Background
Crosstalk between cancerous and stromal cells in an inflammatory-based microenvironment not only affects the aggressive behavior of cancer cell, but also influences the efficacy of anti-cancer drugs [
1‐
4]. On the other hand, drug impacts are usually assayed in cell culture or immunosuppressed mice models [
5], neither of which provides a clear picture of the complicated interactions between tumor cells and their surrounding inflammatory stroma, eventually resulting in inconsistencies in the outcome of clinical trials [
6]. It is thus naïve to assess the sole effect of inhibiting a drugable target in the aggressive cancer cells while ignoring the context of micro-environmental changes [
4]. Communication between cancer cells and their surrounding stroma is through secretion by these cells of soluble factors including growth factors, chemokines and cytokines [
7]. The secreted factors reprogram the surrounding stroma with the aim of neutralizing the impact of various intruders disrupting the survival of the cancer cells [
8,
9].
One of the most influential cytokines in tumor microenvironment is Tumor Necrosis Factor-alpha (TNF-α) [
10,
11], long known for its dual effects cancer [
12]. Depending on the cellular context as well as its concentration, the effects of TNF-α may vary from inducing necrosis to survival benefits, endowing invasive properties and driving epithelial-mesenchymal transition (EMT) [
13].
Considering the pleiotropic and conflicting effects of TNF-α, various efforts have been made to exploit this cytokine as a therapeutic target in cancer. In this regard, some approaches have used TNF-α administration to tumor tissue to induce cell death due to its necrotizing effects [
14]. However, using TNF-α as a necrotizing agent has been associated with significant lethal side effects limiting the applicability of TNF-α administration to patients [
15]. On the other side, TNF-α activates NFKB transcription factor along with JNK and p38 in a cellular context-dependent manner [
16]. This canonical pathway induces the expression of several downstream targets of NFKB leading to the increased invasiveness of cancer cells. A different approach that has attracted a great of attention is the use of TNF-α inhibitors, such as Etanercept, as candidates for cancer biotherapy [
17].
Etanercept is a dimeric soluble receptor for TNF-α, which competes with the cell surface receptors For TNF binding. A phase II clinical trial evaluating the therapeutic efficacy of Etanercept in patients with advanced metastatic breast cancer revealed a reduction in the concentration of TNF-α related cytokines such as CCL2; however, disease stabilization was detected in only one patient [
18]. In order to elucidate the underlying reason for this outcome, we aimed at integrating molecular-cellular experiments with systems modeling approaches to compare the effect of Etanercept in breast cancer cells alone and in the context of host immune cells.
Methods
High through-put data analysis and enrichment-based tests
Proteomics data providing the secretome of MDA-MB-231 cells were extracted from GSE51938 dataset in GEO. The secretion profile was submitted to Enrichr online web (
http://amp.pharm.mssm.edu/Enrichr/). “LINCS L1000 Ligand Perturbations up” library was used for enrichment analysis. This library consists of gene expression changes induced by various ligands (e.g. growth factors and cytokines) and displays ligands inducing similar gene expression profiles.
Gene-set enrichment tests including gene ontology (GO) biological process, transcription factors and signaling pathways were carried out using the corresponding libraries embedded in Enrichr online web tool.
Boolean network simulation
Cellular signaling network was constructed based on the literature survey [
19] and translated into Boolean rules. Asynchronous simulation mode was used to predict the effect of macrophages on the anti-cancer efficacy of Etanercept. BoolNet R package was used to perform the dynamic modeling and analysis of attractor states [
20].
Cell culture and condition medium preparation
MDA-MB-231 human breast cancer cell line and THP-1 human monocytic cells were purchased from Pasteur Institute of Iran (Tehran, Iran). A total of 2 × 105 THP-1 cells were seeded in 2 ml RPMI supplemented with 10% FBS in the upper chamber of 6-well insert plates (SPL Life Sciences, Seoul, South Korea). Cells were treated with 10 ng/ml of PMA (Phorbol 12-myristate 13-acetate) for 24 h to allow differentiation of THP-1 cells to M0 macrophages and their subsequent attachment to the inserts. After 24 h, 2 × 105 of MDA-MB-231 cells were seeded in the lower chamber in 2 ml of DMEM. The co-culturing system was kept in 5% CO2 incubator for 72 h.
To prepare condition medium (CM), the upper insert was withdrawn and MDA-MB-231 cells were washed 3 times with PBS and starved for 24 h in the serum free DMEM. The condition medium was then filtered using 0.2 μm filter and kept at − 70 °C until use. CM was diluted with 50% of fresh medium for further experiments.
