.

Live animal fluorescence imaging was utilized to study the tumor targeting and in vivo distribution of various nanovesicles in K562/ADR tumor-bearing mice via tail vein injections. As shown in Fig. 5A, strong mPEG-ICG fluorescence was detected in the tumor sites of mice 2 h postinjection and was maintained up until 24 h postinjection. After that, the fluorescence intensity within the tumor sites began to decrease, suggesting the metabolism and biodegradability of the nanovesicles. Meanwhile, the tumors and major organs (heart, liver, spleen, lung and kidney) were obtained 24 h postinjection from the sacrificed mice, and the biodistributions of DOX⋅HCl and mPEG-ICG were analyzed through their fluorescence signals. As depicted in Fig. 5B, the intratumoral DOX⋅HCl accumulation of DOX⋅HCl-loaded nanoparticles was significantly higher than that of the DOX⋅HCl group, indicating the excellent drug encapsulation and target delivery function of the nanovesicles. The PIDA group showed enhanced accumulation of DOX⋅HCl compared with the PID group, which was in accordance with the results of cellular uptake. The ICG distribution of different formulations was almost the same, which was consistent with the result of mPEG-ICG cellular uptake (Fig. S17). The fluorescence detected in the liver suggested that nanoparticles were mainly metabolized in the livers of mice. Briefly, PID and PIDA not only possessed prospective performance for fluorescence guidance but also enhanced drug accumulation in the desired site, demonstrating the great potential of the nanovesicles in multimodal therapy.

Fig. 5.

(A) In vivo fluorescence imaging of K562/ADR-bearing mice after intravenous injections of PI, PID and PIDA. (B) Ex vivo fluorescence images of tumors and major organs for DOX⋅HCl distribution and ICG distribution for various formulations.

3.11. In vivo antitumor efficacy

In vivo antitumor experiments were carried out on K562/ADR tumor-bearing nude mice. All nude mice were randomly divided into eight groups and treated with PBS, DOX⋅HCl, PI, PID and PIDA. As shown in Fig. 6A, PBS, DOX⋅HCl or different formulations were intravenously administered on Day 1, 3 and 5, and NIR irradiation was conducted on Day 2, 4 and 6. First, the mild photothermal effects of PI, PID and PIDA were verified by an infrared thermal camera. The temperature changes within the tumor site were recorded 24 h postinjection. Fig. 6B indicates that the temperatures of the tumor sites increased to ∼42 °C within 5 min, which was beneficial to the ablation of the tumor. As shown in Fig. 6C, tumors grew markedly in the PBS and PI groups, while the tumor growth in the PI NIR and DOX⋅HCl groups was suppressed, with tumor inhibition rates (TIRs) of 25.6% and 33.1%, respectively. In contrast, the treatments of the PID and PIDA groups suppressed tumor growth, with TIR values of 53.4% and 65.8%, respectively, which were analogous to the results reported previously [60]. Taking advantage of the combination of MPTT, PDT and chemotherapy, the PID NIR group clearly exhibited much stronger inhibition of tumor growth, with a TIR of 71.1%. In addition, we found that the PIDA NIR groups displayed the most significant anticancer efficacy, with a TIR of 80.8%, showing that NO generated from the nanoparticles could enhance the therapeutic efficacy of multimodal therapy.

Fig. 6.

(A) Drug administration and NIR irradiation protocol of the antitumor investigation. (B) Infrared thermal images of the tumors in mice under NIR irradiation. (C) Tumor volume and (D) body weight changes in each group over the treatment period. (E) H&E staining of tumors from different groups (scale bar = 100 µm).

To detect the systemic toxicity of various formulations, the body weights of nude mice were recorded throughout the treatment period. As displayed in Fig. 6D, mice treated with DOX⋅HCl experienced drastic body weight loss. Fortunately, this problem was alleviated by encapsulation of DOX⋅HCl using PEP, which resulted in less weight loss in the PID and PIDA groups. Additionally, the safety profile was confirmed by H&E staining of tumors and organ tissues (Fig. S19). The DOX⋅HCl-loaded polymersomes successfully avoided the toxicity toward the heart induced by DOX⋅HCl, and the major organs remained normal. These results indicated that PIDA exhibited safe and efficient cancer therapy in vivo.

4. Conclusion

An NIR-stimulated NO-releasing PIDA nanovesicle was developed to incorporate the photosensitizer mPEG-ICG, the NO donor L-Arg and the chemotherapeutic drug DOX·HCl via amphiphilic PEP copolymers and further exerted multimodal anticancer therapy. Both in vitro and in vivo experiments demonstrated the vigorous effect of PIDA on drug resistance reversal. These advantages demonstrate that PIDA offers huge potential in synergistic anticancer therapy.