The following antibodies were used for the animal experiments: mCD152 (mCTLA-4) monoclonal antibody (9H10; Bio X Cell), mPD-1 monoclonal antibody (RMP1-14; Bio X Cell), anti-mouse CD8 monoclonal antibody (116-13.1, Bio X Cell), polyclonal Syrian hamster IgG (Isotype Controls of anti-mouse CTLA-4, Bio X Cell ), rat IgG2a isotype control (Isotype Controls of anti-mouse PD-1, Bio X Cell ), and mouse IgG2a isotype control (Isotype Controls of anti-mouse CD8, Bio X Cell).

4T1 (murine breast cancer cells) and CT26 (murine colorectal adenocarcinoma) were purchased from American Type Culture Collection (ATCC). Both tumor cell lines were grown in RPMI-1640 medium supplemented with 10% fetal bovine serum at 37°C, 5% CO₂. For hypoxic culture, cells were cultured in a hypoxia chamber (37°C with an atmosphere of 1% O₂ and 5% CO₂, balanced with N₂, and humidified). For normoxic culture, cells were cultured in an atmosphere containing 21% O₂ (37°C and an atmosphere of 5% CO₂ equilibrated with atmospheric O₂ in a humidified incubator).

The designed shRNA construct, as previous studies [16], contained a unique 19-nt double-stranded AGT target sequence that presented as an inverted complementary repeat, a loop sequence (5’-CTCGAG-3’), the RNA PloIII terminator (5’-TTTTTT-3’), and 5’ single-stranded overhangs for ligation into AgeI- and EcoRI-digested Pglv-u6-Puro lentivirus vector (GenePharma, Shanghai, China). The recombinant vector was named pGLV–AGT–shRNA. The negative control vector (pGLV–NC–shRNA) contained a nonsense shRNA insert in order to control any effects caused by non-RNAi mechanisms. We co-transfected the 293 T cells with three optimized packaging plasmids (pGag/Poll, pRev and pVSV-G) and the pGLV-AGT-shRNA or Pglv-NC-shRNA expression clone construct, which produced lentiviral stocks with a suitable titer. Stably transduced 4T1 and CT26 cells were selected using puromycin, adding the minimum concentration of puromycin required to kill untransduced 4T1 or CT26 cells. The efficiency of knockdown was detected by real-time qPCR and western blotting.

Total RNA was extracted from cells using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Then, reverse transcription was performed with the PrimeScript® RT reagent kit (Takara), followed by real-time PCR using an ABI 7500 Sequence Detection System with a SYBR® Premix Ex Taq™ kit (Takara). The sequences of the specific PCR primers were as follows:

AGT: Forward: 5′-TGAAGGATACACAGAAGCAA-3′

Reverse: 5′-TGGTAAAGGAGATGGAAGG-3′

GAPDH: Forward: 5′- TGTCGTGGAGTCTACTGGTG-3′

Reverse: 5′-GCATTGCTGACAATCTTGAG-3′

GAPDH was used as the internal control.

Whole cell lysates were harvested and Western blotting was conducted as described previously [16]. Primary antibodies are listed in Table 1. Blots were visualized using an ECL detection kit (Millipore). Antibodies for Western blotting were listed in Additional file 1: Table S1.Tab1

Cytokines influenced (alteration>2folds) by AGT silence

Cell proliferation ability was detected by MTT assays. Cells were plated into 96-well plates (2×10³ cells/well) in 200 μL of medium. After incubation for 8 h, cells were exposed to candesartan (10μm/well, a selective AT1R blocker), PD123319 (10μm/well, a selective AT2R blocker), a combination of these (10μm/well candesartan and 10μm/well PD123319), and control DMSO for 0, 24, 48, 72, and 96 h. At each time point, 20 μL of 5 mg/mL MTT solution in PBS was added to each well for 4 h at 37 °C. Subsequently, a 20% sodium dodecyl sulfate (SDS) solution in 0.01% HCl (150 μL) was added to each well, and the absorbance at 570 nm was measured using a spectrophotometric plate reader (ND-1000, Wilmington, DE). Two independent experiments were performed, each in triplicate.

