Daurisoline Suppresses Growth of Esophageal Squamous Cell Carcinoma by Inhibiting MEK1/2 In Vitro and In Vivo

Esophageal squamous cell carcinoma (ESCC) accounts for 90% of esophageal cancer and has a high mortality rate worldwide. The clinical treatment of ESCC is mainly surgical resection. The ve-year survival rate of ESCC patients in developing countries is less than 20%. Therefore, identifying new and effective drugs that can prevent the occurrence and recurrence of ESCC is clinically signicant. Here, daurisoline, a bis-benzylisoquinoline alkaloid, was found to have an anticancer effect on ESCC. We investigated the effects of daurisoline on ESCC cell growth and proliferation using ESCC cell lines (KYSE150 and KYSE450 cells) and tumor growth in an ESCC patient-derived xenograft model. Phosphoproteomics was used to identify changes in protein phosphorylation after daurisoline treatment. Molecular docking simulation, pull down assay and amino acid mutation experiments were conducted to determine the target proteins and specic amino acid binding sites of daurisoline. In vitro kinase assay was used to determine the effect of daurisoline on protein phosphorylation. The correlation between MEK1/2 and ERK1/2 expression levels in ESCC was analyzed using TCGA database.


Background
Esophageal squamous cell carcinoma (ESCC) accounts for 90% of esophageal cancer and has a high mortality rate worldwide. The clinical treatment of ESCC is mainly surgical resection. The ve-year survival rate of ESCC patients in developing countries is less than 20%. Therefore, identifying new and effective drugs that can prevent the occurrence and recurrence of ESCC is clinically signi cant. Here, daurisoline, a bis-benzylisoquinoline alkaloid, was found to have an anticancer effect on ESCC.

Methods
We investigated the effects of daurisoline on ESCC cell growth and proliferation using ESCC cell lines (KYSE150 and KYSE450 cells) and tumor growth in an ESCC patient-derived xenograft model.
Phosphoproteomics was used to identify changes in protein phosphorylation after daurisoline treatment. Molecular docking simulation, pull down assay and amino acid mutation experiments were conducted to determine the target proteins and speci c amino acid binding sites of daurisoline. In vitro kinase assay was used to determine the effect of daurisoline on protein phosphorylation. The correlation between MEK1/2 and ERK1/2 expression levels in ESCC was analyzed using TCGA database.

Results
In vitro experiments showed that daurisoline inhibited the proliferation and anchorage-independent growth of ESCC cells. In vivo experiments indicated that daurisoline signi cantly inhibited tumor growth. Phosphoproteomics analysis revealed that daurisoline reduced ERK1/2 phosphorylation. A pull down assay showed that daurisoline could bind to MEK1/2. In vitro kinase assay con rmed that daurisoline inhibited the biological functions of MEK1/2. We observed a signi cant correlation between MEK1 and ERK2 in ESCC from the TCGA database.

Conclusion
Daurisoline is a MEK1/2 inhibitor that suppressed ESCC growth in vitro and in vivo.

Background
Esophageal cancer (ESCA) ranks seventh and sixth in incidence and mortality, respectively, among all cancers worldwide [1]. Esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC) are two main pathological classi cations of ESCA [2]. It has been reported that 90% of ESCA cases are ESCC worldwide [3]. Clinical treatment of ESCC is mainly surgery, chemotherapy and radiotherapy, but the prognosis of ESCC is poor [4]. The ve-year survival rate of ESCC patients in developing countries is less than 20% [5]. Therefore, it is of great clinical signi cance to nd drugs that can prevent the occurrence and recurrence of ESCC [6].
Activation of the MEK1/2-ERK1/2 pathway plays an essential role in the proliferation, differentiation, invasion, and metastasis of cancer cells [7]. A large number of studies showed that the changes in the structure and expression level of MEK1/2 are closely related to the occurrence of tumors [8]. Therefore, MEK1/2 inhibitor is considered to be one of the promising areas of cancer research [9]. Currently, a variety of MEK1/2 inhibitors have been found, some of which have been used in the treatment of cancer. Selumetini is used to treat melanoma, and selumetinib is used to treat advanced differentiated thyroid cancer and plexus neuro broma [10][11].
Screening natural compounds is an effective way to nd anticancer drugs. Bis-benzylisoquinoline alkaloids are a kind of natural alkaloids. They are widely distributed in the plants and have a variety of physiological activities [12][13]. Daurisoline is a bis-benzylisoquinoline alkaloid extracted from the rhizome of menisperum dauricum, a traditional Chinese medicine [14]. Previous studies showed that daurisoline has potential pharmacological effects on a variety of diseases such as focal ischemia, reperfusion injury, arrhythmia, platelet aggregation, thrombotin A2 inhibition, and so on [13][14][15][16][17][18]. A few studies found that daurisoline can inhibit the tumor cells proliferation [19][20]. However, the precise mechanism underlying the anticancer effects of daurisoline remains unclear.
Here, quantitative phosphoproteomics was used to explore the potential molecular mechanism of daurisoline in ESCC [21]. Additionally, effect of daurisoline on ESCC tumor growth was evaluated in vivo using the ESCC patient-derived xenograft (PDX) model. Our ndings indicated that daurisoline is a MEK1/2 inhibitor and suppresses ESCC growth.

