研究背景:糖尿病肾病(Diabetes Nephropathy, DN)是糖尿病最常见的微血管并发症，见于 30%-40%的糖尿病患者，是导致终末期肾病(End stage renal disease，ESRD)的重要病 因，也是预测慢性肾脏病(Chronic Kidney Disease, CKD)患者死亡的独立危险因素。 控制血糖、血压及抑制肾素-血管紧张素-醛固酮系统(Renin-Angiotensin System，RAS)虽然有助于降低DN发病并延缓其进展，但并未改善DN的诊疗现状，迫切需要加强机制研究，寻找新的防治策略。近期研究发现，固有免疫识别内源性损伤信号和外源性抗原分子相关的慢性炎症在DN中发挥重要作用。动物实验中肠道异常增殖的致病微生物，可直接被肠固有免疫细胞识别，并可通过白介素22(Interleukin 22, IL22)减轻肺的炎症反应；而IL22减轻DN肾小管间质损伤在动物实验中已得到证实，那么肠道菌群变化是否可能也通过IL22参与 DN的发生发展？肠道菌群与药物之间的相互作用备受关注，药物可能导致肠道菌群失调，而肠道菌群失调又反过来影响降糖药物的疗效和副作用。肉桂醛(Cinnamaldehyde, CIN) 是新型天然降糖药物，已有报道可减轻DN蛋白尿；同时具有抗菌作用，那么其对DN的保护作用，是否是通过调节肠道菌群，进而影响肾脏固有免疫及IL22相关炎症反应而实现？因此本研究将基于DN大鼠模型，探索DN进展过程中的肠道菌群结构与功能改变，寻找可能的关键微生物。并观察肉桂醛干预对DN大鼠肠道菌群的影响，初步探究其可能的机制。研究目的: 1. 建立 DN 大鼠模型，观察 DN 进展过程中的肠道菌群改变，及其与肾损伤(蛋白尿和病理改变)的相关性，寻找可能参与发病的关键微生物；2. 肉桂醛治疗DN大鼠，观察其对肾损伤和肠道菌群的影响，并初步观察其可能的作用机制。研究方法: 1. DN大鼠模型构建及肉桂醛干预:同周龄、血糖、体重相匹配的大鼠，随机分成 4 组，16 只/组：(1)糖尿病肾病(DN)组：通过注射链脲佐菌素(Streptozotocin，STZ)构建 DN 大鼠模型，72h 后，测 定空腹血糖≥16.7mmol/L 者确定纳入 DN 组；(2)对照(Normal Control，NC)组：注射等量 1%柠檬酸钠缓冲液，72h 后空腹血糖＜7mmol/L 者确定纳入 NC 组；(3)DN+ 肉桂醛( DNC )组：建模同 DN 组，第 4 天开始每天肉桂醛灌胃；(4)NC+肉桂醛( NCC )组：建模 NC 组，第 4 天开始每天肉桂醛灌胃。0d、4 周、8 周、12 周为观察时间点，监测血糖、体重、24h 尿蛋白(24-hour Urinary Protein，24hUP)，24h 尿量(24-hour Urinary Volume, 24hUV), 留新鲜粪便。分别在建模 8 周后(8 只/组，分别为 NC8、 DN8、NCC8、DNC8 组)、12 周后(8 只/组，分别为 NC12、DN12、NCC12、DNC12组)处死，留取血、肾组织标本。通过HE、PAS、Masson染色及电镜评估肾病理损 伤。2. 肠道菌群 16s rDNA 测序通过可操作性分类单位(Operational Taxonomic Units，OTU)聚类分析各组间共有和特有的 OUT。α多样性分析：Chao1 指数和 ACE 指数评估群落丰富度，Shannon 指数和 Simpson 指数评估群落多样性。β多样性分析：计算 Weighted Unifrac 距离，并行降纬主坐标分析(Principal Co-ordinates Analysis，PCoA)。用 Metastats 和线性判 别分析(Linear Discriminant Analysis，LDA)方法分析组间差异物种。用 Spearman 进行肠道菌群物种丰度与肾损伤指标的相关性分析。并通过Tax4Fun进行肠道菌群功 能预测。3. 肉桂醛改善 DN机制探索:通过免疫组化及半定量分析检测各组大鼠肾脏 Megalin 蛋白表达情况；通过 Western Blot 检测各组大鼠肾脏固有免疫通路 IL22/IL17A，以及纤维化指标Fibronectin、TGF-β蛋白的表达。比较 NC8 与 DN8 组，DN8 与 DNC8 组，NC8 与NCC8组，组间蛋白表达量差异。4. 统计方法: 计量数据使用 Kolmogorov-Smirnov 检验进行正态性检验。符合正态分布的计 量数据以均值±标准差(mean± SD)表示，非正态的计量数据以中位数(四分位距)表 示，即Median(Q25，Q75)。比较两组差异：对于正态分布，并且方差齐的计量数据， 用t 检验；对于非正态分布、方差不齐的计量数据，用非参数Mann-Whitney 检验。 双侧检验p<0.05 为具有统计学差异。研究结果:一、 DN大鼠肠道菌群改变特点:1. 建立 DN 大鼠模型:与对照组相比，STZ 注射 72h 后，DN 组大鼠血糖明显升高，且出现多饮、多 尿、多食等表现。建模后第4周开始，与NC组相比，DN组体重显著降低(p＜0.001)，血糖及 24hUV 均显著增高(p＜0.001)。建模第 8 周开始，DN 组 24hUP 显著高于 NC组(p＜0.001)，说明 DN 大鼠模型构建成功。 建模 8 周后，DN 组开始出现肾脏病理损伤，光镜下表现为肾小球增大，伴系膜细胞及基质增生；肾小管肿胀、管腔狭窄。电镜下可见肾小球基底膜增厚，足突融合；肾小管基底膜增厚、线粒体肿胀、部分刷状缘脱落。2. DN大鼠的肠道菌群特点: 肠道菌群结构：OTU分析示第8、12周的 DN 组共享764个OTU，特有的 OTU分别为 159 个(17.2%)、218 个(22.2%)，提示随着 DN 进展，肠道菌群种类会发生改变。与对照相比，DN8周的物种多样性、丰富度无明显改变；DN12周物种多样性，Simpson 指数显著降低(0.95±0.01 vs 0.90±0.01，p＜0.05)。β 多样性分析提示，与NC8 周与 DN8 周的组间差异相比，NC12 周与 DN12 周的组间群落结构差异更大。差异物种：DN8周与同龄对照组间差异小，而DN12周与同龄对照组间差异大。Metastats 分析显示，NC8 周与 DN8 周之间，共 1 个门、1 个纲、0 个目、0 个科、 0 个属、9 个种存在显著差异(q＜0.05)。NC12 周与 DN12 周之间，共 3 个门、7 个 纲、10个目、18个科、19个属、14个种存在显著差异。而第8、12周的DN组间，共8个门、10个纲、10个目、16个科、39个属、30个种存在显著差异。说明随着DN 进展，肠道菌群结构发生显著改变。