The IL1β-IL1R signaling is involved in the stimulatory effects triggered by hypoxia in breast cancer cells and cancer-associated fibroblasts (CAFs)

Gene expression analysis was performed using the publically available The Cancer Genome Atlas (TCGA) and Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) datasets [28, 29]. Data was downloaded on the 10 May 2020. The patients clinical information along with the mRNA expression data (RNA Seq V2 RSEM) reported in the Invasive Breast Cancer Cohort of the TCGA project were downloaded from UCSC Xena (https://​xenabrowser.​net/​). The clinical information and the microarray gene expression data (Log2 transformed intensity values) of the METABRIC cohort (n. 2509) were retrieved from cBioPortal for Cancer Genomics (http://​www.​cbioportal.​org/​). Samples of the TCGA cohort (n. 1247) were filtered by the “sample type” in order to obtain exclusively the tumor tissues (n. 1101). Thereafter, patients of both TCGA and METABRIC were classified on the basis of the presence or the absence of the estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2), detected by immunohistochemistry. Gene expression and clinical information were also filtered for missing values. The final filtering resulted in 771 patients of TCGA and 1904 patients of METABRIC.

The Pearson correlation coefficients (r-values) between the mRNA levels of HIF-1α and the other genes of the TCGA (n. 20,530) dataset in TNBC cohort of patients were assessed in R Studio (version 3.6.1) using the cor.test() function and setting the method as “Pearson”. The statistical analysis was performed by using the t-tests considering significant the coefficients obtained with p < 0.001. The first 250 most correlated genes were selected for the next evaluations. In particular, aiming to cluster these genes in pathways, we uploaded our list on the Database for Annotation, Visualization and Integrated Discovery (DAVID) functional annotation analysis website (https://​david.​ncifcrf.​gov/​). We analyzed the genes selecting the functional annotation tool and the option Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, choosing the official gene symbol as “select identifier” and gene list as “list type” in the options of the upload and selecting a limit species of “Homo sapiens” in the background.

Gene Set Enrichment Analysis (GSEA) was performed using the gsea() function of the phenoTest package (https://​bioconductor.​org/​packages/​release/​bioc/​html/​phenoTest.​html) in R Studio. The KEGG “Cytokine-cytokine receptor interaction” pathway (KEGG entry = hsa04060) and “HIF-1 signaling” pathway (KEGG entry = hsa04066) were selected as the reference gene sets. We ranked the genes of the “Cytokine-cytokine receptor interaction” pathway in accordance with the differential expression within HIF-1α high and low (median expression value as threshold assessment) samples, and the “HIF-1 signaling” pathway genes in accordance with the differential expression within IL-1β high and low (median expression value as threshold assessment) samples. Both analysis were performed only in the TNBC subgroup of patients, verifying if the selected sets of genes were enriched at the bottom or the top of the ranked lists. We calculated the enrichment score (ES) that reflects the degree to which a set of genes is overrepresented at the extremes (top or bottom) of the entire ranked list. The score was calculated by walking down a list of genes ranked by their correlation with the selected phenotype (high or low HIF-1α/ IL-1β levels), increasing a running-sum statistic when a gene in that gene set is encountered (each vertical line underneath the enrichment plot) and decreasing it when a gene that isn’t in the gene set is encountered. The magnitude of the increment depends on the correlation of one gene with the phenotype. In this analysis, 20,000 simulations were used (B = 20,000). p < 0.05 was considered significant.

The ROS scavenger N-acetyl-L-cysteine (NAC) (used at a 300 μM concentration) and the proteasome inhibitor MG132 (used at a 10 μM concentration) were purchased from Merck Life Science (Milan, Italy). PD98059 (PD) and LY294,002 (LY) (both used at a 1 μM concentration) were obtained from Calbiochem (Milan, Italy). All compounds were dissolved in DMSO, except NAC that was solubilized in water. Recombinant human IL-1β (used at a 10 ng/mL concentration) was purchased from Thermo Fisher Scientific (Life Technologies Italia, Monza, Italy) and solubilized in PBS with 1% BSA. The IL1R1 antagonist (IL1R1a) human recombinant protein (used at a 50 ng/mL concentration) was purchased from Thermo Fisher Scientific and solubilized in 20 mM TBS, pH 8, with 50% glycerol. Anti-IL-1β neutralizing antibody (MAB601) was purchased from R&D Systems (Bio-Techne, Milano, Italy).

