AGTR1 promotes lymph node metastasis in breast cancer by upregulating CXCR4/SDF-1α and inducing cell migration and invasion

The angiotensin II type I receptor (AGTR1) has a strong influence on tumor growth, angiogenesis, inflammation and immunity. However, the role of AGTR1 on lymph node metastasis (LNM) in breast cancer, which correlates with tumor progression and patient survival, has not been examined. AGTR1 was highly expressed in lymph node-positive tumor tissues, which was confirmed by the Oncomine database. Next, inhibition of AGTR1 reduced tumor growth and LNM in orthotopic xenografts by bioluminescence imaging (BLI). Losartan, an AGTR1-specific inhibitor, decreased the chemokine pair CXCR4/SDF-1α levels in vivo and inhibited AGTR1-induced cell migration and invasion in vitro. Finally, the molecular mechanism of AGTR1-induced cell migration and LNM was assessed by knocking down AGTR1 in normal cells or CXCR4 in AGTR1high cells. AGTR1-silenced cells treated with losartan showed lower CXCR4 expression. AGTR1 overexpression caused the upregulation of FAK/RhoA signaling molecules, while knocking down CXCR4 in AGTR1high cells downregulated these molecules. Collectively, AGTR1 promotes LNM by increasing the chemokine pair CXCR4/SDF-1α and tumor cell migration and invasion. The potential mechanism of AGTR1-mediated cell movement relies on activating the FAK/RhoA pathway. Our study indicated that inhibiting AGTR1 may be a potential therapeutic target for LNM in early-stage breast cancer.


INTRODUCTION
Breast cancer is considered to have the characteristic of easily metastasizing to lymph nodes, which is a major prognostic factor in early-stage breast cancer [1,2]. Patients with lymph node metastasis have a higher risk of loco-regional relapse and systemic metastases [3,4]. The number of involved regional lymph nodes, along with tumor size and distant metastasis, forms the basis of the majority of tumor staging schemes [5,6]. Axillary lymph node dissection (ALND) has been a classic adjuvant treatment for AGING early-stage breast cancer patients with sentinel-node involvement. However, according to clinical trials IBCSG 23-01 [7] and ACOSOG Z0011 [8], ALND does not provide a survival benefit because of complications such as lymphoedema or limited arm movement. Therefore, radiotherapies, chemotherapies and targeted metastatic lymph node therapies may be prospective ways to control nodal disease and future distant dissemination.
Lymph node metastasis is a complex process that involves the migration of tumor cells towards tumorassociated lymphatic vessels, cell trafficking by chemokines and cytokines, successful seeding into draining lymph nodes and remodeling of the lymph node microenvironment. Chemokines [9], a superfamily of cytokine-like molecules, are secreted from organs and induce the directional migration of tumor cells that express specific chemokine receptors [10,11]. A growing number of chemokine ligands and their receptors, including CCL19 and CCL21 for CCR7, CXCL13 for CXCR5, CCL1 for CCR8 [12,13] and CXCL12 (also known as SDF-1α) for CXCR4, have been demonstrated to attract tumor cells to lymph nodes. Muller's study [14] suggested that CXCR4/SDF-1α particularly have prominent roles in primary and metastatic breast cancer, as well as a number of other important malignancies, including lung, brain, and prostate cancers. CXCR4 is highly expressed in human breast cancer cells, while SDF-1α exhibits peak levels of expression in organs such as the lymph node, lung and brain [15]. In vivo, neutralizing the interactions of CXCR4/SDF-1α significantly impairs breast cancer cell metastasis to regional lymph nodes [16][17][18].
A variety of biological molecules, including SIX1 [19], PDGF-D [18], and SEMA4C [20], have been implicated in lymph node metastasis. Apart from these factors, the angiotensin II (Ang II) type I receptor (AGTR1) may also affect lymph node metastasis. AGTR1, a component of the renin angiotensin system (RAS), which classically regulates cardiovascular homeostasis, has displayed the potential to stimulate cell growth [21][22][23][24], migration or invasion [25,26] and to promote angiogenesis [27], inflammation and immunity [28,29]. Accumulating evidence has revealed that the inhibition of AGTR1 by antagonists such as losartan [30,31] and candesartan [32] can suppress angiogenesis, thereby contributing to the suppression of tumor growth and blood metastasis. However, the role of AGTR1 in lymph node metastasis of breast cancer has rarely been described.
There is a close correlation between Ang II/AGTR1 and CXCR4/SDF-1α, both of which are G protein-coupled receptor pathways, have the same downstream molecules and mediate cell migration [33]. In intrahepatic cholangiocarcinoma, activated hematopoietic stem cells (HSCs) promote tumor fibrogenesis, tumor progression and distant metastasis by mediating the epithelialmesenchymal transition (EMT) via the Ang II/AGTR1 and CXCR4/SDF-1α axes [34]. Therefore, we hypothesized that the relationship between AGTR1 and CXCR4/SDF-1α has a great influence on lymph node metastasis in breast cancer.
To explore these unanswered questions, in the present study, we sought to identify the AGTR1 expression profile from clinical breast cancer samples. Then, we defined the oncogenic role of AGTR1 in lymph node metastasis in vivo. In addition, we examined the effect of the AGTR1 inhibitor losartan on CXCR4/SDF-1α expression in mouse models and tumor cells. Finally, we explored the potential mechanism of AGTR1mediated cell migration and invasion and aimed to clarify whether losartan is beneficial for lymph node metastasis prevention and treatment in breast cancer.