其中，与同龄对照组相比，解纤维素拟杆菌(s__Bacteroides_cellulosilyticus)与多形拟杆菌(s__Bacteroides_thetaiotaomicron)在DN8、DN12组均显著性降低。菌群相关性分析：有13个属、12个种的相对丰度与24hUP显著相关(p＜0.05)，其中解纤维素拟杆菌(-0.39，p＜0.01)、多形拟杆菌(-0.42，p＜0.01)与 24hUP 显著负相关。肠道菌群功能：随着DN进展，肠道菌群功能发生显著改变。DNA修复和重组蛋白(DNA_repair_recombination_proteins)、tRNA 合成(Transfer_RNA_biogenesis)、嘌 呤代谢(Purine_metabolism)等通路在 DN12 周显著富集。二、肉桂醛对早期DN的影响及机制初探 1. 肉桂醛对 DN肾损伤的影响肉桂醛灌胃 8 周后，与 DN8 周相比，DNC8 周的大鼠 24hUP(86.65±10.08 vs 52.53±7.03 mg，p=0.02)、24hUV 显著降低，血糖、体重无显著差异。灌胃 12 周，与 DN12 周相比，DNC12 周大鼠 24hUP、24hUV、血糖及体重均无统计学差异。光 镜下，DN8周、DNC8周的大鼠肾病理损伤无显著差别。电镜下，与DN8周相比，DNC8 周的肾小管基底膜厚度显著降低(276.30±10.16 vs 239.50±12.43 nm，p=0.02)，两组间肾小球基底膜厚度无显著差异。2. 肉桂醛对DN大鼠肠道菌群的影响肠道菌群结构：肉桂醛干预使大鼠肠道菌群物种发生改变，DN8周与DNC8周共享 774 个 OTU，特有的 OTU 分别为 149 个(16.1%)、247 个(24.2%)。与 DN8 周相比，DNC8 周的 Simpson 指数显著降低(0.96±0.01 vs 0.94±0.01，p＜0.01)，提示肉桂醛会降低物种多样性，但不影响物种丰富度。通过PCoA分析直观显示，与NC8 周、DN8周之间的距离相比，NC8周、DNC8周之间的距离更近；提示肉桂醛可能部分恢复DN大鼠的群落组成结构。差异物种：与 DN8 相比，DNC8 中共 1 个门、2 个纲、0 个目、5 个科、13 个属、10 个种的物种丰度存在显著差异。属水平，乳杆菌属(g__Lactobacillus)、拟普雷沃菌属(g__Alloprevotella)、g__Oscillospira、g__Weissella 等益生菌在 DNC8 显著增加。种水平，与NC8周相比，共3个菌种在DN8周中降低，且经肉桂醛干预后，丰度恢复, 包括马赛拟杆菌(s__Bacteroides_massiliensis), s__Oscillibacter_sp_ER4， s__Lachnospiraceae_bacterium_A2。与 NC8 周相比，共 2 个菌种在 DN8 中升高，经 肉 桂 醛 干 预 后 ， 丰 度 降 低 ， 包 括 s__Ruminococcus_sp_ID1 、 s__Comamonas_denitrificans。功能预测：肉桂醛干预使得DN大鼠的肠道菌群功能发生显著改变。与DN8周相 比 ， 转 运 子 (Transporters) 、 嘌 呤 代 谢 (Purine_metabolism) 、 ABC 转 运 子 (ABC_transporters)等通路在 DNC8 周富集。3. 肉桂醛改善早期DN的机制初探免疫组化及半定量分析显示，与 NC8 周相比，DN8 周大鼠近端小管刷状缘的Megalin 表达显著减低(0.017±0.001 vs 0.006±0.001，p＜0.01)；与 DN8 周相比， DNC8 周的 Megalin 表达显著恢复(0.006±0.001 vs 0.011±0.001，p＜0.01)。与 NC8 周相比，DN8周的肾脏纤粘蛋白FN表达显著升高(1.00±0.28 vs 4.57±1.20，p=0.04)， 肾脏 IL22 表达显著降低(1.00±0.12 vs 0.61±0.11，p=0.04)，两组间 TGF-β无显著 差异(p=0.08)。肉桂醛干预的 DNC8 周，FN (4.57±1.20 vs 0.68±0.06，p=0.03)、TGF-β(1.85±0.37 vs 0.57±0.15，p=0.03)蛋白的表达均较 DN8 周显著降低，但 IL22(p=0.08)、IL17A(p=0.88)蛋白表达均无显著改变。结论: 本研究条件下：1. DN 大鼠存在病程依赖的肠道菌群失调。与对照相比，建模 12 周后物种多样性显著降低，且组间差异物种增多，与24hUP显著相关； 2. 肉桂醛部分改善 DN 大鼠的肠道菌群失调，可增加乳杆菌属(g__Lactobacillus)、拟普雷沃菌属(g__Alloprevotella)等抗炎益生菌的丰度；3. 肉桂醛可降低DN大鼠早期蛋白尿，改善其肾小管基底膜增厚，并且可增加近端小管 Megalin 的表达，抑制肾脏 FN、TGFβ，但不影响 IL22、IL17A 的表达。

BACKGROUND: Diabetic nephropathy (DN) is the most common microvascular complication of diabetes, which occurs in 30% - 40% of diabetic patients. It is an important cause of the end-stage renal disease (ESRD) and an independent risk factor for predicting the death of patients with chronic kidney disease (CKD). Although the control of blood glucose, blood pressure, and inhibition of the renin-angiotensin system (RAS) can help to reduce the incidence and delay the progress of DN, it does not improve the present situation of diagnosis and treatment in DN. It is urgent to strengthen the mechanism research and search for new prevention and treatment strategies. Recent studies have found that chronic inflammation played an important role in DN， may be caused