Gossypol has anti-cancer effects by dual-targeting MDM2 and VEGF in human breast cancer
- 1.5k Downloads
Mouse double minute 2 (MDM2) and vascular endothelial growth factor (VEGF) are important molecules involved in tumor progression. We researched potential inhibitors that simultaneously target MDM2 and VEGF. In our recent study involving the performance of high-throughput screening with a fluorescence polarization assay, gossypol was identified as one of the top hits that inhibit protein-RNA binding activity. Because MDM2 is an RNA-binding protein and its targets include VEGF mRNA, we investigated whether gossypol has an inhibitory effect on MDM2-VEGF.
UV cross-linking and RNA binding assay, isothermal titration calorimetry assay, and ubiquitination assay were performed to determine mechanisms by which gossypol functions as a dual inhibitor of MDM2 and VEGF. The effect of gossypol on MDM2 and VEGF expression, cancer cell apoptosis, tumor growth and VEGF-mediated angiogenesis were studied in vitro and in vivo in different human breast cancer models with a different p53 status.
We observed that gossypol inhibited expression of both MDM2 and VEGF in human breast cancer cells with either wild-type or mutant p53. A nechanistic study further demonstrated that, through disrupting the interaction between MDM2 protein and VEGF mRNA, gossypol induced MDM2 self-ubiquitination and decreased VEGF translation simultaneously, which resulted in both apoptosis and anti-angiogenesis effects. In vitro, regardless of p53 status, gossypol induced cancer cell apoptosis. In nude mouse xenograft in vivo models, gossypol suppressed tumor growth and VEGF-mediated angiogenesis.
Gossypol has anti-cancer effects by dual-targeting MDM2 and VEGF in human breast cancer. Our study reveals a novel mechanism by which gossypol functions as an anticancer agent. We believe that MDM2-VEGF targeting represents a novel strategy for improving cancer outcome.
KeywordsBreast neoplasms MDM2 VEGF Gossypol
Enzyme-linked immunosorbent assay
Isothermal titration calorimetry
Mouse double minute 2
Vascular endothelial growth factor
Mouse double minute 2 (MDM2) and vascular endothelial growth factor (VEGF) are important molecules involved in tumor progression. Increasingly there has been a broad consensus that they should be used as key targets for drug development. For example, selective small-molecule MDM2 antagonist nutlin-3a has been discovered and is under pre-clinical development . The recombinant humanized VEGF monoclonal antibody, bevacizumab, is recommended by the National Comprehensive Cancer Network (NCCN) guidelines for treatment of metastatic cancer .
MDM2 is a multifunctional oncoprotein that is overexpressed in a variety of human malignancies, including breast cancer [3, 4, 5, 6, 7, 8, 9]. Recently published studies, including ours, demonstrate that MDM2 is able to interact with specific RNA sequences or structures [10, 11]. The COOH-terminal RING finger domain of MDM2 protein was found to bind to the AU-rich sequence within the 3′untranslated region (3′UTR) of VEGF mRNA, to stabilize VEGF mRNA, and to increase its translation .
MDM2 and VEGF have been considered as potential cancer targets. Various MDM2 inhibitors have been found to have anticancer activity. The most promising of these is nutlin-3a. However, the majority of these inhibitors exert their effects through restoring the p53 apoptotic response . But actually, more than 50% of all solid tumors carry the p53 mutation , and there is a higher incidence of p53 mutation or loss of p53 observed in aggressive cancer . Moreover, activated p53 in turn induces MDM2 expression through the p53-MDM2 feedback loop. These compounds do not decrease expression of MDM2 and may even increase it. In light of the p53-independent mechanism through which MDM2 regulates VEGF-mediated tumor angiogenesis, no significant VEGF inhibition or anti-angiogenesis activity has been observed after treatment with nutlin-3a .
Bevacizumab is the first Food and Drug Agency (FDA)-approved anti-angiogenic drug for clinical use. There is preliminary evidence of its efficacy when combined with cytotoxic agents. However, even low-dose bevacizumab is associated with increased risk of venous thromboembolism and some patients with a clinical response to VEGF blockade ultimately develop progressive disease . Therefore, numerous studies remain committed to testing novel anti-cancer drugs targeting MDM2 and VEGF, or both simultaneously.
