Genetic mutations underlie the majority of diseases seen by pediatric hematologists/oncologists. In hematology, the mutations are often inherited and lead to lifelong debilitating diseases such as sickle cell disease and hemophilia. In oncology, the mutations are usually acquired and lead to diseases such as acute lymphoblastic leukemia that are lethal if not treated with aggressive chemotherapy. Our broad goal is to develop ways to manipulate the genome in a sequence specific manner in order to develop specific therapies for pediatric hematologic and oncologic diseases. Specifically, we are interested in understanding DNA double-strand break repair as a way toward developing innovative tools to treat patients with inherited mutations and as way towards understanding the acquisition of mutations that lead to oncologic disease.

Gene Correction for Inherited Genetic Diseases such as Sickle Cell Disease

    The major approach to gene therapy has been through gene addition type of approaches. With the effort of hundreds of labs over the last 10-20 years progress has been made in this approach. The cure of a handful of children with severe combined immunodeficiency (SCID) seemed to validate the effort. The development of leukemia in two patients by insertional mutagenesis highlighted a potential serious fundamental limitation to the gene addition approach. A second approach to gene therapy is gene correction whereby the mutation is corrected rather than complemented. The gene correction approach has obvious theoretical advantages. One potential way of performing gene correction is by gene targeting. Gene targeting is the process by which an exogenous segment of DNA replaces an endogenous genomic target by homologous recombination. In mammalian cells, the rate of spontaneous gene targeting is extremely low (1 in a million). If a DNA double-strand break is created in the genomic target, however, the rate of gene targeting can increase by 5 orders of magnitude (to 1 in 20). While the efficiency of spontaneous gene targeting is too low to be of therapeutic use, the efficiency of double-strand break mediated gene targeting is high enough to be potentially curative of many genetic diseases, including sickle cell disease. These studies of double-strand break mediated gene targeting were generally performed in immortalized cell lines using the I-SceI endonuclease (a yeast endonuclease with a specific 17 basepair recognition sequence). Thus, to develop double-stand break mediated gene targeting as a treatment option two developments need to occur: 1) develop reagents that can create specific double-strand breaks at essentially any site in the genome and 2) understand gene targeting and homologous recombination in human hematopoietic stem cells.

 

 

 made to stimulate gene targeting at different sequences. Theoretically it may be possible to design ZFNs to stimulate at gene targeting at any location in the genome. As part of our development of this new technology we have made ZFNs that stimulate gene targeting in the GFP gene and the human CD8a gene. In collaboration with Sangamo Biosciences (Richmond, CA), we have made ZFNs to target exon 5 of the interleukin receptor common-gamma chain (“common-gamma chain”). In cell lines, the common-gamma chain ZFNs can stimulate gene targeting in human hematopoietic cell lines at the endogenous locus to a frequency of 5-18%. We are in the process of writing this data up for publication. In a further collaboration with Sangamo, we are now developing ZFN’s to stimulate gene targeting at the human b-globin locus. ZFNs differ from the I-SceI nuclease, however, because they have cytotoxicity when expressed in mammalian cells. This cytotoxicity is most likely the result of ZFN mediated cleavage at off-target sites in the genome. A major focus of the laboratory is on minimizing ZFN cytotoxicity. We are approaching the problem of ZFN cytotoxicity in two ways. The first is to try to identify the potential off-target sites that the ZFNs are cleaving at. Work over the last several years has shown that double-strand breaks can often serve as sites of integration of exogenously introduced DNA. We plan to use this phenomenon to capture the sites of ZFN cleavage. The second is to empirically develop ways to minimize cytotoxicity. Our approach to minimizing toxicity include using in vitro techniques to improve the specificity of the ZFNs for their target site compared to off-target sites and developing ways to regulate the ZFN expression to maintain high levels of targeting while minimizing toxicity.

