|Home | About | Journals | Submit | Contact Us | Français|
Although the remarkable versatility and efficacy of recombinant adeno-associated virus 2 (AAV2) vectors in transducing a wide variety of cells and tissues in vitro, and in numerous pre-clinical animal models of human diseases in vivo, have been well established, the published literature is replete with controversies with regard to the efficacy of AAV2 vectors in hematopoietic stem cell (HSC) transduction. A number of factors have contributed to these controversies, the molecular bases of which have begun to come to light in recent years. With the availability of several novel serotypes (AAV1 through AAV12), rational design of AAV capsid mutants, and strategies (self-complementary vector genomes, hematopoietic cell-specific promoters), it is indeed becoming feasible to achieve efficient transduction of HSC by AAV vectors in a murine serial bone marrow transplantation model in vivo, where stable integration of the proviral AAV genome does not lead to any overt hematological abnormalities. Thus, a better understanding of the AAV-HSC interactions, and the availability of a vast repertoire of novel serotype vectors, are likely to have significant implications in the use of AAV vectors in high-efficiency transduction of HSCs as well as in gene therapy applications involving the hematopoietic system.
Gene therapy with vectors based on retroviruses and adenoviruses has been attempted in a number of clinical trials [Edelstein et al., 2007]. Initially, retroviral vectors yielded encouraging results, but the development of T cell lymphoma in non-human primates, and more recently, the development of T-cell leukemia in three children in a clinical trial for gene therapy of X-linked severe-combined immunodeficiency with these vectors has raised serious safety concerns [Deichmann et al., 2007; Hacein-Bey-Abina et al., 2003; Kohn et al., 2003]. Similarly, questions were raised with reference to efficacy of adenoviral vectors in gene therapy of cystic fibrosis, and subsequently, the death of a patient in a clinical trial for gene therapy of was attributed to the use of first-generation adenoviral vectors [Raper et al., 2003].
A third group of viruses, termed parvoviruses, have to date not been associated with any malignant disease. In fact, parvoviruses have been shown to possess anti-tumor properties [Asokan and Samulski, 2005; Li et al., 2005]. Parvoviruses are among the smallest of known viruses, which contain a single-stranded DNA genome, and infect all vertebrates. One parvovirus of human origin, the adeno-associated virus 2 (AAV2), has been studied extensively [Berns and Bohenzky, 1987; Muzyczka et al., 1984]. Although ~90% of the human population is sero-positive for AAV2, no known disease has thus far been associated with AAV2 infection. AAV2 requires co-infection with a helper-virus, such as adenovirus, for its optimal replication, but in the absence of a helper-virus, the AAV2 genome integrates site-specifically into the human chromosome 19. The non-pathogenic nature of AAV2, coupled with the remarkable site-specificity of the proviral genome, were instrumental in the development of recombinant AAV2 vectors. Recombinant AAV2 vectors have been shown to efficiently transduce a variety of genes in a number of cell types in vitro, and their safety and efficacy in mediating sustained transgene expression has been documented in several small and large animal models in vivo [Griffey et al., 2006; Hermonat and Muzyczka, 1984; McCown et al., 1996; Rolling et al., 2006]. A number of Phase I/II clinical trials with AAV2 vectors are currently underway, and additional trials will soon be initiated [Aitken et al., 2001; Edelstein et al., 2007; Flotte et al., 1996; Flotte et al., 2004; Kay et al., 2000; Manno et al., 2003; Snyder and Francis, 2005; Wagner et al., 2002].