Reagents and antibodies
Daurisoline was purchased from Shanghai Liding Biotechnology Co., Ltd., (Shanghai, China; purity ≥98%, Cat No DR10996) for the in vitro and in vivo analyses. MEK1/2 and ERK1/2 antibodies from Cell Signaling, p-ERK1/2 T185/Y187 antibody from Thermo Fisher Scienti c, p-MEK1/2 antibody from A nity Biosciences Ltd., t-MEK1 and t-MEK2 antibody from Sino Biological, Ki67 antibody from Abcam, β-tubulin antibody from Hangzhou HuaAn Biotechnology Co., Ltd., and GAPDH antibody from ProteinTech. In vitro kinase assay The active MEK1 (100 ng) or MEK2 (50 ng) protein was mixed with daurisoline or vehicle (DMSO) at different concentrations in the reaction buffer at 25 ℃ for 15 min. Then, the inactive ERK2 protein (250 ng) and ATP (100 μM) were added and incubated at 30 ℃ for 30 min. Finally, the proteins were analyzed using western blotting.

GSEA analysis
The GSEA (gene set enrichment analysis) V4.0.3 (http://www.broadinstitute.org/gsea/) was used to analyze the phosphoproteome dataset to observe the difference of protein expression between the daurisoline treatment and control groups. See the references for more details [23].

Establishment of the ESCC PDX model
The tumor sample from an ESCC patient was EG20 (ESCC, T 2 N 0 M 0 II, moderately differentiated, obtained from Linzhou Cancer Hospital, Henan Province, China). The daurisoline treatment group was given 20 and 40 mg/kg, respectively. More experimental details are borrowed from our previous method [25].

Immunohistochemical staining
Immunochemical staining was used to detect the molecular changes in EG20 tumors. The indexes were Ki67, p-MEK1/2, and p-ERK1/2. See our previous instructions for the complete experiment [25].
Statistical analyses SPSS 24.0 software (IBM, USA) was used for data analysis in this experiment, and the results were expressed as Mean ± SD. Student T-test, one-way ANOVA or Kruskal-Wallis H-test was used according to different conditions of the data. The experiment was repeated three times, and p < 0.05 indicated statistically signi cant difference.

Daurisoline inhibits ESCC cell proliferation in vitro
By drug screening, 40 μM daurisoline was found to have obvious cytotoxicity to KYSE450 cells (Fig. 1a).
Daurisoline is a bis-benzylisoquinoline alkaloid (Fig. 1b). The IC 50 of KYSE150 and KYSE450 cells were 20.733 and 13.560 μM at 24 h, and 11.0149 and 9.936 μM at 48 h, respectively. Under daurisoline treatment for 24 h and 48 h, IC 50 of immortalized esophageal epithelial cells (SHEE cells) was 41.102 μM and 11.406 μM, respectively. With the increase in daurisoline concentration, the survival rate of cells decreased gradually (Fig. 1c). In the cell proliferation assay, we used the following drug concentrations: 0, 2.5, 5, 10, and 20 μM. Daurisoline inhibited the proliferation of KYSE150 and KYSE450 cells in a dosedependent manner. The proliferation inhibition rates of KYSE150, KYSE450, and SHEE cells were 44.51%, 68.83%, and 25.50%, respectively, after treatment with 20 μM daurisoline for 96 h (Fig. 1d). These results indicated that daurisoline was less toxic to SHEE cells compared with ESCC cells. To verify this, we further performed the anchorage-independent cell growth assay using the same drug concentration.
Daurisoline inhibited the clonal size and number of clones formed for KYSE150 and KYSE450 cells. At 20 μM, the clonal formation rates of KYSE150 and KYSE450 cells were 30.25% and 42.36%, respectively, compared with that of the control cells (Fig. 1e). These results indicated that daurisoline inhibited ESCC cell proliferation in vitro.

