OSW-1 Induces Apoptosis and Cyto-Protective Autophagy, and Synergizes with Doxorubicin on Spontaneous Metastasis of Triple Negative Breast Cancer

Triple-negative breast cancer (TNBC) is the most malignant subtype of breast cancer and chemotherapy using drugs such as doxorubicin is the main treatment strategy. However, drug resistance and dose-dependent toxicities restricted its clinical use. Natural products are major sources of anti-tumor drugs. OSW-1 is a natural compound that has shown strong anti-cancer effects in several cancer types, but its effects on the ecacies of chemotherapy in TNBC and the underlying mechanism remain unclear. We investigated the effects of OSW-1 on the viabilities, apoptosis and autophagy in 4T1 and MDA-MB-231 cells. The synergistic effects of combination with OSW-1 and doxorubicin on TNBC growth and spontaneous metastasis were examined in vitro and in vivo.


Background
Triple-negative breast cancer (TNBC) is the most malignant subtype of breast cancer and chemotherapy using drugs such as doxorubicin is the main treatment strategy. However, drug resistance and dosedependent toxicities restricted its clinical use. Natural products are major sources of anti-tumor drugs. OSW-1 is a natural compound that has shown strong anti-cancer effects in several cancer types, but its effects on the e cacies of chemotherapy in TNBC and the underlying mechanism remain unclear.

Methods
We investigated the effects of OSW-1 on the viabilities, apoptosis and autophagy in 4T1 and MDA-MB-231 cells. The synergistic effects of combination with OSW-1 and doxorubicin on TNBC growth and spontaneous metastasis were examined in vitro and in vivo.

OSW-1 induces mitochondria-dependent apoptosis and cyto-protective autophagy in TNBC cells in vitro.
In addition, OSW-1 and doxorubicin exhibits strong synergistic anti-TNBC capabilities both in vivo and in vitro. The combination treatment can strongly inhibit spontaneous metastasis in 4T1 model. Consistent with this inhibition, the proportion of CD8 + T lymphocytes in the lung microenvironment were signi cantly increased after combination treatment.

Conclusions
This study indicates that the combination therapy with natural compound and chemotherapeutic drug could be a promising anti-TNBC strategy which deserves further investigation.