The TNBC MDA-MB 231 breast cancer cells were provided by ATCC (Manassas, VA, USA), used less than 6 months after resuscitation, routinely tested and authenticated according to the ATCC suggestions. MDA-MB 231 cells were maintained in DMEM/F12 (Dulbecco’s modified Eagle’s medium) with phenol red, supplemented with 5% fetal bovine serum (FBS) and 1% penicillin/streptomycin (Thermo Fisher Scientific). CAFs were isolated, cultured and characterized as previously described [30] from 10 invasive mammary ductal carcinomas and pooled for the subsequent studies. Briefly, specimens were cut into 1–2 mm diameter pieces, placed in a digestion solution (400 IU collagenase, 100 IU hyaluronidase, 10% FBS, antibiotics and antimycotics) (Thermo Fisher Scientific) and incubated overnight at 37 °C. Cells were then separated by differential centrifugation at 90×g for 2 min. The supernatant containing fibroblasts were centrifuged at 485×g for 8 min, the pellet obtained was suspended in fibroblasts growth medium (Medium 199 and Ham’s F12 mixed 1:1 and supplemented with 10% FBS and 1% penicillin) (Thermo Fisher Scientific) and cultured at 37 °C, 5% CO2. CAFs were then expanded into 10-cm Petri dishes and stored as cells passaged for three population doublings within total 7 to 10 days after tis-sue dissociation. Primary cell cultures of fibroblasts were characterized by immunofluorescence with human anti-vimentin (V9; 1:500) and human anti-cytokeratin 14 (LL001) (Santa Cruz Biotechnology, DBA, Milan, Italy; 1:250). FAPα antibody (H-56; Santa Cruz Biotechnology, DBA, Milan, Italy; 1:500) was used to assess fibroblast activation (data not shown). We used CAFs passaged for up to 10 population doublings for the experiments, to minimize clonal selection and culture stress, which could occur during extended tissue culture. All cell lines were grown in a 37 °C incubator with 5% CO2 and switched to medium without serum and phenol red the day before treatments to be processed for immunoblot and RT-PCR assays.

Total RNA was extracted, and cDNA was synthesized by reverse transcription as previously described [31]. The expression of selected genes was quantified by real-time PCR using platform Quant Studio7 Flex Real-Time PCR System (Thermo Fisher Scientific). Gene-specific primers were designed using Primer Express version 2.0 software (Applied Biosystems) and are as follows: 5′-ACCTATGACCTGCTTGGTGC-3′ (HIF-1α forward) and 5′-GGCTGTGTCGACTGAGGAAA-3′ (HIF-1α reverse); 5′-TTAAAGCCCGCCTGACAGA-3′ (IL-1β forward) and 5′-GCGAATGACAGAGGGTTTCTTAG-3′ (IL-1β reverse); 5′-TTACAGAGGGAAAACGACACCT-3′ (GPER forward) and 5′-GTGGGTCTTCCTCAGAAGGG-3′ (GPER reverse); 5′-AGTCCCTGAGCATCTACGGT-3′ (COX2 forward) and 5′-CATCATCAGACCAGGCACCA-3′ (COX2 reverse); 5′-AAGCCACCCCACTTCTCTCTAA-3′ (ACTB forward) and 5′-CACCTCCCCTGTGTGGACTT-3′ (ACTB reverse). Assays were performed in triplicate and the results were normalized for actin beta (ACTB) expression and then calculated as fold induction of RNA expression.

PCR arrays were performed using a TaqMan™ Human Tumor Metastasis Array (Thermo Fisher Scientific) according to the manufacturer’s instructions. The amplification reaction and the results analysis were carried out using platform Quant Studio7 Flex Real-Time PCR System (Thermo Fisher Scientific).