Daurisoline signi cantly downregulates the phosphorylation of ERK1/2 in ESCC cells
We used quantitative phosphoproteomics to investigate the molecular mechanism underlying daurisoline inhibiting ESCC cell proliferation. We treated KYSE150 cells with 20 μM daurisoline for 24 h. (Fig. 2a). The rst-order mass error of most spectrograms was within 10 ppm, which conformed to the high precision characteristic of orbital well MS (Fig. S1a). Most peptides were 7-20 amino acids in length. The distribution of the peptide lengths identi ed by MS conformed to the quality control requirements (Fig.  S1b). The number of secondary spectra obtained was 383,273. After screening, the number of available effective spectrograms was 94,520. We identi ed 12,965 phosphorylation sites on 3130 proteins, of which 7415 sites on 2500 proteins contained quantitative data. To ensure a high credibility of the results, p > 0.75 was used to lter the identi cation data. Consequently, we identi ed 8549 sites on 2861 proteins, of which 6601 sites on 2408 proteins included quantitative data (Fig. S1c). The protein quanti cation group was normalized to remove the in uence of protein expression on the modi ed signal. The screening criteria for different sites were as follows: 1.5 times the change threshold and two-sample twotailed t-test (p < 0.05). Based on these data and standards, we found that the phosphorylation of 176 sites was enhanced and that of 340 sites was reduced in the daurisoline treated cells (Fig. S1d).
A total of 340 down regulated phosphorylation sites were enriched in the Kyoto Encyclopedia of genes and genomes (KEGG) pathway. KEGG pathway was selected according to Fisher's exact test (p < 0.05). Ultimately, nine KEGG pathway targets were screened: viral carcinogenesis, spliceosome, RNA transport, adipocytokine signaling pathway, herpes simplex virus infection, synaptic vesicle cycle, Salmonella infection, chronic myeloid leukemia and Th17 cell differentiation (Fig. 2b). The heat map shows the levels of the phosphorylation sites in the viral carcinogenesis pathway, and most of the phosphorylation sites are downregulated (Fig. 2c). In six KEGG pathways, multiple proteins appeared repeatedly in the different pathways. Therefore, Wayne diagrams are made based on the six KEGG pathways to nd the key proteins in these pathways (Fig. 2d). We found that NFκB1 was enriched in six pathways, ERK2 in four pathways, and STAT3 in three pathways. The original mass spectra of p-ERK2 T185/Y187 are shown in Supplement Figure (Fig. S1e). In the volcanic map, according to the -log 10 ( P-value), phosphorylation of the four phosphorylation sites is ranked from top to bottom as p-ERK2 Y187 (3.0177), p-ERK2 T185 (2.4976), p-STAT3 S727 (1.7433), and p-NFκB1 S907 (1.5686) (Fig. 2e). In addition, the protein expression ratios of the above four phosphorylation sites are 0.504, 0.458, 0.566, and 0.503, respectively (Fig. 2f). In conclusion, we believe that p-ERK2 T185/Y187 are important phosphorylation sites in the above nine KEGG pathways. In addition, although ERK1 was not enriched in the KEGG pathway, the phosphorylation of ERK1 was also downregulated in the phosphoproteomics data. The phosphorylation level of ERK1/2 in KYSE150 and KYSE450 cells was signi cantly downregulated after daurisoline (20 μM) treatment for 24 h (Fig. 2g). These results indicated that daurisoline inhibited ERK1/2 phosphorylation in ESCC cells.
Daurisoline directly binds with MEK1 /2 We used IGPS1.0 software to predict upstream kinases for speci c phosphorylation sites in the above nine KEGG pathways. The prediction results showed that the upstream kinase proteins of p-ERK2 T185 and p-ERK2 Y187 were MEK1-7 (MAP2K1-7), and ALK and LTK, respectively (Table. S1). In order to verify the accuracy of the upstream kinase prediction, we performed the CNBr-activated Sepharose 4B pull down assay (Fig. S2). The protein binding to the drug was detected by Mass spectrum. The results of mass spectrometry showed that the protein binding to daurisoline were MEK1, MEK2, MEK3, and MEK7 with Unused ProtScores of 38.69, 17.28, 15.76, and 4.83, respectively (Table S2). Previous studies on MAPK signaling pathway showed that the sole substrate of MEK1/2 protein kinase is ERK1/2, the substrate of MEK3 is p38MAPK, and the substrate of MEK7 is JNK [26]. Therefore, based on the Unused ProtScores, daurisoline was likely to affect the biological function of MEK1/2 and downregulate ERK1/2 phosphorylation. Autodock 4.0 software was used for molecular docking in order to explore the speci c binding sites of daurisoline with MEK1/2 proteins. Daurisoline could bind to MEK1 (binding energy: -11.06 kcal/mol). The binding sites were Asn78 and Lys97 (Fig. 3a). Daurisoline could also bind to MEK2 (binding energy: -8.50 kcal/mol). The binding sites were Asp194 and Asp212 (Fig. 3b). Pull down assays displayed that daurisoline could directly bind to MEK1/2 (Fig. 3c, d). In addition, daurisoline also binds directly to MEK1 and MEK2 in KYSE150 and KYSE450 cells (Fig. 3e, f). To verify the accuracy of the molecular simulation results, we substituted the above-mentioned amino acids of MEK1 and MEK2 proteins with alanine (Ala, A). The results showed that after Asn78 substitution on MEK1 protein, the binding e ciency of the protein with daurisoline was signi cantly reduced (Fig. 3g). Furthermore, the binding e ciency of MEK2 protein with daurisoline was signi cantly reduced after Asp194 substitution (Fig. 3h).