by innate immune cells recognizing the endogenous damage signals and exogenous antigen molecules. In animal experiments, pathogenic microorganisms with abnormal proliferation can be recognized directly by intestinal innate immune cells, reducing the inflammatory response of the lung through interleukin 22 (IL22). It has also been confirmed that IL22 can alleviate renal tubulointerstitial injury in DN. Is it possible that the changes of intestinal flora may also participate in the occurrence and development of DN through IL22? The interaction between intestinal flora and drugs has attracted much attention. Drugs may lead to dysbiosis of intestinal flora, which in turn affects the efficacy and side effects of antidiabetic drugs. Cinnamaldehyde (CIN) is a new antidiabetic drug, which can reduce DN proteinuria. At the same time, it has an antibacterial effect. Is it possible that the protective effect on DN is realized by regulating intestinal flora and affecting IL22 related inflammatory response? Therefore, based on the DN rat model, this study will explore the changes of intestinal flora structure and function in the process of DN, and search for possible key microorganisms. And we will observe the effect of cinnamaldehyde on DN and intestinal flora，and initially explore its possible mechanism. OBJECTIVE: 1) To establish DN rat model and observe the changes of intestinal microflora during the progression of DN, also its correlation with renal injury (proteinuria and pathological changes), trying to find the key microorganisms that may participate in the pathogenesis; 2) To observe the effect of cinnamaldehyde on renal injury and intestinal flora in DN rats and its possible mechanism. METHODS: 1) Establishment of DN rat model and intervention of Cinnamaldehyde Rats with the same age, blood glucose, and body weight were randomly divided into 4 groups (16 rats/group): (1) Diabetic nephropathy (DN) group: DN rat model was established by injection of streptozotocin (STZ). After 72 hours, the rats with fasting blood glucose ≥ 16.7 mmol / L were included in DN group; (2) Normal control (NC) group: 72 hours after injection of 1% sodium citrate buffer，those with fasting blood glucose less than 7 mmol / L were included in NC group; (3) DN + cinnamaldehyde (DNC) group and (4) NC + cinnamaldehyde (NCC) group: the modeling was the same as above, and cinnamaldehyde was given by gavage every day from the 4th day. Blood glucose, bodyweight, 24-hour urinary protein (24hUP), 24-hour urinary volume (24hUV), and fresh feces were monitored at 0d, 4W, 8W, and 12W. At 8th weeks (NC8, DN8, NCC8, DNC8) or 12th weeks (NC12, DN12, NCC12, DNC12) after modeling, rats were sacrificed (8 rats/group). The blood and kidney tissue samples were collected. HE, PAS, Masson staining, and electron microscope were used to evaluate the renal pathological injury. 