The association of AGTR1 with clinicopathological data in breast cancer tissues
To investigate the relationship and significance between AGTR1 and clinical pathological factors, we conducted IHC staining in breast cancer tissue sections, and positive AGTR1 was observed in 275 cases. As indicated in Table 1, the number of lymph node-positive patients with AGTR1 high was 87, compared with 71 lymph node-negative patients with AGTR1 low . Therefore, higher expression of AGTR1 protein significantly correlated with lymph node metastasis (P<0.05) but not with age, histological grade or tumor size. Representative images of IHC in lymph node positive and negative tissues are shown in Figure 1A. Quantification of AGTR1 by the HSCORE method (shown in Figure 1B) suggested a significant increase in its expression in breast cancer tissues with lymph node metastasis (2.16 in LN+ vs. 1.52 in LN-), while similar results were observed in ER+ or HER2-samples (Supplementary Figure 1).
We searched the Oncomine database for AGTR1 expression in breast cancer in lymph node-positive and lymph node-negative breast cancer patients and found that high expression of AGTR1 mRNA also occurred in lymph node-positive patients compared with lymph node-negative patients [0.12 vs. -0.88 (log2 mediancentered intensity)] ( Figure 1C). These results suggested that the upregulation of AGTR1 in primary tumors was associated with lymph node metastasis in breast cancer.

Inhibition of AGTR1 reduces tumor growth and lymph node metastasis
To study the effect of AGTR1 on tumor growth and lymph node metastasis, we prepared two types of orthotopically implanted tumors, Balb/c-nu nude mice bearing MDA-MB-231 tumors and Balb/c mice bearing 4T1 tumors, and chose losartan, an AGTR1 antagonist, to block AGTR1. Saline or losartan (40 mg/kg, body weight) was administered by gavage (i.g.) every day in the control and losartan groups, respectively, for two weeks (mice with MDA-MB-231 tumors) or one week (mice with 4T1 tumors) after implantation. The growth curves showed that losartan administration resulted in a significant reduction in tumor growth (Figure 2A and 2B). In parallel, we used bioluminescence imaging (BLI) to monitor tumor growth in MDA-MB-231 and 4T1 tumors. Cells were transfected with lentivirus expressing the firefly luciferase gene and puromycin resistance gene. After selection with puromycin, the surviving colonies were screened under an in vivo imaging system ( Figure 2C). We observed that orthotopically implanted tumors in the control group displayed significantly stronger firefly luciferase signals than those in the losartan group (Ctrl: 4482±947. To detect lymph node metastasis in mice bearing MDA-MB-231 and 4T1 tumors, we used indirect BLI to monitor four groups of lymph nodes (proper axillary, accessory axillary, subiliac, and popliteal lymph nodes) resected from mice. We observed that mice bearing MDA-MB-231 tumors in the control group developed metastases to lymph nodes with significantly stronger firefly luciferase signals (Ctrl: 430±169.8 vs. LOS: 31±31) ( Figure 2E). Calculation of the number of positive lymph nodes from each group (total number of LN=32) showed that the proportion was 62.5% in the Ctrl group compared with 2.5% in the losartan group ( Figure 2H). The same results were also observed in 4T1 tumors in which losartan treatment significantly decreased lymph node metastasis by quantitative analysis of the luciferase signal and the ratio of lymph node metastasis ( Figure 2F and 2I).