Background
Cancer statistics for 2021 reported that three most common cancers in women include breast cancer, lung cancer and colorectal cancer, with breast cancer alone accounting for 30% of cancers in women 1 .
Breast cancer is the most commonly diagnosed cancer and the leading cause of death among women worldwide. Previous studies associated this high mortality with late metastases, with a 5-year survival rate reduced from 99% for patients with only local lesions to 27% when metastases occur 2 . Breast cancer mainly comprises three subtypes: hormone receptor positive/ERBB2 negative (HR + /ERBB2 − ), ERBB2 positive (ERBB2 + ), and triple negative breast cancer (TNBC). TNBC has the highest likelihood of recurrence compared with the other two subtypes. Main current treatment options for TNBC in China include surgery, radiation therapy, chemotherapy and systemic treatment 3 . Although TNBC treatment strategies have been intensively investigated recently, the 5-year survival rate of TNBC patients is still low due to high incidence of recurrence and metastasis 4 . The incidence of lung metastasis is more frequent in TNBC compared with non-TNBC 5 . Due to risk factors, population size and age structure, the incidence and mortality of breast cancer continue to increase in China, and the trend is not likely to stop by 2030 6 . Therefore, exploration and development of effective new drugs or combination therapy regimens for metastatic TNBC are urgently needed.
Chemotherapeutic agents are currently the most commonly used treatment options for TNBC 7 .
Doxorubicin (DOX), one of the most widely used anti-TNBC drugs, is a topoisomerase II (TOP II) inhibitor that exert DNA damages by introducing double-strand break (DSB) and causing chromatin damages in the genome 8 . However, as with other small molecule chemotherapeutic agents, complete tumor eradication is usually not achieved by DOX monotherapy due to rapid development of multidrug resistance in cancer cells 9,10 . Moreover, DOX causes several severe side effects, among which dosedependent irreversible cardiotoxicity is the most serious one. Previous studies reported that approximately 11% of cancer patients treated with DOX developed acute cardiotoxicity. In addition, other vital organs such as kidney, liver, and brain are adversely affected following DOX therapy 11,12 .
Limitations of DOX monotherapy have shifted the focus to combination chemotherapy as a treatment option against TNBC. Combination therapy, namely the combination of drugs with different pharmacological activities to treat disease, has some advantages compared to monotherapy because it may exert synergistic e cacies and decrease the dose-limiting side effects of single drugs 13 . Therefore, several combination therapy strategies, such as the co-admibistration of more than two anti-cancer agents, have gradually been used to enhance chemosensitivity.
Biologically active natural compounds are potential anticancer drug sources. Due to their negligible side effects and excellent anticancer activities, some of them have been serving as potent adjuvants to conventional chemotherapeutic anticancer drugs. OSW-1[3β,16β,17α-trihydroxycholest-5-en-22-one-16-O- , the main active component of saponins, was isolated from simple bulbs of Ornithogalum saundersiae in 1992 14 . Several previous studies have found that OSW-1 inhibits several types of malignant cells, including colon cancer cells, liver cancer cells, and leukemia [15][16][17] . Moreover, previous studies showed that OSW-1 is approximately 10-100 times more cytotoxic compared with some clinical anticancer agents such as paclitaxel, adriamycin, cisplatin, mitomycin C and camptothecin 18 .
The e cacy of OSW-1 to inhibit breast cancer has been investigated previously 19 . However, the underlying mechanism of OSW-1 and its effects on current standard chemotherapy of TNBC are still illusive 19 . In the current study, we explored the effects of OSW-1 on mitochondria-mediated intrinsic apoptosis. Autophagy plays important roles in determining the fate of cancer cells. Therefore, we explored how autophagy regulates the e cacies of OSW-1 to suppress TNBC. The e cacies of combining OSW-1 with DOX were evaluated in vitro. Moreover, the effects of this combination strategy on ve levels: Slight synergism (0.85<CI<0.9), moderate synergism (0.7<CI<0.85), synergism (0.3<CI<0.7), strong synergism (0.1<CI<0.3), and very strong synergism (CI<0.1) 23,24 .

Western blot
After treatment with OSW-1, DOX or their combinations for the indicated time, protein levels were measured by western blotting analysis as described previously 25 . Cells were harvested and lysed on ice in RIPA buffer containing protease inhibitors cocktail to extract proteins. Subsequently, the cell lysate was centrifuged at 13,000g, 4℃ for 15 min. The protein concentration in the cell lysate was determined using the BCA Protein Assay Kit (Pierce, Rockford, IL, USA). Equal amounts of protein (30 μg) were separated by electrophoresis on SDS-PAGE gels and transferred onto polyvinylidene di uoride (PVDF) membranes. The membranes were then blocked with 5% nonfat milk in TBS/T for 1h and then incubated overnight with the speci c primary antibody at 4°C. The membrane was washed with TBS/T and then incubated with HRP-labeled secondary IgG antibodies. The protein bands were visualized by enhanced chemiluminescent substrate (Amersham, Piscataway, NJ). The intensity of protein bands was determined using Image J software for three independent experiments.

Morphological analysis of cell nuclei
After treatment with OSW-1 for 48 h, cells were washed with PBS and xed with methanol for 15 min and then stained with Hoechst 33342 (10 µg/mL) for 30 min in the dark. The morphology of the nuclei was observed with an inverted uorescence microscope (Olympus, Tokyo, Japan). Each assay was replicated three times.

Plasmid transfection and immuno uorescence analysis
The 4T1 or MDA-MB-231 cells at the exponential growth stage were transfected with GFP-RFP-LC3 plasmid using the Lipofectamine 3000. After 24 h of transfection, cells were seeded in Millicell EZ SLIDE (Merckmillipore) and treated with OSW-1 or CQ for 48 h. They were then xed with 4% paraformaldehyde for 30 min at room temperature and washed with PBS. The cells were stained with the DAPI solution (5 µg/mL) for 5 min at 37℃. The quanti cation of GFP-RFP-LC3 in transfected cells were examined and imaged by confocal microscopy.

