NH₂-PLGA-NH₂ (lactide:glycolide, 50/50, MW: 8 kDa) was obtained from Xi’an Rui-Xi Biological Technology Co. Ltd. (Xi’an, Shaan Xi, China). 1-Butyl-3-methylimidazolium-L-lactate (BML, MW: 228.29 Da) was provided by Cheng Jie Chemical Co. Ltd. (Shanghai, China). Hyaluronic acid (HA, MW: 3.8 kDa) was purchased from Dalian Meilun Biotechnology Co. Ltd. (Dalian, Liao Ning, China). Paclitaxel (Ptx, MW: 853.93 Da) was purchased from Shanghai Aladdin Biotechnology Co. Ltd. (Shanghai, China). 1,2-Dipalmitoyl-sn-glycerol-3-phosphocholine (DPPC, MW: 734.1 Da), polyvinyl alcohol (PVA, MW: 30–70 kDa), 2-(N-morpholino) ethane sulfonic acid (MES), N-hydroxysuccinimide (NHS), and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) were obtained from Sigma-Aldrich Corporation (St. Louis, MO, USA). All reagents used in this work were of analytical grade without further purification.

The double emulsification method was used to synthesize Ptx and BML-loaded DPPC-PLGA hybrid nanoparticles (BNPs@Ptx) following a previous report (Xu et al., 2019b). Briefly, 10 mg of DPPC, 20 mg of NH₂-PLGA-NH₂, and 2 mg of Ptx were thoroughly dissolved in 2 ml of dichloromethane (DCM). BML, 400 µl (0.5 mg/ml, in deionized water), was mixed in the solution. The Ptx and BML solution was emulsified in an ice bath (2 min, with a 5-s on–off duty cycle) using an ultrasonic oscillation instrument (SCIENTZ-IID, Ningbo Scientz Biotechnology Co., Ltd., Ningbo, Zhejiang, China) at 12% power (24 kHz, 600 W). PVA, 8 ml (2% w/w, in water) solution, was further added to the emulsion and emulsified as before. DCM was then evaporated in a fume hood for 3 h under magnetic stirring. Finally, BNPs@Ptx was collected by centrifugation at 10,000 × g for 10 min and washed twice with double distilled water. Nanoparticles targeting CD44 were prepared by covalently coupling hyaluronic acid to BNPs@Ptx via amino groups on PLGA (Alam et al., 2017). Briefly, 2 mg of HA was dissolved in 195 mg of MES buffer (pH = 5.5). EDC (27 mg) and 8.6 mg of NHS were then added to the HA solution. Subsequently, the mixture was incubated with 1 ml of BNPs@Ptx suspension under continuous stirring for 24 h on ice bath. HA-BNPs@Ptx was collected and purified by centrifugation and washed with double distilled water two times.

To examine cellular uptake and intracellular trafficking, BNPs@Ptx and HA-BNPs@Ptx were labeled with a red fluorescence probe Dil. Endocytosis/phagocytosis experiments were performed in 4T1 cells and J774 cells (murine macrophages), respectively. 4T1 cells per well (1 × 10⁴) were seeded in confocal dishes. After 24 h, 100 µl of DiI-BNPs@Ptx and DiI-HA-BNPs@Ptx was added to co-incubate with 4T1 cells for 1, 3, and 6 h, respectively. At each time point, 20 µl of Lyso Tracker Green was added to each confocal dish to stain endo/lysosomes. One and a half hours after incubation with Lyso Tracker Green, all cells were fixed with 4% paraformaldehyde for 10 min. 4,6-Diamidino-2-phenylindole (DAPI, blue, 10 µl) was finally added in the fixed cells to stain the nuclei. Confocal fluorescence images of the fixed cells were obtained by a laser scanning confocal microscope (LSCM, A1R + MP, Nikon Co., Tokyo, Japan).

J774 cells were treated in the same way as 4T1 cells, except that 5 × 10⁴ cells and 10 µl of PMA (100 ng/ml) were added to each dish and incubated for 48 h.

