Available online at www.sciencedirect.com
R
Molecular and Cellular Neuroscience 22 (2003) 25–36
www.elsevier.com/locate/ymcne
Xrx1 controls proliferation and multipotency of retinal progenitors
Simona Casarosa,a,b Marcos A. Amato,a,b Massimiliano Andreazzoli,a,b Gaia Gestri,a,b
Giuseppina Barsacchi,a,b and Federico Cremisia,b,c,*
a
Sezione di Biologia Cellulare e dello Sviluppo, Dipartimento di Fisiologia e Biochimica, Università di Pisa,
via Carducci, 13-56010 Ghezzano (Pisa), Italy
b
Università degli Studi di Pisa, Centro di Eccellenza AmbiSEN, Pisa, Italy
c
Scuola Normale Superiore, piazza dei Cavalieri, 7-56100, Pisa, Italy
Received 9 July 2002; revised 10 October 2002; accepted 25 October 2002
Abstract
We investigated the function of Xrx1 during Xenopus retinogenesis. Xrx1 overexpression lengthens mitotic activity and ectopically
activates the expression of markers of undifferentiated progenitors in the developing retina. We assayed Xrx1 ability to support proliferation
with a cell-autonomous mechanism by in vivo lipofection of single retinal progenitors. Xrx1 overexpression increases clonal proliferation
while Xrx1 functional inactivation exerts the opposite effect. We also compared the effects of Xrx1 with those of the cyclin-dependent kinase
cdk2, a strong mitotic promoter. Despite the similar increase in clonal proliferation displayed by both factors, Xrx1 and cdk2 act differently
on retinal cell fate determination. cdk2/cyclinA2 lipofected retinas show a decrease in early-born cell types as ganglion cells and cones and
an increase in late-born types such as bipolar neurons. On the contrary, Xrx1 lipofected retinas show no changes in the proportions of the
different cell types, thus suggesting a role in supporting multipotency of retinal progenitors.
© 2003 Elsevier Science (USA). All rights reserved.
Introduction
In the nervous system, the mechanisms coordinating cell
proliferation and fate commitment of neuronal progenitors
are fundamental for the construction of layered structures,
such as the retina. In these structures, the generation of
definite cell types in distinct layers is also influenced by the
time when precursor cells exit the cell cycle (reviewed by
Livesey and Cepko, 2001; Ohnuma et al., 2001). The development of the vertebrate retina recapitulates most of the
general mechanisms governing neuronal development. Retinal cell types are generated through temporal waves that
are distinct (even if partially overlapping) and conserved in
the retinogenesis of different vertebrate species. One single
multipotent retinal precursor can generate all retinal cell
types, by means of multiple cell divisions, while undergoing
a progressive restriction of its differentiation potential. Ganglion cells are generated first, followed by cones and hori-
* Corresponding author. Fax: ⫹39-050-878486.
E-mail address: cremisi@dfb.unipi.it (F. Cremisi).
zontal cells, amacrine cells, rods, bipolar cells, and, finally,
Müller glia (reviewed by Harris, 1997; Livesey and Cepko,
2001). Understanding how multipotent retinal precursors
self-renew, or restrict their potential for generating the right
number of cell types at the right time and in the appropriate
retinal layer, is one of the main topics in the study of eye
development.
Both extrinsic cues and intrinsic factors regulate proliferation versus differentiation during retinogenesis. Among
extrinsic growth regulators, Notch and Delta (ArtavanisTsakonas et al., 1999; Baker, 2000), Shh (Jensen and Wallace, 1997), and FGF family members (McFarlane et al.,
1998) support mitotic activity. This effect is achieved either
by downregulating genes that favor exit from the cell cycle
and/or by sustaining the expression of genes driving the cell
cycle. These same factors also contribute to cell fate decisions (Ezzeddine et al., 1997; Henrique et al., 1997; Furukawa et al., 2000; Lillien and Wancio, 1998; McFarlane et
al., 1998). Intrinsic factors are part of the cell cycle machinery or belong to a numerous family of determination/differentiation factors that, in turn, regulate the cell cycle ma-
1044-7431/03/$ – see front matter © 2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S1044-7431(02)00025-8
26
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
chinery in concert with the above-mentioned extrinsic cues
(Ohnuma et al., 2001). As intrinsic determination/differentiation factors, bHLH transcription factors are both activators and inhibitors of neuronal differentiation, acting to
balance proliferation and cell differentiation (Kageyama
and Nakanishi, 1997; Cepko, 1999). Most bHLH factors are
expressed in all the nervous system and control neuronal
differentiation with no evident cell-type specificity. Conversely, some of them are expressed in specific subsets of
neuronal precursors and affect the differentiation of definite
cell types of the retina, such as Math5 in mouse (Brown et
al., 2001) and Xath5 and neurogenin in Xenopus (Kanekar
et al., 1997; Perron et al., 1999).
The precise relationships among control of the cell
cycle machinery, extrinsic and intrinsic cues, and generation of specific retinal cell types are not yet well understood in the retina. Moreover, little is known about patterning genes that affect cell proliferation of specific
vertebrate CNS structures cell-autonomously, such as
genes coding for transcription factors that have an effect
on those same cells in which they are expressed. The
Xenopus homeobox-containing gene Xrx1 is a possible
candidate for this function in the retina (Casarosa et al.,
1997; Furukawa et al., 1997). Xrx1 is one of the master
regulators of vertebrate eye development, together with
Pax6 and Six3 (Mathers et al., 1997; Chow et al., 1999;
Zhou et al., 2000). Among them, Xrx1 shows the most
eye-specific pattern of expression, covering first most of
the presumptive forebrain at early neurula stage and later
the eye vescicle, the pineal gland, and a small ventral
diencephalic region (Casarosa et al., 1997). Xrx1 expression progressively disappears from differentiating retinal
cells, while it is retained in the ciliary marginal zone
(CMZ), which is a source of retinal stem cells in the
adulthood of fish and amphibians (Harris and Perron,
1998). Gain- and loss-of-function studies of rx in different species have shown this gene to be necessary for the
formation of anterior brain structures and eye, and sufficient to induce retina and neural tube hypertrophy
(Mathers et al., 1997; Andreazzoli et al., 1999; Winkler
et al., 2000; Chuang and Raymond, 2001; Loosli et al.,
2001). However, the mechanisms of rx action are still
poorly understood. Moreover, since these studies were
performed on early embryos, they do not analyze the role
of rx1 in the proliferation and differentiation of retinal
progenitors.
Here we report that Xrx1 overexpression is able to induce
retina and neural tube hypertrophy through a territory-restricted increase in cell proliferation, which is due to a
longer persistence of dividing precursor cells during development. Moreover, we show that Xrx1 promotes clonal
proliferation and maintains multipotency of single retinal
progenitors throughout retinogenesis, without affecting
their differentiation. We thus suggest a role for Xrx1 in
supporting retinal stem cells.
