Abstract
In general, a fetal approach has a number of conceptual advantages compared with postnatal interventions. When reviewing the experimental data of postnatal gene or stem cell therapies, several problems become apparent. These include the following: (1) the underlying defects usually have already caused extensive or even irreversible damage; (2) sufficient transgene expression requires prohibitively large amounts of gene delivery vector; (3) most adult tissues proliferate less than fetal cells, and hence are typically less optimally transduced; and (4) host immune responses, either preexisting or secondary to vector delivery, may rapidly eliminate transgene expression and prevent future interventions. Early gene transfer, in the neonatal or even fetal period, may overcome some if not all of these obstacles (Abi-Nader et al., 2009; Wagner et al., 2009). The fetus enjoys a protective environment in the uterus, which usually prevents the pathophysiological changes induced by the condition. Its organs and tissues are relatively small, necessitating a smaller amount of the therapeutic agent. The presence of many progenitor and stem cells provides the stage for lifelong correction if they can be targeted in a permanent manner. Early in gestation the lack of a mature, adaptive immune system may limit or prevent a destructive immune response against the life-saving transgene.
The majority of experiments in the field of fetal gene therapy have been performed in small animal models. The obvious advantages of these include limited ethics constraints, wide availability, well-documented embryogenesis and fetal development, short gestation, large litter size, and the availability of transgenic disease models. This, however, cannot preclude at some stage the use of a large animal model. Higher species mimic human organogenesis more closely. One such example is lung development. Mice do not alveolize in utero (Thurlbeck, 1975), whereas the human fetus has already done so by about 36 weeks, which is 1 month before birth (Burri, 2006). Also, the maturation of the immune system differs significantly between rodents and humans (Daston et al., 2004). Even though experimental data from small animal models have been instrumental in our understanding of diseases, the inherent differences bring into question the relevance of some observations in lower species. Large animal models such as sheep or primates are better suited for the testing of invasive procedures applicable in the human fetus, but they will have a lower litter size and longer pregnancy duration, imposing limits on cohort size and internal controls. These precious experiments inherently create the need for long-term follow-up methods that obviate or limit the need to sacrifice study subjects.
Blood samples can be used to monitor levels of secreted transgenes or their therapeutic effect (e.g., factor IX and coagulation time in the case of hemophilia), and tissue biopsies can provide the samples for conventional histology, immunohistochemistry, or molecular biological assays (ELISA, quantitative PCR, etc.) to detect the effect of expressed transgenes or engrafted stem cells. Several noninvasive imaging methods can be used to trace or quantify reporter genes or to measure their biological effect. These include magnetic resonance imaging (MRI), computerized tomography (CT), ultrasound scanning (US), positron emission tomography (PET), single-photon emission computerized tomography (SPECT), and bioluminescence imaging (BLI) (Massoud and Gambhir, 2003). The latter is the technique of choice in the study by Tarantal and Lee (2010) in this issue of Human Gene Therapy. The concept of BLI is based on the native light emission from one of several organisms that bioluminesce. The luciferase enzyme can convert its natural substrate in an enzymatic reaction and emit photons that can be detected and quantified with charge-coupled device (CCD) cameras. BLI has proven its value in a variety of applications using small animal models (Dothager et al., 2009).
We highlight the study by Tarantal and Lee (2010) for two particular reasons. One is that they challenge the traditional view that adeno-associated virus (AAV) cannot be used for long-term gene expression in proliferating tissue. The development of recombinant AAV technology has been a major breakthrough for gene therapists. The combination of the nonpathological nature of AAV with the availability of various serotypes with unique tropisms and the flexibility to incorporate a variety of transgenes has endowed researchers with an elegant biological toolkit (Gao et al., 2005; Wu et al., 2006). Studies have reported improvements in the vision of patients with Leber's congenital amaurosis, a genetic degenerative disease of vision, subsequent to injection of AAV into the subretinal space (Bainbridge et al., 2008). These results have infused the gene therapy community with renewed enthusiasm after the latest setbacks (Alton et al., 2007).
The observation by Tarantal and Lee (2010) that AAV can induce long-term gene expression after prenatal injection into nonhuman primates opens an interesting line of research. AAV is still considered an episomal vector system, expected to cause only transient gene expression due to the loss of vector genome by cell division in the growing organs. This is the case in experiments in which the transduction takes place in the fetal or newborn period. Different organs have different proliferation rates, which can in part explain the variance in data concerning long-term gene expression. Some studies showed only transient expression after transduction of the liver (Flageul et al., 2009) whereas others have detected long-term expression after targeting muscles (Wang et al., 2005). Further research is warranted to elucidate the mechanisms by which AAV retains efficient expression over time (e.g., the transduction of stem cells, the induction of immune tolerance, etc.) and which factors are species, tissue, vector, or transgene dependent.
The second important message of this study is that they pioneered the use of BLI in a large animal model. The physical properties of photons render their detection more challenging. Unlike high-energy photons (>1 MeV), such as the gamma rays used in PET scans, the photons emitted by the enzymatic BLI reaction have a low-energy value (<3 eV) and are readily scattered or absorbed in an opaque medium. This makes the interpretation of BLI data quantitatively and qualitatively difficult if the signal emanates from deep structures in the studied organism. The implementation of bioluminescence tomography compensates for this heterogeneity (Wang et al., 2008), using a method similar to CT, producing high-resolution three-dimensional image reconstructions that enable three-dimensional localization and quantification of bioluminescence. The system is based on a complex image reconstruction process, which integrates the modeling of light–tissue interactions combined with anatomical data to compensate for the signal attenuation. The concept is still under development and far from being a standard method in larger organisms. Further improvements in luciferase biochemistry and validation of the reconstructed images with other imaging techniques as well as histological data will enhance the use of this technique in preclinical research.
Clinicians in the field of fetal interventions currently have all the necessary tools and techniques at their disposal to access the fetus confidently with minimal risk to both the mother and child. They welcome studies like these that bring experimental treatments one step closer to the clinic, especially at a time when gene and stem cell therapies have redeemed themselves and regained their status as treatments for the future.