Polygonum Barbatum Extract Reduces Colorectal Cancer Cell Proliferation, Migration, Invasion, and Epithelial-mesenchymal Transition via Regulation of the YAP and β-catenin Pathways

Colorectal cancer (CRC) is the third most common cancer worldwide, but the development of novel therapeutics for CRC remains a challenge. Polygonum barbatum has anticancer potential, but its mechanism of action requires further investigation. This study was designed to investigate the inhibitory effect of Polygonum barbatum on human CRC cells. The HPLC ngerprints of the Polygonum barbatum extract (PBE) and quercetin standard were determined using analytical RP-HPLC and evaluations were completed using the human colon cancer cell line HCT-116 (KRASG13D mutation) and HT-29 cells. After treatment with PBE, cell viability, colony formation, migration, invasion, and apoptosis were analyzed using CCK-8, colony formation, wound healing, Transwell invasion, and ow cytometry assays, respectively. RNA-sequencing, western blotting, and co-immunoprecipitation were also used to analyze changes in the whole-transcriptome of these cells and identify possible mechanisms of action for PBE in CRC cells.


Background
Colorectal cancer (CRC) is the third most commonly diagnosed cancer in the world, and both its incidence and mortality rates are increasing in Asia [1]. Treatments for unresectable metastatic CRC are designed to facilitate tumor shrinkage and control metastatic lesions, and a combination of targeted therapy and cytotoxic chemotherapies is commonly applied as the primary treatment for metastatic CRC. These treatments have resulted in a signi cant improvement in the median overall survival, from 12 to 30 months, over the last two decades [2]. Most patients experience some initial response to treatment, but many experience some degree of drug resistance over time, reducing e cacy. In addition, the high degree of toxicity associated with the chemotherapy options for CRC also limit their long term application. This means that there is still an urgent need to develop novel therapeutic agents for CRC.
The Hippo pathway, including core kinase complexes MST1/2 and LATS1/2 and downstream effectors Yes-associated protein (YAP) and transcriptional coactivator with PDZ-binding motif (TAZ), regulates cell growth and differentiation acting as a tumor suppressor pathway [3]. Once activated, the Hippo pathway suppresses the nuclear translocation of YAP, which acts as an oncogene in most settings. The Hippo pathway is downregulated in a variety of cancer cells where YAP is known to be activate. YAP promotes the expression of various target genes, including connective tissue growth factor (CTGF) and cysteinerich angiogenic inducer 61 (CYR61), which are associated with mesenchymal differentiation [4] and poor prognosis in CRC patients [5]. In addition, the Wnt signaling pathway regulates cell growth, epithelialmesenchymal transition (EMT), and self-renewal, and aberrant WNT signaling has been associated with progression in CRC tissues [6,7]. This makes these pathways ideal targets for new CRC therapies.
Polygonum barbatum, a perennial herb belonging to the Polygonaceae family, is widely distributed across Southeast Asia and generally grows in marshy ground near riversides and other aquatic environments [8]. Polygonum barbatum is known to possess antimicrobial activity [9] and its bioactive compounds have demonstrated anti-proliferative activity against non-small cell lung carcinoma (NCI-H640), breast cancer (MCF-7), and cervical cancer (HeLa) cells [10]. However, the effect of Polygonum barbatum treatment on CRC cells has not been described, and mechanistic insights into its action remain scarce. This study was designed to clarify the effects and underlying mechanism of Polygonum barbatum extracts (PBE) on CRC cells.

Methods
Instrumentation and analytical conditions HPLC analysis of PBE and quercetin was performed on a Hitachi L-7100 instrument with a UV detector (L-2400), pump (L-2130), and autosampler (L-2200). Reverse-phase separation of the marker compound was performed using a Lichro CART® RP-18e (4.0 × 250 mm i.d., 5 μm) column and a gradient elution was achieved using two solvents, namely water (A) and methanol (B), at a ow rate of 1 mL/min. The gradient program consisted of an initial linear increase from 30% B to 60% B over 15 min, followed by an increase to 80% B over 15 min. This was then maintained for 10 min, and then returned to the initial gradient condition over the next 10 min before being maintained for another 5 min. The injection volume was 10 μL and the UV absorption spectra were recorded online at 370 nm, then the data were processed using the Hitachi Model D-2000 Elite Chromatography Data Station Software.
Cell culture and viability assays Human colon cancer cell lines HCT-116 (KRAS G13D mutation) and HT-29 were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in DMEM with 10% fetal bovine serum (FBS) at 37 °C. Cell viability was assayed using a Cell Counting kit-8 according to the manufacturer's instructions (CCK-8, Dojindo, Japan) [11].

