Background
Ovarian cancer is the second most common and the most lethal gynecologic cancer [
1]. This lethality is due to late diagnosis of the disease, with > 70% of patients diagnosed with an advanced/metastatic stage (stage III and IV). The disease progresses aggressively throughout the peritoneal cavity in an asymptomatic manner [
2]. Unlike most solid tumors, ovarian cancer rarely metastasizes via the blood, but rather disseminates throughout the peritoneal cavity. Dissemination is highly aggressive, in a feed-forward manner, posing a unique treatment challenge for advanced/metastatic ovarian cancer. Ovarian carcinoma metastasizes to other neighboring organs by direct contact, for example to the bladder or colon [
3]. Cancer cells also detach from the primary ovarian tumor and spread throughout the peritoneum, affecting multiple organs.
Since ovarian cancer can metastasize through the peritoneal fluid, metastasis is markedly different than the hematogenous route of other cancers. Recent findings have demonstrated metastasis to be linked to epithelial-to-mesenchymal transition (EMT), the formation of multicellular spheroids (MCS), and the development of stem cell properties [
4]. In this process, detached tumor cells aggregate as MCS within the abdominal cavity to overcome anoikis. These spheroids spread throughout the peritoneal cavity, invade the peritoneum, and implant in pelvic organs and the omentum. Moreover, MCS in the ascites of ovarian cancer patients is a major impediment to effective treatment and correlate with poor clinical outcomes.
Recent evidence has demonstrated that angiotensin II type I receptor (AGTR1), is involved in tumor progression and metastasis. AGTR1 is a well-known receptor that regulates the cardiovascular system. In fact, AGTR1 is proposed to be involved in several types of gynecological malignancies including endometrial cancer [
5] and cervical carcinoma [
6]. Expression of AGTR1 has been found in several human malignant tumors, including breast [
7], skin [
8], and prostate [
9], as well as gynecologic cancers [
10,
11]. Further, angiotensin II (ANGII) and AGTR1 play essential roles in tumor survival, angiogenesis, and metastasis. For example, an antagonist of AGTR1 inhibits the migration and invasion of human lung adenocarcinoma cells through inactivation of the PI3K/AKT and MAPK signaling pathways [
12]. ANGII increases metalloproteinase-2 (MMP2) and MMP14 activities via the MAPK pathway [
13]. In vitro, ANGII stimulates cellular proliferation and vascular endothelial growth factor (VEGF) secretion by gynecologic cancer cells [
14].
In ovarian cancer, the frequency of AGTR1-positive cells is very high, with 85% of invasive adenocarcinomas expressing AGTR1 [
15]. This finding suggests that ANGII and AGTR1 may play crucial roles in the biology of ovarian cancer and could be promising therapeutic targets. A recent study has demonstrated serum angiotensin-converting enzyme (ACE) levels to be significantly higher in epithelial ovarian cancer (EOC) patients than in a control group [
16]. Further, the expression of AGTR1 can be induced by BRCA1 [
17]. In summary, current data indicate that the study of AGTR1 pathways may reveal new avenues for the investigation of ovarian cancer pathogenesis. However, the molecular basis for how the ANGII/AGTR1 axis influences ovarian cancer is unclear. Herein, a role for ANGII and AGTR1 in ovarian cancer spheroid formation is identified. A clearer understanding of the molecular mechanistic basis for ovarian cancer will provide for development of new and efficacious therapies for this deadly malignancy. Moreover, a role for ANGII has been suggested in other cancers and the data herein may provide insight into the role of ANGII in other cancers, with regard to MCS formation.