Sirius Red staining was performed as previously described [28]. Formalin-fixed tissues were embedded in paraffin and sectioned at 5 μm thickness.Sirius Red staining for collagen was performed using 0.1% Sirius Red (Direct Red80; Sigma) and counterstained with Weigert’s hematoxylin.

The following anti-mouse antibodies were used for flow cytometry: Live/Dead Fixable Near IR Dead Cell Stain (Life Technologies), CD45-PerCP-Cy5.5 (BD Biosciences), CD3-PE (BD Biosciences), CD44-PE (BD Biosciences), CD8a-APC (BD Biosciences), CD4-APC (BD Biosciences), Foxp3-PE (BD Biosciences), CD11b-PE-Cy7 (BD Biosciences), F4/80-APC (BD Biosciences), Ly6C-PE (BD Biosciences), Ly6G-FITC (BD Biosciences), and Ly6G-APC-Cy7 (BD Biosciences). Flow cytometry was performed with Canto II (BD) and the data were analyzed with FlowJo 7.6 software (TreeStar). Live/dead cell discrimination was performed using Live/Dead Fixable Aqua Dead Cell Stain Kit (Life Technologies). All of cell surface or intracellular stainings were done according to the manufacturer's instructions. T effector cells (T_eff) were phenotyped as CD45⁺CD8⁺CD44⁺, regulatory T cells (T_regs) as CD45⁺CD4⁺Foxp3⁺, macrophages as CD45⁺CD11b⁺F4/80⁺, tumor-associated macrophages (TAMs) as CD45⁺CD11b⁺F4/80⁺CD206⁺, monocytic myeloid-derived suppressor cells (Mo-MDSCs) as CD45⁺CD11b⁺Ly6G^lowLy6C^high, and granulocytic myeloid-derived suppressor cells (G-MDSCs) as CD45⁺CD11b⁺Ly6G^highLy6C^low.

AngII, IL-10, GM-CSF, G-CSF, Eotaxin-2, and CXCL11 ELISA kits were all obtained from Ray Biotech, Inc. (Norcross, GA, USA). TNFSF14 ELISA kits was obtained from CUSABIO Inc. (Wuhan, China). The ELISA for them was carried out according to the manufacturer’s instructions. Two independent experiments were performed, each in triplicate. For AngII ELISA, the microplate in the kit, which is pre-coated with anti-rabbit secondary antibody, was incubated with an anti-AngII antibody, under such conditions that both biotinylated AngII peptide and a peptide standard or targeted peptide in the samples interacted competitively with the AngII antibody. Unbound biotinylated AngII peptide was then allowed to interact with streptavidin-horseradish peroxidase (SA-HRP), which catalyzes a color development reaction. The intensity of the colorimetric signal is directly proportional to the amount of the biotinylated peptide-SA-HRP complex and inversely proportional to the amount of the AngII peptide in the standard or samples. A standard curve of known concentration of AngII peptide was established, and the concentration of AngII peptide in the samples was then calculated by interpolation onto the standard curve. All groups of tumor cells were all seeded at a density of 200,000 cells/well in 1 ml medium in 24-well plates.

Initially, 5-μm frozen tissue serial sections were fixed with cold acetone for 15 minutes at 4°C and then washed three times for five minutes each with phosphate-buffered saline. Slides were blocked for one hour in 5% bovine serum albumin at room temperature. Primary antibodies were incubated overnight at 4°C. Following additional washes, the slides were incubated for one hour at room temperature with the appropriate secondary antibody and 4',6-diamidino-2-phenylindole (DAPI) counterstaining. To reduce autofluorescence, the slides were subsequently incubated in 0.3 M glycine for 10 minutes and then mounted in Hydromount aqueous mounting medium (Fisher Scientific). Images were acquired with a fluorescence microscope (Olympus BX51). Antibodies used for the immunofluorescence are listed in Additional file 1: Table S1.