Daurisoline inhibits the MEK1/2-ERK1/2 signaling pathway in ESCC
We performed an in vitro kinase assay. The phosphorylation e ciency of active MEK1 and MEK2 against inactive ERK2 was in turn reduced at 2.5, 5, 10, and 20 μM daurisoline (Fig. 4a, b). These results indicated that daurisoline could inhibit ERK2 phosphorylation by inhibiting MEK1/2 activity. Moreover, Coomassie blue staining showed the location and expression of MEK1/2 and ERK2 proteins.
We used gene set enrichment analysis (GSEA) to enrich the KEGG pathway associated with altered expression of all proteins identi ed in the phosphoproteomics analysis. We found that ERK1/2 was enriched and signi cantly downregulated in the Erbb signaling pathway (Fig. 4c) and prostate cancer pathway (Fig. 4d). In addition, MEK2 (MAP2K2) was also enriched in the aforementioned two pathways. Western blotting showed that daurisoline inhibited ERK1/2 phosphorylation in KYSE150 and KYSE450 cells. However, daurisoline had no signi cant effect on MEK1/2 phosphorylation (Fig. 4c, d). The above results indicate that daurisoline inhibits the MEK1/2-ERK1/2 signaling pathway in ESCC cells.
MEK1/2 were signi cantly overexpressed in ESCC MEK1/2 have been reported to be highly expressed in ESCC [27]. The Cancer Genome Atlas (TCGA) database revealed that MEK1 (Fig. 5a) and MEK2 (Fig. 5b) were both signi cantly overexpressed in a variety of cancers, including ESCA. In addition, MEK1 (Fig. 5c) and MEK2 (Fig. 5d) are also overexpressed in EAC and ESCC. We used ESCC data from the TCGA database to conduct a multi-gene correlation analysis. We observed a signi cant correlation between MEK1 and ERK2 in ESCC (Fig. 5f, g).

Daurisoline inhibits ESCC tumor growth in vivo
To investigate the anticancer effect of daurisoline in vivo, we established the PDX model of ESCC (Fig.  6a). There was no signi cant difference in the body weight between mice treated with daurisoline and control, indicating that daurisoline had no toxic or side effects on mice (Fig. 6b). The nal tumor volume was measured at the end of the treatment (Fig. 6c). The tumor growth rate of daurisoline treated group was signi cantly lower than that of control group (Fig. 6d, g). The tumor weight also decreased with the increase of the dose of daurisoline (Fig. 6e). We calculated the tumor weight inhibition rate of daurisoline treated group. The growth inhibition rates in the 20 and 40 mg/kg daurisoline treatment groups were 52.77% and 84.74%, respectively (Fig. 6f). Immunohistochemical staining showed that the Ki67 and ERK1/2 phosphorylation decreased after daurisoline treatment. The MEK1/2 phosphorylation did not change signi cantly (Fig. 6h). In conclusion, daurisoline inhibited ESCC tumor proliferation and ERK1/2 phosphorylation in tumor cells in vivo.