Results
Xrx1 promotes proliferation of retinal progenitors and
delays neuronal differentiation
During retinogenesis, Xrx1 is expressed in undifferentiated progenitors and in CMZ cells, where its expression
persists as retinogenesis is terminated (Mathers et al., 1997;
Perron et al., 1998). Double labeling with a Xrx1 probe and
an antibody to phosphohistone-H3 (which recognizes cells
undergoing mitotic prophase) (Hendzel et al., 1997) shows
that all M-phase cells of the CMZ express Xrx1 (Fig. 2A,
arrows). Thus, Xrx1 seems to be expressed in undifferentiated, proliferating retinal precursors. To understand whether
Xrx1 plays a role in the generation and/or in the support of
undifferentiated retinal precursors, we performed an overexpression study making use of markers of either cell proliferation or differentiation. Previous studies showed that
Xrx1 overexpression at early stages generates an expanded
neural retina (Mathers et al., 1997; Andreazzoli et al., 1999).
We analyzed the effect of the injection of 100 pg of Xrx1
mRNA in one dorsal animal blastomere of four-cell-stage
embryos (Andreazzoli et al., 1999) on the expression of
cyclin D1, a proliferation marker, and N-tubulin, a marker of
neuronal differentiation, at different stages of retinogenesis.
At stage 37 we can see an increase of cyclinD1 and a
decrease in N-tubulin expression, in both the diencephalon
and the retina of the injected side compared with the control
side (Figs. 1A, 1B, 1E, 1F, 1Q, 1R). In the control side,
N-tubulin marks mainly the external, differentiated region
of the neural tube (Fig. 1R) and, in the eye, the ganglion cell
layer (GCL) and the most central region of the inner nuclear
layer (INL), where differentiation has already occurred (Fig.
1B). Conversely, in the injected side at this stage N-tubulin
is down-regulated in the neural tube as well as in the retina
(Figs. 1F, 1R). In the control retina, cyclinD1 expression is
somewhat complementary to that of N-tubulin, marking the
dividing precursors of the CMZ (Fig. 1A). In the injected
retina the domain of expression of cyclinD1 is largely expanded with respect to the control, comprising not only the
CMZ but also a good portion of the central retina (Fig. 1E).
A similar situation is found in the anterior neural tube (Fig.
1Q). On the whole, the expression pattern of these two
markers at stage 36/37 can be interpreted as an enlargement
of the proliferating retina, after Xrx1 injection. In accordance with the induction of cyclin D1 expression, the domain of Xotch expression is also enlarged in the injected
retina compared with the control (Figs. 1D, 1H). At this
stage, Xotch is downregulated in the GCL of the control
retina (Fig. 1D), which contains differentiated ganglion
cells, while it is still expressed in the GCL of Xrx1-overexpressing retinas (Fig. 1H).
To better evaluate the state of commitment of cells, we
then analyzed the expression of markers of either general
neuronal commitment (Xdelta, not shown) or more specific
retinal commitment (Xath5). We observed a small but con-
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
27
Fig. 1. Effects of Xrx1 overexpression on retina and neural tube. In situ hybridizations of markers of cell proliferation and differentiation on embryos at stages
37 (A–H and Q–T) and 45 (I–P). Probes for cyclinD1 (A,E,I,M,Q), N-tubulin (B,F,J,N,R), Xath5 (C,G,K,O,S), and Xotch (D,H,L,P,T) were hybridized to
cross sections of control retinas (A–D, I–L), injected retinas (E–H, M–P), and anterior neural tube (Q–T). Arrows in (Q)–(T) point at duplicated neural tubes.
GCL, ganglion cell layer; INI, inner nuclear layer; CMZ, ciliary marginal zone; ini, injected side; con, control side.
sistent increase in the expression level of both markers in
the injected retina. Xath5 is expressed in a larger number of
cells compared with the control (compare Fig. 1C with 1G).
We also show that Xath5 is not expressed in the hypertrophic neural tube (Fig. 1S). This observation indicates that
the hypertrophic tissue generated by Xrx1 overexpression in
the anterior region, outside of the eye, has a neural, but not
a retinal identity. Interestingly, when we analyze these
markers at the end of retinogenesis, we observe that the
domains of expression found in the injected side are comparable to those found in the uninjected side (Figs. 1I–1P).
So, Xrx1 is able to extend the proliferation of neuronal
precursors and to delay their differentiation, but does not
cause a total block of differentiation.
Regional competence of Xrx1 proliferative activity
We wanted to further understand the relationship between increased cell proliferation and the hypertrophic neural tissue produced by Xrx1 overexpression. To this aim, we
analyzed the expression of phosphohistone-H3 in the eye
and in the neural tube, after unilateral injection of Xrx1
mRNA in four-cell-stage embryos. We focused our analysis
on stage 42 embryos, since at this stage proliferation of
neuronal precursors is almost completed in the retina, except for the CMZ (Perron et al., 1998). Our findings show
that, in the injected side, the number of phosphohistone-H3
labeled cells is higher in both the eye and anterior neural
tube, with respect to the corresponding structures in the
control side (Figs. 2B, 2E, 2F). In addition, phosphohistoneH3-positive cells are widespread through all retinal regions
(Fig. 2E), and not confined to the CMZ (Fig. 2F). On the
contrary, analysis of more posterior sections of the neural
tube shows no evident difference between the injected and
control sides (Fig. 2C).
We carried out a quantitative analysis on serial sections
of 20 injected embryos. We counted phosphohistone-H3positive cells of both retina and neural tube and compared
their numbers in the injected side with respect to the control
side (Fig. 2G). An anteroposterior (A/P) subdivision of the
28
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
neural tube was taken into account for the comparison: we
analyzed forebrain and hindbrain separately (Figs. 2B, 2C,
2G). The resulting figures show that Xrx1 induces a significant increase in mitotic activity both in the retina (6.1
mitoses/section in injected retina vs 4 mitoses/section in
control) and in the forebrain (5 mitoses/section in injected
neural tube vs 3.2 mitoses/section in control), but not in the
hindbrain (1.9 vs 1.8 mitoses/section, not significant). This
latter result does not depend on a differential distribution of
the injected mRNA along the A/P axis, since the co-injected
tracer (GFP mRNA) diffused up to the very caudal regions
of the neural tube in all embryos analyzed (not shown).
As a further control of our experiments, we tested the
ability of activators of the cell cycle machinery, namely
cdk2 and cyclinA2, to increase the mitotic rate in the hindbrain. Overexpression of these molecules by means of
mRNA injection is known to cause an enlargement of the
neural tube in Xenopus embryos (Zuber et al., 1999), as they
inhibit exit from the cell cycle of neural precursors. In fact,
when analyzed at stage 42, sections of both forebrain and
hindbrain highlighted a significant increase in mitoses in the
cdk2/cyclinA2-injected side (Figs. 2D, 2G). The average
number of mitoses per section increased significantly from
2.63 to 3.91 in the forebrain and from 2.40 to 3.54 in the
hindbrain.
Finally, we performed control experiments to assay the
ability of Xrx1 overexpression to inhibit naturally occurring
programmed cell death. After Xrx1 mRNA injection in one
blastomere at two cell stage, we compared cell death between injected and control side by TUNEL assay. Analysis
carried out at stage 31–35 showed no decrease of apoptosis
in the injected side (not shown), thus excluding a direct
involvement of Xrx1 in the control of the apoptotic program
during retinogenesis.