Subcutaneous tumor models
In vivo mouse experiments were performed at Sichuan University (State Key Laboratory of Biotherapy and Cancer Center) after approval by the Ethics Committee of Sichuan University. Female (8-10 weeks old) BALB/c mice were bought from HFK Bioscience (Beijing, China). To generate 4T1 subcutaneous tumors, 1×10 5 luciferase-expressing 4T1 cells were inoculated subcutaneously into the left-anks of mice. The cells were allowed to grow and generate tumors. When the tumors reached a volume of approximately 100mm 3 (mean ± S.E.M), mice were randomly divided into groups and treated with (1) Vehicle, (2) OSW-1, (3) DOX, or (4) Combination of OSW-1 and DOX at a dose of 10 μg/kg OSW-1 once every 2 days via intraperitoneal(i.p.) injection and 5 mg/kg DOX once every 5 days via intravenous (i.v.) injection. The cages were labeled with cage cards, and different treatments groups were clearly written after randomization. The drug-treated mice and the control-treated mice were in different cages. During treatment, only one cage was taken at a time for dosing. The vehicle mice were dosed rst. After the dosing of the vehicle mice was completed, the drug-treated mice were taken out. The tumor volume was calculated using the formula: V(mm 3 )=L(mm)×W 2 (mm 2 ) ×0.5. The tumor volumes and body weights of the mice were measured every 3 or 4 days.

Spontaneous lung metastasis models
In BALB/c mice, 4T1 cells tend to induce spontaneous lung metastasis 26,27 . When the average tumor volume of mice reached approximately 1000mm 3 , the tumors were carefully dissected, and the wounds were sutured. Metastasis to the lung was monitored using a non-invasive in vivo imaging system (IVIS, PerkinElmer). The data were collected and analyzed with the Living Image®4.5.2 software.

Immunohistochemical staining of tumor sections
The immunohistochemical (IHC) assay of tumor tissues was performed as described previously with some modi cations 28 . Brie y, tumor tissues were collected from mice, xed in 4% paraformaldehyde for 48 h and embedded in para n. Expression level of cleaved caspase 3 and Ki-67 in tumor tissues was detected with IHC analysis.

Analysis of the immune cells in lung microenvironment
The lung tissues were cut into small pieces, and then enzymatically digested with collagenase at 37℃ for 2 h. The cell suspension was ltered through a 70 μm strainer to remove large debris and obtain a single cell suspension. The red blood cells in the suspension were lysed with a red blood cell lysis buffer (BL503B, Biosharp, Hefei, China) at room temperature for 5 min. Finally, the cells were stained with speci c antibodies in the dark at 4℃ for 30 min, and then analyzed with ow cytometry using ACEA NovoCyte (ACEA Biosciences Inc., San Diego, USA).

Statistical analysis
The data were represented as mean ± S.D. or mean ± S.E.M. from independent replicates and analyzed by GraphPad Prism 8.0 software. Multiple comparisons between two groups were undertaken using Student t-test or Mann-Whitney U tests. Signi cant statistical p-values were labeled as ***p<0.001, **p<0.01 and *p<0.05.

OSW-1 inhibited the growth of TNBC cells in vitro
Previous studies have shown that OSW-1 inhibited the growth of pancreatic cancer, hepatocellular carcinoma, colon cancer and breast cancer cells in vitro, and had lower toxicity on normal endothelial cells [15][16][17]19 . The current study focused on the anticancer activities of OSW-1 to suppress TNBC. We selected human TNBC cell line (MDA-MB-231) and mouse TNBC cell line (4T1) for further experiments. MTT assay showed that OSW-1 strongly reduced the viabilities of both TNBC cell lines in a timedependent manner (Fig. 1a). Colony formation was measured after OSW-1 treatment to further determine the inhibitory effects of OSW-1 on TNBC cell proliferation. The data clearly showed that OSW-1 inhibited the number and size of TNBC cell colonies at very low concentrations (Fig. 1b, c). In summary, OSW-1 signi cantly inhibited TNBC cell growth in vitro with very high activities in the current study.

