Gene therapy of streptozotocin-induced diabetic rats with a regulatable insulin expression vector

对于糖尿病基因治疗的工程胰岛素生成替代细胞，有两种类型的细胞可用：神经内分泌细胞和非神经内分泌细胞。神经内分泌细胞具有将前肽转化为成熟的生物活性肽的加工机制，以及在细胞外刺激下分泌成熟肽的调节分泌系统。因此，当神经内分泌细胞（如垂体前皮质来源的内分泌细胞AtT20）用于糖尿病基因治疗时，主要的焦点是设计一种胰岛素分泌机制，以响应生理范围内的葡萄糖刺激。另一方面，当使用非神经内分泌细胞作为胰岛素产生细胞时，需要对细胞进行加工机制和调节分泌系统的改造。在胰腺β细胞中，胰岛素原通过激素原转化酶PC2和PC3（也称为PC1）和羧肽酶H（更多CPH）加工成成熟的胰岛素。这两种蛋白水解反应似乎只针对含有分泌颗粒的神经内分泌细胞。然而，当我们将胰岛素原的加工位点替换为酵母Kex2家族内源性蛋白酶furin可切割的位点时，包括COS-7、HepG2、CHO和NIH3T3在内的非神经内分泌细胞系产生的胰岛素大小与合成的人胰岛素相同。接下来，糠蛋白裂解的胰岛素原需要羧基肽酶去除基本残基以使其成熟。虽然非神经内分泌细胞表达不同数量的羧肽酶HmRNA，但这些细胞含有相当水平的羧肽酶活性。由这些细胞系产生的胰岛素在阳离子交换色谱柱上被洗脱为成熟胰岛素位置的单峰。因此，非神经内分泌细胞能够产生正确加工的胰岛素，如果胰岛素原首先突变为具有糠蛋白可切割的加工位点。为了开发一种可受细胞外刺激调节的胰岛素表达系统，最理想的系统是利用葡萄糖作为刺激剂。另一种调节因子可能是胰岛素本身下调表达。由于非神经内分泌细胞不携带分泌颗粒，也不通过细胞质中的组成途径分泌蛋白质和多肽，因此胰岛素产生的调控步骤仅限于基因转录水平。许多基因的表达都受到葡萄糖的调控。为了调控这些基因的表达，靶细胞需要配备葡萄糖转运蛋白2型和葡萄糖激酶来响应生理范围内的葡萄糖浓度。对于这个标准，主要培养的肝细胞是评估这些基因表达的细胞选择。然而，通过传统的基因转移方法，肝细胞的DNA转移效率相对较低。此外，目前还没有葡萄糖转运蛋白2型和葡萄糖激酶的肝细胞培养细胞系。因此，我们试图利用肝癌细胞系和磷酸烯醇丙酮酸羧激酶（PEPCK）启动子，通过胰岛素本身产生一个调节表达系统。PEPCK在肝细胞中表达，并通过从GTP中摄取磷酸，催化草酰乙酸合成磷酸烯醇丙酮酸。这种酶是肝脏葡萄糖生成的关键步骤。在糖尿病患者中，这种酶因缺乏胰岛素而高度激活。已知PEPCK启动子可被cAMP、糖皮质激素、视黄酸（RA）上调，并被胰岛素和酚酯下调。我们在H4Ell细胞中使用PEPCK启动子和胰岛素DNA结构来检测胰岛素的产生。H4llE细胞以2-3 fmol/10^6 cells/h的水平持续分泌免疫反应性胰岛素（IRI）。0.5 mM cAMP刺激后IRI增加约2倍，cAMP依赖性磷酸二酯酶抑制剂IBMX刺激后IRI增加3倍。5 ~ 700 nM地塞米松组IRI升高2倍，但进一步添加cAMP和IBMX组IRI升高10 ~ 12倍。视黄酸诱导IRI的4倍。Northern blot结果显示，在培养基中加入外源胰岛素显著降低了胰岛素mRNA的表达。但与cAMP、IBMX、地塞米松共同作用，外源性胰岛素的抑制作用减弱。在产生胰岛素的H4llE细胞中，添加wortmannin（一种磷脂酰肌醇(PI) -3激酶抑制剂）可以增加胰岛素的产生，这表明抑制胰岛素信号传导到PEPCK启动子可能是通过PI-3激酶介导的。当细胞被灌注时，胰岛素分泌在刺激2h后升高。因此，H4llE细胞的分泌模式可能与中间作用胰岛素的作用模式相对应。然后，我们将胰岛素表达单元引入腺病毒载体Adex。利用重组腺病毒载体，我们证实了从培养的大鼠肝细胞中产生成熟胰岛素。肝细胞的胰岛素分泌也受到dbcAMP、IBMX、胰高血糖素和地塞米松的良好调节。我们开始用链脲佐菌素（STZ）诱导的糖尿病大鼠进行体内实验。注射100或150mg/Kg体重的STZ后，大鼠出现400-600mg/dl的高血糖。我们从大鼠尾静脉注射重组腺病毒，大多数病毒被认为是在肝脏中积聚的。注射后10天左右胰岛素水平上升至2-3ng/ml，血糖下降至300-400mg/dl。虽然我们不能在重组病毒注射的大鼠中获得正常的血糖，但我们认为这种胰岛素表达系统可以轻度改善糖尿病。我们目前正在确定最佳的病毒滴度和适当的病毒给药方法。少

For engineering insulin-producing surrogate cells for diabetes gene therapy, two types of cells are available : neuroendocrine and non-neuro-endocrine cells. Neuroendocrine cells are equipped with a processing mechanism that converts a propeptide into a mature bioactive peptide, as well as a regulatory secretion system by which a mature peptide is secreted in response to extracellular stimuli. Thus, when neuroendocrine cells such as anterior pituitary corticotroph-derived endocrine cells AtT20 are used for diabetes gene therapy, the primary focus has been oriented to engineer an insulin secretion mechanism that responds to a physiological range of glucose stimuli. On the other hand, when non-neuroendocrine cells are used as insulin-producing cells, the cells need to be engineered with a processing mechanism and a regulatory secretion system.In pancreatic beta cells, proinsulin is processed to mature insulin by prohormone convertases PC2 and PC3 (also named PC1) and carboxypeptidase H ( … More CPH). These two proteolytic reactions appeared specific to secretory granule-containing neuroendocrine cells. However, when we replaced the processing sites of proinsulin with those cleavable by a yeast Kex2 family endoprotease furin, the non-neuroendocrine cell lines including COS-7, HepG2, CHO,and NIH3T3 produced insulin with the same size as synthetic human insulin. Next, the furin-cleaved proinsulin required the removal of basic residues by carboxypeptidases for its maturation. Although non-neuroendocrine cells expressed different quantities of carboxypeptidase HmRNA,these cells contained considerable levels of carboxypeptidase activity. The insulins resulting from these cell lines were eluted as a single peak at the mature insulin position on a cation-exchange chromatography column. Thus, non-neuroendocrine cells are able to produce correctly processed insulin if proinsulin is first mutated to possess furin-cleavable processing