Experimental Eye Research 84 (2007) 577e590
www.elsevier.com/locate/yexer
Characterization of early retinal progenitor microenvironment:
Presence of activities selective for the differentiation of retinal
ganglion cells and maintenance of progenitors
Ganapati V. Hegde, Jackson James, Ani V. Das, Xing Zhao,
Sumitra Bhattacharya, Iqbal Ahmad*
Department of Ophthalmology and Visual Sciences, University of Nebraska Medical Center, Omaha, NE 68198, USA
Received 12 June 2006; accepted in revised form 9 November 2006
Available online 16 January 2007
Abstract
The maintenance and differentiation of retinal progenitors take place in the context of the microenvironment in which they reside at a given time
during retinal histogenesis. To understand the nature of the microenvironment in the developing retina, we have examined the influence of
activities present during the early stage of retinal histogenesis on enriched retinal progenitors, using the neurosphere model. Early and late retinal
progenitors, enriched as neurospheres from embryonic day 14 (E14) and E18 rat retina, respectively, were cultured in embryonic day 3 (E3)
chick retinal conditioned medium, simulating the microenvironment present during early retinal histogenesis. Examination of the differentiation
and proliferation of retinal progenitors revealed that the early microenvironment contains at least three regulatory activities, which are partitioned in different size fractions of the conditioned medium with different heat sensitivity. First, it is characterized by activities, present in
heat stable <30 kDa fraction, that promote the differentiation of retinal ganglion cells (RGCs), the early born neurons. Second, it contains
activities, present in heat-sensitive >30 kDa fraction, that regulate the number of early born neurons and maintain the pool of retinal progenitors.
Third, it possesses activities, present in heat-sensitive <30 kDa fraction, that prevent the premature differentiation of early retinal progenitors
into the late born neurons. Thus, our observations demonstrate the regulatory influence of microenvironment on the maintenance and differentiation of retinal progenitors and establish neurospheres as a viable model system for the examination of such influences.
Ó 2006 Elsevier Ltd. All rights reserved.
Keywords: microenvironment; niche; extrinsic factors; RGC; rod photoreceptor; retina; progenitor; stem cells
1. Introduction
It is becoming increasingly evident that the regulation of
stem cells/progenitors takes place in the context of their
microenvironment. Therefore, understanding the biology of
stem cells/progenitors and their successful clinical use will
require information about the properties of their microenvironment. The developing retina, which has been successfully used
to provide insights into cellular and molecular mechanisms of
* Corresponding author. Tel.: þ1 402 559 4091; fax: þ1 402 559 3251.
E-mail address: iahmad@unmc.edu (I. Ahmad).
0014-4835/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.exer.2006.11.012
neurogenesis, can serve as a model for characterization of the
microenvironment because retinal cell types, which are limited
in number, are known to be generated in response to changing
environment. For example, different experimental approaches
like lineage analysis and cell ablation studies have demonstrated that retinal progenitors are multi-potential, and the decision taken by these cells to differentiate into a particular
retinal cell type depends on local cellecell interactions (Altshuler et al., 1991). Several in vitro studies have suggested that
some of the known and unknown factors may mediate celle
cell interactions during retinal neurogenesis (Yang, 2004).
Also, evidence emerged that, besides diffusible factors,
cellecell interactions mediated by membrane-bound receptors
578
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
and ligand interactions, such as those exemplified by Notch
signaling, play a critical role in retinal neurogenesis (Livesey
and Cepko, 2001; Ahmad et al., 2004).
These cell-extrinsic cues, elaborated by progenitors,
precursors, differentiated neurons and glia, constitute the milieu of the microenvironment. However, how the activities
within the microenvironment maintain retinal progenitors
and influence the outcomes of their differentiation remain unknown. Given the evolutionarily conserved temporal nature of
cell-fate specification during the two stages of retinal histogenesis (Rapaport et al., 2004), it is likely that at any given
time of retinal histogenesis, activities within the microenvironment will be permissive for the differentiation of progenitors into predominant cell types of that stage, and inhibiting
for those that are born later, and yet maintain progenitor pool.
To test this premise, we examined the influence of microenvironment, created by the presence of cells in the early stage
of retinal histogenesis, on the differentiation of retinal progenitors into early and late born retinal neurons, using the
neurosphere model (James et al., 2003; Reynolds and Rietze,
2005).
We have used the neurosphere model of retinal progenitors
for the following reasons: although not entirely homogeneous,
w90% of cells in neurospheres are proliferative, as demonstrated by their ability to incorporate BrdU and express pan
neural (i.e., Nestin and Musashi) as well as retinal progenitor
(i.e., Rx, Pax6 and Chx10) markers (Ahmad et al., 2004).
Examination and interpretation of results regarding cell proliferation and differentiation in neurosphere are less ambiguous
than in explant culture due to limited cellular heterogeneity.
The reason for using E3 chick retinal cells to simulate the
environment representing the stage of early histogenesis included the fact that stages E3eE5 in chick retina represent
a period characterized by the birth of early born neurons
such as the RGCs (Prada et al., 1991). Since RGCs are the first
retinal cell types to differentiate regardless of species, cellextrinsic cues that regulate their differentiation are likely to
be conserved and read by rat retinal progenitors (James
et al., 2003). Besides, tissues are less of a limitation with E3
chick retina than with rat retina at E13eE14, the peak of
RGC generation.
Three different activities were discerned in the early retinal
progenitor microenvironment. These were partitioned into different size fractions with differential heat sensitivity and influence on cell differentiation and proliferation. These
constituted (1) activities that promoted the differentiation of
retinal progenitors into early born neurons, such as RGCs
(2) activities that regulated the number of early born neurons
and maintained the pool of retinal progenitors and (3) activities that prevented the premature differentiation of early retinal
progenitors into late born neurons, such as rod photoreceptors,
before the onset of the late stage of histogenesis. Together,
these cell-extrinsic activities are likely to regulate the efficiency and fidelity of cell differentiation during the early stage
of retinal histogenesis while maintaining a pool of retinal progenitors for subsequent rounds of differentiation. In addition,
our observations present the neurospheres as a viable model
for examining the influence of microenvironment on stem
cells/progenitors in controlled conditions.