Xrx1 promotes clonal proliferation of retinal progenitors
Fig. 2. Analysis of mitotic activity in Xrx1-injected embryos. (A) Double
staining on a cross section of a stage 42 CMZ. Xrx1 mRNA was detected
by in situ hibridization as red fluorescence; phosphohistone-H3 was revealed by DAB (brown). Arrows point to mitotic nuclei also expressing
Xrx1. (B–F) Cross sections of stage 42 embryos stained with anti-phosphohistone-H3. Embryos were injected with Xrx1 (B, C, E) or with cyclinA2/cdk2 (D) mRNAs. (B) Forebrain. (C, D) Hindbrain. (E) Injected
eye. (F) Control eye. In blue, (B). Hoechst nuclear staining; arrows point
to red phosphohistone-H3-positive cells. In (E), note that the morphology
of the injected retina shows dramatic disorganization of cell layering due to
Xrx1 overexpression. L, lens; CMZ, ciliary marginal zone; inj, injected
side; con, control side. (G) Statistical analysis of phosphohistone-H3positive cells in the injected and control sides of embryos injected with
Xrx1 or with cyclinA2/cdk2. The Y axis shows the average numbers of
mitoses/section in either the injected or the control sides of both eye and
neural tube. FB, forebrain, HB, hindbrain. Xrx1 injection in eye and FB and
One way to further investigate the proliferative activity
of Xrx1 in retinal precursors is the analysis of cell number
in clones overexpressing it. We thus compared the clone
size of control retinal progenitors with that of progenitors
misexpressing Xrx1 or other genes controlling cell proliferation. To this aim we cotransfected appropriate expression
constructs, with a GFP plasmid as a clonal tracer, in the
optic vesicle of stage 17/18 embryos, by means of lipofection (Holt et al., 1990). Previous studies described that
isolated clusters of transfected cells may actually correspond to the clonal progeny originated by single retinal
cyclinA2/cdk2 injection in FB and HB, respectively, cause significant
increase of mitotic activity with respect to control, as evaluated by Student’s t-test analysis (P ⬍ 0.02). Error bars indicate SEM. Cell count: inj
eye, n ⫽ 609; con eye, n ⫽ 400; inj FB, n ⫽ 1011 (Xrx1), n ⫽ 180
(cyclinA2/cdk2); con FB, n ⫽ 655 (Xrx1), n ⫽ 121 (cyclinA2/cdk2); inj HB,
n ⫽ 245 (Xrx1), n ⫽ 170 (cyclinA2/cdk2); con HB, n ⫽ 227 (Xrx1) n ⫽
115 (cyclinA2/cdk2).
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
29
Fig. 3. Clusters of clonally related cells of lipofected retinas. Clusters of cells lipofected with GFP (A), Xrx1 (B), or Xrx1–EnR (C). The four pseudo-colors
red, green, blue, and white, which stain lipofected cells, were obtained by digital elaboration of the GFP reporter signal from four 12-␮m serial sections,
respectively. Photographs corresponding to the four serial sections were overlaid to show an entire cluster of lipofected cells. Each of the three clusters shown
(encircled by dashed lines) contains all the lipofected cells of an entire retina. Asterisks indicate transfected cells in the lens. The three examples represent
typical results obtained in the clonal analysis (see Experimental methods). ONL, outer nuclear layer.
progenitors, under certain experimental conditions (Zuber et
al., 1999) (see Experimental methods). We analyzed the
transfected cells as detected by fluorescence on cryostat
Fig. 4. Proliferation and differentiation of retinal cells after lipofection.
Retinal sections at stage 42 after cotransfection of GFP and different
combinations of constructs. In blue, Hoechst staining of nuclei; in green,
GFP reporter activity. (A) Control pCS2 vector. (B) Xrx1. (C) Xrx1–EnR.
(D, E) Same cluster of cells cotransfected with cyclinA2/cdk2 detected with
two different reporters. Transfected cells are GFP positive in (D), while a
myc tag carried by both cyclinA2 and cdk2 vectors allows red fluorescence
staining of lipofected cells in (E). (F) Digital overlay of (D) and (E), all
cells display both reporter activities, resulting in different degrees of
yellow staining. (G) Cotransfection of cyclin A2/cdk2 and Xrx1–EnR. (H)
Bcl2. (I) Cotransfection of Bcl2 and Xrx1–EnR. Xrx1 lipofection (B)
increases the number of transfected cells in all retinal layers, while cotransfection of cyclinA2/cdk2 (D–F) decreases the number of cells in the
ONL and GCL. Note that the decreased number of transfected cells after
Xrx1–EnR lipofection (C) is rescued after Bcl2 cotransfection (I) but not
after cyclinA2/cdk2 cotransfection (G). GCL, ganglion cell layer; INL,
inner nuclear layer; ONL, outer nuclear layer.
sections of stage 42 retinas. Figs. 3 and 4 show examples of
clusters cotransfected with GFP plus the control pCS2 plasmid (Figs. 3A, 4A), pCS2Xrx1 (Figs. 3B, 4B) or the dominant-negative construct pCS2Xrx1–EnR (Figs. 3C, 4C).
Each of the three pictures of Fig. 3 has been obtained by
superimposing four layers corresponding to four serial sections of 12 ␮m each, covering an entire cluster of transfected cells. The four different pseudocolors—white, blue,
green and red— correspond to the GFP fluorescence detected in the four different sections. The figure illustrates the
effects of Xrx1 on clone size: its overexpression makes
more than twice as many cells per cluster than the control
transfection (compare B with A), while the clonal size is
reduced to 30% of the control after Xrx1–EnR transfection
(compare C with A). Fig. 5A reports the statistical analysis
on clone size after transfection. The average clone size is
10.4 cells per clone in the control, 21.38 in Xrx1 transfected
retinas, and 3.62 after Xrx1–EnR transfection. Transfection
of a control plasmid containing the EnR domain alone
affects neither clone size nor cell type differentiation (results not shown).
Our observations indicate that Xrx1 overexpression promotes clonal proliferation of retinal precursors, while its
functional impairment produces the opposite effect. Notably, the increase in clone size due to Xrx1 lipofection is very
similar to that obtained after cdk2/cyclinA2 lipofection
(Figs. 4D– 4F, 5A).
The decreased clone size after Xrx1–EnR lipofection
could be due to an impairment of mitotic activity, to increased cell death, or to both. We previously showed Xrx1–
EnR expression to induce apoptosis (Andreazzoli et al.,
1999). Here we show that cdk2/cyclinA2 cotransfection cannot rescue Xrx1 functional inactivation by Xrx1–EnR (3.92
cells/clone) (Figs. 4G, 5A). Such observations are in accordance with a possible apoptotic effect of Xrx1 functional
inactivation. Nonetheless, a main goal remains to investigate the proliferative potency of retinal progenitors after
Xrx1 functional inactivation. To this purpose, we aimed to
rescue apoptosis induced by Xrx1–EnR transfection by
30
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
means of the antiapoptotic Bcl2 gene, which is an inhibitor
of specific cell death pathways in the nervous system (Chen
et al., 1997). Bcl2 structural and functional conservation has
been demonstrated, and the human gene was found to rescue
apoptosis induced by inhibitors of transcription or translation in Xenopus embryos (Hensey and Gautier, 1997).
