背景：宫颈癌（Cervical Cancer，CC）是一种常见的妇科恶性肿瘤，全球范围内造成的女性患者死亡率居第四位，严重威胁女性健康。高危型人乳头瘤病毒（High-risk Human Papilloma Virus，HR-HPV）的持续感染是宫颈癌发病的主要原因。宿主遗传变异和病毒之间复杂的相互作用是决定病毒感染结局以及疾病进展的关键因素。本研究拟从全转录组测序、DNA 甲基化修饰和免疫相关基因变异等多个层面探究宿主遗传因素在HPV 感染发生以及肿瘤进展过程中的作用，以期为宫颈癌的预防和治疗提供理论支撑。

方法：（1）选取9例宫颈癌组织与相对应的癌旁正常组织进行全转录组测序，用DESeq2或DEGseq进行差异基因检测，条件为Qvalue ≤ 0.05或Qvalue ≤ 0.001；使用Phyper对差异基因进行GO及KEGG富集分析，以 Qvalue ≤ 0.05或Qvalue ≤ 0.001为阈值，满足此条件的定义为在候选基因中显著富集；利用GEPIA对差异表达的mRNA进行宫颈癌预后分析；分别利用Cytoscape中的stringAPP插件和cytoHubba插件进行蛋白质互作分析和Hub基因分析；利用Cytoscape软件构建 circRNA/lncRNA-miRNA-mRNA ceRNA调控网络；利用免疫细胞丰度综合分析平台ImmuneCellAI进行免疫细胞浸润丰度分析。（2）通过公共数据库TCGA和GTEx、宫颈癌组织转录组测序数据和启动子高甲基化基因筛选宫颈癌中异常甲基化基因；通过RT-qPCR、启动子区甲基化注释筛选启动子受高甲基化调控的基因GFRA1；通过BSP检测宫颈癌细胞系中GFRA1启动子甲基化水平；通过MSP检测宫颈癌细胞系和癌组织中GFRA1启动子甲基化水平；通过RT-qPCR检测宫颈癌组织及癌旁正常组织中GFRA1的表达情况；通过慢病毒感染构建基因稳定表达或表达敲低的宫颈癌细胞系；通过RNA测序（RNA-sequencing，RNA-seq）查找GFRA1作用的信号通路及下游作用分子；通过Transwell检测GFRA1对宫颈癌细胞迁移能力的影响；通过构建裸鼠和小鼠肺转移模型评价GFRA1对肿瘤转移的作用；通过TIMER在线平台分析基因表达与免疫细胞浸润的关系。（3）以中国汉族人群中的健康对照者、癌前病变患者（Cervical Intraepithelial Neoplasia，CIN）和宫颈癌患者为研究对象，利用下一代测序技术（Next-Generation Sequencing，NGS）对HLA Ⅰ 类基因（A、B和C）和HLA Ⅱ 类基因（DRB1、DPB1和DQB1）进行基因分型；使用Tellgenplex HPV DNA检测试剂盒对HPV进行分型；使用PyHLA软件和χ²检验检测宫颈癌、CIN和健康对照组之间等位基因频率的差异；采用χ² 检验检测宫颈癌、CIN和健康对照组HLA单倍型频率差异以及宫颈癌和CIN患者中HPV基因型频率的差异。

结果：（1）宫颈癌组织全转录组测序获得 DEmRNA 3,763 个（上调 2,343 个，下调 1,420 个），DEcircRNA 6,022个（上调2,737个，下调3,285个），DElncRNA 2,532个（上调1,473个，下调1,059个），DEmiRNA 174个（上调117个，下调57个）。对差异表达的mRNA进行富集分析发现差异基因主要集中在信号分子和相互作用、癌症、细胞生长和死亡、DNA复制和修复等通路上；预后分析鉴定到16个宫颈癌预后相关基因；蛋白互作分析鉴定到74个Hub基因。对差异表达的miRNA进行靶向分析，获得2,638个靶向mRNA，637个靶向lncRNA，980个靶向circRNA。对具有靶向关系且差异表达的基因构建circRNA/lncRNA-miRNA-mRNA网络，获得23个核心miRNA和以及404个靶向的mRNA，以及可能具有ceRNA效应的16个circRNA和6个lncRNA。免疫细胞浸润分析发现，相比较癌旁正常组织，宫颈癌组织中具有较强的免疫抑制细胞的浸润。