Gossypol, a polyphenol derived from cotton seeds, exhibits potent anti-cancer activities and has completed phase II clinical trials for treatment of human prostate cancer, with promising initial results . However, information on its mechanism is still limited [18, 19, 20, 21]. In our recent study involving the performance of a high-throughput screening (HTS) with fluorescence polarization (FP) assay in four chemical libraries containing a total 141,394 known reagents, drugs, and small-molecule compounds, gossypol was identified as one of the top hits that inhibit protein-RNA binding activity . Because MDM2 is an RNA-binding protein and its targets include VEGF mRNA, we investigated whether gossypol has a similar inhibitory effect on MDM2-VEGF. We found that gossypol is a potent inhibitor of MDM2-VEGF. Through disrupting the interaction between MDM2 protein and VEGF mRNA, gossypol induces MDM2 self-ubiquitination and decreases VEGF translation simultaneously, which results in not only cancer cell apoptosis but also suppression of tumor angiogenesis. Our study helps to elucidate the mechanism by which gossypol functions as an anticancer agent. We believe that MDM2-VEGF targeting represents a novel strategy for improving cancer outcome.
Cell lines and reagent
Five human breast cancer cell lines (MCF-7, ZR-75-1, MDA-MB-231, MDA-MB-468 and T47D) were used in this study. The first two have wild-type (wt) p53, while the remaining three are p53-mutant. All five cell lines were obtained from the China Center for Type Culture Collection (CCTCC). The gossypol was purchased from Sigma-Aldrich (St. Louis, MO, USA), and stock solution was prepared at 10 mM in dimethyl sulfoxide (DMSO). In vitro biological activity of gossypol was determined after incubation at various concentrations (2.5, 5.0, 7.5, and 10.0 μM) for different periods of time (2, 4, 8, and 24 h). Equivalent concentrations of DMSO were used as vehicle controls.
Plasmids and gene transfection
MDA-MB-468 cells with MDM2 knockdown or overexpression were described previously . To determine whether gossypol-induced ubiquitination of MDM2 required the intrinsic self-ubiquitination E3 ligase activity of MDM2 itself, a QuickChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA) was used to mutate MDM2 464; this leads to cysteine being substituted by alanine to generate the plasmid pCMV-MDM2 C464A, which exhibits loss of E3 ligase activity. MDM2 gene promoters 1 and 2 were cloned into the pGL3-Basic vector to generate the MDM2 p1 and p2-Luciferase plasmids. MDA-MB-468 cells were transiently transfected with these MDM2 promoter luciferase plasmids, treated with gossypol and then prepared for testing the luciferase activity of MDM2 promoter reporters using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) according to the manufacturer’s instructions.
Quantitative RT-PCR was used to measure the expressions of MDM2 and VEGF at mRNA level. The primer sequences were as follows: MDM2 forward 5′-TGTTGGTGCACAAAAAGACACTT-3′, MDM2 reverse 5′-GCACGCCAAACAAATCTCCTA-3′; VEGF forward 5′-GATGAAAGGCGGCATACGG-3′, VEGF reverse 5′-CAGGGCTATTCTTCTTAGTGTGC-3′.
The protein expression levels of MDM2, p53, VEGF, caspase-3, cleaved caspase-3, and poly(ADP-ribose) polymerase (PARP) were analyzed by western blot. The following antibodies were used: anti-MDM2 antibody (1:1000; Sigma), anti-p53 antibody (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-VEGF antibody (1:200; Santa Cruz Biotechnology), anti-caspase-3 antibody (1:1000; Santa Cruz Biotechnology), anti-cleaved caspase-3 p11 antibody (1:1000; Santa Cruz Biotechnology), and anti-PARP antibody (1:1000; Cell Signaling, Danvers, MA, USA).
In vivo protein-RNA binding assay
To detect in vivo binding between MDM2 protein and VEGF mRNA, VEGF mRNA was co-immunoprecipitated with MDM2 antibody from whole-cell extracts by using a previously described method . The resulting RNA was analyzed by quantitative RT-PCR using VEGF-specific primers.
