-
环境内分泌干扰物(EDCs)[1]是20世纪90年代以来引起广泛关注的一类新型化学物质.分为天然和人工合成两类,除铅、汞、砷等金属类外,其余均为有机化学物.按其用途可分为工业原材料、洗涤用品、化妆品、农药、塑料包装和儿童玩具等,因工业生产和使用被不断释放到空气、水体、土壤等环境介质中,可经呼吸空气、饮水、食物摄入及皮肤接触等途径进入人体,影响生物体内激素的合成、释放和代谢,甚至可与激素竞争结合相应受体,或者影响激素受体上下游的调控基因,干扰激素正常的调节功能,造成内分泌系统的紊乱,从而引发各种疾病,对人类健康产生潜在的危害和不利影响[2].
已有大量研究报道EDCs可产生类雌激素效应[3],如邻苯二甲酸脂类、多氯联苯类、双酚类等化合物.这些EDCs在环境浓度下发挥着与天然雌激素类似的生物学活性,从而影响女性的生殖健康[4].已知的不利健康危害包括:月经周期变化、子宫内膜异位症、子宫肌瘤、多囊卵巢综合征、不孕不育、乳腺癌、子宫内膜癌及卵巢癌.此外,近年来的流行病学调查 报告进一步证实女性雌激素敏感癌症发病率与环境中EDCs的浓度存在密切相关性.
EDCs诱导产生的类雌激素效应主要通过经典的核雌激素受体ERα与ERβ介导[5].随着结构生物学研究的深入,有报道证实存在膜雌激素受体(GPER),属于7-跨膜G蛋白偶联受体(GPCR)家族[6]的一员,作为雌激素信号的靶蛋白,参与了雌激素在生殖、神经、内分泌、免疫和心血管系统中的介导作用,对于包括癌症在内的一系列疾病,GPER正成为一种新的治疗靶点和预后指标[7-9].G蛋白偶联雌激素受体(GPER)最初被称为GPR30[10],其结构未被结晶实验确认,激动和拮抗的分子机制及其结构特征的研究仍处在初步阶段.不同于经典的核雌激素受体亚型,GPER被认为是介导快速细胞信号的媒介[11],参与介导了雌激素快速非基因组效应[12].内源性化合物,如雌激素可与细胞膜上的GPER结合,激活快速细胞效应反应[13],包括环磷酸腺苷的产生、细胞内钙离子的动员、多种激酶的激活,如细胞外信号调节激酶、肌醇磷脂3激酶、离子通道以及内皮型一氧化氮合酶等途径.影响下游效应分子在相应的靶组织中发挥其生物学效应[14],从而导致细胞的增值与分化.类似地,环境污染物也可模拟内源性物质结合GPER[15].如BPA可在低浓度下激活GPER,产生较强的雌激素效应[16].
因此,构建快速筛查环境污染物是否结合GPER的分类预测模型,可为理论评估化合物健康风险与毒性效用提供重要的依据.虽然针对环境污染物的毒性预测,已有报道显示机器学习算法可表现出良好的分类预测性能 [17-20],但是结合GPER的小分子却未有可用的分子数据库和已知的预测模型,这限制了构建针对GPER分类预测模型的发展.为了解决这一问题,本研究系统地总结了已报道的有机小分子结合GPER的数据,并进一步评测了随机森林(random forest,RF)、支持向量机(support vector machine,SVM)、神经网络(neural network,ANN)、K-最近邻(K-nearest neighbour,KNN)、朴素贝叶斯(naive bayes,NB)、逻辑回归(logistic,LG)等6种典型机器学习算法的分类预测性能,其中基于RF算法构建的分类模型展示出了优秀的分类预测表现.
