ARTICLE IN PRESS
Progress in Retinal and Eye Research 27 (2008) 260–283
www.elsevier.com/locate/prer
The genesis of retinal architecture: An emerging role
for mechanical interactions?
Lucia Galli-Restaa,, Paola Leoneb, David Bottaria, Monica Ensinia,
Elisa Rigosia, Elena Novellib
a
Istituto di Neuroscienze CNR, Via G. Moruzzi 1, 56100 Pisa, Italy
b
Fondazione Bietti Onlus I.R.C.C.S., 00198 Rome, Italy
Abstract
Patterns in nature have always fascinated human beings. They convey the idea of order, organization and optimization, and, to the
enquiring mind, the alluring promise that understanding their building rules may uncover the forces that shaped them.
In the retina, two patterns are outstanding: the stacking of cells in layers and, within the layers, the prevalent arrangement of neurons
of the same type in orderly arrays, often referred to as mosaics for the crystalline-like order that some can display. Layers and mosaics
have been essential keys to our present understanding of retinal circuital organization and function. Now, they may also be a precious
guide in our exploration of how the retina is built.
Here, we will review studies addressing the mechanisms controlling the formation of retinal mosaics and layers, illustrating common
themes and unsolved problems. Among the intricacies of the building process, a world of physical forces is making its appearance. Cells
are extremely complex to model as ‘‘physical entities’’, and many aspects of cell mechanotransduction are still obscure. Yet, recent
experiments, focusing on the mechanical aspects of growth and differentiation, suggest that adopting this viewpoint will open new ways
of understanding retinal formation and novel possibilities to approach retinal pathologies and repair.
r 2008 Elsevier Ltd. All rights reserved.
Contents
1.
2.
3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
An outline of retinal development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1. A single proliferating cell layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2. Two layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3. Four layers and at least four mosaics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4. The mature pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The development of retinal mosaics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1. Early retinal mosaics and why we care about them . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2. The spatial rule that we call mosaic regularity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3. Predominance of homotypic-cell interactions in mosaic formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4. Potential players contributing to cell spacing within mosaics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
261
261
262
263
264
264
265
265
265
266
266
Abbreviations: AB, apico-basal; AJ, adherent junction; Crb, Crumb; DAH, differential adhesion hypothesis; DG, dystroglycan; ECM, extracellular
matrix; GCL, ganglion cell layer; GFP, green fluorescent protein; ILM, inner limiting membrane; INL, inner nuclear layer; IOP, intraocular pressure; IPL,
inner plexiform layer; NBL, neuroblastic layer; OLM, outer limiting membrane; ONL, outer nuclear layer; OPL, outer plexifom layer; RPE, retinal
pigment epithelium.
Corresponding author.
E-mail address: galli@in.cnr.it (L. Galli-Resta).
1350-9462/$ - see front matter r 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.preteyeres.2008.02.001
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
4.
5.
6.
3.4.1. Cell death in mosaic formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.2. Does cell fate determination contribute to mosaic regularity? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.3. Tangential migration in mosaic assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5. Hypotheses on the nature of homotypic interactions controlling mosaic formation: a role for dendrites?. . . . . . . . . .
3.6. Mosaics as building blocks of retinal assembly? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.7. Exploring the link between mosaics and layer formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.8. Mosaic development in short form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Investigating the mechanisms controlling cell layering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1. Role of classical adhesion molecules and basal lamina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1. Basal lamina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.2. Receptors for extracellular matrix components: integrins and distroglycan . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.3. Cadherins and their intracellular ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2. The crb and Apical Polarity complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3. Diffusible factors and the role of RPE: retinal re-aggregates and transplants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4. A summary of layer development: common end points for common mechanisms? . . . . . . . . . . . . . . . . . . . . . . . . .
An emerging role for mechanical forces in retinal development?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1. Mechanical components in the control of eye size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2. Naı̈ve considerations on retinal cohesion during development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3. The mechanical properties of retinal tissue and cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4. The beginning of retinal layers: a mechanical explanation for RGCs basal migration . . . . . . . . . . . . . . . . . . . . . . .
5.5. Cell sorting and the differential adhesion hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.6. Tension induces process outgrowth and stabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.7. The cell ‘‘tactile set point’’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.8. The strain-stiffening behaviour of cells and biological polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9. A mechanical link between mosaics and layers? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction
The vertebrate retina is an outstanding example of
modular neural circuitry. Its neurons are stacked in layers,
separated by plexiform strata where processes and connections are confined (Fig. 1). Different cellular layers contain
different neurons, and, within each layer, most neuronal types
are arranged in regular arrays, known as retinal mosaics.
This modular organization has been the key to our current
understanding of retinal function and organization. The
stratification of neurons and processes in the mature retina
has allowed the elucidation of a basic ‘‘functional unit’’
comprised of photoreceptors, bipolar, and ganglion cells,
which, in each point in the visual field, operates from the
capture of light to the delivery of visual information to the
brain. The serial organization of this basic unit is directly
reflected in the sequential position of its cellular components
across the retinal depth, while the parallel arrangements of
many such units allows visual processing throughout the
visual field. Further exploration has uncovered the integrating mechanisms involving horizontal and amacrine cells, the
different circuitry for rods and cones, and the ON and OFF
pathways—to mention a few—all of which are wired in
modular sub-circuits within the retina (Wässle and Boycott,
1991). Therefore, very schematically, the retina can be viewed
as a tri-dimensional array of like functional units. Layers and
mosaics are two main orthogonal views of such tridimensional organization (Fig. 1).
261
266
266
266
267
268
268
268
269
269
269
269
270
271
271
272
273
274
274
275
276
277
277
278
279
279
279
280
280
This review looks at retinal development through the
lens provided by the studies investigating mosaic and layer
formation. It is organized in four sections. The first section
is a brief overview of the major steps in retinal development. The second section deals with retinal mosaics, their
potential blue-print role in retinal development and our
current knowledge of the mechanisms controlling their
formation. The third section reviews studies addressing the
molecular and cellular players that control layer formation
in the retina. As it will be seen, many studies reveal an
important role of adhesion molecules and their binding
partners in controlling retinal patterning. This suggests
that mechanical forces may have a significant role in retinal
development or, for the sceptical, that a mechanical view to
retinal development might provide interesting novel insights. For this reason, the last section will consider what
we know about the mechanics of retinal development,
alongside recent challenging data revealing unexpected
effects of mechanical forces on cell growth and differentiation. This last section is inevitably highly speculative, but it
is intended to illustrate new ways of approaching retinal
formation, and to highlight their potential in dealing with
retinal pathologies and repair.
2. An outline of retinal development
This section sketches an outline of vertebrate retinal
development to be used as a reference frame for the
ARTICLE IN PRESS
262
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
Fig. 1. (a) The retinal organization as depicted by Ramon y Cajal (1892): the layers occupied by cell bodies (D, F, H) alternate to layers where cell
processes and synaptic contacts are confined (E, G). A Muller glia cell (ň) and an astrocyte (o) are separately represented on the right. A ¼ pigment
epithelium, B ¼ photoreceptor inner and outer segments, C ¼ outer limiting membrane, I ¼ nerve fibre layer, J ¼ inner limiting membrane. (b) A
schematic drawing of the regular arrays, or mosaics, formed by individual cell types within each retinal layer.
experiments described in the following sections. Data from
multiple species are referred to, trying to draw a general
common picture, and ignoring species-specific peculiarities.
A more detailed presentation of retinal development can be
found elsewhere (Sernagor et al., 2006).
2.1. A single proliferating cell layer
The eye develops from a lateral out-pouching of the
neural tube which, through a sequence of inductive events,
grows in size, folds inwards and develops into a cup formed
by two cell monolayers (Fig. 2), the retinal pigment
epithelium (RPE) in the outside, and the retinal neuroepithelium in the inside. The two layers are initially
separated by a ventricle. In a first phase, both epithelia
are proliferating, which leads to a rapid increase in the size
of the cup. At this stage, the retinal neuroepithelium is an
array of parallel-elongated cells, the neuroblasts, densely
packed one next to the other. On the basal side, the cells are
anchored to a basal lamina of extracellular matrix (ECM),
the inner limiting membrane (ILM). At the opposite, apical
side, which faces the RPE, the cells are ‘‘stitched’’ to one
another by transmembrane protein complexes, the adherens junctions (AJs), which link the cytoskeletons of
adjacent cells (reviewed in Tepass and Harris, 2007).
While neuroblasts are actively incorporating DNA and
dividing, their nuclei perform a sort of back and forth dance,
which brings the nucleus to a most basal position when
DNA duplication is complete, and back to the apical region
to engage in mitotic division (reviewed in Baye and Link,
2007). Throughout the cell cycle, cells retain their apical and
Fig. 2. Schematic representation of the two-layered optic cup during early
eye development. Outside (top) lies the retinal pigmented epithelium
(RPE), inside (bottom) the developing retinal neuroepithelium. The two
are separated by a ventricle. Adjacent retinal neuroepithelial cells are
joined by a belt of adherens junctions (AJs) on their apical (A) side, and
rest on a lamina of ECM, the inner limiting membrane (ILM) on their
basal (B) side.
basal links. Electron microscopy shows that dividing cells
are linked through AJs to the adjacent cells (Kuwabara and
Weidman, 1974). Real-time imaging (Fig. 3) shows that after
a cell division, one of the daughter cells maintains the basal
process of the progenitor, while the second daughter rapidly
develops a new one (Das et al., 2003).
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
263
Fig. 3. Real-time imaging of cell division in the retina of a living zebrafish. Many cells express GAP-GFP, and are thus fluorescent. The dividing cell is
outlined in green. Apical is up, basal is down. The time-lapse series shows the cell dividing, while the basal process (arrows) remains in place during the
entire event. In preparation for division, the nucleus and all the cytoplasm move to the apical side of the retina. After division (t_400 ), the cell on the left
seems to have inherited the basal process while the cell on the right seems to be growing a new process alongside its sibling (arrowheads). Reproduced from
Das et al. (2003), with permission.
Fig. 4. Real-time imaging of the genesis of a retinal ganglion cells (RGC) in the retina of a living Zebrafish. The highlighted cell is an ath5:GFP progenitor
transplanted in a wild-type environment. t_0 corresponds to the time of appearance of ath5:GFP. Daughter cells have been highlighted with different
colours (red or yellow). Apical is up, basal is down. After division (t_45 min), both daughter cells migrate toward the basal surface. At t_5 h, one daughter
cell (red) migrates toward the apical surface, whereas the other one (yellow) retracts the apical process and starts to grow an axon toward the basal surface.
The white arrows point at the retracting apical process and the growth cone at the tip of the axon. The dashed line indicates the basal edge of the retina.
AP, apical cell; RGC, retinal ganglion cell. Reproduced from Poggi et al. (2005a, b), with permission.
2.2. Two layers
The first cells to leave the mitotic cycle are the retinal
ganglion cells (RGCs). Real-time imaging shows that a
newly born RGC develops a growth cone from its basal
process and from this an axon, which runs along the ILM,
toward the optic nerve (Fig. 4). In the meantime, the RGC
cell body is moved to the basal side of the retina, as if
towed by the leading growth cone, and the apical process is
retracted (Poggi et al., 2005a; Zolessi et al., 2006). Later,
the cell starts to develop dynamic processes from the apical
portion of the cell body, from which the dendritic tree
develops (Kay et al., 2004). The basal migration of RGCs
creates the first cellular stratification in the retina, which
becomes subdivided into a neuroblastic layer (NBL) and a
ganglion cell layer (GCL).
ARTICLE IN PRESS
264
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
2.3. Four layers and at least four mosaics
Soon after the RGCs, horizontal cells, cones, and
amacrine cells start being generated (Rapaport, 2006). All
have migratory behaviours different from the RGCs. Both
the cones and the horizontal cells remain in the NBL. The
latter cells, which can be discriminated morphologically
from the other NBL cells, have processes directed to the
apical side of the retina (Ramon y Cajal, 1892, 1929), and
cell bodies that move across the depth of the NBL (Edqvist
and Hallbook, 2004). The amacrine cells migrate basally,
towards the GCL, apparently retaining neither the apical
nor the basal process (Fig. 5). Rather, they produce short
dynamic processes in all directions, as if exploring the
neighbourhood. Once they get close to the GCL, they stop
the cell body migration and continue to ‘‘explore’’ the
environment in all directions with short dynamic processes.
The processes that run towards the GCL are selectively
stabilized, grow longer, ramify, and eventually give rise to
the cell dendritic tree. The thin layer of processes containing the amacrine cells dendritic trees, found at the basal
side of the NBL, and apical to the GCL, is the first
rudiment of the inner plexiform layer (IPL). Interestingly,
the dendritic trees of amacrine cells are immediately aimed
at either the apical or basal region of the IPL, where they
are targeted by the RGC dendrites (Godinho et al., 2005).