Cells were transfected using X-treme GENE 9 DNA Transfection Reagent (Roche Diagnostics, Merck Life Science) for 24 h before treatments with a control vector and a specific shRNA sequence or antisense vector for each target gene. The short hairpin (sh) RNA constructs to knock down the expression of HIF-1α and the control shRNA construct were purchased form SABioscience Corporation. Antisense vector for GPER (AsGPER), which was generated by cloning of the entire open reading frame of the receptor cDNA in the reverse orientation in pcDNA3.1 Hygro (−), was a kind gift from Eric R Prossnitz (University of New Mexico Health Science Center, Albuquerque, USA). The plasmid DN/c-fos, which encodes for c-fos mutant that heterodimerizes with c-fos dimerization partners but does not allow DNA binding, was a kind gift from Dr. C. Vinson (NIH, Bethesda, MD, USA).

Cells were grown in 10-cm dishes, exposed to treatments for 16 h, and then cross-linked with 1% formaldehyde and sonicated. Supernatants were immuno-cleared with salmon DNA/protein A-agarose (Merck Life Science) and immunoprecipitated with anti-HIF-1α or anti-GPER antibody or nonspecific IgG. Pellets were washed, eluted with a buffer consisting of 1%SDS and 0.1 mol/L NaHCO3, and digested with proteinase K. DNA was obtained by phenol/chloroform extractions and precipitated with ethanol. The yield of target region DNA in each sample after ChIP was analyzed by real-time PCR. The primers used to amplify a region containing a HRE site located into the GPER promoter sequence were: 5′-TGCAGCACTTCAAAACAATAACC − 3′ (Fw) and 5′-GGGTTTGAGTTGTTTTTCCTTTGG-3′ (Rv); the primers used to amplify a region containing a HRE site located into the IL-1β promoter sequence were: 5′-ACAGACAGGGAGGGCTATTG-3′ (Fw) and 5′-GGGCAAGGAGTAGCAAACTA-3′ (Rv). Data were normalized to the input for the immunoprecipitation and the results were reported as fold changes respect to nonspecific IgG.

Cells were grown in 10-cm dishes, exposed to treatments, and then lysed as previously described [32]. Equal amounts of whole-protein extract were resolved on a 10% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane (Amersham Biosciences, Sigma-Adrich, Milan, Italy), which were probed with primary antibodies (1:1000) against HIF-1α and IL-1β (R&D Systems, Bio-Techne, Milano, Italy), GPER (AB137479) (Abcam, DBA, Milan, Italy), phosphorylated ERK1/2 (E-4), ERK2 (C-14), p-AKT1/2/3 (Ser 473)-R, AKT/1/2/3 (H-136), c-fos (E-8) and β-actin (AC-15; 1:4000) (Santa Cruz Biotechnology, DBA, Milan, Italy) and then revealed using the chemiluminescent substrate for western blotting Westar Nova 2.0 (Cyanagen, Biogenerica, Catania, Italy). For nuclear extracts, cells were lysed using 300 μl of cytosolic buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA, pH 7.5, 10% glycerol) with protease inhibitors (1.7 mg/ml aprotinin, 1 mg/ml leupeptin, 200 mmol/liter phenylmethylsulfonyl fluoride, 200 mmol/liter sodium orthovanadate and 100 mmol/liter sodium fluoride). Following centrifugation (14,000 g, 4 °C, 10 min), the supernatant was referred to as cytoplasmic fraction and the pellet containing nuclei was resuspended in high salt buffer (20 mM HEPES pH 7.9, 25% [v:v] glycerol, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA and protease inhibitors). For the extraction of nuclear proteins, the obtained solution was vortexed thoroughly, incubated overnight with agitation and centrifugated at 14000 g, 4 °C for 10 min. Equal amounts of the collected supernatant, which represent the nuclear fraction, were then run on 10% SDS-PAGE and western blot analysis was performed as described above. The purity of the nuclear fraction was confirmed by immunoblotting with primary antibodies against β-actin (AC-15; 1:4000) and anti-LMNB/Lamin (M-20; 1:2000) (Santa Cruz Biotechnology, DBA, Milan, Italy).