2. Materials and methods
2.1. Enrichment and culture of retinal
progenitors as neurospheres
Timed-pregnant E14 and E18 SpragueeDawley rats were
obtained from Sasco (Wilmington, MA), and the gestation
day was confirmed by the morphological examination of embryos (Christie, 1964). Fertilized hen eggs (Hoovers, Omaha,
NE) were incubated in a humidified incubator at 38 C for 3
days, and embryonic stages were confirmed by morphological
examination (Hamburger and Hamilton, 1951). Embryos were
harvested at appropriate stages, and eyes were enucleated. Retina was dissected out, and dissociated by trypsin digestion, as
previously described (Ahmad et al., 1999). Rat retinal dissociates were cultured in retinal culture medium (RCM; DMEM/
F12, 1 N2 supplement, 2 mM L-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin) supplemented with either
the combination of EGF (20 ng/ml; Collaborative Research),
FGF2 (10 ng/ml; Collaborative Research) and 0.1% FBS for
E14 cells, or with EGF (20 ng/ml) for E18 cells, at a density
of 105 cells/cm2 for 4e6 days to generate neurospheres.
Neurospheres were exposed to 10 mM 5-bromo-20 -deoxyuridine (BrdU) (Sigma) for the final 48 h of culture to tag the
dividing cells. BrdU-tagged neurospheres, obtained from
E14/E18 retina, were plated on poly-D-lysine-(250 mg/ml)
and laminin-(5 mg/ml) coated 12 mm glass cover slips, and
co-cultured either across a 0.4 mm pore size semi-permeable
membrane (Millipore, Bedford, MA) with E3 chick retinal/
forebrain cells, or in the presence of E3CM/RCM supplemented with 1% FBS for 5e6 days. To ascertain the effects
of E3CM on the generation of neurospheres, E14 retinal
dissociates were cultured in the presence of 1% FBS or
E3CM fractions in a 96-well plate at a density of 10,000
cells/well for 2e3 days. The number of neurospheres generated/cm2 was quantified.
2.2. Preparation, fractionation and
treatment of conditioned medium
The dissociated E3 chick retinal cells were cultured in
RCM supplemented with 1% FBS for 3 days. The supernatant
containing enriched activities was separated by centrifugation
(5000 rpm, 4 C, 10 min), and defined as E3 chick retinal conditioned medium (E3CM). E3CM and RCM containing 1%
FBS were fractionated into >30 kDa and <30 kDa fractions
using 30 kDa centricon tubes according to the manufacturer’s
instruction (Millipore). The fractionated E3CM was reconstituted by adding the missing fraction from fractionated RCM
containing 1% FBS as shown in the respective flow charts
(Figs. 4A and 7A). To determine the heat stability of activities,
E3CM fractions were subjected to heat treatment (70 C for
30 min), and the heat-treated E3CM fractions were
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
reconstituted as shown in the respective flow charts (Figs. 5A,
E and 8A).
2.3. Explants culture
E14 retinal explants were cultured in 1% FBS or E3CM on
a 0.4 mm pore size semi-permeable membrane for 5 days. Explants were either fixed for immunohistochemical or frozen for
RT-PCR analyses as previously described (James et al., 2003).
2.4. Immunofluorescence analysis
Immunocytochemical analysis was carried out for the
detection of cell specific markers and BrdU as previously
described (Ahmad et al., 1999). Briefly, 4% paraformaldehyde-fixed cells were incubated in PBS containing 5% normal
goat serum (NGS) and 0.2/0.4% Triton X-100 followed by
an overnight incubation in Brn3b (1:300, BabCo)/Nestin
(1:4, DSHB) and BrdU (1:100) antibodies at 4 C. Since
BrdU immunocytochemistry interfered with fluorescence
specific for opsin-immunoreactivity, immunocytochemistry to
detect opsin was carried alone. TUNEL staining was carried
out as per the manufacturer’s instruction (Promega). Immunohistochemical analysis of explants was carried out on paraformaldehyde-fixed, 10-mm cryostat sections as previously described
(Ahmad et al., 1998), and counterstained with DAPI. Cells
were examined for epifluorescence following incubation in
species specific IgG conjugated to Cy3/FITC. Images were
captured using cooled CCD-camera (Princeton Instruments)
and Open lab software (Improvision Inc.). To determine the
percentage of specific cell types in a particular condition,
579
number of BrdUþ and cell specific antigen positive cells
were counted in 10e15 randomly selected fields in twoethree
different cover slips. Each experiment was repeated at least
three times. Values are expressed as mean SEM. Statistical
analysis was done using Student’s t-test to determine the
significance of the differences between various conditions.
2.5. RT-PCR analysis
Total RNA was isolated from neurospheres in differentiating conditions using RNA isolation kit (Qiagen RNeasy mini
kit), and the genomic DNA was removed by RNase-free DNase I digestion as per the manufacturer’s instruction (Ambion
DNA free kit). RNA (2 mg) was transcribed to cDNA using
superscript reverse transcriptase as previously described
(James et al., 2003). The cDNA was normalized against
a house keeping gene, b-actin, whose primers span an intron,
which was used to detect the presence of traces of genomic
DNA even after DNase digestion. To detect any cDNA contamination in other reagents in the PCR mix, the amplification
was carried out using water instead of cDNA. We were unable
to do a negative RT because of limited amount of samples.