At this stage, the retina has four morphologically
distinguishable layers—the NBL, a rudiment of inner
nuclear layer (INL) basal to and continuous with the
NBL, the IPL and the GCL—and at least four detectable
cell mosaics (Fig. 6)—a regular distribution of cholinergic
amacrine cells in the GCL, a second one in the INL
rudiment (Galli-Resta et al., 1997), and, in the NBL, an
orderly subpopulation of cones (Bruhn and Cepko, 1996;
Bumsted et al., 1997; Larison and Bremiller, 1990;
Raymond et al., 1995; Wikler and Rakic, 1991, 1997),
and a non-random array of horizontal cells (Novelli et al.,
2007; Raven et al., 2005; Scheibe et al., 1995).
2.4. The mature pattern
The next step, in terms of strata, is the appearance of the
outer plexiform layer (OPL), which is initially composed by
the processes (and the first synaptic contacts) of the
horizontal and the photoreceptor cells. At this stage, the
retina is still missing its central ‘‘ingredient’’, the bipolar
cells, somewhat like a sandwich without part of its filling.
In the last phase of neurogenesis, mostly rods, bipolar
and Muller cells are born (Rapaport, 2006). The bipolar
cells differentiate from neuroepithelial-like cells that
develop short lateral processes from the basal process first,
then from the apical process. Of these, the processes
Fig. 5. Real-time imaging of amacrine cell migration: (a) Migrating amacrine cells show undirected neurite outgrowth. Time-lapse images of a YFP+
amacrine cell (arrow) migrating towards the GCL. Multiple neurites emerge from the cell, but they do not appear to be polarized towards the GCL. The
dashed line represents the position of the future IPL, at the interface between the forming INL and GCL. (b) Amacrine cells display an early bias for the
IPL sublamina in which they will ultimately stratify. Confocal time-lapse images of a pair of GFP+ amacrine cells (green). The red indicates the IPL.
From the outset, either cell restricts dendritic ramification to either the basal or the apical sublamina of the IPL. Reproduced with permission from
Godinho et al. (2005).
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
265
Fig. 6. Early appearance of retinal mosaics: (a) A cone mosaic in the embryonic monkey retina. Reproduced from Wikler et al. (1997), with permission.
(b–c) A regular array of cholinergic amacrine cells (Islet-1+) in the embryonic rat retina. The mosaic is viewed in a flat mount (b) and in a radial section
(c). Note that at this stage only three layers (NBL, IPL, and GCL) can told apart.
growing in the IPL and OPL become selectively stabilized
and develop into axon and dendrites, respectively, while
the main apical and basal processes of the cells are
retracted (Morgan et al., 2006).
Very schematically, the development of the retina now
requires only a few additional steps: the maturation of the
photoreceptors, which form their ‘‘light detecting units’’,
the inner and outer segments; the lateral migration and
compaction of central photoreceptors to form the fovea in
primates; and the maturation and refinement of synaptic
connections between the neurons. These processes are
treated in details elsewhere (Usukura and Shuichi, 1996;
Curcio et al., 1990; Wong, 1999).
‘‘construction work’’ that goes on simultaneously within
the different layers.
Some retinal mosaics are among the earliest detectable
mature-like ‘‘patterned units’’ assembled during this
building process. For this reason, we will start exploring
the logic of retinal construction by reviewing what we
currently know about mosaic formation. We will first
consider which spatial (geometrical) rules underlie the
regular spacing of cells within mosaics, then analyse studies
addressing the potential contribution of cell genesis, death
and migration to mosaic formation, and end up with a
speculation on the nature of the cell interactions that
underlay the assembly of cell mosaics.
3.2. The spatial rule that we call mosaic regularity
3. The development of retinal mosaics
3.1. Early retinal mosaics and why we care about them
As we have seen, the retina is initially a tissue made by
apico-basal (AB)-oriented elements that gradually converts
into a structure characterized by cellular patterns—the
layers and the cell mosaics—which are orthogonal to the
AB axis. This patterning is progressively laid down by a
Retinal mosaics gained their name from the almost
crystalline regularity of some cone arrays. Other cell types
do not tile the retina as regularly, yet they still appear
orderly when selectively stained in retinal whole-mounts.
Since the seminal study by Wässle and Riemann (1978), a
number of ways have been devised to explore the geometry
of retinal mosaics. Essentially, they all show that mosaics
are non-random cell distributions. Furthermore, modelling
ARTICLE IN PRESS
266
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
and statistical analysis have shown that, in most cases, the
geometrical rule underlying the non-random spacing of
like-cells within their layer is a ‘‘minimal spacing rule’’ by
which cells of the same type avoid being closer to one
another than a given distance. This minimal distance is
usually comparable to several cell body diameters,
and varies with cell type (reviewed in Cook and Chalupa,
2000; Galli-Resta, 2002). The model of mosaic formation
based on the implementation of a simple spacing rule
has generated many interesting experiments, as it suggests
that, in taking position within their layer, cells only care
for similar (homotypic) cells, disregarding cells of other
types.
3.3. Predominance of homotypic-cell interactions in mosaic
formation
A number of studies support the view that each cell type
forming a mosaic does so mostly independently of cells of
different types. In the adult retina mosaics formed by
different cells types are not spatially correlated (Rockhill
et al., 2000). Furthermore, experiments where the populations of pre- or postsynaptic partners are either ablated or
genetically incremented do not affect mosaic formation in
the case of the horizontal cells (Raven et al., 2007; Raven
and Reese, 2003; Reese et al., 2005) or the cholinergic
amacrine cells (Galli-Resta, 2000).
3.4. Potential players contributing to cell spacing within
mosaics
As it is commonly observed in biology, there seems to
be no single mechanism controlling cell spacing within
retinal mosaics, but rather multiple processes appear to
contribute to its development. In particular, investigations
have been focused on cell death, cell fate control, and cell
displacement.
3.4.1. Cell death in mosaic formation
Cell death has been proposed to contribute to the
regularity of retinal mosaics on the basis of the following
observations: (1) The mosaics of developing b-RGCs in the
cat increase their regularity after undergoing a 20% cell
loss (Jeyarasasingam et al., 1998). (2) In the bcl-2 overexpressing mice, some cell populations are increased as a
result of the death sparing action of the transgene; in
particular, the dopaminergic amacrine cells are about 10
times as numerous as in normal wild-type animals and
display a more disorganized spatial arrangement than
normal (Raven et al., 2003). (3) The minimal spacing
between neighbouring cholinergic cells is decreased when
cell death in this population is prevented by degradation of
extracellular ATP or by blockade of the purinergic P2X
receptor for ATP in the neonatal rat (Resta et al., 2005). (4)
Finally, time-lapse analysis of the development of green
fluorescent protein (GFP) labelled amacrine cells in the
zebrafish, shows the frequent occurrence of cell death, with
a higher likelihood to die for amacrine cells that come close
to one another, in line with the view that death might
contribute to like-cell spacing during development (reviewed in Poggi et al., 2005b).
While these experiments strongly suggest a role for death
in mosaic formation, it is important to consider that cell
elimination can improve the regularity of a distribution of
cells only by removing wrongly positioned cells and not by
correctly positioning cells. Thus, death should not be
expected to create a regular distribution, unless it is
complemented by ‘‘positive mechanisms’’, such as continual cell addition and/or cell displacement. Two ‘‘positive’’ processes have been proposed to contribute to the
spacing of mosaic cells, spatial order at the time of cell fate
determination, and lateral cell migration.
3.4.2. Does cell fate determination contribute to mosaic
regularity?
There are contrasting views as to the role of cell fate
mechanisms in determining minimal cell spacing among
like cells in the mature vertebrate retina. Newly born chick
RGCs have a non-random spatial distribution when still in
the apical region of the retina, suggesting that their genesis
undergoes spatial constraints (McCabe et al., 1999). Yet, at
least in the mammalian retina, most cells forming mosaics
migrate to their final layer moving both along (radial) and
perpendicular (tangential) to the AB axis, so that their final
position is not related to the position in which they were
originally born (Reese et al., 1995).
3.4.3. Tangential migration in mosaic assembly
Heterozygous female mice expressing the transgene LacZ
linked to the chromosome X have been used to analyse the
spatial relation between retinal neurons and their clone of
origin. In these animals, the random X inactivation, which
occurs before neurogenesis in the retina, creates a retina
from an initial random pattern of LacZ+ (blue) and LacZ(white) progenitors. If cells in a clone remain together, and
only take different AB positions, these clones should be
columns arranged along the AB axis. If some cell moves
out of the clone, it will stand out as a blue cell in a white
column or vice versa. Muller glia, rod bipolar cells
and rods, all appear to just migrate radially (i.e.
perpendicular to the retinal layers) from the location
of their original progenitor. On the contrary, mosaicforming cells (i.e. cones, horizontal, amacrine, and ganglion cells) are displaced tangentially (i.e. parallel to the
retinal layers) away from their clone of origin. The average
tangential cell displacement is several tens of microns, and
within each mosaic forming population, all the cells
move with respect to their clone of origin (Reese and
Galli-Resta, 2002; Reese et al., 1995, 1999b). Real-time
imaging of cell migration in the zebrafish retina has directly
shown tangential cell movements for both horizontal and
amacrine cells (Godinho et al., 2007; Poggi et al., 2005b).
But which mechanisms drive cell displacement away from
like-cells?
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
3.5. Hypotheses on the nature of homotypic interactions
controlling mosaic formation: a role for dendrites?
The finding that in most retinal mosaics neurons avoid
being closer to one another than a minimal set distance
(reviewed in Cook and Chalupa, 2000; Galli-Resta, 2002),
combined with the observation of the almost perfect
dendritic tiling that RGC mosaics achieve with minimal
dendritic superimposition (Wässle et al., 1983), have
suggested the hypothesis that dendrites might be instrumental to cell spacing in retinal mosaics. The following
evidence supports this view: (1) Modelling experiments
have shown that minimizing dendritic superimposition
would suffice to create an orderly cell array (Eglen et al.,
2000). (2) In retinas with supernumerary RGCs, the size of
the RGC dendritic tree scales in accordance to the
increased cell density (Kirby and Chalupa, 1986). Similarly,
the dendritic trees of horizontal cells scale with density
when different mouse strains are compared (Reese et al.,
267
2005). (3) If all RGCs are killed in a retinal sector, the
dendrites of nearby RGCs grow into the lesion, suggesting
that interactions between neighbouring RGC dendrites
normally control the extent of RGC dendritic trees (Eysel
et al., 1985; Perry and Linden, 1982). A similar mechanism,
mediated by the cadherin Flamingo, controls dendritic tree
development in subsets of Drosophila neurons tiling the fly
body (reviewed in Parrish et al., 2007). (4) The cholinergic
and amacrine cell mosaics become regular when and
where their cells are connected by a continuous dendritic
net (Fig. 7). If dendritic stability is inhibited in these
mosaics during development, mosaic disorganization occurs (Fig. 8, Galli-Resta et al., 2002).
Dendritic interactions, however, are unlikely to be the
only mechanism mediating cell spacing in retinal mosaic,
since recent experiments show that a minimal cell spacing
is observed in populations of ganglion and amacrine
cells that have been so massively depleted in the neonate
as to make it unlikely that any dendritic interaction might
Fig. 7. Regular cell spacing in the mosaics of cholinergic amacrine cells and horizontal cells appears when and where a continuous net of dendrites link
neighbouring cells: (a) When an incomplete net of processes is observed in the mosaic of cholinergic amacrine cells in the GCL of the rat retina, irregular
cell spacing is observed close to the gaps in the net of processes (e.g. cell clusters indicated by arrows). (b) At birth, the cholinergic cells of the GCL are
regularly spaced. A continuous net of dendrites links neighbouring elements. (c) On postnatal day 4 (P4), irregular cell spacing and gaps in the net of
horizontal cell processes are observed in the peripheral rat retina. (d) At the same age, the horizontal cells are regularly spaced in the central retina, and
linked by a continuous net of processes. Confocal images. Scale bar 10 mm. Reproduced with permission from Galli-Resta et al. (2002).
ARTICLE IN PRESS
268
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
Fig. 8. Perturbing cell dendrites disrupts retinal mosaics: (a) At birth, the cholinergic cells in the GCL are already regularly spaced in the rat retina. (b–d)
Perturbing the cell dendrites (by using antisense oligonucleotides to microtubule-associated proteins (MAPs)) disrupts the regular spacing of these cells
(b), and scatters them to different retinal depths [(c) before, and (d) during treatment]. Scale bar 50 mm (a, b), 10 mm (c, d).
have occurred between like-cells (Farajian et al., 2004; Lin
et al., 2004).