Discussion
It is of great clinical signi cance to nd drugs that can prevent the occurrence and recurrence of ESCC [6]. Traditional Chinese medicine has long been used in clinical settings. Compounds extracted from traditional Chinese medicine represent a promising resource for anticancer drug development [28]. By screening natural compounds, we found that daurisoline exhibits anticancer effects on ESCC. We found that daurisoline inhibited the proliferation of ESCC cells. In addition, daurisoline was less toxic to SHEE cells (Fig. 1c, d). These results indicate that at the cellular level, the anticancer effect of daurisoline is more bene cial than harmful. Furthermore, PDX model can retain the histological, molecular, and genetic characteristics of the original tumor; thus, it is superior to the traditional cell lines and their derivatives in vitro [29][30]. In this study, we observed that daurisoline displayed a signi cant anticancer effect in vivo (Fig. 6e, f). The anticancer effect of daurisoline in ESCC PDX model was also signi cantly better than that of CDX model of lung cancer [19].
Phosphorylation regulates many aspects of protein function. It is a dynamic post-translational modi cation, which alters protein structure, localization, interaction, and stability [31]. Although a few literatures have reported the anticancer effects of daurisoline in vitro and in vivo, the mechanistic target of daurisoline remains undetermined [18][19][20]. Therefore, in this study, we explored the mechanism underlying daurisoline inhibiting ESCC cell proliferation through phosphoproteomics. The results of mass spectrometry showed that daurisoline inhibited ERK1/2 phosphorylation in ESCC cells (Fig. 2g). We nally identi ed the molecular targets of daurisoline as MEK1/2. Moreover, In vitro kinase assay showed that daurisoline inhibited ERK2 phosphorylation by inhibiting MEK1/2 activity (Fig. 4a, b). In addition, we performed a polygenetic correlation analysis on ESCC clinical samples from the TCGA database, and found that MEK1 and ERK2 were signi cantly correlated (Fig. 5e, f). MEK1/2 kinases can speci cally phosphorylate and activate ERK1/2 [32]. Abnormal regulation of the MEK1/2-ERK1/2 pathway will inevitably lead to cell cancerization [33]. As a key nodal protein in this signaling pathway, MEK1/2 plays an important role in tumorigenesis [34]. MEK1/2, as the target of tumor therapy, is highly selective and unique [35][36]. MEK1 and MEK2 share 80% of the amino acid composition in the kinase domain and are highly conserved evolutionarily [37]. The interaction between natural compounds and target proteins can be simulated by molecular docking technology [38]. We observed that daurisoline bound to MEK1 at Asn78 and Lys97 sites, and MEK2 at Asp194 and Asp212 sites (Fig. 3a, b). It was found that Asn78 residue of MEK1 and Asp194 residue of MEK2 were important binding sites for daurisoline (Fig. 3g, h). Therefore, our ndings showed that daurisoline inhibited the growth of ESCC tumor by inhibiting the MEK1/2-ERK1/2 signaling pathway (Fig. 6i). Daurisoline may prevent ESCC in high-risk individuals and reduce the postoperative recurrence rate of ESCC.

Conclusions
In this study, we demonstrated that daurisoline is a MEK1/2 inhibitor, which led to a decrease in ERK1/2 phosphorylation. This in turn leads to inhibition of ESCC cells proliferation and tumor growth in the ESCC PDX model.     database. e A heat map of the correlation between MEK1/2 and ERK1/2 of ESCC. The horizontal and vertical coordinates represent genes, and different colors represent correlation coe cients, and the darker the color represents the two stronger correlation. f Spearman correlation analysis of MEK1 expression and ERK2 expression. The horizontal axis in the gure represents the expression distribution of MEK1, and the ordinate is the expression distribution of ERK2. The density curve on the right represents the distribution trend of ERK2; the upper density curve represents the distribution trend of MEK1; In the top side the value represents the correlation P value, correlation coe cient and correlation calculation method. (*P < 0.05, **P < 0.01, ***P < 0.001) Figure 6 Daurisoline inhibit tumor growth of ESCC in vivo. a Schematic diagram of establishing PDX models of ESCC. b Changes in body weight after treatment of the daurisoline. c Photographs of the EG20 tumor, d changing curve of EG20 tumor volume, e tumor weight and f tumor weight inhibition rate after treatment of daurisoline. g EG20 tumor volume changes of each mouse with different doses of daurisoline. h Immunohistochemical Staining of EG20 tumor tissue (bottom graph) and statistic graph (upper graph). i