2) 16S rDNA sequencing of intestinal flora After clustering operational taxonomic units (OTUs), αdiversity was analyzed using Chao1 and ACE for richness, Shannon and Simpson for diversity. Andβdiversity was analyzed by Weighted Unifrac distance with principal co-ordinates analysis (PCoA). Metastats and linear discriminant analysis (LDA) were used to analyze the difference in intestinal flora among groups. Spearman was used to analyze the correlation between the abundance of intestinal flora and renal injury indicators. The function of intestinal flora was predicted by Tax4fun. 3) Mechanism of cinnamaldehyde alleviating DN The expression of the megalin protein in the kidney was detected by immunohistochemistry and semi-quantitative analysis. Western blot was used to detect the innate immune pathway IL22/IL17A, inflammation pathway caspase-1/IL18, and fibrosis pathway fibronectin and TGF- β protein expression. The differences in protein expression between NC8 and DN8 group，DN8 and DNC8 group, NC8 and NCC8 were analyzed. 4) Statistical analysis Kolmogorov Smirnov test was used to test the normality of measurement data. The continuous variable with normal distribution was represented by mean value ± standard deviation (mean ± SD); the non-normal distribution variable was represented by median (Q25, Q75). For the continuous variable with normal distribution, a t-test was used to compare the differences between the two groups; For those with non-normal distribution, the differences between the two groups were compared by non-parametric Mann Whitney test. The difference was statistically significant if bilateral test p< 0.05. RESULTS: 1. Characteristics of intestinal flora changes in diabetic nephropathy 1) Establishment of DN rat model Compared with the control group, 72 hours after STZ injection, the blood glucose of the DN group was significantly increased, showing polydipsia, polyuria and polydipsia in DN rats. From the 4th week after modeling, the bodyweight of the DN group was significantly lower than that of the NC group (p<0.001), and the blood glucose, 24hUV of the DN group were significantly higher than that of NC group (p<0.001). From the 8th week, the 24hUP of the DN group was significantly higher than that of the NC group (p<0.001), indicating that the DN rat model was successfully constructed. There was pathological damage of the kidney in the DN group after 8 weeks of modeling, which showed glomerular enlargement, accompanied by mesangial cells and matrix proliferation; renal tubular swelling, and lumen stenosis. Electron microscope showed glomerular basement membrane thickening, podocytes fusion; tubular basement membrane thickening, mitochondrial swelling, and brush border partial falling off. 2) Characteristics of intestinal flora in DN rats Structure of intestinal flora: OTU analysis showed that DN8/DN12 shared 764 OTUs, with 159 (17.2%) and 218 (22.2%) unique OTUs respectively. Compared with the control group, the diversity and richness of DN8 had no significant change; but the Simpson index was decreased significantly (0.95±0.01 vs 0.90±0.01，p＜0.05) in DN12. β diversity analysis showed that the community structure difference between NC12/DN12 group was larger than that of the NC8/DN 8 group. Species Difference: There was little difference between DN8 and NC8, but there was a great difference between DN12 and NC12. Metastats analysis showed that there were significant differences between NC8 and DN8 groups in 1 phylum, 1 class, 0 order, 0 family, 0 genus, and 9 species (q< 0.05). There were significant differences between NC12 and DN12 groups in 3 phyla, 7 classes, 10 orders, 18 families, 19 genera, and 14 species. There were significant differences between DN8 and DN12 groups in 8 phyla, 10 classes, 10 orders, 16 families, 39 genera, and 30 species. It indicated that the structure of intestinal flora changed significantly with the development of DN. Compared with the same age control group, s__ Bacteroides_ Cellosilyticus and s__ Bacteroides_Thetaiotaomicron were significantly decreased in DN8 and DN 12 groups.Correlation analysis: The relative abundance of 13 genera and 12 species were significantly correlated with 24hUP (p<0.05), among which, s__Bacteroides_Cellosilyticus (- 0.39, p<0.01) and s__ Bacteroides_Thetaiotaomicron (-0.42, p< 0.01)were significantly negatively correlated with 24hUP.Intestinal flora function: With the development of DN, intestinal flora function changed significantly. DNA_repair_and_recombination_Protein, transfer_RNA_Purine metabolism, purine_ Metabolism were significantly enriched in DN12. 2． Effect of cinnamaldehyde on early DN and its mechanism 1) The effect of cinnamaldehyde on renal injury in DN After 8 weeks of intragastric administration of cinnamaldehyde, compared with DN8, the 24hUP (86.65±10.08 vs 52.53±7.03 mg, p=0.02), and 24hUV of DNC8 were decreased significantly, but blood glucose and bodyweight showed no significant difference. After 12 weeks of gavage, there were no significant differences in 24hUP, 24hUV, blood glucose, and body weight between DNC12 and DN12. Under the light microscope, there was no significant difference in renal pathological changes between DN8 and DNC8 groups. Under the electron microscope, compared with DN8, the thickness of the renal tubular basement membrane (TBM) of DNC8 was significantly decreased (276.30±10.16 vs 239.50±12.43 nm, p=0.02). There was no significant difference in glomerular basement membrane (GBM) thickness between the two groups. 