• <tr id="yyy80"></tr>
  • <sup id="yyy80"></sup>
  • <tfoot id="yyy80"><noscript id="yyy80"></noscript></tfoot>
  • 99热精品在线国产_美女午夜性视频免费_国产精品国产高清国产av_av欧美777_自拍偷自拍亚洲精品老妇_亚洲熟女精品中文字幕_www日本黄色视频网_国产精品野战在线观看 ?

    Identification of prognostic markers by integrating the genome and transcri ptomics in ovarian cancer

    2019-11-01 03:01:14XinWuDongXinTangXiaNanSangKuiLongWangMinHaoGangCao
    TMR Modern Herbal Medicine 2019年4期

    Xin Wu, Dong-Xin Tang,Xia-Nan Sang,Kui-Long Wang,Min Hao,Gang Cao*

    1College of Pharmaceutical Sciences,Zhejiang Chinese Medical University,Hangzhou,Zhejiang,China.

    2First Affiliated Hospital of Guiyang College of Traditional Chinese Medicine,Guiyang,Guizhou,China.

    Abstract

    Keywords:Ovarian cancer,TCGA,GEO,5-gene signature,Prognostic marker

    Background

    Ovarian cancer is one of the common malignant tumors that seriously threaten the physical and mental health of women.Early ovarian cancer is so hidden that patients often do not have obvious symptoms or signs.Most patients are already at an advanced stage of cancer when they find themselves have ovarian cancer.In recent years,there has been no substantial reduction in the incidence and mortality of ovarian cancer[1].Therefore,the early diagnosis and treatment of ovarian cancer has become a topic of widespread concern.At present,the first-line treatment of ovarian cancer is tumor cell reduction and platinum-based chemotherapy[2].Although most patients are sensitive to the initial treatment,more than 60%of the patients still have tumor recurrence,metastasis and eventually die of ovarian cancer after the initial treatment[3].In order to improve the prognosis of ovarian cancer,a growing number of researchers shift their research focus to molecular biomarkers,such as genetic mutations and immune checkpoints,expecting to achieve early detection and treatment of ovarian cancer[4,5].

    In recent years,researchers have conducted many studies related to ovarian cancer biomarkers by using mass spectrometry and other techniques,and have found a large number of biomarkers[6].For example,CA125 is a conventional serum marker for ovarian cancer,but its specificity is limited.It is commonly used to monitor treatment responses and detect disease recurrence[7].Pratet al[8]evaluated the prognostic effect of the lowest value of CA125 within the reference range(<35U/ml).After initial treatment,the absolute value of CA125 increase≥5U/ml,which is a strong predictive indicator of ovarian cancer recurrence.Human epididymis 4(HE4)is also one of the common biomarkers of ovarian cancer.A study by Anastasiet al[9]shows that in patients with ovarian cancerrecurrence,the expression of HE4 increases in 5 to 8 months before CA125,suggesting that HE4 could be a good early marker for monitoring ovarian cancer recurrence.Kuket al[10]first reported in 2010 that Nidogen-2 was a new serum biomarker for ovarian cancer that can be used for early diagnosis.They studied 100 patients with ovarian cancer,100 patients with benign gynecological diseases and 100 healthy control samples.The results show that the average levels of serum Nidogen-2 in normal control group and benign gynecological disease group are 13.2mg/L and 12.1 mg/L,respectively.There was no significant difference between the two groups.However,the average concentration of serum Nidogen-2 in patients with ovarian cancer was 18.6 mg/L,which was significantly different from that in normal control group or benign gynecological disease group(P<0.001).In all samples,CA125 and Nidogen-2 concentrations were found to be closely related(p<0.001).This suggested that serum Nidogen-2 may be a new serum marker of ovarian cancer.

    Although hundreds of biomarkers have been found to be associated with ovarian cancer,the challenges for early diagnosis of ovarian cancer remain due to the limited sensitivity and specificity ofsingleovarian cancer markers.Their use in early diagnosis has not yet been determined [11].High throughput multi-omics sequencing data has laid a solid foundation for the identification of genes related to cancer prognosis.Multi-omics data analysis can reveal the mechanism of cancer occurrence and development from many aspects[12,13].Based on genomics data of GEO database,Li,Zhi[14]found that four gene signatures were closely related to the drug sensitivity of paclitaxel in the treatment of breast cancer through logistic regression,cox regression and correlation analysis with bootstrapping.In short,it is pretty important to integrate multi-omics data in bioinformatics to identify the gene signature associated with the prognosis of ovarian cancer.

    Therefore,in this paper,we propose a method to integrate multi-omics data to identify biomarkers related to ovarian cancer.First,gene expression data,single nucleotide mutation and copy number variation data of ovarian cancerpatientsare obtained from TCGA(https://www.cancer.gov/tcga) and GEO(https://www.ncbi.nlm.nih.gov/gds/) databases.The 5-gene signature,internal test set and external verification set,most relevant to prognosis,have been obtained through the screening ofintegrated genomics and transcripto me data,which have proved their ability to predict the prognosis of patients.5-gene signature shows a strong clinical independence.The results of GSEA analysis also show that the pathway of 5-gene signature enrichment is significantly related to the pathway and biological process of the occurrence and development of ovarian cancer.It suggests that the model has potential clinical application value and can provide potential therapeutic targets for the diagnosis of clinical patients.

