Abstract
Purpose.:
Because of its role in cell cycle regulation and apoptosis, p53 may be involved in maintaining the post-mitotic state of the adult eye. To shed light on the role of p53 in retinal development and maintenance, this study investigated the pattern of expression of p53, its family members, and its regulators during the development of the mouse eye.
Methods.:
Relative quantitative real-time PCR (qRT-PCR) was used to determine the steady-state levels of target transcripts in RNA extracted from wild-type mouse whole eyes or retinas between embryonic day (E) 15 and post-natal day (P) 30. Immunoblotting was used to compare the steady-state levels of the protein to that of the transcript.
Results.:
Transcript and protein levels for p53 in the eye were highest at E17 and E18, respectively. However, both p53 transcript and protein levels dropped precipitously thereafter, and no protein was detected on immunoblots after P3. Expression patterns of p63, p73, Mdm2, Mdm4, and Yy1 did not follow that of p53. Immunohistochemistry analysis of the developing eye showed that both p53 and Mdm2 are abundantly expressed at E18 in all layers of the retinal neuroblast.
Conclusions.:
Downregulation of p53 in the post-mitotic retina suggests that, although p53 may be involved in ocular and retinal development, it may play a minimal role in healthy adult retinal function.
The p53 tumor suppressor plays a critical role in regulating DNA damage repair, cell cycle progression, apoptosis, and autophagy.
1 It is also involved in the regulation of significant physiological steps during development, aging, and metabolism (for review, see ref. 1). During retinal genesis, developmental apoptosis is observed from embryonic day (E) 15 until post-natal day (P) 11.
2 This indicates that at least during development, there may be a role for p53 in the retina.
p53 is negatively regulated by murine double minute 2 (Mdm2),
1 Mdm4,
3 and yin yang 1 (Yy1).
4 Downstream targets of p53 include p21, a Cip/Kip cyclin-dependent kinase (Cdk) inhibitor
5 essential for p53-dependent cycle arrest,
6 the retinoblastoma protein (pRb),
7 and the E2F family of transcription regulators.
8
Although p53 has been shown to be expressed in several murine ocular tissues,
9,10 p53 knockout mice on C57BL × CBA
11 and 129/Sv × C57BL/6
12 backgrounds do not exhibit any ocular abnormalities, implying that p53 may not be essential for ocular development.
11 However, severe abnormalities are present in p53 null mice on the pure BALB/c OlaHsd
13 and C57BL/6
14 backgrounds, suggesting that alleles from the C57BL/6 genetic background interact with p53 in ocular development and contribute to the abnormal phenotype observed in the absence of p53.
14
In terms of ocular stress and disease, p53 has been shown to be involved in the retinal response to oxidative stress
15,16 and to lead to G
1 arrest upon retinal exposure to ionizing radiation.
17 −19 p53 may also play a role in hypoxia due to ischemia
20,21 and the development of age-related macular degeneration.
22 In contrast, p53 is not directly involved in the apoptotic response to intense light damage
23,24 or retinitis pigmentosa.
25–27 Finally, although p53 mutations are not associated with retinoblastoma, p53 does play a role in the disease development. In retinoblastoma cells, the p53 negative regulator Mdm4 is overexpressed, leading to p53 degradation and allowing cancerous cells to bypass the p53 checkpoint.
28 For a review of the known roles of p53 in ocular stress, please see ref.
29.
To begin to understand the role of p53 and its family members, regulators, and downstream target genes in the eye, the current study examined the levels of the aforementioned genes during early mouse eye development and in the post-mitotic retina. Using quantitative real-time PCR (qRT-PCR), immunoblotting, and immunohistochemistry (IHC) analysis, we determined the steady-state levels of the transcript and protein of p53 and of its regulators Mdm2, Mdm4, and Yy1, either in whole eyes from E15 through P3 or in retinal samples after P3 through P30 from C57BL/6 mice, which have been shown to develop ocular abnormalities in the absence of p53.
14 To determine whether other family members play any potential role in absence of p53, levels of p63 and p73 were also examined. Finally, expression levels of downstream p53 target genes were investigated.
C57BL/6 mice were housed and fed as previously described.
