GDF15 promotes the proliferation of cervical cancer cells by phosphorylating AKT1 and Erk1/2 through the receptor ErbB2
- 965 Downloads
Growth differentiation factor 15 (GDF15) is a member of the TGF-β superfamily, and evidence suggests that a substantial amount of GDF15 is secreted in various human cancers, such as ovarian cancer, prostate cancer, and breast cancer, among others. However, the function of GDF15 in cervical cancer has not yet been reported.
Immunohistochemistry was used to detect GDF15 expression in normal cervix and in different cervical cancer lesions. Cell growth curves, MTT, tumor formation assays and flow cytometry were utilized to observe the effects of ectopic GDF15 expression on the proliferation and cell cycle of cervical cancer cells. Real-time PCR, western blotting and immunoprecipitation assays were conducted to measure the expression of genes related to the cell cycle and the PI3K/AKT and MAPK/ERK signaling pathways. A chromatin immunoprecipitation assay was performed to confirm whether C-myc bound to a specific region of the GDF15 promoter. Inhibitor treatment and immunoprecipitation assays were employed to identify the association between GDF15 and ErbB2.
GDF15 expression gradually increased during the progression of cervical carcinogenesis. GDF15 promoted cervical cancer cell proliferation via exogenous rhGDF15 treatment or the use of gene editing technology in vitro and in vivo and significantly accelerated the cell cycle transition from G0/G1 to S phase. The expression of p-ErbB2, p-AKT1, p-Erk1/2, CyclinD1 and CyclinE1 was up-regulated and the expression of p21 was down-regulated in GDF15-overexpressing and rhGDF15-treated cervical cancer cells. C-myc trans-activated GDF15 expression by binding to the E-box motifs in the promoter of GDF15 and contributed to the positive feedback of GDF15/C-myc/GDF15. Furthermore, GDF15 bound to ErbB2 in a protein complex in cervical cancer cells.
Our data demonstrated that GDF15 promoted the proliferation of cervical cancer cells via the up-regulation of CyclinD1 and CyclinE1 and the down-regulation of p21 through both the PI3K/AKT and MAPK/ERK signaling pathways in a complex with ErbB2.
KeywordsGDF15 Proliferation ErbB2 Cervical cancer
Growth differentiation factor 15
Mitogen-activated protein kinase
Squamous cell carcinoma
Transforming growth factor β
Currently, cervical cancer is the second most common cancer in women in developing countries after breast cancer, and recent findings indicate an increased risk for young women under 25 years of age who may be predisposed to develop cervical cancer . Multistep factors such as proto-oncogene activation, tumor suppressor gene inactivation, epigenetic alterations, and genomic instability are involved in the development of cervical cancer [2, 3]. Substantial evidence has indicated that various oncogenes (such as LGR5 and DAX1) and suppressor genes (such as KLF4 and SLUG) exhibit abnormal expression during the development and progression of cervical carcinoma [4, 5, 6, 7], which suggests the important role of gene activity in the carcinogenesis of cervical cancer. Although it is known that cervical carcinogenesis is a process in which normal tissues progress to squamous intraepithelial lesions and then cervical cancer, the underlying mechanisms of cervical carcinogenesis have not been fully elucidated.
Growth differentiation factor 15 (GDF15, also known as MIC-1, NAG-1, PTGF-β, and PDF) belongs to the transforming growth factor beta (TGF-β) superfamily. The GDF15 gene is located on chromosome 19 and comprises two exons that encode a 308-amino-acid polypeptide, which consists of a 29-amino-acid signal peptide, a 112-amino-acid mature protein, and a 167-amino-acid pro-peptide. GDF15 is expressed in the cytoplasm as a precursor 35-kDa protein that is cleaved to produce a mature 17-kDa secreted cytokine . GDF15 is weakly and stably expressed in most tissues under normal physiological conditions but is substantially up-regulated under pathological conditions such as injury, inflammation and carcinoma, among other diseases . Evidence shows that GDF15 plays an important role in carcinogenesis-related activities, including proliferation, migration, invasion, and angiogenesis in various types of tumors. For example, GDF15 knockdown in malignant gliomas reduced cell proliferation in vitro and tumorigenesis in vivo , while GDF15 overexpression promoted tumorigenesis and progression in oral squamous cell carcinoma . GDF15 was identified as a novel potential biomarker of cervical cancer in a previous study . Unfortunately, there is no evidence demonstrating the effects of GDF15 expression on the development and progression of cervical carcinoma; the molecular mechanisms of GDF15 in cervical carcinoma are also largely unknown. This study aimed to fully explore the function and mechanisms of GDF15 in cervical carcinogenesis.
