C3a-C3aR signaling promotes breast cancer lung metastasis via modulating carcinoma associated fibroblasts

The mouse breast cancer cell lines 4 T1 and EO771 used in this study were obtained from the American Tissue Culture Collection (ATCC) and CH3 BioSystems respectively. Cells were routinely cultured in the recommended high-glucose DMEM medium (Hyclone) supplemented with 10% fetal bovine serum (Hyclone), L-glutamine (2 mM), penicillin (100 mg/mL), and streptomycin (100 mg/L). Primary CAFs were isolated from 4 T1 breast tumor tissues of BalB/c mice or C3a receptor-deficient (C3aR^−/−) mice by FACS. All cultures were incubated at 37 °C in a humidified atmosphere of 5% CO_2.

Six- to 8-week-old female Wild-type BalB/c mice were purchased from the Chinese Academy of Medical Sciences (Beijing, China). C3aR^−/− mice with a BalB/c background were kindly provided by Dr. Zhou Hong (Department of Immunology, Nanjing Medical University). The mice were kept under specific pathogen-free conditions at the Animal Center of Third Military Medical University. Six- to 8-week-old female nude mice were purchased from the Animal Institute of the Academy of Medical Science (Beijing, China). The MMTV-PyMT mice were kindly provided by Dr. Liu Xiaolong (Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences). For sacrificing the mice, they were kept in a chamber with isoflurane (concentration: 3–4%) for 2–3 min. After the induction of anesthesia, mice were immediately sacrificed by cervical dislocation. The animal studies have been conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of the Third Military Medical University [10].

The 4 T1 cells were routinely cultured with DMEM containing 10% fetal bovine serum until they reached 70–80% confluency. Thereafter, 6 to 8-week-old female C3aR^−/− (Balb/c background) or Balb/c mice were orthotopically injected with 1 × 10⁵ 4 T1 cells in 100 μl phosphate-buffered saline per mice in the third mammary fat pad at day 0. Tumor sizes were monitored two or three times per week from day 7,and tumor volume was calculated as follows: V = (length × width²) × 0.5 [25]. The mice were anesthetized and sacrificed 28 days after tumor inoculation. Primary tumors were dissected from the body and weighed, and tumor weight and volume were analyzed. There were six to eight mice in each group.

For the orthotopic model of spontaneous metastasis, 4 T1 cells were inoculated into the third mammary fat pad of WT or C3aR^−/− mice on day 0, and tumor sizes were monitored three times a week. The mice were anesthetized and sacrificed on day 28 post tumor injection. Indian ink was injected into the lungs of sacrificed mice with a syringe through the trachea until the lungs were completely filled with ink. The lungs were then removed immediately and immersed in Fekete’s solution for several minutes to show the pulmonary nodules; white dots on the black lungs were observed and counted for tumor metastasis. The MMTV-PyMT mice were sacrificed at 16 weeks of age, and tumor nodules in the lungs were counted after India ink injection as indicated above.

For the co-implantation assay, a total of 5 × 10⁴ 4 T1 cells alone or mixed with 2.5 × 10⁵ sorted CAFs from either WT or C3aR^−/− mice were co-injected into the third mammary fat pad of 6 week old female Balb/c nude mice. Tumor sizes were measured twice or three times a week and tumor volumes were calculated 28 days later, mice were anaesthetized and sacrificed. Filling the lungs with ink so that metastatic nodules were visible, the number of metastases was compared among three groups, each of which used 5–7 mice.

A selective antagonist of C3aR (C3aRA, SB290157, Cayman) was used for C3aR blockade. 4 T1 tumor-bearing mice were randomly assigned to control or experimental groups on day 1 post tumor inoculation and were intraperitoneally injected with C3aRA (10 mg/kg body weight) or PBS twice weekly, the mice were euthanized and sacrificed 28 days after tumor injection and the lung nodules were calculated and analyzed.

The MMTV-PyMT mice were treated with C3aRA(10 mg/Kg bodyweight) in 200 ul or PBS as early as 4-weeks of age, and they were sacrificed at 16 weeks of age. Lung metastases were identified and compared between the C3aRA treatment and control groups.

Protein was extracted from the cells or tumor tissues with RIPA buffer, dissolved with SDS–polyacrylamide gels and transferred to PVDF membranes (Millipore, 0.45 μm). The membranes were then blocked with 5% BSA at 37 °C for 1 h and incubated with primary antibodies at 4 °C overnight. Primary antibodies against a-SMA (clone: E184, Abcam,1:1000 diluted), p-AKT-Ser473 (clone: D9E, CST, 1:1000 diluted), AKT (CST, 1:1000), GAPDH (Beyotime, AF0006, 1:1000 diluted), Vimentin (Beyotime, AF0318, 1:1000 diluted), and E-Cadherin (Beyotime, AF0138,1:1000 diluted),α-Tubulin Rabbit Polyclonal Antibody (Beyotime, AF0001, 1:1000 diluted), were used. After washing with 0.1% TBST several times, the membranes were incubated with goat anti-rabbit secondary antibodies (Beyotime, 1:5000) or goat anti-mouse secondary antibodies (Beyotime, 1:5000) for 1 h at room temperature. The protein expression levels were visualized by the enhanced chemo-luminescence assay (ECL, Beyotime). Images were captured by using FluorChem HD2 system.