The data herein suggest ANGII treatment to significantly increase the spheroid formation, growth and the invasiveness of multiple ovarian cancer cell lines. These activities are mediated by classic direct activation of the MAPK/ERK pathway and transactivation of the epidermal growth factor receptor (EGFR). Interestingly, a dramatic increase in ANGII production and release are found in ovarian cancer cells. This suggests that ANGII forms a positive feedback loop that enhances cancer progression and metastasis. Matched with our hypothesis, the ANGII levels in ascites of ovarian cancer cell patients are significantly higher than in non-cancerous patients. In a xenograft model, intraperitoneal (i.p.) injection of ANGII significantly increases the tumorigenicity and metastasis of ovarian cancer cells, whereas an AGTR1 antagonist, losartan, suppresses this effect. Moreover, a proteomic approach was used to describe the molecular ANGII/AGTR1 axis in ovarian cancer and those data suggest that ANGII regulates lipid homeostasis. Of the highly upregulated genes, we focused on Stearoyl-CoA Desaturase (SCD1), which is involved in the actions of ANGII in terms of spheroid formation, endoplasmic reticulum (ER) stress, as well as tumorigenicity and metastasis in the xenograft model. Hence, our data suggest that the positive feedback loop of ANGII is one of the major pathways that promotes ovarian cancer development, by enhancing spheroid formation within the peritoneal cavity. The increase in spheroid formation is mediated by upregulation of SCD1 gene expression, affecting lipid homeostasis and suppressing ER stress within the spheroid.
Materials and methods
Cell lines and cell culture
The cancer cell line A2780 was obtained from the American Type Culture Collection (ATCC). The Ovca429 cell line was kindly provided by Prof. S.W. Taso of The University of Hong Kong. The isogenic highly metastatic (HM) and non-metastatic (NM) cells were generated from SKOV3.ip1 as described previously [
18]. For the transwell assay, migrated cells on the lower surface of the filter and non-migrated cells were separately harvested, and digested into single-cell suspensions. Isolated single-cell clones were selected and verified by several in vitro and in vivo assays, including the spheroid formation assay, xenograft experiments, and spectral karyotyping. The HM cells exhibited a strong metastatic signature, unlike the NM cells, which failed to form detectable metastases. Therefore, this HM/NM model offers a well-controlled experimental system for studying the metastasis of ovarian cancer. Ovca429, A2780, and HM were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 100 U/ml of penicillin and 100 μg/ml of streptomycin (Gibco) at 37 °C with 5% CO
2.
Human samples
The research protocol was approved by the Harbin Medical University Secondary Affiliated Hospital research ethics committee. Patients diagnosed with ovarian carcinoma between December 2016 and April 2018 were included in this study. The tumor samples were fixed with formalin. Informed consents were obtained prior to the following experimental and clinical data analysis and integration.
Quantitative real-time polymerase chain reaction (RT-qPCR)
Total RNA was isolated using TriPure Isolation Reagent (Roche). Total RNA (1 μg) was reverse-transcribed with an oligo-dT primer and Superscript III reverse transcriptase (Invitrogen). RT-qPCR analyses of target genes were performed by the SYBR Green PCR Master Mix (Applied Biosystems) with gene specific primers (Additional file
1: Table S1). The fluorescence signals were measured with an ABI 7300 Real-Time PCR System (Applied Biosystems). The ratio change in the target gene relative to the GAPDH house keeping gene control was determined by the 2
-ΔΔCt method [
19].
Agarose (0.5%) was pre-coated on 100 mm culture plates (Corning). To generate spheroids, ovarian cancer cells (50000) (Ovca429, A2780, or HM) were seeded on the pre-coated plates and cultured in medium supplemented with 10% FBS for 8–10 days. Spheroids were collected by centrifugation and re-seeded in 6-well plates and incubated for 12 h for cell attachment. After attachment, unattached multicellular aggregates were removed. The adhered MCS were stained with 0.5% crystal violet and the number of spheroids calculated by ImageJ software.