（2）宫颈癌异常甲基化基因功能研究中，通过BSP和MSP分析，发现在宫颈癌细胞和癌组织中GFRA1启动子区甲基化水平升高，宫颈癌组织中GFRA1甲基化率为83.3%（30/36）。通过RT-qPCR发现相比较癌旁正常组织，癌组织中 GFRA1表达降低，且GFRA1的表达与其启动子区的高甲基化相关（P < 0.05）。体外实验发现GFRA1上调能够抑制宫颈癌细胞的迁移；体内实验证实GFRA1上调抑制小鼠中宫颈癌细胞的肺转移能力。转录组测序发现SERPINE1为GFRA1下游作用分子，GFRA1负调控SERPINE1的表达；体外实验发现SERPINE1上调能够从促进宫颈癌细胞的迁移；体内实验证实SERPINE1上调促进小鼠中宫颈癌细胞的肺转移能力。

（3）等位基因 HLA Ⅰ-A*33:03、B*58:01 和 C*03:02 在 CC 组和CIN2+ 组中的频率均低于对照组（Pc < 0.05），而 B*55:02和 C*04:03 的频率高于对照组 （Pc < 0.05）。单倍型 HLA Ⅰ-A*33:03-C*03:02-B*58:01在 CC 和 CIN2+ 组中的频率均低于对照组 （Pc < 0.05）。等位基因 HLA Ⅱ-DPB1*05:01 和  DRB1*12:01 在 CC 组和 CIN2+ 组中的频率均低于对照组（Pc < 0.05）。CC 组中 DQB1*06:03 和 DRB1*04:06 频率低于对照组（Pc < 0.05）。CC组中，单倍型HLA-DRB1*16:02-DQB1*05:02-DPB1*05:01频率低于对照组（Pc < 0.05），HLA-DRB1*12:02-DQB1*03:01-DPB1*296:01频率高于CIN2+组（Pc < 0.05）。在HPV16感染患者中，CC组A*33:03、C*03:02、DPB1*05:01的频率明显低于对照组（Pc<0.05），而B*55:02的频率显著高于对照组（Pc < 0.05）；此外，CIN2+组的C*03:02和DPB1*05:01频率显著低于对照组（Pc < 0.05）。在HPV16感染患者中，单倍型 A*33:03-C*03:02-B*58:01 在CC 组和CIN2+组中的频率均低于对照组（Pc < 0.05），而CIN2+组中A*11:01-C*14:02-B*51:01和A*24:02-C*03:04-B*13:01的频率的高于CC组和对照组（Pc < 0.05）。

结论：（1）宫颈癌组织全转录组测序发现宫颈癌组织特异的DEGs、Hub基因和预后基因，宫颈癌中异常表达的circRNA和lncRNA有可能通过ceRNA效应参与宫颈癌的调控；宫颈癌组织中的免疫抑制细胞浸润水平高于癌旁正常组织。（2）GFRA1基因启动子区在宫颈癌组织和宫颈癌细胞中甲基化水平升高，其高甲基化有可能导致宫颈癌组织和宫颈癌细胞中GFRA1低表达； GFRA1作用于下游分子SERPINE1，通过抑制肿瘤细胞转移来发挥抑癌作用。GFRA1有望成为宫颈癌转移风险预测和评估的生物标志物。（3）HLA Ⅰ 类和HLA Ⅱ 类等位基因及其单倍型不仅与 CIN 和宫颈癌的风险相关，还与宫颈癌的组织学类型以及分期相关。此外，HLA基因在等位基因和单倍型水平上也与HPV16感染相关。将来，这些HLA Ⅰ 类和HLA Ⅱ 类基因在CIN和宫颈癌发展中的作用应进行进一步的研究和验证。

