Objective: To investigate the effect of intravitreal injection of neural stem cells (NSC) induced from human umbilical cord mesenchymal stem cells, and to provide a theoretical basis for the clinical treatment of blood-retinal barrier damage due to diabetic retinopathy (DR). Methods: Experimental study. Sixty male Sprague-Dawley rats were randomly divided into control group, DR group and NSC group. Diabetic rats were induced by injection of streptozotocin, and the control rats were injected with an equal volume of solvent. Three months after the establishment of diabetic models, the NSC group was injected with 2 μl of NSC in the right vitreous, and the DR group was injected with 2 μl of phosphate-buffered saline. One month later, all the rats were sacrificed. The retinal vessels and leakage were examined with flat-mounted retinas. Vascular permeability was quantified by analyzing albumin leakage using the Evans blue (EB) method. Retina was examined by hematoxylin and eosin staining. Results: Retinal blood vessels of the control rats were normal, with no EB leakage outside the vessels. The background fluorescence was enhanced and focal leakage and focal dilated vessels were detected in the DR group. In the NSC group, background fluorescence was enhanced slightly and EB leakage area decreased significantly compared with the DR group. The average EB in control group, DR group and NSC group were (9.91±1.53), (24.67±2.26) and (12.85±2.58)μg/g, The EB leakage in the NSC group decreased significantly compared with the DR group (q=9.748, P<0.05). Pathological hematoxylin and eosin staining showed that the retinal layer structure was normal and clear in the control group, the retina was thin, the cell arrangement was in disorder and the nucleus was swelling in the DR group, the status of the NSC group was between the other two groups. Conclusions: Transferring human umbilical cord mesenchymal stem cells-induced NSC in vitro to diabetic rat models by intravitreal injection could reduce leakage of blood vessels and attenuate blood-retinal barrier breakdown induced by diabetes. (Chin J Ophthalmol, 2017, 53: 53-58).
目的： 探讨由人脐带间充质干细胞（hUCMSC）体外诱导的神经干细胞（NSC）经玻璃体腔注射途径移植对糖尿病视网膜病变(DR)大鼠血-视网膜屏障（BRB）的保护作用。 方法： 实验研究。应用随机数字表法将60只Sprague-Dawley（SD）大鼠随机分为正常对照组、DR模型对照组及NSC治疗组。DR模型对照组和NSC治疗组大鼠腹腔注射STZ建立糖尿病模型。于糖尿病成模后3个月，NSC治疗组大鼠右眼玻璃体腔注射2 μl NSC悬液；DR模型对照组大鼠右眼注射等容量PBS；上述两组大鼠左眼及正常对照组大鼠双眼不给予任何干预。干预后1个月，采用伊文思蓝（EB）灌注血管视网膜铺片观察视网膜血管形态及渗漏情况；EB定量检测BRB破坏情况；观察视网膜组织病理学改变情况。组间总体差异采用单因素方差分析，组间两两比较采用Dunnett t检验。 结果： EB灌注视网膜铺片检查显示，正常对照组大鼠视网膜血管无明显异常，无EB渗漏到血管外；DR模型对照组背景荧光增强，可见明显的EB渗漏高荧光区；NSC治疗组背景荧光略增强，EB渗漏区较DR模型组明显减少。EB渗漏定量分析显示，正常对照组、DR模型对照组及NSC治疗组EB平均渗漏量分别为（9.91±1.53）、（24.67±2.26）及（12.85±2.58）μg/g，组间总体差异有统计学意义（F=103.801，P<0.01）。DR模型对照组大鼠视网膜平均EB渗漏量较正常对照组增加（q=15.306，P<0.05）；NSC治疗组较DR模型对照组明显减少（q=9.748，P<0.05）。视网膜组织病理学检查示正常对照组视网膜各层结构清晰、排列整齐、形态正常；DR模型对照组视网膜细胞排列较紊乱，神经纤维层水肿明显，细胞核肿胀、体积增大，细胞密度降低；NSC治疗组视网膜细胞较DR模型对照组略整齐，介于两组之间。 结论： hUCMSC诱导的NSC经玻璃体腔途径移植，能够改善BRB功能，减少早期DR大鼠血管渗漏，从而达到防治DR进展的治疗作用。（中华眼科杂志，2017，53：53-58）.
Keywords: Blood-retinal barrier; Diabetes mellitus, experimental; Diabetic retinopathy; Intravitreal injections; Mesenchymal stem cell transplantation; Neural stem cells.