After exposure to treatments, cells were washed and lysed using 500 μl RIPA buffer with protease inhibitors (1.7 mg/ml aprotinin, 1 mg/ml leupeptin, 200 mmol/liter phenylmethylsulfonyl fluoride, 200 mmol/liter sodium orthovanadate and 100 mmol/liter sodium fluoride). Samples were then centrifuged at 13,000 rpm for 10 min and protein concentrations were determined using Coomassie (Bradford) protein assay. Proteins (200 μg) were then incubated for 2 h with 900 μl of immunoprecipitation buffer with inhibitors, 2 μg of anti-c-fos or anti-GPER antibodies and 20 μl of Protein A/G agarose immunoprecipitation reagent (Santa Cruz Biotechnology, DBA, Milan, Italy). Samples were then centrifuged at 13,000 rpm for 5 min at 4 °C to pellet beads. Pellets were washed four times with 500 μl of PBS and centrifuged at 13,000 rpm for 5 min at 4 °C. Supernatants were collected, resuspended in 20 μl RIPA buffer with protease inhibitors, 2X SDS sample buffer and heated to 95 °C for 5 min. Samples were then run on 10% SDS-PAGE, transferred to nitrocellulose and probed with primary antibodies. Western blot analysis and ECL detection were performed as described above.

Cells were grown on a cover slip, serum deprived for 18 h and then exposed to treatments for 16 h. Next, cells were fixed in 4% paraformaldehyde in PBS, permeabilized with 0.2% Triton X-100, washed 3 times with PBS and incubated at 4 °C overnight with a primary antibody (1:250) against GPER (AB137479) (Abcam, DBA, Milan, Italy) or Phospho-Myosin Light Chain 2 (Ser19) (p-MLC) (Cell Signaling, Euroclone, Milan, Italy). After incubation, the slides were extensively washed with PBS, probed with alexa Fluor 594 goat anti-rabbit IgG (1:300, Thermo Fisher Scientific) and 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) (1:1000; Sigma-Aldrich). Then, the images were obtained using the Cytation 3 Cell Imaging Multimode reader (BioTek, AHSI, Milan Italy) and analyzed by the Gen5 software (BioTek, AHSI, Milan Italy).

Cells were exposed to treatments for 16 h, washed twice with PBS, fixed in 4% paraformaldehyde in PBS for 10 min, washed briefly with PBS, then incubated with Phalloidin-Fluorescent Conjugate (Santa Cruz Biotechnology). The images were obtained using the Cytation 3 Cell Imaging Multimode reader (BioTek, AHSI, Milan Italy) and analyzed by the Gen5 software (BioTek, AHSI, Milan Italy).

The concentrations of IL-1β in supernatants from hypoxia-treated MDA-MB-231 cells were measured using human IL-1β ELISA Kit (Thermo Scientific, Monza Italy), according to the manufacturer’s instructions. The plates were read at 450 nm on a Microplate Spectrophotometer Epoch™ (BioTek, AHSI, Milan Italy).

The non-fluorescent 2′,7′-dichlorofluorescin diacetate (DCFDA) probe was used to evaluate intracellular ROS production. Briefly, cells were incubated with 10 μM DCFDA (Sigma-Aldrich) at 37 °C for 30 min, washed with PBS, and then exposed to treatments for 15 min, as indicated. Cells were then washed with PBS, and the images were obtained using the Cytation 3 Cell Imaging Multimode reader (BioTek, AHSI, Milan Italy) and analyzed by the Gen5 software (BioTek, AHSI, Milan Italy).

MDA-MB-231 cells were cultured under normal or low oxygen tension (2%) for 16 h. Thereafter, the supernatants were collected, centrifuged at 3500 rpm for 5 min to remove cell debris and used as conditioned medium in the appropriate experiments.