基于膜雌激素受体(GPER)结合化合物能力的分类预测模型
Classification prediction model based on GPER binding ability of membrane estrogen receptor
-
摘要: 近年来,计算毒理学的方法被广泛应用于潜在的环境内分泌干扰物(EDCs)的筛选.膜雌激素受体(GPER),作为一种可以快速响应内源性配体雌激素的关键靶蛋白,调控其介导的多项生理学功能.但是针对GPER的化合物毒性预测模型仍未见报道.因此,本研究收集了130个化合物对GPER的结合活性数据,主要包括双酚类、多溴联苯类以及农药杀虫剂类环境污染物.利用随机森林(RF)、支持向量机(SVM)、人工神经网络(ANN)、K最近邻(KNN)、朴素贝叶斯(NB)以及逻辑回归(LG)等6种机器学习算法构建二分类模型.结果显示,所有被测试算法的测试集准确率均达到85%以上,其中SVM、RF、ANN、KNN等4种算法的训练集准确率高于93%,10折交叉验证准确率高于80%,说明得到的模型具有优秀的分类预测性能.因此,本研究基于机器学习算法构建的分类模型,可以用来快速、准确地预测环境污染物是否通过结合GPER产生内分泌干扰效应.为评估环境污染物的潜在健康风险提供了理论依据.Abstract: In recent years, computational toxicology has been widely applied to the screening of potential environmental endocrine disruptors(EDCs).Membrane estrogen receptor(GPER), as a key target protein that can rapidly respond to endogenous ligand estrogen, regulates its mediated multiple physiological functions. However, the prediction model of compound toxicity for GPER has not been reported. Therefore, the binding activity data of 130 compounds against GPER were collected in this study, mainly including bisphenols, polybrominated biphenyls and environmental pollutants like pesticides and insecticides.Six machine learning algorithms, including random forest (RF), support vector machine(SVM), artificial neural network(ANN), K-nearest neighbor(KNN), Naive Bayes(NB)and logistic regression(LG), were used to construct the dichotomous model.The results showed that the test set accuracy of all the algorithms tested reached more than 85%, among which the training set accuracy of SVM、RF、ANN and KNN algorithm was 93% higher than that of the four algorithms, and the 10-fold cross validation accuracy was 80% higher than that of the four algorithms, indicating that the model obtained had excellent classification prediction performance. Therefore, the classification model built based on machine learning algorithm in this study can be used to quickly and accurately predict whether environmental pollutants will produce endocrine disrupting effects by combining with GPER.It provides a theoretical basis for evaluating the potential health risks of environmental pollutants.
-
表 1 模型选定的描述符
Table 1. Selected descriptors for models
变量
Variables描述
DescriptionGATS4c Geary autocorrelation - lag 4 / weighted by charges GATS4s Geary autocorrelation - lag 4 / weighted by I-state ETA_Eta Composite index Eta AATSC6i Average centered Broto-Moreau autocorrelation - lag 6 / weighted by first ionization potential MATS2i Moran autocorrelation - lag 2 / weighted by first ionization potential 表 2 模型预测结果
Table 2. Model prediction results
数据集
Dataset化合物数
Chemical number
(n)真阳性
True positive
(TP)真阴性
True negative
(TN)假阴性
False negative
(FN)假阳性
False positive
(FP)模型 RF 训练集 104 73 30 0 1 测试集 26 17 6 1 2 模型 SVM 训练集 104 72 29 1 2 测试集 26 16 7 2 1 模型ANN 训练集 104 72 27 1 4 测试集 26 15 7 3 1 模型 KNN 训练集 104 71 28 2 3 测试集 26 15 7 3 1 模型 NB 训练集 104 69 16 4 15 测试集 26 16 6 2 2 模型 LG 训练集 104 70 13 3 18 测试集 26 16 7 2 1 表 3 模型化合物预测情况
Table 3. Prediction of model compounds
序号
No.化合物
Compound参考文献
ReferenceGPER配体分子(是/否)
GPER ligand molecule(Y/N)观测值