3.6. Mosaics as building blocks of retinal assembly?
Independently of the many open questions on mosaic
formation, even a very conservative investigator would
agree on a few points: (a) most retinal cell types form
mosaics; (b) some mosaics form early in development;
(c) most mosaics form independently of one another. This
allows some speculations. When arrayed in mosaics, likecells are each warranted an individual limited domain
within the retina. If each cell explores a domain centred
around its own soma, as many recent experiments suggest
that growing dendrites do (reviewed in Lohmann et al.,
2002; Wong and Ghosh, 2002), then each neuron can
search for synaptic partners only within a limited region.
This search will undergo limited competitive interactions
between like neurons, each of which will be searching in a
region centred on its own domain. Furthermore, if the
synaptic partners are also arranged in a mosaic, then
topographical order in the connections is ensured, as
neighbouring cells in the first array will be connected
preferentially to cells that are neighbours in the second
array. In this scheme, the capability of cells to recognize
one another and two sets of local mechanisms—the
minimal spacing rule between like neurons and the local
search for synaptic partners—would suffice to lay down a
blue-print of retinal architecture (Galli-Resta, 2001, 2002).
3.7. Exploring the link between mosaics and layer formation
Speaking of mosaics as the two-dimensional arrays that
can be seen in retinal flat mounts, one tends to forget the
basic question of how mosaic cells become confined within a
single cell monolayer. This question is still unanswered, but
there are interesting indications. The horizontal cells are
initially moving across the retinal thickness (Edqvist and
Hallbook, 2004), and later converge to a single retinal layer,
acquiring extra-regularity as they do so (Novelli et al., 2007;
Raven et al., 2005). Similarly, the early cholinergic amacrine
cell mosaics occasionally display cells out of the mosaic
layers, before acquiring their adult-like regularity (GalliResta et al., 2002, 1997). Furthermore, if the dendrites of
these cells are temporarily perturbed (Fig. 8), this disrupts
both the regular spacing of the cells and their alignment in a
single cell monolayer (Galli-Resta et al., 2002). These data
indicate a strong correlation between mosaic and layer
formation, but before discussing this point further, we need
to consider what is currently known about layer formation.
3.8. Mosaic development in short form
Mosaics are orderly arrays of like-cells that occupy a
single cell monolayer. Commonly, their geometrical
regularity derives from the presence of a minimal spacing
between their cells. The first amacrine cells and cone
mosaics appear before their layers can be told apart
(Fig. 6). The formation of mosaics seems predominantly
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
due to interactions that are restricted to the mosaic cells, as
they appear indifferent to perturbations affecting other cell
types. Several processes collaborate in controlling mosaic
formation, including cell death, and lateral tangential
migration. A number of studies show a strong relationship
between dendritic development and cell spacing in retinal
mosaics. Some mosaics become regular where and when
their cells are linked by a continuous net of dendrites
(Fig. 7), and perturbing this net disrupts regular cell
spacing within the mosaic, as well as its alignment in a
monolayer (Fig. 8). These observations suggest a link
between mosaic and layer formation.
4. Investigating the mechanisms controlling cell layering
The arrangement of cells in layers is the first feature
of the retina described in textbooks. From the point of
view of building strategies that each cell type has a layer
to migrate to, as a part of its differentiation program
seems very efficient. If we consider the building of the
retina as an incessant sorting process taking place while
many events occur simultaneously (cell migration, cell
divisions, cell birth, cell death, differentiation, process
outgrowth, etc.), a strategy of ‘‘bucket sorting’’ seems
advantageous. As it is generated, each element goes to its
‘‘bucket’’/layer, where it is spatially close to the elements to
which it will eventually connect. In many severe retinal
pathologies, retinal layering is disrupted or altered, which
is the most obvious indication of the importance of this
organization. But how are layers built in the making of the
retina?
During the last few years, a considerable number of
studies have shown the effect of genetic or other experimental manipulations on the retinal organization in layers,
or, as some refer to it, its lamination. As we will see, an
emerging aspect is that neurogenesis and lamination can be
viewed as separate processes, regulated by different control
mechanisms. For this reason, here we will purposefully
ignore experiments affecting the generation of specific cell
types, although they have obvious consequences on
lamination, and focus on manipulations that affect
lamination without apparent effects on the genesis of the
different retinal neurons. Neurogenesis and cell fate control
in the retina have been the subject of recent reviews to
which the reader is referred (Cayouette et al., 2006; Livesey
and Cepko, 2001; Rapaport, 2006).
Very schematically, the manipulations leading to
alterations affecting the normal development of retinal
layers identified so far can be subdivided in three major
categories: (a) alterations targeting classic adhesion molecules or the basal lamina, (b) alterations affecting the
proteins of the Crumbs (crb)- or the Apical Polarity
complex, and (c) alterations removing diffusible signalling
factors. Following is a detailed illustration of these effects.
The reader wishing to skip this presentation can directly go
to the summary at the end of this Section 4.4.
269
4.1. Role of classical adhesion molecules and basal lamina
4.1.1. Basal lamina
Since the stage of proliferating neuroepithelium, the
retina is delimited on its vitread side by a basal lamina, the
ILM, a ‘‘fabric’’ made by laminin and collagen fibres and
other ECM molecules. A number of studies illustrate
alterations in the inner retinal architecture due to mutations or manipulations affecting the ILM. Several hereditary diseases in humans, such as Fukayama muscular
dystrophy, muscular-eye-brain diseases and Walker–
Warburg syndrome show breaches in the ILM and
associated retinal ectopias—i.e. clumps of neurons exiting
the inner retina boundary toward the vitreous (Haltia et al.,
1997; Nakano et al., 1996; Sertie et al., 2000; Williams
et al., 1984). Random ruptures of the ILM and associated
retinal ectopias (Fig. 9) are also observed in mice KO for
components of the ILM (e.g. nidogen (Candiello et al.,
2007), or laminin g-1 (Halfter et al., 2005)). These mice
often display extensive retinal haemorrhages due to retinal
vascular leakage, but ectopias seem likely to be due to ILM
ruptures, as the two are spatially associated. The causal
link between inner retinal disorganization and ILM loss is
further supported by experiments in chick, where the lack
of retinal vasculature allows using enzymatic digestion to
target the ILM without vascular complications. Intraocular injection of collagenase during early retinal neurogenesis in chick embryos completely digest the ILM in 24 h and
leads to an altered stratification of RGC somata within a
few days (Halfter and Schurer, 1998).
4.1.2. Receptors for extracellular matrix components:
integrins and distroglycan
Integrins are the major family of cell adhesive receptors
binding components of the ECM. Their binding to ligand
leads to the formation of focal adhesions that link the
ECM to the intracellular cytoskeleton (reviewed in Hynes,
2002).
In the developing retina, integrins are found throughout
the retinal layers (Cann et al., 1996). Mouse embryos
lacking integrin a6 (Georges-Labouesse et al., 1998) or
both a3 and a6 (De Arcangelis et al., 1999) display retinal
abnormalities, with ectopic neuroblastic outgrowths in the
vitreous body in the eye and altered ILM deposition. In the
embryonic chick retina blocking integrin b1 prior to
neurogenesis dramatically reduces retinal growth and
disrupts retinal neuroepithelial integrity (Svennevik and
Linser, 1993). Similarly, blocking integrin b1 in Xenopus
embryos causes anomalous retinal lamination, with abnormal cell clusters in the outer retina, commonly known as
rosettes (Li and Sakaguchi, 2004). Rosettes (Fig. 10) are
sphere-like structures where the relative stratification of
neurons is locally correct, but it is organized inside-out,
with the apical segments of the photoreceptors pointing
inward.
In addition to integrins, another transmembrane receptor linking the ECM to the cytoskeleton is dystroglycan
ARTICLE IN PRESS
270
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
Fig. 9. RGC ectopias associate to ILM breaches: (a, b) Sections through the eye of a control (a) and a laminin-g1 KO mouse (b) at 16.5 days of gestation.
Note the haemorrhages in (b–d). Radial sections of the normal (c) and laminin-g1 KO retina (d). The control retina has a smooth vitreal (basal) surface
(black arrow) with multiple hyaloid arteries (white arrows), whereas cellular ectopias (E) are seen along the vitreal surface of the mutant retina (d–f).
Electron microscopic images of the basal surface of the retina of the normal (e) and laminin-g1 KO retina (f). A continuous, uninterrupted ILM (arrow) is
seen in the control retina (e), while mutant mice have a discontinuous ILM (f). R ¼ retina, L ¼ lens, ON ¼ optic nerve. Bar: (a, b) 100 mm; (c, d) 50 mm;
(e, f) 50 nm. Reproduced with permission from Halfter et al. (2005).
(DG). DG is expressed in the developing retina, and is
specifically localized to the neuroepithelial endfeet at the
vitreal border (Blank et al., 1997). Antagonizing DG
function in the embryonic chick retina disrupts the AB
orientation of neuroepithelial cells, and causes hyperproliferation, while opposing effects are obtained by overexpressing DG (Schroder et al., 2007). In Xenopus, loss of
DG function leads to ocular malformations including
microphtalmia, early disorganization of the basal lamina,
and clustering of rosette-like structures (Fig. 10) in the
mature animal (Lunardi et al., 2006).
4.1.3. Cadherins and their intracellular ligands
Cadherins form a superfamily of transmembrane
molecules, many members of which appear to act as
Ca2+-dependent adhesion molecules (reviewed in
Takeichi, 1995). Several cadherins are expressed in the
vertebrate eye, and some (N-cadherin, R-cadherin, cadherin-11, cadherin-8, and cadherin-6) dynamically change
their expression during retinal development (Honjo et al.,
2000).
N-cadherin is a major component of the apical belt of
AJs joining the neuroepithelial cells of the developing
retina. In the presence of functional blocking antibodies
against N-cadherin, explanted chick retinas of early
embryos tend to dissociate and cannot be maintained as
a tissue mass (Fig. 11a and b). Older retina do not
disaggregate but develop abnormal arrangements of cells in
the retina, leading to the appearance of rosette-like cell
clusters (Matsunaga et al., 1988). Similarly, severe alterations of retinal architecture have been observed in different
Zebrafish mutations affecting the coding region of the
N-cadherin gene (Erdmann et al., 2003; Malicki et al.,
2003; Masai et al., 2003; Pujic and Malicki, 2001). In these
mutants, the AB polarity of the proliferating retinal
neuroepithelium is dramatically compromised (Fig. 11d
and e) in the embryo, and rosette-like patterns form in the
mature retina (Fig. 11f and g).
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
271
4.2. The crb and Apical Polarity complexes
Fig. 10. Loss of retinal lamination and formation of rosette-like patterns
after loss of the ECM receptor DG: (a, b) Sections of a normal Xenopus
embryonic retina, stained for a photoreceptor specific nuclear factor (grey
in (a)) and with Hoechst nuclear staining (blue in (b)). (c, d). The normal
lamination of the retina is lost after treatment with morpholino for the
ECM receptor DG. Note the rosette-like patterns in the outer retina
(arrows in (c); staining as in (a)) and the abnormal layering evidenced by
the Hoechst nuclear staining. Reproduced with permission from Lunardi
et al. (2006).
A highly conserved molecular machinery positions AJs
at the interface between apical and basolateral plasma
membrane domains in epithelia. Major components of
such machinery are the crb and Apical protein complexes
(reviewed in Richard et al., 2006; Tepass and Harris, 2007).
Not surprisingly, mutations affecting either complex affect
the proper development of retinal layers.
In Drosophila, crb is necessary for the cohesion of the
epithelia, and for proper eye development (Tepass et al.,
1990). Mouse mutants in crb1 (one of the two crb
homologues), show a fragmented outer limiting membrane
(OLM, the mature form of the AJs belts), photoreceptor
degeneration, rosettes formation and retinal folding
(Mehalow et al., 2003; van de Pavert et al., 2007). In
humans, mutations in the crb1 gene are observed in 4% of
patients affected by retinitis pigmentosa and in 10–15% of
Leber congenital amaurosis (Richard et al., 2006). Interestingly, analysis of the retina in a patient carrying a crb1
mutation showed abnormal layering (Jacobson et al.,
2003). Mutations or pharmacological manipulations affecting elements of the crb complex in Zebrafish cause
disruption of the AB polarity in the retinal neuroepithelium, and severe alteration of retinal lamination, rosette
formation and abnormal photoreceptors in the mature
animal (Hsu et al., 2006; Malicki and Driever, 1999; Omori
and Malicki, 2006; Wei and Malicki, 2002; Wei et al.,
2006).
The Apical Polarity complex (comprising Par3, Par6 and
atypical PCK), appears to physically interact with the crb
complex in vertebrates. All known mouse and zebrafish
mutations affecting proteins of the Apical Polarity complex, display retinal delamination and rosettes (Koike
et al., 2005; Malicki et al., 1996; Richard et al., 2006; Wei
et al., 2004).
4.3. Diffusible factors and the role of RPE: retinal
re-aggregates and transplants.