Transfection of Bcl2 alone increases the average clone size
(17.47 cells/clone, Figs. 4H, 5A) compared with control
(10.4 cells/clone). Nevertheless, Bcl2/Xrx1–EnR cotransfected clones show lower size than Bcl2 transfected clones
(10.91 cells/clone) (Figs. 4I, 5A), thus suggesting that Xrx1
functional inactivation decreases the mitotic activity of retinal progenitors.
Xrx1 supports multipotency of retinal progenitors
We addressed the question of whether Xrx1 could be
involved in cell fate commitment of neural precursors. We
analyzed the cell types generated after lipofection of Xrx1 or
cdk2/cyclinA2. Fig. 4 shows examples of the results obtained after transfection of the different constructs assayed
in the present work. The relative percentages of the different
retinal cell types in control transfected clones are similar to
those already reported (Fig. 6B) (Kanekar et al., 1997;
Zuber et al., 1999). Transfection of Xrx1 had no effect on
the relative proportions of the various cell types, compared
with the control (Figs. 4A, 4B, 5B). In some experiments we
have further confirmed the identity of the cells transfected
with Xrx1 using specific antibodies. Fig. 6C shows that Xrx1
lipofected cells in the GCL effectively express Islet-1, a
marker of ganglion cells (Dorsky et al., 1997). Analysis of
R5 immunoreactivity (a marker of Müller glial cells)
(Drager et al., 1984) confirms that Xrx1 does not increase
Müller glia (Fig. 6D), conversely to what was observed in
mouse (Furukawa et al., 2000) (see Discussion). Transfection of cdk2/cyclinA2 generated a significant decrease in
ganglion cells, and cones, as shown by calbindin immunoreactivity (Fig. 6B), and an increase in bipolar cells (Figs.
4A, 4D– 4F, 5B). The different phenotypes exhibited by
Xrx1 and cdk2/cyclinA2 may point to different roles for
these two kinds of molecules, even though both are involved
in stimulating cell proliferation (see Discussion).
The ability of Xrx1 transfected cells to generate all retinal cell types in the same proportions as the control cells do
favors the possibility that Xrx1 activity is compatible with
the maintenance of precursors in a multipotent state. To
investigate this point, we analyzed the pattern of BrdU
incorporation of the different cell types generated by Xrx1
transfected progenitors. BrdU was administered at stage
35–36, when all early-born cells, namely, ganglion cells,
have already left the cell cycle and have started differentiation. Fig. 7A shows a section of a nontransfected retina at
stage 42 after BrdU immunostaining (green), over Hoechst
counterstaining (blue). In all retinas analyzed (n ⫽ 23)
BrdU is consistently distributed in a gradient decreasing
from the CMZ toward the central part of the retina, where
labeled cells are mostly in the INL. Notably, labeled cells
were never detected in the GCL. Conversely, BrdU-positive
cells, mostly bipolar cells of the INL and photoreceptors,
are detectable in the central retina, thus suggesting that the
BrdU was incorporated by later precursors.
We then analyzed retinal cells transfected at stage 17
either with pCS2 or with pCS2Xrx1, together with a myctag vector suitable for immunodetection. Figs. 7B–7E show
examples of results. Again, none of the control transfected
retinas analyzed (n ⫽ 11) show BrdU incorporation in the
GCL. Arrows in Figs. 7B and 7C point to one controltransfected cell of the GCL that did not incorporate BrdU.
Some control-transfected cells belonging to the same clone
are BrdU-positive (Figs. 7B, 7C) in the INL and in the
photoreceptor layer, thus demonstrating that BrdU incorporation was effective. Typical examples of Xrx1 transfected
cells are shown in Figs. 7D and 7E. As in the controltransfected clones, in Xrx1 transfected clones we find BrdUpositive cells in both the INL and photoreceptor layer.
Nonetheless, unlike the control, most of the Xrx1 transfected retinas analyzed (n ⫽ 12) show BrdU-positive transfected cells in the GCL as well (arrows in Figs. 7D, 7E).
Thus, Xrx1 over expression is likely to maintain early retinal
precursors, i.e., those fated to generate also ganglion cells,
in a proliferating condition beyond the due time (Fig. 8) (see
Discussion).
Discussion
In this article we show evidence that Xrx1 supports
proliferation and multipotency of retinal progenitors by cellautonomous mechanisms and with a region-specific competence. The way Xrx1 acts on retinal progenitors is different
from that of a general signal promoting cell proliferation,
such as the cdk2 cyclin-dependent kinase. Thus, Xrx1 represents a new, intrinsic cue acting on the control of retinal
proliferation.
Xrx1 controls cell proliferation in retina and forebrain
The net effect of Xrx1 overexpression in the early embryo is an overgrowth of neural tissue restricted to the
territories of Xrx1 expression, retina and forebrain (Mathers
et al., 1997; Andreazzoli et al., 1999; this article). Similarly,
expansion of neural retina has been shown when overexpressing rx1 and rx2 in zebrafish and rx3 in medaka, although the authors propose different mechanisms for these
effects (Chuang and Raymond, 2000; Loosli et al., 2001). In
agreement with the results obtained in medaka, we show
direct evidence that the retinal overgrowth is achieved in
Xenopus by increased mitotic activity. Moreover, the fact
that Xrx1 overexpression is not sufficient to induce expression of Xath5, a marker of retinal progenitors, in the ectopic
neural tube confirms that the hypertrophic neural tissue
induced by Xrx1 is not ectopic retina. This supports the idea
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
that the Xrx1 phenotype is due to a region-restricted increase in proliferation rather than to a transformation toward
retinal fate of other embryonic territories. Interestingly,
Xrx1 overexpression does not affect neuronal commitment
since neuronal cell differentiation is recovered by stage 45,
when the expression of the markers mentioned above becomes comparable to that of control retina.
Another key regulator of retinal development, the transcriptional repressor Xoptx2, induces retinal overgrowth by
supporting clonal proliferation of retinal progenitors. Notably, Xoptx2 is turned on after Xrx1 onset of expression
(Zuber et al., 1999) and Xrx1 overexpression is able to
induce Xoptx2 expression (M.A. et al., submitted), thus
suggesting the view that Xoptx2 might be downstream of
Xrx1 in a regulatory cascade. We also found that Xrx1 can
positively regulate both the Delta/Notch pathway and the
cell cycle machinery, by delaying the arrest of expression of
Xotch and cyclinD1 in the developing retina. These observations are examples of how tissue-specific genes such as
Xrx1 can regionally regulate cell proliferation by controlling
general extrinsic and intrinsic cues.
Two effects can account for the increased mitotic activity
of neural progenitors: a shortening of the cell cycle or a
lengthening of the proliferative state. Xrx1 overexpression
increases phosphohistone-H3 immunoreactivity in retina
and forebrain at stage 42, when normally most of the progenitors of these regions have already left the cell cycle.
This observation suggests that Xrx1-overexpressing cells
lengthen their proliferative condition. The same conclusion
is also supported by the analysis of markers of undifferentiated cells, such as cyclinD1 and Xotch (see Results and
Fig. 1).
