**7. Safety: Vector genotoxicity, transposon vectors and other issues**

The genotoxicity issue is currently the most immediate and direct safety concern related to the gene therapy using HSCs. Several otherwise successful gene therapy trials of severe combined immunodeficiency using retroviral vectors have resulted in occurrence of leukemia in a significant percentage of patients. Substantial efforts were thus devoted to elucidation of integration patterns and clonal population structure in the hematopoietic compartment after viral transduction, both in experimental models and in clinical trials. The obtained results, although not unanimous, demonstrate nevertheless a frequent occurrence of oligoclonal hematopoiesis after gene therapy, with viral integration sites tending to concentrate in the vicinity of a limited number of genes preferentially involved in growth and proliferation control such as above mentioned Evi-1, PRDM16 or HMGA2. Although upregulation of these genes rarely led to overt neoplastic transformation, it is nevertheless clear that the patients with oligoclonal hematopoiesis are at substantial risk of acquiring leukemias at some future time point.

Various strategies are being currently developed to minimize the risk of neoplastic transformations of HSCs after viral transduction. The most promising approaches include using lentiviral instead of retroviral vectors, and insulators to shield cellular oncogenes from activation by strong viral promoters (Puthenveetil et al., 2004). Insulators, however, tends to significantly reduce viral titers (Nielsen et al., 2009), relatively inefficient (Uchida et al., 2011) and do not provide guarantee against insertional activation of potential oncogenes such as HMGA2 (Cavazzana-Calvo et al., 2010). Another approach is to use promoters specific for differentiated cells that are expected to produce negligible activation of oncogenes in stem cells. However, such promoters tend to provide comparably lower expression levels, and although this might be improved by addition of strong enhancers (Gruh et al., 2008), it is far from certain that such combinations would not activate nearby cellular promoters.

Transposon vectors offer an exciting alternative to retro- and lentiviral vectors. The transposon-based gene delivery combines advantages of integrating viral vectors with those of plasmid vectors. Permanent genomic integration of transposon vectors provides longterm expression, whereas there are significantly fewer constraints on vector design and use of various function elements like insulators. Transposon systems are inherently less immunogenic than viral delivery systems, whereas their cargo capacity generally exceeds that of retro- and lentiviral vectors (Zayed *et al.*, 2004). Initial experiments with transposons were plagued by low efficiency of integration, but continuous improvements in molecular design of transposases have significantly increased the efficiency of integration process (Mátés et al., 2009). Currently, transposons based on Sleeping Beauty (SB) system represent the most advanced version of this technology (reviewed by Ivics Izsvák, 2011), although other system such as piggyBac are also being perfected (Yusa et al., 2011) and may offer some advantages, such as larger cargo capacity, over the SB system (Lacoste et al., 2009).

Although stable SB transposon-mediated gene transfer into hematopoietic cells was reported (Xue et al., 2009), efficient vector delivery to HSCs remains poorly resolved issue, which is currently being addressed by using electroporation or hybrid lentiviral-transposon vectors (Staunstrup et al., 2009). Although certain undesired effects such as SB transposase cytotoxicity were observed, it seems that they might be minimized by controllable mRNA delivery (Galla et al., 2011). Compared to lenti- and retroviral vectors that show preferential integration near active genes, SB transposon vectors demonstrate nearly random integration profiles (Moldt et al., 2011), although this property might not be shared by other transposon systems (Huang et al., 2010).

Another serious safety concern is a direct consequence of a current low efficiency of transduction of LTR HSCs, which necessitates the use of myeloablative pre-conditioning and negative selection strategies to eliminate competing endogenous HSCs and increase chimerism levels. Negative selection strategies using in particular alkylating drugs place a significant stress upon hematopoietic system. However, as demonstrated by Xie et al., 2010, repetitive hematopoietic stress by busulfan administration in a nonhuman primate may rapidly lead to reduction of polyclonality and eventually to cytopenia. In addition, potential long term mutagenic effects of alkylating agents are largely unknown, thus adding more uncertainty as to correct assessment of risks and benefits of this strategy. Apparently, in order to tackle efficiently the problem of low transduction efficiency, it is not sufficient to rely on the use of negative selection only, but is also important to achieve substantial improvements in ex vivo stem cell culturing, expansion and transduction efficiency. Promising approaches also involve use of positive ex vivo and in vivo selection and in situ transduction strategies.