Knock down of R-cadherin in Zebrafish leads to a
spectrum of effects, where the most severely affected
animals have small eyes, increased cell death and lack of
retinal lamination (Babb et al., 2005). Silencing the
Xcadherin-6 gene in Xenopus, leads to abnormal eyes
showing lamination defects with rosettes on the outer
retina, a reduced GCL and a fragmented RPE (Ruan et al.,
2006).
In the AJs, cadherin molecules are linked to the actin
cytoskeletal network. An essential component of this link is
b-catenin, which is abundantly expressed in the developing
retina (Liu et al., 2002; Mu et al., 2001). Conditional KO
mice that silence b-catenin expression in the retina just
before neurogenesis display a normal proliferation rate, but
retinal progenitor cells loose their normal AB arrangement.
During subsequent development, these retinas exhibit a
higher than normal rate of cell death and display rosettelike cell aggregates (Fu et al., 2006).
The mechanisms controlling the development of retinal
architecture have also been investigated by disaggregating
embryonic retinal cells and allowing their re-aggregation
and development in tri-dimensional cultures. Retinal cells
re-aggregate in spheroids, where mitotic divisions are
localized in rosette-like patterns. These give rise to all
different retinal cell types that locally stratify according to
the normal order observed in the retina. The differentiated
re-aggregates however do not present the continuous
laminar retina-like organization, because often photoreceptors are found pointing inside-out, in rosette-like
patterns (Fig. 12a and b). A continuous correct lamination
is observed when the re-aggregates are allowed to grow in
media conditioned by Muller glia or RPE cells (reviewed in
Layer et al., 2002), or (Fig. 12c and d) when cultured
together with explants of the embryonic retinal anterior
rim (Nakagawa et al., 2003). These experiments indicate
ARTICLE IN PRESS
272
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
Fig. 11. Effect of N-cadherin perturbation on retinal cohesion and lamination: (a, b) Explanted chick embryonic retinas 24 h after incubation in control
medium (a) or in a medium containing N-cadherin antibodies (b). The latter treatment totally disaggregates the retina (b), and the cells can be dispersed by
gentle agitation. Reproduced with permission from Matsunaga et al. (1988). (c, d) The retinal neuroepithelium in a normal Zebrafish embryo (c) and in an
N-cadherin mutant (d) where AB polarity is lost. (e, f) The mature retina in a normal Zebrafish embryo (e) and in an N-cadherin mutant (f), which displays
a sever loss of retinal lamination. The blue staining labels photoreceptors. le ¼ lens. Reproduced with permission from Malicki et al. (2003).
the need of diffusible factors to orchestrate global retinal
organization. The nature of these factors is still unknown
(but see Nakagawa et al., 2003). The importance of the
RPE in controlling the development of retinal architecture
has been confirmed in vivo, by selective genetic ablation of
the RPE cells. Loss of RPE during retinogenesis is
associated to microphthalmia, disorganization of the
retinal architecture, and the appearance of rosettes.
Interestingly, where some patches of RPE escaped ablation, the correct laminated structure of the retina was
maintained adjacent to the RPE cells (Raymond and
Jackson, 1995).
Interestingly, rosette-like patterns are also commonly
observed when embryonic retinas are transplanted ectopically in the brain. These transplants produce all the normal
retinal cell types, are usually smaller than normal retinas
and connect to the retino-recipient visual centres of the
host brain (reviewed in Lund et al., 1988). Behavioural and
electrophysiological studies show that transplanted retinas
transmit light information to the host brain (Coffey et al.,
1989; Klassen and Lund, 1987) albeit lacking topographical mapping in their connections (Galli et al., 1988). This
agrees with the view suggested by their anatomical
organization, i.e. that the circuitry is locally correctly
wired, but it is not organized in a global map.
4.4. A summary of layer development: common end points
for common mechanisms?
Reviewing the effects of mutations and manipulations
affecting layer development, some common themes emerge.
When the basal side of the retina is affected (experiments
on the ILM and some integrins), the main effects are on the
inner retina, with local cell delamination and ectopias, as if
the retina had lost its inner (basal) border (Figs. 9 and 14).
Note however, that some experiments affecting integrins
and DG have more severe effects, interesting the entire
retina, as does the expression pattern of these proteins.
When the retinal apical side is affected, and in particular,
the AJs, dramatic effects may ensue, from total retinal deaggregation (loss of N-cadherin in early retinas; Fig. 11a
and b), to the lack of any retinal lamination (R-cadherin
mutants), or the loss of neuroepithelium AB polarity, and
subsequent retinal delamination (Fig. 11c–f; N-cadherin
Zebrafish mutants, Zebrafish mutants in crb-complex
components, mouse and Zebrafish mutants in Apical
Polarity complex components, b-catenin conditional KO
mouse, RPE ablation). Not all the mutations affecting the
apical side are so severe. Some display their effects mostly
in the phase of photoreceptor maturation (e.g. crb1 mice
and human mutants), but this might depend on the partial
redundancy of the molecules involved (Richard et al.,
2006).
By far, the most common abnormality in retinal
laminations is the appearance of rosette-like cell clusters,
which can either affect only cells of the outer (apical) retina
or all retinal layers. Rosettes, which are commonly
observed in a number of retinal pathologies, are characterized by the presence of a locally correct cell stratification,
which is not continuous throughout the tissue, as photoreceptors point inside-out (Figs. 10 and 12). In retinal reaggregates this defect is prevented by soluble factors
secreted by the RPE (Fig. 12b). Conversely, retinas lacking
the RPE show rosettes.
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
273
Fig. 12. Retinal re-aggregates develop rosette-like patterns that are prevented by RPE derived factors: (a, b) Retinal re-aggregates normally develop
rosette-like patterns (a, nuclear staining; b, immunostaining for the major cell types, green ¼ photoreceptors, red ¼ amacrine (Am) and ganglion cells
(GC), blue ¼ bipolar cells). (c, d). A correct lamination of the retina is obtained in the presence of media conditioned with RPE cells or of the anterior
retinal rim (c, d). The inset shows a higher magnification of the outer region of the retina. Scale bars, 20 mm. Reproduced with permission from Nakagawa
et al. (2003).
A very schematic hypothesis to order these effects could
be as follows: AB polarity in the neuroepithelium seems
essential for the development of the appropriate continuous retinal stratification in layers. Normally, the RPE acts
as a ‘‘polar star’’, producing factors that allow positioning
and alignment of the apical sides of the neuropithelial cells.
The main feature of the apical side appears to be the
presence of the AJs. Manipulations affecting AJ components (N-cadherin, b-catenin, etc.), or the complexes
controlling their formation (crb and Apical complexes),
compromise layer formation. Finally, a normal retinal
layering also requires that an appropriate basal lamina is
deposited at the basal side, and that cells have the right
type and amount of integrins to connect to it.
Independently of the validity of this schematic view,
there is a clear predominance of adhesion molecules and
their binding partners among the molecules so far known
to affect retinal patterning. Thus, it seems reasonable that,
alongside biochemical and genetic investigations, one
important avenue to explore is the role of adhesion and,
more generally, of mechanical interactions in shaping the
retinal tissue. This exploration is just beginning. The
following section outlines some of its important and
unexpected aspects.
5. An emerging role for mechanical forces in retinal
development?
Retinal pathologists and ophthalmic surgeons experience
the impact of mechanical forces on the retina on a daily
basis, from retinal detachment, to excavation of the optic
nerve head in glaucoma, or retinal tears caused by vitreoretinal traction, to mention a few. Yet, mechanical forces
are largely neglected in the study of normal retina
development.
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
274
7
cannula
control
glass rod
Eye diameter (nm)
6
5
4
3
2
1
5
6
7
8
age (days)
Fig. 13. Intraocular pressure (IOP) and ILM integrity participate in eye growth control: (a) Reducing IOP during development almost abolishes eye
growth in the chick embryo. Redrawn with permission from Coulombre (1956). (b) Enzymatic disruption of the ILM in the chick embryo leads to
increased eye size within a few days: compare in (b) the enzyme-injected (right) and the non-injected contralateral control eye (left). Four days after
treatment the eye size was increased almost 1.5 folds. Scale bar 1 cm. Reproduced with permission from Halfter et al. (2006).
5.1. Mechanical components in the control of eye size
A role for mechanical forces in eye development has
been suggested by classical embryology. In the 1950s,
Coulombre showed that intraocular pressure (IOP) promotes eye growth. Inserting a hollow cannula into the eye,
which allowed vitreous efflux and thus reduced IOP, he was
able to almost abolish the 6-fold increase observed in the
eye of the chick embryo between embryonic day 4 (E4) and
E8 (Fig. 13a). The same effect was not obtained using a
solid rod of the same size as the cannula (Coulombre,
1956). The force exerted by IOP is counterbalanced by the
sclera (reviewed in McBrien and Gentle, 2003), and, at least
during embryonic development, also by the ILM. Enzymatic digestion of the ILM with collagenase injection in the
E4 chick embryo causes a 50% increase in eye size within 4
days (Fig. 13b), an increase which is almost completely
prevented by promoting ILM reconstitution with a chasing
injection of laminin (Halfter et al., 2006). These experiments, and the growing field of studies on the mechanisms
controlling eye growth in animal models of myopia,
indicate an important role for mechanical forces in global
eye growth. But what about the retinal tissue and its
organization during development?
5.2. Naı¨ve considerations on retinal cohesion during
development
To explore whether mechanical forces have a role in
retinal development, we will start with some naive
considerations. Electron microscopy studies show that the
retinal neuroepithelial cells are anchored basally to an
ECM lamina (the ILM), while apically, they are joined to
one another by belts of AJs. These connections are
maintained throughout the cell cycle, so that even during
mitosis dividing cells remain in contact with adjacent cells,
and clones generated by single precursors radiate as
compact distributions of cells from the centre to the
periphery of the retina (Reese et al., 1995). Can we envision
cell adhesion to the ILM and cell–cell adhesion through
AJs as the main forces that hold the retinal neuroepithelium together while it grows?
The most striking evidence in support of this view is the
observation that early functional removal of N-cadherin,
the core protein of AJs, leads to a total de-aggregation
of the retina (Fig. 11a and b Matsunaga et al., 1988).
Subsequently, this or other manipulations affecting
N-cadherin, are not so effective (Fig. 11c–f), as AJs are
not lost, but rather become spatially disorganized. This
disorganization suffices to prevent the global orderly
development of layers and mosaics, and commonly
rosette-like patterns form (Erdmann et al., 2003; Malicki
et al., 2003; Masai et al., 2003; Matsunaga et al., 1988;
Pujic and Malicki, 2001). Intuitively, this is not surprising.
If adhesion between cells is not homogeneous and/or not
adequate to balance the forces that expand the tissues, one
should expect tissue disruption at one extreme, and in
general, tissue irregularities. We could even naively
envision that, if AJs are stronger on one side of a cell than
on the opposite side, the cell might be pulled sideways, and
out of the AB orientation. Once the AB orientation of the
tissue is locally perturbed, this could easily evolve into folds
and rosette-like patterns as cells proliferate. This is of
course a speculation, but it serves to consider how
mechanical unbalance should be expected to generate
tissue irregularities.
There are also striking indications for a mechanical
action of the ILM in maintaining tissue integrity. Mice or
men with defective ILM and/or integrins (the receptors
that cells use to anchor the ILM) show a similar phenotype
(Candiello et al., 2007; De Arcangelis et al., 1999; GeorgesLabouesse et al., 1998; Halfter et al., 2001; Haltia et al.,
1997; Nakano et al., 1996; Sertie et al., 2000; Williams
et al., 1984). In the absence of the ILM, the cells of the
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
275
Fig. 14. ILM degradation disorganizes the GCL and causes the retraction of the basal endfeet of neuroepithelial cells: (a, b). Sections from a normal 5-day
(E5) chick embryo retina (a) and an E5 retina 24 h after digestion of the ILM with collagenase. The ILM is immunolabelled in red (a). Note the
disorganization of the GCL. (c, d). Single neuroepithelial cells (green) in a normal (c) and collagenase treated E5 chick retina. The basal endfeet of
the neuroepithelial cells reach the ILM (red) in normal retinas (c), but withdraw from the basal surface when the ILM is degraded (d). The arrows indicate
the retinal apical margin. Reproduced with permission from Halfter et al. (2006) (a, b), Halfter et al. (2001) (c, d).
inner retina are disorganized (Fig. 14a and b), while in the
case of a discontinuous ILM, they bulge out of the ILM
breaches into the vitreous body (Fig. 9). These behaviours
strongly suggest that normally the ILM mechanically
constraints the retinal neuroepithelium. Interestingly,
hours after ILM digestion in the chick embryo eye, the
neuroepithelial cells retract their basal processes (Fig. 14c
and d), as if, having lost their attachment to the ILM, they
had lost tension and recoiled (Halfter, 1998). To understand whether these are just metaphors or real events, one
needs to explore the mechanical properties of the developing retina.