    Methods

    Data download and preprocessing

    We download RNA-Seq data of TCGA,clinical follow-up information and copy number variation data of SNP6.0 chip from UCSC cancer browser(https://xenabrowser.net/datapages/).The mutation annotation file(MAF)is downloaded from the GDC client.The GSE17260 expression profile data and clinical follow-up information are downloaded from the GEO database.We obtain the TCGA training set and the test set,and scale the samples.The TCGA training set contains 189 samples,the test set contains 189 samples,and the GSE17260 contains 110 samples.We obtain the specific distribution of patient age,survival status,staging,lymphatic invasion,venous invasion and Tumor Stage in three data sets(Supplement Table 1).

    Univariate Cox proportional risk regression analysi s

    As Jin Chenget al[15]did a univariate Cox proportional risk regression analysis of each immune gene,we screen out the genes that are significantly related to the patient OS in the training data set.p<0.01 is selected as the threshold.

    Analysis of copy number variation data

    GISTIC is widely used and detects both broad and focal(potentially overlapping)recurring events.We use the GISTIC 2.0 [16]soft wareto identify genes with significanta mplification or deletion.The parameter thresholdsare the fragments with amplification or deletion length greater than 0.1 and p<0.05.

    Gene mutation analysis

    In order to identify the significantly mutated genes,we use the Mutsig 2.0 software to identify the significantly mutated genes in the maf file of the TCGA mutation data,with a threshold of p<0.05.

    Construction of prognostic immune gene signature

    We choose the genes that are significantly related to patient’soverallsurvival(OS)and genes thatare amplified,deleted,and mutated.Random survival forest algorithm is further used to rank the importance of prognostic genes [17].We use R package random Survival Forest[18]to screen genes,setting Monte Carlo iterations to 100 and forward steps to 5.The genes with relative importance greater than 0.4 are identified as feature genes.The multivariate Cox regression analysis is further carried out,and the following risk scoring models are constructed:

    In the formula,N represents the number of prognostic immune genes.Expk represents the expression value of prognosticimmune genes,which istheestimated regression coefficient of immune genes in the multivariate Cox regression analysis.

    Functional enrichment analysis

    Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)pathway enrichment analysis are performed using the R package clusterprofiler[19],to identify over-represented GO terms in three categories(biological processes,molecular function and cellular component),and KEGG pathway.Inthisanalysis,FDR < 0.05 is considered to have statistical significance.

    GSEA analysis

    GSEA[20]analysis is performed by the JAVA program(http://software.broadinstitute.org/gsea/downloads.jsp)

    using the MSigDB[21]C2 Canonical pathways gene set collection,which contains 1320 gene sets.Gene sets with the false discovery rate(FDR)value less than 0.05 after performing 1000 permutations are considered to be significantly enriched.Statistical analysis

    The Kaplan-Meier(KM)curve is drawn when using the median risk score in each dataset as cutoff to compare the survival risk between high-risk group and low-risk group.Multivariate Cox regression analysis is performed to test whether genetic markers are independent prognostic factors.The significance is defined as P<0.05.All of these analyses are carried out in R 3.4.

    Results

    Prognostic genes obtained from preliminary analysi s of multi-omics data

    For the samples of TCGA training set,we establish the relationship between patients’overall survival(OS)and gene expression with univariate regression analysis,identifying 2097 genes of univariate cox regression with p value less than 0.05 as candidate prognostic genes.The first 20 genetic information of the 2097 candidate prognostic genes are shown in Table 1.

    Table 1.Information of top 20 prognosis-related genes

    Gene HR coefficient Z-score P-value ENSG00000114742 0.6710 -0.3990 -3.7834 0.0002 ENSG00000064270 0.6529 -0.4263 -3.7722 0.0002 ENSG00000276612 1.3696 0.3145 3.7371 0.0002

    Copy number variation data analysis

    For the copy number variation data in TCGA,we use GISTIC 2.0 to identify the genes with significant amplification ordeletion.Parame terthresholds are fragments with amplification or deletion lengths greater than 0.1 and p<0.05.The significantly amplified fragments of ovarian cancer genome are shown in Figure 1A.Supplement Table 2 records the genes that are significantly amplified on each fragment.For example,CCNE1 on 19q12 is significantly amplified(q value=2.06E-98),KRAS on the 12p12.1 issignificantly amplified(q value=2.64E-18),and CD7 on 17q25.3 is significantly amplified(q value=0.00080632).A total of 447 geneshavea mplification.The fragmentswith significant deletion in the ovarian cancer genome are shown in Figure 1B.Supplement Table 3 records the genes that have significant deletion on each fragment.For example,there is a significant deletion of CCNB1 on 5q13.3(q value=2.44E-58),a significant deletion of RB1 on 13q14.2(q value=2.06E-10),and a significant deletion of CDKN2A on 9p21.3(q value=0.04325).A total of 1069 genes have gene deletions.

    Table 2.Five genes significantly associated with the overall survival in the training set

    Figure 1.A:Significantly amplified fragments in the ovarian cancer genome;B:Significantly deleted fragments in the ovarian cancer genome

    Table 3.Identification of clinical factors and clinical independence associated with prognosis by univariate andmultivariate COX regression analysis

    Mutation Data analysis

    For the mutation annotation data of TCGA,we use Mutsig2 to identify the genes with significant mutations,selecting p<0.05 as a threshold.A total of 654 genes are obtained which have significant mutation frequencies,as shown in mutation_gene.txt.Figure 2 shows the distribution of synonymous mutations,missense mutations,frame insertion or deletion,frame movements,nonsense mutations,shear locis,and other non-synonymous mutations in the 50 genes with the most significant p values in TCGA ovarian cancer samples.We identify that some of the 654 genes are closely related to the previous studies that have been reported to be related to the occurrence and development of cancer,such as RB1,TP53,BRCA1,and CDK12 etc.