3D culture assay
Agarose (0.5%, 30 μL per well) was used to pre-coat 96-well plates. Ovarian cancer cells were seeded to the plates together with Matrigel (500 cells in 100 μL culture medium with 100 μL Matrigel; BD). For ANGII and losartan treatments, a 2 x drug concentration was added directly into the culture medium. After 30–60 min matrix gel polymerization, another 100 μL of medium containing a 1 x concentration of the corresponding drug was added. The plates were then incubated at 37 °C at 5% without disturbance for 10 days. The growth of MCS was observed by light microscopy and the size of the MCS (diameter) calculated. To measure the cell proliferation of the spheroids, they were reseeded to culture plates and incubated for 2 days. The cells were then stained with 0.5% crystal violet and growth area assessed.
Western blotting
Briefly, cells were collected and lysed with radioimmunoprecipitation assay buffer (RIPA) (50 mM Tris pH 7.4, 0.25% Na-deoxycholate, 1% NP-40, 150 mM NaCl and 1 mM EDTA) with proteinase and phosphatase inhibitors (Roche). Protein concentration was measured by Coomassie® Brilliant Blue (Bio-Rad). 20–30 μg total protein was separated by 8–10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinyl difluoride (PVDF) membranes (Bio-Rad) with a Trans-Blot® Turbo™ Transfer System. For membrane blotting, 5% non-fat dry milk in Tris-buffered saline with Tween 20 (TBST) was used. After incubation with primary antibodies in TBST (The details of the antibodies and the dilutions are listed in Additional file
1: Table S2), the membranes were subsequently incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (1:3000 dilution, Bio-Rad). The signals were detected by the Clarity Western ECL substrate (Bio-Rad) and captured by ChemiDoc™ XRS+ Imaging Systems (Bio-Rad). Protein relative levels were quantified by analyzing the ratio of band intensity of target protein versus loading control using ImageJ.
Immunohistochemistry
Formalin-fixed tissue sections were prepared as described previously [
20]. Deparaffinized tissue sections were stained with AGTR1 antibody after epitope retrieval by microwave. Immunohistochemistry (IHC) was performed by a standard automated IHC procedure (Dako Autostainer Universal System). The primary antibody for AGTR1 (rabbit polyclonal, MBS244122) was used at a dilution of 1:200, and negative controls were treated with PBS. The secondary antibody was anti-rabbit-HRP conjugated and used at a dilution of 1:500. The investigator conducting IHC (Yu Shan) was blinded to patient clinical data. The grading criteria were as follows: location of signal (cytoplasm negative 0, cytoplasm 1, mixed 2, and membrane 3); signaling strength (negative 0, weak 1, moderate 2, and strong 3). IHC scores were calculated as the sum of location scores and signaling strength scores.
Construction of constitutive active mutant (CAM) of AGTR1 stable expression cells
A site-directed mutagenesis kit (NEB, E0554s) was used to introduce point mutations (AGTR1-N111S and AGTR1-L305Q) into the pEGFP-N1-AGTR1 plasmid. The mutations were confirmed by Sanger sequencing. The mutated plasmids were transfected into Ovca429 cells and selected with G418 for one month to establish stable cell lines. The surviving cells after G418 selection were further selected from a single colony and validated by PCR identification. AGTR1 modified cells were used for spheroid formation and xenograft experiments.
Xenograft model of ovarian cancer
NOD-SCID 6–8 week-old females (Charles River Laboratories, Wilmington, MA) were used for tumor xenograft experiments. All animal studies were performed according to the guidelines and regulations set by the Animal Research Ethics Committee of University of Macau (UMARE-029-2017). Ovarian cancer cells (Ovca429), 4 × 106, were suspended in 0.2 mL Hanks’ balanced salt solution and were injected into the peritoneal cavity. The day after inoculation, the mice were randomly separated into groups and treated with different drugs dissolved in Hanks’ buffer. After one month, the mice were sacrificed and the number and weight of disseminated tumor nodules within the peritoneal cavity measured.