OSW-1 induced apoptosis of TNBC cells in vitro
Inducing apoptosis is a very common mechanism of anticancer agents. OSW-1 has been shown to cause apoptosis of cancer cells. We examined the apoptosis of TNBC cell lines after OSW-1 exposure. As shown in Fig. 2a, OSW-1 treated TNBC cell lines showed the signs of cell contraction, a typical characteristic of apoptotic cells, indicating the occurrence of cell apoptosis. Hoechst 33342 staining were performed after incubating with OSW-1 for 48 h. Bright-blue uorescent condensed nuclei, reduction of cell volume and nuclear fragmentation were observed and further showed the features of apoptosis (Fig. 2b). FCM was used to quantitatively analyze OSW-1-induced apoptosis. Treatment with OSW-1 for 48 h increased the percentage of apoptosis of 4T1 and MDA-MB-231 cells in a concentration-dependent manner (Fig. 2c, d). Expression analysis of some key proteins related to apoptosis also con rmed that OSW-1 treatment induced cleavage of caspase 3 and PARP, two hallmark events during apoptosis (Fig. 2g-i). We also found that OSW-1 reduced the expression of phosphorylated AKT (Fig. 2g-i), which is consistent with previous studies 29 . In summary, these ndings show that the inhibitory activity of OSW-1 on TNBC cells occurs through apoptosis.

Osw-1 Likely Induced Apoptosis Via The Mitochondriamediated Intrinsic Pathway
We determined the expression levels of some apoptosis-related proteins to explore the mechanism of apoptosis. Bcl-2 family proteins play pivotal roles in apoptosis by maintaining mitochondrial homeostasis. Previous studies reported that Bcl-2 family proteins maintain balance between antiapoptotic and pro-apoptotic proteins, which are crucial for regulating mitochondrial integrity 30 . Expression of Bcl-2 family proteins in TNBC cells was determined after treatment with OSW-1 for 48h and the data showed that OSW-1 increased the expression of Bax and decreased that of Bcl-2, leading to signi cant increase in Bax/Bcl-2 expression ratio ( Fig. 2g-i). These ndings suggest that OSW-1 might induce apoptosis through the mitochondria-mediated intrinsic pathways. To verify this hypothesis, we determined the changes of ΔΨm by FCM after OSW-1 treatment using Rh123. Our ndings showed that OSW-1 caused loss of ΔΨm in 4T1 and MDA-MB-231 cells (Fig. 2e, f). Together, our data indicate that mitochondria-mediated intrinsic apoptotic pathway plays important roles in OSW-1-induced cell death.
OSW-1 induced cyto-protective autophagy and impared autophagy ux in TNBC cells Autophagy is a double-edged sword in cancer progression and cancer treatment. Previous studies showed that OSW-1 treatment activated autophagy in colorectal cancer and pancreatic cancer cells 31 . To elucidate the role of autophagy in OSW-1 induced TNBC cell death, we conducted several assays. Conversion of LC3-I into lipid LC3-II and formation of LC3 puncta are regarded as two biomarkers of autophagy. The current study established that OSW-1 treatment induced increase in LC3-II expression (Fig. 3a) and accumulation of LC3 puncta (Fig. 3b, c), indicating activation of autophagy. Expression level of autophagy-speci c substrate p62 is an indicator of autophagy ux, which is involved in autophagy regulation and maintains intracellular homeostasis and signal transduction. We observed that p62 levels increased in TNBC cells after OSW-1 treatment (Fig. 3a), suggesting that OSW-1 might impair autophagy ux. To investigate whether the impaired autophagic ux is due to failure in autophagosome-lysosome fusion, a tandem RFP-GFP-labeled LC3B construct was used. The data showed that most LC3B puncta showed RFP + GFP + signals (autophagosomes) rather than RFP + GFP − signals (autolysosomes) (Fig. 3d-f), indicating that OSW-1 treatment caused accumulation of autophagosomes and decreased autolysosome formation. The above data showed that OSW-1 blocked autophagy ux in TNBC cells. To elucidate how autophagy mediated the anti-TNBC effects of OSW-1, MDA-MB-231 and 4T1 cells were treated with OSW-1 in combination with chloroquine (CQ), an autophagosome-lysosome fusion inhibitor, or Wortmannin (WORT), a PI3K inhibitor which prevent autophagy initiation and autophagosome formation. The data clearly showed that OSW-1 induced cell growth suppression was signi cantly enhanced after combination with CQ and Wortmannin in a time-dependent manner. Overall, these data showed that OSW-1 induced protective autophagy and inhibition of autophagy could aggravate OSW-1's anti-cancer activities in TNBC cells.

