Abstract
purpose. To determine the expression patterns of the retinoblastoma protein and
the E2F transcription factor families in limbal and corneal epithelia
and in corneal keratocytes in situ during corneal development and
differentiation.
methods. Retinoblastoma protein (pRb) and its family members p107 and p130;
E2F-1, -2, and -4, members of the E2F family of transcription factors;
and Ki67, a marker of actively cycling cells, were localized by
indirect immunofluorescence microscopy, in corneas of neonatal,
juvenile, and adult rats. Presence of mRNA for pRb, p107, p130, and E2F
types 1 to 5 in adult corneal epithelium was determined by reverse
transcription–polymerase chain reaction.
results. mRNA for all members of pRb and E2F families was present in adult
corneal epithelium. The greatest number of Ki67-positive corneal and
limbal epithelial cells were present at days 13 to 19, and
Ki67-positive stromal keratocytes at day 2. pRb and E2F-2 were
localized to all cells in neonatal, juvenile, and adult corneas. With
age, p130 localization became more intense and nuclear in stromal
keratocytes and suprabasal cells of corneal and limbal epithelia; p107,
initially nuclear in limbal and corneal epithelia, became increasingly
cytoplasmic in corneal epithelium. E2F-1 was initially nuclear in
keratocytes and diminished after day 10. E2F-1 was localized in the
basal cell layer of limbal and corneal epithelia after day 10. E2F-4
was always nuclear in limbal epithelium and cytoplasmic in corneal
epithelium.
conclusions. Expression patterns of pRb and E2F family proteins vary with corneal
cell differentiation, but are most apparent with p130 and p107. Nuclear
localization of p130 appears to correlate with terminal differentiation
in epithelium and entrance into a quiescent state by keratocytes. In
contrast, p107 is nuclear in the undifferentiated limbal basal cells
and is cytoplasmic in the remainder of the corneal epithelial
cells.
In the present study, the cornea was used to examine alterations
in proteins associated with regulation of cellular proliferation and
differentiation.
1 The developing rodent cornea provides an
opportunity to examine these alterations in both epithelium and stromal
keratocytes.
2 3 4 5 At birth, rat corneal and limbal
epithelia consist of one or two layers of cells. At 10 days of age, the
number of cell layers increases to two or three, and by 14 days, just
after eyelid opening, four to five layers of cells are
observed.
6 In the adult, stromal keratocytes are thought
to be quiescent; however, 5% to 10% of the stromal keratocytes appear
to proliferate during corneal development.
7 8 Proliferation is maximal at 1 to 3 days of age and declines until the
time of eyelid opening (12–14 days). This suggests that during
development the keratocytes either become quiescent or exit the cell
cycle.
8 9 10
One of the key regulatory proteins of the cell cycle is the
retinoblastoma gene product (pRb). The product of the retinoblastoma
gene is a ubiquitously expressed, 105-kDa nuclear
phosphoprotein.
11 12 The Rb family consists of three
members, pRb, p107, and p130.
13 14 In their
hypophosphorylated state, pRb, p107, and p130 complex with members of
the E2F family of transcription factors. After their phosphorylation,
the complex with E2F is disrupted, allowing free E2F to activate the
transcription of genes necessary for DNA synthesis. Binding of DNA and
transcription occurs only when E2F is complexed with the DP family of
transcription factors
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 The pRb and E2F families also
appear to regulate entrance and exit from the cell cycle and have been
associated with muscle and neuronal
differentiation.
25 32 33 34
Recent reports suggest that translocation of cell cycle–associated
proteins from the cytoplasm to the nucleus may be another method of
cell cycle regulation. It has been generally assumed that cell
cycle–associated proteins, which function at least in part by binding
DNA, would be present primarily in the nucleus. However, cyclin
B1,
35 36 CDK 4,
37 E2F-1, and
E2F-4
38 have all been demonstrated to be translocated
between the cytoplasm and the nucleus in a cell cycle–dependent
manner. This would allow them to have maximal affect in the nucleus
when they are most needed to regulate cell cycle entrance, progression,
and exit. It is unclear whether these proteins also have a functional
role when they are localized in the cytoplasm.
Although there have been numerous studies of pRb and E2F function in
cell lines and cultured cells, very few investigators have examined
these proteins in vivo. In the present study, we used the developing
cornea to correlate the expression and localization of the pRb and E2F
families in cells at different states of differentiation.