We show by in vivo lipofection of single retinal progenitors that Xrx1 misexpression affects cell proliferation
with a cell-autonomous mechanism. However, while the
role of Xrx1 in supporting mitotic activity in gain-offunction experiments has been elucidated (see above), the
same cannot be easily predicted in functional inactivation
experiments. In fact, we previously showed that Xrx1 loss
of function by means of Xrx1–EnR mRNA injection induces apoptosis at early neural plate stage (Andreazzoli
et al., 1999), which does not allow analysis of the proliferative potency of Xrx1 targeted cells. To overcome
this technical restriction, we aimed to keep cells alive in
the absence of Xrx1 function by taking advantage of the
antiapoptotic Bcl2 gene. Bcl2 lipofection increases retinal clone size. This indicates that a Bcl2-sensitive pathway of cell death contributes to regulate the clone size of
progenitors, as a proportion of retinal cells normally
undergoes cell death during development (Bahr, 2000;
Gonzalez-Hoyuela et al., 2001). Nonetheless, clone size
after Bcl2/Xrx1–EnR cotransfection is lower than clone
size after transfection of Bcl2 alone. These data could be
interpreted by considering that Xrx1 loss of function
inhibits mitotic activity, since apoptosis by itself cannot
31
entirely account for the smaller clone size of Xrx1–EnR
clones.
Xrx1 supports proliferation without affecting cell fate
commitment
Although transfection of cdk2/cyclinA2 produced an increase in clone size that is comparable to that induced by
Xrx1, the role of Xrx1 does not seem to be restricted to
control of the cell cycle. Indeed, cotransfection of cdk2/
cyclinA2 cannot rescue the reduced clone size due to Xrx1–
EnR. This would suggest that, in the absence of normal Xrx1
function, cdk2/cyclinA2 are not limiting factors in retinal
cell proliferation or survival. Moreover, it also suggests that
Xrx1 can influence molecular pathways different than those
controlled by cdk2/cyclinA2. This hypothesis is also supported by the different effects exerted by Xrx1 and cdk2/
cyclinA2 on the differentiation of retinal cell types. Unlike
Xrx1 lipofection, the lipofection of cdk2/cyclinA2 generates
an alteration of the normal proportions of cell types,
namely, an increase in bipolar cells and a decrease in cones
and ganglion cells. In principle, cdk2/cyclinA2 could either
interact with other factors affecting neuronal cell fates or
influence neuronal fate commitment by altering the normal
timing of neuronal generation. Genes involved in the control of cell proliferation and differentiation, such as Xotch,
Delta, p27Xic1, and Xath5, were reported both to affect cell
differentiation and to modify the relative proportion of the
cell types of the Xenopus retina (Dorsky et al., 1995, 1997;
Kanekar et al., 1997; Ohnuma et al., 1999). In some instances, models have been proposed to explain the effect of
these genes on cell fate as a consequence of changes in the
“timing” of cell type generation (Kanekar et al., 1997;
Ohnuma et al., 1999; Moore et al., 2002). Then, a possible
interpretation of the results obtained after cdk2/cyclinA2
lipofection is that a prolonged proliferation of progenitors
during retinogenesis depletes multipotent precursors and
increases the number of neuronal precursors committed to
later fates. As a result, at the time when precursors eventually exit the cell cycle they would produce more late-type
neurons, such as bipolar neurons, and fewer early-type neurons, such as ganglion cells and cones.
We show that Xrx1 supports cell proliferation without
affecting cell fate commitment, since its overexpression
generates no alteration of the proportions of the different
retinal cell types. Different results were reported after overexpression of the Xrx1 mouse homologue, rax, which increases the number of Müller glial cells in transduced clones
(Furukawa et al., 2000). However, different types of retinal
progenitors were targeted in the two approaches, as retinal
progenitors in newborn mouse retina have a restricted cell
fate commitment and generate late-born cell types, mostly
rods, bipolar cells, and Müller glia (Furukawa et al., 2000).
Moreover, gliogenesis was recently shown to be differentially regulated in mouse and Xenopus retinas (reviewed by
Vetter and Moore, 2001), with key molecules like Notch
32
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
pression marks an early step of neuronal differentiation
common to multipotent retinal progenitors (Kanekar et al.,
1997). In the Xenopus retina, Xotch lipofection maintains
transfected cells as undifferentiated proliferating progenitors, while Xath5 increases the number of differentiated
ganglion cells. Thus, the capability of Xrx1 to induce the
expression of both genes would suggest Xrx1 sustains proliferating progenitors like those that generate ganglion cells,
that is, multipotent progenitors that are competent to differentiate into all retinal neurons. In summary, we propose that
Xrx1 acts as an intrinsic cue in supporting multipotency and
mitotic activity of embryonic retinal progenitors.
Potential role of Xrx1 in retinal stem cells
Fig. 5. Statistical analysis of retinal progenitors proliferation and differentiation after lipofection. Statistics on the results shown in Fig. 4 and 5. In
(A), the Y axis reports the average clone size after lipofection with different
combinations of vectors (X axis). Error bars represent SE. Clone count:
pCS2, n ⫽ 52; Xrx1, n ⫽ 24; Xrx–1Enr, n ⫽ 84; Bcl2, n ⫽ 15; Bcl2 ⫹
Xrx1–Enr, n ⫽ 44; cyclinA2/cdk2, n ⫽ 9; cyclinA2/cdk2 ⫹ Xrx1–Enr, n ⫽
33. Student’s t test was performed to evaluate clone size increase or
decrease compared with control (pCS2). Clone size was significantly increased by Xrx1, Bcl2, and cyclinA2/cdk2 (P ⬍ 0.001, P ⬍ 0.02, and P ⬍
0.02, respectively) and decreased by Xrx1–Enr and cyclinA2/cdk2 ⫹ Xrx1–
Enr (P ⬍ 0.001). (B) Percentage of specific retinal cell types, after lipofection of control (n ⫽ 1245 cells), Xrx1 (n ⫽ 1101 cells), and cyclinA2/
cdk2 (n ⫽ 1743 cells) vectors. Error bars represent SEM. The number of
cells lipofected with Xrx1–EnR was not sufficient for a statistical analysis
of the cell types. cyclinA2/cdk2 lipofection significantly increased the
number of bipolar cells (P ⬍ 0.02) and decreased the number of ganglion
cells and photoreceptors (P ⬍ 0.02) with respect to control, as evaluated by
Student’s t test.
and p27 playing different roles in Müller cell specification
in these two species (Dorsky et al., 1995; Onhuma et al.,
1999; Furukawa et al., 2000; Dyer and Cepko, 2001; Moore
et al., 2002). Our results can be explained considering that
Xrx1 supports multipotent progenitors. In fact, the results of
BrdU incorporation after lipofection indicate that Xrx1
overexpression delays the time of exit from the cell cycle of
cells fated to the GCL, namely, the multipotent retinal
precursors (see Results and Fig. 8). This supports the idea
that Xrx1-overexpressing progenitors generate early cell
types at a late time of retinogenesis. Interestingly, Xrx1
overexpression supports transcription of both Xotch and
Xath5, which are markers of proliferating and retina-committed progenitors, respectively. In particular, Xath5 ex-
We show that Xrx1 is also expressed in the dividing cells
of the CMZ, as other genes involved in the control of cell
proliferation (Perron et al., 1998). This observation suggests
that Xrx1 could also support proliferation of multipotent
stem cells of CMZ. The precise contribution of Xrx1 to the
capability of the stem cells of the CMZ to self-renew and
maintain their potential to generate all retinal cell types
remains to be established.