Apoptosis/necrosis assay
The cells were harvested after drug treatment and resuspended in Annexin V-binding buffer at 1 × 106 cells/mL. For each 100 μL sample, 5 μL of the Annexin V FITC conjugate was added and incubated at room temperature for 15 min. Next, 5 μL of Propidium Iodide (PI, 1.0 mg/mL) was added for 15 min for the staining of dead cells. After the incubation, 400 μL of Annexin V-binding buffer was added with samples kept on ice until analysis. The stained cells were analyzed by flow cytometry (BD Accuri™ C6 Flow Cytometer, with at least 10,000 events).
Liquid chromatography (LC)-tandem mass spectrometry (MS/MS) analysis
Cells were treated with ANGII in 6-well plates with 200 μL of lysis buffer (0 .1M Tris-HCL, pH 7.5, 4% SDS, 0 .1M DTT) used for each well. The cells were collected with a cell scraper and vortexed 3 times for 5 min each. The lysates were sonicated with pulses of 30s on and 30s off, 20 times, at 4 °C and then incubated at 56 °C for 30 min. After centrifugation (16,000 x g for 5 min) supernatants were collected for digestion. The protein amount was measured using the Bio-Rad RC-DC protein assay, as instructed by the product manual. (Bio-Rad RC DC™ Protein Assay Kit II, 5000122). Total protein (12.5 μg) was digested with trypsin (trypsin, enzyme to protein ratio 1:20) and vortexed for 1 min and incubated at 37 °C for 16 h. After drying the peptide extract completely with a Speed Vac, 15 μL of 5% acetonitrile/0.1% formic acid was used to reconstitute the samples, which were then sonicated for 5 min. All contents were transferred to a new HPLC vial-Thermo #MSCERT5000-36LVW. A total of 6 μL was injected for LC-MS analysis [
21]. Label-free quantification was obtained by MaxQuant 1.5.3.17. A database search was performed using PEAKS Studio 8.5.
RNA interference
In this study, RNA interference was used to silence target genes (AGTR1, SCD1). The silencing of AGTR1 was accomplished by transfecting siRNAs against AGTR1 into cancer cells. siRNA smart pool (L-005428-00-0005) from Dharmacon for AGTR1 interference was used for transfection. During transfection 20–40 pmol/well (6-well plates) of siRNA were transfected into the target cells with lipofectamine 3000. For the knock-down of SCD1, we used short hairpin RNA (shRNA) silencing plasmids. The SCD1-shRNA plasmids were purchased from Genepharma (Shanghai, China). SCD1-shRNA was constructed into a pGPU6/GFP/Neo vector. The sequences were as follows: shSCD1–1#: GAGATAAGTTGGAGACGATGC, shSCD1–2#: GGTACTACAAACCTGGCTTGC; shSCD1–3#: GCGATATGCTGTGGTGCTTAA; shSCD1–4#: GCACATCAACTTCACCACATT; the non-targeting shRNA (shNC): TTC TCC GAA CGT GTC ACG T. Similar to the siRNA, the shRNA plasmids were transfected into the target cells in 6-well plates (2 μg / well) with lipofectamine 3000. After 24 h, the cells were harvested for real-time PCR analysis, western blotting or functional assay.
We compared AGTR1 expression in normal ovary tissues and in ovarian cancer tissue by searching gene symbol “AGTR1” and source type of “tissues” (
http://medicalgenome.kribb.re.kr/GENT/
) [
22]. The gene expression profiles of ovarian cancer and normal ovary were extracted and presented as scatter plots. The outlier analysis tool in the Oncomine Platform was employed to identify the signature genes that were significantly upregulated in independent studies of ovarian cancer (
https://www.oncomine.org/resource/login
) [
23,
24]. The AGTR1 gene expression correlation with prognosis was analyzed by using a Kaplan Meier (KM) plotter (
http://kmplot.com/analysis/
) [
25].
