Culican Lab / Projects

Introduction:

Amblyopia, or "lazy eye", is the most common cause of vision loss in childhood, affecting approximately 5% of school-aged children. Patching the stronger eye to change the relative strength of the lazy eye can result in spectacular improvement if performed during a critical period of development when the brain's visual circuits are still being modulated. However, patching once a child has out grown this period results in only modest or no improvement. This type of plasticity is found throughout the developing nervous system, and the cellular mechanisms that underlie this developmental modulation of neuronal circuits is what interests me.

To begin to understand what cellular mechanisms account for the modulation of visual circuits as a result of inter-ocular competition, I am studying the development of connections between the retina and the dorsal lateral geniculate nucleus of the thalamus (LGN) in mice. Early in development, inputs from the two eyes occupy overlapping territories in the LGN. Over time the projections from different eyes become segregated into eye specific layers or areas. This has been demonstrated histologically (the area of overlapping input is reduced) and physiologically (the number of binocularly driven cells is fewer). How this reduction in convergence occurs at the cellular level is not well understood. Determining where on a LGN cell inputs from different eyes are distributed will help explain what cues post-synaptic cells use to maintain some synapses while eliminating others.

The only system where this process of synapse elimination has been shown anatomically for single cells is at the neuromuscular junction. There are many reasons why it has been difficult to study single cells in the visual system, or the CNS in general. First, unlike the neuromuscular system where a few motor neurons contact a single muscle cell in a specific location, there are many hundreds of synapses onto LGN cells. Additionally the brain is much less accessible than muscle, and the anatomy of the visual circuitry is more complex. Lastly, there have been limitations due to a paucity of robust labeling methods to identify pre- and post-synaptic cells in the developing visual system. But with new tools available to us now, including transgenic technology to create animals that express fluorescent proteins in subsets of neurons, and new imaging methods such as two photon microscopy, some of these limitations seem much less daunting.

By understanding the changes in synaptic connectivity that occur during inter-ocular competition, we can identify the cellular mechanisms that define the critical period. By genetically labeling individual retinal cells with fluorescent proteins, I can trace their projections to the brain during the period of inter-ocular competition. In this way I can begin to identify the changes in synaptic morphology that correspond to changes in functional visual circuits during development. Ultimately, by understanding the anatomic rearrangements of synaptic inputs during inter-ocular competition we can begin to describe the cellular mechanisms that define the critical period, which could lead to interventions that lengthen the therapeutic window to treat children diagnosed with amblyopia.

Imaging RGC projections to the LGN during Normal Development:

To image single retinal ganglion cells (RGCs) I have generated a transgenic mouse that expresses fluorescent protein in RGCs. In collaboration with Dr. Rachel Wong, we have used the Cre/loxP system to conditionally express yellow fluorescent protein (YFP) in subsets of RGCs. By crossing our animals (Thy1-newstop/YFP) to a line that expresses Cre only in the retina (αPax/Cre), we achieve cleavage of the stop sequences only in retinal cells, resulting in retina specific expression of YFP (Figure 1). In these animals individual RGCs become genetically labeled with YFP, allowing visualization of entire axonal arbors in the LGN unobscured by corticothalamic projections. The origin of each projection can be determined by co-labeling with fluorescent cholera toxin. Axons that are double labeled with YFP/Alexa 555 originate from the right eye, and those with YFP/Alexa 647 originate from the left eye. The extent of the segregated and unsegregated regions of the LGN can simultaneously be determined by examining the bulk projection from both eyes labeled with the fluorescent cholera toxins. Individual axonal arbors can be imaged and reconstructed using either confocal microscopy in serial brain sections, or imaged in situ using two-photon microscopy (Figure 2). I will follow individual YFP-labeled RGC projections from the retina to the LGN and assess changes in the distribution of these inputs as they undergo monocular segregation.

Activity Independent Regulators of Topographic specificity of retino-geniculate connections:

Phr1 is the mouse ortholog to esrom in zebrafish, and highwire in drosophila. These molecules have been implicated in mechanisms involved with topographic mapping in the tectum and synaptic morphology and synaptic strength, respectively. The mouse ortholog, Phr1, is necessary for normal respiratory function and knock-outs of this molecule result in neonatal death. Aaron DiAntonio (http://molecool.wustl.edu/diantonio.htm) and colleagues have generated a conditional knock-out of Phr1 that affects retinal cells only. Our examination of these animals demonstrates normal retina activity, normal segregation of inputs in the developing dLGN, but abnormal topographic location of retinal inputs within the dLGN (Figure 4). Topographic changes in retinogeniculate projections have been described in many systems where retinal activity has been manipulated by eyelid suture, monocular enucleation, or pharmacologic blockade (see Shatz 1996 for review). There are also a number of mutant animals that have altered retinal activity that have abnormal retino-geniculate projections. This mutant, however, has topographic errors despite normal activity. For this reason, the Phr1 mutant is an excellent model for studying the molecular mechanisms that specify topographic address independently from altered activity.