5.3. The mechanical properties of retinal tissue and cells
Atomic force microscopy probes the mechanical properties of tissues on very small spatial scales, using cantilevers
with nano- to micro-metric tips that, ‘‘tapping’’ on the
sample, allow the determination of its deformation in
response to an applied force. When so tested, the chick
ILM displayed a predominantly elastic behaviour, with an
elastic modulus (a measure of the ‘‘spring like’’ properties
of the material) similar to the values obtained for
cartilagineous tissues (around 5–10 kPa Candiello et al.,
2007). This suggests that the ILM is a flexible membrane
ARTICLE IN PRESS
276
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
Fig. 15. A balance between cell adhesion and inner cellular tension appears to control the patterning of lens cells in the Drosophila eye: (a) The
Drosophila eye, formed by multiple ommatidia. (b, c) The clusters of four lens cells normally found within each ommatidium (b) closely resemble clusters
of soap bubbles (c). Adapted from Janmey and Discher (2004) by permission from Macmillan Publishers Ltd.: Nature, copyright 2004.
that can transmit forces across the tissue, and resist cell
traction, acting as an external scaffold, provided that
neurons are softer than the ILM. Is this the case?
Testing the mechanical properties of acutely isolated
retinal cells by scanning force microscopy, has shown that
both retinal neurons and Muller glia have a dominant
elastic component (i.e. they deform in proportion to the
applied stress and tend to recover their initial shape once
stress is over), and are very soft (elastic modulus
200–600 Pa), indeed among the softest tissue of the body.
Similar properties are observed when macroscopically
testing retinal explants (Lu et al., 2006). Unfortunately,
there are no direct measurements of the mechanical
properties of individual retinal cells during development,
but they are likely to be as soft or softer than adult neurons
since whole retinal measures showed a stiffening of the
retina with age (Reichenbach et al., 1991). In support of
this, culture studies show that typical neuronal elastic
modules are around hundreds of PA, suggesting that this
may be a general feature of neurons (Georges and Janmey,
2005; Georges et al., 2006). Putting these data together we
can look at the cells forming the retinal neuroepithelium as
soft springs anchored on their basal side to a much stiffer
continuous scaffold, the ILM. Furthermore, considering
the total retinal de-aggregation induced by early
N-cadherin perturbation, adhesion at the levels of the
AJs seems even more important than the ILM in holding
the tissue together, at least in the young retina, when the
ILM is thinner and weaker than at later stages (Candiello
et al., 2007). However, the mechanical role of AJs in
keeping the vertebrate retina together during development
still needs to be explored in more details.
The view that AJs provide direct physical coupling
between adjacent cells and endow cells with the capacity to
act as a syncytium is a common developmental theme (see
for example Gumbiner, 1996; Halbleib and Nelson, 2006).
Compelling data come from the Drosophila, where the
dynamic regulation of AJs is an essential feature of
epithelial cohesion and growth. Recent studies show that
epithelia cell patterns might minimize mechanical interactions between the cells (reviewed in Lecuit and Lenne,
2007). Remarkable examples are the clusters of four lens
cells found in each ommatidium of the Drosophila eye. At
maturity, the cluster of lens cells is strikingly similar to a
cluster of soap bubbles, which assumes its configuration by
minimizing surface tension (Fig. 15). This similarity is also
observed in mutants presenting clusters of 3–6 lens
cells. Genetic manipulations and modelling studies suggest
that patterning of these cell clusters depends on a
force equilibrium in which adhesion mediated by E- and
N-cadherin opposes the inner cell tension generated by the
action-myosin cytoskeleton (Hayashi and Carthew, 2004;
Janmey and Discher, 2004; Kafer et al., 2007). It will be
interesting to explore whether similar mechanisms control
the lateral migration and compaction of the cone photoreceptors forming the primate fovea (reviewed in Curcio
et al., 1990), where the cone mosaics assumes an almost
crystalline geometry.
5.4. The beginning of retinal layers: a mechanical
explanation for RGCs basal migration
Thus far, we have been dealing with dynamic mechanical
forces that hold the retina together while it grows. In a
way, the appearance of layers and mosaics occurs as cells
make their controlled escape from this strong AB cohesive
mechanism, since neuronal differentiation in the retina
begins with cells losing contact with either one (e.g. the
RGCs) or both (e.g. amacrine cells) of the AB edges.
Is a mechanical view also interesting to investigate these
processes? On the basis of their real-time observations of
RGC differentiation, Zolessi and co-authors suggest that,
once the axon is formed, the RGC body moves to the basal
surface because that is where the cell remains attached
when the apical side retracts (Zolessi et al., 2006).
Similarly, a century ago Cajal hypothesized that the
growth cone pulls the cell body towards the basal surface
to which it is attached (Ramon y Cajal, 1929). In
agreement with this view, ILM degradation leads to RGCs
scattering away from the ILM (Fig. 14a and b), and
disorganizes the course of their axons (Halfter, 1998;
Halfter et al., 2005). Furthermore, studies in cultures of
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
277
Fig. 16. Sorting of cell populations in tri-dimensional cultures depends on the cell differential adhesive properties: (a, b). Two cell clones, transfected to
express N-cadherin at their surfaces in the ratio of 2.4 (green cells) to 1 (red cells), were mixed in equal proportions and cultured as hanging drops: (a) After
4 h of incubation, the aggregate is a mixture of cells of the two clones. (b) After 24 h of incubation, the cells expressing the lower level of N-cadherin (red)
envelop those expressing higher amounts of N-cadherin (green cells). Reproduced with permission from Foty and Steinberg (2005).
different neuronal types support this view. In an elegant
series of studies, Heidemann and collaborators have
investigated the role of mechanical tension in regulating
axonal growth, using calibrated needles to produce or resist
measurable forces (reviewed in Heidemann et al., 1995).
These studies have shown that, as the growth cone
advances, it pulls on the trailing neurite (Heidemann,
1996; Heidemann and Buxbaum, 1991; Heidemann et al.,
1990; Zheng et al., 1994), and this tension can drag the
soma, like a dog on a leash (Lamoureux et al., 1989). In
accordance with this view, studies in zebrafish mutants
show that in cases of altered AB polarity, RGC axons
reach the closest ECM basal lamina, either the ILM or the
Bruch membrane behind the RPE, and the cell bodies
accordingly follow (Zolessi et al., 2006).
5.5. Cell sorting and the differential adhesion hypothesis
While RGCs may be dragged towards the basal lamina
by their growth cones, nascent amacrine cells seem to
behave rather differently. They migrate without retaining
their anchoring to the apical and basal side, (although they
might move along AB-oriented neuroepithelial cells), and
stop when they get close to the RGCs (Godinho et al.,
2005). How do they know where to stop? It could be
conceived that molecular markers exist that distinguish
different depths along the AB axis. In the mature retina,
the search for markers that could differentiate sub-laminae
within the IPL, has shown that the Sidekicks adhesion
molecules and several known cadherins have a laminated
or regionalized pattern of expression (Yamagata et al.,
2002). A regional code based on differential expression
patterns of adhesion molecules might get even more
diversified when considering the protocadherins (Takeichi,
2007). These differential expression patterns have been
observed only in the mature retina, but are undoubtedly
suggestive of a potential role of differential adhesion (DA)
in positioning different cells at different depth during
retinal development.
Cell sorting based on DA has been hypothesized
following classic embryological experiments where cells
disaggregated from two different tissues (e.g. liver and
retina) initially start as a mixture and gradually become
sorted in a tri-dimensional aggregate where one of the cell
type usually envelops the aggregate formed by the cells of
the second type (Fig. 16, Steinberg, 1996). The ‘‘DA
hypothesis (DAH)’’, explains this behaviour by comparing
cells to liquid molecules, and suggests that tissues formed
by cells endowed with different surface adhesion properties
sort out like immiscible fluids (reviewed in Steinberg, 2007).
In agreement with this view, culture experiments show that
cells of different types always segregate according to the
relative surface tension of their aggregates (Fig. 16).
Moreover, when cell lines are used that only express
cadherins, the surface tension of their aggregates is directly
proportional to the number of cadherin molecules expressed on the cell, thus linking the global property that
dictates the sorting behaviour to the individual adhesive
characteristics of the single cells (Foty and Steinberg,
2005). There is still controversy about the DAH, and it is
possible that a model assuming a liquid-like cell sorting
behaviour will need the complement of the cell elastic
components. Yet, at present, this hypothesis is rather
appealing, and it will be interesting to understand whether
and how much it may explain differential cell positioning in
the developing retina.
5.6. Tension induces process outgrowth and stabilization
The hypothesis that axons elongate in response to the
mechanical tension exerted by the growth cone is more
than a century old, but its first experimental evidence is
much more recent. In 1984, Bray showed that axons of
cultured cells can be elongated by experimentally applied
tension (Bray, 1984). Then, Heidemann and collaborators
performed a systematic quantitative analysis of this
problem in cultured neurons (Chada et al., 1997; Heidemann, 1996; Heidemann and Buxbaum, 1990, 1994;
Heidemann et al., 1995, 1999; Joshi et al., 1985; Lamoureux et al., 1992; Zheng et al., 1991, 1993). They found that
there are two major regimes of tension effects on neurons, a
non-growth, elastic regime, and a fluid-like growth regime.
ARTICLE IN PRESS
278
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
The elastic behaviour (i.e. a deformation from which the
cell recovers as soon as tension is lost) is observed below a
certain threshold. When tension goes beyond threshold, it
initiates process outgrowth if applied on the cell body,
or stimulates outgrowth if exerted on existing processes
(Fig. 17). In all cases, a surprisingly robust linear relationship is observed between the process elongation rate and
the tension applied. These studies have also shown that
neurites are constantly producing an internal tension, and
that this underlies their retraction when anchoring is lost
(reviewed in Heidemann, 1996; Heidemann et al., 1995).
Interestingly, the strength of cadherin mediated adhesion, as determined by single bond rupture measurements,
has an average force per bond around 50 pN (Leckband
1
and Prakasam, 2006), which is about 12 10
of the force
necessary to initiate or stimulate neurite outgrowth in
culture (10–50 mdyn, reviewed in Heidemann et al., 1995).
Thus, it is conceivable that as soon as the developing
dendrites of retinal cells are engaged by a membrane
presenting a few cadherin molecules on its surface, this may
suffice to stabilize them and promote their elongation. In
line with this view, when N-cadherin is selectively ablated
in horizontal cells, their dendrites shrink (Tanabe et al.,
2006), as do the dendrites of RGCs when integrin-mediated
cell adhesion is selectively hampered in these cells (Marrs
et al., 2006).
5.7. The cell ‘‘tactile set point’’
The growth promoting effect of tension cannot explain
the selective stabilization of the cell processes in the
plexiform layers that is observed in vivo (Godinho et al.,
2005; Morgan et al., 2006), unless we go back to the
hypothesis of an adhesion-molecule-code for selective
locations (Yamagata et al., 2002). This hypothesis is
undoubtedly fascinating, but admittedly, a subdivision of
the IPL in terms of adhesion molecule expression is evident
in the mature retina (Yamagata et al., 2002), and not before
the IPL forms. Yet, amacrine cell processes selectively
stabilize in a region that thus becomes the IPL (Godinho
et al., 2005). There is still no explanation for this behaviour,
but mechanical studies of neuronal growth in response to
matrix stiffness provide an attractive hypothesis.
Recent findings show that cell behaviour depends on the
stiffness of the substrate on which the cell grows. This
includes whether the cell lives, dies, proliferates and
whether and how it differentiates. In particular, each cell
type seems to have a preferential interval of substrate
stiffness that optimizes cell differentiation. This ‘‘tactile set
point’’ is close to the stiffness of the tissue the cell belongs
to in the organism (reviewed in Discher et al., 2005).
Neurons prefer soft substrates, and when all other
parameters are alike, the stiffness of the substrate
determines whether neurons or astrocytes survive in
primary cultures obtained by dissociating different brain
regions. When neurons find their optimal, soft substrate
they develop numerous processes (Georges et al., 2006).
Fig. 17. Tension initiates neurite outgrowth. Tension is applied to the
margin of the cell body of a forebrain neuron using a calibrated needle.
The needle on the right is used as a reference. Forty minutes later, a
uniform calibre process has formed, with the calibrated needle attached to
the distal end of the process, where a growth cone appears to have formed.
At 140 min following frame b, the neurite has reached the final length of
75 mm and was manipulated onto the dish surface. A fluorescent image of
the same neuron as in (c), fixed and immuno-labeled for b-tubulin.
Reproduced with permission from Chada et al. (1997).