Cytokine profiles (containing 308 cytokines) of 4T1 cells in normoxic or hypoxic conditions were analyzed using the RayBio Mouse Cytokine Antibody Array. Protein extraction: total protein was extracted from the cells with ice-cold Cell & Tissue Protein Extraction Reagent (KangChen. Cat. # KC-415, China), which contains inhibitors for protein degradation (5ul Protease Inhibitor Cocktail, 5μl PMSF and 5μl Phosphotase Cocktail were added into the 1 ml Protein Extraction Reagent). Determining the protein concentration: the protein concentration was determined by using a BCA Protein Assay Kit (KangChen KC-430, China). Blocking and incubation: a. Protein array membranes were blocked in blocking buffer for 30 min and then incubated with samples at room temperature for 1 to 2h (or incubated at 40 °C overnight). b. Samples were then decanted, and the membranes were washed with washing buffer. After that, the membranes were incubated with diluted biotin-conjugated antibodies at room temperature for 1–2 h. Chemiluminescent detection: the membranes were washed with washing buffer and then reacted with HRP-conjugated streptavidin (1:1000 dilution) at room temperature for 2 h. The membranes were then washed thoroughly and exposed to detection buffer in the dark before being exposed to X-ray film. Afterwards, the membranes were exposed to X-ray film and the image was developed using a film scanner. Data analysis: by comparing the signal intensities, relative expression levels of cytokines were obtained. The intensities of the signals were quantified by densitometry. Positive controls were used to normalize the results from different membranes being compared. Fold changes in protein expression were calculated.

All experimental procedures were approved and overseen by Southern Medical University Institutional Animal Care and Use Committee, and were performed in accordance with the guidelines and regulations for animal experiments set down by Southern Medical University. Four-week-old female BALB/c mice, BALB/c nude mice and NOD/SCID mice were used for the different animal experiments. 4T1 cells (5 × 10⁶ cells) with or without AGT silencing via AGT shRNA lentiviral transduction were inoculated into the mammary fat pads of each mouse. CT26 cells (5 × 10⁶ cells) with or without AGT silencing were inoculated s.c. into the flank of each mouse. When the tumors were allowed to grow for the indicated time, different drugs or control agents were administered by intraperitoneal injection. Antibodies used for the in vivo immune checkpoint blockade experiments were given intraperitoneally at a dose of 250 μg/mouse every three days (four times total) and included: anti-CTLA4 (9H10), anti-PD-1 (RMP1-14), polyclonal Syrian hamster IgG, and rat IgG2a isotype control. Mice CD8-depletion in vivo was performed according to a previous study [29]. In brief, anti-CD8 antibody or an IgG2a isotype control at a dose of 200 μg/mouse was given 2 days prior to tumor implantations (day −2), day 0, and then every 4 days for the duration of the experiment. To explore the influence of Ang II blockers on tumor growth in vivo, candesartan (5mg/kg, i.p., Takeda Pharmaceutical Company Limited, Japan) and PD123319 (5mg/kg, i.p., Selleck Chemicals) were given every other day starting from day 5 after cell injection. Perpendicular tumor diameters were measured using calipers. Tumor volume was calculated using the formula L × W² × 0.5, where L is the longest diameter and W is the shortest diameter.

Statistical analyses of AngII, GM-CSF and G-CSF levels by ELISA assay, the percentages of Teffs, Tregs, TAMs and MDSCs by Flow cytometry and Teffs / Tregs ratio were performed by using Student’s t test with SPSS 13.0 software. Two-way cell growth in vitro and tumor growth in vivo were evaluated using SPSS 13.0 software with two-way repeated-measures ANOVA. Analyses of survival patterns in tumor-bearing mice were performed using the Kaplan-Meier method, and statistical differences were evaluated according to the Mantel-Cox log-rank test. A p value < 0.05 was considered statistically significant.