sites.For developing an insulin expression system that can be regulated by extracellular stimili, most desirable system is to use glucose as a stimulator. Alternative regulator may be insulin itself for down-regulating the expression. Since non-neuroendocrine cells do not carry secretory granules and secrete proteins and peptides through a constitutive pathway without their retention in the cytoplasm, a regulatory step for the production of insulin is limited to a gene transcription level. Many genes are known to be regulated their expression by glucose. For the regulatory expression of these genes target cells are required to equip with both glucose transporter type 2 and glucokinase for responding to a physiological range of glucose consentrations. For this criterion the primarily cultured hepatocytes are the cell of choice for assessing the expression of these genes. However, hepatocytes are relatively inefficient for DNA transfer by conventional gene transfer methods. Furthermore, hepatocyte culture cell lines with glucose transporter type 2 and glucokinase are not avalable at the present time. Thus, we attempted to generate a regulatory expression system by insulin itself using a hepatoma cell line and a phosphoenolpyruvate carboxykinase (PEPCK) promoter. The PEPCK is expressed in hepatocytes and catalyzes oxaloacetate to phosphoenolpyruvate by taking phosphate from GTP.This enzyme acts as a key step for hepatic glucose production. In diabetic subjects this enzyme is highly activated by the lack of insulin. The PEPCK promoter is known to be up-regulated by cAMP,glucocorticoids, retinoic acid (RA), and down-regulated by insulin and phorbol ester. We examined the insulin production using a PEPCK promoter plus insulin DNA construct in H4Ell cells.The H4llE cells secreted immunoreactive insulin (IRI) constantly at a level of 2-3 fmol/10^6 cells/h. IRI was increased about 2-fold upon stimulation with 0.5 mM cAMP,further 3-fold with cAMP-dependent phosphodiesterase inhibitor IBMX.IRI was elevated only 2-fold by 5 to 700 nM dexamethasone, but 10-12-fold by the further addition of cAMP and IBMX.Retinoic acid induced IRI by 4-fold. Addition of exogenous insulin to the culture medium decreased insulin mRNA expresion strikingly on Northern blot. However, together with cAMP,IBMX,and dexamethasone, the inhibitory effect of exogenous insulin became weakened. In the insulin-producing H4llE cells the production was augmented by the addition of wortmannin, a phosphatidylinositol (PI) -3-kinase inhibitor, suggesting that inhibitory insulin signaling to the PEPCK promoter may be mediated through PI-3-kinase. When the cells were perifused, insulin secretion was elevated 2h after the stimulation. Thus, the secretory pattern from the H4llE cells may correspond to the action mode of intermediate-acting insulin.We then introduced the insulin expression unit into an adenoviral vector Adex. Using a recombinant adenoviral vector, we confirmed the production of mature insulin from primarily cultured rat hepatocytes. The insulin secretion from hepatocytes was also well regulated by dbcAMP,IBMX,glucagon, and dexamethasone. We moved onto an in vivo experiment using streptozotocin (STZ) -induced diabetic rats. After injecting STZ at 100 or 150mg/Kg body weight, rats exhibited hyperglycemia of 400-600mg/dl blood glucose. We administered the recombinant adenovirus from a rat tail vein, through where most of viruses are thought to accumulate in the liver. Insulin level went up to 2-3ng/ml around 10 days after the injection, and blood glucose went down to 300-400mg/dl. Although we could not attain normoglycemia in the recombinant virus-injected rats, we think that this insulin expression system improves diabetes mildly. We are currently determining a best virul titer and an appropriate method of virus administration. Less