• <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.

    麻豆久久精品国产亚洲av| АⅤ资源中文在线天堂| 欧美日韩中文字幕国产精品一区二区三区| 观看美女的网站| 黄片大片在线免费观看| 在线播放国产精品三级| 欧美+日韩+精品| 精品一区二区三区人妻视频| 九九久久精品国产亚洲av麻豆| 亚洲av电影在线进入| 嫁个100分男人电影在线观看| 国产成人啪精品午夜网站| 中文字幕av在线有码专区| 欧美日韩精品网址| 中文字幕人成人乱码亚洲影| 最近视频中文字幕2019在线8| 男人的好看免费观看在线视频| 国产成人欧美在线观看| 色综合婷婷激情| 亚洲成人久久性| 亚洲片人在线观看| 夜夜爽天天搞| 精品久久久久久成人av| 成人国产综合亚洲| 午夜福利在线在线| 国产三级黄色录像| 别揉我奶头~嗯~啊~动态视频| 日本三级黄在线观看| 91九色精品人成在线观看| 中文字幕av成人在线电影| 亚洲激情在线av| 天天躁日日操中文字幕| 成人鲁丝片一二三区免费| 99热精品在线国产| 97碰自拍视频| 午夜福利免费观看在线| 可以在线观看的亚洲视频| 国产又黄又爽又无遮挡在线| 黄色日韩在线| 国产精品免费一区二区三区在线| 亚洲精品久久国产高清桃花| 亚洲在线自拍视频| 国产一区二区三区视频了| 热99在线观看视频| 桃红色精品国产亚洲av| 九色国产91popny在线| 在线观看66精品国产| 成人三级黄色视频| 午夜福利免费观看在线| 国产精品亚洲av一区麻豆| 可以在线观看毛片的网站| 亚洲中文字幕一区二区三区有码在线看| 18禁黄网站禁片免费观看直播| 九九热线精品视视频播放| 久9热在线精品视频| 日本三级黄在线观看| 中出人妻视频一区二区| 99国产精品一区二区三区| 伊人久久大香线蕉亚洲五| 亚洲内射少妇av| 成人无遮挡网站| 亚洲第一欧美日韩一区二区三区| 亚洲一区二区三区不卡视频| 婷婷精品国产亚洲av在线| 国产精品,欧美在线| 午夜老司机福利剧场| 精品日产1卡2卡| 成人高潮视频无遮挡免费网站| 男人舔奶头视频| 床上黄色一级片| 可以在线观看的亚洲视频| 国产成人av激情在线播放| 又黄又爽又免费观看的视频| av欧美777| 日韩欧美免费精品| 免费看十八禁软件| av女优亚洲男人天堂| 首页视频小说图片口味搜索| 级片在线观看| 每晚都被弄得嗷嗷叫到高潮| 有码 亚洲区| 真人一进一出gif抽搐免费| 成年女人毛片免费观看观看9| 色综合婷婷激情| 大型黄色视频在线免费观看| 亚洲av电影在线进入| 国产69精品久久久久777片| 精品久久久久久久人妻蜜臀av| 日本黄大片高清| 九九久久精品国产亚洲av麻豆| 免费看美女性在线毛片视频| av视频在线观看入口| 免费搜索国产男女视频| 免费搜索国产男女视频| a在线观看视频网站| 熟女人妻精品中文字幕| 日本撒尿小便嘘嘘汇集6| 欧美中文日本在线观看视频| 精品99又大又爽又粗少妇毛片 | 91久久精品电影网| 国产免费一级a男人的天堂| 在线观看av片永久免费下载| av中文乱码字幕在线| 久久精品国产自在天天线| 国产欧美日韩精品亚洲av| 九九热线精品视视频播放| 日日干狠狠操夜夜爽| 少妇裸体淫交视频免费看高清| 色播亚洲综合网| 操出白浆在线播放| 国产精品美女特级片免费视频播放器| 国产精品久久电影中文字幕| 日本在线视频免费播放| 精品国内亚洲2022精品成人| 国产精品野战在线观看| 国产毛片a区久久久久| 国产99白浆流出| 91av网一区二区| 亚洲国产精品sss在线观看| 久久欧美精品欧美久久欧美| 国产主播在线观看一区二区| 十八禁网站免费在线| 国产精品精品国产色婷婷| 久久天躁狠狠躁夜夜2o2o| aaaaa片日本免费| 久久久色成人| 久久久久久大精品| 老熟妇乱子伦视频在线观看| 国产一区二区三区视频了| 免费av毛片视频| 少妇裸体淫交视频免费看高清| 啪啪无遮挡十八禁网站| 国产麻豆成人av免费视频| 特大巨黑吊av在线直播| a级一级毛片免费在线观看| www.