The development of metastasis represents the most crucial challenge in the treatment of cancer and remains the major cause of cancer related death. Metastatic spread of cancer represents a complex program, during which cancer cells must undergo many changes associated with increased motility, invasiveness, and loss of cell-cell junctions. This conversion is an adaptation of the normal developmental process called epithelial-mesenchymal transition (EMT).
Although EMT term suggests the strict binary conversion from one phenotype to another, new studies showed that EMT is an extremely complex and dynamic process that gives rise to the wide spectrum of intermediate states. The intermediate or hybrid epithelial/mesenchymal phenotypes are characterized by the co-existence of both epithelial and mesenchymal traits and differ in combined expression of epithelial and mesenchymal markers reflecting the ability of the cells to acquire an EMT-associated phenotype manifested by stemness, tumorigenicity, metastatic ability, and resistance to therapy [1
]. The switch in gene expression and associated phenotypic conversion is achieved by the orchestrated and coordinated action of transcription factors (TFs), which bind to their corresponding sequences, leading to the repression of epithelial markers (such as E-cadherin) and conversely to the activation of genes associated with the mesenchymal phenotype [4
The three families of TFs SNAIL, TWIST, and ZEB occupy a privileged position in the triggering of EMT [6
]. ZEB1 protein (zinc finger enhancer binding protein, δEF1), similarly to the SNAIL and TWIST family, includes two dominant members, ZEB1 and ZEB2, that regulate the expression of genes involved in the control of cell polarity and adhesiveness [7
]. Moreover, these proteins have been described as the important players regulating various processes during an organism’s development, as well as fibrosis and cancer progression, including the regulation of metastasis by the regulation of EMT [9
]. Higher levels of ZEB1 are associated with aggressive cancer features, high tumor grade, resistance to therapy, metabolic plasticity, increased incidence of metastasis, and worse clinical prognosis in the vast majority of human cancers [10
]. As a transcription factor, ZEB1 binds directly to specific DNA sequences known as E-boxes (CANNTG) in the promoter region of target genes [11
Anterior gradient protein 2 (AGR2) is highly conserved in vertebrates, playing an important function in the development of cement gland in Xenopus laevis
]. In humans, altered expression of AGR2 was described in various adenocarcinomas, such as breast, esophagus, pancreas, lung, and ovary [14
] and has been shown to contribute to the acquisition of several cancer cells hallmarks, such as tumor proliferation, anchorage-independent tumor growth, formation of metastasis, and resistance to apoptosis and chemotherapy. Although the functions of AGR2 in cancers have been studied intensively in recent years, so far only a few strategies have been demonstrated to regulate AGR2 expression, such as hormone dependent regulation [15
], various microRNAs [16
], shortening of 3’UTR (3’untranslated region) mRNA [18
], and endoplasmic reticulum stress [19
]. However, the regulatory mechanism responsible for the alterations in AGR2 expression during the reversible transition between the epithelial and mesenchymal phenotype still remains obscure.