Sprague–Dawley rats were used, and all procedures conformed to
the ARVO Statement for the Use of Animals in Ophthalmic and Vision
Research. Rats were killed with an overdose of pentobarbital sodium.
Neonatal (1–14 days of age), juvenile (15 days to 6 weeks of
age), and adult (3–4 months of age) rats were killed. Eyes with or
without eyelids were enucleated, placed in optimal cutting temperature
compound (Tissue Tek II; Laboratory Tek, Napierville, IL) and frozen on
dry ice. Six-micrometer cryostat sections were placed on gelatin-coated
slides and air dried overnight at 37°C. Slides were fixed in 100%
methanol for 10 minutes at −20°C, washed with phosphate-buffered
saline (PBS) for 10 minutes, and then blocked in 1% bovine serum
albumin solution in PBS for 10 minutes. Antibody against pRb was
obtained from Boehringer Mannheim (Indianapolis, IN), antibodies
against p107 (C-18), p130 (C-20), E2F-1 (KH95), E2F-2 (L-20), and E2F-4
(C-20) were obtained from Santa Cruz Biotech (Santa Cruz, CA), and
anti-Ki67 (MM1) from Novacastra (Newcastle-upon-Tyne, UK). The
following dilutions were used for primary antibodies: 2 μg/ml pRb,
E2F-2, E2F-4, and Ki67; 1 μg/ml p107 and p130; and 4 μg/ml E2F-1.
Slides were incubated for 1 hour in a moist chamber at room temperature
for all antibodies except pRb and E2F-1, which were incubated overnight
at 4°C. Slides were rinsed with PBS and then blocked with 1% bovine
serum albumin for 10 minutes. Fluorescein isothiocyanate–conjugated
affinity-purified donkey anti-rabbit IgG (for p107, p130, E2F-2, and
E2F-4) and anti-mouse IgG (for pRb, Ki67, and E2F-1; Jackson
ImmunoResearch, West Grove, PA) were applied at a dilution of 1:50 and
incubated for 1 hour at room temperature. Slides were rinsed in PBS,
and coverslips were mounted with Vectashield (Vector, Burlingame, CA).
Sections were viewed and photographed with a microscope
(Axiophot; Carl Zeiss, Thornwood, NY). Negative controls
consisted of secondary antibody alone, irrelevant monoclonal
antibodies, or primary antibody preadsorbed with its respective
antigen. Sections from at least four rats were examined for each
antibody and time point. Antibody binding was examined in the limbal
and corneal epithelia and in stromal keratocytes. In addition, staining
patterns were examined to determine whether localization was
cytoplasmic, nuclear, or both. Relative percentages of Ki67 binding
were determined by comparing Ki67-labeled cells with total cells by
means of by propidium iodide counterstaining. Cells were counted in the
basal cell layers of limbal and corneal epithelia and in the corneal
stroma. Ki67-labeled keratocytes were not quantitated in the limbal
stroma because this area contains a mixture of cell types.
Isolation of Total RNA and Performance of Reverse
Transcription–Polymerase Chain Reaction
Corneal epithelium was harvested, from a central 4-mm area
of 12 adult rat eyes. Tissue was immediately frozen in liquid nitrogen,
and total RNA was isolated using TRIzol Reagent (Gibco–Life
Technologies, Grand Island, NY). Before further use, the total RNA was
treated with DNase I, amplification grade (1 U DNase/1 μg total RNA;
Life Technologies).
Reverse transcription–polymerase chain reaction (RT-PCR) was performed
as previously described
39 using specific primers
(Table 1) . The primer sets for E2F-1, -2, -4, and -5 are
published.
40 The primer sets for pRb, p107, p130, and
E2F-3 were devised using commercial software (Oligo Primer Analysis;
National Biosciences, Plymouth, MN) that selects primers based on
minimal hairpin formation, minimal duplex formation, and guanine
cytosine (GC) composition. Primer sets for
glyceraldehyde-3-phosphate dehydrogenase (G3PDH) were purchased from
Clontech (Palo Alto, CA). Samples were denatured for 1.5 minutes at
94°C, followed by 25 (G3PDH), 30 (p107, p130, pRb), or 35 (E2F-1, -2,
-3, -4, and -5) PCR cycles of denaturation for 1.5 minutes at 94°C,
annealing for 1 minute
(Table 1) , and extension for 1 minute at 72°C.