Cells were treated for 16 h, trypsinized and seeded onto fibronectin (5 μg/ml) coated 96-well plate at a density of 2 × 10⁵ cells/ml and incubated at 37 °C in a humidified incubator containing 5% CO2. Phase-contrast images were captured after 15 min and 60 min. As previously reported [33], round bright cells were considered unspread, whereas spread cells were defined as those cells that had lost their phase-bright appearance and had readily distinguishable nucleus and cytoplasm and quantified by counting six randomly selected fields in each well under a phase-contrast microscope.

Transwell 8 μm polycarbonate membrane (Costar, Sigma-Aldrich, Milan, Italy) was used to evaluate in vitro cell invasion. 5 × 10⁴ cells in 300 μL serum-free medium were seeded in the upper chamber, coated with Corning® Matrigel® Growth Factor Reduced (GFR) Basement Membrane Matrix (Biogenerica, Catania, Italy) (diluted with serum-free medium at a ratio of 1:3). For invasion assays in MDA-MB-231 cells, complete medium was added to the bottom chambers in the presence of treatments where required, then cells were cultured under normoxia or hypoxia for 16 h. For invasion assays in CAFs, conditioned medium collected from MDA-MB-231 cells previously exposed to hypoxia for 16 h was added to the bottom chambers in the presence of treatments, where required. Cells on the upper surface of the membrane were then removed by wiping with Q-tip, and invaded cells were fixed with 100% methanol, stained with Giemsa (Sigma-Aldrich, Milan, Italy), photographed using Cytation 3 Cell Imaging Multimode Reader (BioTek, AHSI, Milan Italy) and counted using the WCIF ImageJ software.

For spheroid generation, 100 μL/well of MDA-MB 231 cell suspensions (1 × 10⁴) were dispensed into 2% agar-coated 24-well plates. Three days after seeding, tumor spheroids (a single spheroid per well) were exposed to treatments and a 50% medium and treatment replenishment was performed every 2 days. Images were obtained on day 20 using a conventional inverted microscope, thereafter cell number per spheroid was determined by trypsinizing three different spheroids, mixing the cell suspension with trypan blue and counting the number of viable cells. The total number of cells obtained was divided by the number of trypsinized spheroids.

The statistical analysis was performed using ANOVA followed by Newman-Keuls’ test to determine differences in means. The bioinformatics analyses, including t-tests, box plots and scatter plots, were performed using the R tidyverse package (https://​www.​tidyverse.​org/​packages/​). p < 0.05 and p < 0.01 were considered statistically significant.

It is well acknowledged that the transcription factor HIF-1α acts as a molecular sensor within the hypoxic breast tumor microenvironment toward aggressive cancer features [7]. Hence, we began our study evaluating the gene signature associated with HIF-1α in TNBC patients that are characterized by a poor prognosis [34]. Exploring the TCGA cohort of TNBC patients, the 250 genes mostly correlated with HIF-1α were first ranked by Pearson correlation coefficient. In order to investigate the biological significance of these genes, KEGG pathway analysis was then performed using DAVID. The top 250 HIF-1α correlated genes were enriched in a set of pathways, as schematically reported in Fig. 1a. The “Cytokine-cytokine receptor interaction” transduction pathway, including IL1B, IL1A, IL1R, IL6, IL6ST, IL7R, IL1RAP, INHBA, OSMR and TGFBR2 that are strongly associated with breast cancer progression [35‐41], was found significantly correlated with HIF-1α. Next, we performed gene set enrichment analysis (GSEA) in order to explore the “Cytokine-cytokine receptor interaction” expression profile according to the high and low HIF-1α phenotypes of both TCGA and METABRIC cohorts of TNBC patients. It is worth noting that the genes included in the “Cytokine-cytokine receptor interaction” pathway were found enriched in the group of patients showing high HIF-1α levels (Fig. 1b). As IL-1β may contribute to breast cancer development and metastasis [37] and may regulate the migratory phenotype of TNBC cells through HIF-1α [42, 43], we sought to evaluate whether IL-1β could be involved in the activation of a hypoxia-related gene signature in TNBC patients. In this regard, we ascertained that the genes included in the “HIF-1 signaling pathway” of the KEGG database are significantly enriched in the TNBC cohorts displaying high IL-1β expression of both TCGA and METABRIC datasets (Fig. 1c). Further corroborating these observations, a positive correlation between IL-1β and HIF-1α levels was found in the TNBC cohorts of both databases (Fig. 2a). Exploring the clinical significance of IL-1β in TNBC patients, we assessed that the IL-1β expression levels are significantly higher in TNBC compared to non-TNBC in both TCGA and METABRIC datasets (Fig. 2b). Overall, the aforementioned data suggests that IL-1β is engaged by HIF-1α toward the development of the aggressive TNBC.