Specific transcripts were amplified with gene specific forward
and reverse primers (Table 1) using a step-cycle program as
previously described (James et al., 2003). PCR amplification
was carried for 25e30 cycles. Products were visualized by
ethidium bromide staining after electrophoresis on a 2% agarose gel. The bands corresponding to specific transcripts, were
scanned and normalized against those corresponding to bactin, amplified from the same batch of cDNA. Mean fold
change for the respective genes in the experimental groups
Table 1
List of primers used for RT-PCR analyses
Gene name
GenBank accession number
Primer sequences
Temperature ( C)
Product size (bp)
b-Actin
XM_037235
50
548
Ath5
AF071223
52
231
Brn3b
AF390076
60
141
RPF1
XM_344604
56
359
Thy1
NM_012673
58
415
Mash1
XM_227409
56
393
Opsin
U22180
64
422
Rhodopsin kinase
U63971
56
238
Arrestin
M60737
56
324
Syntaxin 1
NM_053788
60
385
S-opsin
NM_031015
58
444
Chx10
L34808
(F) 50 -GTGGGGCGCCCCAGGCACCA-30
(R) 50 -CTCCTTAATGTCACGCACGATTTC-30
(F) 50 -TGGGG(I)CA(GA)GA(CT)AA(GA)AA(GA)-30
(R) 50 -CAT(I)GG(GA)AA(I)GG(CT)TC(I)GG(CT)TG-30
(F) 50 -GGCTGGAGGAAGCAGAGAAATC-30
(R) 50 -TTGGCTGGATGGCGAAGTAG-30
(F) 50 -TTTCAGGGGATTCTGGTGTGC-30
(R) 50 -CGCTTTTTGGATGGCTCACTC-30
(F) 50 -TGCCTGGTGAACAGAACCTT-30
(R) 50 -TCACAGAGAAATGAAGTCCGTGGC-30
(F) 50 -GCCTTGCTGTGCTTGAACAAAC-30
(R) 50 -CACTGGGAATGGTCTCATCACTATG-30
(F) 50 -CATGCAGTGTTCATGTGGGA-30
(R) 50 -AGCAGAGGCTGGTGAGCATG-30
(F) 50 -GCTGAACAAGAAGCGGCTGAAG-30
(R) 50 -TGCTGTGTAGTAGATGGCTCGTGG-30
(F) 50 -GCTCGTGAAGGGGAAGAAAGTG-30
(R) 50 -TCTCTGATGTCTGTGGCAAATGC-30
(F) 50 -AAGAGCATCGAGCAGAGCATC-30
(R) 50 -CATGGCCATGTCCATGAACAT-30
(F) 50 -GGATACTTCCTCTTTGGTCGCC-30
(R) 50 -CCGTTCAGCCTTTTGTGTCG-30
(F) 50 -TCCGATTCCGAAGATGTTTCC-30
(R) 50 -GACTTGAGGATAGACTCTGGCAGG-30
56
350
F: forward primer, R: reverse primer.
580
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
was calculated with respect to controls (fold value ¼ 1).
Values are expressed as mean fold change SEM from three
different experiments. Statistical analysis was done using
Student’s t-test to determine the significance between different
conditions.
3. Results
3.1. Activities elaborated by cells representing the early
stage of retinal histogenesis regulate the differentiation
of retinal progenitors into early born neurons
To demonstrate the influence of the microenvironment on
retinal progenitors during the generation of specific retinal
cell types, BrdU-tagged E14 neurospheres were cultured in
RGC differentiating conditions, defined by the presence of
E3CM and absence of mitogens. Controls include E14 neurospheres, similarly cultured in the presence of 1% FBS, minus
mitogens. In this condition, neurospheres, characterized by
BrdUþ and Nestinþ cells (Fig. 1AeD), spread out on the
surface and differentiate (Fig. 1EeH). The differentiation of
early retinal progenitors along RGC lineage was examined
by the expression of Brn3b, a Pou domain transcriptional
factor, known to regulate RGC differentiation and maturation
(Xiang et al., 1993). In addition, levels of transcripts corresponding to Ath5, a bHLH transcription factor that plays
a role in the specification of RGCs (Brown et al., 2001),
RPF1, another Pou domain regulator of RGC differentiation
(Zhou et al., 1996), and Thy1, a marker of matured RGCs
(Barnstable and Drager, 1984) were examined. We observed
a w2-fold increase in the proportion of BrdUþ and Brn3bþ
cells when E14 neurospheres were cultured in E3CM, compared to those in FBS controls, suggesting a positive influence
of E3CM on the differentiation of early retinal progenitors
along RGC lineage (Fig. 1EeM). RT-PCR analysis revealed
that levels of Ath5, Brn3b, RPF1 and Thy1 transcripts were increased in neurospheres exposed to E3CM as compared to
those in controls (Fig. 1N, O), corroborating results obtained
by immunocytochemical analysis, and suggesting that RGC
differentiation in neurospheres involved normal mechanism.
The specificity of E3CM influence on retinal progenitors
was determined as follows: we examined the effects of
E3CM on the expression of s-cone opsin and syntaxin 1,
markers for S-cone photoreceptors and amacrine cells, respectively (Fig. 1N, O). As expected, levels of s-cone opsin
transcripts were higher in E3CM-exposed neurospheres than
in controls, given the fact that cones birth date closely parallels
that of RGCs (Rapaport et al., 2004) and E3CM may contain
cone-promoting activities as it is obtained from cone-rich
retina. A slight difference (w0.9-fold) in levels of syntaxin 1
transcripts was observed between the two groups, suggesting
that E3CM may not influence the differentiation of amacrine
cells whose temporal generation lags that of RGCs and cone
photoreceptors (Rapaport et al., 2004). To ascertain that the
influence on the proportion of Brn3bþ cells was attributed
directly to E3 chick retinal cells, E14 neurospheres were
co-cultured in the presence of these cells and cells derived
from E3 chick forebrain across a semi-permeable membrane.
We observed a similar increase in the proportion of Brn3bþ
in the presence of E3 chick retinal cells as observed with
E3CM, compared to FBS controls (Fig. 1P). No significant
change in the proportion of Brn3bþ cells was observed
when E14 neurospheres were co-cultured with E3 chick
forebrain cells (Fig. 1P). Together, these observations suggested a positive influence of these cells and factors elaborated
by them on the differentiation of E14 retinal progenitors into
RGCs.
To address the issue that the observed effects of extrinsic
factors on cell differentiation could be a function of dissociated cell culture, we examined the effects of E3CM on RGC
differentiation in E14 retinal explants. In retinal explants the
dimensionality and cell interaction of the intact retina are
maintained and cells are born in a spatio-temporal order, comparable to that in vivo (Sparrow et al., 1990). The proportion
of Brn3bþ cells and the relative levels of Brn3b transcripts
were higher in explants in the presence of E3CM than in those
cultured in the presence of FBS only (Fig. 2AeK), thus corroborating the results obtained using the neurosphere culture
paradigm (Fig. 1).
