Gene Therapy

Gene transfer into hematopoietic stem cells has been used successfully for correcting lymphoid but not myeloid immunodeficiencies. Here we report on two adults who received gene therapy after nonmyeloablative bone marrow conditioning for the treatment of X-linked chronic granulomatous disease (X-CGD), a primary immunodeficiency caused by a defect in the oxidative antimicrobial activity of phagocytes resulting from mutations in gp91(phox). We detected substantial gene transfer in both individuals' neutrophils that lead to a large number of functionally corrected phagocytes and notable clinical improvement. Large-scale retroviral integration site-distribution analysis showed activating insertions in MDS1-EVI1, PRDM16 or SETBP1 that had influenced regulation of long-term hematopoiesis by expanding gene-corrected myelopoiesis three- to four-fold in both individuals. Although insertional influences have probably reinforced the therapeutic efficacy in this trial, our results suggest that gene therapy in combination with bone marrow conditioning can be successfully used to treat inherited diseases affecting the myeloid compartment such as CGD.

gamma-Retroviral vectors (gammaRVs), which are commonly used in gene therapy, can trigger oncogenesis by insertional mutagenesis. Here, we have dissected the contribution of vector design and viral integration site selection (ISS) to oncogenesis using an in vivo genotoxicity assay based on transplantation of vector-transduced tumor-prone mouse hematopoietic stem/progenitor cells. By swapping genetic elements between gammaRV and lentiviral vectors (LVs), we have demonstrated that transcriptionally active long terminal repeats (LTRs) are major determinants of genotoxicity even when reconstituted in LVs and that self-inactivating (SIN) LTRs enhance the safety of gammaRVs. By comparing the genotoxicity of vectors with matched active LTRs, we were able to determine that substantially greater LV integration loads are required to approach the same oncogenic risk as gammaRVs. This difference in facilitating oncogenesis is likely to be explained by the observed preferential targeting of cancer genes by gammaRVs. This integration-site bias was intrinsic to gammaRVs, as it was also observed for SIN gammaRVs that lacked genotoxicity in our model. Our findings strongly support the use of SIN viral vector platforms and show that ISS can substantially modulate genotoxicity.

Recombinant adeno-associated viruses (AAV) are among the most promising vectors for gene therapy of genetic diseases, including cystic fibrosis (CF). However, because of its small genome size, the capacity of AAV to package a therapeutic gene is limited. The efficiency of packaging the cystic fibrosis transmembrane conductance Regulator (CFTR) gene into AAV will be an important factor in determining whether recombinant AAV can be developed as a vector for transferring CFTR cDNA to the airway epithelia of patients with CF. Current understanding of the AAV biology suggests that AAV can package a genome slightly larger than the size of a wild-type genome. The precise range of the genome size and the efficiency of packaging have not been defined. Using a series of AAV vectors with progressively-increasing genome size, we were able to analyze quantitatively the packaging efficiency in relation to the vector size and to determine the size limit for packaging. The packaging efficiencies of AAV vectors of variable sizes were determined directly by assaying DNA contents of viral particles, and indirectly by analyzing their efficiency in transfer of a chloramphenicol acetyltransferase (CAT) reporter gene into target cells. Our studies showed that the optimal size of AAV vector is between 4.1 and 4.9 kb. Although AAV can package a vector larger than its genome size, up to 5.2 kb, the packaging efficiencies in this large size range were sharply reduced. When the AAV genome size was smaller than 4.1 kb, the packaging efficiency was also suboptimal. In contrast, when the size of the genome was less than half the length of the wild-type genome, two copies of the vector were packaged into each virion, suggesting that the copy number control during packaging is a "head-full" mechanism. Because the length of the minimal cDNA of CFTR is about 4.5 kb, these results suggest it is possible to package the CFTR gene into AAV if the combined length of transcriptional elements and ITRs is kept under 500 bp. The results of this study are important for directing the design of AAV vectors for efficient gene transfer, as well as for a better understanding of the mechanism of AAV genome packaging.