The final elongation step was performed at 72°C for 7 minutes. Twenty
microliters of the PCR product was then resolved on 1.5% agarose gel
containing 0.5 μg/ml ethidium bromide. The gel was photographed
(Digital Science DC40 camera; Eastman Kodak, Rochester, NY). G3PDH
primers were used as a positive control of cDNA quality. Samples with
no cDNA were also amplified and served as negative controls.
Specificity of the PCR products was checked by cycle sequencing at the
Tufts DNA Sequencing Facility (Boston, MA).
Ki67, a marker of actively cycling cells,
41 42 was
used to demonstrate the relative proliferative activity of limbal and
corneal epithelial cells and central corneal keratocytes. As seen in
Figure 1 , Ki67 is localized in the basal layer of corneal and limbal epithelium.
The number of labeled epithelial cells peaks at days 14 to 15 (
Figs. 1C 1D ;
Fig. 2 ) and then declines until it reaches adult levels (
Figs. 1E 1F ;
Fig. 2 ). Although the relative number of labeled cells varied with age, the
corneal epithelium consistently had a higher percentage of labeled
cells than did the limbal epithelium
(Fig. 2) . No Ki67 was detected in
either limbal or corneal suprabasal cells. These data are consistent
with previous reports that the limbal epithelium contains a mixture of
slow-cycling stem cells and more rapidly cycling transient amplifying
cells, that the central corneal epithelium contains a rapidly
proliferating basal cell layer, and that the suprabasal cells have
exited the cell cycle and become terminally
differentiated.
4 5 31 Ki67 was also localized in stromal
keratocytes in the developing cornea, with the maximal number of
labeled cells in rats 1 to 3 days of age (
Fig. 1A ;
Fig. 2 ). The number
of labeled keratocytes decreased dramatically with age
(Fig. 2) , and by
15 days, less than 0.5% of the cells were labeled (
Fig. 1C ;
Fig. 2 ).
These localization patterns suggest that keratocytes are proliferative
in the rat cornea at birth, then become quiescent or exit the cell
cycle.
RT-PCR was used to assay for mRNA for members of the pRb and E2F
families in adult corneal epithelium. The data shown in
Figure 3 indicate that mRNA for all members of both families was expressed. No
bands were detected in samples with no cDNA (data not shown). Based on
these results, indirect immunofluorescence was then used to localize
pRb, p107, p130, and E2F-1, -2, and -4. E2F-3 and -5 were not examined,
because no antibody could be obtained that elicited reproducible
results on rat cornea.
As seen in
Figures 4 5 and 6 , members of the retinoblastoma family, pRb, p130, and p107 each had a
distinct pattern of localization. pRb was present in the nuclei of all
epithelial cells regardless of proliferative or differentiative state
(Fig. 4) . Binding intensity appeared to be greater in basal cells than
in suprabasal cells. The pattern of pRb localization was the same
during development as in the adult. No differences were observed
between limbal and corneal epithelium. pRb was also present in the
nuclei of all keratocytes, regardless of whether the cells were in a
quiescent or proliferative state.
In the developing rat, p130 was initially present at low levels in the
cornea
(Fig. 5A) and limbus
(Fig. 5E) . As the epithelium began to
stratify (days 7–9), the level of anti-p130 binding appeared to
increase
(Fig. 5B) . Until day 7, p130 exhibited a weak cytoplasmic
localization but became increasingly nuclear after days 9 to 11. At 15
days, nuclear localization of p130 in suprabasal cells could be clearly
seen
(Fig. 5C) . In adult corneal epithelium, cytoplasmic p130
localization was seen in all cell layers; however, nuclear localization
was present only in the suprabasal cell layers (
Fig. 5D1 ). In the
adult, limbal epithelium exhibited nuclear localization of p130 in the
suprabasal cells and in some basal cells, with diffuse cytoplasmic
localization in all layers
(Fig. 5F) . The staining intensity of
anti-p130 also was altered in keratocytes during development. Anti-p130
binding was initially present at low levels
(Figs. 5A 5E) and became
increasingly intense after days 9 to 11
(Figs. 5B 5C) . In the adult
cornea, p130 exhibited a staining pattern similar to that of propidium
iodide (data not shown), indicating a nuclear localization in the
keratocytes (
Figs. 5D1 5F ).