    Figure 2.Distribution of 50 genes with the most significant P value in patients with ovarian cancer

    Pathways and biological processes involved in copy number variation genes and mutant genes

    By identifying amplification and deletion genes with TCGA copy number variation,as well as mutant gene integration,we identify a total of 1946 genes involved in biological processes and pathways.In Figure 3A,1946 genes are significantly enriched in the pathways of occurrence and development of cancer,including Hepatocellular carcinoma,Prostate cancer,Chronic myeloid leukemia,and B cell receptor signaling pathway.In Figure 3B,1946 genes are significantly enriched in developmental process,positive regulation of biological process,positive regulation of cellular process,regulation of signal transduction and other biological processes related to the occurrence and development of cancer.

    Prognostic genes sequencing of the importance of random survival forest

    We identify 2097 candidateprog nosticgenes with amplification,deletion and mutation,and obtain 158 genes.We use random survival forest algorithm to rank the importance of prognostic genes (R package random Survival Forest),setting parameter as nrep=100 and nstep=5,which means that the Monte Carlo iteration is 100,and forward step is 5.We identify the genes with a relative importance greater than 0.5 as the final signature.Figure 4A shows the relationship between error rate and the number of classification trees.Figure 4B shows the importance sequencing of the first five genes out-of-bag.

    Figure 3.A.KEGG pathways participated by 1946 genes with copy number variation and mutation;B.Biological processes involving 1946 genes with copy number variation and mutation

    Figure 4.A.The relationship between error rate and the number of classification trees;B.Importance sequencing of the first five genes out-of-bag

    Establishment of 5-gene signature to divide sample s in TCGA training set

    As for the identified 5-gene signature mentioned above,Table 2 lists the HR,Z score and p value of univariate regression,as well as the importance and relative importance of the first five genes out-of-bag.

    Then,the 5-gene signature is established using the multivariate COX regression analysis method.The model is established as follows:

    The scoring formula for each sample is the sum of the*ordinal numbers of the above gene expression values.Then we select the median score of thesample-0.0002004649 as cutoff,to divide the samples into high-risk group and low-risk group.Figure 5A shows the classification effect in the TCGA training set.95 patients are divided into the low risk group and 94 patients are divided into the high risk group.There are significant differencesbet ween the two groups.Log-rank p=5.977197e-07.Figure 5B shows the ROC curve,in which the AUC is 0.75.Figure 5C shows that the survival time of death samples decreases significantly as the patient's risk score increases.High-risk group has more death samples.The expression changes offive different signature genes with the increase of risk value show that the high expression of PSMB1,COL6A6,and SLC22A2 is related to high risk of them.These three genes are risk factors.The high expression of KLHL23 and CD3G is related to the low risk of them.These two genes are protective genes.

    Figure 5.A.KM survival curve distribution of 5-gene signature in TCGA training se;B.ROC Curve and AUC of 5-gene signature classification;C.The risk score,survival time,survival status,and the expression of five genes in TCGA training set

    Robustness detection of 5-gene signature in TCGA test set

    In order to determine the robustness of the model,we use the same model and the same cutoff as the TCGA training set,and conduct verification in the TCGA test set.Figure 6A shows the classification effect in TCGA test sets.100 patients are divided into the low risk group and 89 patients are divided into the high risk group.There are significant differences between the two groups.Log-rank p=0.01690201.Figure 6B shows the ROC curve,in which the AUC is 0.68.Figure 6C produces results similar to the TCGA training set that the survival time of death samples decreases significantly as the risk value increases.High-risk group has more death samples.The expression changes of five different signature genes with the increase of risk value show that the high expression of PSMB1,COL6A6,and SLC22A2 is related to high risk of them.These three genes are risk factors.The high expression of KLHL23 and CD3G is related to the low risk of them.These two genes are protective genes.

    Verifying the robustnessof5-gene signature in external independent dataset GSE17260

    Similarly,we verify the robustness of the model in the external independent dataset GSE17260 using the same model and the same cutoff as the TCGA training set.Figure 7A shows the classification effect in GSE17260.44 patients are divided into the low risk group and 37 patients are divided into the high risk group.There are significant differences between the two groups.Log-rank p=0.03500249.Figure 7B shows the ROC curve,in which the AUC is 0.68.Figure 7C produces results similar to the TCGA training set that the survival time of death samples decreases significantly as the risk value increases.High-risk group has more death samples.The expression changes of five different signature genes with the increase of risk value show that the high expression of PSMB1,COL6A6,and SLC22A2 is related to high risk of them.These three genes are risk factors.The high expression of KLHL23 and CD3G is related to the low risk of them.These two genes are protective genes.

    Analysis of clinicalin dependence of the 5-gene signature model

    In order to identify the independence of 5-gene signature model in clinical application,we have used univariate and multivariate COX regression to analyze the relevant HR,95%CI of HR,and p value in the clinical information carried by TCGA training set,TCGA test set and GSE17260 data.We systematically analyze age,stage,lymphatic invasion,venous invasion,tumor stage and grouping information Table 3 of the 5-gene signature of patients recorded in TCGA and GSE17260.

    In the TCGA training set,univariate COX regression analysis has found that high risk group,age,and clinical stage II are significantly related to survival.However,the corresponding multivariate COX regression analysis has showed that only high risk group(HR=4.63,95%CI=1.37-15.54,p=0.013),venous invasion YES(HR=0.09,95%CI=0.01-0.73,p=0.025),Clinical stage II(HR=0.04,95%CI=0.001-0.82,p=0.0372)and clinical stage III(HR=0.05,95%CI=0.004-0.62,p=0.02)have clinical independence.

    In the TCGA test set,univariate COX regression analysis has found that high risk group is significantly related to survival.However,the corresponding multivariate COX regression analysis has found that none of the factors have clinical independence.High risk group corresponds to HR=2.094,95%CI=0.72-6.01,and p=0.170.

