Cepko Lab
Research
SUMMARY
The mechanisms that cells use when they are choose their fate during the development of the central nervous system is the main problem under study in our lab. We have focussed our studies on the retina, a tractable model for the rest of the central nervous system. In addition, we are interested in why photoreceptor cells die in many forms of retinal degeneration.
The retina is the thin sheet of central nervous system tissue that lines the back of the eye (Figure 1). The retinal cell types are arranged in laminae and are conserved in terms of their overall function, orderly appearance during development, and morphology (Figure 2).
Figure 1. A drawing of a
human eye is shown. The retina is the thin layer of tissue lining the eyeball.
An area of high acuity vision, the macula, is even thinner, and is particularly
susceptible to degeneration (e.g.age related macular degeneration). The
information processing that occurs in the retina is quite sophisticated, far
more than would be carried out by a camera and thus the retina is better
analogized to a computer than a camera, though the photoreceptor layer is
somewhat similar to the light sensing elements in a camera. The results of the
information processing that occurs in the retina are communicated to the brain
through the optic nerve. The retina performs information processing to extract
such features as direction of motion, changes in intensity, differences in
wavelength of light, and edges and shapes.
Figure 2. The arrangement of
retinal cells is shown in a cross section. The cells that make up the retina
are the primary sensory neurons (rod and cone photoreceptors), the
interneurons (amacrine, bipolar, and horizontal cells), and the output or
macroneurons (the ganglion cells). The ganglion cell axons project out of the
eye through the optic nerve to the brain.
RETINAL
CELL FATE DETERMINATION
Over
20 years ago, we used a lineage marking technique
to discover that the mitotic progenitor cells that generate the various
neuronal and glial cell types of the retina are multipotent. Products of a
final division can be as different as a specialized sensory neuron, such as a
rod photoreceptor, and an interneuron, such as a bipolar cell. We are
continuing these lineage analyses by sharpening our characterization of the
subtypes of neurons in each clone e.g. examining whether the distribution of
subtypes of horizontal cells in a clone is random or whether there are patterns
in the occurrence of subtypes (See Current Lineage
Studies).
We
have been addressing several questions concerning the mechanisms that lead each
retinal cell to choose its fate in the retina. One question concerns the
properties of the progenitor cells themselves. We have found that they change
over time in terms of their ability to divide, and in terms of their ability to
make different types of retinal cells. We have used genomics methods to
discover the genes that are differentially expressed among individual retinal progenitor cells. This cataloguing
method has shown that there is an enormous amount of heterogeneity among
progenitor cells from different ages and even from the same age. The level of
expression of genes that differ among progenitor cells is being manipulated to
provide information about the function of a gene. Effects on cell fate,
proliferation, differentiation, and survival are being assessed.
In
addition to profiling single progenitor cells, we have profiled newborn neurons of various types from the developing
retina, again at the single cell level, including 32 amacrine cells. These data also show a
great deal of heterogeneity among newborn amacrine cells and have
provided insights into the differentiation processes of newborn neurons. In addition,
they are providing molecular markers for different types of neurons previously
identified using only morphological criteria. We are using some of these
distinctive markers of newborn neurons to determine how much of the neuronal
heterogeneity is derived from progenitor cells. For example, we have analyzed
these data for genes that are differentially expressed among mitotic multipotent progenitor cells and differentially expressed in newborn neurons. Some of
these loci will be modified in mice to encode recombinases, such as Cre or Flp,
so that we can track how these progenitor differences play out in terms of the
cell types generated by different types of progenitor cells. This approach
might enable us to learn more about the lineage patterns of retinal progenitor
cells, as well as provide genes that might direct the diversification processes
in retinal neuron production. Although considered a relatively simple CNS
structure, the retina comprises over 60 cell types, with the majority of the
diversity contributed by the interneurons, such as the amacrine cells.
We
are investigating several particular processes in cell fate determination and
development that have appeared to be important from our preliminary work,
and/or from the literature. The role of Notch and its ligands, and the role of alternative splicing, are two such topics.
PHOTORECEPTOR DEVELOPMENT
Rod
and cone photoreceptors mediate the first step in vision, by capturing light
and performing phototransduction, in dim and bright light, respectively. We
have been investigating the mechanisms by which photoreceptors
are determined, diversify, and differentiate. Investigation of the genes
regulated by photoreceptor transcription factors using knock-out mice and
microarrays has revealed the direct and indirect target genes, and cis-acting
regulatory sequences, controlled by several transcription factors that are very
high up in the photoreceptor hierarchy. In addition, we have identified several
new genes that are required for proper photoreceptor development and are examining
their position in the hierarchy through gain and loss of function approaches.
An independent approach to understanding photoreceptor diversification and
development is also being pursued. We are defining the transcription factor
networks that control expression of genes that are expressed in rods vs. cones
to gain insight into how these two types of photoreceptors become distinct from
each other.
PATTERNING
OF THE RETINA
The retina is not a uniform sheet of cells, but
exhibits various types of patterns. For
example, in humans and chicks, there is a central region (the fovea in humans) that is devoid of
rod photoreceptors. We have identified several genes that control some aspects
of retinal pattern. They are expressed asymmetrically across the retina of mice
and chicks early in development when the patterns are set up. For example. the
homebox protein, Vax, is expressed in a ventral pattern early in development.
Overexpression of Vax ventralizes gene expression patterns of the retina. In addition to transcription factors,
we have begun to explore the role of two small molecules, retinoic acid and
thyroid hormone, in the formation of retinal pattern as the genes that regulate
or respond to them are patterned during critical stages of retinal development.
Thyroid hormone
components have shown dynamic patterns of expression across the developing
retina. For example, mRNA for the deiodinase 2 enzyme, which creates active
thyroid hormone, and mRNA for the deiodinase 3 enzyme, which destroys active
thyroid hormone, are expressed in waves across the developing retina. The
function of these enzymes and other components of thyroid hormone are being
investigated using gain and loss of function protocols.
RETINAL DEGENERATION
In
many forms of human retinal degeneration, rod photoreceptors express a mutant
gene, while cone photoreceptors do not. Interestingly, the cones still die.
This implies a non-autonomous process in the death of cones. Humans rely most
heavily on cones, and thus we would like to understand why the cones die so
that gene therapy or a pharmaceutical intervention can be developed. To this
end, we are using microarrays to examine several mouse models of human disease
to discover the non-autonomous process leading to cone death. These studies
have revealed that cones are undergoing autophagy, which might lead to their
death. We are currently investigating this possibility. In addition, we have
discovered that HDAC4, a histone deacetylase, is in the cytoplasm of developing
retinal neurons, including rod photoreceptors. HDAC4 is required for rod
survival, as we demonstrated by loss of function experiments in vivo in wild
type mice. In addition, it can prolong the survival of rods in a disease model in
which rods die due to intrinsic genetic defects. We are investigating the
mechanisms of this effect.
RESEARCH SUPPORT
We are grateful to the following
institutions for current and past support of our studies: Howard Hughes Medical
Institute, The National Eye Institute, The Foundation for Retinal Research, The
Macular Disease Research Foundation, and Merck.
Some Ongoing
Projects
Retinal Progenitor Cell Heterogeneity
The Formation of Retinal Pattern
The Role of Splicing in Retinal Development
Comprehensive Expression Profiles of Single Retinal
Neurons