Recombinant AAV2 vectors have been shown to transduce certain cell types, such as brain and muscle, exceedingly well. However, controversies abound with reference to their efficacy in transducing hematopoietic stem cells (HSCs). For example, in 1994, we first reported successful transduction of CD34+ human primitive hematopoietic cells by recombinant AAV2 vectors at a relatively low vector:cell ratio of 1,000 [Zhou et al., 1994]. Subsequently, similar results were reported by a number of investigators [Nathwani et al., 2000; Santat et al., 2005]. However, others reported that AAV2-mediated transgene expression in HSCs occurred only when the vector particle:cell exceeded 106 [Hargrove et al., 1997; Malik et al., 1997], and one group concluded that human CD34+ cells were impervious to transduction by recombinant AAV2 vectors [Alexander et al., 1997]. This group also claimed that transgene expression observed by others was due to ‘pseudo-transduction’ mediated by contaminants in the vector stocks [Alexander et al., 1997].
In systematic studies undertaken in our laboratory spanning the past decade have unraveled the underlying bases of many, if not all, of the controversies related to HSC transduction by AAV2 vectors, and has been described previously [Ponnazhagan et al., 1997; Zhong et al., 2004; Zhong et al., 2006a], and reviewed recently [Srivastava, 2005; Zhong et al., 2006b]. In essence, the following obstacles have been identified that limit high-efficiency AAV2-mediated transduction of human HSCs: (i) variable levels of expression of the cell surface receptor and co-receptor for AAV2; and (ii) lack of viral second-strand DNA synthesis to yield transcriptionally-active vector genomes. The impaired viral second-strand DNA synthesis also impacts on the efficiency of the proviral genome integration with the host chromosomal DNA. For instance, two groups have reported that AAV2-mediated transgene expression in CD34+ cells is transient presumably because of inefficient integration. On the other hand, efficient integration of the proviral genome in human CD34+ cells and their progenitors has been reported by others using a variety of techniques such as polymerase chain reaction (PCR), Southern blotting, and fluorescence in situ hybridization (FISH). Santat et al. have documented successful transduction of primitive human HSCs capable of serial engraftment in immune-deficient mice, and Paz et al. recently demonstrated that recombinant AAV2 vectors integrated more efficiently in the more quiescent subpopulations of human CD34+ HSCs [Paz et al., 2007; Santat et al., 2005]. We have molecularly cloned and determined the nucleotide sequences spanning the junctions between vector genomes and host chromosomal DNA from long-term cultures of human CD34+ cells (data not shown), which provide further evidence for successful transduction of, and proviral integration into, primary human HSCs.
It was nearly two decades ago that transduction of primary murine HSCs by AAV2 vectors was first reported by LaFace et al. [LaFace et al., 1988]. However, transgene expression occurred in only ~1.5% of progenitor cell colonies presumably because of the use of relatively crude preparations of vector stocks, heterogeneous population of target cells, and sub-optimal promoter and transgene cassette. A decade and a half ago, our laboratory embarked on systematic studies to characterize the AAV2-murine HSC interaction in vitro and in vivo. During the course of these studies, we observed that the following obstacles limit high-efficiency AAV2-mediated transduction of murine HSCs: (i) sub-optimal levels of expression of the cell surface receptor and co-receptor for AAV2; and (ii) impaired intracellular trafficking and nuclear transport of AAV2 vectors; (iii) inefficient viral uncoating in the nucleus; and (iv) lack of viral second-strand DNA synthesis to yield transcriptionally-active vector genomes [Zhong et al., 2004; Srivastava, 2005; Zhong et al., 2006a; Zhong et al., 2006b].
In retrospect, it has become increasingly clear that murine HSC transduction has been attempted with AAV2 vectors, shown schematically in Fig. 1A, that are single-stranded, and transcriptionally-inactive, contain a sub-optimal promoter, such as the cytomegalovirus (CMV) immediate-early gene, which is known to be subjected to transcriptional silencing in murine tissues, and the vector serotype, for which there is sub-optimal expression of the receptor/co-receptor on HSC surface [Cordier et al., 2001; Zhong et al., 2004; Zhong et al., 2006b]. Thus, we decided to alter each of these three parameters and generated recombinant AAV vectors containing the enhanced green fluorescent protein (EGFP) reporter gene, shown schematically in Fig. 1B, where we used double-stranded AAV genome, which bypasses the need for viral second-strand DNA synthesis; used cellular (β-globin gene) and viral (parvovirus B19) promoters known to be active in hematopoietic progenitor cells; and some of the recently developed novel AAV serotypes (AAV1 through AAV12) [Gao et al., 2002], having previously shown that single-stranded (ss) AAV1 vectors transduce murine HSCs more efficiently than ssAAV2 vectors in vivo, and that self-complementary (sc) AAV1, AAV7, AAV8, and AAV10 vectors transduce murine HSCs more efficiently than scAAV2 vectors in vitro [Maina et al., 2008].