Inhibition of lymph node metastasis by AGTR1 antagonism through decreasing CXCR4/SDF-1α levels in vivo
Tumor cells preferentially disseminate to specific distant organs through chemokine guidance, which shares the same mechanisms as those regulating leukocyte homing and trafficking. The higher expression of CXCR4 on tumor cells and intensive generation of SDF-1α from target organs form a major pathway for breast cancer metastasis [35]. Therefore, in our study, we examined CXCR4 and SDF-1α expression in mice with orthotopically transplanted tumors using IHC and RT-PCR. Inhibition of AGTR1 decreased CXCR4 protein levels in tumors (1.50 in Ctrl vs. 0.55 in LOS) ( Figure 2J and 2K) and SDF-1α levels in lymph nodes ( Figure 2L). Taken together, these data suggest that blocking AGTR1 inhibits lymph node metastasis via the CXCR4/SDF-1α pathway.

AGTR1 promotes tumor cell migration and invasion
To investigate the role of AGTR1 in breast cancer cells, MDA-MB-231 and MCF7 cells were stably transfected with lentivirus expressing the AGTR1 gene (AGTR1 expression in breast cancer cells; see Supplementary Figure 2). Western blot and RT-PCR analyses were performed to confirm the stable overexpression of AGTR1 ( Figure 3A and 3B). To evaluate the effect of losartan on cell viability, cells were treated with various concentrations of losartan (0, 100, 200, 300 μM) for 72 h. A significant killing effect on cell viability was observed at 300 μM of losartan in MOCK-MCF7 and MOCK-MDA-MB-231 cells ( Figure 3C and 3D). We next examined the effect of AGTR1 overexpression on cell viability upon stimulation with losartan for 72 h. AGTR1-induced cell proliferation was suppressed by losartan challenge (200 μM) in AGTR1 high cells ( Figure  3C and 3D). The results were consistent with the conclusions that AGTR1 promotes tumor growth in mouse models.
Cell migration and invasion are the essential mechanisms of lymph node metastasis in tumors, which may be the initial step in dissemination from primary tumors. To investigate the effect of AGTR1 overexpression on cell migration and invasion, transwell assays were performed. MDA-MB-231 (5×10 4 cells/well) and MCF7 (2×10 5 cells/well) cells were seeded into transwell upper chambers with 200 μl of serum-free medium and/or 100 μM losartan for migration and Matrigel invasive assays, while 600 µl of 20% FBS medium and/or 100 μM losartan was added to the lower chambers. After incubation for 16 h (MDA-MB-231) or 48 h (MCF7) at 37°C, as shown in Figure  3E and 3F, the number of AGTR1 high -MDA-MB-231 and AGTR1 high -MCF cells increased significantly on the bottom of the transwell membrane. In contrast, losartan treatment inhibited migratory and invasive abilities in both the MOCK and AGTR1 high groups. Similar to the results reported by Seo et al [36], AGTR1 accelerated the aggressiveness of breast cancer cells.