On the basis of the aforementioned findings and in order to provide novel insights on the regulation of IL-1β by hypoxia in TNBC, we ascertained that a low oxygen tension (2% O2, thereafter mentioned as hypoxia) induces the mRNA (Fig. 3a) and protein (Fig. 3b) expression of IL-1β in MDA-MB-231 cells, which are a well-recognized model system of TNBC. Upon hypoxic conditions, the mRNA levels of HIF-1α did not display any change (Fig. 3a) whereas an increased protein expression was observed (Fig. 3b), in accordance with the known ability of hypoxia to decrease the HIF-1α ubiquitination and degradation [9]. Next, performing ELISA assays we determined that the secretion of IL-1β increases in the medium of MDA-MB-231 cells cultured under hypoxia (Fig. 3c). Remarkably, the silencing of HIF-1α abolished the expression of IL-1β prompted by hypoxia, suggesting that HIF-1α is involved in the up-regulation of IL-1β in MDA-MB-231 cells exposed to hypoxic conditions. In accordance with previous findings showing that HIF-1α can be induced by proinflammatory stimuli including IL-1β [43], we ascertained that IL-1β boosts the accumulation of HIF-1α in MDA-MB-231 cells (Additional file 2). Worthy, these results suggest that the HIF-1α-dependent induction of IL-1β upon hypoxic conditions may generate a feed-forward loop that further contributes to the stimulatory interaction between HIF-1α and IL-1β signaling. Reminiscing our previous data on the action of GPER in hypoxic conditions [13, 14], we assessed that hypoxia triggers the up-regulation of GPER mRNA (Fig. 3e) and protein (Fig. 3f) levels in a time-dependent manner in MDA-MB-231 cells. By chromatin immunoprecipitation (ChIP) assay, we then found that HIF-1α is recruited to the GPER promoter region in MDA-MB-231 cells exposed to hypoxia (Fig. 3g). Moreover, the silencing of HIF-1α prevented the protein induction of GPER upon a 16 h exposure to hypoxic conditions (Fig. 3h), indicating that HIF-1α is required for the transcription of GPER in MDA-MB-231 cells exposed to hypoxia. Next, the hypoxia increased HIF-1α levels were not altered by the silencing of GPER that instead abolished the induction of IL-1β (Fig. 3i). Together, these findings suggest that upon hypoxia HIF-1α is involved in the up-regulation of GPER, which then cooperates with HIF-1α toward the regulation of IL-1β. In order to evaluate the transduction mechanism mediating the aforementioned responses, we first assessed that a short exposure of MDA-MB-231 cells to hypoxia does increase the levels of reactive oxygen species (ROS) (Fig. 4a), which are required for the stabilization of HIF-1α and the subsequent transduction of the hypoxia-mediated signaling [44]. In the aforementioned experimental condition, we next ascertained that the activation of two main transduction regulators of HIF-1α, namely ERK1/2 and Akt [45, 46], is prevented in the presence of the ROS scavenger NAC (Fig. 4b). As c-fos expression is a suitable molecular sensor of both GPER and HIF-1α transduction signaling [14, 47, 48], we also determined that the induction of c-fos protein levels observed upon a 4 h exposure to hypoxia is no longer evident in the presence of the MEK inhibitor PD, the PI3K inhibitor LY (Fig. 4c) and NAC (Fig. 4d). Likewise, the increase of HIF-1α, GPER and IL-1β protein levels upon a 16 h exposure to hypoxia using either NAC (Fig. 4e) or PD and LY (Fig. 4f), suggesting that the ROS activity triggers the ERK1/2/Akt/c-fos transduction signaling toward the hypoxic regulation of HIF-1α, GPER and IL-1β expression. In accordance with these findings, the transfection in MDA-MB-231 cells of a plasmid encoding a mutant of c-fos, namely dominant negative c-fos (DN/c-fos), abrogated the up-regulation of HIF-1α, GPER and IL-1β triggered by a 16 h exposure to hypoxia (Fig. 4g). In order to evaluate whether the reduction of HIF-1α protein levels triggered in the aforementioned experimental condition by the DN/c-fos construct is associated with the HIF-1α proteasome-dependent degradation, the MDA-MB-231 cells were treated for 16 h with the proteasome inhibitor MG132. Notably, the MG132 rescued the effects of the DN/c-fos construct on the HIF-1α protein levels (Fig. 4h), suggesting that c-fos is involved in the complex HIF-1α biological regulation at the proteasomal level. By co-immunoprecipitation studies, we also found that hypoxia stimulates a direct interaction between c-fos and HIF-1α in MDA-MB-231 cells (Fig. 4i), indicating that the action of c-fos on HIF-1α stabilization at least in part occurs through a physical interaction between these two factors. Considering that in our previous study the functional cooperation between HIF-1α and GPER triggered the expression of certain genes in hypoxic conditions [13], we first ascertained that HIF-1α co-immunoprecipitates with GPER in MDA-MB-231 cells exposed to hypoxia (Fig. 4j). Thereafter, we evidenced the accumulation of GPER in the nuclear compartment of MDA-MB-231 cells exposed to hypoxia, as evidenced by both immunofluorescence (Fig. 4k) and subcellular fractionation studies (Fig. 4l). Intriguingly, by ChIP assays we then demonstrated that the exposure to hypoxia induces the recruitment of both HIF-1α and GPER to a HRE site located within the IL-1β promoter sequence (Fig. 4m). Overall, these data suggest that a functional interplay between HIF-1α and GPER may be stimulated by hypoxia toward the transcription of IL-1β in MDA-MB-231 cells.

