目的：探讨甲氨蝶呤囊泡对小鼠实验性牙周炎的治疗作用。方法：从人脐带间充质干细胞（hUC-MSC）中分离出细胞外囊泡（EV）。构建了负载甲氨蝶呤的囊泡（MTX-EV），并使用扫描电子显微镜和粒度分析仪分析了其形态和大小。蛋白质印迹用于鉴定其表面特异性蛋白质。选取4~5周龄C57BL/6J雄性小鼠(第四军医大学实验动物中心提供)，其中盲抓法随机选取8只，不进行处理，正常喂养作为正常组，其余诱导通过将脂多糖（LPS）局部注射到牙周组织中来建立牙周炎模型。 LPS每天注射一次，浓度为2 g/L，体积为5 μl，持续两周。将诱导牙周炎成功的小鼠采用盲抓法随机分为4组，每组8只。 LPS组不进行任何治疗，其余三组分别进行牙周局部注射MTX、EVs或MTX-EVs治疗。两周后采用酶联免疫吸附法(ELISA)检测牙龈组织中炎性细胞因子白细胞介素(IL)-1β、IL-6、肿瘤坏死因子-α(TNF-α)的表达量。采用显微CT扫描和HE染色检测四组牙槽骨吸收量。采用流式细胞仪分析牙龈组织中炎症因子的表达比例。结果：扫描电镜结果显示EV和MTX-EV的形状均为圆形或椭圆形。动态光散射（DLS）粒径分析表明，EVs 的粒径约为 200 nm，而 MTX-EVs 的粒径约为 300 nm。 ELISA结果显示正常组、LPS组、LPS MTX组、LPS EVs组、LPS MTX-EVs组IL-1β含量分别为（28.86±2.76）、（51.50±2.04）、（35.26±2.40）、（45.49）。 ±2.04) 和 (35.77±3.49) ng/L。即LPS MTX组、LPS EVs组、LPS MTX-EVs组IL-1β浓度显着低于LPS组（P<0.05）； LPS MTX-EVs组IL-1β质量浓度显着低于LPS EVs组（P<0.05）。正常组、LPS组、LPS MTX组、LPS EVs组、LPS MTX-EVs组IL-6浓度分别为（125.44±4.12）、（221.64±10.59）、（178.16±16.90）、（181.09±18.22） ）和（170.15±9.04）ng/L，其中后三组IL-6浓度显着低于LPS组（P<0.05）。 LPS MTX-EVs组IL-6质量浓度显着低于LPS MTX组和LPS EVs组（P<0.05）。正常组、LPS组、LPS MTX组、LPS EVs组、LPS MTX-EVs组TNF-α浓度分别为（320.27±38.68）、（479.62±40.94）、（342.18±25.89）、（415.88±12.01） ）和（325.75±30.83）ng/L，其中后三组浓度显着低于LPS组（P<0.05）； LPS MTX-EVs组TNF-α质量浓度显着低于LPS EVs组和LPS MTX组（P<0.05）。显微CT结果显示，正常组、LPS组、LPS MTX组、LPS EVs组、LPS MTX组第一磨牙与牙根（M1R1）骨水泥-牙釉质连接处-牙槽骨嵴（CEJ-ABC）距离-EVs组小鼠分别为（0.11±0.03）、（0.28±0.02）、（0.23±0.03）、（0.20±0.04）和（0.18±0.03）mm。与LPS组相比，LPS MTX组、LPS EVs组和LPS MTX-EVs组M1R1的CEJ-ABC均受到不同程度的抑制，差异有统计学意义（P<0.05）。其中，LPS MTX-EVs组骨吸收抑制效果优于LPS MTX组和LPS EVs组，差异有统计学意义（P<0.05）。流式细胞术结果显示，LPS组干扰素γ（IFN-γ）阳性细胞比例为（11.77±1.02）%，LPS EVs组为（6.87±0.65）%，LPS EVs组为（4.15±0.92）%。分别属于 LPS MTX-EV 组。 LPS EVs组和LPS MTX-EVs组IFN-γ阳性细胞比例显着低于LPS组（P<0.05），而LPS MTX-EVs组IFN-γ阳性细胞比例显着低于 LPS EVs 组（P<0.05）。结论：MTX-EVs可有效缓解牙周炎模型小鼠牙周局部炎症环境，减少牙槽骨骨吸收。

Objective: To investigate the therapeutic effect of methotrexate loaded vesicles on experimental periodontitis in mice. Methods: Extracellular vesicles (EVs) were isolated from human umbilical cord mesenchymal stem cells (hUC-MSC). Methotrexate loaded vesicles (MTX-EVs) were constructed, whose morphology and size were analyzed by using scanning electron microscopy and particle size analyzer. Western blotting was used to identify their surface specific proteins. C57BL/6J male mice of 4-5 weeks (provided by Experimental Animal Center of the Fourth Military Medical University) were selected, among which 8 were randomly selected by blind grasp method without treatment and fed normally as normal group, and others were induced to periodontitis models by local injection of lipopolysaccharide (LPS) into the periodontium. The LPS was injected once every day with a concentration of 2 g/L and a volume of 5 μl, lasting for two weeks. The mice with successfully induced periodontitis were randomly divided into 4 groups by blind grasping method, with 8 mice in each group. The LPS group was with no treatment, and the other three groups were treated with periodontal local injection of MTX, EVs or MTX-EVs, respectively. Two weeks later, enzyme-linked immunosorbent assay (ELISA) was used to detect the expressions of inflammatory cytokine interleukin (IL)-1β, IL-6 and tumor necrosis factor-α (TNF-α) in gingival tissue. The amount of alveolar bone resorption of four groups