目的 基于数据挖掘技术探析名老中医治疗2型糖尿病（type 2 diabetes mellitus，T2DM）的辨证用药及组方规律，并通过动物实验验证核心方剂的疗效。方法 从中国知网、万方数据知识服务平台、维普数据库等渠道收集200例名老中医治疗T2DM的医案，利用中医传承计算平台（V3.5）构建数据库，并引入轮廓系数评价K-means聚类效果，结合层次聚类法进行交叉验证以筛选核心方剂。将40只SD雄性大鼠随机分为对照组、模型组、核心方组和二甲双胍组，每组10只。除对照组外，其余组别采用高脂、高糖饲料喂养联合每日1次15.75 g/kg青皮附子水持续ig 4周，随后ip 30 mg/kg链脲佐菌素诱导T2DM气阴两虚证大鼠模型。造模成功后，核心方组大鼠ig核心方10.5 g/kg药液，二甲双胍组大鼠ig 200 mg/kg二甲双胍，对照组和模型组则给予等体积生理盐水，各组每日给药1次，干预4周。监测大鼠在给药期间体质量及空腹血糖变化；实验结束后通过酶联免疫吸附法检测血清胰岛素（serum insulin，INS）、糖化血红蛋白（glycated hemoglobin，IFCC-HbA1c）、白细胞介素-1β（interleukin-1β，IL-1β）、IL-6、超氧化物歧化酶（superoxide dismutase，SOD）、丙二醛（malondialdehyde，MDA）等指标；通过免疫组化染色观察胰岛素阳性面积比例；油红O染色观察肝脏脂质沉积；PAS染色评估肾脏组织病理改变；运用Western blotting检测肝组织中磷脂酰肌醇-3-羟激酶（phosphatidylinositol-3-hydroxykinase，PI3K）/蛋白激酶B（protein kinase B，Akt）、核因子-κB（nuclear factor-κB，NF-κB）及核因子E2相关因子2（nuclear factor erythroid-2-related factor 2，Nrf2）/血红素氧合酶-1（heme oxygenase-1，HO-1）通路相关蛋白表达。结果 数据挖掘显示T2DM主要证型为气阴两虚证；高频药物为黄芪、黄连、茯苓等；药性以寒、温、平为主，药味以甘、苦、辛为主，主归脾、肺、肝经；补虚、清热、利水渗湿及活血化瘀这几类药物的应用最为广泛；关联规则分析得出黄芪-葛根、黄芪-黄连等为核心药对；聚类筛选出由黄芪、山药、葛根、山茱萸、茯苓组成的类方作为核心方。动物实验证实，核心方组空腹血糖、INS、IFCC-HbA1c较模型组显著降低（P＜0.05、0.01、0.001），体质量逐渐回升，IL-1β、IL-6含量显著降低（P＜0.01、0.001），SOD活性显著升高（P＜0.01），MDA含量显著降低（P＜0.05），胰岛素阳性面积比例显著升高（P＜0.05），肝脏脂质沉积面积比例显著降低（P＜0.01），肾脏病理损伤积分显著降低（P＜0.05），肝组织中p-PI3K/PI3K、p-Akt/Akt、Nrf2、HO-1表达水平显著升高（P＜0.05、0.01），p-NF-κB p65/NF-κB p65水平显著降低（P＜0.05）。结论 核心方在动物实验中降糖疗效确切，其作用机制与激活PI3K/Akt信号通路以增强胰岛素敏感性、抑制NF-κB介导的促炎反应、上调Nrf2/HO-1改善氧化应激有关。该方能有效维持胰岛细胞的数量与结构，保护肝肾组织免受糖尿病病理损伤，为中医药防治T2DM提供了科学依据。

