第一部分 结肠炎小鼠肠道菌群失调与胆汁酸代谢紊乱的特征分析

目的：分析急性结肠炎小鼠肠道菌群和胆汁酸代谢的变化特点，探讨二者之间的相关性。方法：选取C57BL/6小鼠，随机分成正常对照组和DSS建模组，正常组不予干预，DSS建模组予以2.5% DSS饮用水5天诱导结肠炎，7天后处死取材。收集DSS建模组第0天和第7天的新鲜粪便标本进行16S rDNA测序和粪便胆汁酸靶向组学测定，分别分析肠道菌群和胆汁酸代谢的变化特点，并对二者进行关联分析。结果：16S测序数据分析：与建模前相比，DSS诱导的结肠炎小鼠粪便菌群alpha多样性显著降低，多样性指数chao1、observed_species、PD_whole_tree和shannon指数均存在显著差异（P值分别为0.003、0.001、0.041和0.043）。PCoA分析显示结肠炎小鼠粪便菌群与建模前相比存在显著差异。DSS建模后小鼠粪便菌群的组成结构发生明显改变，进一步LEfSe判别分析提示Lactobacillus、Bifidobacterium和Staphylococcus菌属在DSS建模后明显减少，Turicibacter、Lachnospiraceae_NK4A136_group、Bacteroides、Enterococcus、Escherichia-Shigella、Clostridium_sensu_stricto_1和Pasteurella菌属在DSS建模后显著增加。PICRUST2功能预测及PCA分析提示DSS建模后小鼠肠道菌群的功能发生明显变化，KEGG分析得到43条显著改变的信号途径，其中按丰度排名前5的信号途径分别为：安沙霉素类合成途径、次级胆汁酸合成途径、D-谷氨酰胺和D-谷氨酸代谢途径、D-丙氨酸代谢途径、脂肪酸合成途径。粪便胆汁酸靶向代谢组学分析：PCA分析显示，结肠炎小鼠粪便胆汁酸谱较DSS建模前发生明显变化，其中总胆汁酸含量显著降低（P=0.016），非结合型胆汁酸、初级胆汁酸和次级胆汁酸均明显降低（P值分别为0.017，0.008和0.043），次级胆汁酸与初级胆汁酸比值也呈下降趋势（P=0.032）。基于OPLS-DA模型的VIP值（VIP>1）及显著性检验的P值（P<0.05）联合筛选DSS建模前后的差异胆汁酸，共鉴定出11个差异胆汁酸，分别为12-ketoLCA、TUDCA、TDCA、THDCA、isoDCA、βUDCA、TCDCA、6-ketoLCA、TCA、TLCA和NorDCA。进一步Pearson相关性分析显示差异菌属Bifidobacterium、Lactobacillus的丰度与多数差异胆汁酸的含量均呈现明显正相关。结论：DSS诱导的结肠炎小鼠粪便菌群结构和功能发生明显变化，主要表现为菌群多样性下降、有益菌减少、有害菌和条件致病菌增加，菌群的次级胆汁酸合成功能下降。同时，结肠炎小鼠粪便胆汁酸谱改变明显，总胆汁酸水平降低，次级与初级胆汁酸比值降低。结肠炎小鼠粪便中的双歧杆菌和乳杆菌属的丰度与多数差异次级胆汁酸含量呈明显正相关。