In order to provide further insights into the metastatic gene expression profile triggered by hypoxia through the IL-1β/IL1R1 axis, we performed a TaqMan Gene Expression Assay, which consists of a Human Tumor Metastasis Array. In this vein, MDA-MB-231 cells were exposed for 16 h either to hypoxic conditions (Fig. 5a) or the recombinant human IL-1β (Fig. 5b), in the presence or absence of the IL1R1 antagonist IL1R1a. Considering the genes resulting with at least a 0.25 log2 fold change upon either hypoxia respect to normoxia-exposed cells (Fig. 5a) or IL-1β respect to vehicle-treated cells (Fig. 5b), 31 and 15 metastatic genes induced respectively by hypoxia and IL-1β were identified (Fig. 6a). Among the 11 genes up-regulated by either hypoxia or IL-1β as depicted in the Venn diagram (Fig. 6a and Additional file 3), we then focused on the regulation of IL-1β and its target gene cyclooxygenase 2 (COX2) considering that their strong increase was abrogated using the IL1R1 antagonist IL1R1a (Fig.5a-b). First, we ascertained that a positive correlation does exist between IL-1β and COX2 levels in TNBC patients of the METABRIC dataset (Fig. 6b). Thereafter, we confirmed by real-time and western blotting assays that the IL1R1 antagonist IL1R1a prevents the mRNA (Fig. 6c) and protein (Fig. 6d-e) expression of both IL-1β and COX2 induced by a 16 h exposure to either hypoxia or IL-1β treatment, in accordance with previous evidence demonstrating that IL-1β may auto-regulate its own production [49]. Furthermore, we demonstrated that the COX2 protein induction by a 16 h hypoxic condition is prevented by silencing GPER (Fig. 6f), indicating that GPER is involved not only in the above described IL-1β expression but also in the increase of COX2 upon hypoxia.