UV crosslinking and RNA binding assay
The DNA template for synthesis of the VEGF 3′UTR probe was generated using the following sequence: 5′-TAATACGAGTCACTATAGGGAAATTCTACATACTAAATCTCTCTCCTTTTTTAATTTTAATATTTG-3′. This DNA fragment incorporated the T7 promoter sequence (underlined). Internally labeled RNA probes were synthesized by in vitro transcription with T7 polymerase (MAXIScript T7 RNA polymerase kit, Ambion) in the presence of [α-32P] UTP (Amersham). GST-MDM2 protein was mixed with 32P-labeled probes in the presence or absence of gossypol. Then UV crosslinking of the RNA-protein complexes formed was done using a 254-nm UV light source set at 400,000 μJ/cm2. Finally, the UV-irradiated RNA-protein complexes were treated with RNase T1, resolved by 10% SDS-PAGE gel and visualized by autoradiography.
Isothermal titration calorimetry (ITC) assay
The binding activity of gossypol to either MDM2 RING protein or VEGF 3′UTR was examined by ITC assay. The ITC assay was performed using the auto-iTC200 instrument (MicroCal, GE). MDM2 protein or VEGF 3′UTR was loaded into a 96-well deep-well PP plate, and then the compound was titrated stepwise into the protein or RNA sample cell using a syringe, for a total of 16 injections. The equilibrium time between two adjacent injections was 210 s. The binding stoichiometry (n), binding constant (Kd) and thermodynamic parameters (ΔH and ΔS) were determined by fitting the titration curve to a one-site binding mode, using the Origin software provided by the manufacturer.
Polysome preparation and analysis
Polysome profiling was carried out as described previously . Cells were incubated with 100 μg/ml cycloheximide (CHX) for 15 minutes to arrest polyribosome migration. The cells were then lysed to isolate cytoplasmic RNA in a buffer containing 20 mM Tris–HCl at pH 8.0, 100 mM NaCl, 5 mM MgCl2, 0.5% Triton X-100, 500 U/ml RNAsin and a cocktail of protease inhibitors. The cell lysates were fractionated on a 15–45% (wt/vol) sucrose gradient and centrifuged in a SW41Ti rotor at 39,000 rpm for 2 h. Fractions from each gradient tube were collected by upward replacement, and absorption at optical density (OD)254 was monitored using a fractionator. The RNA in each fraction was extracted and subjected to quantitative RT-PCR.
Cells were treated with MG132 (30 μM) for 4 h before collection. The cell pellet was lysed and incubated with Ni2+-NTA beads (Qiagen, Valencia, CA, USA) at room temperature for 4 h. The beads were then washed sequentially for 5 minutes in each of the following buffers: buffer A (6 M guanidinium-HCl, 0.1 M Na2HPO4/NaH2PO4, 0.01 M Tris–HCl (pH 8.0), 5 mM imidazole, 10 mM β-mercaptoethanol); buffer B (8 M urea, 0.1 M Na2HPO4/NaH2PO4, 0.01 M Tris–HCl (pH 8.0), 10 mM β-mercaptoethanol); buffer C (8 M urea, 0.1 M Na2HPO4/NaH2PO4, 0.01 M Tris–HCl (pH 6.3), 10 mM β-mercaptoethanol). This ubiquitinated product was finally eluted with buffer D (200 mM imidazole, 0.15 M Tris–HCl (pH 6.7), 30% glycerol, 0.72 M β-mercaptoethanol, 5% SDS) and analyzed by western blot using an anti-ubiquitin antibody (Cell Signaling).
WST-1 cytotoxicity assay
The cytotoxic effect of gossypol was determined by WST-1 assay. Cells were cultured in 96-well plates along with different concentrations of gossypol for 24 h. WST-1 reagent (25 μg/well; Roche) was then added and incubation continued for an additional 4 h. Following this, the OD of the wells was read with a microplate reader (BioTek Instruments, Winooski, VT, USA). Appropriate controls lacking cells were included to determine background absorbance.
Flow cytometry was conducted to analyze the ability of gossypol to induce apoptosis. Cells with or without gossypol treatment were stained with fluorescein isothiocyanate (FITC)-Annexin V and propidium iodide (BD Pharmingen, San Diego, CA, USA). The stained cells were then detected on a FACScan flow cytometer using the CellQuest software (BD Biosciences, Mountain View, CA, USA).
