Developing Efficient and Safe Gene Transfer to Primate Hematopoietic Stem Cells

摘要：临床和基础实验室研究旨在为包括干细胞和祖细胞在内的造血细胞开发高效、安全的基因转导和体外操纵策略，并使用基因标记技术来回答体内造血的重要问题。在恒河猴模型中，我们专注于优化向原始干细胞和祖细胞的基因转移，并使用遗传标记技术来了解干细胞在体内的行为。我们继续进一步提高基因转移到恒河猴移植细胞的效率，导致早期标记细胞水平高达50%-80%，所有谱系的标记细胞水平稳定在5%-35%，这一范围具有临床实用价值。这些水平可以通过传统的两性MLV载体以及基于SIV的慢病毒载体来实现。我们已经开发了基于噬菌体的禽肉瘤白细胞增多症病毒(ASLV)载体和基于噬菌体的非病毒定位载体，用于造血靶细胞应用，因为有更有利的插入位点分布。ASLV可以转导恒河猴长期再生细胞，这在我们的体内自体移植模型中首次得到证实。我们发现，低氧条件下的转导可以改善造血干细胞的植入和长期修饰。我们继续使用LAM-PCR技术，最近利用了高通量修饰和无偏向限制性内切酶程序来改进插入位点检索和跟踪技术。我们收集并分析了CD34+转导的祖细胞移植后对外周血群体的克隆性贡献。鉴于目前有7名因严重免疫缺陷而接受逆转录病毒转导的造血干细胞基因治疗的患者发生白血病，我们对用MLV或SIV载体转导的细胞移植的恒河猴进行了逆转录病毒插入位点的大规模测序。插入位点分析表明，MLV和SIV在基因内的插入都是非随机的，SIV插入均匀地分布在整个基因的长度上，特别是在高度基因丰富的染色体区域。相反，MLV针对转录起始点周围的区域。这些高度非随机的事件要么表明这些克隆对整合有强烈的非随机偏好，要么表明这些克隆在体内植入或存活/增殖优势。14个独立的插入定位于MDS1/EVI1基因座，该区域以前与自发性白血病和复制能力强的病毒的逆转录病毒突变有关。我们没有发现使用SIV或ASLV载体的MDS1/EVI1插入。与MLV相比，SIV和ASLV载体在原癌基因中的插入率显著降低。这些发现对未来基因治疗的临床应用具有重要意义。我们继续探索原始转导的造血细胞克隆扩增和白血病发生的机制，现在使用过表达载体来研究BCL2A1和MDS1/EVI1对永生化或转化的影响。BCL2A1在小鼠HSCs中的过度表达导致克隆性主要是B细胞白血病，这表明该基因产物第一次导致了白血病。我们最近还发现，转导的CD34+细胞的体外扩增对体内克隆多样性产生了深远的影响，移植前经过长期培养后，可以选择插入MDS1/EVI1。在体内，白花丹的细胞毒性压力显示出含有插入特定基因的载体的细胞的克隆性优势。 我们已经成功地开发了两种自杀基因策略，可以在体内去除含有载体的造血细胞，移植后的转导细胞。第一种是利用优化的高度敏感的疱疹病毒tk突变转基因，该转基因由更昔洛韦激活。我们已经证明，在4-6个月前移植的非人灵长类动物中，使用无毒的21天疗程的更昔洛韦，所有可检测到的含有逆转录病毒载体的细胞都被完全消融，载体标记水平稳定。第二种是利用可被小分子二聚体AP1903激活的可诱导caspase9自杀基因。稳定植入的动物有超过90%的含有载体的细胞被AP1903短期治疗消融，我们正在继续优化这一系统。 我们最近利用“条码”慢病毒载体作为一种替代方法，在体内对造血干细胞及其后代进行克隆跟踪，避免了试图通过插入位点跟踪来量化克隆贡献时的偏倚和效率问题。这种非常强大的方法绕过了LAM-PCR或其他插入检索策略对载体插入位点的非定量检索，并允许在相关的大型动物模型中对体内单个HSPC克隆的输出和行为进行详细的定量评估。

Summary: Clinical and basic laboratory studies are directed at developing efficient and safe gene transduction and ex vivo manipulation strategies for hematopoietic cells, including stem and progenitor cells, and using genetic marking techniques to answer important questions about in vivo hematopoiesis. In the rhesus model, shown to be the only predictive assay for human clinical results, we have focused on optimizing gene transfer to primitive stem and progenitor cells, and using genetic marking techniques to understand stem cell behavior in vivo. We have continued to further enhance gene transfer efficiency into rhesus engrafting cells, resulting in early levels of marked cells as high as 50-80%, with stable levels of 5-35% in all lineages, a range with clinical utility. These levels can be achieved with traditional amphotropic MLV vectors, as well as with SIV-based lentiviral vectors. We have developed avian sarcoma leukocytosis virus (ASLV) vectors and site-specific non-viral vectors based on phage for hematopoietic target cell applications, due to more favorable insertion site profiles. ASLV can transduce rhesus long-term repopulating cells, as first demonstrated in our in vivo autologous transplantation model. We have discovered that transduction under hypoxic conditions can improve engraftment and long-term modification of hematopoietic stem cells. We have continued to utilized the LAM-PCR technology, most recently utilizing a high throughput modification, and a non-biased restriction-enzyme free procedure to improve the technology for insertion site retrieval and tracking. We retrieve and analyze clonal contributions to peripheral blood populations following transplantation of CD34+ transduced progenitor cells. Given the occurence of leukemia in now seven patients receiving