AGTR1 promotes proliferation, migration, invasion and lymph node metastasis via CXCR4
Based on the decrease of migration, invasion, lymph node metastasis and CXCR4 expression by losartan, it was hypothesized that AGTR1 plays its key role via CXCR4. We established an MDA-MB-231 xenograft model and found that the AGTR1 high group displayed a stronger capacity for growth (MOCK: 2299±241.5 vs. AGTR1: 8069±1451) and lymph node metastasis (MOCK: 108.5±80.18 vs. AGTR1: 1564±302.6) compared with MOCK-group tumors ( Figure 4A-4C and 4E). AMD3100, a highly specific CXCR4 antagonist, is approved to suppress the growth and metastasis of many types of tumors. We administered AMD3100 to nude mice carrying AGTR1 high -MDA-MB-231 tumors by intraperitoneal injection. The results revealed that AMD3100 could reverse AGTR1-induced tumor growth (AGTR1: 8069±1451 vs. AMD3100: 2923±557.1) and lymph node metastasis (AGTR1: . IHC results showed that AMD3100 could also impede CXCR4 expression on AGTR1 high tumor tissues (1.37 in MOCK vs. 2.14 in AGTR1 high vs. 1.38 in AGTR1 high +AMD3100) ( Figure  4F). We conducted hematoxylin and eosin (HE) staining in each group of lymph nodes to verify tumor metastasis; representative images are shown in Figure 4D.

AGING
Further researches confirmed the results in vitro by knocking down CXCR4 in MDA-MB-231-AGTR1 high and MCF7-AGTR1 high cells using siRNA. Decreased CXCR4 expression was confirmed by RT-PCR and Western blotting ( Figure 5A and 5B). CCK8 assays exhibited that AGTR1 high cells proliferated significantly faster than their MOCK cells in MDA-MB-231 and MCF7 cells, while siCXCR4 suppressed cells proliferation ( Figure 5C). Transwell assays with MDA-MB-231 (2×10 4 cells/well) and MCF7 (1×10 5 cells/well) cells seeding into upper chambers revealed that the enhanced number of AGTR1 high -MDA-MB-231 and AGTR1 high -MCF cells on the bottom of the AGING transwell membrane were inhibited significantly by the suppression of CXCR4 ( Figure 5D and 5E). All together, AGTR1 accelerates proliferation, migration, invasion and lymph node metastasis through upregulating CXCR4.

AGTR1 enhances CXCR4 expression in vitro
From the results above, we found that losartan inhibited CXCR4 expression in mouse tumors. We next determined the effect of AGTR1 on CXCR4 expression in vitro by Western blotting. The results indicated that CXCR4 levels increased significantly in AGTR1 high MDA-MB-231 cells and MCF cells, which was inhibited by losartan ( Figure 6A).
In tumor cells, high levels of actin polymerization are required for the formation of pseudopodia, which are necessary for the spreading of malignant cells into tissues and colonization. To verify the inhibitory effect of losartan on CXCR4, we treated MDA-MB-231 cells AGING with SDF-1α and found that SDF-1α-treated tumor cells showed a distinct cytoskeletal redistribution of F-actin stress fibres and lamellipodia formation beginning 20 min after SDF-1α exposure by phalloidin staining. Pretreatment of breast cancer cells with 100 μM losartan (24 h, 37°C) inhibited SDF-1α-induced actin polymerization and lamellipodia formation ( Figure 6B). These results confirmed our findings that losartan blocked the effect of SDF-1α by inhibiting CXCR4.
To investigate whether the inhibition of AGTR1 expression can decrease CXCR4 expression, we knocked down AGTR1 in MDA-MB-231 and MCF7 cells using siRNA. Decreased AGTR1 expression was confirmed by RT-PCR and Western blotting ( Figure 6D and 6E). The results showed that AGTR1 knock-down cells exhibited lower CXCR4 levels, indicating that the effect of siAGTR1 was the same as that of losartan ( Figure 6A). Moreover, we searched the TCGA database and found that there was a significant correlation between AGTR1 and CXCR4 (R 2 =0.9814) ( Figure 6C).