The localization pattern of p107 was also altered during development
and differentiation
(Fig. 6) . From birth until day 7, p107 was
primarily localized in the nuclei of basal cells in both corneal and
limbal epithelia
(Figs. 6A 6B) . Diffuse cytoplasmic localization was
also present. This pattern remained relatively unchanged in the limbal
epithelium during development; nuclear binding was seen even in adults
(Figs. 5B 5D 5F) . However, in corneal epithelium, the localization
became increasingly cytoplasmic as the epithelium stratified
(Figs. 5C 5E) . In contrast to pRb and p130, nuclear localization was observed
only in basal cells. Also, the staining intensity of anti-p107 in the
corneal keratocytes appeared to diminish with age
(Figs. 5A 5C 5E) .
No fluorescence was observed in the central cornea in sections when
p130 or p107 antibodies were preadsorbed with blocking peptides
(Figs. 4 5 6) . Low level binding was observed in the limbal vasculature of
both the preadsorption (data not shown) and secondary antibody controls
(see
Fig. 8D2 ).
The binding patterns of E2F-1, -2, and -4 are demonstrated in
Figures 7 and 8 . In developing rats (1–10 days), E2F-1 was present at low, diffuse
levels in both limbal and corneal epithelia
(Figs. 7A 7B) . After 10
days, E2F-1 was present at highest levels in the basal cells of the
corneal and limbal epithelium, showing an intense nuclear and diffuse
cytoplasmic localization
(Figs. 7C 7D) . In both limbal and corneal
epithelia, the number of basal cells with nuclear localization
decreased after day 17 (data not shown). E2F-1 was localized in the
keratocytes of both limbus and central cornea at relatively high levels
until day 10
(Figs. 7A 7B) , decreasing rapidly after day 12 in central
cornea
(Fig. 7C) . This pattern was maintained from 17 days to adulthood
(data not shown). Staining intensity also decreased in the limbal
keratocytes but not to as great an extent as observed in central cornea
(Fig. 7D) . Anti-E2F-2 exhibited nuclear binding in epithelial cells and
keratocytes, with a maximal intensity in the basal cell layer of both
limbal and corneal epithelia
(Figs. 8A 8B) . Some cytoplasmic
localization was observed in all cell layers. No changes were detected
during development (data not shown). In the adult cornea, the
localization pattern of E2F-4 most closely resembled that of p107, with
cytoplasmic binding in central corneal epithelium, occasional nuclear
binding in limbal epithelium, and low-level diffuse localization in the
keratocytes
(Figs. 8C 8D) . No changes were detected during development
(data not shown).
The localization patterns of the pRb and E2F families are summarized in
Figures 9 10 and 11 .
We have examined the expression and localization of the pRb family
of proteins and the E2F family of transcription factors in corneal
cells in situ. These proteins are involved in regulation of the cell
cycle, quiescence, and differentiation. The cornea provides an
excellent model for this type of study, in that it contains actively
cycling cells (corneal epithelial basal cells), slow-cycling cells (a
subset of the limbal epithelial basal cells), terminally differentiated
cells (limbal and corneal suprabasal cells), and quiescent cells
(corneal keratocytes). These varying proliferative states were
confirmed by the localization of Ki67.
41 42
The major finding of the present study is that members of the pRb and
E2F families exhibit differential localization, depending primarily on
the differentiative state of the corneal cells. Findings of particular
note include (1) pRb and E2F-2 are present in cells regardless of
whether they are actively cycling; however, the intensity of antibody
binding diminishes in the terminally differentiated epithelial cells;
(2) p130 expression appears to be highest in terminally differentiated
and quiescent cells; (3) p107 and E2F-4 are nuclear in the relatively
undifferentiated limbal basal cells but primarily cytoplasmic in
corneal epithelial cells; and (4) E2F-1 is concentrated in cells with
highest proliferative capacity.
In this study, pRb and E2F-2 were observed to be present in the nuclei
of corneal cells in developing and adult rats. In both limbal and
corneal epithelia, pRb and E2F-2 were present at highest levels in the
basal layer. This localization is consistent with the need for pRb in
the regulation of cell proliferation in actively cycling cells and
suggests that the need for high levels of nuclear pRb and E2F-2
diminishes during terminal differentiation of the epithelial cells. In
contrast to the apparent change in levels of pRb and E2F-2 during
terminal differentiation of the epithelial cells, there was no obvious
change in expression patterns as corneal keratocytes became quiescent.
This suggests that cells resting in G0 maintain pRb and E2F-2 so that
they may reenter the cell cycle. Another possibility is that these
proteins may be present in Rb-E2F family complexes that bind DNA and
suppress expression of genes necessary for DNA synthesis, thus
preventing cell proliferation.