For in vitro stimulation of fibroblasts assay, sorted fibroblasts were seeded in 24-well plates in DMEM containing 10% FBS and 100 units/ml penicillin/streptomycin until 80% confluency. After washing with PBS, the cells were cultured in fresh serum-free media or stimulated with different doses of recombinant mouse C3a (rmC3a) (R&D). Supernatants were collected 48 h later, and filter sterilized using a 0.22 μm filter. The TGFβ1-ELISA kit (Dakewe, 1,217,102) was used for TGF-β1 detection. All experiments were performed according to the manufacturer’s instructions.

The RNA extraction of C3aR^+/+ or C3aR^−/− CAFs were performed as previously described in [25]. In brief, 500 ng of RNA was reverse-transcribed using the PrimeScript™ RT Master Mix kit (Takakra). Real-time PCR was conducted with SYBR Premix Ex Taq™ II (Takara) to quantify the relative expression of mRNA. All primer sequences for real-time qPCR are shown in (Additional file 1: Table S1). Relative changes were quantified using the 2 − ΔΔCT method [26].

For RNA sequencing, after RNA collection from four WT tumors and four C3aR^−/− tumor tissues, sequencing libraries were generated using NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, USA). The library preparations were sequenced on an Illumina Hiseq platform and was performed by Novogene (Beijing, China). The DESeq2 R package was used for differential expression analysis of two groups. The clusterProfiler R package was used to conduct Gene Ontology (GO) enrichment analysis on differentially expressed genes to correct gene length deviation. The significance of the differential gene expression was as setting the p value threshold at 0.05.

The mRNA profiles for tumor samples of 526 invasive breast cancer patients were downloaded in February 2019 from the cBioPortal for cancer genomics (www.​cbioportal.​com) [27‐29]. Spearman’s correlation analyses were conducted to determine the correlation between C3 expression and expression of CAF markers and functional cytokine transcript levels. P-values < 0.05 were considered statistically significant.

Transwell assays were performed in 24-well inserts (Falcon 8.0-μm pore size, Corning) for migration or invasion assays. 4 T1 or EO771 cells were serum starved overnight. Then, 2 × 10⁴ tumor cells were plated in transwell inserts or matrigel-coated inserts as previously described [30], followed by co-culturing with 1 × 10⁵ CAFs isolated from WT mice or C3aR^−/− mice for 24 h. Cells in the upper part of the transwells were removed with a cotton swab, and the migrated cells were fixed in 4% paraformaldehyde and stained with 0.5% crystal violet. The membrane was observed under the microscope and photographed to calculate the total number of cells. Each experiment was repeated at least three times independently.

For the wound-healing assay, 2 × 10⁵/wells of 4 T1 were seeded in six-well plates. A pipette tip was used to draw a gap on the plates. The sorted WT CAFs or C3aR^−/− CAFs were plated in the upper inserts. The migration of 4 T1 to the blank area was observed under the microscope and imaged at a specific time point.

The mammary tumors were dissected from the mice at the indicated time points. The tissues were cut into pieces and digested with collagen IV (1 mg/ml, sigma) and Dispase II (1 mg/ml, Sigma) and they were shaking for 1 h at 37 °C. The tissue/media mixture was strained using a 70 μm cell strainer for single cell suspension preparation. The cells were added anti-mouse CD16/CD32 (Clone 93, Biolegend) and incubated on ice for 10 min. Thereafter they were labeled with anti-mouse PDGFRα (Clone APA5, Biolegend) and anti-mouse F4/80 antibodies (Clone BM8, Biolegend) at a 1:100 dilution for 30 min on ice. Cells were fixed and permeabilized using Fixation/permeabilization concentrate (Ebioscience) and labeled with anti-Ki67 (Clone SolA15, Ebioscience). The FACS data were acquired using a CantoII flowcytometer (BD) and analyzed with FlowJo software. The living PDGFRα⁺F4/80⁻ CAFs were sorted using Aria II cell sorter (BD Bioscience).

Sorted CAFs were seeded in 24-well plates in DMEM containing 10% FBS and 100 units/ml penicillin/streptomycin. Subsequently, non-adherent cells were removed by extensive washing with PBS and adherent cells were treated as indicated for further analysis.

