Fos-tau-LacZ mice expose light-activated pathways in the visual system

We have employed fos-tau-LacZ (FTL) transgenic mice to examine functional activation in the visual areas of the nervous system. The FTL mice express the marker gene lacZ in neurons and their processes following many different stimuli, and allow the imaging of activation from the level of the entire brain surface through individual neurons and their projections. Analysis of FTL expression in the retinas of mice following diurnal exposure to light shows that bipolar cells, specific classes of amacrine cells, ganglion cells, and a dense network of processes in the inner plexiform layer are functionally activated. In animals deprived of light, there is almost no activity in the retina. In the lateral geniculate nucleus (LGN), light exposure appears responsible for FTL expression in dorsal nuclei, but not for expression in the ventral nuclei or the intergeniculate leaflet. In the superficial layers of the superior colliculus, FTL expression is highly dependent on light exposure. Similarly, light exposure is required for FTL expression in primary visual cortex (area 17), but some expression remains in area 18 of dark-adapted animals. Finally, using mice with one or both eyes missing, we have determined which parts of the visual system are dependent on the presence of a functional connectivity from the eye. These data demonstrate the usefulness of the FTL mice to map functional activation within the entire visual system. Furthermore, we can capture visual activation in a conscious animal. Our findings give an insight into the architecture of activity within the retina and throughout the visual system.

Introduction

Visual processing begins in the retina. Here, contrast, color, shape, and movement of an object are analyzed and processed in discrete circuits and pathways. The processed primary information is then sent via the output cells of the retina, the ganglion cells, to the lateral geniculate nucleus (LGN) of the thalamus, and the superior colliculus of the midbrain. The lateral geniculate nucleus (LGN) is the first relay station on the pathway to the visual cortex, while the superior colliculus (SC) is involved in the control of eye movement. Projections from the LGN arrive in layer 4 of the primary visual cortex (area 17 in cats and rodents). In this region, there are extensive internal connections between layers as well as projections onward to other visual cortical areas.

The anatomy of the retina has been studied in detail by immunohistochemistry combined with serial sectioning and electron microscopy. Moreover, electrophysiological studies have greatly assisted in indirectly understanding functional connectivity of the retina. However, these studies have clear limitations. For example, in vivo extracellular recordings can only be undertaken in unconscious animals and have been restricted to ganglion cells, the most accessible neurons of the retina (Boos et al., 1990, Muller et al., 1992). In vitro patch clamp recordings or intracellular recordings are only possible in single cell or retinal slice preparations, in which the cells are accessible (Jensen, 1991, Karschin and Wassle, 1990). Further, in vivo information of the activity of the interneurons in the retina, which are connected in an intricate network, is not available. Thus, whereas its structural connectivity is known to a great extent, imaging the functional activation within the network of different cells has not been adequately achieved.

In the brain, tracing studies have revealed the structural pathway from the retina to subcortical and cortical regions. In the visual cortex, electrophysiological recordings as well as optical imaging techniques have revealed functionally activated areas in primates and nonprimates, including the mouse (Grinvald, 1992, Grinvald et al., 1986, Marino et al., 2003, Schuett et al., 2002). In human, functional imaging techniques like fMRI and PET have also been useful to image functional activation in the visual system (Wandell and Wade, 2003). All these techniques measure or visualize global activity across large populations of neurons, however, they have limited spatial resolution and cannot resolve individual functionally activated cells or any connectivity between these cells.

Another method to image functional activation even at single-cell resolution is the detection of Fos, the protein product of the immediate early gene c-fos. c-fos is rapidly induced following neuronal activation and thus can be used as a marker for neurons that have been activated (Chaudhuri, 1997, Morgan and Curran, 1991). However, because c-fos is expressed exclusively in the cell nucleus, its localization does not provide information of the connectivity or the morphology of activated neurons within the nervous system. We have generated transgenic mice in which an axon-targeted β-galactosidase reporter system (Callahan and Thomas, 1994) is under the regulation of the promoter of the c-fos gene (Wilson et al., 2002). In these fos-tau-LacZ (FTL) transgenic mice, neurons that express c-fos express β-galactosidase (βgal) in their axons and dendrites, permitting direct visualization of their projections using a simple enzymatic assay (Wilson et al., 2002). Thus, the FTL transgenic mice may have certain advantages in imaging functionally activated areas and circuits in the visual system, in that it may allow the identification of cell bodies and their projections.

In this study, we have used FTL mice to image the entire visual system and to follow stimulation from the retina to the LGN, SC, and visual cortex. This allows us to simply correlate functional activation from visual input to final processing in cortex and surrounding areas. The advantage of this approach is that we can look at the results of visual activation in animals that were conscious (in the absence of anesthesia) and we can follow this activation throughout the visual system at very high spatial resolution. It is simple to unequivocally identify cell types in the retina by double staining for βgal and retinal markers. Further, we can quantitate and map those cells, which are activated by a particular stimulus and look for the downstream effects of these stimulated neurons. Finally, the FTL mouse has the advantage of simple and accurate spatial mapping of activated regions on the surface of the brain, including cortex.