30 To obtain eyes from embryonic stages, female mice were superovulated and mated to proven male breeders.
30 Eyes were obtained from fetuses at the specified time points post-coitus. All experiments involving animals were approved by the local Institutional Animal Care and Use Committees and conformed to the National Institute of Health Guide for the Care and Use of Laboratory Animals and the Association for Research in Vision and Ophthalmology Resolution on the Use of Animals in Research.
Quantitative real-time PCR was performed as previously described
31 with total RNA extracted from C57BL/6 tissue. For E15 through P3, whole eyes were used. For time points after P3, retinas were extracted. Primers for all genes were designed to span introns to avoid amplification from genomic DNA.
Two qRT-PCR assays were performed in triplicate with 0.2 μg of each cDNA sample, using the gene-specific primer pairs shown in
Table 1. ΔcT values were calculated relative to those of the neuronal housekeeping gene encoding hypoxanthine phosphoribosyltransferase (
Hprt).
Hprt was assigned an arbitrary expression level of 10,000. Relative gene expression values were calculated as follows: relative expression = 10,000/2
ΔcT, where ΔcT = (gene cT − Hprt cT). PCR quality and specificity were verified by melting curve dissociation analysis and gel electrophoresis.
Amplification efficiency was determined by the calibration dilution curve and slope calculation. This calculator uses the slope produced by a qPCR standard curve to calculate the efficiency of the PCR. Multiple 5-fold dilutions of the cDNA are made, and PCRs are performed with a primer set. cT values are determined and plotted against log starting cDNA concentration by using a scatter chart (Excel software; Microsoft, Redmond, WA) and the slope was determined. Efficiency (
E) is calculated from the equation: E = −1 + 10
(−1/slope). Slopes between −3.1 and −3.6 giving reaction efficiencies between ∼80% and ∼120% are typically acceptable. Except for p63 (
Table 1), all primer sets produced PCR efficiencies within the acceptable range.
Whole eyes or retinas were isolated (depending on age, as described above) and processed for immunoblotting as previously described.
30 Blots were incubated with a primary antibody at specified dilutions (Table 2), washed, and then incubated with a secondary antibody. Membranes were incubated with SuperSignal enhanced chemiluminescent detection system (Pierce, Rockford, IL) and imaged using a Kodak image station (Carestream Molecular Imaging, Rochester, NY).
Cone photoreceptor-derived 661W cells
32 were grown in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Cellgro, Herndon, VA) and a 1% antimycotic-antibiotic cocktail (Invitrogen) and maintained in a humidified atmosphere of 95% air and 5% CO
2 at 37°C.
MCF-7 breast cancer cells
33 were grown in the same medium and under the same conditions as described above. NIH-3T3 cells, a mouse embryonic fibroblast cell line,
34 is the recommended positive control for p107 and were grown in the same medium and under the same conditions as described above.
IHC analysis was performed with frozen sections of eyes as described previously.
31 Labeling with the antibodies listed in
Table 3 was visualized using an Olympus BX-62 microscope equipped with a spinning disc confocal unit (Olympus America, Center Valley, PA) and an Orca-ER camera (Hamamatsu, Bridgewater, NJ). Images were acquired using Slidebook version 4.2 software (Intelligent Imaging Innovations, Denver, CO). Figures were assembled in Adobe Photoshop CS (Adobe, Mountain View, CA), and brightness, contrast, and threshold were adjusted to highlight specific labeling.
Unless otherwise stated, statistical significance was determined by using one-way ANOVA with Bonferroni post hoc multiple pairwise comparison tests (Prism; GraphPad Software, San Diego, CA), and data are means ± standard error of the means (SEM). Statistical significance was set at a P value of <0.05.
To better understand the developmental role of p53 and its family members, regulators, and target genes, their expression levels during murine retinal development were examined.
The transcript for
Trp53, the murine p53 gene, was detected at E15 with a significant increase at E17, followed by a gradual decrease until P30, when
Trp53 expression reached its lowest level (
Fig. 1A, upper panel). The reduction in
Trp53 expression after P11 may be biologically meaningful because retinal mitosis and normal developmental apoptotic cell death are both completed at this time.
2 This pattern of expression implies that
Trp53 may play a role in either the division and/or differentiation of retinal progenitor cells or in developmental apoptotic death.