Background: Cervical cancer (CC) is a common gynecologic malignancy that poses a serious threat to women's health and causes the fourth highest mortality rate among female patients worldwide. Persistent infection with high-risk human papilloma virus (HR-HPV) is the main cause of CC. Complex interactions between host genetic variants and viruses are key determinants of the outcome of viral infection as well as disease progression. In the current study, we investigated the role of host genetic factors in the development of HPV infection and tumor progression from the perspective of whole transcriptome sequencing, DNA methylation modification and immune-related genetic variation, with the aim of providing theoretical support for the prevention and treatment of CC.

Methods: (1) In the current study, nine cases of CC tissues and matched adjacent normal tissues were selected for whole transcriptome sequencing; DESeq2 or DEGseq was used to detect the differentially expressed genes (DEGs) under the condition of Qvalue ≤ 0.05 or Qvalue ≤ 0.001; Over Phyper for GO and KEGG enrichment analysis of DEGs, with Qvalue ≤ 0.05 or Qvalue ≤ 0.001 as the threshold, and those meeting this condition were defined as significantly enriched in candidate genes. GEPIA was used to analyze the prognosis of differentially expressed mRNA in CC, and the stringAPP and cytoHubba in Cytoscape were used to analyze protein-protein interaction and Hub genes, respectively. Immune cell infiltration abundance analysis was performed by ImmuneCellAI, a comprehensive immune cell abundance analysis platform; circRNA/lncRNA-miRNA-mRNA ceRNA interaction network was constructed by Cytoscape. (2) Abnormally methylated genes in CC were screened by public databases TCGA and GTEx, transcriptome sequencing data of CC tissues, and promoter hypermethylation genes; RT-qPCR and promoter region methylation annotation were used to screen genes whose promoters are regulated by hypermethylation in CC. BSP was used to detect GFRA1 promoter methylation level in CC cell lines; MSP was used to detect the GFRA1 promoter methylation in CC tissues and CC cell lines; RT-qPCR was used to detect the GFRA1 expression in CC tissues and adjacent normal tissues; CC cell lines with GFRA1 stably overexpression or expression knockdown were constructed by lentiviral infection; RNA-sequencing (RNA-seq) was used to detect the signaling pathways and downstream molecules of the action of GFRA1; the effect of GFRA1 on the migration ability of CC cell lines was detected by Transwell; the effect of GFRA1 on tumor metastasis was evaluated by constructing lung metastasis models in mice; and the relationship between genes expression and infiltration of immune cells was analyzed by online platform, TIMER.

(3) Healthy controls, cervical intraepithelial neoplasia (CIN) and CC in the Chinese Han population were genotyped for HLA class I genes (A, B and C) and HLA class Ⅱ genes (DRB1, DPB1 and DQB1) with next-generation sequencing (NGS); a Tellgenplex HPV DNA detection kit was used for HPV typing; differences in allele frequencies between CC, CIN and healthy controls were detected with PyHLA and the χ² test; differences in HLA haplotype frequencies between CC, CIN and healthy controls, and differences in HPV genotype frequencies between CC and CIN patients were detected with the χ² test.

Results:(1) Whole transcriptome sequencing of CC tissues obtained 3,763 DEmRNAs (2,343 up-regulated and 1,420 down-regulated), 6,022 DEcircRNAs (2,737 up-regulated and 3,285 down-regulated), 2,532 DElncRNAs (1,473 up-regulated and 1,059 down-regulated), and 174 DEmiRNAs (117 up-regulated and 57 down-regulated). Enrichment analysis of DEmRNAs mainly concentrated in the pathways of signaling molecules and interactions, cancer, cell growth and death, as well as DNA replication and repair; prognostic analysis identified 16 mRNAs with prognostic ability for CC; and protein-protein interactions analysis identified 74 Hub genes. Targeting analysis of DEmiRNAs obtained 2,638 targeted mRNAs, 637 targeted lncRNAs and 980 targeted circRNAs. circRNA/lncRNA-miRNA-mRNA ceRNA networks were constructed for DEGs with targeting relationships. ceRNA networks obtained 23 core miRNAs and 404 mRNAs, as well as 16 circRNAs and 6 lncRNAs. Immune cell infiltration analysis showed that CC tissues had higher immunosuppressive cell infiltration compared with adjacent normal tissues.