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
Thus, the selective stabilization of processes in the region
of the future IPL could be explained if this region is softer
than the adjacent GCL and NBL regions, mostly occupied
by cell bodies. Interestingly, the measured mechanical
properties of retinal neurons suggest that this could be
the case, as there are sub-cellular variations in elasticity,
with the cell processes being softer than the cell somata,
possibly for the rigidity of the nucleus (Lu et al., 2006).
It will be interesting to directly test this hypothesis in
the future.
5.8. The strain-stiffening behaviour of cells and biological
polymers
Why is it that we talk of elastic behaviour in one case,
proposing that the RGC cell body is dragged as a whole by
the growth cone, while in a different instance we describe
the fluid-like behaviour of processes stimulated by applied
tension? In other words, why can we envision cells
responding to pulling forces with global movement in one
instance, and with just a local process protrusion in a
different context?
The answer to this question lies in the strain-stiffening
behaviour of cells, which is the property by which cells
become stiffer as they are deformed. Strain-stiffening
characterizes cells as well as biological polymers, such as
actin and neurofilaments gels (reviewed in Heidemann and
Wirtz, 2004; Janmey and McCulloch, 2007; Janmey and
Weitz, 2004). Indeed, strain hardening in cells seems to be
mostly due to this same property displayed by their
cytoskeleton. Schematically, and very improperly, this
property derives from the fact that the filaments in
cytoskeletal networks are not straight, and have many
small bends. When pulled, they initially just straighten out,
with little resistance. Once they are stretched out completely, however, additional pulling will encounter much
more resistance, as it now acts against the intrinsic
elasticity of the filaments. This behaviour can be easily
observed using a folded piece of paper: it is much easier to
pull out the folds than to stretch the paper itself.
Strain hardening is a rather advantageous property, as it
protects cells and tissues from tearing. It also allows the cell
to avoid major changes in response to small local forces,
and to respond globally to a larger force. Thus, responding
to small forces by elongating a local process and to
stronger forces by moving as a whole, appear to be both in
the repertoire of neuronal reaction to forces.
5.9. A mechanical link between mosaics and layers?
To conclude this panorama of mechanical themes, we go
back to the idea of potential links between mosaic and
layer formation. Immunohistochemical observation and
quantitative analysis show mosaics that become regularly
spaced where and when a net of dendrites links neighbouring cells (Fig. 7), suggesting that dendrites might control
cell positioning (Galli-Resta et al., 2002). In light of our
279
present knowledge of neuronal responses to mechanical
forces, we can envision a neuronal net formed by like cells,
where mechanical forces act on neurites and somata,
dragging, pushing, and pulling, until an equilibrium is
achieved that minimizes the forces. If we assume that the
cells forming a mosaic are mechanically homogeneous,
which seems reasonable as they are all of the same type,
then their equilibrium configuration will be a monolayer
(to minimize forces perpendicular to the net) of roughly
equally spaced cells (to minimize forces parallel to the net).
This mechanical model also explains why manipulations
affecting the cell dendrites simultaneously disrupt both the
orderly cell spacing and their mono-layered arrangement
(Fig. 8) (Galli-Resta, 2002; Galli-Resta et al., 2002). It also
provides a simple explanation for the observation that
retinal cells forming mosaics move laterally with respect to
their clone of origin at the same time as they become
aligned in a single monolayer (Reese et al., 1999a).
Whether and for which cells a mechanical explanation
accounts for regular cell spacing within a monolayer will
need to be further explored.
6. Future directions
This review is not meant to under-rate the complexity of
the biochemical and genetic mechanisms controlling retinal
development, but simply to suggest that mechanical stimuli
may be as important as chemical stimuli in tissue
development.
Several major ‘‘mechanical’’ questions are still unanswered, in the retina as elsewhere. These include the nature
of the cell mechano-sensors, the extraordinary mechanism
by which mechanical tension can integrate the complex
biochemistry of process elongation, and many others. Yet,
although still mainly speculative and incomplete, a
mechanical view seems very promising, not only to
investigate retinal development, but also for its potential
application in retinal pathologies. There are too many
venues that can be explored from this viewpoint to mention
them all, so we end with just a few examples.
In the field of retinal prosthesis and transplants, knowledge of the effects of matrix stiffness on cell outgrowth and
differentiation (reviewed in Discher et al., 2005) may help
devise specific substrates that selectively promote neuronal
rather than glial growth. Interestingly, recent experiments
show that transplanted stem cells integrate much better in
retinas lacking the mechanically ‘‘stiff’’ astrocytes, than
in normal retinas (Kinouchi et al., 2003), an observation in
line with the importance of soft substrates to promote
neuronal growth and differentiation.
Another interesting area to explore is the role of cell
adhesion in synaptic remodelling and in the atrophy and/or
abnormal outgrowth of dendrites observed in retinal
pathologies such as retinitis pigmentosa (Marc et al.,
2003), or in normal aging retinas (Liets et al., 2006).
A third area of considerable interest derives from the
growing evidence that master genes, such as Pax6, might
ARTICLE IN PRESS
280
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
exert some of their most dramatic effects on eye development by controlling cell adhesion properties, thus regulating for example, the interface between the optic vesicle and
the lens placode during induction (reviewed in Collinson
et al., 2004).
Biomechanical investigations are experiencing a ‘‘renaissance’’, thanks to the ever-improving technical sophistication of the available experimental tools, and to the growing
collaborations between biologists, physicists and engineers.
Future break-throughs are soon to be expected.
Acknowledgements
We thank Andreas Reichenbach, Jan Wijnholds, and
Enrica Strettoi for useful discussions, Giulio Cappagli for
his invaluable technical help, Marino and Giovanni Resta
for preparing Fig. 2, Steve Heidemann, Adriana Fiorentini,
Andreas Reichebach, Lamberto Maffei, and Silvia Bisti for
reading the manuscript. We apologize to the authors whose
work, for reasons of space, we have not been able to
include. Work in our laboratory is supported by grants
from the Telethon Foundation (GGP06031), ASI (DCMC)
and the Institute fur Paraplegia.
References
Babb, S.G., Kotradi, S.M., Shah, B., Chiappini-Williamson, C., Bell,
L.N., Schmeiser, G., Chen, E., Liu, Q., Marrs, J.A., 2005. Zebrafish
R-cadherin (Cdh4) controls visual system development and differentiation. Dev. Dyn. 233, 930–945.
Baye, L.M., Link, B.A., 2007. Interkinetic nuclear migration and the
selection of neurogenic cell divisions during vertebrate retinogenesis.
J. Neurosci. 27, 10143–10152.
Blank, M., Koulen, P., Kroger, S., 1997. Subcellular concentration of
beta-dystroglycan in photoreceptors and glial cells of the chick retina.
J. Comp. Neurol. 389, 668–678.
Bray, D., 1984. Axonal growth in response to experimentally applied
mechanical tension. Dev. Biol. 102, 379–389.
Bruhn, S.L., Cepko, C.L., 1996. Development of the pattern of
photoreceptors in the chick retina. J. Neurosci. 16, 1430–1439.
Bumsted, K., Jasoni, C., Szel, A., Hendrickson, A., 1997. Spatial and
temporal expression of cone opsins during monkey retinal development. J. Comp. Neurol. 378, 117–134.
Candiello, J., Balasubramani, M., Schreiber, E.M., Cole, G.J., Mayer, U.,
Halfter, W., Lin, H., 2007. Biomechanical properties of native
basement membranes. FEBS J. 274, 2897–2908.
Cann, G.M., Bradshaw, A.D., Gervin, D.B., Hunter, A.W., Clegg, D.O.,
1996. Widespread expression of beta1 integrins in the developing chick
retina: evidence for a role in migration of retinal ganglion cells. Dev.
Biol. 180, 82–96.
Cayouette, M., Poggi, L., Harris, W.A., 2006. Lineage in the vertebrate
retina. Trends Neurosci. 29, 563–570.
Chada, S., Lamoureux, P., Buxbaum, R.E., Heidemann, S.R., 1997.
Cytomechanics of neurite outgrowth from chick brain neurons. J. Cell
Sci. 110 (Pt. 10), 1179–1186.
Coffey, P.J., Lund, R.D., Rawlins, J.N., 1989. Retinal transplantmediated learning in a conditioned suppression task in rats. Proc.
Natl. Acad. Sci. USA 86, 7248–7249.
Collinson, J.M., Hill, R.E., West, J.D., 2004. Analysis of mouse eye
development with chimeras and mosaics. Int. J. Dev. Biol. 48, 793–804.
Cook, J.E., Chalupa, L.M., 2000. Retinal mosaics: new insights into an
old concept. Trends Neurosci. 23, 26–34.
Coulombre, A., 1956. The role of intraocular pressure in the development
of the chick eye. J. Exp. Zool. 133, 211–223.
Curcio, C.A., Sloan, K.R., Kalina, R.E., Hendrickson, A.E., 1990.
Human photoreceptor topography. J. Comp. Neurol. 292, 497–523.
Das, T., Payer, B., Cayouette, M., Harris, W.A., 2003. In vivo time-lapse
imaging of cell divisions during neurogenesis in the developing
zebrafish retina. Neuron 37, 597–609.
De Arcangelis, A., Mark, M., Kreidberg, J., Sorokin, L., GeorgesLabouesse, E., 1999. Synergistic activities of alpha3 and alpha6
integrins are required during apical ectodermal ridge formation and
organogenesis in the mouse. Development 126, 3957–3968.
Discher, D.E., Janmey, P., Wang, Y.L., 2005. Tissue cells feel and respond
to the stiffness of their substrate. Science 310, 1139–1143.
Edqvist, P.H., Hallbook, F., 2004. Newborn horizontal cells migrate bidirectionally across the neuroepithelium during retinal development.
Development 131, 1343–1351.
Eglen, S.J., van Ooyen, A., Willshaw, D.J., 2000. Lateral cell movement
driven by dendritic interactions is sufficient to form retinal mosaics.
Network 11, 103–118.
Erdmann, B., Kirsch, F.P., Rathjen, F.G., More, M.I., 2003. N-cadherin
is essential for retinal lamination in the zebrafish. Dev. Dyn. 226,
570–577.
Eysel, U.T., Peichl, L., Wassle, H., 1985. Dendritic plasticity in the early
postnatal feline retina: quantitative characteristics and sensitive period.
J. Comp. Neurol. 242, 134–145.
Farajian, R., Raven, M.A., Cusato, K., Reese, B.E., 2004. Cellular
positioning and dendritic field size of cholinergic amacrine cells are
impervious to early ablation of neighboring cells in the mouse retina.
Vis. Neurosci. 21, 13–22.
Foty, R.A., Steinberg, M.S., 2005. The differential adhesion hypothesis: a
direct evaluation. Dev. Biol. 278, 255–263.
Fu, X., Sun, H., Klein, W.H., Mu, X., 2006. Beta-catenin is essential for
lamination but not neurogenesis in mouse retinal development. Dev.
Biol. 299, 424–437.
Galli, L., Rao, K., Lund, R.D., 1988. Tranplanted rat retinae do not
project in a topographic fashion on the host tectum. Exp. Brain Res.
223, 1–4.
Galli-Resta, L., 2000. Local, possibly contact-mediated signalling
restricted to homotypic neurons controls the regular spacing of cells
within the cholinergic arrays in the developing rodent retina.
Development 127, 1499–1508.
Galli-Resta, L., 2001. Assembling the vertebrate retina: global patterning
from short-range cellular interactions. Neuroreport 12, A103–A106.
Galli-Resta, L., 2002. Putting neurons in the right places: local
interactions in the genesis of retinal architecture. Trends Neurosci.
25, 638–643.
Galli-Resta, L., Resta, G., Tan, S.-S., Reese, B., 1997. Mosaics of Islet-1
expressing amacrine cells assembled by short range cellular interactions. J. Neurosci. 17, 7831–7838.
Galli-Resta, L., Novelli, E., Viegi, A., 2002. Dynamic microtubuledependent interactions position homotypic neurones in regular
monolayered arrays during retinal development. Development 129,
3803–3814.
Georges-Labouesse, E., Mark, M., Messaddeq, N., Gansmuller, A., 1998.
Essential role of alpha 6 integrins in cortical and retinal lamination.
Curr. Biol. 8, 983–986.
Georges, P.C., Janmey, P.A., 2005. Cell type-specific response to growth
on soft materials. J. Appl. Physiol. 98, 1547–1553.
Georges, P.C., Miller, W.J., Meaney, D.F., Sawyer, E.S., Janmey, P.A.,
2006. Matrices with compliance comparable to that of brain tissue
select neuronal over glial growth in mixed cortical cultures. Biophys. J.
90, 3012–3018.
Godinho, L., Mumm, J.S., Williams, P.R., Schroeter, E.H., Koerber, A.,
Park, S.W., Leach, S.D., Wong, R.O., 2005. Targeting of amacrine cell
neurites to appropriate synaptic laminae in the developing zebrafish
retina. Development 132, 5069–5079.