Colony formation assay
This assay was completed as previously described [12]. Brie y, cells were incubated for 21 days at 37 °C using 0.5% agar in DMEM containing 10% FBS, 1 mM glutamine, 100 units penicillin, and 100 μg/mL streptomycin before the colonies were counted and quanti ed using ImageJ software version 1.50a (National Institutes of Health, Bethesda, MD, USA).
The Transwell inserts were coated with Matrigel (BD Biosciences, Bedford, MA, USA) prior to the invasion assay, but not the migration assay, and cells were seeded in the upper chamber and incubated in serumfree RPMI-1640 medium while the lower chamber was lled with medium supplemented with 10% FBS.
Cells were added to the upper chamber and incubated for 24 h before being xed in methanol and stained with crystal violet for 15 min. Cells at the bottom of the inserts were then counted using an inverted microscope.

Scratch assay
These assays were completed as previously described [14]. Brie y, HCT-116 and HT-29 cells were scratched using a 100 μL pipette tip, then washed with PBS and incubated with vehicle or PBE. Wound healing was imaged using photomicrography at various time points (Leica Microsystems, Wetzlar, Germany).

Flow cytometry
Cellular apoptosis was analyzed using an Annexin-V/7-AAD staining kit according to the manufacturer's instructions (BioVision, Inc., CA, USA). Cells were treated with vehicle or PBE and then stained with Annexin-V/7-AAD solution and analyzed using a ow cytometer (BD Biosciences, FACS CaliburTM) [12].

Nuclear extracts and western blotting
These assays were completed as previously described [15]. The details for the primary and secondary antibodies used in these assays are summarized in Table 1. The blots were developed using an enhanced chemiluminescence kit (Amersham Biosciences, Buckinghamshire, UK) and measured using a luminescent image analyzer (LAS-3000; Fuji Photo Film Co., Ltd., Tokyo, Japan).  [16]. Transcript abundance was quanti ed using RSEM (version 1.2.28) [17] and differentially expressed genes (DEGs) were identi ed using EBSeq (version 1. 16.0) [18].
Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were used to evaluate the gene clusters identi ed by the clusterPro ler program in R (version 3.6.0) [19].

Co-immunoprecipitation assays
These assays were performed as previously described [20]. Brie y, cells were lysed and incubated with anti-YAP beads for 4 h at 4°C, then washed before the immunocomplexes were suspended in SDS sample buffer and subjected to western blotting. The inputs were loaded with total cell lysates for comparison.

Statistical Analysis
All data are presented as the mean ± standard error of the mean (SEM) and the differences between groups were evaluated using one-way analysis of variance (ANOVA) and a Bonferroni post-hoc test. Statistical analyses were performed using IBM SPSS Statistics version 22 (IBM® SPSS® Statistics 22) and signi cance was accepted when the p values were less than 0.05.

HPLC analysis of PBE
The retention time of quercetin was 17.13 min (Fig. 1A) and we produced a clear HPLC ngerprint for PBE (Fig. 1B).

PBE signi cantly inhibits cell growth and induces apoptosis in CRC cells
Annexin V/7-AAD staining showed that PBE treatment significantly increased the number of apoptotic Differentially expressed mRNAs regulated by PBE RNA sequencing of PBE-treated HCT-116 cells was used to identify the potential mechanism of action for this compound in CRC. The quality and read data for this sequencing experiment are summarized in Table 2. We identi ed 167 and 330 DEGs in cells treated with 30 μg/mL PBE (117 downregulated and 50 upregulated) and 100 μg/mL PBE (278 downregulated and 52 upregulated), respectively (Fig. 3A). DEG scatterplots visualize the differences between the vehicle control and 30 μg/mL PBE ( Figure 3B), 100 μg/mL PBE ( Figure 3C), or 5-FU ( Figure 3D), respectively.