AGTR1 induces the expression of focal adhesion kinase (FAK)/Ras homolog gene family member A (RhoA) signaling molecules
FAK is a cytoplasmic protein tyrosine kinase involved in cytoskeletal remodeling and cell adhesion structures [37,38]. Together with FAK, RhoA and its downstream

AGING
ROCKs are also involved in cell constriction and pseudopodia formation, which are required for cell migration [39,40].
We therefore determined whether FAK/RhoA signaling pathways are activated in AGTR1-overexpressing tumor cells. Western blotting showed that p-FAK, RhoA, ROCK1 and ROCK2 levels increased significantly in AGTR1 high cells, while losartan decreased these levels in both MOCK cells and AGTR1 high cells ( Figure 7A).

CXCR4 is crucial for AGTR1-induced FAK/RhoA molecule upregulation
For breast cancer, studies have provided evidence for the direct involvement of FAK and members of the Rho family in the CXCR4/SDF-1α-induced migration of tumor cells. The above results showed that AGTR1 regulated the expression of CXCR4 and FAK/RhoA signaling molecules. In this experiment, we attempted to clarify whether AGTR1 can activate the FAK/RhoA signaling pathway through CXCR4.
These results demonstrate that CXCR4 signaling is critical for AGTR1-induced lymph node metastasis.

DISCUSSION
The RAS is frequently dysregulated in malignancy, which correlates with poor patient outcomes [41]. The main enzymatic components of the RAS include renin and angiotensin converting enzyme (ACE). The main product of ACE, Ang II, exerts a major effect via binding with specific AGTR1. Vinson first reported in early 1995 that AGTR1 had a possible link with breast cancer and later confirmed this finding in many tumor types, such as ovarian cancer, prostate cancer, and pancreatic cancer [42]. Losartan, an angiotensinreceptor blocker (ARB) that is widely used in the cardiovascular system, has great effects in preventing tumor progression in mice with induced breast cancer. However, there has been debate about the therapeutic role of ARBs in cancer patients over the years. One meta-analysis [43] suggested that ARBs were associated with a modestly increased risk of a new cancer diagnosis, while a more recent meta-analysis [44] confirmed that there was no difference in the risk of cancer with ARBs. Therefore, additional studies are needed to define the effect of ARBs on cancer incidence and mortality.