ObservedSVM RF ANN KNN 1 SK0* [25] Y Y Y Y Y 2 SK0P [25] Y Y Y Y Y 3 G-1 [6, 21] Y Y Y Y Y 4 G-15 [6,21] Y Y Y Y Y 5 G-36 [6,21] Y Y Y Y Y 6 Oleuropein [24] Y Y Y Y Y 7 Hydroxytyrosol [24] Y Y Y Y Y 8 MIBE [24] Y Y Y Y Y 9 4-hydroxytamoxifen [24] Y Y Y Y Y 10 GPER-L1 [24] Y Y Y Y Y 11 GPER-L2 [24] Y Y Y Y Y 12 17β-estradiol [22] Y Y Y Y Y 13 E3 [24] N N N Y N 14 Tamoxifen [22] Y Y Y Y Y 15 Fulvestrant [22] Y Y Y Y Y 16 Epi* [27] Y Y Y Y Y 17 Epi-prop [27] Y Y Y Y Y 18 Epi-4-prop [27] Y Y Y Y Y 19 Epi-5-prop [27] Y Y Y Y Y 20 Epi-Ms [27] N N N Y N 21 C4PY [26] Y Y Y Y Y 22 7β-OH-EpiA* [24] Y Y Y Y Y 23 G-DOTA [37] Y Y Y Y Y 24 G-Bz-DOTA [37] N N N N N 25 G-Bz-DTPA [37] N N N N N 26 Atrazine [24] Y Y Y Y Y 27 PBX1 [34] Y Y Y Y Y 28 PBX2 [34] Y Y Y Y Y 29 ZINC65156419(1) [29] Y Y Y Y N 30 ZINC65156419(2) [29] N N N N N 31 ZINC65156419(3) [29] N N N N N 32 ZINC65156419(4) [29] N N N N N 33 ZINC65156419(5) [29] Y Y Y Y Y 34 ZINC65156419(6) [29] N N N N N 35 ZINC65156419(7) [29] N N N N N 36 ZINC65156419(8) [29] N N N N N 37 ZINC65156419(9)* [29] Y Y Y N Y 38 E2-NH3+ [13] Y Y Y Y Y 39 E2-COO- [13] Y Y Y Y Y 40 E2-NMe3+ [13] Y Y Y Y Y 41 E2-NB [13] Y Y Y Y Y 42 o,p'-DDE [6] Y Y Y Y Y 43 E1* [24] N N Y N N 44 α-E2 [24] N Y Y Y Y 45 Genistein [6] Y Y Y Y Y 46 p,p'-DDT [6] Y Y Y Y Y 47 BPA* [6] Y Y Y Y Y 48 quercetin [24] Y Y Y Y Y 49 Resveratrol* [24] Y Y Y Y Y 50 Raloxifene [24] Y Y Y Y Y 51 zearalonone [6] Y Y Y Y Y 52 Nonylphenol [6] Y Y Y Y Y 53 kepone [6] Y Y Y Y Y 54 STX [24] Y Y Y Y Y 55 PPT* [24] Y Y Y Y Y 56 2,2',5'-PCB-4-OH [6] Y Y Y Y Y 57 equol [24] Y Y Y Y Y 58 2-methoxye stradiol [24] Y Y Y Y Y 59 niacin [24] Y Y Y Y Y 60 daidzein [24] Y Y Y Y Y 61 BDE-003 [28] N N N N N 62 BDE-007* [28] N N N N N 63 BDE-028 [28] N N N N N 64 BDE-047 [28] N N N N N 65 BDE-049* [28] N N N N N 66 BDE-085* [28] N N N N N 67 BDE-099* [28] N N N N N 68 BDE-100 [28] N N N N N 69 BDE-154 [28] N N N N N 70 BDE-180 [28] N N N N N 71 BDE-187 [28] N N N N N 72 BDE-201 [28] N N N N N 73 2'-OH-BDE-003* [28] Y N N N N 74 3'-OH-BDE-007 [28] Y Y Y Y Y 75 3'-OH-BDE-028 [28] Y Y Y Y Y 76 3'-OH-BDE-047* [28] Y Y Y Y Y 77 3'-OH-BDE-154 [28] Y Y Y Y Y 78 4'-OH-BDE-049 [28] Y Y Y Y Y 79 5'-OH-BDE-099* [28] Y Y Y Y Y 80 2'-OH-BDE-007 [28] N N N N N 81 2'-OH-BDE-028* [28] N N N N N 82 3-OH-BDE-100 [28] Y N Y N Y 83 4-OH-BDE-187 [28] Y Y Y Y Y 84 4'-OH-BDE-201 [28] Y Y Y Y Y 85 5-OH-BDE-047 [28] N N N N N 86 5'-OH-BDE-100 [28] N N N N N 87 6-OH-BDE-047 [28] N N N N N 88 6-OH-BDE-085* [28] N N N N N 89 6'-OH-BDE-099 [28] N N N N N 90 6-OH-BDE-180 [28] Y Y Y Y N 91 BPAF [16] Y Y Y Y Y 92 BPB* [16] Y Y Y Y Y 93 BPF* [16] N Y Y Y Y 94 BPS* [16] Y Y Y N N 95 TBBPA [16] N N N N Y 96 TCBPA [16] Y Y Y Y Y 97 Diethylstilbestro [24] N N N N N 98 2-Hydroxy stradiol* [24] Y N Y Y N 99 Aldosterone [24] Y Y Y Y Y 100 Tectoridin [24] Y Y Y Y Y 101 Apigenin* [24] Y Y Y Y Y 102 Methoxychlor [24] Y Y Y Y Y 103 p,p'-DDE* [24] Y Y Y Y Y 104 o,p'-DDT* [24] Y Y Y Y Y 105 DPN [24] Y Y Y Y Y 106 Ethynylestradiol [24] Y Y Y Y Y 107 3MC [32] Y Y Y Y Y 108 AB-1 [31] N N N N N 109 CIMBA-5* [36] Y Y Y Y Y 110 CIMBA-6 [36] Y Y Y Y Y 111 CIMBA-7 [36] Y Y Y Y Y 112 CIMBA-8 [36] Y Y Y Y Y 113 CIMBA-9 [36] N Y N Y N 114 CIMBA-10 [36] Y Y Y Y Y 115 CIMBA-11 [36] Y Y Y Y Y 116 CIMBA-12 [36] Y Y Y Y Y 117 CIMBA-13 [36] Y Y Y Y Y 118 CIMBA-14 [36] N N N N N 119 CIMBA-15 [36] Y Y Y Y Y 120 CIMBA-16 [36] N N N N Y 121 CIMBA-17 [36] Y Y Y Y Y 122 CIMBA-18 [36] Y Y Y Y Y 123 CIMBA-19 [36] N N N N N 124 CIMBA-20 [36] Y Y Y Y Y 125 CIMBA-21 [36] Y Y Y Y Y 126 CIMBA-22 [36] Y Y Y Y Y 127 CIMBA-23 [36] Y Y Y Y Y 128 CIMBA-24 [36] Y Y Y Y Y 129 CIMBA-25 [36] Y Y Y Y Y 130 Carbhydraz [35] Y Y Y Y Y 注:*测试集化合物.