3.2. Activities elaborated by cells representing the early
stage of retinal histogenesis regulate the differentiation
of retinal progenitors into late born neurons
To know whether or not the early retinal microenvironment
is selective for early born neurons, we examined the effects of
E3 chick retinal cells on the differentiation of the retinal progenitors along rod photoreceptor lineage. Birth dating analyses
have shown that a small subset of rod photoreceptors is generated at E13/14 (Morrow et al., 1998; Rapaport et al., 2004).
We hypothesized that the environment of early retinal neurogenesis may not be conducive for the generation of late born
neurons. We assumed that the full extent of the early microenvironment’s inhibitory influences on rod photoreceptor differentiation may not be apparent in the context of early retinal
progenitors due to the relative lack of their competence to generate late born neurons. Therefore, we first tested the hypothesis on late retinal progenitors, enriched from E18 rat retina as
neurospheres (E18 neurospheres), that preferentially generate
rod photoreceptors (Ahmad et al., 1999; Bhattacharya et al.,
2003). We observed a significant decrease in the proportion
of opsinþ cells in E18 neurospheres, cultured in the presence
of E3CM, compared to those in FBS controls (6.5 0.4% vs.
9.8 0.4%; p < 0.001, Fig. 3AeI). Similarly, levels of
transcripts corresponding to photoreceptor regulator, Mash1
(Ahmad et al., 1998) and phenotype specific markers, opsin,
arrestin and rhodopsin kinase were also observed to be
decreased in E18 neurospheres in the presence of E3CM, compared to those in controls (Fig. 3J, K). No difference in levels
of transcripts characteristic of bipolar cells (Chx10) was
observed between the two groups, demonstrating the cell
type-specific influence of E3CM. Together, these observations
suggested that E3CM or activities present in early microenvironment were inhibitory to rod photoreceptor differentiation.
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
581
Fig. 1. Activities present in E3CM promote the differentiation of retinal progenitors into RGCs. An example of E14 neurosphere, cultured in proliferating
conditions containing BrdU, consisted of BrdUþ and Nestinþ cells (AeD). Such neurospheres, pre-exposed to BrdU, were cultured in the presence of E3CM
or 1% FBS (controls) for 5 days followed by immunocytochemical and RT-PCR analyses of cell type-specific regulators and markers. BrdU-tagged E14 neurospheres contained significantly more BrdUþ and Brn3bþ cells (arrows) in the presence of E3CM (EeH, M) than in controls (IeL, M). Levels of transcripts corresponding to RGCs (Ath5, Brn3b, RPF1 and Thy1) and S-cone photoreceptors (s-cone opsin) were higher in E14 neurospheres, cultured in the presence of E3CM
(N, lane 2; O) than in controls (N, lane 1). A slight decrease in levels of syntaxin 1 transcripts between the two groups was observed. There were significantly more
BrdUþ and Brn3bþ cells when E14 neurospheres were co-cultured across the membrane with E3 chick retinal cells, as compared to those co-cultured with either
E3 chick forebrain (FB) cells or FBS (P). Data are expressed as mean SEM from triplicate cultures of three different experiments. 100 for AeD and 200 for
EeL (***p < 0.001).
To address the issue if these activities influenced the nascent
rod differentiation during early histogenesis (Morrow et al.,
1998), E14 neurospheres were cultured in the presence of
E3CM, followed by the examination of opsin expression.
Our attempt to detect opsin immunoreactivities in cells in
E14 neurospheres was not successful, which might be attributed to the fact that few rod photoreceptors are generated by
early retinal progenitors, and that there is a lag time between
the birth and expression of opsin protein (Morrow et al., 1998).
However, RT-PCR analysis, being more sensitive in detecting
gene expression, could detect opsin transcripts, indicating rod
photoreceptor differentiation (Fig. 3L). A comparison of opsin
expression revealed that the levels of opsin transcripts were
lower in E14 neurospheres in the presence of E3CM, compared to those in FBS controls (Fig. 3L, M), suggesting an
inhibitory effect of E3CM on early retinal progenitors’
differentiation along rod photoreceptor lineage.
3.3. Activities present in heat stable <30 kDa and heat
labile >30 kDa E3CM fractions promote and inhibit the
differentiation of retinal progenitors into RGCs,
respectively
To characterize the regulatory activities present during
early retinal histogenesis, we fractionated E3CM into two
different size fractions (i.e., <30 kDa and >30 kDa), based
582
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
Fig. 2. Activities present in E3CM promote RGC differentiation in explants culture. E14 retinal explants were cultured in the presence of E3CM or 1% FBS
(controls) followed by immunohistochemical and RT-PCR analyses of Brn3b expression. There were significantly more Brn3bþ cells when retinal explants
were cultured in presence of E3CM (AeD, I) than in controls (EeH, I). There was a relative increase in the levels of Brn3b transcripts when retinal explants
were cultured in presence of E3CM (J, lane 2; K), compared to those in controls (J, lane 1). OR ¼ outer retina; GCL ¼ ganglion cell layer. Data are expressed
as mean SEM from triplicate cultures of three different experiments. 100 (***p < 0.001).
on the fact that many reported factors that influence cell
differentiation have molecular weight less (e.g. FGF1,
FGF2, CNTF) or more (e.g. Shh, EGF, GDF11, Wnt2b)
than 30 kDa (Fig. 4A). E14 neurospheres were cultured in
reconstituted fractions as described in Fig. 4A. The proportion of BrdUþ and Brn3bþ cells increased significantly in
<30 kDa E3CM fraction, compared to those in >30 kDa
E3CM fraction or FBS controls (Fig. 4BeN, Table 2). Levels
of Ath5 and Brn3b transcripts were similarly higher in the
presence of <30 kDa E3CM fraction compared to those in
>30 kDa E3CM fraction or FBS controls (Fig. 4O, P). Together, these observations suggested that the RGC promoting
activities of the E3CM reside in <30 kDa E3CM fraction.