We speculate that Xrx1 could exert in the stem cells of
the adult CMZ the same effects observed in embryonic
progenitors. In a possible model, the effect of the intrinsic
Fig. 6. Analysis of cell type-specific markers in lipofected retinas. (A, B)
calbindin staining (red) on control (A) and cyclinA2/cdk2 (B) lipofected
cones (green, GFP reporter). Arrows point at calbindin-negative rods;
arrowheads indicate calbindin-positive cones where double staining with
GFP results in yellow color. cyclinA2/cdk2 lipofection reduces the percentage of cones compared with control. Cones account for 57% of control
lipofected photoreceptors (n ⫽ 181, SE ⫽ 6), 52% of Xrx1 lipofected
photoreceptors (n ⫽ 190, SE ⫽ 4.9, not shown), and 27% of cyclinA2/cdk2
lipofected photoreceptors (n ⫽ 102, SE ⫽ 8). The decrease in cones in
cyclinA2/cdk2 lipofected retinas is reflected in a decrease in the total
number of photoreceptors (see B). (C) Islet-1 staining (red) in the GCL of
Xrx1 lipofected retinas (green, GFP reporter). All Xrx1 transfected cells
(arrowheads) show yellow double staining. (D) R5 staining (red) of Xrx1
lipofected cells (green, GFP reporter). R5 labels Müller glia processes; note
that none of the GFP-positive Xrx1 transfected cells show yellow double
staining. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer
nuclear layer.
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
33
Fig. 7. Time of generation of ganglion cells after Xrx1 lipofection. Cross sections of stage 42 retina after a BrdU pulse given at stage 35–36. In blue, Hoechst
nuclear staining; in green, BrdU detection. (A) Cross section of a stage 42 control retina where the BrdU, incorporated by dividing cells at stage 35–36, is
never retained by cells of the GCL at the time of analysis. (B, C) Control lipofected cells. (D, E) Xrx1 lipofected cells. In (B) and (D), myc detection of
pCS2MT reporter (carrying a myc tag), cotransfected with control pCS2 (B) or Xrx1 (D), is red. Control lipofected cells never show BrdU incorporation in
the GCL layer (compare arrows in B, C). On the contrary, most of the Xrx1-lipofected cells in the GCL show BrdU incorporation (arrows in D, E). GCL,
ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.
Xrx1 signal would integrate with the external cues (see
Introduction), which would push cells to exit cell cycle and
choose the proper differentiation type in the mature central
retina, or keep them proliferating in the CMZ.
Experimental methods
bindin (1:500, Oncogene), and anti-R5 (1:1) (Drager et al.,
1984). Secondary antibodies: TRITC-conjugated donkey
anti-rabbit IgG (1:400, Sigma), TRITC-conjugated goat
anti-mouse IgG (1:100, Sigma).
BrdU in vivo administration, labeling, and detection
were carried out as described in Zuber et al. (1999) with
minor modifications. Detection of the BrdU antibody was
Embryos and histology
Induction of ovulation of females, in vitro fertilization,
and embryo culture were carried out as described by Newport and Kirschner (1982). Staging was according to Nieuwkoop and Faber (1994). For both in situ hybridization and
immunohistochemistry, embryos were fixed 2 h in MEMFA
at room temperature (Harland, 1991), cryoprotected with
30% sucrose in PBS O/N at 4°C, and stored at ⫺70°C until
cryosectioning.
In situ hybridization, immunohistochemistry, and BrdU
staining
Digoxigenin (DIG)-labeled antisense RNA probes were
generated for Xrx1 (Casarosa et al., 1997), Cyclin D1 (provided by Matt Cockerill, Xenopus Molecular Marker Resource), N-tubulin (Richter et al., 1988), Xotch (Dorsky et
al., 1995), X-Delta-1 (Chitnis et al., 1995), and Xath5
(Kanekar et al., 1997) as described in Harland (1991). In
situ hybridization on 12-␮m cryosections was carried out as
described in Kanekar et al. (1997). Signal detection with
either NBT/BCIP or Fast Red substrate (Roche) was performed according to the manufacturer’s instructions.
Immunostaining was performed on 12-␮m cryosections
using the following primary and secondary antibodies. Primary antibodies: anti-phosphohistone-H3 (1:200, Upstate
Biotechnology), anti-myc (9E10, 1:500, Sigma), anti-Islet-1
(1:100, Developmental Studies Hybridoma Bank), anti-cal-
Fig. 8. Model of action of Xrx1 on lipofected progenitors. The model
proposes how lipofection of retinal precursors with cyclinA2/cdk2 or with
Xrx1 could interfere with the normal timing of exit from the cell cycle and
with the time-dependent fate restriction. Colors represent the different
degrees of competence of retinal progenitors in generating the different
retinal cell types. At the beginning of normal retinogenesis, a single
“white” progenitor can generate all retinal cell types (as white is the sum
of all colors). As retinogenesis proceeds, the progenitor restricts its potential (it changes color). cyclinA2/cdk2 lipofection delays exit from the cell
cycle, thus resulting in a decreased number of ganglion cells and cones
generated by targeted progenitors. Conversely, Xrx1 lipofection delays the
time of withdrawal of progenitors from the cell cycle without restricting
their potential, so that they can produce early cell types (e.g., ganglion
cells) later than in normal retinogenesis.
34
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
performed by using FITC-conjugated sheep anti-mouse IgG
F(ab⬘)2 (1:100, Sigma) followed by blocking with goat
anti-mouse IgG F(ab⬘)2 fragment (1:100, Sigma). Myc tag
detection was then carried out as described above.
In all immunohistochemistry experiments, Hoechst
33258 (Sigma, final concentration 0.12 ␮g/ml) was added to
secondary antibody for counterstaining of cell nuclei.
cells per retina. When more than a cluster was detected in
the same retina, we scored them as different clones only
when they were well separated. As an additional criterion to
identify single clones, we considered the spatial distribution
of cells through serial sections. A clonal progeny usually
covered few serial sections of 12 ␮m. Serial collections
lacking even a single section were not considered in this
analysis.
Plasmids and embryo mRNA microinjection
Cell counts
All cDNAs used for either RNA in vitro transcription or
in vivo DNA lipofection were subcloned in the pCS2 vector.
The full-length cDNA sequence of human Bcl2 was provided by J.C. Martinou. The pCS2MT vector, carrying a
myc tag, is a kind gift from M. Zuber. Capped synthetic
RNAs were generated in vitro by SP6 transcription from
pCS2Xrx1 and pCS2Xrx1–EnR (Andreazzoli et al., 1999),
from pCS2mtXcyclinA2 and pCS2mtXcdk2 (provided by T.
Hunt; Paris et al., 1991; Howe et al., 1995) and from
pCS2GFP (provided by M. Zuber). The capability of the
dominant-negative Xrx1–EnR construct to antagonize endogenous Xrx1 function has been described (Andreazzoli et
al., 1999). The following amounts of mRNAs were injected
unilaterally at the four-cell stage: pCS2Xrx1, 100 pg;
pCS2mtXcyclinA2 and pCS2mtXcdk2, 1 ng. Five hundred
picograms of pCS2GFP was always coinjected as a tracer.