Note:*Testing set compounds.表 4 模型评价
Table 4. Model performance
数据集
Dataset化合物数
Chemical number
(n)敏感性
Sensitivity
(Sn)特异性
Specificity
(Sp)精确度
Accuracy
(Q)马修斯相关系数
Matthews correlation coefficient
(MCC)模型 RF 训练集 104 1 0.968 0.99 0.977 测试集 26 0.944 0.75 0.885 0.723 模型 SVM 训练集 104 0.986 0.935 0.971 0.931 测试集 26 0.889 0.875 0.885 0.741 模型ANN 训练集 104 0.986 0.871 0.952 0.884 测试集 26 0.833 0.875 0.846 0.672 模型 KNN 训练集 104 0.972 0.903 0.952 0.884 测试集 26 0.833 0.875 0.846 0.672 模型 NB 训练集 104 0.945 0.516 0.817 0.535 测试集 26 0.889 0.75 0.846 0.639 模型 LG 训练集 104 0.959 0.419 0.798 0.480 测试集 26 0.889 0.875 0.885 0.741 -
[1] FENICHEL P, CHEVALIER N. Is testicular germ cell cancer estrogen dependent? The role of endocrine disrupting chemicals [J]. Endocrinology, 2019, 160(12): 2981-2989. doi: 10.1210/en.2019-00486 [2] BALABANIC D, RUPNIK M S, KLEMENCIC A K. Negative impact of endocrine-disrupting compounds on human reproductive health [J]. Reproduction Fertility and Development, 2011, 23(3): 403-416. doi: 10.1071/RD09300 [3] BRAUN J M. Early-life exposure to EDCs: Role in childhood obesity and neurodevelopment [J]. Nature Reviews Endocrinology, 2017, 13(3): 161-173. doi: 10.1038/nrendo.2016.186 [4] FOWLER P A, BELLINGHAM M, SINCLAIR K D, et al. Impact of endocrine-disrupting compounds (EDCs) on female reproductive health [J]. Molecular and Cellular Endocrinology, 2012, 355(2): 231-239. doi: 10.1016/j.mce.2011.10.021 [5] HARRIS H A, KATZENELLENBOGEN J A, KATZENELLENBOGEN B S J E. Characterization of the biological roles of the estrogen receptors, ERα and ERβ, in estrogen target tissues in vivo through the use of an ERα-selective ligand [J]. Endocrinology, 2002(11): 4172-4177. [6] THOMAS P, DONG J. Binding and activation of the seven-transmembrane estrogen receptor GPR30 by environmental estrogens: A potential novel mechanism of endocrine disruption [J]. The Journal of Steroid Biochemistry and Molecular Biology, 2006, 102(1): 175-179. [7] PROSSNITZ E R, BARTON M. The G-protein-coupled estrogen receptor GPER in health and disease [J]. Nature Reviews Endocrinology, 2011, 7(12): 715-726. doi: 10.1038/nrendo.2011.122 [8] PROSSNITZ E R, BARTON M. Signaling, physiological functions and clinical relevance of the G protein-coupled estrogen receptor GPER [J]. Prostaglandins Other Lipid Mediat, 2009, 89(3-4): 89-97. doi: 10.1016/j.prostaglandins.2009.05.001 [9] WANG C, LI J, YE S, et al. Oestrogen inhibits VEGF expression and angiogenesis in triple-negative breast cancer by activating GPER-1 [J]. Journal of Cancer, 2018, 9(20): 3802-3811. doi: 10.7150/jca.29233 [10] NIRO S, PEREIRA E, PELISSIER M A, et al. The DHEA metabolite 7β-hydroxy-epiandrosterone exerts anti-estrogenic effects on breast cancer cell lines [J]. Steroids, 2012, 77(5): 542-551. doi: 10.1016/j.steroids.2012.01.019 [11] PROSSNITZ E, BARTON M. Estrogen biology: New insights into GPER function