    In GSE17260,univariate COX regression analysis has found that high risk group is significantly related to survival.The corresponding multivariate COX regression analysis shows that high risk group(HR=2.12,95%CI=1.01-4.45,p=0.046)has clinical independence.

    The above conditions suggest that our 5-gene signature model is a prognostic indicator independent of other clinicalfactors,which has independent predictive performance in terms of clinical application.

    Figure 6.A.KM survival curve distribution of 5-gene signature in TCGA test set;B.ROC Curve and AUC of 5-gene signature classification;C.The risk score,survival time,survival status,and the expression of five genes in TCGA training set

    Figure 7.A.KM survival curve distribution of 5-gene signature in GSE17260;B.ROC Curve and AUC of 5-gene signature classification;C.The risk score,survival time,survival status,and the expression of five genes in GSE17260

    Analysisofpathwaysenriched in high-risk and low-risk groups using GSEA

    We use GSEA in TCGA training set to analyze the significantly enriched pathways in high-risk and low-risk groups.The selected gene set is c2.cp.kegg.v6.0.symbols,which includes the pathway of KEGG.The GSEA input file contains the standardized expression profile data of the TCGA training set.The sample label marked by 5-gene signature marks the sample as a high-risk group and a low-risk group.The selected threshold for enriched pathway is 0.05.Then we get the significantly enriched pathway(Table 4).

    Forexample,KEGG COLORECTALCANCER,KEGGNOTCH SIGNALINGPATHWAY,KEGG SMALL CELL LUNG CANCER,and KEGG OXIDATIVE PHOSPHORYLATION,which are significantly enriched in the high-risk group in Figure 8,are significantly related to the occurrence,development and metastasis of ovarian cancer.

    Table 4.GSEA analysis of significantly enriched KEGG pathways in high-risk and low-risk groups

    Discussion

    Ovarian cancer is a highly heterogeneous disease in prognosis.Different types of ovarian tumors,which occur in different populations,may have significant differences in prognosis[22,23].It involves the pathology type of tumor,biochemical function,drug tolerance,migration ability and invasion ability,which has a great influence on the treatment choice and prognosis of patients.Therefore,screening prognostic biomarkers that fully reflect the biological characteristics of ovarian cancer is of great significance for individualized prevention and treatment of ovarian patients[24-26].In this study,we have analyzed the expression profiles of 488 ovarian cancer samples from TCGA database and GEO database.We identify five gene signatures associated with prognosis and they have showed strong clinical independence.

    At present,gene signature has been used in clinical research.For example,Liu,Ronget al[27]analyzed the lncRNA expression profile of 258 ovarian cancer patients in TCGA database,predicting the long non-coding RNA expression characteristics of platinum-based chemotherapy sensitivity in ovarian cancer patients,which is used to explore the biomarkers associated with drug resistance of ovarian cancer.These results have shown that screening new prognostic markers in cancer with gene expression profiles has become the most promising method for high-throughput molecular identification.Meng,Xu[28]used multi-omics data,including protein-coding genes,long non-coding RNA and small RNA data to identify disease-related genes to identify prognostic markersassociated with ovarian cancer.It was found that the maximum area under the ROC curve(AUC)constructed by prognostic genes was 0.69 in the training group and 0.62 in the test group.The AUC of our 5-gene signature in both the training set and the verification set is close to 0.70.In addition,our 5-gene signature based on multi-omics data recognition has strong robustness.It has been proved to have a stable prediction performance in the data sets of different platforms such as TCGA and GEO.The five gene signatures have strong clinical independence,which can maintain stable clinical independence under the influence of many clinical factors.

    Among the 5 gene signatures,PSMB1,COL6A6,and SLC22A2 are risk factors,and KLHL23 and CD3G are protective factors.It has been reported that high COL6A6 expression level is significantly correlated with the early pathological stage of breast cancer(P=0.003)[29],and the KLHL23 expression level is closely related to gastric cancer[30].Huang,Poyin [31]established a risk assessment model using six genes,including KLHL23,to predict the molecular genetic characteristics and prognosis of non-small cell lung cancer.It has been reported that CD3G is a potential biomarker of colorectal cancer risk[32],and the genetic polymorphism of CD3G may be associated with susceptibility to liver cancer[33].These genes are closely related to the risk or prognosis of tumors.PSMB1 and SLC22A2 have not been reported to be associated with tumors,while in this study they are found as new prognosis markers for ovarian cancer for the first time.At the same time,our GSEA analysis shows that the enriched pathways in the high and low risk groups obtained by 5-gene signature are significantly related to the occurrence and development of ovarian cancer.These resultsshow thatthemodelhaspotentialclinical application value,which can provide potential therapeutic targets for the diagnosis of clinical patients.

    Figure 8.The enriched pathways in the high risk group obtained by 5-gene signature

    Although we can identify potential candidate genes for tumor prognosis based on multi-omics data through bioinformatics,we should pay attention to some limitations of this study.First,this study has high requirements for the amount of sample data.Although TCGA,GEO and other databases provide data support for our research,only some cancers can obtain sufficient data such as gene expression data,single nucleotide mutation,and copy number variation data etc.There are also problems,such as insufficient sample size,when it is used in other diseases.Secondly,the results obtained only through bioinformatics analysis are inadequate,requiring follow-up pharmacological experiments to further confirm them.Last but not least,this study has high requirements for bioinformatics and other professional knowledge.There are still some technical difficulties in its promotion in clinical use.