OSW-1 sensitizes TNBC cells to DOX treatment in vitro
DOX is the rst-line drug for TNBC chemotherapy due to its excellent anticancer e cacies. However, the dose-dependent cardiotoxicities restricted its clinical use and should not be neglected. Therefore, it's of great signi cance to nd effective combination strategies to increase the anticancer e cacies of DOX. We explored whether OSW-1 enhances the antitumor effects of DOX. MTT assay was undertaken to determine the effects of combination of OSW-1 and DOX in TNBC cells. Three different concentrations of each drug were selected for the permutation and combination. Our results showed that OSW-1 and DOX synergistically inhibited the viabilities of 4T1 and MDA-MB-231 cells (Fig. 4a). To quantitatively evaluate the synergy between DOX and OSW-1, combination index was calculated using Compusyn software (Fig. 4b). The data demonstrated that the CI points were below 1 in all selected concentration combinations. Moreover, very strong synergistic effects (CI < 0.1) were observed at certain concentrations. Colony formation assay was conducted to further investigate the synergistic effects of OSW-1 and DOX and the drug combinations exhibited better inhibition than either monotherapy (Fig. 4c, d). These data suggested a synergistic effect of DOX and OSW-1 to inhibit TNBC cell viabilities.
Both OSW-1 and DOX exert anti-TNBC effects by inducing apoptosis 32 . We speculated that the synergistic effect of DOX and OSW-1 also resulted from enhanced apoptosis. Accumulation of apoptotic cells was determined by Annexin V/7-AAD staining after treatment with OSW-1, DOX and their combination for 48 h. The proportion of apoptotic cells was 15.67% after OSW-1 monotherapy and 10.99% after DOX monotherapy in 4T1 cells. Notably, the apoptotic rate in combination treatment group was increased to 48.6%, which was nearly 3-4 times of that in monotherapy group (Fig. 4e, f). A similar trend was observed in MDA-MB-231 cells (Fig. 4e, f).
Destruction of ΔΨm is one of earliest intracellular events during intrinsic apoptosis. We found that combination treatment caused more than four times of ΔΨm loss when compared to either DOX or OSW-1 alone, suggesting that the dual treatment might enhanced intrinsic apoptosis (Fig. 4g, h). In addition, the enhanced suppressive effects from combination treatment was accompanied by up-regulation of cleaved caspase 3, down-regulation of phosphorylated AKT, and up-regulation of the Bax/Bcl-2 expression ratio. Together, our data indicated that OSW-1 effectively enhances DOX-induced TNBC suppression.