This study was approved by the Animal Care and Use Committee of Huazhong University of Science and Technology. Breast cancer xenograft models were established. After 7 days, tumors were detected and gossypol (dissolved in DMSO) was administrated intraperitoneally at a dose of 10 mg/kg/day for 4 weeks. Control group received control solution containing the same amount of DMSO. Tumor growth and angiogenesis, as determined by microvessel density (MVD), were evaluated for each of these xenograft tumors.
At least three independent experiments were performed. All data were presented as mean ± SEM. Relative gene expression data were analyzed using the 2 − ΔΔCT method . The statistical significance of differences between groups was determined by one-way analysis of variance (ANOVA) or by the nonparametric Kruskal-Wallis test with the use of SPSS 13.0 software. All statistical tests were two-sided. The significance level was set at P < 0.05.
Gossypol binds to MDM2 RING protein to disrupt its interaction with VEGF 3′UTR, inhibiting the expression of both MDM2 and VEGF
We then investigated the effect of gossypol on MDM2 and VEGF expression. We found that, regardless of the p53 status, gossypol significantly inhibited the cellular expression of both MDM2 and VEGF. Further, gossypol inhibited expression of both MDM2 and VEGF in a dose-dependent and time-dependent manner (Fig. 1c). As controls, other compounds (labeled MX2, MX3, MX25, and MX28) identified as the top hits that inhibit protein-RNA binding activity  did not simultaneously inhibit the expression of both MDM2 and VEGF (Fig. 1d).
In addition, the ITC assay was performed to determine whether gossypol binds to either MDM2 RING protein or to VEGF 3′UTR to disrupt their interaction. The results showed that gossypol bound to the MDM2 RING protein with a binding Kd value of 5.21 μM (Fig. 1e), but not to the VEGF 3′UTR (Fig. 1f). Therefore, gossypol binds to MDM2 RING protein to disrupt its interaction with VEGF 3′UTR and consequently inhibit the expression of both MDM2 and VEGF.
Gossypol induces MDM2 self-ubiquitination and protein-degradation
It is well-known that MDM2 is ubiquitinated through its own E3 ubiquitin ligase activity . To further define the mechanism by which gossypol promotes MDM2 protein degradation, the possibility that gossypol could induce MDM2 self-ubiquitination was studied. As expected, gossypol indeed induced ubiquitination of endogenous MDM2, which was associated with VEGF 3′UTR (Fig. 2f, g). This ubiquitination required the intrinsic self-ubiquitination E3 ligase activity of MDM2 itself. While gossypol induced degradation and ubiquitination of wt MDM2, it was unable to induce degradation and ubiquitination of the C464A mutant MDM2 that lacks E3 ubiquitin ligase activity  (Fig. 2h, i). Therefore, gossypol inhibits MDM2 expression through induction of MDM2 self-ubiquitination and protein-degradation.
Gossypol decreases VEGF mRNA stability and thereby its protein translation
Because of decreased mRNA stability, we took steps to investigate whether there also was a decrease in VEGF protein synthesis. We performed linear sucrose-gradient fractionation to assess the association between polyribosomes and VEGF mRNA in MDA-MB-468 cells subjected to gossypol treatment (Fig. 3c). We found that VEGF mRNA shifted away from fractions enriched with translating polyribosomes to fractions containing translation-dormant complexes, which is indicative of decreased translation. In contrast, gossypol had no effect on the polyribosome profile of a control glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA.
We also investigated whether inhibition of VEGF is associated with MDM2. Gossypol inhibited VEGF in control MDA-MB-468 cells but not in MDA-MB-468 cells with MDM2 knockdown (Fig. 3d), suggesting that gossypol-mediated inhibition of VEGF is MDM2-dependent.
Gossypol induces cancer cell apoptosis, dependent on MDM2 inhibition
To clarify whether the observed cell death was associated with induction of apoptosis, gossypol-treated cancer cells were stained with Annexin V and quantitated by flow cytometry. We found that gossypol-induced cell death was indeed due to apoptosis. Gossypol induced apoptosis in both wt and mutant p53-expressing cancer cells. As the concentration of gossypol increased, the percentage of apoptotic cells gradually increased. Correspondingly, cleavage of caspase-3 and PARP was detected by western blotting after gossypol treatment (Fig. 4b).