At the protein level, P53 was abundant at E18 in quantities comparable to those observed in the continuously dividing cone photoreceptor-derived cell line 661W (
Fig. 1A, lower panel). At P3, P53 is still detectable but at a significantly lower level than that at E18. At subsequent time points during which the retina alone was analyzed, P53 protein was barely visible. This suggests that either p53 is unnecessary for retinal development at later stages or that p53 is more abundantly expressed at earlier time points in other ocular structures.
Mdm2, Mdm4, and Yy1 are well-known negative regulators of p53.
35 Mdm2 transcript levels significantly increased from E15 to P1 and remained steady between P3, when the whole eye was analyzed, and at P7, when analysis of the retina alone began. After P7, levels of
Mdm2 transcript fell, and thereafter, transcript levels remained stable into adulthood. However, in contrast to P53 which peaked at E18, Mdm2 protein levels seem to peak between P3 and P7 and then decline slightly (
Fig. 1B, lower panel). The increase in protein expression at P3 and P7 reflects the peak in transcript levels (
Fig. 1B).
Comparably to
Trp53, transcript levels for
Yy1 increased after E15 in ocular tissues and dropped shortly after birth. In the P7 retina, levels of
Yy1 transcript were insignificantly lower than ocular levels at P3 but continued to drop as the retina progressed in development (
Fig. 1C, upper panel). Yy1 protein, on the other hand, was expressed starting at embryonic stages and increased until P7 to the level at which the protein was maintained at the later time points examined (
Fig. 1C, lower panel). Because the retina became totally post-mitotic by P11,
36 it seems that Yy1 levels increased as the cells differentiated and entered post-mitosis. This is supported by the fact that the continuously dividing 661W cell line does not express any Yy1. This also suggests that Yy1 may have an alternative function in the retina other than p53 regulation.
Mdm4 transcript levels in ocular tissues were high at E15 and gradually decreased thereafter (
Fig. 1D, upper panel). However, at P7, retinal levels of
Mdm4 transcript were slightly higher than those observed in total ocular tissues at P3. At later post-natal stages, retinal
Mdm4 levels declined slowly, and by P30, they were significantly lower than the levels observed at P7. The level of Mdm4 protein reflects the pattern observed for its transcript (
Fig. 1D, lower panel). However, there was an increase at P60 that is hard to explain if the only function for Mdm4 is regulation of p53, because P53 is almost undetectable in adult retinal extracts. Nevertheless, in embryonic and early post-natal stages, Mdm4 behaved in a pattern similar to that of P53, suggesting that Mdm4 rather than Mdm2 may be the actual regulator of retinal p53 .
To determine whether
Trp53 was expressed mostly in the retina or in other ocular tissues during embryonic developmental time points, IHC analysis was performed. As shown in
Figure 2, both P53 and Mdm2 were abundantly present at E18 in the neuroblast layer, which is in agreement with prior immunoblot results (
Figs. 1A, 1B). However, while Mdm2 is clearly present in the retinal pigment epithelium, P53 is absent (
Fig. 2, compare panels A and B). In the retinoblast layer, most cells coexpressed these two proteins. However, there were cells (
Fig. 2D, white arrow) that expressed P53 alone while other cells expressed Mdm2 alone (
Fig. 2D, pink arrow). Furthermore, P53 generally appears to be localized to the nucleus (
Fig. 2D, white and yellow arrows), and Mdm2 seems to be mostly cytoplasmic (
Fig. 2D, yellow and pink arrows).
The p53 family also includes p63
37 and p73.
38,39 As structural and functional homologs of p53, these family members can form oligomers, bind DNA, and activate p53 target genes to regulate the cell cycle and mediate apoptosis.
37–40 p63 is required for ectoderm-derived tissues to develop normally,
41,42 and loss of p73 leads to hydrocephalus, chronic inflammation, and defects in neuronal development and pheromone sensing in zebrafish embryos,
43 suggesting a potential functional redundancy in the roles of p53, p63, and p73 in retinal development. Therefore, the patterns of expression of
Trp63 and
Trp73 transcripts during retinal development were determined.
qRT-PCR showed that
Trp63 was expressed at relatively high levels in ocular tissues at E15 but rapidly and significantly dropped at E17.