熟女人妻精品国产| 又黄又爽又免费观看的视频| 欧美日韩综合久久久久久 | 成人精品一区二区免费| 99久久九九国产精品国产免费| 国产一区二区三区视频了| 3wmmmm亚洲av在线观看| 90打野战视频偷拍视频| 欧美xxxx黑人xx丫x性爽| 51国产日韩欧美| 日韩欧美国产在线观看| 搞女人的毛片| 午夜影院日韩av| 亚洲av一区综合| 国产成人啪精品午夜网站| 一级黄色大片毛片| 亚洲精品成人久久久久久| 欧美zozozo另类| 日韩欧美在线二视频| 国内精品美女久久久久久| 99在线人妻在线中文字幕| 久久久久国内视频| 亚洲在线自拍视频| 两个人的视频大全免费| 听说在线观看完整版免费高清| 久久精品国产99精品国产亚洲性色| 老司机深夜福利视频在线观看| 一进一出好大好爽视频| 嫁个100分男人电影在线观看| 成人精品一区二区免费| 99国产极品粉嫩在线观看| 欧美一区二区亚洲| 精品久久久久久成人av| 国产成+人综合+亚洲专区| 日本五十路高清| a级毛片a级免费在线| 国产日本99.免费观看| 久久久久国产精品人妻aⅴ院| 国产精品av视频在线免费观看| 久久亚洲真实| 真实男女啪啪啪动态图| 99久久久亚洲精品蜜臀av| 丰满的人妻完整版| 日韩欧美精品免费久久 | 91九色精品人成在线观看| 精品国产亚洲在线| 国产一区二区在线av高清观看| 97超级碰碰碰精品色视频在线观看| 欧美不卡视频在线免费观看| 高清日韩中文字幕在线| 欧美日韩福利视频一区二区| 香蕉久久夜色| 欧美bdsm另类| 一本一本综合久久| 窝窝影院91人妻| 国产高潮美女av| 日本免费a在线| 欧美极品一区二区三区四区| 亚洲国产高清在线一区二区三| 免费高清视频大片| 国产麻豆成人av免费视频| 手机成人av网站| 18禁国产床啪视频网站| 99riav亚洲国产免费| 99热精品在线国产| 久久久久亚洲av毛片大全| 天天躁日日操中文字幕| 亚洲人成网站高清观看| 日韩中文字幕欧美一区二区| 成人特级av手机在线观看| 很黄的视频免费| 免费av不卡在线播放| 88av欧美| 在线看三级毛片| 天天一区二区日本电影三级| 国产成年人精品一区二区| 国产成人aa在线观看| 久久久久久久久中文| 欧美bdsm另类| 99riav亚洲国产免费| 波多野结衣巨乳人妻| 亚洲精品日韩av片在线观看 | 久久精品国产99精品国产亚洲性色| 国产精品乱码一区二三区的特点| 久久欧美精品欧美久久欧美| 日本免费a在线| 久久伊人香网站| 国内少妇人妻偷人精品xxx网站| 嫩草影院精品99| 在线看三级毛片| 成年版毛片免费区| 男女床上黄色一级片免费看| 少妇的丰满在线观看| 婷婷精品国产亚洲av| АⅤ资源中文在线天堂| 女人被狂操c到高潮| 婷婷精品国产亚洲av| 淫秽高清视频在线观看| 欧美中文日本在线观看视频| 少妇熟女aⅴ在线视频| 亚洲美女视频黄频| 久久久色成人| 欧美日韩瑟瑟在线播放| 色吧在线观看| 欧美3d第一页| 欧美不卡视频在线免费观看| av中文乱码字幕在线| 五月玫瑰六月丁香| 一个人观看的视频www高清免费观看| 