第二部分 婴儿双歧杆菌对结肠炎症及肠道菌群-胆汁酸代谢的影响

目的：分析婴儿双歧杆菌干预对小鼠结肠炎症、肠道菌群、胆汁酸代谢及结肠组织转录组的影响，初步揭示婴儿双歧杆菌通过调控肠道菌群及胆汁酸代谢来缓解结肠炎症。方法：C57BL/6小鼠随机分成正常对照组、DSS建模组以及婴儿双歧杆菌灌胃组，DSS建模组予建模的同时予以生理盐水连续灌胃7天，婴儿双歧杆菌灌胃组予以建模的同时予以婴儿双歧杆菌连续灌胃7天。建模结束后，评估小鼠炎症情况，并分别收集第0天和第7天的新鲜粪便标本进行16S rDNA测序和粪便胆汁酸靶向测定，处死小鼠留取结肠组织行RNA测序，多组学联合分析婴儿双歧杆菌在结肠炎发病过程中的作用。结果：与DSS建模组相比，婴儿双歧杆菌灌胃组小鼠疾病活动度评分和结肠组织学评分均显著降低（P值分别为0.009、0.046），结肠炎症明显缓解。16S测序数据分析：婴儿双歧杆菌灌胃组小鼠粪便菌群alpha多样性较DSS建模组无明显变化（P值分别为0.200, 0.343，0.343，0.486），PCoA分析提示两组粪便菌群存在一定差异。与DSS建模组相比，婴儿双歧杆菌灌胃组粪便Bifidobacterium和Lactobacillus菌属丰度呈现增加趋势，Bacteroides、Enterococcus、Escherichia-Shigella、Clostrium_sensu_stricto_1和Pasterurella菌属丰度呈下降趋势，但差异均无统计学意义（P值分别为0.886、0.343、0.886、0.886、0.514、0.886和>0.999）。粪便胆汁酸靶向代谢组学分析：与DSS建模组相比，婴儿双歧杆菌灌胃组的粪便总胆汁酸、非结合型胆汁酸和次级胆汁酸均呈升高趋势，但差异无统计学意义（P值分别为0.114，0.114和0.200）。其中，DSS建模组粪便胆汁酸THDCA含量较正常对照组降低（P=0.016），而婴儿双歧杆菌灌胃组THDCA含量显著高于DSS建模组（P=0.029）。结肠组织转录组学分析：婴儿双歧杆菌灌胃组和DSS建模组RNA-seq差异表达分析显示两组间有450个差异表达基因（Differential Expression Genes, DEGs），其中在婴儿双歧杆菌灌胃组上调的DEG有401个，下调的DEG有49个。功能富集分析显示多种免疫炎症相关的生物学过程如白细胞迁移、白细胞趋化和白细胞-细胞黏附等，以及KEGG途径如细胞因子-细胞因子受体相互作用、IL-17信号途径和细胞黏附分子途径均参与了婴儿双歧杆菌介导的肠道炎症缓解过程。进一步，通过GEO数据库分析胆汁酸受体在UC患者临床样本中的表达情况，法尼酯X受体（Farnesoid X Receptor，FXR）在UC患者结肠组织中表达水平显著降低。同时，Western blot结果显示DSS建模组小鼠结肠FXR的表达水平明显低于正常对照组，而婴儿双歧杆菌灌胃组FXR表达水平明显高于DSS建模组。结论：婴儿双歧杆菌干预显著缓解DSS诱导的小鼠结肠炎症。多组学分析结果显示婴儿双歧杆菌对结肠炎小鼠的肠道菌群和胆汁酸代谢均有一定的调节作用。RNA-seq显示婴儿双歧杆菌对结肠炎小鼠肠道的免疫炎症相关信号传导途径产生了显著影响。胆汁酸受体FXR在UC患者和结肠炎小鼠的结肠组织中均表达下调，而婴儿双歧杆菌干预能上调结肠炎小鼠肠道FXR表达水平。FXR可能在婴儿双歧杆菌介导的结肠炎症缓解中发挥重要作用。