KEGG Pathway and GO Enrichment Analysis
KEGG and GO analysis were then used to further clarify the regulatory signaling pathway involved in PBE mediated inhibition of CRC cell proliferation, and the top 10 KEGG pathways for the DEGs in each group are shown in Table 3-5. We found that extracellular matrix (ECM)-receptor interactions and focal adhesion (FA) were the most signi cantly enriched pathways in PBE-treated cells ( Figure 4A and 4B) and that these pathways included thrombospondin 1 (THBS1), glycoprotein Ib platelet subunit beta (GP1BB), laminin subunit alpha 5 (LAMA5), AGRN, integrin subunit beta 8 (ITGB8), tenascin XB (TNXB), heparan sulfate proteoglycan 2 (HSPG2), Fraser extracellular matrix complex subunit 1 (FRAS1), integrin subunit alpha 2 (ITGA2), bronectin 1 (FN1), laminin subunit beta 2 (LAMB2), collagen type IV alpha 5 (COL4A5), integrin subunit alpha V (ITGAV), and SHC adaptor protein 3 (SHC3). In addition, GO enrichment analysis of the DEGs from the PBE-treated groups displayed signi cant enrichment for cellular proliferation and extracellular matrix organization ( Figure 4C and 4D). Whole blood vessel morphogenesis was shown to be signi cantly enriched in the 5-FU group ( Figure 4E). When we evaluated the GO terms associated with molecular function we found that both integrin binding and extracellular binding were signi cantly enriched in the PBE-treated cells ( Figure 4F and 4G), while RNA polymerase II core promoter proximal region sequence-speci c DNA binding was enriched in the 5-FU-treated cells ( Figure 4H). In addition, both endocytic vehicle lumen and extracellular cellular components from the cellular component category were signi cantly enriched in the PBE-treated cells ( Figure 4I and 4J), while adheren junctions was the most signi cantly enriched term in the 5-FU-treated cells ( Figure 4K).

PBE signi cantly inhibits cell migration, invasion, and epithelial-to-mesenchymal transition (EMT) in CRC cells
Wound healing, cell migration, and invasion assays were used to elucidate the effects of PBE on CRC cells. We found that PBE signi cantly inhibited the wound healing rate in both HCT-116 (Figs. 5A and 5 B) and HT-29 (Figs. 5C and 5D) cells. Similar inhibitory effects were observed in the migration assays ( Figure 5E and 5F) and we found that the invasion rates of PBE-treated cells were signi cantly lower than those of the vehicle control ( Figure 5G and 5H). In addition, we went on to evaluate the EMT-associated markers using western blotting. We found that the levels of epithelial markers ZO-1 and E-cadherin were increased in PBE-treated cells when compared with the control in both HCT-116 (Figs 5I, 5J, and 5 K) and HT-29 cells (Figs 5P, 5Q,  We then used western blot to explore the molecular mechanisms underlying the effects of PBE in CRC cells. We found that PBE (100 μg/mL) signi cantly increased the S127 phosphorylation of YAP in both cell lines (Figs. 6A-6D). In addition, the levels of p-YAP increased in a dose-dependent manner following the addition of PBE (Figs. 6E-6G). This was then con rmed by evaluating both cytoplasmic and nuclear extracts in more detail. PBE treatment signi cantly reduced the nuclear translocation of YAP, but did not change the cytosolic retention of this protein (Figs. 6I-6N). Meanwhile, PBE treatment reduced the expression of YAP target genes CTGF and CYR61 (Figs. 6O-6R). These results reveal that PBE treatment signi cantly suppresses YAP signaling. In addition, co-immunoprecipitation revealed that the interaction between glycogen synthase kinase 3β (GSK3β), β-catenin, and YAP increased in PBE-treated cells (Fig.  6S). This implies that PBE also regulates the WNT signaling pathway.

PBE suppresses the GSK3β/β-catenin signaling pathway
We then went on to con rm the role of PBE in the regulation of WNT signaling by evaluating the GSK3β/ β-catenin signaling pathway. As expected, PBE signi cantly reduced the phosphorylation of GSK3β at Ser9 in both a time (Figs. 7A-7D) and dose dependent manner (Figs. 7E-7H) in both cell lines. Consistently, the nuclear and cytosolic protein levels of β-catenin were also shown to be modulated by PBE treatment (Figs. 7I-7N). We further investigated the downstream targets of the Wnt pathway and demonstrated that PBE treatment signi cantly increased the phosphorylation of β-catenin and its targets, including cyclin D1, c-Myc, and c-Jun, which were downregulated in both cell lines (Figs. 7O-7X). These ndings indicate that PBE inhibits the WNT/β-catenin signaling pathway in CRC cells.