We first established syngeneic tumor models with 4T1 breast cancer cells in immune-competent BALB/c mice. To test the effect of AngII signaling on the 4T1 tumors, BALB/c mice bearing 4T1 tumors of moderate sizes were repeatedly treated with the AngII-receptor blockers candesartan for AT1R and PD123319 for AT2R. Although 4T1 tumor growth was slightly retarded by PD123319, significant inhibition of tumor growth was only observed when mice were treated by candesartan alone or a combination of them (Fig. 1a). To determine whether the anti-tumor growth effect of AngII signaling blockage was caused by directly inhibiting the proliferation of the 4T1tumor cells, the effect of AngII signaling blockage on 4T1 cell proliferation was evaluated in vitro by a MTT assay. We observed no difference in cell proliferative ability in vitro between the cells treated with candesartan, PD123319, combination of both, and DMSO (Fig. 1b). Furthermore, we performed the same in vivo experiment using BALB/c nude mice that were T-cell immunodeficient. Neither candesartan nor PD123319 could inhibit tumor growth in these T-cell immunodeficient mice (Fig. 1c). These results indicate that AngII signaling may be involved in the immune escape of 4T1 tumor cells in BALB/c mice.Fig1Fig. 1

Local AngII in tumor microenvironment is involved in immune escape of tumor cells. a The influence of AngII-receptor blockers on 4T1 tumor growth in syngeneic BALB/c mice; candesartan (cand, 5mg/kg, i.p.; starting at day 5 after cell injection, daily ) for AT1R and PD123319 (5mg/kg, i.p.; starting from day 5 after cell injection, every other day) for AT2R, n=5. b The influence of AngII-receptor blockers on 4T1 cells proliferation in vitro as detected by MTT assay; candesartan (10μm/well) and PD123319 (10μm/well). c The influence of Ang II-receptor blockers on 4T1 tumor growth in T-cell immunodeficient BALB/c nude mice, n=5. d The expression of AGT, a precursor of AngII, was silenced in 4T1 cells and CT26 cells by shRNA. Hypoxia remarkably induced generation of Ang II in 4T1 and CT26 cells as detected by ELISA. AGT-silencing greatly decreased Ang II levels, especially under hypoxic condition. Data are presented as mean±SEM, n=3. e AGT silencing did not influence cell-proliferation of 4T1 cells in vitro as detected by MTT assay. f AGT silencing obviously inhibited tumor growth of 4T1cells in BALB/c mice comparing to negative control, and the depletion of CD8⁺ T cells in mice by in vivo administration of monoclonal antibody against CD8 greatly attenuated this effect (n=6). Ns, no significance; *, P < 0.05; **, P < 0.01

Because it has been well-documented that 4T1 tumors are highly resistant to PD1 and CTLA-4 checkpoint immunotherapy [26, 30], we sought to determine whether an AngII signaling blockage could enhance the response of 4T1 tumors to checkpoint immunotherapy. BALB/c mice with 4T1 tumors were repeatedly treated by intraperitoneal injection (i.p.) with anti-PD-1 or anti-CTLA-4 antibodies as single agents or in combination with candesartan. We did not observe any significant inhibition of tumor growth when they were treated with anti-PD-1 or anti-CTLA-4 alone (Fig. 2a ). Interestingly, candesartan could remarkably improve the effect of these immune checkpoint blockers on 4T1 tumors (Fig. 2a ) and prolonged the survival of 4T1 tumor-bearing mice (Fig. 2b ). To further verify this effect of AngII signaling blockage on checkpoint immunotherapy, BALB/c mice bearing AGT-silenced 4T1 tumors were also treated by intraperitoneal injection of anti-PD-1 antibody (Additional file 1: Figure S3). After four administrations of anti-PD-1 antibody, all observable tumors derived from AGT-silenced 4T1 cells completely disappeared (Fig. 2c and Additional file 1: Figure S4A). More importantly, no recurrence was noted in these mice which have achieved long-term survival over one-year of observation (Fig. 2d). Similar results were obtained in the CT26 colon tumor model in syngeneic BALB/c mice (Additional file 1: Figure S4B). These results suggest that Ang II signaling blockage sensitizes tumors to checkpoint immunotherapy in mouse models.Fig2Fig. 2