第三部分 婴儿双歧杆菌通过FXR/CCL21轴参与小鼠结肠炎症缓解

目的：分析胆汁酸受体FXR对急性结肠炎小鼠肠道菌群和胆汁酸代谢的影响，探讨FXR与其下游靶基因相互作用参与缓解结肠炎症的潜在机制。方法： 选取C57BL/6野生型小鼠，将其随机分成野生型DSS建模组和野生型婴儿双歧杆菌灌胃组，另选取FXR敲除小鼠随机分成敲除鼠DSS建模组和敲除鼠婴儿双歧杆菌灌胃组。四组小鼠均予2.5% DSS饮用水5天诱导结肠炎，同时对野生鼠和敲除鼠婴儿双歧杆菌灌胃组均予婴儿双歧杆菌灌胃、DSS建模组均予生理盐水灌胃，连续7天。评估小鼠炎症情况，分别收集四组小鼠第0天和第7天的新鲜粪便标本进行16S rDNA测序和粪便胆汁酸靶向测定，处死小鼠后留取结肠组织行RNA测序。联合多组学分析FXR在婴儿双歧杆菌缓解结肠炎症过程中的作用。通过生物信息学分析预测FXR的靶基因，并在rhTNF-α刺激诱导的炎症细胞模型中初步验证激活FXR对候选靶基因的影响，进一步通过CUT&Tag技术验证了二者之间的靶向结合。结果：FXR敲除后，婴儿双歧杆菌灌胃组小鼠的体重、疾病活动度评分和结肠组织学评分较DSS建模组均无明显变化（P值分别为0.185、0.421、0.857），结肠炎症无显著缓解。16S测序数据分析：实验第0天，FXR敲除小鼠与野生型小鼠相比，粪便菌群alpha多样性明显降低；PCoA分析提示粪便菌群组成存在明显差异；LEfSe分析发现，与野生鼠相比，FXR敲除小鼠中Ruminococcaceae_UCG_014菌属显著富集，而Bifidobacterium、Turicibacter菌属则显著减少。实验第7天，与敲除鼠DSS建模组相比，敲除鼠婴儿双歧杆菌灌胃组粪便菌群alpha多样性无明显变化；PCoA分析显示两组间粪便菌群略有区分，但不明显。LEfSe分析发现在FXR敲除鼠诱导结肠炎后，Turicibacter、Escherichia-Shigella、Romboutsia、Enterococcus、Akkermansia菌属丰度增加，而Jeotgalicoccus和Lachnoclostridium菌属丰度减少，其中Turicibacter、Escherichia-Shigella、Romboutsia和Akkermansia菌属丰度在婴儿双歧杆菌灌胃后呈降低趋势，但差异无统计学意义。粪便胆汁酸靶向代谢组学分析：实验第0天，与野生型小鼠相比，FXR敲除小鼠粪便次级与初级胆汁酸比值降低（P=0.005）；第7天，FXR敲除鼠诱导结肠炎后，粪便结合型胆汁酸和初级胆汁酸水平升高（P值分别为0.049，0.007），非结合与结合型胆汁酸比值、次级与初级胆汁酸比值降低（P值分别为0.028，0.014）；敲除鼠婴儿双歧杆菌灌胃后粪便结合型胆汁酸和初级胆汁酸水平呈降低趋势，非结合与结合型胆汁酸比值呈增加趋势，差异均无统计学意义（P值分别为0.057，0.629，0.229），次级与初级胆汁酸比值则无增加趋势。实验第0天，通过FXR敲除小鼠和野生型小鼠的粪便胆汁酸差异分析筛选出5种差异胆汁酸，分别为βMCA、dehydroLCA、isoLCA、alloLCA和LCA。实验第7天，FXR敲除后，婴儿双歧杆菌灌胃组和DSS建模组间无差异显著的粪便胆汁酸。肠道转录组学分析：敲除鼠DSS建模组与野生型DSS建模组之间有471个DEGs，其中上调的DEGs 有242个，下调的DEGs有229个。功能富集分析显示两组间有19条差异显著的KEGG途径，其中包括数条免疫炎症相关的信号通路如细胞黏附分子、细胞因子与细胞因子受体相互作用和白细胞跨内皮迁移等。生物信息学分析筛选出7个FXR候选靶基因，包括CCL21、TIE1、CDC25B、MMRN2、SLCLA3、PLIN1和SLC1A1。进一步，采用rhTNF-α刺激建立炎症细胞模型，对上述候选靶基因进行验证，qRT-PCR结果显示CCL21在rhTNF-α刺激的24h显著升高（P=0.0001），同时予以FXR激动剂GW4064处理后CCL21的mRNA水平明显降低（P＜0.0001）。CUT&Tag-qPCR结果显示，实验组CCL21的水平显著高于IgG阴性对照组。结论：FXR敲除后，婴儿双歧杆菌干预对小鼠结肠炎症的缓解作用减弱。FXR敲除对小鼠的肠道菌群和胆汁酸代谢均产生了显著影响，FXR敲除后婴儿双歧杆菌干预对结肠炎小鼠肠道菌群和胆汁酸代谢的调节作用减弱。生物信息学分析提示FXR可能通过靶向调控CCL21的表达来缓解结肠炎症，细胞水平结果证实激活FXR能抑制炎症细胞模型中CCL21的表达，CUT&Tag实验进一步验证了二者之间的靶向结合。