OSW-1 sensitizes TNBC tumor growth and spontaneous metastasis to DOX treatment in vivo
Having demonstrated that OSW-1 enhanced the inhibitory of DOX in vitro, we next sought to determine the in vivo effects of this combination. We generated 4T1 TNBC models in vivo by subcutaneous inoculation of the cells into BALB/c mice. Notably, both OSW-1 and DOX modestly inhibited tumor growth in vivo (Fig. 5a), while combined DOX/OSW-1 treatment showed the best inhibitory effects on tumor growth without causing obvious weight loss of the animals (Fig. 5a, b). To determine the effects of the drugs on cancer cell proliferation and apoptosis in vivo, tumor tissue was stained with Ki-67 and cleaved caspase 3. We found that the abundance of Ki-67 was decreased most while the staining score of cleaved caspase 3 was the strongest in the combination treatment (Fig. 5c, d).
Metastasis is the main reason of cancer related death. Spontaneous lung metastasis and systemic metastasis in mice were evaluated in the current study because lung is a major metastatic site of TNBC. The primary 4T1 tumors were surgically removed from the mice, after which the metastases were visualized using IVIS imaging. While lung metastasis was inhibited by OSW-1 and DOX monotherapy, the combination of the two agents exhibited the strongest inhibition (Fig. 5e). Additionally, the combination treatment exerted better inhibition of systemic metastasis than monotherapy as evidenced by the quanti cation of luminescence intensity. Together, our data demonstrate that the addition of OSW-1 augmented the anticancer effects of DOX, especially in suppressing metastasis (Fig. 5f).

Combination of OSW-1 and DOX promoted the in ltration of CD8 + T cell in the lung
The immune microenvironment plays critical roles in cancer metastasis and signi cantly in uences treatment response and overall outcome of cancer patients. It has been shown that DOX could regulate the in ltration of some key immune cells in the TME 33 . Since immune cells play vital roles in cancer metastasis, we speculated that regulating immune cells in the TME are involved in the inhibition of metastasis.
CD8 + T cells play important roles in adaptive cancer immunity to eliminate cancer cells. The e cacy of most immunotherapy depends on the activities of CD8 + T cells. Cytotoxic CD8 + T cells entered tumor microenvironment, and exhibited cytotoxicities to tumor cells 34 . CD8 + T cells inhibit metastasis. When the balance between CD8 + T cells and platelets is disrupted, platelets produce CXCL4, which induces MDSCs thereby inhibiting the CD8 + T-cell function 35 . While neither DOX nor OSW-1 showed obvious effects on the proportion of CD8 + T cells in the lung, we found that its frequencies were augmented after combination treatment (Fig. 6a). We also detected the expression of the activation marker (CD69) and exhaustion marker (PD-1) on CD8 + T cells. The data showed that its activation was not in uenced after treatment. While DOX decreased the expression of PD-1 to some extent (p > 0.05), the addition of OSW-1 did not shown synergistic effects (Fig. 6g, h). CD4 + T helper cells are another group of key regulators of anticancer immunity in TME, which mediate adaptive immune response to cancer mainly by activating cytotoxic CD8 + T cells 34 . DOX treatment increased the proportions of CD4 + T cells in the lung. However, we didn't see a synergism between DOX and OSW-1 to in uence the frequencies of CD4 + T cells (Fig. 6b).
Neutrophils mainly promote cancer progression through several mechanisms, including promoting angiogenesis, immunosuppression and metastasis 36 . Targeting neutrophils represent a promising anticancer strategy. We observed that combination of OSW-1 and DOX signi cantly reduced the frequencies of neutrophils among CD45 + leukocytes compared with vehicle and OSW-1 monotherapy (Fig. 6c). However, no signi cant differences were observed between OSW-1 and DOX groups, indicating that the addition of OSW-1 did not affect the effects of DOX.

Dendritic cells (DCs) are the most powerful antigen-presenting cells and play central roles in regulating
the balance between CD8 + T cell immunity and tolerance to tumor antigens 37 . Although DOX treatment increased the frequencies of DCs in the lung, there were no signi cant differences between the combination treatment and DOX treatment group.
Macrophages play complicated roles in the progression of lung metastasis. Among them, M2 (CD206 + ) subtypes play promoting role in more cases 38,39 . In this study, we showed that DOX monotherapy caused a non-signi cant increase in the percentage of total F4/80 + macrophages in leukocytes. Intestinally, DOX treatment signi cantly decreased the proportions of M2 subtypes. Notably, OSW-1 treatment did not affect the effects of DOX to in uence the number of macrophages. In summary, our data suggest that CD8 + T might play important roles in the suppression of lung metastasis from the combination treatment. Notably, the addition of OSW-1 didn't exert inhibitory effects to DOX's anticancer immunity.