    In conclusion,we propose a method to integrate multi-omics data to identify biomarkers related to ovarian cancer in this paper.We finally identify 5-gene signature,and construct a prognostic assessment model for patients.The model has a good AUC in both test sets and external independent datasets.5-gene signature shows strong clinical independence.Compared with clinical characteristics,the model can better predict the survival risk of patients,which proves its potential clinical application value and provides new biomarkers for the diagnosis of clinical patients.

    Conclusion

    In conclusion,weproposeamethod to integrate multi-omics data to identify biomarkers related to ovarian cancer in this paper.We finally identify 5-gene signature,and construct a prognostic assessment model for patients.The model has a good AUC in both test sets and external independent datasets.5-gene signature shows strong clinical independence.Compared with clinical characteristics,the model can better predict the survival risk of patients,which proves its potential clinical application value and provides new biomarkers for the diagnosis of clinical patients.

    99久久人妻综合| 肉色欧美久久久久久久蜜桃| 女人高潮潮喷娇喘18禁视频| 亚洲欧美日韩另类电影网站| 午夜久久久在线观看| 一区二区av电影网| 免费人妻精品一区二区三区视频| 成人免费观看视频高清| 精品人妻一区二区三区麻豆| 女人被躁到高潮嗷嗷叫费观| 老司机靠b影院| 欧美日韩精品网址| 精品少妇久久久久久888优播| 2018国产大陆天天弄谢| av在线观看视频网站免费| 中文天堂在线官网| 国产成人精品久久二区二区91 | 9热在线视频观看99| 久久久久国产一级毛片高清牌| 久久精品人人爽人人爽视色| 韩国高清视频一区二区三区| 中国国产av一级| 国产一区二区三区av在线| 美女中出高潮动态图| 成年人午夜在线观看视频| 成人国产麻豆网| 国产精品一区二区在线观看99| 日韩av免费高清视频| 日韩制服丝袜自拍偷拍| 国产成人啪精品午夜网站| 人体艺术视频欧美日本| 亚洲av综合色区一区| 午夜激情久久久久久久| 国产成人精品无人区| 久久婷婷青草| 999久久久国产精品视频| 精品国产超薄肉色丝袜足j| 亚洲欧美精品综合一区二区三区| 成年人午夜在线观看视频| 一级黄片播放器| 97精品久久久久久久久久精品| 丝袜脚勾引网站| 国产精品国产三级国产专区5o| 中文乱码字字幕精品一区二区三区| 狂野欧美激情性bbbbbb| 亚洲精品日本国产第一区| 久久久久国产精品人妻一区二区| 老司机亚洲免费影院| 亚洲精品国产av成人精品| 97精品久久久久久久久久精品| 性色av一级| 丰满饥渴人妻一区二区三| 韩国av在线不卡| 亚洲熟女毛片儿| 久久久久久久国产电影| 极品少妇高潮喷水抽搐| 十八禁高潮呻吟视频| 91成人精品电影| 高清在线视频一区二区三区| 国产在线一区二区三区精| av在线老鸭窝| 巨乳人妻的诱惑在线观看| 亚洲视频免费观看视频| 秋霞伦理黄片| 天天躁狠狠躁夜夜躁狠狠躁| 国产日韩欧美亚洲二区| 日韩视频在线欧美| 18禁国产床啪视频网站| 欧美黑人欧美精品刺激| 成年人免费黄色播放视频| 久久女婷五月综合色啪小说| 亚洲av欧美aⅴ国产| 男人操女人黄网站| 99久久精品国产亚洲精品| 午夜激情久久久久久久| 丰满迷人的少妇在线观看| 国产成人免费无遮挡视频| 久久精品国产综合久久久| 日韩免费高清中文字幕av| 热99国产精品久久久久久7| 一级a爱视频在线免费观看| 日韩一区二区三区影片| 国产亚洲一区二区精品| 国产亚洲一区二区精品| www.