The injected embryos were collected under an epifluorescence microsope following detection of the GFP reporter
activity.
In vivo DNA lipofection
DNA was lipofected into the anterior neural region of
stage 17 embryos as previously described (Holt et al., 1990;
Dorsky et al., 1995). DNA was lipofected at a final concentration of 0.25 mg/ml with a 1:3 (w/w) DNA:DOTAP
(Roche) ratio. pCS2GFP or pCS2MT reporter vectors were
cotransfected with a 1:3 (w/w) ratio with respect to the
construct(s) of interest, to mark the transfected cells. Either
GFP-positive or myc-tagged cells were counted and the cell
types were identified based on their laminar position and
morphology, as described in Dorsky et al. (1995, 1997).
For the clone size analysis, the amount of lipofected
DNA was reduced to 0.05 mg/ml to obtain isolated clones of
transfected cells. At limiting DNA dilution, most retinas
contained no transfected cells or one single cluster, with an
average number of cells that did not vary when the percentage of transfected retinas became very low, thus suggesting
targeting of clonal progeny. When DNA was lipofected at
such low doses, each transfected cluster appeared compact
and oriented in a vertical column, as it is observed in clonal
analysis performed by using retroviral vectors, or direct
injection of enzymatic/fluorescent tracers (Price et al., 1987;
Turner and Cepko, 1987; Wetts and Fraser, 1988). When
analyzing clone size, we considered batches of injected
embryos with one or less than one cluster of transfected
For the analysis of mitosis in mRNA-injected embryos,
we counted phosphohistone-H3 positive cells in either the
control or injected side, as detected by the analysis of GFP
reporter activity. The analysis was carried out on 12-␮m
serial cross sections, by scoring the number of cells per
section in either retina or neural tube. Three or more sets of
experiments were carried out for each type of mRNA injected. Weighted averages of the number of cells per section
for both injected and control sides were then compared as
shown in Fig. 2.
Lipofected cells were considered to have a clonal origin
as described here and in Zuber et al. (1999). According to
previous studies (Dorsky et al., 1995, 1997; Kanekar et al.,
1997; Ohnuma et al., 1999; Zuber et al., 1999), we scored
the different cell types on the basis of their morphology and
laminar position. We sometimes confirmed cell identity
with specific antibodies such as: Islet-1 for ganglion cells
(Dorsky et al., 1997), R5 for Müller glia (Drager et al.,
1984), calbindin for cones (Chang and Harris, 1998). Cell
type percentages were calculated as weighted averages, after analysis of three or more independent sets of experiments (Kanekar et al., 1997; Zuber et al., 1999).
Acknowledgments
We are grateful to W.A. Harris, A. Viczian, and M.E. Zuber
for providing pCS2MT, pCS2GFP, pCS2mtXcyclinA2, and
pCS2mtXcdk2 plasmids and antibodies, and for teaching
S.C. the lipofection protocol. We acknowledge T. Hunt for
permission to use the full-length XcyclinA2 and Xcdk2 sequences. Xath5 probe is a kind gift of M. Cockerill. We
acknowledge W.A. Harris, M. Götz, P. Malatesta, and R.
Vignali for insightful comments on the manuscript and for
helpful discussions. Finally, we are indebted to M. Fabbri,
D. de Matienzo, and S. di Maria for technical assistance.
This work was supported by grants from M.U.R.S.T and
from EEC, Biotechnology Program (Grant QLRT-200001460).
References
Andreazzoli, M., Gestri, G., Angeloni, D., Menna, E., Barsacchi, G., 1999.
Role of Xrx1 in Xenopus eye and anterior brain development. Development 126, 2451–2460.
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
Artavanis-Tsakonas, S., Rand, M.D., Lake, R.J., 1999. Notch signaling:
cell fate control and signal integration in development. Science 284,
770 –776.
Bahr, M., 2000. Live or let die: retinal ganglion cell death and survival
during development and in the lesioned adult CNS. Trends Neurosci.
10, 483– 490.
Baker, N.E., 2000. Notch signaling in the nervous system: pieces still
missing from the puzzle. Bioessays 22, 264 –273.
Brown, N.L., Patel, S., Brzezinski, J., Glaser, T., 2001. Math5 is required
for retinal ganglion cell and optic nerve formation. Development 128,
2497–2508.
Casarosa, S., Andreazzoli, M., Simeone, A., Barsacchi, G., 1997. Xrx1, a
novel Xenopus homeobox gene expressed during eye and pineal gland
development. Mech. Dev. 61, 187–198.
Cepko, C.L., 1999. The roles of intrinsic and extrinsic cues and bHLH
genes in the determination of retinal cell fates. Curr. Opin. Neurobiol.
9, 37– 46.
Chang, W.S., Harris, W.A., 1998. Sequential genesis and determination
of cone and rod photoreceptors in Xenopus. J. Neurobiol. 35, 227–
244.
Chen, D.F., Schneider, G.E., Martinou, J.C., Tonegawa, S., 1997. Bcl-2
promotes regeneration of severed axons in mammalian CNS. Nature
385, 434 – 439.
Chitnis, A., Henrique, D., Lewis, J., Ish-Horowicz, D., Kintner, C., 1995.
Primary neurogenesis in Xenopus embryos regulated by a homologue
of the Drosophila neurogenic gene Delta. Nature 375, 761–766.
Chow, R.L., Altmann, C.R., Lang, R.A., Hemmati-Brivanlou, A., 1999.
Pax6 induces ectopic eyes in a vertebrate. Development 126, 4213–
4222.
Chuang, J.C., Raymond, P.A., 2001. Zebrafish genes rx1 and rx2 help
define the region of forebrain that gives rise to retina. Dev. Biol. 231,
13–30.
Dorsky, R.I., Rapaport, D.H., Harris, W.A., 1995. Xotch inhibits cell
differentiation in the Xenopus retina. Neuron 14, 487– 496.
Dorsky, R.I., Chang, W.S., Rapaport, D.H., Harris, W.A., 1997. Regulation
of neuronal diversity in the Xenopus retina by Delta signalling. Nature
385, 67–70.
Drager, U.C., Edwards, D.L., Barnstable, C.J., 1984. Antibodies against
filamentous components in discrete cell types of the mouse retina.
J. Neurosci. 4, 2025–2042.
Dyer, M.A., Cepko, C.L., 2001. p27 kip1 and p57 kip2 regulates proliferation in distinct retinal progenitor cell populations. J. Neurosci. 21,
4259 – 4271.
Ezzeddine, Z.D., Yang, X., DeChiara, T., Yancopoulos, G., Cepko, C.L.,
1997. Postmitotic cells fated to become rod photoreceptors can be
respecified by CNTF treatment of the retina. Development 124, 1055–
1067.
Furukawa, T., Kozak, C.A., Cepko, C.L., 1997. rax, a novel paired-type
homeobox gene, shows expression in the anterior neural fold and
developing retina. Proc. Natl. Acad. Sci. USA 94, 3088 –3093.