AGING
It has been extensively published that AGTR1 levels may correlate with tumor size (T stage) and distant metastasis (M stage) in some cases [45][46][47][48]. However, with regard to lymph node metastasis (N stage), Puddefoot [49] and Rodrigues [26,50] revealed that Ang II reduces cell adhesion and metastasis, which correlates with lymph node metastasis indirectly. Here, we showed a direct function by which AGTR1 is involved in promoting lymph node metastasis. We found that AGTR1 expression in lymph node-positive patients was higher than that in lymph node-negative patients in our clinical samples and the Oncomine database. Moreover, the inhibition of AGTR1 decreased lymph node metastasis in orthotopic MDA-MB-231 and 4T1 breast carcinomas.
Interestingly, in our study, we did not find significant changes in VEGF-C mRNA in different mouse tumors (Supplementary Figure 3). A well-known hypothesis called 'seed and soil' describes the crosstalk between tumor cells and the microenvironment [56]. During this process, chemokines play a critical role in guiding tumor cells (the 'seed') to lymph nodes (the 'soil') [57]. Among all of these chemokines in breast cancer, CXCR4/SDF-1α is a major chemokine pair involved in lymph node metastasis [1]. CXCR4 is expressed by almost all cancer types, suggesting that the CXCR4/SDF-1α pair may be involved in site-specific metastasis formation in a large number of malignant diseases [58]. Therefore, we first investigated whether AGTR1 regulated lymph node metastasis through CXCR4/SDF-1α. In vivo, we found that losartan decreased CXCR4 levels in tumor tissues and SDF-1α at the mRNA level in lymph nodes. We found that blocking CXCR4 using AMD3100 on the basis of AGTR1-overexpression tumors could abolish AGTR1induced lymph node metastasis, suggesting that CXCR4 signaling directly mediates AGTR1-induced metastasis.
Previous studies have shown that neutralizing the interactions of CXCR4/SDF-1α leads to a significant inhibition of lymph node and lung metastasis [1,59]. Therefore, our data strongly suggest that blocking AGTR1 reduces lymph node metastasis in vivo and that this effect is likely mediated via CXCR4/SDF-1α. In addition, SDF-1α binds to CXCR7, another chemokine receptor that is highly expressed on breast cancer cells, and enhances CXCR7-mediated tumor migration and metastasis by activating STAT3, MMP9, MMP2 and VCAM-1 [60]. Apart from CXCR4/SDF-1α, CCR7-CCL19/CCL21 [61] are also key players in cell www.aging-us.com 3980 AGING dissemination via the lymphatic system, but the level of CCL21 in lymph nodes was not influenced by losartan in our study (Supplementary Figure 4).
Another essential mechanism for inducing lymphatic metastasis is the migratory and invasive capacity of tumor cells [9]. Our observations revealed that AGTR1 accelerated breast cancer cell migration and invasion. There is evidence that in certain cancer types, such as gastric cancer, ovarian cancer, lung cancer and choriocarcinoma, Ang II/AGTR1 signaling is associated with the upregulation of a range of target genes that play a role in MMP-2 and MMP-9 activation and the induction of ICAM-dependent adhesion, inducing cell migration and EMT. EMT exhibits a disruptive effect on cell-cell junctions and promotes invasion into lymphatics, which was first revealed in studies of embryo implantation and embryogenesis [62,63]. Our findings were consistent with those results.
Using orthotopically implanted mice, we found that losartan decreased CXCR4 expression. Therefore, in vitro, we validated those results and further clarified the relationship between AGTR1 and CXCR4 using RNA interference to knock down AGTR1. Both losartan and siAGTR1 decreased CXCR4, which indicated that AGTR1 may regulate the downstream molecule CXCR4. Rizzo's research has also revealed the close relationship between Ang II/AGTR1 and SDF-1α/CXCR4 in kidney diseases, indicating that increased Ang II could activate podocytes to release SDF-1α, which binds to its receptor CXCR4 located on parietal epithelial cells (PECs) [64]. A variety of CXCR4 inhibitors, including TI40, TNI4003, AMD3100, and AKBA, have also shown encouraging anticancer effects in pre-clinical studies [65]. In fact, CXCR4 antagonists have been translated into clinical trials, as presented at the 2017AACR-NCI-EORTC International Conference. X4P-001-IO, a CXCR4 inhibitor, in combination with axitinib has exhibited a promising disease control rate (DCR) of 93% in clear cell renal cell carcinoma, which was announced by Atkins' team. In HER2-negative metastatic breast cancer, the CXCR4 inhibitor balixafortide in combination with eribulin in phase 1 clinical trials exhibited a longer progression-free survival (PFS) than that for eribulin only; the one-year overall survival rate was 76% (balixafortide plus eribulin) vs. 54% (eribulin) [66]. From our study, we determined that losartan could inhibit CXCR4 expression in both mouse tumors and cells and may exhibit a similar therapeutic effect as CXCR4 inhibitors. Losartan may provide a safer and more economic strategy for CXCR4induced lymph node metastasis and relapse.
SDF-1α mediates actomyosin cytoskeletal rearrangements and thereby regulates cell migration [67]. After pretreating MDA-MB-231 cells with losartan and then SDF-1α for 20 min, we found that losartan reduced SDF-1α-induced pseudopod formation by phalloidin staining. This effect validated that losartan decreased CXCR4 expression on cell membranes, which reduced the interaction between CXCR4 and SDF-1α. Actually, SDF-1α expression is higher in lung, pancreatic, and esophagogastric tumors and is associated with hypoxia-induced angiogenesis [68]. However, SDF-1α expression is lower in breast tumors, and these differences are reasonable because high levels of SDF-1α in primary tumors promote local invasion, which contributes to poor survival [69]. Low SDF-1α in breast cancer promotes its migration to distant organs. In our study, we showed that losartan could decrease both CXCR4 in primary tumors and SDF-1α in lymph nodes, which is an essential mechanism for reducing lymph node metastasis in breast cancer.
FAK/RhoA signaling molecules are widely accepted as important promoters of directional cell movement [38,39]. Therefore, we investigated whether AGTR1 affects the FAK/RhoA signaling pathway. Our results showed that AGTR1 increased FAK/RhoA molecules and their downstream effectors ROCK1, ROCK2 and MLC, which agreed with previous studies showing that FAK/RhoA is activated by Ang II/AGTR1. FAK, which upon Ang II-induced activation translocates to sites of focal adhesion with the extracellular matrix and promotes the phosphorylation of cytoskeletal proteins such as paxillin and talin, which together play a role in the regulation of cell morphology and migration. In vascular smooth muscle cells (VSMCs), Ang II rapidly increases the phosphorylation of FAK, and losartan inhibits the phosphorylation of FAK and further Ang IImediated VSMC migration [70]. AGTR1 also influences RhoA, whose function is to activate G12/13 and the subsequent interaction between its G subunit and RhoGEF, thereby further activating RhoA and leading to the stimulation of ROCK [71].
CXCR4/SDF-1α activates members of the FAK family of protein tyrosine kinases and further upregulates RhoA to stimulate ROCK, which promotes cell contractility and migration [72][73][74][75]. Therefore, we clarified the role of CXCR4 in AGTR1-mediated cell migration through the FAK/RhoA pathway by knocking down CXCR4 in AGTR1 high cells. For the first time, we revealed that AGTR1 overexpression increased FAK/RhoA signaling molecules, which could be decreased by siCXCR4 (Figure 8).
In conclusion, we found a new role of AGTR1mediated tumor metastasis in our study. For the first time, we reveal that inhibition of AGTR1 by losartan reduces lymph node metastasis through CXCR4/SDF-AGING 1α. Our findings provide evidence for clinical treatments using ARBs by targeting lymph node metastasis in breast cancer patients with hypertension. Both ARBs and CXCR4 inhibitors have been translated into clinical trials in cancer patients, and additional clinical trials are needed to accelerate the development of breast cancer treatment strategies.