Synergistic anti-cancer activity of thermal chemotherapy was evaluated with 4T1 cells in vitro. The cells were seeded into a six-well plate (7.5 × 10⁵ cells per well) for 24 h and then incubated with 1) PBS, 2) HA-BNPs@Ptx, 3) MW, 4) BNPs@Ptx + MW, an5) HA-BNPs@Pd tx + MW for 6 h. In order to optimize the thermal effects of nanoparticles, HA-BNPs@Ptx/4T1 cells were further exposed to MW irradiation at different power (0.4, 0.8, 1.2, and 1.6 W cm⁻²) for 1, 2, 3, and 4 min, respectively (see Supplementary Material, Methods 1.3 for details). After incubation, the cells of the last three groups were exposed to MW irradiation (0.8 W cm⁻², 4 min). The anti-cancer activity of each group was qualitatively examined by using calcein-AM/propidium iodide (PI) double stain kit (Beyotime Biotechnology® Inc., Suzhou, Jiangsu, China). After LIVE/DEAD staining of the cells, all cell samples were imaged under an inverted fluorescence microscope.

To quantitatively assess the synergistic chemo-thermal therapy, 4T1 cells were stained by Annexin V-(FITC)/PI apoptosis detection kit (4A Biotech Co., Ltd., Peking, China) and analyzed by flow cytometer.

BALB/c mice (female, 18–20 g) were provided by Dashuo Biological Technology (Chengdu, Sichuan, China). All processes were in accordance with the Chinese Society of Laboratory Animals on animal welfare and approved by the Animal Use and Care Management Advisory Committee of West China Hospital of Sichuan University (Approval No. 2017014A). 4T1 cells (1 × 10⁶/wells) were implanted into the second axillary mammary fat pad on the right side. When the volume of tumor reached about 100 mm³, 4T1 breast tumor-bearing mice were randomly assigned into four groups (six mice per group): 1) DiI-BNPs@Ptx, 2) DiI-HA-BNPs@Ptx, 3) MW + DiI-BNPs@Ptx, and 4) MW + DiI-HA-BNPs@Ptx to compare the effect among non-targeted (EPR), CD44-targeted (active targeting), and MW irradiation (physical targeting). MW + DiI-BNPs@Ptx and MW + DiI-HA-BNPs@Ptx were treated with mild hyperthermia (4-min MW exposure at 0.8 W cm⁻²). Then nanoparticles (0.5 μl/g, each microliter of nanoparticle suspension contained 0.434 μg of Ptx and 1.45 μg of nanoparticles) were injected into 4T1 breast tumor-bearing mice via tail vein. In vivo fluorescence images were collected before and after injection of nanoparticles at 1, 3, 6, and 24 h using the IVIS Spectrum system (Lumina XR, Caliper Life Sciences, Boston, MA, USA). To examine the biodistribution of nanoparticles, the main organs (heart, liver, spleen, lungs, and kidneys) and tumors were isolated from mice to perform ex vivo imaging using the same IVIS Spectrum.

Blood chemistry analysis of BALB/c mice was made 7 days after intravenous injection with HA-BNPs@Ptx (Supplementary Material, Methods 1.4 for details).

The first MW irradiation aimed to generate mild hyperthermia to alter the TME to increase the uptake of nanoparticles. Six mice in each group received mild hyperthermia at the tumor site with a MW power of 0.8 W cm⁻². The temperature of tumor was monitored in real time using an infrared thermal mapping instrument. After mild hyperthermia, DiI-HA-BNPs@Ptx (0.5 μl/g, each microliter of nanoparticle suspension contained 1.45 μg of nanoparticles) was immediately injected into 4T1 breast tumor-bearing mice via tail vein. The mice were euthanized, and the tumor tissue was removed after 24 h. The retention of nanoparticles (red fluorescence) in tumor tissues was revealed using a pathological section scanner Pannoramic DESK (P-MIDI-P250, 3D HISTECH, Budapest, Hungary).

To further explore changes in TME, intratumoral perfusion and micro-vessel density of tumor were evaluated using contrast-enhanced ultrasound imaging (CEUS) and CD31 immunohistochemical analysis. Microbubbles (0.2 ml/kg, SonoVueTM, Bracco, Italy) were injected via tail vein 24 h after the first MW irradiation. CEUS was performed using an ultrasound scanner (iU22, Koninklijke Philips N.V., Eindhoven, Netherlands) with a 12- to 5-MHz transducer. Subsequently, mice were euthanized, and tumors were sectioned and stained with CD31 (dilution 1:50, Wuhan Service Bio Co., Ltd., Wuhan, Hubei, China) to evaluate the micro-vessel density. The pathological sections were imaged using a fluorescence microscope.