Godinho, L., Williams, P.R., Claassen, Y., Provost, E., Leach, S.D.,
Kamermans, M., Wong, R.O., 2007. Nonapical symmetric divisions
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
underlie horizontal cell layer formation in the developing retina in
vivo. Neuron 56, 597–603.
Gumbiner, B., 1996. Cell adhesion: the molecular basis of tissue
architecture and morphogenesis. Cell 84, 345–357.
Halbleib, J.M., Nelson, W.J., 2006. Cadherins in development: cell
adhesion, sorting, and tissue morphogenesis. Genes Dev. 20,
3199–3214.
Halfter, W., 1998. Disruption of the retinal basal lamina during early
embryonic development leads to a retraction of vitreal end feet, an
increased number of ganglion cells, and aberrant axonal outgrowth.
J. Comp. Neurol. 397, 89–104.
Halfter, W., Schurer, B., 1998. Disruption of the pial basal lamina during
early avian embryonic development inhibits histogenesis and axonal
pathfinding in the optic tectum. J. Comp. Neurol. 397, 105–117.
Halfter, W., Dong, S., Balasubramani, M., Bier, M.E., 2001. Temporary
disruption of the retinal basal lamina and its effect on retinal
histogenesis. Dev. Biol. 238, 79–96.
Halfter, W., Willem, M., Mayer, U., 2005. Basement membranedependent survival of retinal ganglion cells. Invest. Ophthalmol. Vis.
Sci. 46, 1000–1009.
Halfter, W., Winzen, U., Bishop, P.N., Eller, A., 2006. Regulation of eye
size by the retinal basement membrane and vitreous body. Invest.
Ophthalmol. Vis. Sci. 47, 3586–3594.
Haltia, M., Leivo, I., Somer, H., Pihko, H., Paetau, A., Kivela, T.,
Tarkkanen, A., Tome, F., Engvall, E., Santavuori, P., 1997. Muscleeye-brain disease: a neuropathological study. Ann. Neurol. 41,
173–180.
Hayashi, T., Carthew, R.W., 2004. Surface mechanics mediate pattern
formation in the developing retina. Nature 431, 647–652.
Heidemann, S., 1996. Cytoplasmic mechanisms of axonal and dendritc
growth in neurons. Int. Rev. Cytol. 165, 235–296.
Heidemann, S.R., Buxbaum, R.E., 1990. Tension as a regulator and
integrator of axonal growth. Cell Motil. Cytoskel. 17, 6–10.
Heidemann, S.R., Buxbaum, R.E., 1991. Growth cone motility. Curr.
Opin. Neurobiol. 1, 339–345.
Heidemann, S.R., Buxbaum, R.E., 1994. Mechanical tension as a
regulator of axonal development. Neurotoxicology 15, 95–107.
Heidemann, S.R., Wirtz, D., 2004. Towards a regional approach to cell
mechanics. Trends Cell Biol. 14, 160–166.
Heidemann, S.R., Lamoureux, P., Buxbaum, R.E., 1990. Growth cone
behavior and production of traction force. J. Cell Biol. 111, 1949–1957.
Heidemann, S.R., Lamoureux, P., Buxbaum, R.E., 1995. Cytomechanics
of axonal development. Cell Biochem. Biophys. 27, 135–155.
Heidemann, S.R., Lamoureux, P., Buxbaum, R.E., 1999. Mechanical
stimulation of neurite assembly in cultured neurons. In: Haynes, L.W.
(Ed.), The Neuron in Tissue Culture. IBRO Wiley, New York,
pp. 105–119.
Honjo, M., Tanihara, H., Suzuki, S., Tanaka, T., Honda, Y., Takeichi,
M., 2000. Differential expression of cadherin adhesion receptors in
neural retina of the postnatal mouse. Invest. Ophthalmol. Vis. Sci. 41,
546–551.
Hsu, Y.C., Willoughby, J.J., Christensen, A.K., Jensen, A.M., 2006.
Mosaic Eyes is a novel component of the Crumbs complex and
negatively regulates photoreceptor apical size. Development 133,
4849–4859.
Hynes, R.O., 2002. Integrins: bidirectional, allosteric signaling machines.
Cell 110, 673–687.
Jacobson, S.G., Cideciyan, A.V., Aleman, T.S., Pianta, M.J., Sumaroka,
A., Schwartz, S.B., Smilko, E.E., Milam, A.H., Sheffield, V.C., Stone,
E.M., 2003. Crumbs homolog 1 (CRB1) mutations result in a thick
human retina with abnormal lamination. Hum. Mol. Genet. 12,
1073–1078.
Janmey, P.A., Discher, D.E., 2004. Developmental biology: holding it
together in the eye. Nature 431, 635–636.
Janmey, P.A., McCulloch, C.A., 2007. Cell mechanics: integrating cell
responses to mechanical stimuli. Annu. Rev. Biomed. Eng. 9, 1–34.
Janmey, P.A., Weitz, D.A., 2004. Dealing with mechanics: mechanisms of
force transduction in cells. Trends Biochem. Sci. 29, 364–370.
281
Jeyarasasingam, G., Snider, C.J., Ratto, G.-M., Chalupa, L.M., 1998.
Activity regulated cell death contributes to the formation of ON and
OFF a ganglion cell mosaics. J. Comp. Neurol. 394, 335–343.
Joshi, H.C., Chu, D., Buxbaum, R.E., Heidemann, S.R., 1985. Tension
and compression in the cytoskeleton of PC 12 neurites. J. Cell Biol.
101, 697–705.
Kafer, J., Hayashi, T., Maree, A.F., Carthew, R.W., Graner, F., 2007.
Cell adhesion and cortex contractility determine cell patterning in the
Drosophilaretina. Proc. Natl. Acad. Sci. USA 104, 18549–18554.
Kay, J.N., Roeser, T., Mumm, J.S., Godinho, L., Mrejeru, A., Wong,
R.O., Baier, H., 2004. Transient requirement for ganglion cells during
assembly of retinal synaptic layers. Development 131, 1331–1342.
Kinouchi, R., Takeda, M., Yang, L., Wilhelmsson, U., Lundkvist, A.,
Pekny, M., Chen, D.F., 2003. Robust neural integration from retinal
transplants in mice deficient in GFAP and vimentin. Nat. Neurosci. 6,
863–868.
Kirby, M.A., Chalupa, L.M., 1986. Retinal crowding alters the
morphology of alpha ganglion cells. J. Comp. Neurol. 251, 532–541.
Klassen, H., Lund, R.D., 1987. Retinal transplants can drive a pupillary
reflex in host rat brains. Proc. Natl. Acad. Sci. USA 84, 6958–6960.
Koike, C., Nishida, A., Akimoto, K., Nakaya, M.A., Noda, T., Ohno, S.,
Furukawa, T., 2005. Function of atypical protein kinase C lambda in
differentiating photoreceptors is required for proper lamination of
mouse retina. J. Neurosci. 25, 10290–10298.
Kuwabara, T., Weidman, T.A., 1974. Development of the prenatal rat
retina. Invest. Ophthalmol. 13, 725–739.
Lamoureux, P., Buxbaum, R.E., Heidemann, S.R., 1989. Direct evidence
that growth cones pull. Nature 340, 159–162.
Lamoureux, P., Zheng, J., Buxbaum, R.E., Heidemann, S.R., 1992. A
cytomechanical investigation of neurite growth on different culture
surfaces. J. Cell Biol. 118, 655–661.
Larison, K.D., Bremiller, R., 1990. Early onset of phenotype and cell
patterning in the embryonic zebrafish retina. Development 109,
567–576.
Layer, P.G., Robitzki, A., Rothermel, A., Willbold, E., 2002. Of layers
and spheres: the reaggregate approach in tissue engineering. Trends
Neurosci. 25, 131–134.
Leckband, D., Prakasam, A., 2006. Mechanism and dynamics of cadherin
adhesion. Annu. Rev. Biomed. Eng.
Lecuit, T., Lenne, P.F., 2007. Cell surface mechanics and the control of
cell shape, tissue patterns and morphogenesis. Nat. Rev. Mol. Cell
Biol. 8, 633–644.
Li, M., Sakaguchi, D.S., 2004. Inhibition of integrin-mediated adhesion
and signaling disrupts retinal development. Dev. Biol. 275, 202–214.
Liets, L.C., Eliasieh, K., van der List, D.A., Chalupa, L.M., 2006.
Dendrites of rod bipolar cells sprout in normal aging retina. Proc.
Natl. Acad. Sci. USA 103, 12156–12160.
Lin, B., Wang, S.W., Masland, R.H., 2004. Retinal ganglion cell type, size,
and spacing can be specified independent of homotypic dendritic
contacts. Neuron 43, 475–485.
Liu, X., Sarthy, V.P., Craft, C.M., Honda, Y., Ide, C., 2002. Developmental expression of beta-catenin in mouse retina. Anat. Sci. Int. 77,
182–188.
Livesey, F.J., Cepko, C.L., 2001. Vertebrate neural cell-fate determination: lessons from the retina. Nat. Rev. Neurosci. 2, 109–118.
Lohmann, C., Myhr, K.L., Wong, R.O., 2002. Transmitter-evoked
local calcium release stabilizes developing dendrites. Nature 418,
177–181.
Lu, Y.B., Franze, K., Seifert, G., Steinhauser, C., Kirchhoff, F., Wolburg,
H., Guck, J., Janmey, P., Wei, E.Q., Kas, J., Reichenbach, A., 2006.
Viscoelastic properties of individual glial cells and neurons in the CNS.
Proc. Natl. Acad. Sci. USA 103, 17759–17764.
Lunardi, A., Cremisi, F., Dente, L., 2006. Dystroglycan is required for
proper retinal layering. Dev. Biol. 290, 411–420.
Lund, R.D., Hankin, M.H., Rao, K., Radel, J.D., Galli, L., 1988. Retinal
transplants and plasticity of the primary optic projections in rodents.
In: Petit, T.L., Ivy, G.P. (Eds.), Neural Plasticity: A Lifespan
Approach, vol. 39. A. Liss, New York, pp. 1–19.
ARTICLE IN PRESS
282
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
Malicki, J., Driever, W., 1999. oko meduzy mutations affect neuronal
patterning in the zebrafish retina and reveal cell–cell interactions of the
retinal neuroepithelial sheet. Development 126, 1235–1246.
Malicki, J., Neuhauss, S.C., Schier, A.F., Solnica-Krezel, L., Stemple,
D.L., Stainier, D.Y., Abdelilah, S., Zwartkruis, F., Rangini, Z.,
Driever, W., 1996. Mutations affecting development of the zebrafish
retina. Development 123, 263–273.
Malicki, J., Jo, H., Pujic, Z., 2003. Zebrafish N-cadherin, encoded by the
glass onion locus, plays an essential role in retinal patterning. Dev.
Biol. 259, 95–108.
Marc, R.E., Jones, B.W., Watt, C.B., Strettoi, E., 2003. Neural
remodeling in retinal degeneration. Prog. Retin. Eye Res. 22, 607–655.
Marrs, G.S., Honda, T., Fuller, L., Thangavel, R., Balsamo, J., Lilien, J.,
Dailey, M.E., Arregui, C., 2006. Dendritic arbors of developing retinal
ganglion cells are stabilized by beta 1-integrins. Mol. Cell Neurosci. 32,
230–241.
Masai, I., Lele, Z., Yamaguchi, M., Komori, A., Nakata, A., Nishiwaki,
Y., Wada, H., Tanaka, H., Nojima, Y., Hammerschmidt, M., Wilson,
S.W., Okamoto, H., 2003. N-cadherin mediates retinal lamination,
maintenance of forebrain compartments and patterning of retinal
neurites. Development 130, 2479–2494.
Matsunaga, M., Hatta, K., Takeichi, M., 1988. Role of N-cadherin cell
adhesion molecules in the histogenesis of neural retina. Neuron 1,
289–295.
McBrien, N.A., Gentle, A., 2003. Role of the sclera in the development
and pathological complications of myopia. Prog. Retin. Eye Res. 22,
307–338.
McCabe, K.L., Gunther, E.C., Reh, T.A., 1999. The development of the
pattern of retinal ganglion cells in the chick retina: mechanisms that
control differentiation. Development 126, 5713–5724.
Mehalow, A.K., Kameya, S., Smith, R.S., Hawes, N.L., Denegre, J.M.,
Young, J.A., Bechtold, L., Haider, N.B., Tepass, U., Heckenlively,
J.R., Chang, B., Naggert, J.K., Nishina, P.M., 2003. CRB1 is essential
for external limiting membrane integrity and photoreceptor morphogenesis in the mammalian retina. Hum. Mol. Genet. 12, 2179–2189.
Morgan, J.L., Dhingra, A., Vardi, N., Wong, R.O., 2006. Axons and
dendrites originate from neuroepithelial-like processes of retinal
bipolar cells. Nat. Neurosci. 9, 85–92.