第四部分 溃疡性结肠炎炎癌转化的机制初探

目的：探讨UC癌变过程中基因表达及生物学过程的早期异常变化；分析FXR对AOM/DSS诱导的炎癌小鼠肠道菌群的影响，初步探讨FXR对肠道菌群的影响在小鼠结肠炎癌转化中的作用。方法：GEO数据库挖掘方面，通过加权基因共表达网络分析探讨UC癌变过程中早期基因表达及生物学过程的异常变化。动物实验方面，选取野生型C57BL/6小鼠和FXR敲除小鼠，随机分成野生型对照组、野生型建模组和敲除鼠建模组。野生型对照组予常规饲养不加干预，野生型建模组和敲除鼠建模组均予AOM腹腔注射+3轮2%DSS饮用水诱导炎癌模型，分别收集实验前后的新鲜粪便标本进行宏基因组测序及分析，12周后处死取材，评估结肠成瘤情况。结果：本研究纳入了GSE47908数据集中的13例正常对照、18例UC和6例结肠炎相关不典型增生（colitis associated dysplasia, CAD）患者的结肠基因表达数据。CAD组和正常对照组差异分析筛选出4929个DEGs（log₂ FC阈值为0.3），基于4929个DEGs的WGCNA将其分为6个基因模块（灰色模块除外），其中绿色和蓝色模块分别与UC和CAD呈现方向一致的相关性，分别与表型Progress之间具有最高的负相关系数（r = -0.75, P=9e-08）和正相关系数（r = 0.71, P = 9e-07）。绿色模块和蓝色模块基因的功能富集分析提示，线粒体功能、细胞-细胞间连接和免疫反应相关GO-BP的改变与UC炎癌转化密切相关。绿色模块和蓝色模块共筛选出13个关键基因，包括SLC25A3、ACO2、AIFM1、ATP5A1、DLD、TFE3、UQCRC1、ADIPOR2、SLC35D1、TOR1AIP1、PRR5L、ATOX1和DTX3。动物实验中，FXR敲除鼠建模组的体重较野生型建模组明显下降，结肠成瘤总数目和长径分别为1-2mm和>2mm的瘤体数目均显著高于野生型建模组。FXR敲除鼠与野生鼠粪便菌群宏基因组测序分析：建模前FXR敲除鼠与野生鼠相比，肠道菌群基因数目明显下降（P=0.002），NMDS分析显示两组间菌群组成结构发生明显改变，LEfSe分析筛选出差异菌属21个，差异菌种67个；功能丰度差异分析筛选出128条差异显著的KEGG途径。野生鼠AOM/DSS建模前与建模后粪便菌群宏基因组测序分析：野生鼠建模后粪便菌群基因数目较建模前显著下降（P=0.004），NMDS分析显示菌群组成结构发生明显改变，LEfSe分析筛选出差异菌属23个，差异菌种64个；功能丰度差异分析筛选出126差异显著的KEGG途径。上述两次宏基因组LEfSe差异分析和功能丰度差异分析分别取交集，筛选出13个共同的差异菌属，36个共同的差异菌种和71条共同的KEGG途径。13个菌属分别为：Mucispirillum、Prevotella、Unclassified_Lachnospiraceae、Roseburia、Oscillibacter、Clostridioides、Bilophila、Unclassified_Firmicutes、Butyrivibrio、Eubacterium、Ruminococcus、Acetivibrio和Flavonifractor，上述菌属均为野生型炎癌小鼠中丰度显著减少的菌属，而FXR敲除后其丰度也明显减少。结论：WGCNA分析发现线粒体功能障碍、细胞-细胞间连接和免疫调节异常可能是促进炎癌转化进程的重要生物学过程，其中涉及的关键基因包括SLC25A3、ACO2、AIFM1、ATP5A1、DLD、TFE3、UQCRC1、ADIPOR2、SLC35D1、TOR1AIP1、PRR5L、ATOX1和DTX3。FXR敲除促进小鼠结肠的炎癌转化。AOM/DSS诱导结肠炎症炎癌转化过程中，宏基因测序分析发现小鼠粪便菌群基因数目减少，菌群组成结构及功能发生明显改变。敲除FXR后，小鼠粪便菌群的组成结构及功能也发生紊乱。进一步分析显示，13个菌属和36个菌种同时受到FXR敲除和AOM/DSS诱导炎癌的影响。