and clinical opportunities [J]. Molecular and Cellular Endocrinology, 2014, 389(1): 71-83. [12] ALBANITO L, MADEO A, LAPPANO R, et al. G protein-coupled receptor 30 (GPR30) mediates gene expression changes and growth response to 17β-estradiol and selective GPR30 ligand G-1 in ovarian cancer cells [J]. Cancer Research, 2007, 67(4): 1859-1866. doi: 10.1158/0008-5472.CAN-06-2909 [13] REVANKAR C M, MITCHELL H D, FIELD A S, et al. Synthetic estrogen derivatives demonstrate the functionality of intracellular GPR30 [J]. Acs Chemical Biology, 2007, 2(8): 536-544. doi: 10.1021/cb700072n [14] TSUGAWA Y, HIRAMOTO M. IMAI T. Estrogen induces estrogen receptor α expression and hepatocyte proliferation in late pregnancy [J]. Biochemical & Biophysical Research Communications, 2019, 511(3): 592-596. [15] PROSSNITZ E R. GPER modulators: Opportunity nox on the heels of a class akt [J]. The Journal of Steroid Biochemistry and Molecular Biology, 2018, 176: 73-81. doi: 10.1016/j.jsbmb.2017.03.005 [16] CAO L Y, REN X M, LI C H, et al. Bisphenol AF and bisphenol B exert higher estrogenic effects than bisphenol A via G protein-coupled estrogen receptor pathway [J]. Environmental Science & Technology, 2017, 51(19): 11423-11430. [17] RUSSO D P, ZORN K M, CLARK A M, et al. Comparing multiple machine learning algorithms and metrics for estrogen receptor binding prediction [J]. Molecular Pharmaceutics, 2018, 15(10): 4361-4370. doi: 10.1021/acs.molpharmaceut.8b00546 [18] GRISONI F, CONSONNI V. BALLABIO D. Machine learning consensus to predict the binding to the androgen receptor within the compara project [J]. Journal of Chemical Information and Modeling, 2019, 59(5): 1839-1848. doi: 10.1021/acs.jcim.8b00794 [19] SHEFFIELD T, JUDSON R. Ensemble QSAR modeling to predict multispecies fish toxicity lethal concentrations and points of departure [J]. Environmental Science & Technology, 2019, 53(21): 12793-12802. [20] WANG Z, CHEN J, HONG H. Applicability domains enhance application of PPARγ agonist classifiers trained by drug-like compounds to environmental chemicals [J]. Chemical Research in Toxicology, 2020, 33(6): 1382-1388. doi: 10.1021/acs.chemrestox.9b00498 [21] ARNATT C K, ZHANG Y. G protein-coupled estrogen receptor (GPER) agonist dual binding mode analyses toward understanding of its activation mechanism: a comparative homology modeling approach [J]. Molecular Informmatics, 2013, 32(7): 647-658. doi: 10.1002/minf.201200136 [22] WANG D, HU L, ZHANG G, et al. G protein-coupled receptor 30 in tumor development [J]. Endocrine, 2010, 38(1): 29-37. doi: 10.1007/s12020-010-9363-z [23] BARTON M, PROSSNITZ E R. Emerging roles of GPER in diabetes and atherosclerosis [J]. Trends in Endocrinology & Metabolism, 2015, 26(4): 185-192. [24] PROSSNITZ E, ARTERBURN J. International union of basic and clinical pharmacology. XCVII. G protein-coupled estrogen receptor and its pharmacologic modulators [J]. Pharmacological Reviews, 2015, 67(3): 505-540. doi: 10.1124/pr.114.009712 [25] KHAN S U, AHEMAD N, CHUAH L H, et al. Sequential ligand and structure-based virtual screening approach for the identification of potential G protein-coupled estrogen receptor-1 (GPER-1) modulators [J]. Rsc Advances, 2019, 9(5): 