自偷自拍.com| 国产乱人偷精品视频| 精品国产一区二区三区四区第35| 国产精品国产av在线观看| 男女床上黄色一级片免费看| 叶爱在线成人免费视频播放| 十八禁人妻一区二区| 菩萨蛮人人尽说江南好唐韦庄| 又大又黄又爽视频免费| 国产熟女午夜一区二区三区| 久久精品久久久久久噜噜老黄| 亚洲国产av影院在线观看| 操美女的视频在线观看| 菩萨蛮人人尽说江南好唐韦庄| 欧美中文综合在线视频| 91老司机精品| 亚洲,一卡二卡三卡| 中文字幕精品免费在线观看视频| 欧美精品亚洲一区二区| 999精品在线视频| 日日啪夜夜爽| 亚洲国产毛片av蜜桃av| 久久久精品免费免费高清| 日本黄色日本黄色录像| 色94色欧美一区二区| 国产免费福利视频在线观看| 男女边摸边吃奶| av福利片在线| 国产男女内射视频| 美女扒开内裤让男人捅视频| 亚洲欧洲精品一区二区精品久久久 | 欧美激情 高清一区二区三区| 亚洲男人天堂网一区| 波野结衣二区三区在线| 国产精品久久久久久人妻精品电影 | 欧美老熟妇乱子伦牲交| av视频免费观看在线观看| 亚洲第一青青草原| www.av在线官网国产| 国产伦人伦偷精品视频| 一级,二级,三级黄色视频| 国产一级毛片在线| 久久精品久久久久久久性| 亚洲伊人色综图| 一区二区日韩欧美中文字幕| 国产一卡二卡三卡精品 | 青草久久国产| 极品人妻少妇av视频| 亚洲精品国产一区二区精华液| 久久久精品94久久精品| 制服人妻中文乱码| 日日爽夜夜爽网站| 9191精品国产免费久久| 国产 精品1| 国产欧美日韩一区二区三区在线| 无遮挡黄片免费观看| 精品一区二区三卡| 亚洲,欧美精品.| 叶爱在线成人免费视频播放| 伊人亚洲综合成人网| 在线观看免费午夜福利视频| av网站在线播放免费| 黑人欧美特级aaaaaa片| 免费日韩欧美在线观看| 99热全是精品| videos熟女内射| 另类亚洲欧美激情| 亚洲中文av在线| 精品少妇黑人巨大在线播放| 成年人免费黄色播放视频| 亚洲七黄色美女视频| 青草久久国产| 精品久久蜜臀av无| 啦啦啦中文免费视频观看日本| 王馨瑶露胸无遮挡在线观看| a 毛片基地| 女人被躁到高潮嗷嗷叫费观| 日韩精品免费视频一区二区三区| 精品人妻熟女毛片av久久网站| 欧美变态另类bdsm刘玥| 日韩视频在线欧美| av线在线观看网站| 亚洲欧美精品综合一区二区三区| 2021少妇久久久久久久久久久| 人人澡人人妻人| 大码成人一级视频| 国产熟女午夜一区二区三区| 天天躁夜夜躁狠狠久久av| 久久国产精品大桥未久av| 欧美97在线视频| 天天躁夜夜躁狠狠躁躁| 国产乱来视频区| 成年人午夜在线观看视频| 99久久99久久久精品蜜桃| 亚洲av电影在线观看一区二区三区| 国产成人欧美在线观看 | 日本黄色日本黄色录像| 99精国产麻豆久久婷婷| 久久天躁狠狠躁夜夜2o2o | 十分钟在线观看高清视频www| 久久久久精品国产欧美久久久 | 国产日韩欧美在线精品| av视频免费观看在线观看| 一级毛片黄色毛片免费观看视频| 只有这里有精品99| 美女扒开内裤让男人捅视频| 亚洲精品国产av成人精品| 国产成人欧美| 97在线人人人人妻| 欧美人与善性xxx| 这个男人来自地球电影免费观看 | 久久人人爽人人片av| 免费黄频网站在线观看国产| 国产在线视频一区二区| 男女边摸边吃奶| 大片电影免费在线观看免费| 久久精品国产综合久久久| 欧美亚洲日本最大视频资源| 国产av一区二区精品久久| 中文字幕人妻丝袜制服| 亚洲欧美成人综合另类久久久| 人人澡人人妻人| 亚洲熟女毛片儿| 久久精品aⅴ一区二区三区四区| 天天添夜夜摸| 欧美日韩福利视频一区二区| 免费av中文字幕在线| 日韩免费高清中文字幕av| 男人舔女人的私密视频| 深夜精品福利| 伊人亚洲综合成人网| 一级毛片 在线播放| 精品第一国产精品| 97人妻天天添夜夜摸| 如何舔出高潮| 久久婷婷青草| 一级毛片黄色毛片免费观看视频| 大陆偷拍与自拍| 99国产综合亚洲精品| 国产成人精品久久二区二区91 | 午夜av观看不卡| 成人黄色视频免费在线看| 一个人免费看片子| 一本大道久久a久久精品| 97人妻天天添夜夜摸| 国产精品久久久久久精品电影小说| 欧美乱码精品一区二区三区| 欧美日韩国产mv在线观看视频| 亚洲精品久久午夜乱码| 国产亚洲午夜精品一区二区久久| 日韩大片免费观看网站| 一本—道久久a久久精品蜜桃钙片| 少妇人妻 视频| 色精品久久人妻99蜜桃| 99精国产麻豆久久婷婷| 久久精品人人爽人人爽视色| 90打野战视频偷拍视频| 超色免费av| 观看美女的网站| 九色亚洲精品在线播放| 在线观看一区二区三区激情| www.av在线官网国产| 国产熟女午夜一区二区三区| 天美传媒精品一区二区| 亚洲精品在线美女| av视频免费观看在线观看| 亚洲一码二码三码区别大吗| 视频区图区小说| 在线观看免费高清a一片| 国产爽快片一区二区三区| 欧美日本中文国产一区发布| 久久99一区二区三区| 欧美中文综合在线视频| 水蜜桃什么品种好| 无遮挡黄片免费观看| 亚洲成人av在线免费| 国产成人免费观看mmmm| 丰满少妇做爰视频| 新久久久久国产一级毛片| 天堂中文最新版在线下载| 中国三级夫妇交换| 