Part 1  Intestinal dysbiosis and bile acid dysmetabolism of colitis in mice

Objective: To explore the correlation between the intestinal dysbiosis and bile acid metabolic disorders in mice with acute colitis. Methods: C57BL/6 mice were randomly divided into normal control group and DSS group. The normal control group received no intervention. The DSS group was treated with 2.5% DSS drinking water for 5 days to induce colitis, and sacrificed after 7 days. Fresh stool specimens on day 0 and day 7 of the DSS group were collected for 16S rDNA sequencing and fecal bile acid targeted determination. Analyze the characteristics of intestinal microbiota dysbiosis and bile acid metabolic disorders, and conduct correlation analysis between them. Results: 16S rDNA sequencing analysis: Compared with that before modeling, the alpha diversity of the fecal flora of DSS-induced colitis mice was significantly reduced and the diversity index chao1, observed_species, PD_whole_tree, and shannon index reduced significantly (P values ​were 0.003, 0.001, 0.041 and 0.043, respectively). PCoA analysis showed that the fecal flora of mice with colitis was significantly different from that before modeling. After DSS modeling, the composition and structure of mouse fecal flora changed significantly. Further LEfSe discriminant analysis showed that Lactobacillus, Bifidobacterium and Staphylococcus were significantly reduced after DSS modeling, and Turicibacter, Lachnospiraceae_NK4A136_group, Bacteroides, Enterococcus, Escherichia-Shigella, Clostridium_sensu_stricto_1 and Pasteurella were significantly increased after DSS modeling. PICRUST2 function prediction and PCA analysis indicated that the function of mouse intestinal flora changed significantly after DSS modeling. KEGG analysis identified 43 significantly changed signal pathways, among which the top 5 pathways are Biosynthesis of ansamycins, Secondary bile acid biosynthesis, D-Glutamine and D-glutamate metabolism, D-Alanine metabolism and Fatty acid biosynthesis pathways. Fecal bile acid targeted metabolomics analysis: PCA analysis showed that the fecal bile acid profile of colitis mice changed significantly compared with that before DSS modeling. The total bile acid content was significantly reduced (P=0.016). Specifically, the unconjugated bile acid, primary bile acid and secondary bile acid were significantly reduced (P values ​​were 0.017, 0.008 and 0.043), the ratio of secondary bile acid to the primary bile acid decreases (P=0.032). According to the VIP value (VIP>1) based on OPLS-DA model and the P value of significance test (P<0.05), the differential bile acids before and after DSS modeling were jointly screened. A total of 11 differential bile acids were identified including 12-ketoLCA, TUDCA, TDCA, THDCA, isoDCA, βUDCA, TCDCA, 6-ketoLCA, TCA, TLCA and NorDCA. Pearson correlation analysis showed that the abundance of Bifidobacterium and Lactobacillus were significantly positively correlated with the content of most differential bile acids. Conclusion: The structure and function of the fecal flora of mice with DSS-induced colitis changed significantly, mainly manifested as decreased diversity, decreased beneficial bacteria, increased harmful bacteria and conditional pathogenic bacteria, and decreased secondary bile acid synthesis function. The fecal bile acid profile of colitis mice changed significantly, the total bile acid level decreased, and the ratio of secondary to primary bile acid decreased. The abundance of Bifidobacterium and Lactobacillus in the feces of mice with colitis was significantly positively correlated with the content of most differential secondary bile acids.

Part 2  Effect of Bifidobacterium infantis on colonic inflammation and intestinal flora-bile acid metabolism