Discussion
Clinical treatment of TNBC remains unsatisfactory despite great efforts invested in health care and more effective treatment strategies are urgently needed to improve the survival of TNBC patients. DOX-based chemotherapy has become a standard treatment strategy for TNBC. However, toxicities and drug resistance restricted its clinical use. The natural compound OSW-1 is a potent anti-breast cancer drug candidate, which has shown good anti-tumor activity in several tumor types. Combination therapy provides more options for effective therapy from existing drugs, and is commonly used in clinical practice. In the current study, this strategy was used to explore the e cacy of combination treatment of TNBC with OSW-1and DOX.
In the current study, we con rmed that the destructive effects of OSW-1 on TNBC cell lines were exerted through promoting apoptosis. Apoptosis is an important mechanism of many anticancer drugs and it can be activated via two main pathways: the extrinsic apoptotic pathway, which is activated after the binding of the Fas ligand to the death receptors, and the mitochondrial-dependent intrinsic apoptotic pathway, which is mediated through mitochondria 40 . Previous studies established that OSW-1-induced apoptosis is mediated by mitochondria and involves cleavage of Bcl-2. Bcl-2 is cleaved by caspases during apoptosis, which leads to inactivation of Bcl-2 41 , or conversion of Bcl-2 into Bax-like fragments 42 . Some studies indicated that caspase-3 is involved in Bcl-2 cleavage 41,42 . Inhibition of caspase-induced Bcl-2 cleavage leads to inhibition of apoptosis 43 .
The intrinsic apoptosis pathways can be activated by several stimuli, which cause changes in inner mitochondrial membrane and opening of mitochondrial permeability transformation pores and the loss of ΔΨm 44 . We found that ΔΨm decreased after OSW-1 treatment, indicating that the integrity of mitochondrial membrane was impaired. This was further con rmed by decreased expression of Bcl-2 protein. Therefore, mitochondria-induced intrinsic apoptotic pathway was involved in OSW-1 induced apoptosis of TNBC cells.
In order to meet metabolic needs of cells and renewal of some organelles, cells degrade and recycle proteins and organelles in intracellular lysosomes through autophagy. Previous studies aver that autophagy in tumor cells is a double-edged sword, which may exert tumor-suppressive or tumorpromoting under different conditions 45 . Therefore, both promotion and inhibition of autophagy could be therapeutic strategies in advanced cancer depending on the speci c situations 46 . Autophagy-related genes (ATG) are critical regulators of autophagy process, and a series of ATG protein complexes coordinate to form double-membrane vesicles of autophagosomes 47 . These vesicles capture damaged or redundant proteins, lipids, glycogen, as well as organelle, which are further recognized by autophagy receptors (for example, p62) after ubiquitination 47 . Cargo receptors bind cargo and autophagosome membrane components LC3-II, thereby facilitating cargo separation. Autophagosome then binds to lysosome to form autophagolysome, where cargo is degraded. The term autophagy ux denotes total autophagy degradation. LC3 and p62 are two important and reliable markers to monitor autophagy ux 48 . We observed that OSW-1 induced accumulation of autophagy speci c substrate protein p62 and autophagosomes in TNBC cells. This indicated that autophagy ux might be affected and involved in the anti-TNBC effects of OSW-1. Two autophagy inhibitors were further used to elucidate the function of OSW-1 induced autophagy. The PI3K inhibitor Wortmanin inhibits early autophagy initiation, whereas CQ inhibits binding of lysosomes and autophagosomes during late autophagy. Our study showed that both Wortmanin and CQ aggravated the inhibitory effects of OSW-1, suggesting that OSW-1 induced cytoprotective autophagy in TNBC cells.
In vivo e cacy is an important parameter to evaluate the anticancer potential of a drug. In the current study, 4T1 subcutaneous tumor model was established and metastasis was measured after surgical resection of the primary tumor. We determined that the addition of OSW-1 augmented DOX's inhibition of the primary tumor growth, and did not cause additional weight loss of the mice. Most morbidity and mortality from breast cancer are due to metastatic disease, especially lung metastasis. Spontaneous lung metastasis and systemic metastasis were clearly seen after the primary tumor resection, which re ected the clinical practice of breast cancer progression. Notably, OSW-1 signi cantly enhanced the inhibitory effects of DOX on spontaneous metastasis in 4T1 model.