Sample collection
Breast cancer samples were acquired from the pathology department, and the patients did not receive preoperative radiotherapy or chemotherapy. The obtained specimens and clinicopathological data (age, histological grade, tumor size, lymph node metastasis, ER, PR, HER2 status) were conducted in accordance with the ethical standards and according to the Declaration of Helsinki and according to national and international guidelines and were approved by the Ethics Committee of the Medical Faculty of Tongji Medical College. The samples were used to prepare tissue microarray slides and were then subjected to AGTR1 IHC staining (Supplementary Table 1).

Immunohistochemistry and HE staining
Sections were deparaffinized in xylene three successive times and then hydrated in different percentages of ethanol (100%, 90% and 75%). Slides were soaked in citrate buffer and boiled for 15 min at 100°C in a microwave oven. After cooling to room temperature, the slides were placed into 3% hydrogen peroxide for 20 min to block endogenous peroxide activity and then incubated with 5% BSA for 30 min. The samples were stained with primary antibodies at 4°C overnight. After treating with secondary antibody and diaminobenzidine (DAB) for 4 min in a dark room, hematoxylin was used to stain the nucleus. The slides were placed on coverslips and visualized under a light microscope.
The immunohistochemical score (IHS) was calculated by multiplying an estimate of the percentage of immunoreactive cells (quantity score) by an estimate of the staining intensity (0, negative; 1, weak; 2, moderate and 3, strong). An HSCORE >2 was considered high expression, while an HSCORE ≤2 was considered low expression.
Lymph nodes were embedded in paraffin for histopathological examination. After tissue sectioning, mounting onto slides and staining with HE, histological changes were assessed by two independent pathologists and were visualized using a light microscope.