啦啦啦中文免费视频观看日本| 丰满饥渴人妻一区二区三| 久久久久精品人妻al黑| 中文字幕人妻熟女乱码| 精品少妇久久久久久888优播| 另类精品久久| 国产成人午夜福利电影在线观看| 青草久久国产| 中文字幕av电影在线播放| 国产爽快片一区二区三区| 亚洲人成77777在线视频| 极品人妻少妇av视频| av电影中文网址| 丝袜喷水一区| 一级毛片黄色毛片免费观看视频| 日韩大片免费观看网站| 国产精品嫩草影院av在线观看| 精品亚洲成国产av| 精品国产一区二区三区久久久樱花| 久久久精品国产亚洲av高清涩受| 久久青草综合色| 中文字幕亚洲精品专区| 久久天堂一区二区三区四区| 亚洲人成电影观看| 国产高清国产精品国产三级| 亚洲精品第二区| 老汉色av国产亚洲站长工具| 别揉我奶头~嗯~啊~动态视频 | 精品少妇黑人巨大在线播放| 一区二区三区精品91| 丰满迷人的少妇在线观看| 建设人人有责人人尽责人人享有的| 考比视频在线观看| 国产精品无大码| 亚洲精品久久午夜乱码| 女的被弄到高潮叫床怎么办| 国产在线视频一区二区| 亚洲欧洲国产日韩| 国产成人免费无遮挡视频| 伊人久久大香线蕉亚洲五| 黄色视频不卡| 精品一区二区免费观看| 老司机影院成人| 美女午夜性视频免费| 亚洲第一区二区三区不卡| 亚洲国产欧美在线一区| 亚洲精品日本国产第一区| 亚洲人成网站在线观看播放| 视频在线观看一区二区三区| 国产亚洲午夜精品一区二区久久| 国产精品99久久99久久久不卡 | 免费久久久久久久精品成人欧美视频| 在线免费观看不下载黄p国产| 少妇被粗大的猛进出69影院| 日韩av不卡免费在线播放| 国产人伦9x9x在线观看| 麻豆精品久久久久久蜜桃| 中文字幕人妻丝袜一区二区 | 亚洲欧美中文字幕日韩二区| 蜜桃在线观看..| 99精品久久久久人妻精品| av卡一久久| 十八禁高潮呻吟视频| 制服丝袜香蕉在线| 国产亚洲午夜精品一区二区久久| 中文乱码字字幕精品一区二区三区| 老熟女久久久| 一区二区日韩欧美中文字幕| 亚洲av欧美aⅴ国产| 成人影院久久| 美女视频免费永久观看网站| 波野结衣二区三区在线| 爱豆传媒免费全集在线观看| 色婷婷久久久亚洲欧美| 另类亚洲欧美激情| 日韩伦理黄色片| 夜夜骑夜夜射夜夜干| 久久天堂一区二区三区四区| 亚洲熟女精品中文字幕| 女人爽到高潮嗷嗷叫在线视频| 美女中出高潮动态图| 1024视频免费在线观看| 久久人人爽av亚洲精品天堂| 亚洲av在线观看美女高潮| 亚洲av日韩在线播放| 国产片内射在线| 亚洲国产成人一精品久久久| 男女之事视频高清在线观看 | 久久久久久免费高清国产稀缺| 成年人免费黄色播放视频| 色婷婷av一区二区三区视频| 99久国产av精品国产电影| 日韩欧美一区视频在线观看| 麻豆av在线久日| 久久午夜综合久久蜜桃| 国产精品久久久久久久久免| www.av在线官网国产| 狂野欧美激情性xxxx| 一区二区三区激情视频| 国产免费现黄频在线看| 大码成人一级视频| 亚洲av综合色区一区| 你懂的网址亚洲精品在线观看| 成人影院久久| 十八禁高潮呻吟视频| 久久免费观看电影| 国产免费视频播放在线视频| 亚洲国产毛片av蜜桃av| 只有这里有精品99| 大片免费播放器 马上看| 黄片小视频在线播放| 如日韩欧美国产精品一区二区三区| 我要看黄色一级片免费的| 国产在线视频一区二区| 可以免费在线观看a视频的电影网站 | 丰满饥渴人妻一区二区三| 亚洲人成77777在线视频| √禁漫天堂资源中文www| 韩国av在线不卡| 国产成人一区二区在线| 亚洲综合色网址| 国产精品99久久99久久久不卡 | 亚洲精品美女久久久久99蜜臀 | 91国产中文字幕| 精品久久久精品久久久| 日韩欧美一区视频在线观看| 又大又黄又爽视频免费| 国产精品久久久久久精品电影小说| 黄色一级大片看看| 777米奇影视久久| 免费看av在线观看网站| 成人亚洲欧美一区二区av| 欧美精品av麻豆av| 丝瓜视频免费看黄片| 天天躁夜夜躁狠狠久久av| 啦啦啦在线免费观看视频4| 美女午夜性视频免费| 在线观看免费午夜福利视频| 午夜激情av网站| 久久久久久久久久久免费av| 亚洲欧洲精品一区二区精品久久久 | 亚洲精品一区蜜桃| 欧美在线一区亚洲| 久久久久久久久免费视频了| 国产成人系列免费观看| 国产极品天堂在线| 亚洲婷婷狠狠爱综合网| 巨乳人妻的诱惑在线观看| 国产熟女午夜一区二区三区| 久久久久网色| 久久av网站| bbb黄色大片| 亚洲欧美激情在线| 成人亚洲欧美一区二区av| 亚洲情色 制服丝袜| 天天影视国产精品| 国产黄频视频在线观看| 热99国产精品久久久久久7| 欧美黄色片欧美黄色片| 一本—道久久a久久精品蜜桃钙片| 蜜桃国产av成人99| 国产高清不卡午夜福利| 我的亚洲天堂| 欧美97在线视频| 操美女的视频在线观看| 一区二区三区激情视频| 亚洲av日韩在线播放| 国产免费一区二区三区四区乱码| 纯流量卡能插随身wifi吗| 精品一区在线观看国产| 欧美亚洲日本最大视频资源| 中文天堂在线官网| 卡戴珊不雅视频在线播放| 波多野结衣av一区二区av| 黄频高清免费视频| 啦啦啦视频在线资源免费观看| 久久人人97超碰香蕉20202| 