Both immunocytochemical and RT-PCR analyses of the
reconstitution of fraction experiments revealed that
>30 kDa E3CM fraction contained activities that inhibited
RGC differentiation. For example, the proportion of BrdUþ
and Brn3bþ cells was significantly lower in the presence of
>30 kDa E3CM fraction than in FBS controls (Fig. 4FeI,
Table 2). Similarly, levels of Ath5 and Brn3b transcripts
were relatively lower in the presence of >30 kDa E3CM
fraction as compared to FBS controls (Fig. 4O, P). The presence of RGC inhibiting activities in >30 kDa E3CM fraction
explained the observation that the proportion of BrdUþ and
Brn3bþ cells was relatively higher in <30 kDa E3CM
fraction than in un-fractionated E3CM (47 1.4% vs.
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
583
Fig. 3. Activities present in E3CM inhibit the differentiation of retinal progenitors into rod photoreceptors. E18 neurospheres were cultured in the presence of
E3CM or 1% FBS (controls) for 5 days, followed by immunocytochemical and RT-PCR analyses cell type-specific regulators and markers expression. There
were significantly fewer opsinþ cells (arrows) when neurospheres were cultured in presence of E3CM (AeD, I) than in controls (EeH, I). Levels of transcripts
corresponding to rod photoreceptor regulators (Mash1) and markers (opsin, arrestin and rhodopsin kinase) were decreased in neurospheres cultured in the presence
E3CM (J, lane 2; K) than in controls (J, lane 1). No difference in levels of transcripts corresponding to bipolar cells (Chx10) was observed between the two groups.
There was a relative decrease in levels of transcripts corresponding to opsin, when E14 neurospheres were cultured in the presence of E3CM (L, lane 2; M),
compared to controls (L, lane 1). Data are expressed as mean SEM from triplicate cultures of three different experiments. 200 (***p < 0.001).
40.1 0.8%; p < 0.001) (Figs. 4BeN; 1EeM). To ensure
that the observed results were not due to differential cell survival in different conditions, TUNEL assay was carried out
which demonstrated no difference in the number of apoptotic
cells between the groups (Fig. 4QeV).
These results suggested that the early retinal microenvironment possesses both RGC promoting and RGC inhibiting
activities. We were interested in knowing the nature of these
different regulatory activities that are partitioned in different
fractions. To begin with, we examined the temperature sensitivity of these activities by incubating <30 kDa E3CM and
>30 kDa E3CM fractions at 70 C for 30 min, a condition
known to inactivate proteins and used for the characterization
of factors (Waid and McLoon, 1998). E14 retinal progenitors,
cultured either in presence of reconstituted heat-treated or
untreated <30 kDa E3CM, showed no significant difference
in the proportion of BrdUþ and Brn3bþ cells (Fig. 5A, B,
Table 2). Also, levels of Ath5 and Brn3b transcripts remained
the same between the two groups (Fig. 5C, D), suggesting that
the RGC promoting activities, present in the <30 kDa E3CM
fraction, were heat stable. The reconstitution protocol for examining the heat sensitivity of >30 kDa E3CM fraction was
modified since it contained inhibitory activities, whose
absence can be more pronounced against the presence of differentiation-promoting activities (Fig. 5E). Therefore, controls
included fractionated E3CM, reconstituted without heat treatment. The experimental group was similarly reconstituted
except that >30 kDa E3CM fraction was heat-treated. The
proportion of BrdUþ and Brn3bþ cells increased significantly
in reconstituted fraction that contained heat-treated >30 kDa
584
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
Fig. 4. Activities that promote and inhibit the differentiation of retinal progenitors into RGCs are present in <30 kDa and >30 kDa E3CM fractions, respectively.
BrdU-tagged E14 neurospheres were cultured in presence of reconstituted >30 kDa or <30 kDa E3CM fractions (A) or 1% FBS (controls), followed by immunocytochemical and RT-PCR analyses for RGC-specific markers and regulators. There was a significant increase in the proportion of BrdUþ and Brn3bþ cells
(arrows) when neurospheres were cultured in presence of <30 kDa E3CM (BeE, N), compared to those in >30 kDa E3CM (FeI, N) or controls (JeM, N). There
was a significant decrease in the proportion of BrdUþ and Brn3bþ cells (arrows) when neurospheres were cultured in the presence >30 kDa E3CM, compared to
those in controls (JeM, N). There was a relative increase in levels Ath5 and Brn3b transcripts in neurospheres cultured in the presence of <30 kDa (O, lane 3; P),
compared to those in controls (O, lane 1). In contrast, their levels decreased in neurospheres cultured in the presence of >30 kDa E3CM (O, lane 2; P) as compared
to those in controls (O, lane 1). TUNEL analysis revealed only few TUNEL positive cells and their number remained similar in different conditions (QeV). Square
symbols in the reconstitution scheme correspond to conditions described in the graph. Data are expressed as mean SEM from triplicate cultures of three different
experiments. 200 for BeM and 100 for QeV (***p < 0.001; **p < 0.01).
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
585
Table 2
Summary of the differentiation of E14 and E18 retinal progenitors into RGCs (Brn3bþ cells) and rod photoreceptors (opsinþ cells), respectively, in the presence of
different E3CM fractions
E3CM fraction
E14 neurosphere (%Brn3bþ cells)
Fraction
>30 kDa
<30 kDa
<30 kDa (HT)
>30 kDa (HT)
16.5 0.9**
47 1.4***
46.4 0.8
51.8 0.9***
E18 neurosphere (%opsinþ cells)
Control
a
20.8 1.1
20.8 1.1a
47 1.4b
40.6 0.4c
Fraction
Control
10.2 0.8
6.1 0.5***
9 0.4***
NA
9.8 0.4a
9.8 0.4a
6.1 0.5b
NA
**p < 0.01; ***p < 0.001; NA: not applicable because this fraction has no effect on the proportion of opsinþ cells.
a
RCM þ 1% FBS.
b
Untreated <30 kDa E3CM fraction.
c
Untreated >30 kDa E3CM fraction reconstituted with <30 kDa E3CM fraction.
E3CM fraction, compared to controls (Fig. 5F, Table 2). Similarly, the levels of Ath5 and Brn3b transcripts were higher in
E14 retinal progenitors in the experimental group, compared
to those in controls (Fig. 5G, H). These observations suggested
that RGC inhibiting activities in >30 kDa E3CM fraction
were heat labile. Taken together, it appeared that the E3CM
contains heat stable (<30 kDa) and heat labile (>30 kDa)
RGC promoting and inhibiting activities, respectively, for
the regulated differentiation of retinal progenitors along
RGC lineage.
