Synergistic strategy with hyperthermia therapy based immunotherapy and engineered exosomes−liposomes targeted chemotherapy prevents tumor recurrence and metastasis in advanced breast cancer

Abstract Advanced breast cancer with recurrent and distal organ metastasis is aggressive and incurable. The current existing treatment strategies for advanced breast cancer are difficult to achieve synergistic treatment of recurrent tumors and distant metastasis, resulting in poor clinical outcomes. Herein, a synergistic therapy strategy composed of biomimetic tumor‐derived exosomes (TEX)‐Liposome‐paclitaxel (PTX) with lung homing properties and gold nanorods (GNR)‐PEG, was designed, respectively. GNR‐PEG, with well biocompatibility, cured recurrent tumors effectively by thermal ablation under the in situ NIR irradiation. Meanwhile, GNR‐mediated thermal ablation activated the adaptive antitumor immune response, significantly increased the level of CD8+ T cells in lungs and the concentration of serum cytokines (tumor necrosis factor‐α, interlekin‐6, and interferon‐γ). Subsequently, TEX‐Liposome‐PTX preferentially accumulated in lung tissues due to autologous tumor‐derived TEX with inherent specific affinity to lung, resulting in a better therapeutic effect on lung metastasis tumors with the assistance of adaptive immunotherapy triggered by GNR in vivo. The enhanced therapeutic efficacy in advanced breast cancer was a combination of thermal ablation, adaptive antitumor immunotherapy, and targeted PTX chemotherapy. Hence, the synergistic strategy based on GNR and TEX‐Liposome provides selectivity to clinical treatment of advanced breast cancer with recurrent and metastasis.

advanced breast cancer have a shorter survival time, and the mortality rate is approximately 90% within 5 years after diagnosis. [4][5][6] The recurrent tumors or distant organs metastasis (lung, bone) can be produced via system blood, forming advanced breast cancer. Current clinical oncologic methodolodies for advanced breast cancer mainly include local treatment, immunotherapy and systemic chemotherapy, including single or combination chemotherapy. 7 Generally, existing treatments do not achieve complete cure and usually end in failure due to the insurmountable targeting issues, drug resistance and systemic side effects. Meanwhile, patients in the advanced stage are poorly tolerated, leading to further failure in clinical. Thus, an innovative and effective precise treatment strategy for advanced breast cancer with recurrent tumors and distal organ metastasis is urgently needed, aiming to achieve maximum efficacy accompanied with minimum toxicity.
Nanomaterial-based photothermal therapy (PTT) has been extensively explored as an emerging strategy against malignant tumors in clinical. 8 Compared with conventional treatment such as surgical resection, radiotherapy and chemotherapy, patients with advanced tumors have higher compliance with PTT. As reported, various nanomaterials, such as gold nanorods (GNR), [9][10][11][12] graphene oxide, 13 carbon nanotubes, 14 and indocyanine green, have been widely explored for PTT. As highly effective and noninvasive treatment models, PTT has a direct cancer cell killing effect for local tumors through hyperthermia ablation. [15][16][17][18][19] Simultaneously, hyperthermia ablation of tumors can release tumor-associated antigens and these antigens are recognized by dendritic cells (DC) and presented to native T cells, thus promoting the activation of adaptive antitumor immune responses, and further enhancing antitumor efficacy through immunetherapy. 20,21 Although many investigations have demonstrated the remarkable antitumor efficacy of PTT combined with immunotherapy, 22,23 it is not able to cure the distal metastasis lesions. 24 Currently, most of the distant metastatic tumors are treated by systemic chemotherapy in clinical. However, lack of active targeting property makes the treatment need to face the situation of drug resistance, side effects, and failure. Exosomes, with a bilayer membranelike structure similar to liposome, [25][26][27] are naturally secreted by most cell types. 28 As reported, exosomes secreted by breast cancer cells possess excellent lung targeting capability due to their functional surface integrins (α6β4, α6β1), 29 which colocate in the laminin-rich lung microenvironment. 30,31 The characteristic endows enormous potential for employing exosomes as targeted drug delivery vehicles. 32,33 In the early stage, exosomes derived from breast cancer were adopted as a gene delivery vector for the delivery of siS100A4, exhibiting excellent lung targeting capability and realizing the treatment of lung metastasis. 34 However, the low drug loading efficiency of exosomes limits their potential applications as drug delivery platforms. 35 To overcome the targeting defects of lipid-based vectors, a biomimetic hybrid system composed with tumor-derived exosomes (TEX) and liposomes for specific delivery to lung was proposed.
Inspired by the above technological progresses, a synergistic strategy based on GNR-PEG and TEX-Liposome-paclitaxel (PTX) is constructed for targeted treatment of recurrent tumors and distant metastases of advanced breast cancer, respectively (Scheme 1). In this delivery system, GNR-PEG-mediated hyperthermia served as the pioneer to kill recurrent tumors by thermal ablation via in-situ NIR irradiation. Meanwhile, the apoptosis of recurrent tumor cells allowed DC to be recruited, and then the tumor antigens were presented to native T cells, stimulating the adaptive immunity via CD8 + T cell pathway.
Then, TEX membrane with inherent lung homing affinity was hybrid into Liposome-PTX, forming TEX-Liposome-PTX. Sequentially, TEX-Liposome-PTX was adopted for targeted chemotherapy of distal lung metastasis tumors by intravenous injection. In addition, the adaptive immune response triggered by hyperthermia ablation exhibited adjuvant therapy for lung metastatic tumors, thereby enhancing the efficacy of chemotherapy, and providing a potential effective targeted therapy strategy for advanced recurrent and metastatic breast cancer in clinical. CD3-FITC, CD4-APC, CD8-Percp-Cy5.5, CD80-APC, and CD86-PE antibodies were purchased from BD Pharmingen. ELISA kits for interlekin-6 (IL-6), interferon-γ (IFN-γ), and tumor necrosis factor-α (TNF-α) were obtained from Multi Sciences Biotech, Co., Ltd. The other chemical reagents with analytical grade were obtained from Solarbio Life Sciences.  Hydrochloric acid (2.1 ml) and 0.07 mM ascorbic acid were added to acidify the growth solution, and 0.8 ml of the seed solution was introduced into the growth solution subsequently. The solution containing GNR was obtained by standing for 12 h.