Therefore, based on recent findings demonstrating the contribution of AGR2 to the EMT and cancer progression, in addition to in silico analysis of AGR2
promoter predicting a binding site for ZEB1 [20
], we aimed to confirm regulatory effect of ZEB1 on the expression of AGR2. To minimize the impact of hormone regulation on AGR2 expression, we analyzed the ZEB1/AGR2 relationship in lung adenocarcinomas, where AGR2 expression was previously described [21
]. In our present study, we demonstrate the ZEB1-mediated repression of AGR2
and propose the existence of a negative feedback regulatory mechanism through which AGR2 controls the stability of ZEB1
Since the presence of metastasis remains the most common cause of cancer-related death, identification of genes associated with metastasis development and characterization of their regulation mechanism may lead to the development of new, more effective therapies targeting metastatic disease. Epithelial-mesenchymal transition and its reverse process, mesenchymal-epithelial transition (MET), play an important role in embryogenesis, organ fibrosis, stem cell biology, and cancer progression [39
]. These processes are regulated by coordinated changes in the expression of core transcription factors ZEB1/2 SNAI1/2, TWIST1/2, acting as repressors of the epithelial phenotype. Although increased efforts are devoted to understanding the machinery of EMT and MET, many regulators of these processes, which represent potential therapeutic targets, still remain elusive. Previous reports have shown that both AGR2 and ZEB1 are involved in the regulation of EMT/MET [10
], processes closely related to metastasis promotion [41
]; however, the mutual relation between AGR2 and ZEB1 had not been suggested so far. This constitutes the reason why we aimed to investigate the relationship between AGR2 and ZEB1 and how their crosstalk could contribute to the process of EMT. Epidermal growth factor receptor (EGFR) mutations are one of the key characteristics of lung adenocarcinomas and their presence predicts treatment decision [42
]. Interestingly, the study by Zhang et al. showed that EGFR mutations positively correlate with the loss of ZEB1 [38
]. Furthermore, increased expression of AGR2 could serve as a positive biomarker of mutated EGFR predicting the sensitivity to anti-EGFR therapy [44
]. When linked together, these studies present yet another suggestion for the inverse relationship between ZEB1 and AGR2. In the present work, we demonstrate for the first time that AGR2 and ZEB1 expression shows an inverse correlation in lung adenocarcinomas, since AGR2 is usually overexpressed in tumor tissue as compared to the normal tissues, while ZEB1 shows the opposite trend. However, in the late and metastatic stages of lung cancer, expression profiles are changed in favor of ZEB1 [12
], which is associated with a decreased level of AGR2 (Figure S4
Although the involvement and molecular function of AGR2 in tumorigenesis has been exponentially reported over the past decade, its regulation during cancer development and progression has not been completely elucidated. In addition to stress conditions such as hypoxia and ER stress that regulate AGR2 protein expression, several transcription factors have been described to be involved in AGR2
promoter activation, predominantly FOXA1/2, FOXM1, and recently TWIST1 [22
]. Similar to ZEB1, TWIST1 belongs to the EMT-associated TFs, but the effect on AGR2 expression remains unclear, as the only two articles published so far present contradictory results on TWIST1 and AGR2 [46
]. In contrast, our data clearly demonstrate that AGR2 expression is suppressed by ZEB1 expression in lung cancer cells. At the same time, these results do not exclude that the various TFs associated with EMT may influence AGR2 in various manners and to a different extent.
Interestingly, we confirmed the existence of a double-sided regulation that involves AGR2, miR-200c, and ZEB1. The ChIP assay confirmed the previous prediction that ZEB1 binds to the AGR2
promoter and represses the transcription of AGR2
]. Similarly, ZEB1 binds to the promoter of miR-200c
and represses transcription of this miRNA [49
]. However, miR-200c is also one of the essential repressors of ZEB1 in cancer cells, since it binds directly to the 3’UTR region of ZEB1
mRNA and enhances its degradation [34
]. Complementary to these findings, we show that AGR2 protein significantly contributes to miR-200c-dependent suppression of ZEB1, and in this way, AGR2 together with miR-200c
prevent the acquisition of an aggressive phenotype, since decreased ZEB1
expression is tightly associated with the attenuation of migratory and invasive properties of cancer cells [51
]. In line with these findings, we also show that knockout of AGR2
not only upregulates ZEB1 expression and activity (Figure 4
A–C), but it also enhances the metastatic dissemination of lung cancer cells (Figure 6
C). An observed mutual relationship between AGR2/miR-200c is supported by Ljepoja et al., showing that knockout of miR-200c leads to significant downregulation of AGR2, which is associated with advanced cancer-subtypes due to activated EMT, which is in turn associated with increased migration and chemoresistance [53