3.4. Activities present in >30 kDa E3CM fraction
influence the proliferation of retinal progenitors
One of the mechanisms by which activities in different
E3CM fractions can regulate RGC differentiation is by maintaining cells in a proliferative state. To determine the influence
of activities present in different E3CM fractions on cell proliferation, E14 retinal cells were cultured in reconstituted
fractions as described above (Fig. 4A). A significant increase
in the number of neurospheres/cm2 was observed when E14
retinal cells were cultured in presence of >30 kDa E3CM fraction compared to FBS controls (1192 35 vs. 367 35;
p < 0.001), suggesting that activities present in >30 kDa
E3CM fraction promote the proliferation of retinal progenitors
(Fig. 6AeD). There was no significant difference in the
number of neurospheres/cm2 generated when E14 cells were
cultured in the presence of <30 kDa E3CM fraction, compared
to FBS controls. In addition to an increase in the frequency of
neurospheres, there was an increase in the overall size of
neurospheres (w35% bigger) in the presence of >30 kDa
E3CM fraction, compared to those in FBS controls. Taken
together, these observations suggested that activities present
in >30 kDa E3CM promote the proliferation of retinal
progenitors.
3.5. Activities present in heat labile <30 kDa E3CM
fraction inhibit the differentiation of retinal progenitors
into rod photoreceptors
To characterize activities present in E3CM, which influence
the differentiation of retinal progenitors into rod photoreceptors, E18 neurospheres were cultured in reconstituted fractions
as described in Fig. 7A. We observed that the proportion of
opsinþ cells was decreased significantly in the presence of
<30 kDa E3CM fraction, compared to those in >30 Da
E3CM fraction or FBS controls (Fig. 7BeN, Table 2). There
was no significant difference in the proportion of opsinþ cells
when E18 neurospheres were cultured in the presence of
>30 kDa E3CM or FBS (Fig. 7BeN, Table 2). Results
obtained by immunocytochemical analysis were corroborated
by RT-PCR analysis; levels of Mash1 and opsin transcripts
were lower in cells in neurospheres that were cultured in the
presence of <30 kDa E3CM fraction, compared to those
cultured in the presence of >30 kDa E3CM fraction or in
FBS controls (Fig. 7O, P). These observations suggested that
the rod photoreceptor inhibiting activities of the E3CM were
localized in <30 kDa E3CM fraction. To determine the
temperature sensitivity of activities present in <30 kDa
E3CM fraction, we carried out heat treatment and reconstitution experiments as described above (Fig. 8A). We observed
a significant increase in the proportion of opsinþ cells in the
presence of heat-treated <30 kDa E3CM, compared to those
in untreated <30 kDa E3CM (Fig. 8B, Table 2). Consistent
with the immunocytochemical results, levels of Mash1 and
opsin transcripts were increased in the presence of heat-treated
<30 kDa E3CM (Fig. 8C, D). These observations suggested
that rod photoreceptor inhibiting activities in <30 kDa
E3CM fraction were heat labile and therefore, distinct from
RGC promoting activities that reside in the same fraction
but are heat stable.
4. Discussion
The stem cell or progenitor microenvironment in the
developing retina, unlike stem cell niche in adult tissues
like bone marrow (Arai et al., 2005), intestinal crypt (Leedham et al., 2005), hair follicle (Christiano, 2004) and brain
(Wurmser et al., 2004), is ever changing and therefore, relatively more complex due to the demand of the ongoing histogenesis. Here, we demonstrate a complex regulatory nature
of early retinal progenitor microenvironment. For example,
the microenvironment is characterized by three non-overlapping cell regulatory activities, partitioned into different
fractions of the CM, derived from cells in the early stage
of retinal histogenesis. The activities that promote the
586
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
Fig. 5. Heat sensitivity of activities present in E3CM fractions that influence the differentiation of retinal progenitors into RGC differentiation. BrdU-tagged E14
neurospheres were cultured in the presence of reconstituted untreated or heat-treated <30 kDa E3CM fractions (A), followed by immunocytochemical and RT-PCR
analyses for RGC-specific markers and regulators. There was no significant difference in the proportion of BrdUþ and Brn3bþ cells between the two groups (B).
Similarly, there was no difference in relative levels of Ath5 and Brn3b transcripts between the two groups (CeD). There was a significant increase in the proportion
of BrdUþ and Brn3bþ cells when neurospheres were cultured in the presence of heat-treated >30 kDa E3CM fraction as compared to those in untreated fraction
(EeF). Similarly, there was an increase in the relative levels of Ath5 and Brn3b transcripts when progenitors were cultured in the presence of heat-treated >30 kDa
E3CM fraction (G, lane 2; H), compared to those in untreated fraction (G, lane 1). Square symbols in the reconstitution scheme correspond to conditions described
in the graph. Data are expressed as mean SEM from triplicate cultures of three different experiments (***p < 0.001).
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
587
Fig. 6. Activities present in >30 kDa E3CM fraction promote the proliferation of retinal progenitors. To determine the influence of activities present in >30 kDa
and <30 kDa E3CM fractions on proliferation of retinal progenitors, freshly dissociated E14 retinal cells were cultured in the presence of >30 kDa E3CM fraction/
<30 kDa E3CM fraction/1% FBS, and the number of neurospheres generated was counted. The number of neurospheres was significantly greater when cells were
cultured in the presence of >30 kDa E3CM (C, D) than those cultured in the presence of <30 kDa E3CM (B, D) fraction or FBS (A, D). Data are expressed as
mean SEM from triplicate cultures of three different experiments. 200 (***p < 0.001).
differentiation of early retinal progenitors into RGCs are
localized in <30 kDa fraction of the E3CM and appeared
to be heat stable. Amongst the known factors that could
contribute to such activities may include FGFs, that are
known to promote RGC differentiation (Zhao and Barnstable,
1996; Yang, 2004), and whose molecular weight is <30 kDa.