Bone marrow transplantation of murine HSCs transduced ex vivo with scAAV serotype vectors into lethally-irradiated syngeneic recipient mice revealed the following: (i) the transduction efficiency of AAV1 vectors was higher than that of AAV2 vectors, (ii) the transduction efficient of AAV7 vectors was comparable to that of AAV1 vectors, (iii) the B19p6 promoter was more efficient than the HS2-β-globin enhancer/promoter, and (iv) AAV8 and AAV10 vectors were not significantly more efficient than AAV2 vectors [Maina et al., 2008]. Lineage analysis of transgene expression mediated by scAAV serotype vectors in murine hematopoietic progenitor cells 6-months post-primary transplantation, shown in Table 1, clearly revealed that the use of the erythroid cell lineage-specific promoters led to the transgene expression which was restricted to erythroid cells regardless of the AAV serotype vector used. Secondary transplantation of bone marrow mononuclear cells from a scAAV7-B19p6-EGFP vector-transduced primary recipient mouse used to transplant four lethally-irradiated syngeneic mice revealed erythroid lineage-restricted expression in up to 30% of erythroid cells three months post-secondary transplantation. (Table 1).
Southern blot analysis using EGFP DNA as a probe further corroborated previously published studies on stable integration of the proviral AAV genomes into host cell chromosomal DNA, and morphological analyses of peripheral blood and bone marrow cells following transplantation of HSCs following mock-transduction, or transduction with scAAV serotype vectors indicated no deleterious effects of transduced transgene expression [Maina et al., 2008]. The used of scAAV1 and scAAV7 serotype shuttle vectors for transduction of HSCs in a murine bone marrow transplant model further allowed examination of AAV proviral integration pattern in the mouse genome, as well as recovery and nucleotide sequence analyses of AAV-HSC DNA junction fragments. The proviral integration pattern was observed to be random, and recombination sites were localized to different chromosomes (Figure 2). None of the integration sites was found to be in a transcribed gene or near an oncogene, and all animals followed for up to one year exhibited no pathological abnormalities [Han et al., 2008]. Thus, AAV proviral integration-induced risk of oncogenesis was not found in our studies, which provide functional confirmation of stable transduction of self-renewing multipotential HSCs by scAAV vectors as well as show promise for the use of these vectors in the potential treatment of disorders of the hematopoietic system.