精品国产一区二区三区久久久樱花| 久久久久久久精品精品| 久热这里只有精品99| 成人国语在线视频| 亚洲精品中文字幕在线视频| 国产精品久久久久久人妻精品电影 | 各种免费的搞黄视频| 国产精品久久久久成人av| 考比视频在线观看| 少妇人妻 视频| 如何舔出高潮| 麻豆av在线久日| 欧美国产精品va在线观看不卡| 久久久久久久久久久免费av| 久久人人爽av亚洲精品天堂| 国产精品香港三级国产av潘金莲 | 国产精品亚洲av一区麻豆 | 亚洲人成77777在线视频| 久久人人爽av亚洲精品天堂| 日韩一区二区视频免费看| 69精品国产乱码久久久| 一边摸一边做爽爽视频免费| 肉色欧美久久久久久久蜜桃| 国产精品麻豆人妻色哟哟久久| 制服诱惑二区| 一级黄片播放器| 国产成人精品无人区| 一区二区三区乱码不卡18| 啦啦啦 在线观看视频| 91老司机精品| 国产精品国产av在线观看| 成人国产麻豆网| 欧美成人午夜精品| 在线看a的网站| 欧美 亚洲 国产 日韩一| 午夜福利免费观看在线| 日本av手机在线免费观看| 国产又爽黄色视频| 日韩不卡一区二区三区视频在线| 一区二区三区激情视频| 肉色欧美久久久久久久蜜桃| 国产成人a∨麻豆精品| 另类亚洲欧美激情| av又黄又爽大尺度在线免费看| 亚洲av成人不卡在线观看播放网 | 精品第一国产精品| 日韩成人av中文字幕在线观看| 可以免费在线观看a视频的电影网站 | 人人妻人人爽人人添夜夜欢视频| 在线观看人妻少妇| 美女中出高潮动态图| 欧美最新免费一区二区三区| 色播在线永久视频| 国产精品二区激情视频| 777米奇影视久久| 无遮挡黄片免费观看| 国产国语露脸激情在线看| 国产精品久久久久久久久免| 日韩av在线免费看完整版不卡| 亚洲欧美中文字幕日韩二区| 久久精品人人爽人人爽视色| 日本爱情动作片www.在线观看| 国产午夜精品一二区理论片| 777米奇影视久久| 十八禁高潮呻吟视频| 亚洲av成人精品一二三区| 999精品在线视频| 欧美日韩成人在线一区二区| 久久久久精品性色| 国产在线免费精品| 国产精品免费大片| 日韩精品免费视频一区二区三区| 在线观看免费午夜福利视频| 少妇 在线观看| 欧美精品一区二区大全| 五月天丁香电影| 亚洲精品aⅴ在线观看| 一本久久精品| 大陆偷拍与自拍| 咕卡用的链子| 亚洲av日韩精品久久久久久密 | 麻豆av在线久日| 最近2019中文字幕mv第一页| 纵有疾风起免费观看全集完整版| 亚洲一区中文字幕在线| 午夜福利网站1000一区二区三区| 一区二区三区乱码不卡18| 亚洲人成77777在线视频| 在线观看免费高清a一片| 日韩熟女老妇一区二区性免费视频| 国产成人免费无遮挡视频| 看十八女毛片水多多多| 久久韩国三级中文字幕| 久久久久久久久久久免费av| 天美传媒精品一区二区| 亚洲精品美女久久av网站| 观看美女的网站| 欧美激情 高清一区二区三区| 国产亚洲午夜精品一区二区久久| 久久国产精品男人的天堂亚洲| 一级毛片 在线播放| 久久久国产欧美日韩av| 菩萨蛮人人尽说江南好唐韦庄| 新久久久久国产一级毛片| 又大又黄又爽视频免费| 久久久久久人妻| 国产精品一国产av| 欧美日韩av久久| 日韩av在线免费看完整版不卡| 亚洲五月色婷婷综合| 超色免费av| 国产乱来视频区| 日韩av在线免费看完整版不卡| 韩国av在线不卡| 一区二区三区精品91| 国产精品国产三级国产专区5o| 国产精品久久久久久人妻精品电影 | 欧美国产精品一级二级三级| 宅男免费午夜| 国产亚洲午夜精品一区二区久久| 日韩精品免费视频一区二区三区| www.精华液| 熟女av电影| 国产99久久九九免费精品| www.av在线官网国产| 国产成人免费无遮挡视频| 国产精品av久久久久免费| 欧美日韩国产mv在线观看视频| 国产有黄有色有爽视频| 国产片特级美女逼逼视频| 国产成人精品久久久久久| 国产精品免费视频内射| 日韩大码丰满熟妇| 天堂俺去俺来也www色官网| 亚洲一卡2卡3卡4卡5卡精品中文| 午夜精品国产一区二区电影| 亚洲色图 男人天堂 中文字幕| 亚洲国产成人一精品久久久| 亚洲熟女精品中文字幕| 黑人巨大精品欧美一区二区蜜桃| 国产97色在线日韩免费| 久久精品熟女亚洲av麻豆精品| 亚洲精品国产色婷婷电影| 中文字幕人妻丝袜一区二区 | 少妇被粗大的猛进出69影院| 成人漫画全彩无遮挡| 岛国毛片在线播放| 国产成人a∨麻豆精品| 人体艺术视频欧美日本| 一边摸一边抽搐一进一出视频| 国产女主播在线喷水免费视频网站| 少妇 在线观看| 国产熟女午夜一区二区三区| 久久精品国产a三级三级三级| 日韩大片免费观看网站| 80岁老熟妇乱子伦牲交| 国产成人av激情在线播放| 国产男女超爽视频在线观看| 老司机影院成人| 亚洲一区二区三区欧美精品| 亚洲精品中文字幕在线视频| 国产一区二区在线观看av| 青春草视频在线免费观看| 久久久久人妻精品一区果冻| 国产在线免费精品| 亚洲精品国产av成人精品| 亚洲激情五月婷婷啪啪| 在线亚洲精品国产二区图片欧美| av片东京热男人的天堂| 亚洲国产毛片av蜜桃av| 亚洲精品国产色婷婷电影| 啦啦啦 在线观看视频| 午夜91福利影院|