| Preparation and characterization of GNR-PEG
The above obtained GNR solution was centrifuged at 10,000 rpm for 20 min and resuspended in water. The 6 mg/ml PEG-SH solution was introduced into the GNR suspension solution under the condition of vigorous stirring at 50 C. The mixture was stirred for 12 h at 50 C and centrifuged at 10,000 rpm for 20 min. The obtained GNR-PEG was resuspended in water, cleaned by ultrasonication and centrifuged to remove the CTAB, and unreacted PEG-SH. GNR-PEG was prepared successfully and resuspended in ethanol.
Particle size and zeta potentials of GNR-PEG were evaluated by Malvern Zetasizer Nano ZS90 (Malvern Instruments Ltd.). The morphology of GNR-PEG was observed by high-resolution transmission electron microscopy (TEM; JEOL). Absorption spectra of GNR-PEG was detected by the UV-Vis spectrophotometer (JingHua Technological Instrument Corporation). The photothermal effect of GNR-PEG was evaluated by the 808 nm NIR laser irradiation (808 nm; LEO-Photoelectric).

| Preparation of Liposome-PTX
The thin-film hydration method was adopted to prepare Liposome-PTX. Briefly, soy lecithin, cholesterol, and PTX (10:1:1 wt/wt) were dissolved in 15 ml chloroform. Subsequently, chloroform was evaporated to form the thin film under vacuum at 37 C for 2 h and the S C H E M E 1 Illustration of the combination strategy of GNR-PEG and TEX-Liposome-PTX to suppress the advanced recurrent and metastatic breast cancer. GNR, gold nanorod; PTX, paclitaxel; TEX, tumor-derived exosome prepared thin film was further dried under vacuum overnight. Finally, the film was hydrated with 60 ml phosphate-buffered saline (PBS), stirred at 37 C for 3 h, and downsized by sonication under the condition of 300 W.