欧美xxxx黑人xx丫x性爽| 国产色婷婷99| 久久久久亚洲av毛片大全| 国产一区二区激情短视频| 亚洲人与动物交配视频| 国产激情欧美一区二区| 亚洲在线观看片| 老司机福利观看| 午夜精品一区二区三区免费看| 少妇熟女aⅴ在线视频| 日韩欧美国产一区二区入口| 国产私拍福利视频在线观看| 亚洲国产精品成人综合色| www国产在线视频色| 亚洲色图av天堂| 啦啦啦免费观看视频1| www.999成人在线观看| 国产精品嫩草影院av在线观看 | 色吧在线观看| 国产精品久久视频播放| 最新美女视频免费是黄的| 91字幕亚洲| 日韩高清综合在线| 成人欧美大片| 日本五十路高清| 成人特级av手机在线观看| 国产亚洲av嫩草精品影院| 日韩欧美国产在线观看| 成人特级黄色片久久久久久久| 久久精品影院6| 俄罗斯特黄特色一大片| 国产亚洲精品久久久com| 欧美中文日本在线观看视频| 亚洲国产欧美网| 精品不卡国产一区二区三区| 久久精品国产亚洲av香蕉五月| 女人十人毛片免费观看3o分钟| 非洲黑人性xxxx精品又粗又长| 免费在线观看日本一区| 色哟哟哟哟哟哟| 久久久久久大精品| 国产中年淑女户外野战色| 啦啦啦韩国在线观看视频| 中文亚洲av片在线观看爽| 国产主播在线观看一区二区| 亚洲精华国产精华精| 尤物成人国产欧美一区二区三区| 国产视频一区二区在线看| 精品国产美女av久久久久小说| 天天躁日日操中文字幕| 色尼玛亚洲综合影院| 麻豆久久精品国产亚洲av| 99在线人妻在线中文字幕| 成人性生交大片免费视频hd| 99久国产av精品| 精品一区二区三区人妻视频| 18禁美女被吸乳视频| 久久久久精品国产欧美久久久| 亚洲成人精品中文字幕电影| 亚洲五月婷婷丁香| 国语自产精品视频在线第100页| 亚洲无线在线观看| 国产久久久一区二区三区| 国产真实伦视频高清在线观看 | 亚洲,欧美精品.| 欧美丝袜亚洲另类 | 国产精品99久久久久久久久| 精品国产美女av久久久久小说| 欧美乱妇无乱码| 国产精品永久免费网站| 变态另类成人亚洲欧美熟女| 国产高清视频在线观看网站| 99在线视频只有这里精品首页| 欧美最黄视频在线播放免费| 99久久久亚洲精品蜜臀av| 超碰av人人做人人爽久久 | 亚洲国产日韩欧美精品在线观看 | 51国产日韩欧美| 91在线观看av| 国产成人系列免费观看| 欧美xxxx黑人xx丫x性爽| 高潮久久久久久久久久久不卡| 免费在线观看影片大全网站| 国产一区二区亚洲精品在线观看| 两人在一起打扑克的视频| 99久久无色码亚洲精品果冻| 成人亚洲精品av一区二区| 一级黄色大片毛片| 色综合亚洲欧美另类图片| 亚洲av免费高清在线观看| 一区二区三区国产精品乱码| 国产一区在线观看成人免费| 亚洲精品影视一区二区三区av| 一卡2卡三卡四卡精品乱码亚洲| 国产午夜福利久久久久久| 亚洲国产欧美人成| 老司机午夜福利在线观看视频| 日本黄色视频三级网站网址| 国产精品日韩av在线免费观看| 女人被狂操c到高潮| 欧美区成人在线视频| 亚洲av日韩精品久久久久久密| 老司机福利观看| 日韩人妻高清精品专区| 国产黄色小视频在线观看| 高清在线国产一区| av中文乱码字幕在线| 18美女黄网站色大片免费观看| 欧美3d第一页| 全区人妻精品视频| 一夜夜www| 精品久久久久久久久久免费视频| 亚洲中文字幕一区二区三区有码在线看| 51午夜福利影视在线观看| 特大巨黑吊av在线直播| 久久精品91蜜桃| 少妇丰满av| 国产亚洲av嫩草精品影院| 亚洲天堂国产精品一区在线| 白带黄色成豆腐渣| 亚洲国产精品sss在线观看| 