2) Effect of cinnamaldehyde on intestinal flora in DN rats Structure of intestinal flora: Cinnamaldehyde intervention changed the intestinal flora of DN rats. DN8 and DNC8 groups shared 774 OTUs, and the unique OTUs were 149 (16.1%) and 247 (24.2%) respectively. Compared with the DN8 group, the Simpson index of the DNC8 group was significantly decreased (0.96±0.01 vs 0.94±0.01，p＜0.01), suggesting that cinnamaldehyde can reduce species diversity, but does not affect species richness. Compared with NC8/DN8 group, the distance between NC8/DNC8 group was closer，suggesting that cinnamaldehyde may partially restore the community structure of DN rats. Species difference: Compared with DN8, there were significant differences in the relative abundance of 1 phylum, 2 classes, 0 order, 5 families, 13 genera, and 10 species in DNC8. At genera, g__Lactobacillus ， g__Alloprevotella ， g__Oscillospira, g__Weissella, and some other probiotics increased significantly in DNC8. At species, the abundance of 3 species decreased in DN8 and recovered after cinnamaldehyde intervention, including s__Bacteroides_massiliensis, s__Oscillibacter_sp_ER4,and s__Lachnospiraceae_ bacterium_ A2. The abundance of 2 species increased in DN8, and decreased after cinnamaldehyde intervention, including s__Ruminococcus_sp_ID1 and s__ Comamonas_denitrificans。Function prediction: Cinnamaldehyde can significantly change the intestinal flora function of DN rats. Compared with the DN8 group, transporters, purine metabolism, and the ABC transporter pathway were significantly enriched in the DNC8 group. 3. Mechanism of cinnamaldehyde improving early DN Immunohistochemistry and semiquantitative analysis showed that compared with NC8, the expression of megalin in the brush border of the proximal tubule in the DN8 group was significantly decreased (0.017±0.001 vs 0.006±0.001，p＜0.01). Compared with DN8, megalin in DNC8 was significantly increased (0.006±0.001 vs 0.011±0.001，p＜0.01).Compared with NC8, the expression of fibronectin (FN) in the DN8 group was significantly increased (1.00±0.28 vs 4.57±1.20，p=0.04), and the expression of IL22 in the kidney was significantly decreased (1.00±0.12 vs 0.61±0.11，p=0.04), but there was no significant difference of TGF-β. Compared with DN8, FN (4.57±1.20 vs 0.68±0.06， p=0.03) and TGF-β(1.85±0.37 vs 0.57±0.15，p=0.03) in DNC8 was significantly decreased. There was no significant difference in IL-22 and IL-17A protein expression between DN8 and DNC8. CONCLUSION: 1. There was intestinal flora dysbiosis in DN rats, related to the severity of DN. The diversity of intestinal flora decreased significantly at 12th weeks, with more significant difference species correlated with 24hUP. 2. Cinnamaldehyde can partially restore the intestinal flora dysbiosis of DN rats, and selectively increase the abundance of anti-inflammatory probiotics, such as g__Lactobacillus and g__Alloprevotella. 3. Cinnamaldehyde can not only reduce the early proteinuria and decrease the thickness of TBM in DN rats but also increase the expression of Megalin in the kidney and inhibit FN and TGF-β. But it did not affect the expression of IL22 and IL17A.