RNA isolation and RT-PCR
Lymph node (ground into cell suspension) and breast cancer cell RNAs were extracted and reverse transcribed to cDNA using a reverse transcriptional kit (TaKaRa Bio, Inc.

BLI
Cells expressing firefly luciferase were seeded into 96well plates. After adherence and the addition of Dluciferin (Xenogen, 150 mg/kg) to each well, the cells were screened by IVIS. Mice with firefly luciferase were anaesthetized and injected with D-luciferin (i.p.). After 10 min, the mice were placed in the IVIS to monitor the tumor size. To detect lymph node metastasis, proper axillary, accessory axillary, subiliac and popliteal lymph nodes of each mouse were resected for indirect BLI. The number of positive lymph nodes and firefly luciferase signals in each group were calculated.

Cell proliferation assay
MOCK and AGTR1 high cells (8×10 3 cells/well) were seeded in 96-well plates and treated with different doses of losartan (0, 100, 200, 300 μM) for 24 h, 48 h, and 72 h (Supplementary Figures 5-9). CCK8 reagents were added to each well for 1 h, and the OD value was expressed as the absorbance at 450 nm.

Transwell assay
MDA-MB-231 and MCF7 cells were seeded into transwell (Corning Costar) upper chambers with 200 μl of serumfree FBS medium and/or 100 μM losartan for migration and Matrigel invasive assays, while 600 µl of 20% FBS medium and/or 100 μM losartan was added to the lower chambers. After incubation for 16 h (MDA-MB-231) or 48 h (MCF7) at 37°C, the upper chambers were soaked in 4% paraformaldehyde and then stained with crystal violet. The cells in the upper chamber were completely removed with cotton swabs, and the cells on the reverse side of the membrane were imaged and counted under a microscope.

siRNA transfection
Cells were seeded into 6-well plates at a density of 2 × 10 5 cells/well and incubated overnight. A total of 5 μl of siRNA was added to 500 μl of Opti-MEM with RNAiMAX in each well according to the manufacturer's instructions. After 48 h, RNA and protein were extracted for testing the transfection efficiency. The siRNA sequences were as follows: siAGTR1-1: GCAGTAGCCAGCAATTTGA siAGTR1-2: ATAAGAAGGTTCAGATCCA siCXCR4-1: CAGCUAACACAGAUGUAAATT siCXCR4-2: GCGUGUAGUGAAUCACGUATT

Lentivirus transfection
AGTR1 lentivirus (Huameng Biotechnology, Ltd.) was transfected into luciferase-expressing or non-luciferaseexpressing MDA-MB-231 cells for animal experiments or cell experiments, respectively. After transfection, puromycin (2 μg/ml) was used to select luciferaseexpressing cells and hygromycin (600 μl/ml) for AGTR1 high cells. The transfection efficiency was detected using PCR and Western blotting.

Actin polymerization assay
The seeded MDA-MB-231 cells on sterile cover slips were untreated or treated with 100 μM losartan for 24 h, and SDF-1α was added for 20 min. Then, the cells were fixed in 4% formaldehyde for 20 min and incubated in 0.1% Triton X-100 with 5% BSA for 30 min. After staining with 100 nM rhodamine phalloidin, the actin filaments in the cells were captured and analyzed using a Zeiss confocal photomicroscope (Olympus FluoView FV1000).

Reagents and antibodies
One losartan potassium tablet (50 mg, 461.01 g/mol, Merck Sharp & Dohme (Australia) Pty., Ltd.) was dissolved in 9375 μl of saline for intragastric administration, and one tablet was dissolved in 1084.6 μl of PBS for cell experiments.

Statistical analyses
All data are presented as the mean±standard deviation (SD). Student t test (2-tailed, unpaired) was used for significance analysis. The P value<0.05 was considered significant.

CONFLICTS OF INTEREST
The authors declare no conflict of interest with the current manuscript.