Direct and indirect retinohypothalamic projections to the supraoptic nucleus in the female albino rat.

Earlier studies have shown that retinohypothalamic projections terminate extensively within the hypothalamus of the rat. Recently, we identified a light retinal projection to the supraoptic nucleus as well as a larger, well-focused projection resulting in a peri-supraoptic nucleus terminal field. In this study, we employed a double labeling method with cholera toxin conjugated to horseradish peroxidase (CT-HRP) and pseudorabies virus, a transsynaptic neural tracer, to evaluate retinorecipient neurons in both the supraoptic nucleus and peri-supraoptic nucleus terminal field. In addition, we looked for evidence that cells in the peri-supraoptic nucleus terminal field project into the supraoptic nucleus. Three strains of pseudorabies virus were compared. A direct retinosupraoptic nucleus circuit was confirmed with all three strains. Retinorecipient neurons in the peri-supraoptic nucleus terminal field were also confirmed. However, there was a strain-based difference in the identification of these neurons. The wild-type Becker strain labeled cells in the peri-supraoptic nucleus terminal field in a manner paralleling the early, intermediate and late stages of infection of the suprachiasmatic nucleus. This parallel time course suggests that retinal ganglion cells terminate directly on cells in the peri-supraoptic nucleus terminal field. Conversely, the Bartha and PRV-91 strains showed appreciable labeling of peri-supraoptic neurons only at long survival times. This longer time course suggests that these mutant strains label neurons in the peri-supraoptic nucleus terminal field indirectly, after passing through additional neurons. In addition, experiments with monocular injection of CT-HRP and posterior pituitary injection of pseudorabies virus showed retrogradely labeled second-order cells in the peri-supraoptic nucleus amidst the CT-HRP labeled terminal field of the retinohypothalamic tract. These results demonstrate a direct projection from the retina to the supraoptic nucleus and provide evidence for an indirect circuit from the retina to the supraoptic nucleus via neurons located in the peri-supraoptic nucleus terminal field. The strain-based differences imply that those retinal ganglion cells that project to the peri-supraoptic nucleus terminal field differ from those that project to the suprachiasmatic nucleus. In addition, these results suggest a neuroanatomic basis for photic effects on physiological mechanisms that are not mediated by the circadian timing system.