Tetrazolium-based viability test
Co-culturing of MDA-MB-231 with THP-1 cells was performed as described in the previous section. Co-cultured as well as MDA-MB-231 cells were treated with 1 or 10 μg/ml of Etanercept (Aryogen Pharmed, Tehran, Iran) or control buffer for 72 h. The concentration of Etanercept was selected based on the plasma concentration obtained after weekly administration of the drug to the arthritis rheumatoid patients [
7]. After 72 h, THP-1 containing inserts were discarded, and 20 μl of 5 mg/ml of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) stock solution was added to each well. Following 3 h of incubation, the supernatants were aspirated, and 200 μl of dimethylsulfoxide (DMSO) were used to dissolve purple formazan in each well. Absorbance was measured at 545 nm using microplate Reader (Stat Fax-2100, ST. Louis, USA).
Cell cycle analysis
MDA-MB-231 cells were harvested after 72 h of co-culturing with the differentiated macrophages in the presence and absence of 1 or 10 μg/ml Etanercept. Cells were washed 3 times with cold PBS, detached using 0.025% trypsin and re-suspended in 70% ethanol and stored at − 20 for at least 24 h to allow cell fixing. Cells were pelleted by centrifugation at × 400 g and ethanol was discarded. They were then stained with 200 μl of PI (Propidium Iodide) (1 mg/ml) stock solution supplemented with 150 μl of stock solution of DNAse-free RNAse A (CinaClone, Tehran. Iran) (2 mg/ml) and 0.1% Triton X-100 in 10 ml PBS, and then analyzed by Flow cytometry (PARTEC GmbH, Munster, Germany). FlowJo software was used for data analysis and measuring fraction of the cells in four stages of the cell cycle, i.e. G1, S, G2 and subG1.
Annexin-PI apoptosis assay
To measure the percentage of apoptotic cells induced by Etanercept either in MDA-MB-231 cells alone or co-cultured with the macrophages, Annexin-V staining with PI was used based on the manufacture’s instruction. Briefly, cells were washed with PBS and harvested with 0.025% trypsin. Detached cells were centrifuged at 400×g for 5 min and re-suspended in Annexin-PI solution containing 2 μl annexin-V-FLUOS labeling agent, 2 μl PI solution and 1 ml incubation buffer, and incubated at 37 °C for 15 min, and then analyzed by flow cytometry.
NFKB activity determination
A total of 1 × 10
6 MDA-MB-231 cells were seeded into 100 mm cell culture plates and allowed to attach for 24 h. Cells were then treated with 50% diluted CM or normal medium as control for 72 h. Activity of NFKB was determined using TransAM® NFκB Transcription Factor ELISA Kit (#40096, Active Motif, Belgium) as described in the kit manual. In brief, following the incubation period, cells were washed twice with PBS supplemented with PhosStop phosphatase inhibitor cocktails (Roche, USA) and 0.05% NP-40. Ultra-centrifugation was used to extract nucleus from the cytoplasmic fraction. The nucleus pellet was further treated with nucleus lysis buffer and centrifuged at × 14,000
g to extract nuclear proteins. Protein assay was performed using Bradford method [
21], and 10 μg of nuclear extract was added to each well of DNA-coated wells. To calculate the fraction of DNA-attached NFKB, HRP-conjugated anti-phospho-p65 primary antibody was added to each well and the fluorescence intensity was measured at 630 nm (Stat Fax-2100, ST. Louis, USA).
Quantitative real time PCR for the analysis of macrophage cytokines
A total of 1 × 10
6 of differentiated THP-1 cells using10 ng/ml of PMA were treated with CM for 48 h and incubated at 37 °C. Cells were washed with PBS and lysed using RNX - Plus solution (CinaClone, Tehran, Iran). Total mRNA was extracted and subjected to cDNA synthesis using Revert Aid first strand cDNA synthesis kit (ThermoFisher Scientific, USA) according to the manufacturer’s instructions. PCR was carried out using Taqman mastermix [
21]. Forward (F) and reverse (R) primer sequences used were as following:
TNF-Α-F: 5′-CCCTGACATCTGGAATCTGGAG-3′,
TNF-Α-R: 5′-TCAAGGAAGTCTGGAAACATCTGG-3′,
CCL22-F: 5′-TGGGTGAAGATGATTCTCAATAAGC-3′,
CCL22-R: 5′-CTATAATGGCAGGGAGGTAGGG-3′,
IL10-F: 5′-CTTGCTGGAGGACTTTAAGGGTTAC-3′,
IL10-R: 5′-CTTGATGTCTGGGTCTTGGTTCTC-3′.
ΔΔCT was calculated and normalized according to the expression of GAPDH using the following primers, GAPDH-F: 5′-ACATCAAGAAGGTGGTGAAGCAG-3′, GAPDH-R: 5′- GCGTCAAAGGTGGAGGAGTG-3.