Discussion
ANGII has been reported to promote cell proliferation in monolayer cell cultures of cell lines A2780 and HEYA8. However, micromolar concentrations of ANGII (10 to 100 μM) were required [
26]. Our data are consistent with that report in that ANGII at nanomolar concentrations (10 to 90 nM) only slightly increased proliferation of Ovca429, A2780, and HM cells. Therefore, we hypothesized that ANGII may have other functions that enhance ovarian tumor development other than direct stimulation of cancer cell proliferation. As ovarian cancer preferentially spreads and metastasizes through the peritoneal cavity, MCS formation is essential for ovarian cancer peritoneal metastasis [
29,
38,
39]. Hence, we hypothesized that AGTR1 activation would promote tumorigenesis and peritoneal metastasis by enhancing MCS formation and MCS growth. ANGII was found, even at nanomolar concentrations, to significantly enhance MCS growth and spheroid formation (Fig.
2a &c).
In addition, ANGII was shown to stimulate ovarian cancer cell migration, with the migrated cells expressing higher AGTR1 than parental cells (Fig.
2h). Moreover, migrated cell morphology was more mesenchymal-like compared to parental cells (Additional file
2: Figure S2d). Therefore, ANGII may induce cancer stem cell (CSC) like properties and be more metastatic. Bioinformatics analysis confirmed this hypothesis. AGTR1 expression is positively correlated with epithelial-mesenchymal transition (EMT) scores (Additional file
2: Figure S4b). GSEA analysis supported the expression of an EMT gene set including; E-cadherin, N-cadherin, and TGF-beta. Each was upregulated in patients with high AGTR1 expression (Additional file
2: Figure S4c). Epithelial ovarian cancer (EOC) cells undergo EMT transition, which liberates cancer cells from the primary tumor site and allows migration to distant organs [
40]. Furthermore, AGTR1 is highly expressed by the metastatic subtype (Additional file
2: Figure S4a) with high expression of AGTR1 associated with poor patient outcomes (Fig.
1). Taken together, these results suggest that higher expression of AGTR1 is likely to be greater in metastatic ovarian cancer cells.
Large spheroids consist of an outer proliferating region, an intermediate region of quiescent cells, and a necrotic core due to a lack of oxygen and nutrients [
41]. Herein, the growth of the MCS dramatically increased necrotic cell death, but ANGII reduced necrosis (Fig.
3j). Interestingly, co-treatment with losartan and ANGII induced greater cell death than losartan alone. This result is similar to the tumor xenograft results in which co-treatment with losartan and ANGII resulted in a further decrease in both tumor volume and the number of tumor nodes (Fig.
3g). These results suggest that not only are signaling pathways regulated by AGTR1 responsible for tumorigenesis but also that other signaling pathways activated by ANGII or related products are involved in the suppression. From the RAS pathway, possible receptors are AGTR2 and MAS1 that have opposing effects to that of AGTR1 in several cancers [
42‐
45]. Cell death analysis suggested MAS1 rather than AGTR2 to be involved (Fig.
3k). When losartan was combined with ANGII or ANG(1–7), the percentage of necrotic cells was greatly elevated compared to control. Combined treatment with losartan and CGP42112 had no effect. Activation of AGTR2 has been reported to reduce the effect of ANGII on ovarian cancer cell survival [
26]. In this study, no significant effect of the AGTR2 agonist, CGP42112A, on cell death was observed. This may be due to the extremely low expression of AGTR2 compared to AGTR1 and MAS1 (Additional file
2: Figure S3d). Since ACE2 can cleave ANGII forming ANG(1–7) and since ANG(1–7) is able to activate the MAS1 receptor [
46], co-treatment with ANGII and losartan enhanced the necrosis of ovarian cancer cells. This likely results from activation of MAS1 and the inhibition of AGTR1, reducing survival and growth of the spheroids as well as cancer metastasis.
Proteomic analysis revealed the molecular pathways by which AGTR1 enhances MCS formation and MCS growth (Fig.