Objective: To analyze the effects of Bifidobacterium infantis on colon inflammation, intestinal flora, bile acid metabolism and intestinal transcriptome in mice, and to preliminarily reveal that Bifidobacterium infantis can relieve colon inflammation by regulating the intestinal flora and bile acid metabolism. Methods: C57BL/6 mice were randomly divided into normal control group, DSS group and Bifidobacterium infantis group. The DSS group was treated with saline by gavage for 7 days while modeling, and the Bifidobacterium infantis group was treated with Bifidobacterium infantis by gavage for 7 days while modeling. Evaluate the colonic inflammation of the mice, and collect fresh stool samples on day 0 and day 7 for 16S rDNA sequencing and fecal bile acid targeted determination. The mice were killed and their colon was collected for RNA sequencing. Multi-omics analysis was performed to explore the role of Bifidobacterium infantis in the pathogenesis of colitis. Results: Compared with the DSS group, the disease activity score and colon histological score of the Bifidobacterium infantis group were reduced (P values ​​were 0.009, 0.046), and the colon inflammation is significantly relieved. 16S rDNA sequencing data analysis: The alpha diversity of the fecal flora of mice in the Bifidobacterium infantis group did not change significantly compared with the DSS group (P values ​​were 0.200, 0.343, 0.343 and 0.486,​​respectively). PCoA analysis indicated that there were some differences in the fecal flora between the two groups. The abundance of Bifidobacterium and Lactobacillus in the fecal of the Bifidobacterium infantis group showed an increasing trend compared with the DSS group, and the abundance of Bacteroides, Enterococcus, Escherichia-Shigella, Clostrium_sensu_stricto_1 and Pasterurella had a decreasing trend, but the differences were not statistically significant (P values were 0.886, 0.343, 0.886 0.886, 0.514, 0.886 and >0.999, ​​respectively). Fecal bile acid targeted metabolomics analysis: Compared with the DSS group, the total fecal bile acids, unconjugated bile acids, and secondary bile acids of the Bifidobacterium infantis group showed an increasing trend, but the differences were not statistically significant (P value were 0.114, 0.114 and 0.200, respectively). Compared with the normal control group, the fecal bile acid THDCA content of the DSS group was lower (P=0.016), while the Bifidobacterium infantis group had a significantly higher THDCA content than the DSS group (P=0.029). Intestinal transcriptomics analysis: RNA-seq and differential expression analysis of Bifidobacterium infantis group and DSS group showed that there were 450 differential expression genes (Differential Expression Genes, DEGs) between the two groups, including 401 up-regulated and 49 down-regulated DEGs. Functional enrichment analysis shows that there are a variety of biological processes related to immune inflammation, such as leukocyte migration, leukocyte chemotaxis and leukocyte cell-cell adhesion, etc., and KEGG pathways such as cytokine-cytokine receptor interaction, IL-17 signaling pathway and cell adhesion molecules, participating in the relieving process of colon inflammation mediated by Bifidobacterium infantis. GEO database analysis showed that the bile acid receptor FXR was significantly reduced in UC patients. Western blot results showed that the expression level of colonic FXR was significantly lower than that of the normal control group, while the expression level of FXR in the Bifidobacterium infantis group was significantly higher than that of the DSS group. Conclusion: Bifidobacterium infantis can alleviate DSS-induced colonic inflammation in mice. The results of multi-omics analysis showed that Bifidobacterium infantis can regulate the intestinal flora and bile acid metabolism of colitis mice. RNA-seq showed that Bifidobacterium infantis had a significant impact on the immune and inflammation-related signal transduction pathways in colon of colitis mice. The expression of bile acid receptor FXR is down-regulated in the colon tissues of UC patients and colitis mice, while the intervention of Bifidobacterium infantis can up-regulate the colonic expression of FXR in colitis mice. FXR may play an important role in the alleviation of colon inflammation mediated by Bifidobacterium infantis.

Part 3  Bifidobacterium infantis participates in the alleviation of  colonic inflammation via FXR/CCL21 axis in mice