Matrigel invasion assay
MDA-MB-231 cells were grown in CM for 72 h. Cells were washed 3 times with PBS and starved with serum-free DMEM for 24 h. 24-well insert plates (SPL Life Sciences, Seoul, South Korea) with 8 μm pore size were coated with 1 mg/ml of matrigel (BD Biosciences, San Jose, USA). After 4 h of incubation of the coated plates at 37 °C, when a smooth layer was formed on the insert, a total of 1 × 105 starved cells were added to the upper chamber of each well in serum free medium, and the lower chamber was supplemented with DMEM containing 10% FBS as attractant. After 24 h, the inserts were fixed in 4% formaldehyde and stained using 0.05% crystal violet. Traversed cells across the membrane were visualized under an inverted microscope.
Statistical analysis
For the enrichment tests, adjusted p-value by Bonferroni method < 0.05 was considered as significant. For experimental assays, GraphPad Prism was used for data analysis. Data represent mean ± S.E.M from the indicated replicates for each test. ANOVA was used to determine the significance of difference (p-value < 0.05).
Discussion
NFKB is highly active in the invasive breast cancer cells resulting in the secretion of a number of cytokines, a condition similar to that of cells treated with TNF-α. These data fueled our initial hypothesis that using clinically available TNF-α inhibitors would be a smart approach to alter the secretion profile of invasive breast cancer cells, and thus motivated us to assess the inhibitory effect of Etanercept, a TNF-α inhibitor, to suppress proliferation and survival of breast cancer cells. Our hypothesis is also supported by the findings of other studies indicating the efficacy of Etanercept, a soluble receptor, to capture and neutralize circulating TNF-α in breast cancer treatment [
18,
27]. The present study showed that inhibition of TNF-α signaling by Etanercept results in a reduction in the activity of NFKB transcription factor, leading to reduced survival rates of MDA-MB-231 cells, and induced cell cycle arrest and apoptosis.
Notably, the effect of anti-TNF therapies is still controversial with conflicting outcomes at patient level. In this regard, it has been shown that tumor microenvironment exerts profound effects not only on the behavior of cancer cells but also on the efficacy of treatments that has to be taken into account when evaluating the effect of novel drugs [
28]. To address this issue, using publicly available proteomic data of MDA-MB-231 secretome, a high level of Granulocyte-Monocyte Stimulating Factor (GM-CSF) expression was detected in the secretome, which modulate activation of tumor associated macrophages. Accordingly, our findings on cell cycle analysis, annexin/PI and matrigel assays further confirmed that when co-cultured with MDA-MB-231 cells, macrophages reduce the anti-proliferative and anti-invasive effects of Etanercept in cancer cells. These findings suggested that the anti-cancer effect of Etanercept can be diminished or abrogated by resident macrophages in the tumor site, which would in turn suggest a role for the microenvironment in treatment efficiency.
We also observed that in the co-culture condition, Etanercept does not reduce the activity of NFKB to the extent observed in breast cancer cells alone. These results underscore the importance of stromal cell secretome in resistance to the targeted therapies. Tumor associated macrophages are important components of tumor stroma, which link inflammation and cancer progression by secreting various cytokines. The secreted cytokines from macrophages can trigger PI3K/AKT/mTOR signaling pathway [
29] and also lead to the activation of NFKB as well as STAT3 in cancer cells [
30‐
32]. Interaction between NFKB and STAT3 is a key mediator of crosstalk in the tumor inflammatory microenvironment [
33]. STAT3 is constitutively activated in triple negative breast cancer, and ample evidence imply that persistently- activated STAT3 maintains constitutive NFKB activity [
34,
35]. Collectively, these results suggest that the presence of macrophages in tumor tissues potentiates the activity of NFKB resulting in subsequent resistance to anti-TNF therapies. In other words, the activated signaling pathways can decrease the reliance of NFKB activation solely on TNF-α. This mechanism may apparently reduce the efficacy of TNF-α inhibition strategy.
Moreover, we used Boolean network modeling to simulate the effect of Etanercept in the presence and absence of macrophages. This modeling provides useful insights to confirm our findings that in the presence of macrophage-associated cytokines and chemokines, activation of JAK-STAT and PI3K-AKT pathways can restore survival and proliferation of tumor cells to compensate the decreased activity of NFKB by Etanercept.
We were also interested to assess the effect of Etanercept on the secretion profile of macrophages that were grown in the CM. We showed that Etanercept reduces the transcript levels of TNF-α, IL10 and CCL22, all of which are associated with tumor progression. Thus our results imply that in addition to inhibiting the aggressive behavior of tumor cells within the context of microenvironment, Etanercept is able to alter the activity of tumor associated macrophages.
As a whole, our data suggest that evaluating the anticancer effect of drugs without considering the effect of stromal cells may lead to failure of drug candidates in the later clinical stages. An integrated use of high through-put data available in the literature and systems biology would allow predicting the effect of drugs in the presence of stromal cells. These data should further be validated by in vivo experiments.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.