]. Taken together, these data indicate that the expression of AGR2 may be attenuated in tumor cells directly by binding ZEB1 to the AGR2
promoter or indirectly by inhibition of miR-200c
Our data demonstrate that the alterations in AGR2 expression affect the stability of ZEB1
mRNA, since ZEB1
mRNA is more stable in cells lacking AGR2. In contrast, the presence of AGR2 significantly enhances the degradation of ZEB1
mRNA. These findings are also supported by the protein-protein interaction of AGR2 with hnRNPU, which is part of the complex consisting of actin and histone acetyltransferase p300/CBP-associated factor (PCAF) [37
]. Interestingly, this ribonucleoprotein complex predominantly responsible for the regulation of RNA polymerase II transcription elongation was previously shown to be involved in the regulation of the miR-200 family [35
These findings, together with our in vivo data, offer new insight into cancer progression and metastatic cascade. We show that lung cancer cells A549 with endogenous expression of AGR2, when injected subcutaneously, form more easily primary tumors compared to A549 cells with AGR2
gene knockout which generate significantly smaller tumors. This is in accordance with the previously described function of AGR2 as the inducer of tumor growth [22
]. Moreover, Milewski et al. also showed that AGR2
is transcriptionally activated by FOXM1, resulting in mucinous character, accelerated growth, and invasiveness of these adenocarcinomas [22
]. However, our results rather indicate that presence of AGR2 inhibits EMT, whereas A549 cells with silenced AGR2
are more prone to develop lung micrometastases in mouse xenografts in contrast to cells with endogenous AGR2 overexpression. Therefore, we hypothesize a model in which regulated modulation of AGR2 expression may serve as a driver of metastatic progression. Higher expression of AGR2 in a primary site may serve as the growth promoter, while activation of EMT program decreases AGR2 level, thus helping cancer cells to disseminate, and finally, MET switch may re-activate expression of AGR2, leading to enhanced adhesion and easier colonization of a secondary site.
4. Materials and Methods
4.1. Cell Lines and Reagents
Lung cancer cell lines: A549 (ATCC® CCL-185™) and H1299 (ATCC® CRL-5803™) as well as the epidermoid carcinoma derived A431 (ATCC® CRL-1555™) and immortalized HEK 293 (ATCC® CRL-1573™) derived from human embryonic renal epithelium, were maintained in high glucose Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS, Life Technologies, Darmstadt, Germany), 1% pyruvate, and L-glutamine at 37 °C in a humidified atmosphere of 5% CO2. All cell lines were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). Unless otherwise stated, cells were grown to 70–80% confluence prior to treatment. TGF-β (R&D Systems, Minneapolis, MN, USA) was added to a final concentration of 1 ng/mL for 24 hours and 10 µg/mL actinomycin D (Life Technologies, Darmstadt, Germany).
Cells were transfected with 2 μg of plasmid DNA or 50 pmol of siRNA oligonucleotides (Dharmacon, ThermoFisher Scientific, Pittsburgh, PA, USA) per million cells. The Flp-InTM
System (Invitrogen, Carlsbad, CA, USA) was used to generate H1299-LZ4 cells (hereinafter H1299) containing a single integrated Flp Recombination Target (FRT) site. The coding sequence of the human AGR2
gene was stably inserted into this site using Flp recombinase mediated site-specific DNA recombination to give H1299-LZ4-AGR2 (H1299 AGR2) cells. Cell lines with AGR2
gene knockout were prepared as described previously using CRISPR/Cas9 [24
]. Briefly, A549 cells were transfected with plasmid LentiCRISPR-v2_AGR2 or LentiCRISPR-v2_scrambled serving as a control. Then, the cells were exposed to puromycin for several weeks. Clones were selected from the pool of resistant cells and tested for AGR2 expression and validated by sequencing. Two cells clones were further used: A549 scrambled (A549 scr, with AGR2 expression) and A549 KOAGR2 (without AGR2 expression).
4.2. Gene Expression
Total RNA was isolated using Ribozol reagent (VWR, Lutterworth, UK). The cDNA was synthesized by RevertAid H Minus Reverse Transcriptase (Life Technologies, Darmstadt, Germany). Either SYBR Green MasterMix (Roche, Basel, Switzerland) or TaqMan Universal PCR MasterMix (Life Technologies, Darmstadt, Germany) were used for quantitative PCR. 18S
rRNA and GAPDH
served as parallel endogenous controls. The data represent means of three technical triplicates within each independent biological replicate (n
= 3). The primer sequences are listed in supplementary Table S2
. The relative mRNA expression levels of each gene were calculated using the 2−ΔΔCT
4.3. TaqMAN Advanced microRNA Assay
MicroRNA expression levels were determined using TaqMan Advanced miRNA Assays (ThermoFisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. The final data show means of three technical triplicates within each biological replicate (n = 3) and the obtained average CT values for target miRNA and RNU48 serving as a reference for data normalization were used to calculate relative gene expression using the 2−ΔΔCT method.