(2) In the functional study of abnormal methylated gene in CC, hypermethylation of GFRA1 promoter region was found in CC cells and CC tissues by BSP and MSP, and the rate of GFRA1 methylation was 83.3% (30/36) in CC. With RT-qPCR, GFRA1 expression was reduced in CC tissues compared with adjacent normal tissues, and it was associated with hypermethylation of the promoter region (P < 0.05). In vitro experiments showed that GFRA1 could inhibit the migration of CC cells; in vivo experiments confirmed that GFRA1 could inhibit the lung metastasis of CC cells in mice. Transcriptome sequencing revealed that SERPINE1 was a downstream molecule of GFRA1, and GFRA1 negatively regulated the expression of SERPINE1; In vitro experiments showed that SERPINE1 could promote the migration of CC cells; in vivo experiments confirmed that SERPINE1 could promote the lung metastasis of CC cells in mice.

 (3) The frequencies of alleles HLA Ⅰ-A*33:03, B*58:01 and C*03:02 in CC and CIN2+ groups were lower than the control group (Pc < 0.05), while the frequencies of B*55:02 and C*04:03 were higher than the control group (Pc < 0.05). Haplotype HLA Ⅰ-A*33:03-C*03:02-B*58:01 in both CC and CIN2+ groups was lower than healthy controls (Pc < 0.05). The frequencies of HLA Ⅱ-DPB1*05:01 and DRB1*12:01 in both CC and CIN2+ groups were lower than the controls (Pc < 0.05). Frequencies of DQB1*06:03 and DRB1*04:06 in CC were lower than the controls (Pc < 0.05). Haplotype DRB1*16:02-DQB1*05:02-DPB1*05:01 in CC was lower than the control group (Pc < 0.05), and HLA-DRB1*12:02-DQB1*03:01-DPB1*296:01 in the CC group was higher than in the CIN2+ group (Pc < 0.05). Among HPV16 infected patients, the frequencies of A*33:03, C*03:02, and DPB1*05:01 were significantly lower in the CC group than in the control group (Pc < 0.05), while the frequency of B*55:02 was significantly higher than the control group (Pc < 0.05). In addition, the frequency of C*03:02 and DPB1*05:01 was significantly lower in the CIN2+ group than in the control group (Pc < 0.05). In HPV16 infected patients, haplotype A*33:03-C*03:02-B*58:01 was lower in the CC group and the CIN2+ group than the control group, whereas the frequencies of A*11:01-C*14:02-B*51:01 and A*24:02-C*03:04-B*13:01 in the CIN2+ group were higher than the CC group and the control group.

Conclusion: (1) Whole transcriptome sequencing of CC tissues identified CC-specific DEGs as well as Hub genes and prognostic genes, circRNA and lncRNA, which are likely to be involved in the regulation of CC through the ceRNA effect; and the level of immunosuppressive cell infiltration in CC tissues is higher than in adjacent normal tissue. (2) The promoter region of GFRA1 is hypermethylated in CC tissues and CC cell lines, and its hypermethylation may lead to the low expression of GFRA1 in CC tissues and CC cells; SERPINE1 was the downstream molecule of GFRA1, and GFRA1 exerts a cancer-suppressing effect by inhibiting the metastasis of tumor cells. The methylation of GFRA1 is expected to be a predictive and evaluative biomarker for the prediction and assessment of CC metastasis risk. (3) HLA class I and class Ⅱ alleles and their haplotypes are associated not only with the risk of CIN and CC, but also with the histologic type and stage of CC. In addition, HLA genes were also associated with HPV16 infection at the allele and haplotype level. In the future, the function of these HLA class I and class Ⅱ genes in CIN and CC should be further studied and validated.