| Extraction and separation of TEX
The model of 4T1 tumor-bearing mice was established to acquire the autologous breast cancer cells. The protocol to extract exosome was in accordance with the published article in our laboratory. 34 Exosome was harvested from the autologous breast cancer cells culture medium by ultracentrifugation. Briefly, the autologous breast cells were cultured for 48 h in FBS-free high-glucose DMEM medium. The concomitant culture medium was collected, and a gradient centrifugation was used to isolate the exosome as the following conditions. The supernatant was centrifuged at 305 Â g for 10 min to dislodge the cells, centrifuged at 2005 Â g for 15 min to remove cell debris, and centrifuged at 100,000 Â g for 75 min to isolation exosomes. The exosome pellet was washed with PBS containing PMSF, and again pelleted by centrifugation to purify the exosome. The isolated exosome was resuspended in PBS and feezed at À80 C. All the protocol was conducted at 4 C.
Exosome membrane was isolated by the repeated freezing and thawing. The above obtained exosome was resuspended in cold TM buffer, frozen at À80 C rapidly, and thawed in a water bath at 37 C for 5 min. Repeated freeze-thaw cycle five times in the above condi- To identify the successful fusion between TEX membrane and Liposome, CLSM was used to verify the fusion. DSPE-PEG-FITC phospholipid was adopted for Liposome preparation. Thus, Liposome was labeled by FITC (green). TEX membrane was incubated with Alexa Fluor 647 anti-CD9 antibody. Therefore, TEX was indicated by Alexa Fluor 647 (red). Afterwards, the mixture of TEX membrane and Liposome and the hybrid TEX-Liposome-PTX were incubated with NIH3T3 cells for 4 h. Next, NIH3T3 cells were fixed with 4% paraformaldehyde and subsequently treated with Hoechst 33342 for 15 min.

| Cytotoxicity of TEX-Liposome-PTX
To assess the treatment effect of TEX-Liposome-PTX, MTT assay was performed as the same as the above protocol. After the 4T1 cells were seeded into the 96-well plates, different concentrations (0.1, 1, 5, 20, 40, and 80 μg/ml) of PTX, Liposome-PTX, and TEX-Liposome-PTX were added and incubated for another 24 h. The cytotoxicity of TEX-Liposome-PTX was conducted by MTT assay.

| In vitro cellular uptake and targeting of TEX-Liposome
Cellular uptake of TEX-Liposome was estimated using CLSM and the cellular targeting ability was evaluated through the different ratio of TEX and Liposome. Briefly, NIH3T3 cells were seeded into the confocal dishes at a density of 1.0 Â 10 5 cells/dish and incubated for 24 h in the incubator. Afterward, the culture medium was discarded and replaced with DiO-labeled Liposome formulations. The ratio of Liposome and TEX in the various formulations was 1:0, 100:1, 20:1, and 5:1. After incubation for 0.5, 2, and 4 h, the cellular uptake of NIH3T3 cells was stopped with the cold PBS and fixed with 4% paraformaldehyde for 10 min. Subsequently, the nuclei were stained by incubating with DAPI for 10 min at room temperature. The fluorescent images were obtained by CLSM (TCS SP5; Leica) and semi-quantitative analysis was performed using ImageJ.

| Detection of cytokine levels
After being treated with a single dose of PTX-chemotherapy, the peripheral blood of mice was taken and coagulated at room temperature for 15 min. Then, samples were centrifuged at 3000 rpm for 15 min, and the serum was obtained. The concentration of TNF-α, IFN-γ, and IL-6 in serum was measured according to the manufacture's operating protocol in the ELISA reagent test kit.

| Statistical analysis
The above-mentioned data were processed by GraphPad Prism software (version 8.0) and expressed as mean ± SD. The difference between the two independent groups was performed by a Student's t-test. p < 0.05 was considered statistical significance.

| Preparation and characterization of GNR-PEG
GNR-PEG as an effective PTT agent was synthesized according to the seed growth method. To surmount the aggregation instability and cell cytotoxicity of GNR, PEG was modified on the surface. As presented in TEM images (Figure 1a Finally, the photothermal conversion stability of GNR-PEG was investigated under five on and off cycles with the irradiation of 808 nm. As exhibited in Figure 1h, the stable photothermal conversion capability proved the good stability of GNR-PEG nanoparticles.