Trp63 expression dropped further at P1 but remained steady at P3. A second dramatic drop in expression was observed at P7 in retinal samples, to levels that were almost undetectable, which continued into early adulthood (
Fig. 3A, upper panel). Immunoblot analysis showed that in contrast to transcript levels, P63 protein levels were mostly stable throughout all developmental stages tested. (
Fig. 3A, lower panel). However, there might have been a slight increase in P63 levels at P7. Of interest is the fact that P63 was barely observed in 661W cells, which express reasonable amounts of P53 (
Fig. 1A, lower panel). The pattern of expression of
Trp63 transcript suggests that P63 might play a minimal role in early ocular and/or retinal development, at least in the presence of P53.
Expression of
Trp73 transcript is remarkably different than those of
Trp53 and
Trp63. Although it also started at relatively high levels at E15,
Trp73 transcript levels dropped significantly by E17 and remained at that level through P1. However, at P3, there was a significant increase in transcript steady-state levels before dropping again to a drastically low level by P7. Although
Trp73 transcript levels were virtually undetectable at P7, they rebounded slightly at P15 and P30 (
Fig. 3B, upper panel). At the protein level, P73 started at relatively high levels at E16 but quickly became undetectable after E18 in ocular tissues and the retina (
Fig. 3B, lower panel). This suggests that p73 may not play any significant role in mouse ocular/retinal late development, at least in presence of p53. However, based on the high levels of p73 that were observed at E16, P73 may have a significant role in early ocular development.
Because p53 has been shown to be tightly regulated during retinal development, it was relevant to examine the pattern of expression of p53 targets including p21, pRb, and the E2F family. p53 regulates cell cycle arrest at G
1 through p21-mediated inactivation of pRb.
44 p21-mediated cell cycle arrest can also occur through downregulation of p21 by telomerase reverse transcriptase (TERT) in response to dysfunctional telomeres
45 and through inhibition of proliferating cell nuclear antigen (PCNA), an essential DNA polymerase δ cofactor.
46,47
qRT-PCR analysis showed that transcript levels of
Cdkn1a, the murine p21 gene, were relatively high between E15 and P3, when the whole eye was examined, and were then subsequently reduced to lower levels that were maintained from P7 and beyond, when the retina alone was examined (
Fig. 4A, upper panel). Current data suggest a role for p21 in the development of ocular tissues other than the retina. However, P21 was not observed in any of the protein samples obtained from ocular or retinal tissues (
Fig. 4A, lower panel). Clearly, P21 does not follow the pattern observed for its transcript. It is interesting that P21 is also absent in 661W cells. Altogether, the data suggest that
Cdkn1a may serve a role in
Trp53-dependent stress response such as apoptosis.
The developmental pattern of expression of
Pcna transcript resembled that of
Trp53, with high expression at E17 that was reduced to very low levels by P11 (
Fig. 4B, upper panel). However, the protein levels seemed to be steady when ocular tissues were examined between E16 and P3 but lower in retinal samples at P7 and almost nonobservable at P30. At P7, central portions of the retina have entered post-mitosis,
36 and by P30 all retinal cells stopped dividing.
36 This pattern is in agreement with previous IHC findings
48 and correlates well with retinal cell cycle activity.
Examination of the
Tert expression pattern shows that steady-state transcript levels fluctuated in ocular tissues between E15 and P3 (
Fig. 4C, upper panel). However, levels of
Tert transcript seemed to drop as retinal cells exited the cell cycle by P11 (
Fig. 4, upper panel). On the protein level, Tert levels increased from E16 to P3, but, unlike the transcript levels that seemed moderately high in the retina at P7, the protein levels were reduced at P7. Surprisingly, Tert levels were increased at P15 and then drop to very low levels by P30 (
Fig. 4, lower panel). It is difficult to propose a function for Tert at P15 as the entire retina is post-mitotic by that age.
36 Although there was one band observed in MCF-7 cells, ocular tissues seemed to express multiple forms of Tert, with a prominent larger form.
What is interesting is that despite being downstream targets of p53, expression patterns of Cdkn1a and Tert message did not follow that of Trp53 transcript. Furthermore, although the decline in the transcript levels of p53 target genes was consistent with their role in cell cycle regulation, it is difficult to understand the behavior of their equivalent proteins. Altogether, this suggests that regulation of these genes may be controlled by additional transcription factors.