4.4. Western Blot Analysis
Cells were washed twice with cold phosphate-buffered saline (PBS) and then scraped into NET lysis buffer (150 mM NaCl, 1% NP-40, 50 mM Tris-HCl, pH 8.0, 50 mM NaF, 5 mM EDTA, pH 8.0) supplemented with protease and phosphatase inhibitor cocktails according to the manufacturer’s instructions (Sigma-Aldrich, St. Louis, MO, USA). Following sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), the samples were transferred on to nitrocellulose membranes and incubated overnight at 4 °C with the primary antibodies. The following day, membranes were washed and probed with horseradish peroxidase conjugated with secondary antibodies (1:1000) for 1 hour at room temperature. Chemiluminescent signals were developed using the ECL solution and visualized with GeneTools (Syngene, Cambridge, UK). Either α-tubulin or β-actin was used as a loading control and as a reference for protein expression normalization. Subcellular fractionation using NE-PER Nuclear and Cytoplasmic Extraction Reagent (ThermoFisher Scientific, Waltham, MA, USA) was conducted according to the manufacturer’s instructions.
Antibodies: ZEB1 (Cell Signalling Technology Danvers, MA, USA, Santa Cruz Biotechnology, Dallas, TX, USA); AGR2 (K-31, in-house); Lamin B1, β-actin (Santa Cruz Biotechnology, Dallas, TX, USA), Alexa Fluor 488 goat anti-mouse IgG, Alexa Fluor 532 goat anti-rabbit IgG (both Abcam, Cambridge, UK); horseradish peroxiadase (HRP)-conjugated swine anti-rabbit and HRP-conjugated rabbit anti-mouse (both Dako, Glostrup, Denmark).
4.5. Luciferase Reporter Gene Assay
Cells were grown in 12-well plates and transfected using PEI (polyethylenimine, Sigma-Aldrich, St. Louis, MO, USA). Renilla luciferase reporter plasmid pLuc-CDS (#42100, Addgene, Watertown, MA, USA) bearing ZEB1 coding region was co-transfected with mammalian reporter vector pGL3, allowing for weak constitutive expression of firefly luciferase. Twelve hours later, the cells were treated with TGF-β and luciferase activity was measured using the Dual-Luciferase Assay System (Promega Corporation, Madison, WI, USA) after an additional 24 hours incubation. The pGL3 reporter plasmid with cloned AGR2 promoter sequence from −1584 to +96 was used to analyze the efficiency of AGR2 transcription. In this case, Renilla luciferase vector served as an internal control.
4.6. Cell Invasion Assay
Cell invasion was assessed using a CytoSelect 24-well Cell Invasion Assay kit (Cell Biolabs, San Diego, CA, USA) according to the manufacturer’s protocol. Briefly, at 36 hours post-transfection, 1 × 105 cells in 300 μL serum-free medium were added to the upper chamber precoated with basement membrane matrix solution. Subsequently, 0.5 mL of 10% FBS-containing medium were added to the lower chamber as a chemoattractant. The cells were incubated for 36 hours at 37 °C, then the non-invading cells were removed with cotton swabs. The cells, which migrated to the bottom of the membrane, were fixed and stained with staining solution for 10 minutes. The inserts with stained cells were air-dried, then incubated in the extraction solution for 10 minutes and absorbance was measured at a TECAN spectrophotometer (Tecan, Zürich, Switzerland) at 560 nm.