Cell lines and culture conditions
Five cervical cancer cell lines HeLa, SiHa, C-33 A, HT-3, CaSki and HL-60 were obtained from American Type Culture Collection (ATCC, Rockville, MD, USA) and maintained in recommended media supplemented with 10% fetal bovine serum at 37 °C and 5% CO2. Dulbecco’s modified Eagle’s medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA) was used to culture HeLa, SiHa and C-33 A cells, McCoy 5A medium (Sigma-Aldrich) was used for HT-3 cells, RPMI 1640 (Sigma-Aldrich) was used for the CaSki and HL-60 cells under the identical conditions. Recombinant human GDF15 (rhGDF15; 213-10,128-1) and Human serum albumin solution (HSA; 020682) was purchased from RayBiotech (Norcross, GA, USA) and Vitrolife (Gothenburg, Sweden), respectively. MK-2206 is an AKT inhibitor, SCH772984 is an Erk1/2 inhibitor, SB525334 is a TGF-β receptor inhibitor and CI-1033 is an EGFR/ErbB2 inhibitor, all of which were purchased from biochemical company (Selleck, USA).
Human tissue specimens, immunohistochemistry and immunocytochemistry
A total of 76 samples, 21 normal cervix (NC), 23 high-grade squamous intraepithelial lesion (HSIL), and 32 squamous cervical cancer (SCC) samples without a history of chemotherapy, immunotherapy, or radiotherapy were obtained by means of surgery at the First Affiliated Hospital of Xi’an Jiaotong University Medical College between January 2008 and December 2016. The procedures followed medical-ethics approval practices. Hematoxylin and eosin-stained sections were used for pathological diagnosis. Clinical stages were identified according to the International Federation of Gynecology and Obstetrics (FIGO) classification system. For IHC, the staining procedure was performed as previously described . Images were captured with an Olympus-CX31 microscope digital camera and Leica DFC 500 digital camera and processed with LAS AF software (Leica, Solms, Germany). To determine immuno-reactivity scores (IRSs), specimen sections were scored and evaluated by 2 investigators. Staining intensity was scored on a scale from 0 to 3: negative (0), weak (1), moderate (2), or strong (3). The percentage of positive cells, defined as the relative positive-staining area, was scored on a scale from 0 to 4: 0–5%(0), 6–25%(1), 26–50%(2), 51–75%(3), and 76–100%(4). The IRS was calculated by multiplying the “staining intensity score” by the “percentage of positive cells”. IRSs ≤2 were defined as GDF15-negative, while IRSs between 3 and 12 (3 ≤ IRS ≤ 12) were defined as GDF15-positive.
The cell culture media was collected and prepared as 10 times dilution. The detailed procedures were conducted strictly with the protocol in the Quantikine® ELISA human GDF15 immunoassay kit (#DGD150, R&D system, Minneapolis, MN, USA).
Western blotting analysis
Western blotting analyses were performed as previously described using 30μg lysates from fresh tissues and cells . Proteins were visualized by exposing Chemiluminescent HRP Substrate (Millipore, Billerica, MA, USA) by protein imprinting imaging system (Tanon5200, China). Antibodies against human GDF15, AKT1, Erk1/2, GAPDH were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Anti-PI3K (p-p110a) was from Abcam Technology (Cambridge, MA, USA). Anti-p-AKT was purchased from CST (Littleton, CO, USA). Anti-Ras and Anti-Ras-GTP was from NewEast Biosciences (WuHan, CN). Anti-p-Erk1/2, anti-ErbB2, anti-p-ErbB2, anti-FOXO1 were from Abclonal Technology (Boston, MA, USA).