Furukawa, T., Mukherjee, S., Bao, Z.Z., Morrow, E.M., Cepko, C.L., 2000.
rax, Hes1, and notch1 promote the formation of Muller glia by postnatal retinal progenitor cells. Neuron 26, 383–394.
Gonzalez-Hoyuela, M., Barbas, J.A., Rodriguez-Tebar, A., 2001. The autoregulation of retinal ganglion cell number. Development 128, 117–
124.
Harland, R.M., 1991. In situ hybridization: an improved whole-mount
method for Xenopus embryos. Methods Cell Biol. 36, 685– 695.
Harris, W.A., 1997. Cellular diversification in the vertebrate retina. Curr.
Opin. Genet. Dev. 7, 651– 658.
Harris, W.A., Perron, M., 1998. Molecular recapitulation: the growth of the
vertebrate retina. Int. J. Dev. Biol. 42, 299 –304.
Hendzel, M.J., Wei, Y., Mancini, M.A., Van Hooser, A., Ranalli, T.,
Brinkley, B.R., Bazett-Jones, D.P., Allis, C.D., 1997. Mitosis-specific
phosphorylation of histone H3 initiates primarily within pericentromeric heterochromatin during G2 and spreads in an ordered fashion
35
coincident with mitotic chromosome condensation. Chromosoma 106,
348 –360.
Henrique, D., Hirsinger, E., Adam, J., Le Roux, I., Pourquie, O., IshHorowicz, D., Lewis, J., 1997. Maintenance of neuroepithelial progenitor cells by Delta–Notch signalling in the embryonic chick retina.
Curr. Biol. 7, 661– 670.
Hensey, C., Gautier, J., 1997. A developmental timer that regulates apoptosis at the onset of gastrulation. Mech. Dev. 69, 183–195.
Holt, C.E., Garlick, N., Cornel, E., 1990. Lipofection of cDNAs in the
embryonic vertebrate central nervous system. Neuron 2, 203–214.
Howe, J.A., Howell, M., Hunt, T., Newport, J.W., 1995. Identification of
a developmental timer regulating the stability of embryonic cyclin A
and a new somatic A-type cyclin at gastrulation. Genes Dev. 9, 1164 –
1176.
Jensen, A.M., Wallace, V.A., 1997. Expression of Sonic hedgehog and its
putative role as a precursor cell mitogen in the developing mouse
retina. Development 124, 363–371.
Kageyama, R., Nakanishi, S., 1997. Helix–loop– helix factors in growth
and differentiation of the vertebrate nervous system. Curr. Opin. Genet.
Dev. 7, 659 – 665.
Kanekar, S., Perron, M., Dorsky, R., Harris, W.A., Jan, L.Y., Jan, Y.N.,
Vetter, M., 1997. Xath5 participates in a network of bHLH genes in the
developing Xenopus retina. Neuron 19, 981–994.
Lillien, L., Wancio, D., 1998. Changes in epidermal growth factor
receptor expression and competence to generate glia regulate timing
and choice of differentiation in the retina. Mol. Cell Neurosci. 10,
296 –308.
Livesey, F.J., Cepko, C.L., 2001. Vertebrate neural cell-fate determination:
lessons from the retina. Nat. Rev. Neurosci. 2, 109 –118.
Loosli, F., Winkler, S., Burgtorf, C., Wurmbach, E., Ansorge, W., Henrich,
T., Grabher, C., Arendt, D., Carl, M., Krone, A., Grzebisz, E., Wittbrodt, J., 2001. Medaka eyeless is the key factor linking retinal determination and eye growth. Development 128, 4035– 4044.
Mathers, P.H., Grinberg, A., Mahon, K.A., Jamrich, M., 1997. The Rx
homeobox gene is essential for vertebrate eye development. Nature
387, 603– 607.
McFarlane, S., Zuber, M.E., Holt, C.E., 1998. A role for the fibroblast
growth factor receptor in cell fate decisions in the developing vertebrate retina. Development 125, 3967–3975.
Moore, K.B., Schneider, M.L., Vetter, M.L., 2002. Posttranslational mechanisms control the timing of bHLH function and regulate retinal cell
fate. Neuron 34, 183–195.
Newport, J., Kirschner, M., 1982. A major developmental transition in
early Xenopus embryos. Cell 30, 687– 696.
Nieuwkoop, P.D., Faber, J., 1994. Normal Table of Xenopus laevis. Garland, New York.
Ohnuma, S., Philpott, A., Wang, K., Holt, C.E., Harris, W.A., 1999.
p27Xic1, a Cdk inhibitor, promotes the determination of glial cells in
Xenopus retina. Cell 99, 499 –510.
Ohnuma, S., Philpott, A., Harris, W.A., 2001. Cell cycle and cell fate in the
nervous system. Curr. Opin. Neurobiol. 11, 66 –73.
Paris, J., Guellec, R.L., Couturier, A., Guellec, K.L., Omilli, F., Camonis,
J., MacNeill, S., Philippe, M., 1991. Cloning by differential screening
of a Xenopus cDNA coding for a protein highly homologous to cdc2.
Proc. Natl. Acad. Sci. USA 88, 1039 –1043.
Perron, M., Kanekar, S., Vetter, M.L., Harris, W.A., 1998. The genetic
sequence of retinal development in the ciliary margin of the Xenopus
eye. Dev. Biol. 199, 185–200.
Perron, M., Opdecamp, K., Butler, K., Harris, W.A., Bellefroid, E.J., 1999.
X-ngnr-1 and Xath3 promote ectopic expression of sensory neuron
markers in the neurula ectoderm and have distinct inducing properties
in the retina. Proc. Natl. Acad. Sci. USA 96, 14996 –15001.
Price, J., Turner, D., Cepko, C., 1987. Lineage analysis in the vertebrate
nervous system by retrovirus-mediated gene transfer. Proc. Natl. Acad.
Sci. USA 84, 156 –160.
36
S. Casarosa et al. / Molecular and Cellular Neuroscience 22 (2003) 25–36
Richter, K., Grunz, H., Dawid, I.B., 1988. Gene expression in the embryonic nervous system of Xenopus laevis. Proc. Natl. Acad. Sci. USA. 85,
8086 – 8090.
Turner, D.L., Cepko, C., 1987. A common progenitor for neurons and glia
persists in rat retina late in development. Nature 228, 131–136.
Vetter, M.L., Moore, K.B., 2001. Becoming glial in the neural retina. Dev.
Dyn. 221, 146 –153.
Wetts, R., Fraser, S.E., 1988. Multipotent precursors can give rise to all
major cell types of the frog retina. Science 239, 1142–1145.
Winkler, S., Loosli, F., Henrich, T., Wakamatsu, Y., Wittbrodt, J., 2000.
The conditional medaka mutation eyeless uncouples patterning and
morphogenesis of the eye. Development 127, 1911–1919.
Zhou, X., Hollemann, T., Pieler, T., Gruss, P., 2000. Cloning and expression of XSix3, the Xenopus homologue of murine six3. Mech. Dev. 91,
327–330.
Zuber, M.E., Perron, M., Philpott, A., Bang, A., Harris, W.A., 1999.
Giant eyes in Xenopus laevis by overexpression of Xoptx2. Cell 98,
341–352.