How to keep photoreceptors alive.

O ver the last 40 years, there has been increasing success in the surgical treatment of retinal detachment in that the retinal reattachment could be achieved in a high proportion of cases but visual recovery was frequently poor. An explanation for this poor visual outcome was derived from work on experimental retinal detachment in which it was shown that photoreceptor cell death occurred because of a wave of apoptosis during the first few days of retinal detachment (1). It is recognized that apoptosis or programmed cell death is a genetically encoded potential of all cells (2). It is characterized by cleavage of most of the nuclear DNA into short but well organized chains of nucleosomes in multiples of 200 bp by an endogenous nonlysosomal nuclease (3, 4) and may be triggered by changes in the metabolic environment of the cell. The photoreceptor cells are closely approximated to the retinal pigment epithelium (Fig. 1) and depend on the retinal pigment epithelium for their metabolic sustenance. Physical separation of the two and loss of metabolic exchange as occurs in retinal detachment (Fig. 2) might have been supposed to cause a sequence of events that would inevitably induce cell loss that was not amenable to treatment. However, apoptosis may be manipulated by altering the metabolic environment, and it has been shown that brainderived growth factor injected into the eye reduces the rate of photoreceptor cell loss in experimental retinal detachment (5), although the precise means by which apoptosis was induced and the mechanism of the therapeutic effect were uncertain. In this issue of PNAS, Nakazawa et al. (6) describe convincing evidence that monocyte chemoattractant protein 1 (MCP-1) plays a critical role in inducing photoreceptor apoptosis in experimental retinal detachment in mice. It had been reported some years ago that levels of MCP-1 were high in the vitreous of patients with retinal detachment (7). MCP-1 seemed to be an attractive candidate as an inducer of cell loss because it had been proposed as playing a role in the pathogenesis of a variety of disorders of the central nervous system including Alzheimer’s disease. Nakazawa et al. showed that the level of MCP-1 in mice with retinal detachment was increased 10-fold in the vitreous when compared with normals and that its expression in Muller cells (a class of retinal glial cell) was up-regulated after 72 h of detachment, a time of maximum apoptosis. Apoptosis was reduced by injecting an anti-MCP-1 blocking fragment intravitreally. Apoptosis was also reduced by 80% in MCP-1 knockout mice with retinal detachment. In each case, suppression of apoptosis was accompanied by reduction of CD11b macrophage/microglial cells, invading cells that are found universally in retinal detachment. Interestingly, Nakazawa et al. provide evidence with both in vitro and in vivo experiments that the effect is not mediated by a direct effect of MCP-1 on photoreceptor cells; rather, it was mediated through the activated macrophage/microglial cells. This information is very important because there are clear therapeutic implications for the acute management of retinal detachment in humans. Hopefully it will be possible to transfer these findings into treatment strategies that can suppress the photoreceptor cell loss such patients often experience. These findings also have broad implications for retinal diseases, including diabetic retinopathy, retinal vascular occlusions, and retinal dystrophies. A brief summary of relevant work on the pathophysiology of retinal dystrophies and experimental therapeutics for these dystrophies will serve to put the work of Nakazawa et al. (6) in perspective. In both humans and animals, there has been increasing reason to believe that in inherited retinal diseases, the metabolic defect caused by the mutation does not cause cell death directly. This observation is evident clinically with respect to cone loss in patients with retinitis pigmentosa caused by mutations in the rhodopsin gene, which is expressed exclusively in rod photoreceptor cells. A similar situation exists in mice transfected with a mutant rhodopsin gene (8, 9). It is also illustrated in the setter dog with progressive atrophy of both the rod and cone photoreceptors, with phosphodiesterase activity being defective in rods but not in cones (10). Most striking are the observations on a chimera created of a pigmented mouse transfected with a mutant rhodopsin gene and a WT albino mouse. The setting was achieved by generating chimeric embryos composed of cells from both the albino and pigmented mouse lines. Although there was patchy distribution of cells from the pigmented and nonpigmented origin, the distribution of photoreceptor cell death followed exactly the same pattern as that seen in the pigmented rhodopsin mouse, implying that