was detected by using micro-CT scanning and HE staining. The expression proportion of the inflammatory factor in gingival tissue was analyzed by using flow cytometry. Results: The scanning electron microscopy results showed that EVs and MTX-EVs were circular or elliptical in shape. Dynamic light scattering (DLS) particle size analysis showed that the particle size of EVs was around 200 nm, while that of MTX-EVs was around 300 nm. The ELISA results showed IL-1β levels in the normal group, LPS group, LPS+MTX group, LPS+EVs group and LPS+MTX-EVs group were (28.86±2.76), (51.50±2.04), (35.26±2.40), (45.49±2.04) and (35.77±3.49) ng/L. That is, the IL-1β concentrations in the LPS+MTX group, LPS+EVs group and LPS+MTX-EVs group were significantly lower than that in the LPS group (P<0.05); the mass concentration of IL-1β in the LPS +MTX-EVs group was significantly lower than that in the LPS+EVs group (P<0.05). The concentrations of IL-6 in the normal group, LPS group, LPS+MTX group, LPS+EVs group and LPS+MTX-EVs group were (125.44±4.12), (221.64±10.59), (178.16±16.90), (181.09±18.22) and (170.15±9.04) ng/L, among which the concentration of IL-6 in the last three groups were significantly lower than that in the LPS group (P<0.05). The mass concentration of IL-6 in the LPS+MTX-EVs group was significantly lower than those in the LPS+MTX group and LPS+EVs group (P<0.05). The concentrations of TNF-α in the normal group, LPS group, LPS+MTX group, LPS+EVs group and LPS+MTX-EVs group were (320.27±38.68), (479.62±40.94), (342.18±25.89), (415.88±12.01) and (325.75±30.83) ng/L, among which the concentrations of last three groups were significantly lower than the LPS group (P<0.05); the mass concentration of TNF-α in the LPS+MTX-EVs group was significantly lower than those in the LPS+EVs group and LPS+MTX group (P<0.05). The micro-CT results showed that the distance of cement-enamel junction-alveolar bone crest (CEJ-ABC) of the first molar and root (M1R1) in the normal group, LPS group, LPS+MTX group, LPS+EVs group and LPS+MTX-EVs group of mice were (0.11±0.03), (0.28±0.02), (0.23±0.03), (0.20±0.04), and (0.18±0.03) mm, respectively. Compared with the LPS group, the CEJ-ABC of the M1R1 in the LPS+MTX group, LPS+EVs group and LPS+MTX-EVs group were inhibited to varied degrees with statistically significant differences (P<0.05). Among them, LPS+MTX-EVs group had the best bone resorption inhibitioin effect compared to LPS+MTX group and LPS+EVs group, and the differences were statistically significant (P<0.05). The flow cytometry results indicated that the proportion of interferon-γ (IFN-γ) positive cells was (11.77±1.02)% in the LPS group, (6.87±0.65)% in the LPS+EVs group, and (4.15±0.92)% in the LPS+MTX-EVs group, respectively. The proportions of IFN-γ positive cells in the LPS+EVs group and LPS+MTX-EVs group were significantly lower than that in the LPS group (P<0.05), while the ratio of IFN-γ positive cells in the LPS+MTX-EVs group was found significantly lower than that in the LPS+EVs group (P<0.05). Conclusions: MTX-EVs can effectively alleviate the periodontal local inflammatory environment and reduce bone resorption of alveolar bone in periodontitis model mice.