For the cell immunofluorescence assay sorted CAFs were seeded in a cell culture dish (NEST) and cultured overnight. The CAFs were fixed in 4% paraformaldehyde for 20 min and incubated with solution (including 1% BSA and 0.3% Triton X-100 in PBS) for permeabilization and blockade of unspecific binding at room temperature for 45 min. The primary antibody against mouse C3aR (14D4, Hycult biotech, 1:50 diluted) and rabbit anti-mouse α smooth muscle actin (E184, Abcam, 1:500 diluted) were incubated at 4 °C overnight. For tumor tissues, frozen sections of mammary tissue from WT/C3aR^−/− mice were fixed with ice-cold 4% paraformaldehyde for 15 min at room temperature. After washing with PBS, the sections were blocked with 5% bovine serum albumin-containing PBS for 1 h at room temperature, followed by primary antibody incubation overnight at 4 °C. The following antibodies were used for tissue immunofluorescence staining: E-Cadherin mouse monoclonal antibodies (AF0138, Beyotime, 1:50 diluted) and Vimentin Mouse Monoclonal Antibodies (AF0318, Beyotime, 1:100 diluted). Alexa Fluor 488-labeled goat anti-rabbit IgG (1:200 dilution, Abcam), Alexa Fluor 488-labeled goat anti-Armenian hamster antibody (1:200 dilution, Abcam),or Alexa Fluor647-labeled goat anti-rabbit IgG (1:200 dilution, Abcam) were used as secondary antibodies. After counterstaining with DAPI (Beyotime), sections were imaged under an Olympus Fluorescence Microscope.

RNA-sequencing data has been deposited to NCBI (PRJNA600392).

Previously, we and others demonstrate that C3aR signaling promote tumor growth by promoting immune inhibition [6, 10]. However, its role in metastatic spread of breast cancer has not been explored. To investigated whether C3aR signaling contributed to metastasis, we orthotopicaly injected 4 T1 cells (a mouse breast cancer cell line), which closely mimics stage IV of human breast cancer, into mammary fat pad of Balb/c C3aR^+/+ mice and C3aR^−/− mice, respectively. Our result suggest that C3aR deficiency resulted in a decreased lung metastatic burden (Fig. 1a and b), while it did not significantly affect the growth of primary breast tumors (Additional file 1: Figure S1a-d). Epithelial mesenchymal transition(EMT)induction is one of the most important mechanisms for cancer metastasis [22], the downregulation of epithelial marker E-cadherin and the upregulation of mesenchymal marker vimentin are typical characteristics of EMT. To this end, we detected the EMT markers in tumor tissues and found that the mesenchymal marker vimentin was down-regulated, and the epithelial marker E-cadherin was up-regulated when the 4 T1 cell was inoculated in C3aR^−/− mice (Fig. 1c-f). The results suggest that C3aR signaling promote metastasis of breast cancer via inducing EMT of tumor cells.

To further illustrate the underling mechanism of C3a-C3aR mediated breast cancer metastasis, we applied RNA sequencing to profile the gene expression of 4 T1 tumors from WT mice or C3aR^−/− mice. RNA sequencing data revealed that tumor tissue from C3aR^−/− mice displayed significant differences in gene expression profiles compared with that of WT tumors. Among them, we found that genes associated with extracellular matrix components were significantly downregulated in tumors isolated from C3aR^−/− mice than those from WT mice (Fig. 2a and b). To our knowledge, CAFs is the main source of extracellular matrix components in TME [31]. Hence, we clarified whether C3aR signaling altered the quantity and quality of CAFs within the TME. We firstly compared the percentage of CAFs (PDGFRα⁺F4/80⁻ cells) in tumors isolated from WT and C3aR^−/− mice. We found comparable CAFs infiltration between the two groups (Additional file 1: Figure S2a&b). We next determined the mRNA expression of genes, associated with activation and functionality of CAFs, isolated from the 4 T1 tumors from C3aR^−/− mice or WT mice (Additional file 1: Figure S2c). Our data indicates that C3aR^−/− CAFs expressed a diminished level of conventional fibroblast markers, including a-SMA, PDGFRα, FAP when compared with WT CAFs. Accordingly, C3aR^−/− CAFs also expressed reduced functional cytokines including TGF-β, CXCL12, HGF (Fig. 2c). These results suggest that C3aR signaling promotes lung metastasis of breast cancer possibly by altering the function of CAF, rather than changing its numbers.

Upon analysis of the clinical data from human invasive breast cancer mRNA profiles for tumor samples of 526 invasive breast cancer patients, we found that C3 expression was positively correlated with CAFs markers (Fig. 3.)a-c and its effector cytokines(Fig. 3.d-f) in human breast cancer tissues [32]. Additionally, the ECM components (Fig. 3. g-l) mostly synthesized by CAF were also associated with local C3 expression. To sum up, we concluded that production of C3 complement may contribute to enhancing the function of CAF and promoting the formation of invasive breast cancer.