Plasmid construction and stable transfectants
The vector pIRES2-AcGFP was purchased from Clontech (CA, USA). The full-length CDS of human GDF15 was cloned into the pIRES2-AcGFP vector. GDF15-specific and ErbB2-specific short hairpin RNAs were designed and cloned into pGPU6/GFP vectors to inhibit GDF-15 and ErbB2 expression (Gene Pharmcompany, Shanghai, China). The primers and oligonucleotide sequences are described in Supplementary Additional file 1: Table S1. The pX330-U6-Chimeric_BB-CBh-hSpCas9 plasmid (Plasmid#42230) containing SpCas9 and single guide RNA was obtained from a non-profit plasmid share repository belonging to Feng Zhang (Addgene, Cambridge, MA, USA) . Suitable CRISPR target sites within the positive and negative strands of exon 1 of GDF15 were identified using the ‘CRISPR Design Tool’ (http://crispr.mit.edu/) hosted by the Feng Zhang laboratory (Massachusetts Institute of Technology, MA, USA) to obtain the minimum number of off-target sites in the human genome (Additional file 1: Table S1). To generate stable transfected cell lines, cells were transfected with vectors using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Stable clones from HeLa and SiHa cells were selected using G418 reagent (Calbiochem, Darmstadt, Germany) and identified by western blotting. For transient HT-3 cell lines, protein was collected after being transfected 48 h.
RNA isolation and real-time PCR
Total RNA was isolated using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA), and first-strand cDNA was synthesized using the PrimeScript RT Reagent Kit (TaKaRa, Osaka, Japan). Complementary DNA and SYBR-Green fluorescence- signal-detection assays (TaKaRa, Osaka, Japan) were used for real-time PCR. The resulting data were normalized to housekeeping gene GAPDH. Real-time PCR and data collection were performed on an IQ5 instrument (Bio-Rad, CA, USA). Primers were designed using Perl Primer software, and their oligonucleotide sequences are described in Additional file 2: Table S2. The real time-PCR experiments were quantified by the 2 − △△CT method and repeated at least three times.
Cell growth and cell viability assays
Cells (5 × 104) were seeded in triplicate with 2 mL of media into 35-mm tissue culture dishes for 7 days. The numbers of cells were counted after harvesting every 2 days using a hemocytometer under light microscopy. The cell viability assays were performed by applying 3-(4,5-dimethylthiazole-yl)-2,5-diphenyl tetrazolium bromide (Sigma-Aldrich, St Louis, MO, USA) dye to cells that were seeded in 96-well plates for 1000 cells each in 7 days or 3000 cells in 48 h, as described in a standard protocol. Then, the absorbance at 490 nm was measured (Bio-Rad, CA, USA).
Flow cytometry analysis
Cells for cell cycle analysis were harvested and fixed with 70% cold ethanol at 4 °C overnight. After being washed in PBS, the cells were incubated in 1 mL of staining solution (20 mg/mL propidium iodide; 10 U/mL RNaseA) (Aldrich, St. Louis, MO, USA) at room temperature for 30 min. Then, the samples were measured by FACS Calibur flow cytometry (BD, Franklin Lakes, NJ, USA), and the cell cycle distributions were analyzed by the software “Flowjo_V10” (FlowJo, LLC, Ashland, OR, USA).
Quantitative chromatin immunoprecipitation
HeLa and SiHa cells were subjected to ChIP using the EZ-ChIP Assay kit (Millipore, MA, USA). Briefly, the cells were treated with 37% formaldehyde to crosslink proteins, and the reaction was terminated with 0.125 M glycine. After sonication, chromatin–protein complexes were immunoprecipitated with 5 μg of anti-C-myc antibodies or 1 μg of mouse IgG. Real-time PCR was performed to amplify the regions of interest or internal negative control regions. Each sample was assayed in triplicate, and the fold enrichment ratio was calculated as the value of the ChIP sample versus the corresponding input sample. Samples that yielded a twofold enrichment or better were considered positive targets. The primers used for these studies are listed in Additional file 2: Table S2.