Discussion
PBE signi cantly reduced motility and tumorigenic potential by modulating EMT in both HCT116 and HT29 CRC cells. PBE treatment triggered cellular apoptosis, as established by an annexin V-FITC and 7-AAD double stain assay. Both GO and KEGG analysis of the RNA sequencing data was consistent with these ndings, with this analysis demonstrating that the expression of the genes responsible for extracellular matrix organization (CTGF and CYR61), cell motility (THBS1 and CXCL8), and cell growth (PLXNB1 and FN1) were all inhibited by PBE. These effects are associated with the blockage of both YAP and Wnt signaling (Fig. 8).
Polygonum barbatum has been reported to produce potential anticancer bioactive compounds such as dihydrobenzofuran, sesquiterpene derivatives [10,21], and quercetin [22]. Quercetin has been shown to activate the Hippo pathway and inhibit YAP signaling [23], and sesquiterpene derivatives have been reported to induce ROS-and TRAIL-mediated apoptosis, enhance chemotherapy responses, and inhibit EMT with these effects being accompanied by the downregulation of β-catenin in CRC cells [24]. To the best of our knowledge, this is the rst study to show that Polygonum barbatum can reduce the migration and tumorigenic potential of CRC cells by blocking both YAP and β-catenin signaling. Based on these ndings, these bioactive compounds are thought to synergistically contribute to the anti-CRC properties of PBE. However, most of these ndings are based on in vitro studies, which means that further in vivo evaluations and clinical trials are required to make any de nitive statements on their activity. Interestingly, 2,3-dihydrobenzofuran derivatives have been shown to exhibit microsomal prostaglandin E2 synthase-1 inhibitor activity [25], a key enzyme in prostaglandin E2 (PGE 2 ) synthesis known to boost CRC immune evasion [26]. Therefore, the regulatory effect of PBE on the COX-2/PGE 2 axis, and its mediators such as Janus kinase 2/signal transducer and activator of transcription 3 in CRC, will be investigated in our future work.
The ECM regulates cellular behavior and participates in both cellular adhesion and migration, with the overexpression of ITGAV, ITGA1, ITGB8, and FN genes known to be involved in CRC growth and metastasis [27][28][29] [30]. In addition, collagen XII, FRAS1, LAMA5, and THBS1 are all associated with colorectal liver metastasis, which is the most common distant metastasis in CRC patients [31] [32] [33]. In this study, our GO and KEGG analyses of the RNA sequencing data revealed that both the ECM and FA pathways were signi cantly downregulated in response to PBE. When combined with the results from the apoptosis evaluations we can con rm that these outcomes are consistent with the nding that diminished expression of ITGA2 promotes death and apoptosis in CRC cells [34]. Interestingly, we found that GP1BB, a regulator of epithelial cell adhesion, was upregulated in response to PBE treatment. This regulator has been reported to increase cell-cell contact, leading to the downregulation of EMT [35]. Consistently, cell mobility, invasion, and EMT were also shown to be inhibited by PBE. These ndings indicate that PBE is able to inhibit the ECM-receptor interaction and FA pathways, suppressing cell migration, invasion, and EMT. ECM and FA are also involved in the regulation of many signaling pathways, including the Hippo pathway. Low ECM resistance and FA correspond to reduced mechanics and subsequently regulate the Hippo pathway [36]. YAP/TAZ, the major effector of the Hippo pathway, senses alterations in ECM composition [37] and its downstream targets CYR61 and CTGF are upregulated. This upregulation has been linked to drug resistance and RAS/MAPK blockade in CRC cells, indicating their critical role in RAS-mutated metastasis and drug resistance [38] [39] [40] [41]. Here we found that PBE treatment blocked YAP signaling and its downstream gene expression. This indicates that PBE inhibits CRC progression by regulating YAP signaling, which is closely associated with EMT progression. PBE also inhibited several EMT markers, including SNAIL, TWIST, and SLUG. We also showed that PBE was able to inhibit cell growth, invasion, and migration in both HCT116 and HCT29 cells. Based on these ndings we suggest that PBE might exhibit synergistic effects with 5-Fu and EGFR inhibitors (cetuximab) when used to treat CRC. However, this requires further investigation. YAP is regulated by the Hippo pathway, including MST1/2 and LATS1/2, which interacts with the phosphatase and tensin homolog /AKT/ mechanistic target of rapamycin autophagic axis [42]. However, the role of autophagy in CRC remains unclear and still requires further in-depth mechanistic studies to be fully understood [43]. Given this, it would be interesting to investigate the role of PBE in the modulation of autophagy and the regulation of YAP via the Hippo pathway.
Approximately 50% of patients with CRC present with mutations in the β-catenin gene [44]. This means that WNT/β-catenin signaling is a potential therapeutic target for CRC. We found that PBE treatment decreased the nuclear translocation of β-catenin by increasing β-catenin, GSK-3β, and YAP complex formation, which blocked β-catenin-mediated oncogenic signaling and inhibited invasion, migration, and EMT progression. Hippo signaling also interacts with Notch signaling to suppress liver tumorigenesis [45] and LAMA5 expression is associated with Notch signaling, which promotes EMT by interacting with SLUG and SNAIL [33]. Based on the inhibitory effect of PBE on EMT, it would be interesting to investigate the regulatory role of PBE in Notch signaling, which has already been tightly linked to CRC progression [46]. However, this requires further investigation. In this study, we showed that PBE exerts a signi cant inhibitory effect on the two main signaling pathways (YAP and Wnt/βcatenin) associated with CRC progression. However, KRAS mutations and PI3K/AKT activation play important roles in drug resistance [47], and PBE exerts equal inhibitory effects in both KRAS mutant and wild-type CRC cells. It would be interesting to investigate the synergistic effect of PBE and EGFR inhibitors in the treatment of CRC in future work.