| Preparation and characterization of biomimetic TEX-Liposome-PTX
As previously reported, the PTX-loaded liposomes were prepared by thin-film hydration method. The hydrodynamic size and zeta potential of Liposome-PTX were measured by dynamic light scattering (DLS).
After PTX loaded, an increased size of 20 nm and a decreased zeta potential of 9 mV were observed (Table S1). The morphology of Liposome-PTX was visualized by TEM with negative staining. Both Liposomes and Liposome-PTX presented obvious unilamellar vesicle morphology observed in Figure 2c,d, and the sizes were consistent with hydrodynamic size detected by DLS. Drug loading and encapsulation efficiency detected by HPLC were 9.99 ± 0.96% and 97.58 ± 1.36%, respectively. The dialysis method was adopted to investigate the in vitro release behavior of PTX-loaded liposomes in PBS solution.
As exhibited in Figure 2j, PTX released from liposome slowly, reaching 70% within 100 h and presenting sustained release behavior.
To enable lung-PMN targeting of Liposome-PTX, exosome membranes were extracted from autologous breast cancer cells using ultracentrifugation and repeated freezing and thawing, fused into Liposome by incubation-extrusion method, forming an exosome hybrid liposome delivery system. 37 The harvested exosomes were characterized by DLS and TEM (Figure 2b and Table S1). The particle size of exosome was 106.63 ± 1.24 nm, and the zeta potential was À37.45 ± 1.53 mV. A round-shape morphology with the visible lipid layer of exosomes was observed by TEM, which was consistent with the previous literature, 34 demonstrating the successful extraction of exosomes. Furthermore, according to intuitive evidence from  (Table S1), and the particle size distribution exhibited in Figure 2g,h proved the better dispersion properties of the hybrid nanoparticles again. As presented in Table S1, the particle size of exosome membrane was 336.47 ± 20.33 nm, greater than exosome. With the proportion of exosome membranes increasing in the hybrid nanoparticles and the loading of PTX, the particle size increased gradually (Figure 2i and Table S1). Finally, the successful hybridization of exosome membrane and liposomes could also be confirmed by surface potential. The zeta potentials of hybrid nanoparticles prepared in all proportions were The encapsulation efficiency of TEX-Liposome-PTX with various proportions was~97%, and PTX contents were all above 9.2%, showing negligibly significant difference with that of Liposome-PTX (Table 1). Meanwhile, the in vitro cumulative release behavior of TEX-Liposome-PTX with different proportions was investigated, displaying a sustained release, similar to the behavior of Liposome-PTX.

| In vitro biocompatibility of GNR-PEG and TEX-Liposome-PTX
Cytotoxicity of delivery nanosystem limits the in vivo application. Therefore, an excellent vehicle should have satisfactory biocompatibility. As reported, TEX have inherent affinity to interact with endothelial cells. 38 Thus, we chose HUVEC cells to investigate the biocompatibility of GNR-PEG by MTT assay. As exhibited in Figure 3a

| In vitro cellular uptake and targeting of TEX-Liposome
Pulmonary metastasis is a common phenomenon in advanced breast cancer. As reported, exosomes derived from breast cancer cells have lung homing capability due to the presence of integrins. 29 Thus, we speculated that TEX hybrid liposome had the lung targeting capability.
To evaluate the lung targeting ability, NIH3T3 cells were selected and cell internalization of DiO-labeled liposomes was monitored by CLSM.
As shown in Figure 5a  T lymphocytes are the main force of immunotherapy in vivo. 39 For immune effect, humoral immunity is mediated by CD4 + T cells, and cytotoxic CD8 + T lymphocyte can kill tumor cells straightly. 40

| In vitro antitumor assays
Results of T lymphocyte differentiation in lung tissues were presented in Figure 9c-e. Notably, the proportion of CD4 + T cells had no difference in all groups, indicating that GNR and the combination strategy would not activate the humoral immunity mediated by CD4 + T cells through MHC-II. Similarly, the percent of CD8 + T cells and the concentration of serum cytokines (TNF-α, IL-6, and IFN-γ) in TEX-Liposome-PTX and control groups had no significant difference. While the percent of CD8 + T cells in GNR + TEX-Liposome-PTX group was twofold higher than that in TEX-Liposome-PTX group, suggesting that GNR-mediated thermal ablation increased the percent of CD8 + T cells in lung. In addition, activated CD8 + T cells have an increased ability to produce immune cytokines, including IFN-γ and TNF-α. 41 As expected, the cytokine concentration of TNF-α, IL-6, and IFN-γ in GNR + TEX-Liposome-PTX group was significantly higher than that in control group, indirectly proving that CD8 + T cell-mediated immunity was activated. In summary, the treatment strategy of GNR combined with TEX-Liposome-PTX could activate cell-specific immune responses through MHC-I, play a tumor-killing effect, and enhance the efficacy of local and metastatic tumors.