43
Expression of p130 in situ possesses characteristics similar to cells
in culture, where terminally differentiated and quiescent cells express
the highest levels of p130.
14 32 33 44 This is apparent in
the developing rat, in which p130 was initially localized at low levels
in developing corneal epithelium and then became nuclear in the
suprabasal cells as the epithelium stratified and differentiated
(Fig. 5) . A similar change was observed in keratocytes, where nuclear
localization of p130 increased with age
(Fig. 5) , correlating with the
decrease in keratocyte proliferation. It may play a role in actively
repressing keratocytes from entering the cell cycle and in terminal
differentiation of the epithelial cells. This hypothesis is supported
by Paramio et al.,
45 who found that p130 is restricted to
epidermal cells committed to differentiation. It is further supported
by the recent report of Dagnino et al.,
46 who found, by in
situ hybridization, that expression of E2F-5 (p130’s most common
complex partner
21 27 ) is upregulated in the
nonproliferative, terminally differentiated cells in both epidermal and
intestinal epithelia.
One unexpected finding in our study was the cytoplasmic localization of
p107 in the corneal epithelium of adult rats. Obvious nuclear
localization of p107 in adult corneal epithelium was extremely rare
(less than 1% of the basal cells). This is in stark contrast to the
localization of Ki67, which was present in 42% of corneal epithelial
basal cells. These data suggest that p107 may not play a major role in
homeostatic cell proliferation of the corneal epithelium. One possible
explanation for the cytoplasmic localization of p107 is that the
protein is constitutively expressed in developing and adult ocular
epithelia, and that its activity is regulated by entry into the
nucleus.
38 This would remove the necessity for a layer of
gene regulation to specifically turn off p107 transcription. The
meaning of the high number of limbal cells with nuclear p107 is also
unclear. We have previously shown that limbal basal cells are arrested
in G1.
2 Because p107 levels in serum-stimulated cells in
culture are upregulated late in G1, the present findings are consistent
with the arresting of limbal basal cells after the initiation of p107
expression but before entry into the S phase. Alternatively, p107 may
be repressing the transcription of a protein necessary for limbal basal
cells to differentiate.
It has been suggested that p107 and p130 may be redundant or may
compensate for each other.
14 44 47 48 This concept is
primarily based on the findings that
p107
−/− 47 and
p130
−/− 48 mice are viable, with no
obvious phenotype. However, mice without both proteins die shortly
after birth.
48 By comparison,
pRb
−/− mice die on embryonic day
13.5.
49 50 51 No corneal phenotypes have been reported for
any of the knockout mice, but it is not clear whether the cornea has
been extensively examined. The localization patterns of p107 and p130
demonstrated in this study suggest that the proteins play different
roles in the regulation of corneal cell differentiation. For example,
p130 is localized in the nucleus of the terminally differentiated wing
cells; whereas, p107 has a cytoplasmic localization. This suggests that
p130 may play a more prominent role in terminal differentiation than
does p107. We are currently examining p107
−/− and p130
−/− mice to determine whether p107 or
p130 localization is altered to compensate for the absence of the
other.
One of the intriguing and potentially important observations in this
study is the differential expression of several members of the pRb and
E2F families in limbal and corneal epithelia. This was most obvious in
the expression of p107 and E2F-4, which were nuclear in a subpopulation
of the basal cells in the limbal epithelium but cytoplasmic in the
adult corneal epithelium. A differential localization was shown by
p130, in that a subpopulation of the limbal basal cells exhibited
nuclear localization, whereas few if any basal cells in the corneal
epithelium had nuclear localization. These results may help in
understanding the dynamics of stem cell proliferation in the cornea.
Unfortunately, all the antibodies used for these proteins are rabbit
polyclonal antibodies, and we have not been able to perform
colocalization experiments.
The findings in the present study demonstrate that the members of the
pRb and E2F families are differentially expressed in corneal and limbal
epithelia and in corneal keratocytes. These results suggest that the
level of protein expression and the pattern of localization are
important mechanisms of regulation in corneal cell proliferation and
differentiation and represent an initiation of experiments to
understand the role of the pRb and E2F families. Obviously, however,
because phosphorylation of these proteins is also a major regulatory
event, changes in phosphorylation relative to cell cycle progression
must be examined. In addition, the alteration of pRb–E2F family
complexes should be examined. These experiments represent the next
stage of our investigations.