Recent evidence shows a strong correlation between the functionality of the p53 pathway and the development of retinoblastoma,
28 leading us to examine the developmental patterns of expression of pRb, its family members p107 and p130, and its targets E2F1, E2F3, E2F4, and E2F5. Support for a developmental role of the retinoblastoma pathway is further highlighted by the early age of onset of retinoblastoma in cases where
Rb is mutated.
Expression of the
Rb transcript started at relatively low levels at E15 but increased by E17 before gradually decreasing to a steady level after P11, when the development of the retina was complete (
Fig. 5A, upper panel).
36 On the protein level, pRb levels increased moderately at E18 but then remained detectable as late as P30 (
Fig. 5B, lower panel), suggesting a role for pRb beyond control of the cell cycle in the post-mitotic retina.
p107 transcript levels were high from E15 through P1, with a statistically insignificant reduction at P3 (
Fig. 5B, upper panel). The
p107 levels dropped significantly by P11 and remained low at all later time points. The pattern of protein expression closely followed the pattern observed for the transcript (
Fig. 5B, lower panel). It is clear that P107 levels are modulated with the cell cycle activity more so than pRb, suggesting that P107 may play a significant role during ocular/retinal development in the mouse. This is in agreement with the role described for p107 as a functional alternative to pRb in the Rb protein-deficient mouse.
49
Although
p130 transcript levels fluctuated during ocular/retinal development, they were maintained at relatively high levels. That the levels of
p130 transcript were relatively high at P30, when expression of both
Rb and
p107 were at their lowest during retinal development, may suggest a role for p130 in the post-mitotic retina. This is consistent with previous work showing that p130 is high in arrested cells.
50 This is also consistent with the relatively low, constant levels of P130 observed throughout all stages of ocular/retinal development tested (
Fig. 5C, lower panel).
The retinoblastoma family of proteins regulates the activity of the E2F protein family to arrest the cell cycle at G
1.
51 pRb binds all E2F proteins, while p107 and p130 bind specifically to the E2F3b, −4, −5, −6, −7a, and −7b transcriptional repressors.
52 We detected higher levels of
E2f1 transcript and protein between E15 and P3 (
Fig. 6A). The subsequent decrease in transcript expression at P7 onward was reflected in the protein level (
Fig. 6A). Although pRb levels remained constant during these time points (
Fig. 5A, lower panel), levels of P107 were considerably high (
Fig. 5B, lower panel). pRb is known to bind to and inactivate E2F1,
53,54 implying that the increase in E2F1 may have been the result of gene expression driven by a cell cycle protein not included in this study.
E2f3 transcript levels remained generally steady, with peaks of expression at E17 and P11 (
Fig. 6B, upper panel). The increase in
E2f3 expression at E17 coincided with the E17 increase observed in the
Trp53 transcript (
Fig. 1A, upper panel). The largest amount of protein was observed in the retina at P7 (
Fig. 6B, lower panel).
In contrast,
E2f4 transcript was maintained at constant levels throughout development with a statistically significant reduction apparent only at P30 (
Fig. 6C, upper panel). The protein seemed to be expressed at constant low levels (
Fig. 6C, lower panel).
Finally, although levels of
E2f5 transcript varied with the progression of ocular/retinal development (
Fig. 6D, upper panel), the protein levels seemed to remain constant throughout the time points examined (
Fig. 6D, lower panel).
To examine the role of p53 in ocular/retinal development and maintenance, transcript and protein steady-state levels for p53, its family members, regulators, and downstream targets were analyzed by qRT-PCR and immunoblotting. This study is especially relevant because the retina becomes post-mitotic by P11,
2 suggesting a role for p53 in cell cycle arrest. Furthermore, because programmed cell death is observed during the development of ocular tissues, p53 may also play a role in apoptosis of the developing eye. This study was performed in the C57BL/6 mouse because it is the strain that has been shown to develop ocular abnormalities in the absence of p53.