Primary antibody (anti-GDF15) was incubated with AminoLink plus coupling resin (Pierce Co-IP kit, Rockford USA) for 1 h at room temperature. Equivalent amounts of protein (500 μg) incubated with resin-antibody complex overnight at 4 °C. Protein–antibody complexes were washed in lysis buffer. Denatured protein complexes were separated by western blotting. Another antibody(anti-p-ErbB2) was used to detect interacting proteins. IgG antibodies of corresponding isotype (Santa Cruz, Dallas, TX, USA) were used as controls to exclude unspecific associations.
Tumor xenograft assay
For in vivo assays, nude female mice between 4 and 6 weeks old were purchased from an animal breeding facility (Slac Laboratory Animal Co., Nanjing, CN). Animals were maintained in a specific pathogen free(SPF) room in accord with institutional policies. The mice were randomly divided into four groups, with three repeats for each clone: HeLa-GFP and HeLa-GDF15, SiHa-GFP and SiHa-GDF15. Each mouse received a subcutaneous injection of 1 × 106 cells. The development and progression of solid tumors was monitored longitudinally over 12 weeks. Tumor volume(V) was calculated every 4 days by tumor length(L) and width(W) as: V = L × W2/2.
Each sample was assayed in triplicate, and each experiment was repeated three times. Data are presented as means ± SD, and Student’s t-test (unpaired, two- tailed) or one-way ANOVA was used to compare the means of independent samples. Correlations were analyzed using Pearson linear-regression analysis. Statistical analyses were performed with SPSS for Windows v. 16.0 (SPSS Inc., Chicago, IL, USA). P-values of P < 0.05 were considered statistically significant.
GDF15 expression in normal cervix and different cervical cancer lesions
GDF15 promoted the proliferation of cervical cancer cells and tumor formation in vitro and in vivo
GDF15 promoted the proliferation of cervical cancer cells by accelerating the cell cycle transition from G0/G1 to S phase
Generally, the changes that occur during cell proliferation involve modulation of the cell cycle. To investigate whether GDF15 affected the cell cycle of cervical cancer cells, flow cytometry was performed to analyze differences in the cell cycle between GDF15-modified cervical cells and their control cells. As shown in Fig. 3a and b, the percentage of cells in G0/G1 phase obviously decreased from 59.29% in HeLa-GFP cells to 36.43% in HeLa-GDF15 cells, while the percentage of cells in S phase was markedly increased from 30.51% in HeLa-GFP cells to 50.06% in HeLa-GDF15 cells. Consistently, a similar result was observed in SiHa cells; the percentage of cells in G0/G1 phase decreased from 70.95% in SiHa-GFP cells to 53.20% in SiHa-GDF15 cells, and the percentage of cells in S phase increased from 20.80% in SiHa-GFP cells to 35.80% in SiHa-GDF15 cells (Fig. 3c and d, P < 0.01). These results suggested that the overexpression of GDF15 accelerated the cell cycle transition from G0/G1 to S phase in cervical cancer cells. Conversely, the percentage of G0/G1 cells significantly increased and the percentage of S phase cells significantly decreased in HT-3-shGDF15 cells compared with HT-3-shNC cells (Fig. 3e and f, P < 0.01), which suggested that GDF15 silencing delayed the cell cycle transition from G0/G1 to S phase. Together, these results showed that GDF15 accelerated the cell cycle in cervical cancer cells at the transition from G0/G1 to S phase and suggested that GDF15 promoted cervical cancer cell proliferation through alterations in the cell cycle phases.