We ask how critical oncoprotein MDM2 is to the gossypol-induced cell apoptosis and death. Results as shown in Fig. 4c indicated the importance of MDM2 inhibition. MDM2 siRNA or overexpression plasmid was introduced into MDA-MB-468 cells. Cells with MDM2 knockdown or overexpression were less responsive to gossypol-induced apoptotic effect than control cells.
Gossypol suppresses tumor growth and VEGF-mediated angiogenesis in breast cancer xenograft models
Upon processing tumor tissues, we found that gossypol inhibited in vivo expression of MDM2 (Fig. 5d). In addition, serum levels of VEGF in nude mice were measured by ELISA (Fig. 5e), providing further evidence for the simultaneous inhibition of VEGF production by gossypol. These results were consistent with those observed in vitro in the breast cancer cell lines treated with gossypol. Finally, the effect of gossypol on tumor angiogenesis was evaluated by immunostaining for the endothelial cell marker CD31. Positive staining was digitally quantified and measured as MVD. Anti-angiogenesis activity occurred following administration of gossypol. Compared to control, there was a significant decrease in CD31 staining (Fig. 5f). Accordingly, there was significant suppression of MVD by 46.5% in the MCF-7 xenograft model and by 51.2% in the MDA-MB-468 xenograft model, respectively (Fig. 5g).
There is evidence of the anti-cancer effects of gossypol , and several studies have investigated the possible molecular targets and mechanisms of action. For example, gossypol has been identified as a Bcl-2 homology 3 (BH3)-mimetic, is able to bind to the BH3 domain of Bcl-2 family members, and induces apoptosis . In addition, gossypol modulates the p53 , c-Myc  and NF-κB  signaling pathways. In the present study, we reported that gossypol is a potent inhibitor of MDM2-VEGF. Through disrupting the molecular interaction between MDM2 protein and VEGF mRNA, gossypol induces MDM2 self-ubiquitination and decreases VEGF translation, which results in both apoptosis and anti-angiogenesis effects.
Our study revealed a novel mechanism by which gossypol functions as an anticancer agent. MDM2 and VEGF are important molecules involved in tumor progression. MDM2 acts as an oncoprotein, while VEGF plays crucial roles in tumor angiogenesis. For years, our research has focused on the p53-independent activity and mechanism of MDM2 in the regulation of VEGF . We also try to screen for potential inhibitors that simultaneously target MDM2 and VEGF. Although cytotoxic drugs kill tumor cells, the plentiful blood supply to the tumor can still allow residual tumor cells to continue growing. Therefore, we believe that MDM2-VEGF dual-targeting represents a novel cancer therapy strategy, not only against cancer cells, but also against the tumor microenvironment, especially tumor angiogenesis. This will offer greater control on tumor growth.
For the first time, we described the inhibition of gossypol on MDM2. Compared with other classical MDM2 inhibitors like nutlin-3a, gossypol is distinct in that it inhibits MDM2 and induces apoptosis in cancer cells with a different p53 background. As mentioned, MDM2-p53 interaction inhibitors have several limitations. First, these inhibitors exhibit effects only in cancer cells bearing wild-type p53. Second, these inhibitors do not inhibit MDM2, and expression of MDM2 can even be enhanced through the p53-MDM2 feedback regulatory loop. Third, these inhibitors would not be able to inhibit the p53-independent activities of MDM2. Therefore, gossypol should be particularly useful, especially for high-risk, refractory cancer patients. Studies have shown that most cancer patients, including breast cancer patients, do not express wild-type p53 and mutations in p53 predict worse prognosis and poor treatment outcome [27, 28].