Mu, X., Zhao, S., Pershad, R., Hsieh, T.F., Scarpa, A., Wang, S.W.,
White, R.A., Beremand, P.D., Thomas, T.L., Gan, L., Klein, W.H.,
2001. Gene expression in the developing mouse retina by EST
sequencing and microarray analysis. Nucleic Acids Res. 29, 4983–4993.
Nakano, I., Funahashi, M., Takada, K., Toda, T., 1996. Are breaches in
the glia limitans the primary cause of the micropolygyria in
Fukuyama-type congenital muscular dystrophy (FCMD)? Pathological study of the cerebral cortex of an FCMD fetus. Acta Neuropathol.
91, 313–321.
Nakagawa, S., Takada, S., Takada, R., Takeichi, M., 2003. Identification
of the laminar-inducing factor: Wnt-signal from the anterior rim
induces correct laminar formation of the neural retina in vitro. Dev.
Biol. 260, 414–425.
Novelli, E., Leone, P., Resta, V., Galli-Resta, L., 2007. A threedimensional analysis of the development of the horizontal cell mosaic
in the rat retina: implications for the mechanisms controlling pattern
formation. Vis. Neurosci. 24, 91–98.
Omori, Y., Malicki, J., 2006. oko meduzy and related crumbs genes are
determinants of apical cell features in the vertebrate embryo. Curr.
Biol. 16, 945–957.
Parrish, J.Z., Emoto, K., Kim, M.D., Jan, Y.N., 2007. Mechanisms that
regulate establishment, maintenance, and remodeling of dendritic
fields. Annu. Rev. Neurosci. 30, 399–423.
Perry, V.H., Linden, R., 1982. Evidence for dendritic competition in the
developing retina. Nature 297, 683–685.
Poggi, L., Vitorino, M., Masai, I., Harris, W.A., 2005a. Influences on
neural lineage and mode of division in the zebrafish retina in vivo.
J. Cell Biol. 171, 991–999.
Poggi, L., Zolessi, F.R., Harris, W.A., 2005b. Time-lapse analysis of
retinal differentiation. Curr. Opin. Cell Biol. 17, 676–681.
Pujic, Z., Malicki, J., 2001. Mutation of the zebrafish glass onion locus
causes early cell-nonautonomous loss of neuroepithelial integrity
followed by severe neuronal patterning defects in the retina. Dev.
Biol. 234, 454–469.
Ramon y Cajal, S., 1892. La rétine des vertebrés. La Cellule (English
translation by Thorpe, S and Glikstein, M, Charles C. Thomas,
Sprigfield, IL, USA, 1972).
Ramon y Cajal, S., 1929. Etudes sur la neurogenese de quelques
vertebrates. Madrid (English translation by Guth, L, Charles C.
Thomas, Sprigfield, IL, USA, 1960).
Rapaport, D.H., 2006. Retinal neurogenesis. In: Sernagor, E., Eglen, S.J.,
Harris, W., Wong, R.O.L. (Eds.), Retinal Development. Cambridge
University Press, Cambridge, pp. 30–58.
Raven, M.A., Reese, B.E., 2003. Mosaic regularity of horizontal cells in
the mouse retina is independent of cone photoreceptor innervation.
Invest. Ophthalmol. Vis. Sci. 44, 965–973.
Raven, M.A., Eglen, S.J., Ohab, J.J., Reese, B.E., 2003. Determinants of
the exclusion zone in dopaminergic amacrine cell mosaics. J. Comp.
Neurol. 461, 123–136.
Raven, M.A., Stagg, S.B., Nassar, H., Reese, B.E., 2005. Developmental
improvement in the regularity and packing of mouse horizontal cells:
implications for mechanisms underlying mosaic pattern formation.
Vis. Neurosci. 22, 569–573.
Raven, M.A., Oh, E.C., Swaroop, A., Reese, B.E., 2007. Afferent control
of horizontal cell morphology revealed by genetic respecification of
rods and cones. J. Neurosci. 27, 3540–3547.
Raymond, S.M., Jackson, I.J., 1995. The retinal pigmented epithelium is
required for development and maintenance of the mouse neural retina.
Curr. Biol. 5, 1286–1295.
Raymond, P.A., Barthel, L.K., Curran, G.A., 1995. Developmental
patterning of rod and cone photoreceptors in embryonic zebrafish.
J. Comp. Neurol. 359, 537–550.
Reese, B.E., Galli-Resta, L., 2002. The role of tangential dispersion in
retinal mosaic formation. Prog. Retin. Eye Res. 21, 153–168.
Reese, B.E., Harvey, A.R., Tan, S.-S., 1995. Radial and tangential
dispersion patterns in the mouse retina are cell-class specific. Proc.
Natl. Acad. Sci. USA 92, 2494–2498.
Reese, B.E., Necessary, B., Tam, P., Faulkner-Jones, B., Tan, S., 1999a.
Clonal expansion and cell dispersion in the developing mouse retina.
Eur. J. Neurosci. 11, 2965–2978.
Reese, B.E., Necessary, B.D., Tam, P.P., Faulkner-Jones, B., Tan, S.S.,
1999b. Clonal expansion and cell dispersion in the developing mouse
retina. Eur. J. Neurosci. 11, 2965–2978.
Reese, B.E., Raven, M.A., Stagg, S.B., 2005. Afferents and homotypic
neighbors regulate horizontal cell morphology, connectivity, and
retinal coverage. J. Neurosci. 25, 2167–2175.
Reichenbach, A., Eberhardt, W., Scheibe, R., Deich, C., Seifert, B.,
Reichelt, W., Dahnert, K., Rodenbeck, M., 1991. Development of the
rabbit retina. IV. Tissue tensility and elasticity in dependence on
topographic specializations. Exp. Eye Res. 53, 241–251.
Resta, V., Novelli, E., Di Virgilio, F., Galli-Resta, L., 2005. Neuronal
death induced by endogenous extracellular ATP in retinal cholinergic
neuron density control. Development 132, 2873–2882.
Richard, M., Roepman, R., Aartsen, W.M., van Rossum, A.G., den
Hollander, A.I., Knust, E., Wijnholds, J., Cremers, F.P., 2006.
Towards understanding CRUMBS function in retinal dystrophies.
Hum. Mol. Genet. 15 (Spec. no. 2), R235–R243.
Rockhill, R.L., Euler, T., Masland, R.H., 2000. Spatial order within but
not between types of retinal neurons. Proc. Natl. Acad. Sci. USA 97,
2303–2307.
Ruan, G., Wedlich, D., Koehler, A., 2006. Xenopus cadherin-6 regulates
growth and epithelial development of the retina. Mech. Dev. 123,
881–892.
Scheibe, R., Schnitzer, J., Röhrenbeck, J., Wohlrab, F., Reichenbach, A.,
1995. Development of A-type (axonless) horizontal cells in the rabbit
retina. J. Comp. Neurol. 354, 438–458.
Schroder, J.E., Tegeler, M.R., Grosshans, U., Porten, E., Blank, M., Lee,
J., Esapa, C., Blake, D.J., Kroger, S., 2007. Dystroglycan regulates
ARTICLE IN PRESS
L. Galli-Resta et al. / Progress in Retinal and Eye Research 27 (2008) 260–283
structure, proliferation and differentiation of neuroepithelial cells in
the developing vertebrate CNS. Dev. Biol. 307, 62–78.
Sernagor, E., Eglen, S., Harris, B., Wong, R., 2006. Cambridge University
Press, Cambridge.
Sertie, A.L., Sossi, V., Camargo, A.A., Zatz, M., Brahe, C., Passos-Bueno,
M.R., 2000. Collagen XVIII, containing an endogenous inhibitor of
angiogenesis and tumor growth, plays a critical role in the maintenance
of retinal structure and in neural tube closure (Knobloch syndrome).
Hum. Mol. Genet. 9, 2051–2058.
Steinberg, M.S., 1996. Adhesion in development: an historical overview.
Dev. Biol. 180, 377–388.
Steinberg, M.S., 2007. Differential adhesion in morphogenesis: a modern
view. Curr. Opin. Genet. Dev. 17, 281–286.
Svennevik, E., Linser, P.J., 1993. The inhibitory effects of integrin
antibodies and the RGD tripeptide on early eye development. Invest.
Ophthalmol. Vis. Sci. 34, 1774–1784.
Takeichi, M., 1995. Morphogenetic roles of classic cadherins. Curr. Opin.
Cell Biol. 7, 619–627.
Takeichi, M., 2007. The cadherin superfamily in neuronal connections and
interactions. Nat. Rev. Neurosci. 8, 11–20.
Tanabe, K., Takahashi, Y., Sato, Y., Kawakami, K., Takeichi, M.,
Nakagawa, S., 2006. Cadherin is required for dendritic morphogenesis
and synaptic terminal organization of retinal horizontal cells.
Development 133, 4085–4096.
Tepass, U., Harris, K.P., 2007. Adherens junctions in Drosophila retinal
morphogenesis. Trends Cell Biol. 17, 26–35.
Tepass, U., Theres, C., Knust, E., 1990. crumbs encodes an EGF-like
protein expressed on apical membranes of Drosophila epithelial cells
and required for organization of epithelia. Cell 61, 787–799.
Usukura, J., Shuichi, O., 1996. Morphogenesis of photoreceptor outer
segments in retinal development. Prog. Retin. Eye Res. 15, 113–125.
van de Pavert, S.A., Sanz, A.S., Aartsen, W.M., Vos, R.M., Versteeg, I.,
Beck, S.C., Klooster, J., Seeliger, M.W., Wijnholds, J., 2007. Crb1 is a
determinant of retinal apical Muller glia cell features. Glia 55,
1486–1497.
Wässle, H., Boycott, B.B., 1991. Functional architecture of the
mammalian retina. Physiol. Rev. 71, 447–480.
Wässle, H., Riemann, H.J., 1978. The mosaic of nerve cells in the
mammalian retina. Proc. R. Soc. Lond. B 200, 441–461.
283
Wässle, H., Peichl, L., Boycott, B.B., 1983. A spatial analysis of on- and
off-ganglion cells in the cat retina. Vision Res. 23, 1151–1160.
Wei, X., Malicki, J., 2002. nagie oko, encoding a MAGUK-family protein,
is essential for cellular patterning of the retina. Nat. Genet. 31,
150–157.
Wei, X., Cheng, Y., Luo, Y., Shi, X., Nelson, S., Hyde, D.R., 2004. The
zebrafish Pard3 ortholog is required for separation of the eye fields and
retinal lamination. Dev. Biol. 269, 286–301.
Wei, X., Zou, J., Takechi, M., Kawamura, S., Li, L., 2006. Nok plays an
essential role in maintaining the integrity of the outer nuclear layer in
the zebrafish retina. Exp. Eye Res. 83, 31–44.
Wikler, K.C., Rakic, P., 1991. Relation of an array of early-differentiated
cones to the photoreceptor mosaic in the primate retina. Nature 351,
397–400.
Wikler, K.C., Rakic, P., Bhattacharyya, N., Macleish, P.R., 1997. Early
emergence of photoreceptor mosaicism in the primate retina revealed by a
novel cone-specific monoclonal antibody. J. Comp. Neurol. 377, 500–508.
Williams, R.S., Swisher, C.N., Jennings, M., Ambler, M., Caviness Jr.,
V.S., 1984. Cerebro-ocular dysgenesis (Walker–Warburg syndrome):
neuropathologic and etiologic analysis. Neurology 34, 1531–1541.
Wong, R.O., 1999. Retinal waves and visual system development. Annu.
Rev. Neurosci. 22, 29–47.
Wong, R.O., Ghosh, A., 2002. Activity-dependent regulation of dendritic
growth and patterning. Nat. Rev. Neurosci. 3, 803–812.
Yamagata, M., Weiner, J.A., Sanes, J.R., 2002. Sidekicks: synaptic
adhesion molecules that promote lamina-specific connectivity in the
retina. Cell 110, 649–660.
Zheng, J., Lamoureux, P., Santiago, V., Dennerll, T., Buxbaum, R.E.,
Heidemann, S.R., 1991. Tensile regulation of axonal elongation and
initiation. J. Neurosci. 11, 1117–1125.
Zheng, J., Buxbaum, R.E., Heidemann, S.R., 1993. Investigation of
microtubule assembly and organization accompanying tensioninduced neurite initiation. J. Cell Sci. 104, 1239–1250.
Zheng, J., Buxbaum, R.E., Heidemann, S.R., 1994. Measurements of
growth cone adhesion to culture surfaces by micromanipulation.
J. Cell Biol. 127, 2049–2060.
Zolessi, F.R., Poggi, L., Wilkinson, C.J., Chien, C.B., Harris, W.A., 2006.
Polarization and orientation of retinal ganglion cells in vivo. Neural
Dev. 1, 2.