Are Lentivirus Vectors Safer?

Abstract

In 2003, the field of gene therapy faced serious adverse events, apparently arising in part from insertional activation of the LMO2 gene in a clinical gene therapy trial in Paris that sought to correct deficiency of the interleukin-2 common γ-chain receptor (IL2RG). Although this remains the only human gene therapy trial reporting such side effects (including an almost identical trial in London), and the molecular mechanisms involved in the leukemogenesis are still being hotly debated, the development of leukemia in three participants and the previous demonstration by Baum's group of a mouse leukemia associated with replication-defective Moloney leukemia virus vectors1Murine leukemia induced by retroviral gene marking.Science. 2002; 296: 497Crossref PubMed Scopus (571) Google Scholar added momentum to the development of lentiviruses as vectors for hematopoietic stem cell (HSC) transduction. Lentivirus vectors have been shown to transduce noncycling cells—a potentially important consideration for HSCs that are largely quiescent in the bone marrow. However, Bushman and colleagues2HIV-1 integration in the human genome favors active genes and local hotspots.Cell. 2002; 110: 521-552Abstract Full Text Full Text PDF PubMed Scopus (1400) Google Scholar have shown that lentivirus vectors seem to prefer to integrate near active transcriptional units rather than near the transcriptional start sites preferred by γ-retroviruses. Subsequent studies have shown that both lentivirus and γ-retrovirus vectors seem to integrate in highly expressed genes and in gene-dense regions. The activation or deregulation of genes, including oncogenes, in the vicinity of vector integration sites could thus be a major safety consideration for all integrating vectors. It is critical to examine this issue in greater detail before abandoning more than a decade of safe clinical experience with γ-retrovirus vectors. First, the several studies published thus far use very different target cells and usually rely on cell lines as opposed to primary cells for integration analysis. This could be an important variable if epigenetic modifications such as histone-associated target DNA influence site selection. In addition, in most cases to date, lentivirus vectors using self-inactivating (SIN) configurations have been compared with γ-retrovirus vectors with intact long terminal repeats—hardly a fair comparison. Because enhancers can mediate significant changes in transcription rates over many kilobases, the differences in integration patterns noted above, which are usually in the range of twofold, and statistically significant, could yet be biologically irrelevant. At the 3rd Stem Cell Clonality and Genotoxicity Retreat held in December 2006 in Orlando, it became clear that several groups are addressing this issue experimentally from a more rigorous standpoint. A potential limitation of side-by-side comparisons is that the various members of the Retroviridae may differ in their susceptibility to cellular restriction factors in a vector-specific manner in model species. Nevertheless, more than 3 years after the initial report of vector-related leukemia, no clear answer as to whether lentivirus vectors are safer is yet evident. This is noteworthy, because there are many instances in which workers in gene therapy have shifted en masse after early reports of advances in vector design. One example comes to mind: the transition from amphotropic to gibbon ape leukemia virus (GALV) to RD114 packaging pseudotypes. Where are the data showing that one of these pseudotypes is actually significantly superior in the setting of human HSC targets and clinical trials? These considerations are of more than passing importance. They shape the expectations of both regulators and peer reviewers. Reviewers may come to believe that one or another system is “standard” or “accepted” without a firm basis on which to do so, and such a belief clearly could influence the review process, leading to lack of enthusiasm for continued studies comparing vector systems. For instance, in some study sections, it is now assumed that lentivirus vectors should be developed for clinical usage and γ-retroviruses abandoned, even though no human trial using lentivirus vectors has been approved by the FDA using HSCs as targets. Although in some instances lentivirus may emerge as the preferred vector system, it is equally likely that systems based on γ-retrovirus, foamy virus, or transposable elements with added safety features, such as SIN design, cellular promoters, insulators, and strong poly(A) signals to prevent readthrough, will provide advantages in many diseases. We should individually and collectively expect reasonably designed experimental comparisons and keep our minds open to the data generated from these experiments.