However, the relative heat sensitivity of FGFs may exclude
them as a part of RGC promoting activities in the
<30 kDa heat-stable fraction. We observed that the
>30 kDa E3CM fraction harbors heat labile activities that
are inhibitory to the differentiation of early retinal progenitors into RGCs. Similar RGC inhibiting activities have
been observed in >10 kDa fraction of conditioned medium
obtained from E9 chick retinal cells (Waid and McLoon,
1998). These activities, present in the microenvironment
along with those that promote progenitor differentiation,
suggest that the microenvironment fine-tunes the outcomes
of differentiation process by striking a balance between
differentiation-promoting and differentiation-inhibiting activities. The activities present in the heat labile >30 kDa
E3CM fraction may play an important role in the maintenance of early retinal progenitors pool by promoting their
proliferation. Amongst the known factors that could contribute towards such activities may include Wnts and Shh, that
are heat labile >30 kDa proteins and observed to regulate
proliferation of retinal progenitors (Kubo et al., 2003;
Yang, 2004; Ahmad et al., 2004; Das et al., 2006). Such
activities, which can positively influence proliferation, will
ensure the maintenance of retinal progenitors for subsequent
cell generation.
An intuitive characteristic of the early retinal stem cell/
progenitor microenvironment is to ensure that progenitors
588
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
Fig. 7. Activities that inhibit the differentiation of retinal progenitors into rod photoreceptor are present in <30 kDa E3CM fraction. E18 neurospheres were
cultured in the presence of fractionated and reconstituted >30 kDa or <30 kDa E3CM fractions or 1% FBS (controls) (A) followed by immunocytochemical
and RT-PCR analyses of the expression of opsin and Mash1. There was a significant decrease in the proportion of opsinþ cells (arrows) when progenitors were
cultured in the presence of <30 kDa E3CM fraction (BeE, N), compared to those cultured in the presence of >30 kDa E3CM fraction (FeI, N) or FBS (JeM,
N). There was a decrease in the relative levels of opsin and Mash1 transcripts when progenitors were cultured in the presence of <30 kDa E3CM fraction (lane
3) as compared to those cultured in the presence of >30 kDa E3CM fraction (O, lane 2; P) or FBS (O, lane 1). Square symbols in the reconstitution scheme
correspond to conditions described in the graph. Data are expressed as mean SEM from triplicate cultures of three different experiments. 200
(***p < 0.001).
at that stage do not prematurely differentiate into cell types
that are predominantly generated in the late stage of histogenesis. This is particularly important with the emerging evidence that progenitors within a tissue are plastic and can
respond to conducive microenvironment for generating different cell types (James et al., 2003) and for regeneration
(Conboy et al., 2005). For example, it has been demonstrated
that the late retinal progenitors, when exposed to conducive
microenvironment, can generate early born neurons, demonstrating their inherent plasticity (James et al., 2003).
Evidence suggests that the early retinal progenitors may be
similarly plastic and may possess some capacity to generate
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
589
cells other than those that are predominantly born during
early neurogenesis (Watanabe and Raff, 1990, 1992). That
the generation of such cells can be influenced by microenvironment was demonstrated when E15 retinal cells were
observed to generate significantly more rod photoreceptors
in the presence of cells from postnatal stage than from E15
retina (Watanabe and Raff, 1990, 1992). Based on these observations it could be surmised that the early progenitors are
competent to give rise to early born neurons by virtue of the
selective expression of the genome (James et al., 2004).
However, this competence is unlikely to be absolute, considering the fact that few rod photoreceptors are born during
early histogenesis (Morrow et al., 1998). But the odds are
heavily against such specification as the early progenitors,
besides being partial towards the generation of early born
neurons, are surrounded by cues for their specification and
the presence of inhibitory activities, present in heat labile
<30 kDa, for the generation of late born neurons. It could
be suggested that CNTF, a heat labile <30 kDa protein and
a candidate regulator of rod photoreceptors, may contribute
to such activities. However, based on the observation that
the expression of CNTF is detected much later in histogenesis (Kirsch et al., 1997) it may be ruled out from the
<30 kDa fraction. As the development progresses, due to
differential gene expression (James et al., 2004), progenitors
acquire the ability (competence) to respond more favorably
to cues that promote the generation of the late born neurons
than to those conducive for giving rise to early born neurons.
This process is likely to be aided further by changes in the
microenvironment which progressively attenuate the inhibitory influences on the generation of the late born. Alternatively, it can be proposed that there could be sub pools of
progenitors with variations in levels of competence to generate early and late born neurons during early histogenesis.
Examination of these possibilities will require further studies.
Taken together, our observations reveal the complex nature
of the early retinal progenitor microenvironment for the
maintenance and differentiation of retinal progenitors. It is
characterized by different activities, the combined actions
of which may regulate the efficiency and fidelity of cell
differentiation during early retinal histogenesis.
Acknowledgements
Fig. 8. Heat sensitivity of activities present in E3CM fractions that influence
the differentiation of retinal progenitors into rod photoreceptor. E18 neurospheres were cultured in presence of untreated or heat-treated <30 kDa
E3CM fraction (A), followed by immunocytochemical and RT-PCR analyses for the expression of opsin and Mash1. There was a significant increase
in the proportion of opsinþ cells when progenitors were cultured in the
presence of heat-treated <30 kDa E3CM fraction as compared to untreated
fraction (B). Similarly, there was an increase in the relative levels of Mash1
and opsin transcripts when progenitors were cultured in the presence of
heat-treated <30 kDa E3CM fraction (C, lane 2; D) as compared to
This work was supported by the Glaucoma Foundation; The
Nebraska Tobacco Settlement Biomedical Research Development Fund and Research to Prevent Blindness. We thank Dr.
Colin Barnstable for opsin antibody, and Kavita B. Mallya
and Frank Soto Leon for their excellent technical assistance.
untreated controls (C, lane 2). Square symbols in the reconstitution scheme
correspond to conditions described in the graph. Data are expressed as
mean SEM from triplicate cultures of three different experiments
(***p < 0.001).
590
G.V. Hegde et al. / Experimental Eye Research 84 (2007) 577e590
References
Ahmad, I., Das, A.V., James, J., Bhattacharya, S., Zhao, X., 2004. Neural stem
cells in the mammalian eye: types and regulation. Semin. Cell Dev. Biol.