The second MW irradiation was dedicated to activate the sensitizer BML and to release the chemotherapy agent Ptx in the nanoparticles. Thirty 4T1 breast tumor-bearing mice were divided into five groups: PBS (control group, G1) (Tajbakhsh et al., 2018), first MW + HA-BNPs@Ptx (G2) (Thakur and Kutty, 2019), first MW + HA-BNPs + second MW (G3) (Mu et al., 2017), HA-BNPs@Ptx + second MW (conventional SDDS, G4) (Ding et al., 2021), and first MW + HA-BNPs@Ptx + second MW (G5) (Hu et al., 2020). After treatment with mild hyperthermia (4-min MW exposure at 0.8 W cm⁻², G2, G3, and G5), PBS, HA-BNPs, or HA-BNPs@Ptx (0.5 μl/g, each microliter of nanoparticle suspension contained 0.434 μg of Ptx) was immediately injected into 4T1 breast tumor-bearing mice intravenously. Twenty-four hours after injection, the mice received a second MW irradiation (0.8 W cm⁻²) for 4 min. The tumor temperature was monitored in real time by infrared thermal mapping instrument. The body weight and tumor volumes were recorded every 3 days. The tumor suppression rate (TSR) was calculated using the following formula: TSR ( % ) = ( Vc - Vx ) / Vc × 100 %

where Vc is the volume of the PBS group, and Vx is the volume of the treatment group.

On day 18 after treatment, tumors and main organs were excised and fixed overnight in 10% buffered formalin. The tumors were sectioned and stained with H&E, Ki-67 antibody, and TdT-mediated dUTP nick-end labeling (TUNEL) staining. The main organs were sectioned and stained with H&E. The expression level of CD3⁺, CD4⁺, and CD8⁺ of immune cells in tumor tissues were evaluated by immunofluorescence. The slides of tumor tissues were incubated with Anti-CD3 Rabbit pAb (P22646,1:700), Anti-CD4 Rabbit pAb (P06332 1:800), and Anti-CD8 Rabbit mAb (P10966, 1:500), following the standard procedure of Wuhan Service Bio Co., Ltd. (Wuhan, Hubei, China). The nucleus was labeled with DAPI. The images were obtained by using the pathological section scanner Pannoramic DESK.

In addition, peripheral blood was collected from treated mice before euthanasia to analyze the percentages of tumor antigen-specific CD3⁺, CD4⁺, and CD8⁺ T cells. Serum cells were incubated with 3 µl of FITC anti-mouse CD3 antibody, 5 µl of PE anti-mouse CD4, and 10 µl of APC anti-mouse CD8 (4A Biotech Co., Ltd., Peking, China) for 1 h at room temperature, respectively. Finally, the cells were resuspended in PBS and analyzed by flow cytometry and Flow Jo software.

All data were expressed as mean ± standard deviation (SD). The statistical analysis was carried out with GraphPad Prism Version 8.0 software (GraphPad, USA). Comparisons among multiple groups were performed by one-way analysis of variance (ANOVA). Two-group comparisons were performed by Student’s t-test. Statistical significance was indicated by * for p < 0.05, ** for p < 0.01, and *** for p < 0.001.

Figure 2A illustrates the preparation procedure of nanoparticles HA-BNPs@Ptx. TEM and SEM images showed a core–shell structure of the nanoparticles with a mean diameter of about 150 nm, spherical shaped, and with a smooth surface (Figure 2B). DLS revealed that HA-BNPs@Ptx was 203.30 ± 7.51 nm and relatively homogenous (PDI: 0.172, Figure 2C). HA-BNPs@Ptx was negatively charged with zeta potential of −22 mV (Figure 2D) and stable at neutral pH (Supplementary Figures S1A,B). The vibration absorption peaks of amide bonds at 1,656, 1,546, and 1,334 cm⁻¹ in FTIR spectrums confirmed the successful conjugation of HA to BNPs@Ptx nanoparticles (Figure 2E). Ptx loading capacity of HA-BNPs@Ptx determined by spectrophotometer was 13.17 wt.%, corresponding to encapsulation efficiency of 85.62 wt.% (Figure 2F).