露出奶头的视频| www国产在线视频色| 久久香蕉精品热| 神马国产精品三级电影在线观看| 午夜亚洲福利在线播放| 精品人妻一区二区三区麻豆 | 日韩av在线大香蕉| 一本精品99久久精品77| 日韩精品青青久久久久久| 搡老妇女老女人老熟妇| 国产在视频线在精品| 久久精品人妻少妇| 亚洲精品影视一区二区三区av| 精品一区二区三区av网在线观看| 久久久久亚洲av毛片大全| 观看美女的网站| 最近最新中文字幕大全电影3| 日本黄大片高清| 极品教师在线免费播放| 黄色丝袜av网址大全| 精品人妻偷拍中文字幕| 国内毛片毛片毛片毛片毛片| 色播亚洲综合网| 午夜精品久久久久久毛片777| 特级一级黄色大片| 午夜免费观看网址| 97超级碰碰碰精品色视频在线观看| 又紧又爽又黄一区二区| 丰满乱子伦码专区| 午夜免费成人在线视频| 免费搜索国产男女视频| 人人妻人人看人人澡| 宅男免费午夜| 午夜福利成人在线免费观看| 亚洲av成人不卡在线观看播放网| 午夜免费激情av| 在线十欧美十亚洲十日本专区| 丰满乱子伦码专区| xxx96com| 热99在线观看视频| e午夜精品久久久久久久| 精品一区二区三区视频在线观看免费| 88av欧美| 美女 人体艺术 gogo| 不卡一级毛片| 毛片女人毛片| 日本成人三级电影网站| 男插女下体视频免费在线播放| 国产乱人伦免费视频| a级毛片a级免费在线| 亚洲在线观看片| 日本与韩国留学比较| 精品无人区乱码1区二区| 亚洲中文字幕一区二区三区有码在线看| 国产黄a三级三级三级人| 亚洲第一电影网av| 97人妻精品一区二区三区麻豆| 精品99又大又爽又粗少妇毛片 | 国产综合懂色| 搡女人真爽免费视频火全软件 | 窝窝影院91人妻| 婷婷六月久久综合丁香| www国产在线视频色| 悠悠久久av| 五月玫瑰六月丁香| 他把我摸到了高潮在线观看| 精品一区二区三区视频在线 | 亚洲中文字幕一区二区三区有码在线看| 亚洲性夜色夜夜综合| 一本精品99久久精品77| 亚洲中文字幕一区二区三区有码在线看| 国产欧美日韩精品亚洲av| 一卡2卡三卡四卡精品乱码亚洲| 欧美日韩黄片免| 国产黄色小视频在线观看| 精品一区二区三区av网在线观看| 在线观看一区二区三区| av女优亚洲男人天堂| 99久久九九国产精品国产免费| 国产成人福利小说| 丰满人妻熟妇乱又伦精品不卡| 亚洲av成人不卡在线观看播放网| 欧美黑人巨大hd| 免费av观看视频| 亚洲av成人精品一区久久| 欧美在线黄色| 成人18禁在线播放| 久久精品国产清高在天天线| 精品国内亚洲2022精品成人| 久久久久久久午夜电影| 亚洲va日本ⅴa欧美va伊人久久| av视频在线观看入口| 丁香六月欧美| 欧美一区二区亚洲| 无限看片的www在线观看| 欧美黑人巨大hd| 免费av不卡在线播放| 亚洲欧美日韩高清专用| 91字幕亚洲| 日韩有码中文字幕| av天堂在线播放| 亚洲一区高清亚洲精品| 午夜免费观看网址| 精品一区二区三区av网在线观看| 露出奶头的视频| 亚洲无线观看免费| 亚洲国产欧美人成| 亚洲成av人片在线播放无| 琪琪午夜伦伦电影理论片6080| 美女被艹到高潮喷水动态| 制服丝袜大香蕉在线| 成人高潮视频无遮挡免费网站| 老鸭窝网址在线观看| 麻豆成人av在线观看| 男女之事视频高清在线观看| 国产精品国产高清国产av| 性色avwww在线观看| 身体一侧抽搐| 日韩有码中文字幕| 日韩 欧美 亚洲 中文字幕| 日本在线视频免费播放| 19禁男女啪啪无遮挡网站| 人人妻人人看人人澡| 亚洲av熟女| 小说图片视频综合网站| 男女视频在线观看网站免费| 