Objective: Analyze the effect of bile acid receptor FXR on intestinal flora and bile acid metabolism of colitis mice, and explore the potential mechanism of the interaction between FXR and its downstream target genes in alleviating colon inflammation. Methods: Wild-type C57BL/6 mice were divided into WT DSS group and WT Bifidobacterium infantis group, FXR knockout mice were divided into KO DSS group and KO Bifidobacterium infantis group. The four groups of mice were given 2.5% DSS drinking water for 5 days to induce colitis. The WT and KO Bifidobacterium infantis groups were treated with Bifidobacterium infantis by gavage for 7 days, and the DSS groups were treated with saline by gavage. Evaluate the colon inflammation and collect the fresh fecal specimens on day 0 and day 7 for 16S rDNA sequencing and fecal bile acid targeted determination. The mice were killed on day 7 and colon tissues were collected for RNA sequencing. Multi-omics analysis was conducted to explore the role of FXR in mediating the protective effect of Bifidobacterium infantis on colon inflammation. Bioinformatics analysis was performed to predict FXR target genes, and preliminarily verifies the effect of FXR on candidate target genes in an inflammatory cell model induced by rhTNF-α. Then，CUT&Tag and qPCR was conducted to verify the targeted binding between FXR and CCL21. Results: The weight, disease activity score and colonic histological score of the KO Bifidobacterium infantis group had no significant changes compared with the KO DSS group (P values ​​were 0.185, 0.421 and 0.857, respectively), and the colon inflammation was not relieved. 16S rDNA sequencing data analysis: On day 0, compared with WT groups, the alpha diversity of the fecal flora of KO groups was significally reduced; PCoA analysis showed that there was a significant difference in the composition of the fecal flora between the two groups; LEfSe analysis found that compared with WT groups, Ruminococcaceae_UCG_014 was significantly enriched in KO groups, while Bifidobacterium and Turicibacter were significantly reduced. On day 7, compared with the KO DSS group, the alpha diversity of fecal flora of the KO Bifidobacterium infantis group did not change significantly; PCoA analysis showed that there was a slight distinction of the fecal flora between the two groups. LEfSe analysis found that, for KO DSS group, the abundance of Turicibacter, Escherichia-Shigella, Romboutsia, Enterococcus, and Akkermansia increased, while the abundance of Jeotgalicoccus and Lachnoclostridium decreased, among which the abundance of Turicibacter, Escherichia-Shigella, Romboutsia and Akkermansia showed a decreasing trend after administration of Bifidobacterium infantis, but the difference was not statistically significant. Fecal bile acid targeted metabolomics analysis: On day 0, compared with WT groups, the ratio of secondary bile acid to primary bile acid in FXR KO mice obviously decreased (P=0.005)；On day 7, for KO DSS group, the levels of fecal conjugated bile acid and primary bile acid increased significantly (P values ​​were 0.049 and 0.007, respectively), the ratio of unconjugated to conjugated bile acid and the ratio of secondary to primary bile acid decreased significantly  (P values ​​were 0.028 and 0.014, respectively). For KO Bifidobacterium infantis group, the levels of fecal conjugated bile acid and primary bile acid showed a decreasing trend, and the ratio of unconjugated to conjugated bile acid showed an increasing trend, but their differences were not statistically significant (P value were 0.057, 0.629 and 0.229, respectively), and the ratio of secondary to primary bile acid showed no increasing trend. On day 0, the fecal bile acid differential analysis of KO groups and WT groups identified five differential bile acids, including βMCA, dehydroLCA, isoLCA, alloLCA and LCA. On day 7, there was no significant difference in fecal bile acid between the KO Bifidobacterium infantis group and KO DSS group. Intestinal transcriptomics analysis: There are 471 DEGs between the KO DSS group and WT DSS group including 242 up-regulated DEGs and 229 down-regulated DEGs. Functional enrichment analysis showed that there were 19 significantly differential KEGG pathways between the two groups, including several immuno-inflammation-related signaling pathways such as cell adhesion molecules, cytokine-cytokine receptor interaction, and leukocyte transendothelial migration. Bioinformatics analysis screened out 7 FXR candidate target genes, including CCL21, TIE1, CDC25B, MMRN2, SLCLA3, PLIN1 and SLC1A1. Preliminary verification of CCL21 on HT29 cell level: An inflammatory cell model was established by rhTNF-α (100ng/ml) stimulation. qRT-PCR showed that CCL21 was significantly increased when incubated with rhTNF-α for 24h (P=0.0001). Treatment with FXR agonist GW4064 can significantly reduce the mRNA level of CCL21 in inflammatory cell model (P＜0.0001). CUT&Tag-qPCR results showed that the level of CCL21 in experimental group was significantly higher than that of the IgG negative control group. Conclusion: After FXR knockout, the alleviating effect of Bifidobacterium infantis on colitis in mice is weakened. FXR knockout has a significant effect on intestinal flora and bile acid metabolism of mice. After FXR knockout, the effect of Bifidobacterium infantis on intestinal flora and bile acid metabolism of colitis mice is weakened. Bioinformatics analysis suggests that FXR may alleviate colonic inflammation by targeting CCL21 expression. Results at the cellular level confirm that activating FXR can inhibit the expression of CCL21 in inflammatory cell models. Further, the targeted binding between FXR and CCL21 was verified by CUT&Tag .