In order to optimize the MW thermal effect and Ptx release, nanoparticle suspensions (HA-NPs@Ptx and HA-BNPs@Ptx) were exposed at different MW power and time. The infrared thermal imaging pictures showed that under the same irradiation, the heating effect of HA-NPs@Ptx without BML (Figures 3A,B) was significantly lower than that of HA-BNPs@Ptx with the sensitizer (Figures 3C,D). For HA-BNPs@Ptx, the temperature of the suspension increased rapidly from room temperature to 35.2°C (0.4 W cm⁻²) and reached nearly 60°C (1.6 W cm⁻²) under 4-min MW exposure. Due to the lack of BML, the temperature of HA-NPs@Ptx increased only to 30.4°C at 0.4 W cm⁻² and 41.8°C at 1.6 W cm⁻². These results suggest that HA-BNPs@Ptx has a greater potential to trigger Ptx release in combination with thermal chemotherapy.

Figures 3E,F show the release of Ptx from HA-BNPs@Ptx upon MW irradiation. The Ptx solution standard curve was established according to previous reports to quantify Ptx (Supplementary Figure S2) (Furman et al., 2017). A maximum release of 70% was obtained at 1.6 W cm⁻² for 4 min. At this condition, the temperature was very high (58°C, Figure 3D), and this would directly induce tumor necrosis (Seynhaeve et al., 2020; Pan et al., 2021). In order to avoid excessive heat and demonstrate our synergistic cascade strategy, a lower MW power of 0.8 W cm⁻² was chosen for subsequent in vitro and in vivo studies.

Hyaluronic acid, as a CD44-specific ligand, is widely used in various targeted drug delivery systems for TNBC (Zhang et al., 2020; Sun et al., 2019). The expression of CD44 on 4T1 cells was verified (Supplementary Figure S3A), and the specific targeting ability of nanoparticles to 4T1 cells was compared between HA-BNPs@Ptx and BNPs@Ptx (Supplementary Figure S3B). Supplementary Figure S3A reveals that the nanoparticles contain HA bonded fairly to 4T1 cells, and the binding was specific because the binding could be inhibited by adding the free hyaluronic acid to 4T1 cells before the incubation with the nanoparticles (Cerqueira et al., 2017). Non-specific binding could be observed with BNPs@Ptx, but the number of fluorescent nanoparticles was significantly less than that of HA-BNPs@Ptx, suggesting that the specific binding of HA and CD44 could be used to improve the cellular internalization efficiency of nanoparticles (Supplementary Figure S3B).

The cellular uptake of HA-BNPs@Ptx was evaluated with 4T1 cells and J774 cells. The cell nuclei were stained by DAPI (blue), the endo/lysosomes were stained by LysoTracker Green (green), and the nanoparticles were labeled with DiI (red). Figure 4A shows that BNPs@Ptx and HA-BNPs@Ptx were taken up both by 4T1 and J774 cells, and stayed in endo/lysosomes. However, HA-BNPs@Ptx leads to higher fluorescent intensity compared with BNPs@Ptx. Furthermore, the uptake of HA-BNPs@Ptx by 4T1 cells was quantified using flow cytometry. A maximum uptake was observed after 6 h of incubation (Supplementary Figure S4). A lower fluorescent intensity obtained at 24 h suggests probably a partial degradation of nanoparticles in endo/lysosomes.

The synergistic antitumor effect of MW thermal chemotherapy was studied qualitatively by confocal microscope and quantitatively by flow cytometry. In vitro hyperthermia experiments indicate that MW irradiation alone was sufficient to induce apoptosis of 4T1 tumor cells (Supplementary Figure S5), and the effect increased with increasing exposure power and time. Figure 4B shows that the combination of MW and HA-BNPs@Ptx resulted in the most significant cell damage compared with the other tested conditions (PBS, BNPs@Ptx, MW, MW + BNPs@Ptx). The 4T1 cell apoptosis analyzed by flow cytometry presented similar results (Figures 4C,D). The apoptosis was less than 8% for the control group (PBS), 32% for HA-BNPs@Ptx alone, 35% for MW irradiation only, 55.6% for MW + BNPs@Ptx, and 87% for MW + HA-BNPs@Ptx. These results suggest that there was a synergistic effect by combining hyperthermia and Ptx-loaded nanoparticles, and also confirm the superiority of CD44-targeted nanoparticles compared with non-targeting nanoparticles.