GDF15 altered the expression of cell cycle-related genes via FOXO1 and C-myc
FOXO1  and C-myc [15, 16] are two well-known classical transcription factors that regulate the cell cycle via their target genes, which include p21, CyclinD1, and CyclinE1, among others. Figure 4c to e shows the decreased expression of FOXO1 and the increased expression of C-myc in both HeLa-GDF15 and SiHa-GDF15 cells, compared with their respective controls (HeLa-GFP and SiHa-GFP). These results strongly suggested that GDF15 accelerated cell cycle progression in cervical cancer cells by targeting cdc25A, CDK2, CDK4, p21, CyclinD1 and CyclinE1 through alterations in the expression of FOXO1 and C-myc. Interestingly, pro-GDF15 expression was further elevated after treatment with conditioned medium from GDF15 overexpressing cells (HeLa-GDF15 and SiHa-GDF15, Fig. 4f and g) and treatment with rhGDF15 (Additional file 4: Figure S2a-d). We speculated that there was mechanism regulating the positive feedback of GDF15 expression. Two transcriptional binding elements (CAACTG at − 3421 and CACGTG at − 972) were detected in the promoter of GDF15 through sequence analysis in the UCSC Genome online database; additionally, the E-box motif (CANNTG) is regulated by the typical transcription factor C-myc. To confirm our suspicion, a quantitative ChIP (chromatin immunoprecipitation) assay was applied to demonstrate that C-myc recognized and bound the promoter to trans-activate the expression of GDF15 (Fig. 4h, p < 0.05). The positive feedback between GDF15 and C-myc is shown in Fig. 4i.
GDF15 induced the activation of AKT and ERK1/2 in human cervical cancer cells
GDF15 activated the PI3K/AKT and MAPK/ERK signaling pathways via ErbB2 phosphorylation
Correlation analysis between the expression of GDF15 and p-AKT1, p-Erk1/2, C-myc, and FOXO1 in cervical cancer specimens
GDF15 was first discovered independently by several different laboratories in the late 1990s and was found to be a regulator of macrophage activation and, a putative placental mediator of embryonic development. Recently, GDF15 was reported to have an important function in various types of carcinomas. Large-scale delineation of biomarkers from cancer samples showed elevated expression of GDF15 in the tissue and serum of patients with prostate, breast and colorectal carcinomas . Consistently, GDF15 was reported to be highly expressed in prostate cancer , malignant melanoma , ovarian cancer  and pancreatic ductal adenocarcinoma . GDF15 was thus considered a novel potential diagnostic and prognostic biomarker for the improved risk assessment of cancer progression. In the present study, GDF15 was found to be up-regulated between normal cervix and cervical cancer lesions, which suggests that GDF15 might play a role in cervical carcinogenesis. GDF15 expression was also found to be elevated in cervical cancer compared to adjacent normal tissue in a microarray-based study . Although the expression of GDF15 was higher in invasive squamous cervical cancer than in high-grade squamous intraepithelial lesions in the present study, no significant difference was observed between these two groups. This lack of a difference might be attributed to the limited samples used in the present study, and a large-scale study with more specimens is needed. Therefore, all of these results imply that GDF15 is involved in cervical carcinogenesis.
The present study showed that GDF15-overexpression in HeLa and SiHa cells enhanced tumor formation in vivo. Moreover, rhGDF15, treatment with conditional medium and GDF15-overexpression stimulated HeLa and SiHa cell proliferation in vitro. Furthermore, the down-regulation of GDF15 in HT-3 cells inhibited cell proliferation in vitro. Studies have shown that GDF15 promotes cell proliferation and tumor formation in ovarian cancer, pancreatic ductal adenocarcinoma and prostate cancer [29, 30, 31, 32]. All of these data are in accordance with each other and support the idea that GDF15 is a tumor promoter in several human cancers.
Cell cycle analysis showed that GDF15 overexpression accelerated the transition of cervical cancer cells from G0/G1 phase to S phase, whereas GDF15 down-regulation hindered the transition of cervical cancer cells from G0/G1 phase to S phase. These results suggest that GDF15 regulates the cell cycle through cell cycle checkpoint proteins. Among the cell cycle checkpoint proteins that we tested, the expression of cdc25A, CDK2, CDK4, CyclinD1 and CyclinE1 was up-regulated, and the expression of p21 was down-regulated in GDF15-overexpressing cervical cancer cells, which suggests that the cell cycle transition from G0/G1 phase to S phase in cervical cancer cells is regulated by GDF15 through the cdc25A/CDK4-CyclinD1 and p21/CDK2-CyclinE1 complexes. Consistently, Jin YJ et al. demonstrated that GDF15 promoted cell proliferation via the up-regulation of CyclinD1 and CyclinE1 and the down-regulation of p21 in human umbilical vein endothelial cells . In addition, our current study showed the down-regulated expression of FOXO1 and the up-regulated expression of C-myc in GDF15-overexpressing cervical cancer cells. These data indicate that alterations in cell cycle transition and the expression of cell cycle checkpoint proteins induced by GDF15 are mediated through FOXO1 and C-myc and innovatively demonstrate the positive feedback relationship of GDF15/C-myc/GDF15 in cervical cancer cells.