Part 4  Preliminary investigation of the transformation mechanisms from ulcerative colitis to carcinogenesis

Objective: To explore the early abnormal changes of gene expression and biological processes in the process of UC carcinogenesis. To analyze the characteristics of intestinal dysbiosis in the mouse model of colitis-associated colorectal cancer induced by AOM/DSS, and to preliminarily explore the effect of FXR on intestinal flora and the role of this effect in the transformation from inflammation to cancer in mice. Methods: Through GEO database mining and WGCNA, the early abnormal changes of gene expression and biological processes in the process of UC carcinogenesis are explored. As for animal experiments, wild-type C57BL/6 mice were divided into WT control group and WT AOM/DSS group, FXR knockout mice were used as KO AOM/DSS group. The WT control group was given no intervention, and the WT and KO AOM/DSS groups were administered AOM (12.5mg/kg) intraperitoneally and three cycles of 2% DSS drinking water orally to induce colitis-associated colorectal cancer. Fresh stool samples before and after the experiment were collected for metagenomic sequencing. The mice were sacrificed after 12 weeks and the formation of tumors was evaluated. Results: A total of 13 controls, 18 UC and 6 CAD in GSE47908 were enrolled in WGCNA. Six functional modules were identified based on 4929 DEGs between CAD and control group (log₂ FC cut-off is 0.3). Green and blue modules were selected because of their consistent correlation with UC and CAD, and the highest correlation coefficient with the progress of UC-associated carcinogenesis. Functional enrichment analysis revealed that genes of these two modules were significantly enriched in biological processes, including mitochondrial dysfunction, cell-cell junction, and immune responses. Thirteen hub genes (SLC25A3, ACO2, AIFM1, ATP5A1, DLD, TFE3, UQCRC1, ADIPOR2, SLC35D1, TOR1AIP1, PRR5L, ATOX1, and DTX3) were identified. As for animal experiments，the weight loss of the KO AOM/DSS group was more significant than that of the WT AOM/DSS group. The total number of colon tumors and both of the number of tumors with a long diameter of 1-2mm and >2mm were higher than that of the WT AOM/DSS group. Metagenomic sequencing analysis of fecal flora of FXR knockout mice and wild type mice: Before modeling, FXR knockout mice had a significant decrease in the gene number of flora compared with wild mice (P=0.002). NMDS analysis showed that the composition of the flora was significantly changed in FXR knockout mice, and LEfSe analysis identified 21 differential bacterial genera and 67 differential bacterial species. Then, 128 KEGG pathways with significant differences between FXR knocked mice and wild type mice were selected by differential analysis of functional abundance. Metagenomic sequencing analysis of the fecal flora of wild mice before and after AOM/DSS modeling: The number of fecal flora genes of wild mice after modeling was significantly lower than that before modeling (P=0.004). NMDS analysis showed that the composition of intestinal flora changed obviously after modeling, and LEfSe analysis screened out 23 differential genera and 64 differential species. Functional abundance differential analysis identified 126 significantly differential KEGG pathways. Take the intersection of the above two LEfSe differential analysis, then we found 13 common differential genera and 36 common differential species. Similarly, we also take the intersection of the above two functional abundance differential analysis and then identified 71 common KEGG pathways. The 13 bacterial genera include Mucispirillum, Prevotella, Unclassified_Lachnospiraceae, Roseburia, Oscillibacter, Clostridioides, Bilophila, Unclassified_Firmicutes, Butyrivibrio, Eubacterium, Ruminococcus, Acetivibrio, and Flavonifractor, the abundance of all of which was significantly changed after modeling in wild-type mice, and was also significantly changed in FXR knocked out mice compared with wild type mice before modeling. Conclusion: GEO database mining and WGCNA reveals that the abnormal changes of  mitochondrial function, cell-cell junction, and immune responses may play crucial roles in the process of UC carcinogenesis. And 13 hub genes were identified as the potential key genes, including SLC25A3、ACO2、AIFM1、ATP5A1、DLD、TFE3、UQCRC1、ADIPOR2、SLC35D1、TOR1AIP1、PRR5L、ATOX1, and DTX3. FXR knockout promotes the transformation from colitis to cancer. Both FXR knockout and AOM/DSS-induced colitis-associated colorectal cancer can reduce the number of genes in fecal flora of mice, and significantly change the structure and function of the flora. Metagenomic sequencing analysis revealed that 13 bacterial genera and 36 bacterial species were simultaneously affected by FXR knockout and AOM/DSS modeling.