Adaptive and innate immunity play central roles in cancer inhibition and progression. DOX has been
shown to modulate anticancer immunity in breast cancer and other cancer models 33,49 . We speculated that the antimetastatic mechanism of the combination treatment may be achieved by regulating immune microenvironment of the metastasis. Therefore, the frequencies of some vital immune cells in the lung microenvironment were analyzed. CD8 + T cells play central roles in TME. Cytotoxic CD8 + T cells (CTL) entered tumor microenvironment and exhibited cytotoxicities to tumor cells 34 . The e cacy of most immunotherapy depends on the activities of CD8 + T cells. CD4 + T helper cells are another group of key regulators of TME, which mediate adaptive antitumor immune response mainly by activating cytotoxic CD8 + T cells 34 . As one of the most abundant leukocytes in the immune system, neutrophils play critical roles in promoting cancer progression through several mechanisms, including promotion of angiogenesis, immunosuppression 36 . DCs are the most powerful antigen-presenting cells with central roles in regulating balance between CD8 + T cell immunity and tolerance to tumor antigens 37 . Both alveolar macrophages and interstitial macrophages in the lung have been reported to contribute to the progression of lung metastasis 38, 39 . Macrophages can be simply divided into M1 (CD206 − ) and M2 (CD206 + ) subtypes according to CD206 expression, of which M2 macrophages are considered to promote cancer progression. We observed that the frequencies of CD8 + T cells in the lung were signi cantly augmented after combination therapy. DOX treatment increased the proportions of CD4 + T and DC cell in the lung, whereas the percentages of neutrophils and M2 macrophages in CD45 + leukocytes were decreased after DOX treatment. Notably, the addition of OSW-1 to DOX treatment regimen didn't affect the regulatory effects of DOX on the number of those immune cells. All these trends are favorable signals for the anticancer immunity, demonstrating that the e cacy of combination therapy with OSW-1 and DOX to treat TNBC cell lung metastasis in vivo probably occurs through regulating immune microenvironment.
In conclusion, this study established that OSW-1 has anti-TNBC effects and induces mitochondriamediated apoptosis. In addition, OSW-1 blocks autophagy ux and induces cytoprotective autophagy.
Furthermore, combination therapy with OSW-1 and DOX showed synergistic effects against TNBC both in vitro and in vivo. The addition of OSW-1 augmented the anti-metastasis abilities of DOX and promoted the in ltration of CD8 + T cells in immune microenvironment of lung tissue, suggesting the importance of anticancer immunities in the combination therapy. However, the current study did not clarify the speci c molecular mechanisms and targets of OSW-1 and DOX combination therapy against TNBC. These should be explored for further development of cutting-edge therapies for TNBC.

Declarations
Ethics approval and consent to participate The animal experiments were carried out in accordance with the recommendations of the Ethics Committee of Sichuan University. The protocol was approved by the Ethics Committee of Sichuan University.

Consent for publication
Not applicable.

Availability of data and materials
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any quali ed researcher.      The effects of combination therapy with OSW-1 and DOX on immune cells in ltration in the lung microenvironment. After in vivo luminescence imaging in Fig. 5, the lung tissues of the mice were collected to prepare sing cell suspensions. The cells were then stained with different sets of uoresceinconjugated antibody and analyzed with ow cytometry to determine the frequencies of some important cells involved in anticancer immunity. (a) The percentage of CD8+ T lymphocytes among CD45+