Published at www.nejm.org April 27, 2008 NEJM
Preliminary Results of Gene Therapy for Retinal Degeneration
Joan W. Miller, M.D.

>In this issue of the Journal, two groups of investigators — Bainbridge et al.
1 and Maguire et al. 2 — describe the first results of separate clinical trials investigating the short-term safety and preliminary efficacy of gene therapy for Leber's congenital amaurosis. Both groups present short-term data (12 months and 5 months, respectively) on three patients with Leber's congenital amaurosis in each study; the patients were enrolled in trials of recombinant adeno-associated viral delivery of the human retinal pigment epithelium-specific 65 kDa protein gene (RPE65), which was administered as a subretinal injection during vitrectomy.
Originally described by Leber in 1869, Leber's congenital amaurosis has the earliest onset of all inherited retinal dystrophies causing congenital blindness; it is also the most severe form.
3 , 4 , 5 Infants with Leber's congenital amaurosis have profound visual impairment or blindness at birth. However, the visual impairment is usually recognized only later when parents note the infant's inability to track objects or light. Severe visual impairment persists throughout childhood, resulting in an inability to read or ambulate independently, and finally in total blindness by the third or fourth decade of life. Other findings include nystagmus (roaming eye movements), abnormal pupillary responses, and flat or nearly undetectable signals on electroretinography (ERG). The appearance of the retina is normal early on but progresses to a pigmentary retinopathy over time. There is no treatment for Leber's congenital amaurosis.
Most cases of Leber's congenital amaurosis are inherited in an autosomal recessive fashion, and disease-associated mutations have been found in seven genes expressed preferentially in the photoreceptor or retinal pigment epithelial (RPE) cell.
4 Mutations in RPE65 account for approximately 6% of cases. 4 , 6 RPE65 encodes an enzyme in the RPE that catalyzes the conversion of all-trans-retinyl esters to 11-cis-retinal. The latter is the precursor of rhodopsin, which is required for phototransduction and vision. The biochemical defect that is associated with the RPE65 mutation leads to immediate, severe vision impairment with subsequent degeneration of the rods and cones.
Animal models of Leber's congenital amaurosis exist, including a strain of Briard dogs with a spontaneous defect in RPE65. Proof-of-principle experiments restored visual function in dogs treated with subretinal adeno-associated virus (AAV) carrying RPE65 complementary DNA.
7 Visual function has remained stable for almost 8 years after a single treatment.
Both Bainbridge et al. and Maguire et al. used recombinant AAV in their clinical trials. AAV was selected because it is noninfectious. However, it can elicit an immune response, since most people have been exposed to AAV and may carry antibodies. Both groups therefore administered perioperative local and systemic corticosteroids and observed no significant inflammation in treated patients. Although both groups delivered DNA coding for human RPE65, Bainbridge et al. used a tissue-specific promoter to limit expression to RPE and excluded patients with null mutations to minimize the possibility of an immune response to the transgene product. Maguire et al. did not use a tissue-specific promoter, having demonstrated effective transgene expression in RPE after subretinal injection in animal studies, as well as some expression in optic nerve, chiasm, and brain.
7 , 8 Maguire et al. did not exclude patients with null mutations, although the reported patients all had missense mutations. Maguire et al. used surfactant to reduce adsorption of viral particles to the contact surfaces of delivery devices and processed the vector to remove empty capsid (which reduced empty capsid from >80% to <2%)> 9 Such responses in patients with Leber's congenital amaurosis are diminished, consistent with decreased photoreceptor input into the afferent arc of the reflex. Maguire et al. found that the pupillary light reflex in treated patients became asymmetric, consistent with a marked increase in signal input from the retina of the treated eye. These objective data are very useful in monitoring the outcome of intervention in this group of patients with retinal degenerations and severely limited vision function, in which changes in more standard clinical outcomes (visual acuity and visual fields) are difficult to quantify.
Maguire et al. observed the development of a macular hole in one patient, which they believed was caused by contraction of a preexisting epiretinal membrane after surgery. Alternatively, as the authors note, subretinal injections in atrophic retina may cause complications, and this observation warrants further study. Given the limited central vision in the enrolled patients, the hole was of no clinical significance; however, it would be clinically significant in patients with better baseline retinal function. Both groups of investigators followed patients with optical coherence tomography, which provides some retinal anatomic detail. High-resolution spectral-domain optical coherence tomography provides greater resolution of the retinal-cell layers before and after treatment; this approach should be considered as these studies move forward and in future studies.
The preliminary results from these investigations suggest that in the short term, the procedure is safe. Moreover, the data are suggestive of efficacy. Both groups recognize that longer follow-up and additional subjects are necessary to provide satisfying safety data. Certainly, efficacy data will be available only in larger trials. Some of the remaining issues include the reproducibility and persistence of the improved retinal function and whether further retinal degeneration is delayed or averted. In addition, systemic or ocular complications may yet be encountered with the treatment of additional patients, with higher doses of vector, and with longer follow-up. One might also speculate that treatment of younger patients with less advanced retinal degeneration might allow greater improvement of visual function. Finally, larger studies may show advantages of a particular vector preparation, promoter selection, or technique.
Dr. Miller reports receiving consulting fees and grant support from Genzyme, which has an AAV gene therapy program. No other potential conflict of interest relevant to this article was reported.
Bainbridge JWB, Smith AJ, Barker SS, et al. Effect of gene therapy on visual function in Leber's congenital amaurosis. N Engl J Med 2008;358. DOI: 10.1056/NEJMoa0802268.