15, 53e62.
Ahmad, I., Dooley, C.M., Afiat, S., 1998. Involvement of Mash1 in EGFmediated regulation of differentiation in the vertebrate retina. Dev. Biol.
194, 86e98.
Ahmad, I., Dooley, C.M., Thoreson, W.B., Rogers, J.A., Afiat, S., 1999. In
vitro analysis of a mammalian retinal progenitor that gives rise to neurons
and glia. Brain Res. 831, 1e10.
Altshuler, D., Turner, D.L., Cepko, C.L., 1991. Specification of cell type in the
vertebrate retina. In: Development of the Visual System, vol. 3. MIT Press,
London. 37e58.
Arai, F., Hirao, A., Suda, T., 2005. Regulation of hematopoietic stem cells by
the niche. Trends Cardiovasc. Med. 15, 75e79.
Barnstable, C.J., Drager, U.C., 1984. Thy-1 antigen: a ganglion cell specific
marker in rodent retina. Neuroscience 11, 847e855.
Bhattacharya, S., Jackson, J.D., Kuszynski, C., Joshi, S., Ahmad, I., 2003.
Prospective identification and enrichment of neural progenitors from the
mammalian retina. Investig. Ophthalmol. Vis. Sci. 44, 2764e2773.
Brown, N.L., Patel, S., Brzezinski, J., Glaser, T., 2001. Math5 is required for
retinal ganglion cell and optic nerve formation. Development 128,
2497e2508.
Christiano, A.M., 2004. Epithelial stem cells: stepping out of their niche. Cell
118, 530e532.
Christie, G.A., 1964. Development stages in somite and post-somite rat embryos. Based on external appearance, and including some features of the
macroscopic development of the oral cavity. J. Morphol. 114, 263e286.
Conboy, I.M., Conboy, M.J., Wagers, A.J., Girma, E.R., Weissman, I.L.,
Rando, T.A., 2005. Rejuvenation of aged progenitor cells by exposure to
a young systemic environment. Nature 433, 760e764.
Das, A., Zhao, X., James, J., Kim, M., Cowan, K.H., Ahmad, I., 2006. Neural
stem cells in adult ciliary epithelium express GFAP and are regulated by
Wnt signaling. Biochem. Biophys. Res. Commun. 339, 708e716.
Hamburger, V., Hamilton, H.L., 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49e92.
James, J., Das, A.V., Bhattacharya, S., Chacko, D.M., Zhao, X., Ahmad, I.,
2003. In vitro generation of early-born neurons from late retinal progenitors. J. Neurosci. 23, 8193e8203.
James, J., Das, A.V., Rahnenfuhrer, J., Ahmad, I., 2004. Cellular and molecular characterization of early and late retinal stem cells/progenitors: differential regulation of proliferation and context dependent role of Notch
signaling. J. Neurobiol. 61, 359e376.
Kirsch, M., Lee, M.Y., Meyer, V., Wiese, A., Hofmann, H.D., 1997. Evidence
for multiple, local functions of ciliary neurotrophic factor (CNTF) in
retinal development: expression of CNTF and its receptors and in vitro
effects on target cells. J. Neurochem. 68, 979e990.
Kubo, F., Takeichi, M., Nakagawa, S., 2003. Wnt2b controls retinal cell differentiation at the ciliary marginal zone. Development 130, 587e598.
Leedham, S.J., Brittan, M., McDonald, S.A., Wright, N.A., 2005. Intestinal
stem cells. J. Cell. Mol. Med. 9, 11e24.
Livesey, F.J., Cepko, C.L., 2001. Vertebrate neural cell-fate determination:
lessons from the retina. Nat. Rev. Neurosci. 2, 109e118.
Morrow, E.M., Belliveau, M.J., Cepko, C.L., 1998. Two phases of rod photoreceptor differentiation during rat retinal development. J. Neurosci. 18,
3738e3748.
Prada, C., Puga, J., Perez-Mendez, L., Lopez, R., Ramirez, G., 1991. Spatial
and temporal patterns of neurogenesis in the chick retina. Eur. J. Neurosci.
3, 559e569.
Rapaport, D.H., Wong, L.L., Wood, E.D., Yasumura, D., LaVail, M.M., 2004.
Timing and topography of cell genesis in the rat retina. J. Comp. Neurol.
474, 304e324.
Reynolds, B.A., Rietze, R.L., 2005. Neural stem cells and neurospherese
re-evaluating the relationship. Nat. Methods 2, 333e336.
Sparrow, J.R., Hicks, D., Barnstable, C.J., 1990. Cell commitment and differentiation in explants of embryonic rat neural retina. Comparison with the
developmental potential of dissociated retina. Brain Res. Dev. Brain Res.
51, 69e84.
Waid, D.K., McLoon, S.C., 1998. Ganglion cells influence the fate of dividing
retinal cells in culture. Development 125, 1059e1066.
Watanabe, T., Raff, M.C., 1990. Rod photoreceptor development in vitro:
intrinsic properties of proliferating neuroepithelial cells change as development proceeds in the rat retina. Neuron 4, 461e467.
Watanabe, T., Raff, M.C., 1992. Diffusible rod-promoting signals in the developing rat retina. Development 114, 899e906.
Wurmser, A.E., Palmer, T.D., Gage, F.H., 2004. Neuroscience. Cellular
interactions in the stem cell niche. Science 304, 1253e1255.
Xiang, M., Zhou, L., Peng, Y.W., Eddy, R.L., Shows, T.B., Nathans, J., 1993.
Brn-3b: a POU domain gene expressed in a subset of retinal ganglion cells.
Neuron 11, 689e701.
Yang, X.J., 2004. Roles of cell-extrinsic growth factors in vertebrate eye
pattern formation and retinogenesis. Semin. Cell Dev. Biol. 15, 91e103.
Zhao, S., Barnstable, C.J., 1996. Differential effects of bFGF on development
of the rat retina. Brain Res. 723, 169e176.
Zhou, H., Yoshioka, T., Nathans, J., 1996. Retina-derived POU-domain factor1: a complex POU-domain gene implicated in the development of retinal
ganglion and amacrine cells. J. Neurosci. 16, 2261e2274.