5b). GO analysis suggested activation of the SREBP pathway. The SREBP pathway is well known to upregulate lipogenesis by increasing expression of various lipogenic enzymes including; FASN, ACC, and SCD1 [
47]. For example, the growth of glioblastoma in xenograft models is sharply reduced if the SREBP pathway is inhibited [
48,
49]. The hostile microenvironment of the peritoneal cavity does not favor unregulated growth of cancer cells, therefore cancer spheroids have an increased demand for fatty acids [
50]. The findings herein suggest that the effect of ANGII on MCS and metastasis is via an increase in lipid synthesis. Instead of endogenous production of fatty acids from citrate, cancer cells up-take fatty acids from exogenous sources, such as adipocytes.
The unregulated growth of cancer cells increases ER stress and induces cell death. One of the reasons is the overload of saturated fatty acid results in lipotoxicity [
39,
51,
52]. Therefore, proliferating cancer cells and/or hypoxic cells need to balance their growth rate and unsaturated lipid levels to prevent ER stress and to maintain cell survival [
47]. This can be accomplished by increasing the rate of unsaturated fatty acid synthesis and the use of both endogenous and exogenesis fatty acids. SCD1 is a desaturase localized within the ER, which introduces double bonds into fatty acids [
37,
53]. The data herein suggest that SCD1 is significantly increased after ANGII treatment (Fig.
5c). As reflected by the reductions in BiP and pPERK, ANGII may suppress the ER stress level by upregulation of SCD1. Hence, ANGII treatment would reduce cellular necrosis via an increase in unsaturated fatty acids within ovarian cancer spheroids. This may explain why ANGII only has a slight effect on cell proliferation in monolayer culture conditions. ER stress increases during tumor formation (or in cancer spheroid formation) due to the limited accessibility of nutrients and oxygen. ANGII could reduce the ER stress significantly by increases the production of unsaturated lipids. In contrast, culture conditions during anchorage-dependent growth are optimal and ER stress levels are relative low. Therefore, AGTR1 activation has only a minor effect on proliferation of cells cultured in monolayer.
Activation of PI3K/AKT and/or MAPK/ERK1/2 triggers the SREBP pathway and induces the expression of lipogenesis genes [
54‐
56]. The results herein have demonstrated ANGII to trigger PI3K/AKT and MAPK/ERK1/2 signaling pathways (Fig.
3). Interestingly, transactivation of EGFR by GPCR has been reported in various cancers [
57‐
59]. We also found that ANGII treatment increased the phosphorylation of EGFR with activation of downstream Gab1 and Shc proteins in ovarian cancer cells. Blocking of PI3K/AKT and MEK/ERK pathway reverse ANGII effect on ovarian cancer cell spheroid formation. Therefore, ANGII stimulation of the SREBP pathway is by activation of both the classical Gq signaling pathway as well as by EGFR transactivation. In addition to the SREBP pathway, the proteomic data suggest the suppression of JNK cascade transduction and inactivation by the extrinsic apoptotic cell signaling pathway. Both of the pathways are believed to mediate ER stress [
60,
61] and may influence ER stress during MCS formation.
Exogenous ANGII has been reported to induce mesenchymal stem cell production of ANGII [
62]. Herein, ANGII upregulated expression of its precursor gene, AGT, stimulating the release of ANGII into the culture medium (Fig.
4a-b). In vivo, the concentration of ANGII in the xenograft mice was notably higher than control animals. These data strongly support our hypothesis that the ANGII/AGTR1 axis forms a positive feedback loop, enhancing the effect of ANGII on cell migration and MCS formation. Interestingly, ANGII in ascites was also found in severe ovarian hyper-stimulation syndrome (OHSS) patients. This suggests that ovary dysfunction may lead to ANGII accumulation in ascites [
63]. We have confirmed that ANGII in ascites from ovarian cancer patients is significantly higher than from non-cancerous patients. The basal level of ANGII in peritoneal fluid is extremely low. Hence, there is potential to use ANGII as a biomarker for the diagnosis of ovarian cancer recurrence in the peritoneal cavity. However, more extensive studies with larger patient numbers are required to confirm this hypothesis.