The biocompatibility of HA-BNPs@Ptx and HA-NPs was investigated by cell counting kit8 (CCK8) assay with HUVEC and 4T1 breast cancer cells. A high cell viability (>92%) was obtained after incubation with HA-BNPs@Ptx and HA-NPs (without Ptx and BML) for 24 h for the two cell lines (Supplementary Material, Supplementary Figure S6). Supplementary Figure S7 shows that the hemolysis rate induced by HA-BNPs@Ptx is less than 5%. This confirmed that the nanoparticles had good biocompatibility and also showed good stability, with no leakage of Ptx during 24-h incubation (Nave et al., 2016).

To verify the targeting ability and biodistribution of HA-BNPs@Ptx, 4T1 breast tumor-bearing mice model was established in the second axillary mammary fat pad on the right side. HA-BNPs@Ptx and BNPs@Ptx were injected intravenously into 4T1 breast tumor-bearing mice, respectively. The targeting ability and biodistribution of nanoparticles were studied in real-time using a whole animal fluorescence imaging. The intensity of HA-BNPs@Ptx and BNPs@Ptx at tumor sites reached the maximum level at 24 h (Figure 5A). Compared with the non-targeted BNPs@Ptx group, the HA-BNPs@Ptx group exhibited significantly higher uptake in the tumor, almost two times higher (Figure 5B) due to the presence of HA active targeting. For untargeted BNPs@Ptx, the first MW exposure also improved significantly the EPR effect, the tumor uptake comparable with HA-BNPs@Ptx without the first MW exposure. The first MW + HA-BNPs@Ptx had the highest tumor retention rate, which is 1.7 times that of the first MW + BNPs@Ptx and HA-BNPs@Ptx, and almost four times that of BNPs@Ptx alone (EPR effect) both at 6 and 24 h after the injection of nanoparticles. Tumor tissues and major organs (heart, lung, kidney, liver, and spleen) were harvested for ex vivo imaging (Figure 5C). The fluorescence intensity of tumor tissues was also two times higher for the targeted nanoparticles than for non-targeted nanoparticles (Figure 5D). For other organs, comparable results were observed in the liver and spleen; a stronger fluorescence was observed in the lungs for HA-BNPs@Ptx, probably suggesting a higher CD44 expression in the lungs of the breast cancer model.

The purpose of the first MW irradiation was to pretreat animals to modify TME, so as to improve the tumor uptake of nanoparticles (Yang and Gao, 2017; Hynynen, 2018). In order to understand the mechanism of the pretreatment of MW prior to the injection of HA-BNPs@Ptx, the animal was continuously monitored by infrared thermography to control the temperature change in tumors. Figures 6A,B show that the temperature of tumors was 36.7 ± 0.2°C before the first MW irradiation and increased to 43.8 ± 0.4°C after 4 min of MW irradiation (0.8 W cm⁻²). After the treatment with MW, the nanoparticles DiI-HA-BNPs@Ptx were immediately injected into mice through the tail vein (∼0.5 μl/g). Twenty-four hours after treatment, the hyperthermia effect was measured by one of the following three methods. First, red fluorescence-labeled nanoparticles that accumulated at the tumor site were detected by tissue sections (Figures 6C,D). Second, contrast-enhanced ultrasound (CEUS) imaging was performed to measure the change in intratumoral blood flow (Figures 6E,F). Third, the morphology of tumor vasculature was examined using CD31 immunofluorescent staining (Figures 6G,H).