Herein, a mechanistic study indicated that gossypol induces MDM2 self-ubiquitination and protein degradation. MDM2 has been well-characterized as a member of the RING-finger-type family of E3 ubiquitin ligases. MDM2 regulates ubiquitination of not only p53 but also MDM2 itself . The E3 ligase activity responsible for MDM2 self-ubiquitination is regulated by many cellular signaling pathways and molecular events. Previous studies showed that nucleic acids, such as polyA, polyG and certain cellular small RNA, are able to bind to the RING domain of MDM2 and suppress its self-ubiquitination activity [29, 30]. In the current study, the dissociation between MDM2 protein and VEGF mRNA following treatment with gossypol led not only to decreased translation of VEGF mRNA but also to simultaneous self-ubiquitination of MDM2 protein. These results provide a preliminary indication that binding of VEGF mRNA to MDM2 protein may also suppress its self-ubiquitination activity, which results in increased MDM2 protein stability and expression. We have shown that, in response to hypoxia, MDM2 protein is translocated from the nucleus to the cytoplasm. In the cytoplasm, MDM2 binds to VEGF mRNA and increases VEGF mRNA stability and translation . On the other hand, further detailed studies are required to give direct evidence that MDM2 and VEGF are mutually regulated, and the E3 ubiquitin-ligase activity of the MDM2 RING protein for self-ubiquitination is negatively regulated by the binding of the VEGF mRNA.
VEGF is the most prominent angiogenic factor that plays crucial roles in tumor angiogenesis. Gossypol has previously been reported to modulate the VEGF signaling pathway . Gossypol inhibits VEGF expression in human prostate cancer cells. Further, gossypol blocks multiple steps in VEGF-activated biological events in endothelial cells. For example, gossypol inhibits endothelial cell viability, chemotactic motility, and microvessel sprouting. Gossypol exerts its function through blockade of the VEGF/VEGFR2 signal cascade in vascular endothelial cells. In agreement, in this study, we also observed that there was significant inhibition of VEGF in human breast cancer following treatment with gossypol. And besides that, importantly, we illuminated the mechanism by which gossypol inhibits production of VEGF by cancer cells. Gossypol decreases VEGF mRNA stability and thereby its protein translation. Our results help to better understand the functional property of gossypol in the VEGF signaling pathway.
In summary, gossypol is a dual inhibitor of MDM2 and VEGF that disrupts the molecular interaction between MDM2 protein and VEGF mRNA, induces MDM2 self-ubiquitination and degradation, decreases VEGF mRNA stability and protein translation simultaneously, and therefore exerts anti-cancer effects through apoptotic and anti-angiogenesis pathways in human breast cancer in vitro and in vivo, regardless of the p53 status of the cancer cells. We believe development of these MDM2-VEGF inhibitors as potential anticancer drugs for clinical use is worthwhile and represents a novel strategy for improving cancer outcome.
We thank Prof. Muxiang Zhou (Division of Hematology/Oncology, Aflac Cancer Center and Blood Disorders Service, Emory University School of Medicine, Atlanta, GA, USA) for helpful discussions and comments on the manuscript.
This study was supported by National Nature Science Foundation of China (Grant numbers 81472488 and 81502296).
Availability of data and materials
The datasets supporting the conclusions of this article are included within the article.
SZ and TFS were the principal investigators who conceived, coordinated and oversaw the study and wrote the manuscript. JX and JSL participated in study design and performed the RNA binding assay, ITC assay, and ubiquitination assay. QY and JW performed the WST-1 cytotoxic assay, flow cytometry, and animal studies. SZ and XJ analyzed the data and prepared the figures. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
All animal studies were approved by and conducted in accordance with the Animal Care and Use Committee of Huazhong University of Science and Technology.
Availability of supporting data
- 2.Schneider BP, Wang M, Radovich M, Sledge GW, Badve S, Thor A, Flockhart DA, Hancock B, Davidson N, Gralow J, Dickler M, Perez EA, Cobleigh M, Shenkier T, Edgerton S, Miller KD, ECOG 2100. Association of vascular endothelial growth factor and vascular endothelial growth factor receptor-2 genetic polymorphisms with outcome in a trial of paclitaxel compared with paclitaxel plus bevacizumab in advanced breast cancer: ECOG 2100. J Clin Oncol. 2008;26(28):4672–8.CrossRefPubMedPubMedCentralGoogle Scholar
- 28.Bourdon JC, Khoury MP, Diot A, Baker L, Fernandes K, Aoubala M, Quinlan P, Purdie CA, Jordan LB, Prats AC, Lane DP, Thompson AM. p53 mutant breast cancer patients expressing p53γ have as good a prognosis as wild-type p53 breast cancer patients. Breast Cancer Res. 2011;13(1):R7.CrossRefPubMedPubMedCentralGoogle 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.