Objective To explore the syndrome differentiation and medication, and formulation rules of renowned traditional Chinese medicine (TCM) practitioners in treating type 2 diabetes mellitus (T2DM) using data mining technology, and validate the efficacy of the identified core formula using an animal model. Methods A total of 200 medical cases of T2DM treated by famous TCM practitioners were collected from databases such as China National Knowledge Infrastructure (CNKI), Wangfang, VIP. A database was established using the TCM Inheritance Computing Platform (V3.5). The silhouette coefficient was introduced to evaluate K-means clustering performance, and hierarchical clustering was used for cross-validation to screen for core formulas. For the experimental validation, 40 male Sprague-Dawley (SD) rats were randomly divided into four groups (n = 10 each): control, model, core formula, and metformin. Except for the control group, T2DM models with “qi-yin deficiency” syndrome were induced by a 4-week high-fat and high-sugar diet combined with daily oral administration of Qingpi Fuzi decoction (15.75 g/kg), followed by an intraperitoneal injection of streptozotocin (STZ, 30 mg/kg). After successful modeling, the core formula group received the core formula (10.5 g/kg), and the Metformin group received metformin (200 mg/kg) via gavage daily; Control and Model groups received equal volumes of saline, each group was given the drug once a day for four weeks of intervention. Body weight (BW) and fasting blood glucose (FBG) were monitored during the administration period. Serum levels of insulin (INS), glycated hemoglobin (IFCC-HbA1c), interleukin-1β (IL-1β), IL-6, superoxide dismutase (SOD), and malondialdehyde (MDA) were measured by ELISA. The proportion of insulin-positive areas was assessed via immunohistochemical staining. Oil red O staining was used to observe lipid deposition in the liver. PAS staining was used to evaluate the pathological changes of renal tissue. Protein expressions of the phosphatidylinositol-3-hydroxykinase (PI3K)/protein kinase B (Akt), nuclear factor-κB (NF-κB), nuclear factor erythroid-2-related factor 2 (Nrf2)/heme oxygenase-1 (HO-1) pathways in liver tissues were detected by Western blotting. Results Data mining revealed that qi-yin deficiency was the primary syndrome of T2DM. High-frequency herbs included Huangqi (Astragali Radix), Huanglian (Coptidis Rhizoma), and Fuling (Poria). The medicinal properties were predominantly cold, warm, and neutral, while flavors were mainly sweet, bitter, and pungent, targeting the spleen, lung, and liver meridians. These types of drugs such as tonifying deficiency, clearing heat, promoting diuresis, and activating blood circulation were most prevalent. Association rule analysis identified Astragali Radix-Gengen (Puerariae Lobatae Radix) and Astragali Radix-Coptidis Rhizoma as core herb pairs. Clustering identified a core formula consisting of Astragali Radix, Shanyao (Dioscoreae Rhizoma), Puerariae Lobatae Radix, Shanzhuyu (Corni Fructus), and Poria. In vivo experiments showed that the core formula group had significantly lower FBG, INS, and IFCC-HbA1c levels compared to the Model group (P < 0.05, 0.01, 0.001), with a gradual recovery in BW. Furthermore, IL-1β and IL-6 were significantly decreased (P < 0.01, 0.001), SOD was increased (P < 0.01), and MDA was decreased (P < 0.05). Pathologically, the core formula group showed a significantly higher ratio of insulin-positive area (P < 0.05), reduced hepatic lipid deposition (P < 0.01), and lower renal pathological injury scores (P < 0.05). Western blotting analysis indicated that core formula significantly up-regulated the expression of p-PI3K/PI3K, p-Akt/Akt, Nrf2, and HO-1 (P < 0.05, 0.01), while down-regulating p-NF-κB p65/NF-κB p65 (P < 0.05). Conclusion The core formula exhibits a definite hypoglycemic effect. Its mechanism of action is associated with activating the PI3K/Akt signaling pathway to enhance insulin sensitivity, inhibiting NF-κB-mediated pro-inflammatory responses, and up-regulating the Nrf2/HO-1 pathway to ameliorate oxidative stress. This formula effectively maintains the number and structure of pancreatic islet cells and protects liver and kidney tissues from diabetic pathological damage, providing a scientific basis for the prevention and treatment of T2DM with TCM.

R285

国家自然科学基金项目（82260861）；广西中医基础研究重点实验室项目（19-45045-11）：广西中医药大学“桂派杏林拔尖人才项目”（20220002）；2024年广西壮医自治区大学生创新训练计划项目（S202410600080）