4.7. Chromatin Immunoprecipitation (ChIP) Assay
A549 cells were grown to 80% confluency, collected and cross-linked for 10 minutes at 37 °C with 1% formaldehyde. Glycine was added to a final concentration of 0.5 M for 5 minutes at 37 °C. Cells were then washed with PBS, scraped and centrifuged (2400 rpm, 4 °C for 10 minutes). The supernatant was removed and the pellet was resuspended in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM KCl, 0.1% SDS, 5 mM EDTA, 1% NP-40, 0.5% sodium deoxycholate, and 1% protease inhibitor cocktail from Sigma-Aldrich, St. Louis, MO, USA), left on ice for 10 minutes and snap frozen in liquid nitrogen. Samples were then sonicated at 4 °C (10 × 20 seconds with a 30 seconds pause, VibraCell, Sonics & Materials, Newtown, CT, USA). Supernatants were recovered by centrifugation (13000 rpm, 4 °C for 10 minutes) and precleared for 1 hour at 4 °C with protein G-sepharose beads. Beads were prepared by several wash cycles, followed by 1 hour incubation with salmon sperm at room temperature and then diluted in dilution buffer (45 mM Tris-HCl, pH 7.5, 135 mM KCl, 0.9% NP-40). Immunoprecipitations were performed overnight at 4 °C with ZEB1 (ZEB H-102, Santa Cruz Biotechnology, Dallas, TX, USA) or control IgG. The precipitated samples were washed sequentially 5 times for 10 minutes each at 4 °C in Wash Buffer (50 mM Tris-HCl, pH 7.5, 150 mM KCl, 1%NP-40, 0.25% sodium deoxycholate). Beads were eluted with 100 μL of Elution Buffer (0.1% SDS, 50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 10 mM DTT). Cross-links were reversed by overnight incubation at 65°C, before DNA purification with the PCR cleanup kit (QIAGEN, Hilden, Germany) and qRT-PCR. The sequences of primers for PCR analysis are listed in Table S1
4.8. Immunoprecipitation (IP)
For immunoprecipitation experiments, cells were extracted using lysis buffer supplemented with complete protease inhibitor cocktail. Cell lysates containing 200 µg of whole proteins were incubated with 1 µg/mL anti-AGR2 mouse antibody (Abnova, Heidelberg, Germany) overnight at 4 °C with gentle agitation. Immune complexes were isolated by incubation with Protein G Sepharose 4 Fast Flow beads (GE Healthcare, Chicago, Illinois, USA) at 4 °C for 2 hours, followed by five washes in lysis buffer. Immune complexes were eluted with 2× Laemmli buffer (Invitrogen, Carlsbad, CA, USA), boiled and set aside for immunoblotting.
4.9. Tumor Xenografts
First, 4 × 106 of either A549 scr or A549 KOAGR2 cells were resuspended in 100 µL PBS and injected subcutaneously into the left and right flanks of 5–6 week old female SCID mice. Mice were divided into two groups of 6, injected with either A549 scr or A549 KOAGR2 cells. Tumors were allowed to grow for 9 weeks after the injections. During this period, tumor volumes were measured (as soon as tumor onset was observed) using a calliper and calculated using the formula ½ × height × width × length. At the end of the observation period, mice were sacrificed and tumors were excised and photographed. Primary tumors and lungs were removed, fixed, and embedded in paraffin. Hematoxylin-eosin staining was used for histopathological evaluation.
All experiments with mice were performed in the authorized animal house of the National Hellenic Research Foundation. Experiments complied with the Protocol on the Protection and Welfare of Animals, as obliged by the rules of the National Hellenic Research Foundation, the regulations of the National Bioethics Committee, and article 3 of the presidential decree 160/1991 (in line with 86/609/EEC directive) regarding the protection of experimental animals.
4.10. Statistical Analysis
The error bars represent the standard deviation of corresponding data sets. One-way ANOVA (analysis of variance) with post-hoc Tukey HSD (Honestly Significant Difference) calculator was used to determine statistically significant differences between the groups generated from at least three independent experiments. Statistical analysis was performed using the free online web tool denoted as One-way ANOVA (ANalysis Of VAriance) with post-hoc Tukey HSD (Honestly Significant Difference) Test Calculator for comparing multiple treatments. Tests with p < 0.05 were considered as significant.