Photopharmacology puts the light sensitivity into the chemistry. A light-sensitive switch is written into the drug's structure, and illumination changes the drug's activity in a way that light can also reverse [10]. The reversibility is the whole point. A switch that closed once would report the first photon and nothing afterwards.
The design choice worth attention here is anatomical rather than chemical. When photoreceptors degenerate and die, the deeper retinal layers often survive and remain capable of firing; what they lose is their input [7]. Rosalba Sortino, co-first author, says the group aimed to re-enter the circuit at the same level as the lost photoreceptors instead of bypassing the retina's own processing [12]. If that holds, the surviving retina keeps doing contrast and adaptation itself, and the molecule does not have to encode them. It is a testable prediction, and the material circulated so far does not test it [17].
That material breaks off mid-sentence, right where the results begin [17], so the quantities that decide the result are not on the table: how many animals, at what wavelength and irradiance, and whether the behavioral assay separates light-guided navigation from a startle response. The field's own stated difficulty is performance at ordinary levels of illumination, where light-responsive drugs have looked safe without yet delivering high-quality vision [9]. The claim that no specialized lighting was needed [1] is the sentence I most want a photometric number underneath.
The release frames the stakes with two figures: about 200 million people affected by age-related macular degeneration and retinitis pigmentosa [5], and a global vision-loss burden above US$400 billion a year in healthcare costs and lost productivity [6]. Divide one by the other and you get roughly $2,000 per affected person per year [15]. Do not use it. The burden covers all vision loss while the 200 million counts two diseases, so numerator and denominator describe different populations [15].
My read is that the delivery route, rather than the restored behavior, is the load-bearing finding [2]. Light-responsive drugs are already in clinical testing [9]; the release itself points at drops as the simplification [1]. A compound that reaches inner retinal neurons from the corneal surface changes who can administer it and how often, because there is no vector to manufacture and no device to implant or train a patient on [4][8]. Gene therapy, by the consortium's own account, fits only a small fraction of patients with particular mutations [8].
The condition on that view is the one Pau Gorostiza states plainly: these molecules do not cure blindness, because they do nothing about the cause of degeneration [11]. Photoreceptors keep dying underneath the treatment, which means the treatable window is set by how much inner retina remains [7]. More than ten years of work went into getting a photoswitch onto a molecule that survives an eye drop, putting the program's start no later than 2016 [13][16].