14
Error-free execution of the fundamental process of cell division is essential for the continuity of all living organisms. An increase in ocular P53 at late embryonic stages implies that p53 may play a regulatory role in keeping developing ocular cells in check. As cells reach a post-mitotic state, P53 levels decrease to barely detectable levels. However, the expression pattern of the p53 negative regulators Mdm2, Mdm4, and Yy1 did not follow that of P53. The relatively higher expression levels of these regulatory proteins suggest either that high levels are necessary to effectively negatively regulate p53 or that these proteins play other roles in addition to their regulation of p53.
An interesting observation made in our IHC analysis is that the retinal pigment epithelium (RPE) did not express P53 but abundantly expressed Mdm2. This further supports the notion that Mdm2 may play a role in RPE cellular homeostasis, independent of its role in p53 regulation.
Because knocking out p53 showed no ocular phenotype in the C57BL × CBA and 129/Sv × C57/BL6 backgrounds, investigators concluded that p53 plays no role in retinal development or maintenance.
11,12 However, when p53 was later knocked down in BALB/c OlaHsd and C57BL/6 mice, researchers found severe ocular abnormalities such as hyaloid vasculature that does not regress by P35 or P42, as in wild-type mice, which is similar to persistent hyperplastic primary vitreous in humans, a high frequency of cataracts, fibrous retrolental tissue, vitreal opacities, retinal folding, and hypoplastic optic nerves.
14 This finding led the authors to suggest that alleles from the C57BL/6 background may contribute to the ocular abnormalities seen in the absence of p53.
14 Alternatively, the p53 family members p63 and p73 may compensate, leading to no observed phenotypic abnormalities. Both p63 and p73 have been found to induce cell cycle arrest and apoptosis in a manner similar to that of p53.
55 However, an increase in P73 levels preceded that of P53, and its levels dropped significantly when P53 was high, suggesting that P73 may play a role only in early ocular development. Because P63 is generally expressed at steady low levels throughout all developmental stages examined, it may not play a direct role in ocular/retinal development. However, that does not preclude the potential role for p63 in the absence of p53. In fact, the potential role of p73 in early retinal development and the presence of p63 may provide an explanation as to why retinal development continues unabated in the retinas of some p53 knockouts.
Like p53, pRb, p107, and p130 act as tumor suppressors, in part by inhibiting cell cycle progression.
56 Recent findings have shown that misregulation of the p53 pathway plays a major role in the formation of retinoblastoma tumors.
28 One would therefore expect that p53 and pRb are coregulated. However, the levels of p53 were highest at E18 and dropped precipitously thereafter, while the levels of pRb remained constant throughout the developmental stages studied. This suggests that despite its role in retinoblastoma, p53 plays an insignificant role in the regulation of pRb levels.
One observation from the analysis of transcript and protein levels is that while P53 protein levels more or less reflected that of the message, protein levels for the family members and regulators generally reflected neither the levels nor the pattern of message expression. This suggests translational and/or post-translational regulation of these proteins. This may have biological implications whereby ocular/retinal cells can produce the necessary proteins when needed, such as upon stress. That speaks for the importance of p53 function(s) and the necessity for redundancies.
It has been shown that in response to DNA damage by ionizing radiation,
18,19 p53 is phosphorylated, leading to apoptosis.
18 Therefore, it is possible that the only role for p53 in the adult retina is to induce apoptosis in response to damage. Such apoptotic cell death would reduce or prevent the immunological response that would be mounted should damaged retinal cells die via necrosis. However, not all retinal insults cause p53 activation. For example, the retinal response to light damage is p53-independent,
23,24 suggesting that the role of p53 in the retina may be quite specific.
In conclusion, expression of p53 and its family members, regulators, and downstream targets seems to occur mostly prior to ocular/retinal cells entering the post-mitotic state. This implies that, although some of these genes may play a role in retinal development, they play a minimal role, if any, in adult retinal function under normal unstressed conditions.
The authors thank Steven J. Berberich for the generous gift of Mdm4 antibody and Shannon M. Conley for critical evaluation of the manuscript.
Supported by the Foundation Fighting Blindness (MRA), National Center For Research Resources Grant P20RR017703, and National Eye Institute Grants P30EY12190 and R01EY018137 (MRA). The contents are solely the responsibility of the authors and do not necessarily represent the official views of National Institutes of Health or any of its institutes.