2525-2538. doi: 10.1039/C8RA09318K [26] LAPPANO R, ROSANO C, PISANO A, et al. A calixpyrrole derivative acts as an antagonist to GPER, a G-protein coupled receptor: mechanisms and models [J]. Disease models & mechanisms, 2015, 8(10): 1237-1246. [27] SARMIENTO V, SANCHEZ T R, ULLOA A M, et al. Synthesis of novel (-) -epicatechin derivatives as potential endothelial GPER agonists: Evaluation of biological effects [J]. Bioorganic & Medicinal Chemistry Letters, 2018, 28(4): 658-663. [28] CAO L Y, REN X M, YANG Y, et al. Hydroxylated polybrominated biphenyl ethers exert estrogenic effects via non-genomic G protein-coupled estrogen receptor mediated pathways [J]. Environmental Health Perspectives, 2018, 126(5): 057005. doi: 10.1289/EHP2387 [29] O’DEA A, SONDERGARD C, SWEENEY P, et al. A series of indole-thiazole derivatives act as GPER agonists and inhibit breast cancer cell growth [J]. Acs Medicinal Chemistry Letters, 2018, 9(9): 901-906. doi: 10.1021/acsmedchemlett.8b00212 [30] ALBANITO L, LAPPANO R, MADEO A, et al. Effects of atrazine on estrogen receptor α and G protein-coupled receptor 30-mediated signaling and proliferation in cancer cells and cancer-associated fibroblasts [J]. Environmental Health Perspectives, 2015, 123(5): 493-499. doi: 10.1289/ehp.1408586 [31] REVANKAR C M, BOLOGA C G, PEPERMANS R A, et al. A selective ligand for estrogen receptor proteins discriminates rapid and genomic signaling [J]. Cell Chemical Biology, 2019, 26(12): 1692-1702. doi: 10.1016/j.chembiol.2019.10.009 [32] CIRILLO F, LAPPANO R, BRUNO L, et al. AHR and GPER mediate the stimulatory effects induced by 3-methylcholanthrene in breast cancer cells and cancer-associated fibroblasts (CAFs) [J]. Journal of Experimental & Clinical Cancer Research, 2019, 38(1): 335-353. [33] ALDO M U, DAVID M L, ERNESTO B R, et al. The effects of (-)-epicatechin on endothelial cells involve the G protein-coupled estrogen receptor (GPER) [J]. Pharmacological Research, 2015, 100: 309-320. doi: 10.1016/j.phrs.2015.08.014 [34] MAGGIOLINI M, SANTOLLA M, AVINO S, et al. Identification of two benzopyrroloxazines acting as selective GPER antagonists in breast cancer cells and cancer-associated fibroblasts [J]. Future medicinal chemistry, 2015, 7(4): 437-448. doi: 10.4155/fmc.15.3 [35] ROSANO C, PONASSI M, SANTOLLA M F, et al. Macromolecular modelling and docking simulations for the discovery of selective GPER ligands [J]. The AAPS Journal, 2016, 18(1): 41-46. doi: 10.1208/s12248-015-9844-3 [36] DELEON C, WANG H, GUNN J, et al. A novel GPER antagonist protects against the formation of estrogen-induced cholesterol gallstones in female mice [J]. Journal of Lipid Research, 2020, 61(5): 767-777. doi: 10.1194/jlr.RA119000592 [37] NAYAK T K, DENNIS M K, RAMESH C, et al. Influence of charge on cell permeability and tumor imaging of GPR30-targeted 111in-labeled nonsteroidal imaging agents [J]. ACS Chemical Biology, 2010, 5(7): 681-690. doi: 10.1021/cb1000636 [38] LUO L J, LIU F, LIN Z K, et al. Genistein regulates the IL-1 beta induced activation of MAPKs in human periodontal ligament cells through G protein-coupled receptor 30 [J]. Archives of Biochemistry and Biophysics, 2012, 522(1): 9-16. doi: 10.1016/j.abb.2012.04.007