A hypoxic environment may lower cell-ECM adhesion prompting cell motility and the epithelial to mesenchymal transition toward aggressive breast cancer phenotypes [50, 51]. Hence, we evaluated the contribution of GPER and IL-1β/IL1R1 mediated signaling in the metastasis-related responses induced by the hypoxia-stimulated IL-1β action. In this regard, we assessed that MDA-MB-231 cells cultured under hypoxic conditions exhibit a reduced spreading ability on fibronectin 60 min after seeding respect to cells grown under normoxic conditions (Fig. 7a-b). Notably, this effect was no longer evident silencing GPER as well as using the IL1R1 antagonist IL1R1a (Fig. 7a-b). Results similar to those mentioned above were observed treating cells with IL-1β (Fig. 7c-d). Together, these findings suggest that both GPER and IL-1β/IL1R axis contribute to hypoxia-decreased cancer cell spreading on the ECM protein fibronectin, according to previous data correlating the hypoxia-lowered cell-ECM adhesion to an increased invasion, intravasation and metastasis [52, 53]. Next, the invasive effects prompted by both hypoxia and IL-1β treatment were prevented not only by silencing GPER but also using the IL1R1 antagonist IL1R1a, as evaluated by Transwell Matrigel invasion assays (Fig. 7e-f). Moreover, both GPER silencing and IL1R1a lessened the spheroid expansion upon hypoxia and IL-1β treatment (Fig. 7g). Overall, these findings may suggest that the signaling network of both HIF-1α/GPER and IL-1β/IL1R, may lead to an auto-regulatory loop of IL-1β toward invasive and metastatic properties of hypoxic TNBC cells.

On the basis of the aforementioned findings, we sought to provide mechanistic insights into the paracrine IL-1β feedback on important components of the tumor microenvironment as CAFs. In this vein, CAFs were cultured with the conditioned medium (CM) collected from MDA-MB-231 cells, which were exposed to hypoxia. Interestingly, the up-regulation of both IL-1β and COX2 protein levels occurring in CAFs upon this experimental condition (Fig. 8a) was prevented either using the IL1R1 antagonist IL1R1a (Fig. 8a) or the neutralizing IL-1β antibody (Additional file 4). Findings similar to those above mentioned were observed in CAFs treated with IL1β (Fig. 8b). Considering that hypoxia may stimulate breast cancer cells toward invasive features as cell motility, formation of stress fibers and matrix contraction [54], the CM from MDA-MB-231 cells exposed to hypoxia promoted in CAFs the phosphorylation of myosin light chain (MLC) on serine-19 (pMLC^S19) (Fig. 8c), which is known to be required for the coordination of the actin-myosin contractility [55]. In accordance with these results, an increased MLC phosphorylation was evidenced in CAFs treated with IL1β (Fig. 8d). Moreover, immunofluorescent staining of polymerized actin (F-actin) using FITC-conjugated phalloidin revealed an augmented formation of stress fibers upon exposure of CAFs to the CM from hypoxia-stimulated MDA-MB-231 cells (Fig. 8e) as well as upon treatment with IL-1β (Fig. 8d). Notable, both pMLC^S19 phosphorylation and the increased formation of actin stress fibers triggered in CAFs by the CM from hypoxic MDA-MB-231 cells (Fig. 8c, e) and IL-1β treatment (Fig. 8d, f), were abolished in the presence of the IL1R1 antagonist IL1R1a. Then, we observed that the invasion of CAFs promoted by the CM from hypoxia-exposed MDA-MB-231 cells (Fig. 8g) or by IL-1β treatment (Fig. 8h), was no longer evident in the presence of IL1R1a. Altogether, these results show that the IL-1β production by the hypoxic TNBC cells may trigger a paracrine feed-forward signaling that stimulates malignant features in main components of the breast tumor microenvironment as CAFs.