AngII signaling blockage sensitizes tumors to checkpoint immunotherapy in mice tumor models. a No significant inhibition of tumor growth when treated with anti-PD-1(α-PD1) or anti-CTLA-4 (α-CTLA4) alone; however, candesartan remarkably improved the effect of immune checkpoint blockers on 4T1 tumors (n=5 ); ATR blockers and checkpoint inhibitors were administered starting from day 5 after cell injection. b Checkpoint immunotherapy combined with candesartan significantly prolonged the survival of 4T1 tumor-bearing mice ( n=5 ). c AGT gene-silencing (shRNA-AGT) in 4T1 cells rendered tumors completely responsive to PD1 checkpoint immunotherapy; negative control: shRNA-NC. d AGT gene-silencing in 4T1 tumors combined with PD1 immunotherapy made all of mice (n=5) achieve long-term survival without any recurrence during almost one-year observation. Ns, no significance; *, P < 0.05; **, P < 0.01

To explore the potential mechanism by which an AngII signaling blockage enhances tumor responses to checkpoint immunotherapy, we first performed multicolor flow cytometry analysis (FACS) to detect the proportions of effector T cells ( T_effs ) and regulatory T cells ( T_regs ) in 4T1 breast tumors from BALB/c mice treated with Ang II-receptor blockers. We found that candesartan significantly increased the frequencies of CD3⁺ T cells in 4T1 tumors (Additional file 1: Figure S5A). More importantly, candesartan remarkably increased the frequencies of CD8⁺CD44⁺ T_eff cells and decreased that of CD4⁺Foxp3⁺ T_reg cells in 4T1 tumors (Fig. 3a), leading to an increased T_effs/T_regs ratio (Fig. 3b). Similar results were also observed in tumors treated with a combination of candesartan and PD123319. However, PD123319 alone did not increase the frequencies of T_eff cells in tumors, but it non-significantly decreased the frequency of T_reg cells compared to the control tumors (Fig. 3a and b). Furthermore, AGT expression suppression in tumor cells had the same impact on tumor infiltrating T cells as the candesartan treatment (Additional file 1: Figure S6). These results indicate that AngII signaling, predominantly Ang II-AT1R-mediated signaling, may be involved in forming a tumor immunosuppressive microenvironment .Fig3Fig. 3

AngII signaling blockage reverses immunosuppressive tumor microenvironment. a Representative FACS plot of T_eff (CD8⁺CD44⁺) and T_reg (CD4⁺Foxp3⁺) in 4T1 breast tumors from BALB/c mice treated with different Ang II-receptor blockers. Bar chart (right) indicated statistic difference (**, P < 0.01; ns, no significance). Data are presented as mean±SEM, n=3. b Bar chart indicated T_eff/T_reg rate (**, P < 0.01). Data are presented as mean±SEM, n=3. c Immunofluorescence analysis showed lots of α-SMA positive fibroblasts and less CD8-positive T cells in hypoxic regions ( pimonidazole-positive ) of 4T1 tumors. AGT gene silence in 4T1 cells obviously inhibited the infiltration of fibroblasts and increased CD8-positive T cells. There were an accumulation of CD206-positive (a marker of mouse M2-phenotype TAMs) cells in pimonidazole-positive hypoxic regions of 4T1 tumors while AGT gene silence remarkably decreased the infiltration of CD206-positive cells in 4T1 tumors, especially in hypoxic regions. d Candesartan abated collagen fiber deposition in 4T1 tumors as detected by Sirius Red staining. e Representative FACS plot of TAMs (CD45⁺CD11b⁺F4/80⁺CD206⁺) in 4T1 tumors with or without candesartan treatment. f Representative FACS plot of Mo-MDSCs (CD45⁺CD11b⁺Ly6G^lowLy6C^high) and G-MDSCs (CD45⁺CD11b⁺Ly6G^highLy6C^low) in 4T1 tumors with or without candesartan treatment. g Bar chart indicated the percentages of TAMs, Mo-MDSCs and G-MDSCs in 4T1 tumors (**, P < 0.01). Data are presented as mean ± SEM, n=3