To illustrate the role of C3aR signaling in modulating CAFs function, we firstly asked whether CAFs express C3aR. To this end, we stained C3aR in sorted PDGFRα⁺ F4/80⁻ cells by immunofluorescence. We found that CAFs cells expressed C3aR, a G-protein coupled receptor, both on the membrane and intracellularly (Fig. 4a). To our knowledge, internalization of C3aR usually suggests that the C3aR receptor is functional as it was reported before [33].

The elevated expression of soluble factors in CAFs has been suggested to act as a possible regulator of adjacent cancer cell migration. To test this, we co-cultured 4 T1 cells or EO771 cells with CAFs isolated from C3aR^+/+ or C3aR^−/− tumor-bearing mice, respectively. Scratch assay and transwell migration/invasion assays confirmed that the co-culturing of C3aR^−/− CAFs with 4 T1 cells rarely stimulated breast cancer epithelial cell migration and invasion, while co-culturing of C3aR^+/+ CAFs significantly promoted EO771 cells migration and invasion (Fig. 4b-d). Moreover, recombinant mouse C3a did not only induce CAFs phenotype markers characterized by alpha smooth muscle actin (a-SMA) expression (Fig. 4e), but it also stimulated the expression and secretion of TGFβ, a crucial pro-metastasis cytokine effector of CAFs. Furthermore, the expression of genes associated with phenotype markers and functional cytokines of CAFs were upregulated under C3a stimulation (Fig. 4g). For the analysis of whether C3aR signaling on CAFs altered the ability of promoting metastasis in-vivo, we injected 4 T1 cells alone, with WT CAFs or with C3aR^−/− CAFs orthotopically into the mammary fat pad of immunodeficient nude mice. Co-inoculation of 4 T1 cells with WT CAFs accelerated lung metastasis, compared with 4 T1 alone, while C3aR signaling deficiency in CAFs abrogated its metastasis-promoting effect (Fig. 4h and i). These results indicated that C3a acts directly on CAFs in a C3aR dependent manner.

We found that CAFs were stimulated with rmC3a protein and showed an increased phosphorylation of AKT at Ser473 as early as 5 min, reaching a peak at 15 min (Fig. 5a). However, rmC3a stimulation could not activate PI3K signaling in C3aR^−/− CAFs (Fig. 5b). Pretreatment with PI3K inhibitor (LY294002) inhibited increased expression of α-SMA and TGFβ secretion induced by C3a. To further determine whether C3a-mediated AKT phosphorylation is C3aR-dependent, we used C3aR antibody (14D4) or a C3aR antagonist (SB290157) to block C3aR signaling. Our data suggest that both C3aR antagonist and C3aR antibody pretreatment could inhibit C3a-mediated AKT phosphorylation (Fig. 5c). Similarly, C3a-induced α-SMA and TGFβ1 were greatly suppressed after C3aR blockade or PI3K inhibition (Fig. 5d and f). C3a stimulated CAF facilitated migration capacity of 4 T1 cells, and can be inhibited by C3aR antagonist and PI3K inhibitor. These data suggest C3a plays a crucial role in regulating CAF activation and effector cytokine production via activating PI3K/AKT signaling pathway.

To address the possible utility of targeting C3aR as a translational anti-metastasis strategy, we employed a C3aR antagonist (SB290157), used in studies of reactive airway [34] and leptomeningeal metastasis [12], to treat a 4 T1 orthotopic mouse model (Fig. 6a). Our data suggest that lung metastasis was markedly reduced by treatment with C3aR antagonist (Fig. 6b-d), whereas the difference in the tumor weight was not significantly changed (Additional file 1: Figure S3a). In addition, C3aR inhibition also protected against the development of lung metastasis in the PyMT-MMTV mouse spontaneous breast cancer model (Fig. 6e and f). Additionally, we explored TCGA data and analyzed the correlation between C3aR1⁺PDGFA⁺ expression and the survival of triple negative breast cancer patients, we found that patients with a high level of C3aR1⁺PDGFA⁺ expression had a poorer survival rate (Fig. 6g). These results potentiate C3aR signaling blockade as an effective anti-metastasis strategy in breast cancer management.

In summary, our study showed that C3aR signaling plays a unique role in promoting lung metastasis of breast cancer by modulating CAFs. C3aR deficiency inhibits pro-metastatic cytokines production by CAFs in a 4 T1 tumor model. Moreover, increased expression of TGF-β and CAF markers were found in C3a-treated CAFs in vitro. Mechanically, we identified that PI3K/AKT signaling plays an essential role in C3a-C3aR signaling mediated by CAFs activation. Our data demonstrated that targeting C3aR might be an effective strategy in tumor metastasis control in breast cancer.