 

Study of Gene Targeting in Human Hematopoietic Stem Cells

    In parallel to developing ZFNs to target the b-globin locus and minimizing ZFN toxicity, we are studying gene targeting in human hematopoietic stem cells (HSCs). In these studies we purify CD34+ cells from human umbilical cord blood. We then introduce a mutated GFP reporter gene into the purified HSCs using a lentivirus to create HSC reporter cells. The reporter cells are initially GFP negative but if gene targeting occurs, the cell becomes GFP positive. We detect and quantify the number of GFP positive cells by flow cytometry and can quantify the rate of gene targeting. These reporter cells are similar to the reporter cell lines we used to work out the parameters of gene targeting described above. We plan to use these reporter cells to compare the efficiency of gene targeting using I-SceI compared to ZFNs and to optimize the method of delivery of the nuclease to HSCs. We plan to use in vitro progenitor cell assays and mouse reconstitution assays to determine if HSCs that have undergone targeting maintain their stem cell properties. The long-term goal of the work, of course, is to combine the development of specific and non-toxic b-globin ZFNs with our studies of gene targeting in HSCs to target the b-globin locus in HSCs.

 Determining the Role of Double-Strand Breaks in Triplet Repeat Instability

    There are at least 12 progressive neurologic diseases that are caused by expansions in triplet repeats. Huntington’s disease is the prototypical example of such diseases. Moreover, Fondon and colleagues have shown that changes in triplet repeat lengths may be an important cause of phenotypic variation. We are interested in triplet repeat variation from two perspectives. The first is to develop repeat specific zinc finger nucleases to create double-strand breaks in the midst of a repeat to determine if such breaks stimulate the contraction and/or expansion of the repeats. The second is to develop repeat specific zinc finger nucleases in order to use gene targeting to contract long repeats that cause disease to shorter repeats that do not cause disease.

Determining the Ontogeny of Acquired Mutations that lead to Pediatric Cancers

    Chromosomal translocations are a central element to many pediatric cancers. Laboratory based studies have shown that these translocations are sometimes sufficient to transform normal cells but other times require second events. Independent of the role of translocations in generating cancer, it is also clear that certain translocations carry prognostic significance. In pediatric acute lymphoblastic leukemia (ALL), for example, certain translocations (such as t(12;21)) are associated with a favorable outcome while others are associated with unfavorable outcomes (such as the t(9;22)). Despite the importance of translocations in pediatric oncology, the ontogeny of translocations remains relatively obscure. The prevailing hypothesis is that translocations occur when two simultaneous DNA double-strand breaks are created on different chromosomes and the broken ends are aberrantly joined thereby creating a translocation. Thus, a chromosomal translocation would represent a failure in accurate DNA double-strand break repair. We have developed several assays to measure the integrity of double-strand break repair. In these assays, we can both quantitatively measure different types of double-strand break repair and qualitatively determine the fidelity of repair. For example, we have found that mutations in a gene important in non-homologous end-joining, Ku70, leads to a significant decrease in both the frequency and accuracy of DNA double-strand break repair by non-homologous end-joining. We are in the process of writing this work up for publication. In the best experimental models to study the etiology of translocations by double-strand breaks, double-strand breaks are created by the I-SceI endonuclease at artificially integrated I-SceI target sites. The development of ZFN’s, however, will allow us to create DSB’s at natural target sites, particularly those known to be involved in oncogenic translocations, and thus study the role of aberrant DSB repair in generating oncogenic translocations.

Summary

    In the early 21st century, the importance of genetic mutations and polymorphisms in human disease is unequivocal. Overall, our interest in the role of genetic mutations in human disease has taken us into several overlapping areas of research. We are interested in homologous recombination as a method to potentially cure patients with genetic diseases. We are interested in the regulation of DNA double-strand break repair both because of its importance in understanding double-strand break induced homologous recombination and also because of its importance in understanding where acquired mutations, like translocations, come from. Finally, we are interested in developing and adapting novel agents, such as zinc finger nucleases, that in combination with studies of the mechanism of homologous recombination and regulation of DNA double-strand break repair we hope will lead to improved therapies for a wide variety of patients.