Accumulating evidence has revealed that intratumor hypoxia contributes to tumor immune escape by altering the function of innate and adaptive immune cells or by increasing the intrinsic resistance of tumor cells to cytolytic activity of immune effectors [31–33]. To demonstrate the mechanism of immune activation by AngII signaling blockage, cytokine profiles (containing 308 cytokines) of 4T1 cells in normoxia or hypoxia were analyzed using the RayBio Mouse Cytokine Antibody Array. Hypoxia induced obvious alterations ( more than 2 folds ) of 64/308 cytokines expression, in which 14 cytokines decreased and 50 increased (more than 2-fold changes) (Fig. 4a). Notably, AGT silencing remarkably altered expression of 65 cytokines in 4T1 hypoxic cells compared with control hypoxic cells. In these 65 cytokines, 45 cytokines were obviously decreased (more than 2 folds) and 20 were increased (more than 2 folds) in hypoxic AGT-silenced 4T1 cells (Fig. 4a). Interestingly, the cytokines altered by AGT-silencing were largely overlapped with those influenced by hypoxia exposure (Fig. 4b). To verify the reliability of the cytokine array data, we further used ELISA analysis to detected the levels of 6 defferential cytokines (IL-10, GM-CSF, G-CSF, Eotaxin-2, CXCL11, and TNFSF14) in the culture supernatant of 4T1 cells under different O₂ conditions, and results consistent with the cytokine array were obtained for all of 6 cytokines (Additional file 1: Figure S9A).Fig4Fig. 4

AGT expression silencing triggers an immune-activating cytokine profile in hypoxic 4T1 cells. a Cytokine profiles ( containing 308 cytokines ) of AGT-silenced 4T1 cells (AGTsi) or negative control (NC) in normoxia or hypoxia condition were analyzed using RayBio Mouse Cytokine Antibody Array. Heat-map displayed differential cytokine expression (>2 folds). b The cytokines (64) impacted by AGT-silencing were largely overlaped (40 cytokines) with those (65) influenced by hypoxia exposure. c Functional relationships of the influenced cytokines (>2folds) according to Gene Ontology (GO) analysis. d The levels of immune-activating cytokines and immunosuppressive cytokines were altered by hypoxia and AGT silencing

To investigate whether the improved immune microenvironment in a tumor would contribute to the response of other tumors in the same mouse, 4T1 cells with AGT silencing and a negative control (NC) were respectively injected into two sides (bilateral) of mammary fat pads in BALB/c mice (Fig. 5a). We found that AGT-silenced tumors did not influence the growth of contralateral NC tumors in bilaterally-injected mice (Fig. 5b). However, when anti-PD-1 antibody was administrated to bilaterally-injected mice by i.p. injection, NC-tumor growth in the same mice was remarkably inhibited, with NC tumors in three of six mice completely disappearing (Fig. 5b). These results suggests that AGT silencing in tumor combined with system PD-1 blockage results in an abscopal effect on other tumor focuses that were originally resistant to PD-1 blockage in mice.Fig5Fig. 5

An abscopal effect on resistant tumors is induced by AGT silencing in local tumor combined with PD-1 blockage. a AGT-silenced 4T1 cells (AGT) and negative control (NC) were injected into single (unilateral) or two sides (bilateral) respectively of mammary fat pads in BALB/c mice. b 4T1 tumor-bearing mice were treated as indicated. The volume of each tumor was measured from each mouse (n=6) . Ns, no significance; **, P < 0.01