Conclusions
Our ndings suggest that Polygonum barbatum exerts an inhibitory effect on CRC cell motility and tumorigenic potential, making these extracts a potential therapeutic for this type of cancer. We suggest that this makes it an appropriate candidate for future clinical trials. The datasets used and/or analysed during the current study are available from the corresponding author upon request.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

47.
T. Vu and P.K. Datta. "Regulation of EMT in Colorectal Cancer: A Culprit in Metastasis," Cancers (Basel), vol. 9, no. 12. Figure 1 HPLC ngerprint of the Polygonum barbatum extract (PBE) (A) and quercetin (B). The amount of the active marker (quercetin) was determined using analytical RP-HPLC and a methanol-water gradient. The peaks were detected by UV light, and the analytes were quanti ed at 370 nm.    presented as the mean ± SEM from three independent experiments. *p<0.05, **p<0.01, and ***p<0.001 vs vehicle-treated cells.

Figure 6
Effect of PBE on the YAP signaling pathway Phosphorylation of YAP following PBE treatment was analyzed using western blotting and evaluated in terms of both time (A-D) and dose dependency (E-H).
Nuclear and cytosolic translocation of YAP was also evaluated in both cell lines (I-N). The relative expression of CTGF and CYR61 was determined using quantitative PCR (O-R) and Co-IP was used to evaluate the interactions between YAP, β-catenin, and GSK-3β (S). Data are presented as the mean ± SEM from three independent experiments. *p<0.05, **p<0.01, and ***p<0.001 vs vehicle-treated cells.

Figure 7
Effect of PBE on the GSK3β/β-catenin signaling pathway Phosphorylation of GSK3β was evaluated in terms of both time (A-D) and dose dependency (E-H) using western blotting. Nuclear and cytosolic translocation of β-catenin was also evaluated in both cell lines (I-N) and the protein levels of cyclin D1, c-Myc, and c-Jun were analyzed and quanti ed (O-X). Data are presented as the mean ± SEM from three independent experiments. *p<0.05, **p<0.01, and ***p<0.001 vs vehicle-treated cells.

Figure 8
Proposed mechanisms of action for PBE in CRC cells. PBE increases the phosphorylation of YAP and blocks WNT signaling, decreasing cell adhesion and ECM stiffness, resulting in the inhibition of CRC cell invasion and migration.