gene therapy for severe immunodeficiencies with retrovirally-transduced hematopoietic stem cells, we have performed large scale sequencing of retroviral insertion sites in rhesus macaques transplanted with cells transduced either with MLV or SIV vectors. The insertion site analysis shows non-random preference for insertions within genes for both MLV and SIV, with SIV insertions distributed evenly over the length of genes and particularly being found in highly gene rich chromosomal regions. MLV instead targets the region around transcriptional start sites. These highly non-random events indicate either a strong non-random preference for integration at these sites, or an in vivo engraftment or survival/proliferative advantage for these clones. 14 independent insertions were localized to the MDS1/EVI1 locus, an area previously implicated in spontaneous leukemias and in retroviral mutagenesis with replication competent viruses. We have found no MDS1/EVI1 insertions using SIV or ASLV vectors. SIV and ASLV vectors have a significantly lower rate of insertion clusters in proto-oncogenes as compared to MLV. These findings have important implications for future gene therapy clinical applications. We continue to explore the mechanism of clonal expansion and leukemogenesis in primitive transduced hematopoietic cells, now using overexpression vectors to study the impact of BCL2A1 and MDS1/EVI1 on immortalization or transformation. BCL2A1 over-expressed in murine HSCs results in clonal primarily B cell leukemias, implicating this gene product for the first time as leukemogenic. We have also recently found a profound impact of ex vivo expansion of transduced CD34+ cells on clonal diversity in vivo, with a selection for MDS1/EVI1 insertions after prolonged culture prior to transplantation. In vivo, cytotoxic pressure with busulfan was shown to result in clonal dominance of cells containing vector insertions in specific genes. We have successfully developed two suicide gene strategies allowing ablation of vector-containing hematopoietic cells in vivo, following transplantation of transduced cells. The first utilizes an optimized and highly sensitive herpes tk mutant transgene, which is activated by ganciclovir. We have shown complete ablation of all detectable retrovirus vector containing cells with a non-toxic 21 day treatment course of ganciclovir in non-human primates transplanted 4-6 months previously, with stable vector marking levels pre ganciclovir. The second utilized an engineered inducible caspase 9 suicide gene which can be activated by the small molecule dimerizer AP1903. Stably engrafted animals had greater than 90% of their vector-containing cells ablated with short treatment courses of AP1903, and we continue to optimize this system. We have recently utilized "bar coded" lentiviral vectors as an alternative methodology for performing clonal tracking of HSCs and their progeny in vivo, avoiding the issues with bias and efficiency in attempts to quantify clonal contributions via insertion site tracking. This very powerful approach circumvents the non-quantitative retrieval of vector insertion sites with LAM-PCR or other insertion retrieval strategies, and is allowing detailed quantitative assessment of the output from and behavior of individual HSPC clones in vivo, in a relevant large animal model.