The results showed that the retention rate of nanoparticles in the tumors irradiated by MW increased by seven times compared with the tumors not irradiated (Figures 6C,D). The echo intensity of CEUS showed that the pretreatment of tumor and its surrounding tissue by MW irradiation improved significantly intratumoral perfusion (Figures 6E,F). CD31 immunofluorescence images also showed increased tumor vascular density and/or intratumoral vasodilatation inside the tumor compared with that of tumors not irradiated (Figure 6F). Figure 6G showed that 24 h after MW exposure, the percentage of effective tumor vessels increased by a factor of about fourfold, from 3.8 ± 1.5% to 15.3 ± 4.3%. These results suggest that the pretreatment of tumors by mild hyperthermia can improve the accessibility of nanoparticles to tumors, possibly by reducing interstitial fluid pressure, dilating and increasing tumor blood vessels, and thus improving intratumoral perfusion (Chen et al., 2019).

The synergistic cascade antitumor effect was achieved by combining two-step MW irradiations with target nanoparticles HA-BNPs@Ptx (Figure 7A). The tested conditions included group 1 (G1), PBS; group 2 (G2), first MW + HA-BNPs-Ptx; group 3 (G3), first MW + HA-BNPs + second MW; group 4 (G4), HA-BNPs-Ptx + second MW, and group 5 (G5) with first MW + HA-BNPs-Ptx + second MW. Notice the difference between G2 and G5; G2 did not receive a second MW irradiation (nanoparticles); the difference between G3 and G5 is that the former contained no Ptx (thermal therapy); the difference between G4 and G5 is that G4 did not receive the first MW irradiation (tumor pre-treatment). The group (G3, G4, and G5) that received a second MW was irradiated under the same conditions as the first one (0.8 W cm⁻² for 4 min), and the tumor temperature was monitored in real time by an infrared thermal mapping instrument. The exposed temperature of G3 and G5 under second MW irradiation were all up to 51°C (Supplementary Material, Figures 7B,C). G4 reached to approximately 45°C. This indicates that BML encapsulated in HA-BNPs-Ptx provided supplement heating.

Figure 7D shows that throughout the treatment period, there was no significant change in animal body weight in all groups, indicating that all treatment groups were well tolerated. Compared with the other groups, the tumor volume of the G1 group increased rapidly, up to about 2,000 mm³ (Figures 7E–G). The antitumor effect of the G5 group was the best, with the inhibition rate of tumor growth up to 88%. G2 and G3 inhibited tumor growth by 58.2% and 67.5%, respectively. G4 also has a good antitumor effect; the inhibition rate of tumor growth was about 72.5%.

The degree of tumor tissue damage was evaluated by immunostaining using H&E, proliferation marker Ki-67. and TUNEL essays. The results showed that the most severe morphological change and necrosis from tumor slices were observed in G5 (Figure 8A top). G5 significantly increased the apoptosis of cancer cells (Figure 8A middle) and had the lowest stained cells by Ki-67 (Figure 8A bottom). The pathological results were consistent with the tumor size and volume, indicating that G5 had the best antitumor effect.

Metastasis is a major challenge for TNBC treatment (Li et al., 2020). 4T1 breast tumor-bearing mice treated with a combination of two-step MW irradiation with HA-BNPs@Ptx exhibited a reduction in spontaneous lung metastasis (Figures 8B,C). Intratumoral immune profiling was performed in the five groups through immunofluorescence staining (Figures 9A–D). The proportions of T cells (CD3⁺) were increased in the treated group compared with G1. Among T cells, the percentages of CD4⁺ T cells and CD8⁺ T cells were significantly increased in the G5. Analysis of peripheral blood by flow cytometry was consistent with the above results (Figures 9E,F). These results suggest that sequential treatment of SCS can stimulate and recruit CD4⁺ CD8⁺ T cells in breast cancer tissues and, thus, effectively inhibit spontaneous lung metastasis.

Subsequently, cytotoxicity of HA-BNPs-Ptx was investigated in vivo. There was no difference in liver and kidney function (analysis of serum levels for ALT, AST, BUN, and CRE) between the PBS group and the HA-BNPs-Ptx group (Supplementary Figure S8). This indicates that the nanoparticles have good biocompatibility. The histological analysis staining of main organs tissue was done by hematoxylin and eosin (H&E) (Supplementary Figure S9). After the treatment to study the damage in mice, no tissue necrosis was observed in the main organs (heart, liver, spleen, and kidneys) for all groups, demonstrating that the SCS was safe in vivo.