In my view, the following three broad areas of further investigations deserve greater scrutiny:
Our systematic studies during the past decade have not only unraveled several obstacles that limit efficient transduction of HSCs by recombinant AAV2 vectors, but have also reinforced the need to continue further studies focused on the basic understanding of all aspects of the life cycle of AAV2 vectors at the molecular level [Srivastava, 2001; Srivastava, 2004; Wu et al., 2007; Zhao et al., 2007; Zhong et al., 2006b]. Similar studies are also needed in order to identify and characterize the optimal AAV serotype vector(s) to achieve efficient transduction of HSC. In this context, it is important to note that endogenous AAV sequences are the third most prevalent in the human bone marrow [Gao et al., 2004]. Thus, it would be worthwhile to develop these serotypes into recombinant vectors, which may also be instrumental in achieving high-efficiency transduction of primary human hematopoietic stem and progenitor cells. In this context, it is noteworthy that in our recent studies [Zhong et al., 2008], we have developed novel AAV2 vectors containing mutations in the surface-exposed tyrosine residues that transduce HeLa cells ~10-fold more efficiently in vitro (Figure 3 A-C), and murine hepatocytes nearly 30-fold more efficiently in vivo at a log lower vector dose (Figure 3 D,E). These vectors also allow the production of therapeutic levels of human Factor IX (F.IX) an ~10-fold reduced vector dose (Figure 3 F). In preliminary studies, we have observed that primary murine and human hematopoietic stem/progenitor cells can also be transduced more efficiently by these tyrosine-mutant AAV2 vectors in vitro, and respectively in syngeneic and NOD/SCID mice in vivo (unpublished data). It is also noteworthy that with a few exceptions, these tyrosine residues are highly conserved in AAV serotypes 1 through 12, and we have also begun to generate the tyrosine-mutants of each of these serotypes. The availability of such a vast repertoire of novel tyrosine-mutant AAV serotype vectors should allow us to gain a better understanding of the role of tyrosine-phosphorylation of AAV capsids in various steps in the virus life cycle, which is likely to have important implications in the optimal use of recombinant AAV serotype tyrosine-mutant vectors in HSC transduction. It is also tempting to speculate that it may eventually be feasible to develop recombinant AAV serotype vectors that allow regulated transgene expression in each of the hematopoietic progenitor cell lineage.
It is generally assumed that recombinant AAV vector genomes remain episomal, but those studies have been carried out in tissues that are post-mitotic [Duan et al., 1998; Inagaki et al., 2007; Nakai et al., 2003; Zhong et al., 2006a]. In HSCs, which must divide to give rise to progenitor cells, the proviral genome would be lost if not stably integrated into host cell chromosomal DNA. In our previous studies, we documented stable integration of the proviral genome into chromosomal DNA in pluripotent HSC transduced with ssAAV1 and AAV2 vectors in a small number of animals [Tan et al., 2001; Zhong et al., 2006a]. Therefore, further studies with a larger number of animals are needed to define this issue at the molecular level. In addition, the integration patterns of ssAAV vectors versus scAAV vectors should also be compared and contrasted. Detailed sequence analyses of integration junctions in various cell lineages will be helpful in not only determining clonality, but also in establishing whether the proviral integration occurs in transcriptionally-active versus the non-transcribed regions of the genome. In our recent studies using a shuttle vector [Han et al., 2008], we have attempted to address some of these issues, but further detailed investigations are clearly warranted.
In view of a recent report of the development of hepatocellular carcinoma in mice injected with a relatively high AAV2 vector dose neonatally [Donsante et al., 2007], it is crucial that the safety and efficacy of the AAV serotype vectors be rigorously examined following integration of the proviral genome into the host chromosomal DNA. It should be emphasized, however, that other investigators found no evidence for AAV-induced malignancy in large-scale studies involving adult mice or p53-deficient mice [Bell et al., 2006; Bell et al., 2005; Schuettrumpf et al., 2007]. These discrepancies with the post-mitotic tissues notwithstanding, the biological consequences of stable transduction and long-term transgene expression in hematopoietic progenitor cells should nonetheless be examined further.
My prediction is that despite the lack of enthusiasm in the development of recombinant AAV2 vectors for HSC transduction in the past, a more complete understanding of the virus-host cell interactions, coupled with the availability of an ever-expanding vast repertoire of novel AAV serotype vectors, will lead to high-efficiency transduction of HSC in not too distant a future, which in turn, will provide a safer alternative to the more commonly used retroviral vectors in hematopoietic stem cell gene therapy in humans.
I thank Drs. Richard Jude Samulski and James M. Wilson for generously providing the packaging plasmids for AAV serotypes, and the members of my laboratory for helpful discussions. This research was supported in part by Public Health Service grants R01 EB-002073, R01 HL-065770, HL-076901, and P01 DK-058327 (Project 1) from the National Institutes of Health.
Contract grant sponsor: National Institutes of Health; Contract grant numbers: R01 EB-002073, R01 HL-065770, HL-076901, and P01 DK-058327 (Project 1)