亚洲一区二区三区色噜噜| 午夜精品久久久久久毛片777| 无遮挡黄片免费观看| 精品午夜福利视频在线观看一区| 丰满人妻一区二区三区视频av | 国产三级在线视频| 法律面前人人平等表现在哪些方面| 国产精品av视频在线免费观看| 精品久久久久久成人av| 黄片小视频在线播放| 欧洲精品卡2卡3卡4卡5卡区| 国产精品三级大全| 久久精品国产综合久久久| 亚洲精华国产精华精| 动漫黄色视频在线观看| 美女高潮的动态| 久久精品国产自在天天线| 国内精品久久久久精免费| 欧美性猛交黑人性爽| 激情在线观看视频在线高清| 国产亚洲av嫩草精品影院| 亚洲真实伦在线观看| 亚洲av电影不卡..在线观看| 久久精品夜夜夜夜夜久久蜜豆| 夜夜躁狠狠躁天天躁| 久久久久久大精品| 亚洲专区国产一区二区| 久久久色成人| 搡老岳熟女国产| 中文字幕人妻熟人妻熟丝袜美 | 真人做人爱边吃奶动态| 偷拍熟女少妇极品色| 一个人看的www免费观看视频| 十八禁人妻一区二区| 女人高潮潮喷娇喘18禁视频| 精品久久久久久成人av| 久久久久久久午夜电影| 伊人久久精品亚洲午夜| 久久午夜亚洲精品久久| 国产av在哪里看| 中文在线观看免费www的网站| 亚洲熟妇熟女久久| 欧美日韩瑟瑟在线播放| 给我免费播放毛片高清在线观看| 久久久久亚洲av毛片大全| 成熟少妇高潮喷水视频| 无遮挡黄片免费观看| 亚洲在线自拍视频| 嫩草影院精品99| 99视频精品全部免费 在线| 亚洲欧美日韩高清专用| 久久精品人妻少妇| 1024手机看黄色片| 国产伦在线观看视频一区| 黄色日韩在线| 超碰av人人做人人爽久久 | 搡老熟女国产l中国老女人| 亚洲18禁久久av| 国产在视频线在精品| 中文字幕人成人乱码亚洲影| aaaaa片日本免费| 三级国产精品欧美在线观看| 男人的好看免费观看在线视频| 欧美日韩精品网址| 欧美性感艳星| 亚洲午夜理论影院| 成人特级av手机在线观看| 啦啦啦观看免费观看视频高清| 国产精品一及| 亚洲男人的天堂狠狠| 日韩 欧美 亚洲 中文字幕| 国产69精品久久久久777片| 亚洲专区中文字幕在线| 天天一区二区日本电影三级| 成人特级av手机在线观看| 国产伦精品一区二区三区四那| 欧美一区二区国产精品久久精品| 一个人看视频在线观看www免费 | 身体一侧抽搐| 两人在一起打扑克的视频| 在线免费观看的www视频| 中文字幕av在线有码专区| 乱人视频在线观看| 一级黄色大片毛片| 午夜激情欧美在线| 他把我摸到了高潮在线观看| 亚洲人成网站高清观看| 亚洲av熟女| 老汉色∧v一级毛片| 法律面前人人平等表现在哪些方面| www.999成人在线观看| 亚洲国产精品合色在线| 麻豆国产av国片精品| 久久久久亚洲av毛片大全| 国产精品永久免费网站| 国产av不卡久久| 亚洲精品一卡2卡三卡4卡5卡| 操出白浆在线播放| 91麻豆精品激情在线观看国产| 波多野结衣巨乳人妻| 午夜免费激情av| 免费看a级黄色片| 国产色婷婷99| 亚洲av第一区精品v没综合| 成人av一区二区三区在线看| 国产v大片淫在线免费观看| 免费电影在线观看免费观看| 亚洲欧美日韩东京热| 禁无遮挡网站| 一区福利在线观看| av专区在线播放| 十八禁人妻一区二区| 女人高潮潮喷娇喘18禁视频| 中文字幕久久专区| 一夜夜www| 欧美+亚洲+日韩+国产| 日本三级黄在线观看| 亚洲av一区综合| 精品一区二区三区视频在线观看免费| 亚洲avbb在线观看| 久久香蕉国产精品| 亚洲中文字幕日韩| 国产69精品久久久久777片| 女生性感内裤真人,穿戴方法视频|