Mouse Versus Monkey

The era of gene therapy is coming – there’s no doubt about that. And the eye is a promising candidate, as it provides easy surgical access, good visualization of the treated tissue, and has a (relatively) immune privileged status. Over a dozen gene therapies for retinal disease are currently in clinical trials, and many more are in the pipeline. Adeno-associated virus (AAV) serotype 2 (AAV2) is the vector that, for the most part, has been used safely and successfully in the vast majority of these trials. But there’s a problem. Mouse studies have shown that intravitreal injection (IVI) of AAV2 results in the transduction of the innermost retinal ganglion cells, but not the photoreceptors in the outer retina. Subretinal injection transduces both, but this approach is considerably more challenging, invasive, costlier and riskier to perform than an IVI. In an ideal world, you’d have a vector and genetic payload that can be injected by IVI, and penetrate and target the cell types of interest – even those in the outer retina.

Fortunately, other AAV serotypes have emerged since AAV2’s discovery, and have the ability to target particular cell types and effectively penetrate certain tissues. We now have the capability of performing in vivo-directed evolution whereby the configuration of the virus’ capsid (and therefore the cells it infects and delivers its payload to) can be altered to target a desired cell type (1). The problem is, most of the work with these vectors has been limited to mouse studies – and the anatomical and physiological differences between murine and human eyes means that what might work in mice, might not work in the clinic. A team from the Perelman School of Medicine, Philadelphia, decided to put two eGFP-expressing AAV serotypes to the test, namely AAV7m8 and AAV8BP2, both of which had successfully been used to transfect mouse retinal cells. But instead of using mice, they used non-human primates (NHPs; cynomolgus macaques to be specific) in the hope of providing a better prediction of the outcome in the human retina (2).

They found some important interspecies differences: in mice, AAV7m8 was able to reach photoreceptor and retinal pigment epithelium cells in the outer retina after IVI, whereas in NHPs, this happened only at the highest dose of vector (which was also associated with severe inflammation and cell injury). One of the reasons AAV8BP2 attracted interest was because it transfects cone photoreceptors and bipolar cells after subretinal injection in mice, but in NHPs it only transfected cone receptors efficiently. After intravitreal delivery, both vectors achieved transduction in the anterior chamber and the optic pathway. Notably, AAV8BP2 had the better safety profile, even at higher doses. The upshot? The authors state in the paper: “This study shows that one cannot extrapolate directly between mice and ‘men.’”