Activated AKT1 [34, 35, 36] and Erk1/2 [37, 38] phosphorylate FOXO1 and C-myc and thereby promote cell survival and proliferation. In the present study, the expression of PI3K, p-AKT1, p-Erk1/2 and Ras-GTP was significantly increased by rhGDF15 treatment or GDF15 overexpression but was decreased by GDF15 knock down, which indicates that GDF15 promotes cervical cancer cell proliferation through the PI3K/AKT and MAPK/ERK signaling pathways. Additionally, rhGDF15 and GDF15 overexpression obviously increased the expression of p-ErbB2, p-AKT1 and p-Erk1/2. Furthermore, ErbB2 inhibition significantly blocked the PI3K/AKT and MAPK/ERK signaling pathways, while inhibition of the TGF-β receptor had no effects on these two pathways. Kim KK et al. have reported that GDF15 activates AKT and ERK-1/2 via the transactivation of ErbB2 in human breast and gastric cancer cells in vitro . Similar results were reported by Joshi JP et al. in the breast cancer cells in vitro . In our study, for the first time, an IP assay showed that GDF15 bound to the p-ErbB2 receptor. However, the current study also revealed a suppressed proliferation of HeLa and SiHa cells with TGF-βR inhibitor treatment, which suggests that TGF-βR promotes cervical cancer cell proliferation through other mechanisms. These results suggest that GDF15 activates the PI3K/AKT and MAPK/ERK signaling pathways primarily through ErbB2.
A recent study found that GFRAL was the receptor for GDF15, and this ligand promoted weight loss in mice and nonhuman primates [39, 40, 41]. We also noted that nude mice injected with GDF15-overexpressing cervical cells had less food intake and lighter body weight than mice receiving control cells, although the data are not shown in this article.
This work was supported by a grant to Prof. Peng-Sheng Zheng from the National Natural Science Foundation of China (No.81472728 and 81672910).
Availability of data and materials
Data sharing not applicable to this article as no data-sets were generated or analyzed during the current study.
SL and YMM acquired of data, analyzed data, drafted the article and revised it critically for important intellectual content. PSZ and PZ: substantially contributed to conception and design, final approved of the vision to be published.
Ethics approval and consent to participate
All human samples were obtained at the First Affiliated Hospital of Xi’an Jiaotong University Medical College and the procedures followed medical-ethics approval practices. All animal experiments were approved by the Institutional Animal Care of Xi’an Jiaotong University Medical College.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 7.Cui N, Yang WT, Zheng PS. Slug inhibits the proliferation and tumor formation of human cervical cancer cells by up-regulating the p21/p27 proteins and down-regulating the activity of the Wnt/beta-catenin signaling pathway via the trans-suppression Akt1/p-Akt1 expression. Oncotarget. 2016;7:26152–67.PubMedPubMedCentralGoogle Scholar
- 23.Li C, Wang Q, Wang JF. Transforming growth factor-beta (TGF-beta) induces the expression of chondrogenesis-related genes through TGF-beta receptor II (TGFRII)-AKT-mTOR signaling in primary cultured mouse precartilaginous stem cells. Biochem Biophys Res Commun. 2014;450:646–51.CrossRefPubMedGoogle Scholar
- 33.Jin YJ, Lee JH, Kim YM, et al. Macrophage inhibitory cytokine-1 stimulates proliferation of human umbilical vein endothelial cells by up-regulating cyclins D1 and E through the PI3K/Akt-, ERK-, and JNK-dependent AP-1 and E2F activation signaling pathways. Cell Signal. 2012;24:1485–95.CrossRefPubMedGoogle Scholar
- 36.Xin B, Yamamoto M, Fujii K, et al. Critical role of Myc activation in mouse hepatocarcinogenesis induced by the activation of AKT and RAS pathways. Oncogene. 2017;36:5087-97Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. 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.