Induction of Cell Cycle Entry and Cell Death in Postmitotic Lens Fiber Cells by Overexpression of E2F1 or E2F2

Abstract

purpose. Previous studies have shown that inactivation of the retinoblastoma tumor suppressor protein (pRb) can cause lens fiber cell proliferation and apoptosis. Because pRb is thought to block cell cycle progression by inhibition of E2F transcription factors, experiments were conducted to test whether overexpression of different E2F family members would be sufficient to induce fiber cell proliferation and subsequent apoptosis. The in vivo functions of the transcription factor E2F2 have not previously been analyzed or described in transgenic mice.

methods. Human E2F1 and E2F2 cDNAs were linked to the αA-crystallin promoter. Transgenic mice were generated by microinjection. Changes in cell cycle regulation were assayed by immunohistochemistry for 5-bromo-2′-deoxyuridine (BrdU) incorporation and by in situ hybridization. Cell death was assayed using the TdT-dUTP terminal nick-end labeling (TUNEL) assay.

results. At embryonic day (E)15.5, strong expression of the E2F1 and E2F2 transgenes was detected in lens fiber cells with little or no expression in epithelial cells. BrdU incorporation and TUNEL assays showed that overexpression of either E2F1 or E2F2 in lens fiber cells was sufficient to cause cell cycle entry and subsequent apoptosis. Expression of either E2F1 or E2F2 was sufficient to induce the transcription of cyclins (A2, B1, and E), as well as p53 and Bax in the lens fibercells.

conclusions. Expression of either E2F1 or E2F2 can induce postmitotic lens fiber cells to re-enter the cell cycle. Inappropriate cell cycle entry is recognized by p53 in each case, and programmed cell death ensues.

The retinoblastoma family of tumor suppressers (pRb) and the E2F family of transcription factors (initially identified as cellular factors for induction of the viral gene E2) are central players in cell cycle regulation. The E2F proteins heterodimerize with DP (DRTF1-polypeptide) family members to form active transcription factors. The transcriptional activity is inhibited by binding to hypophosphorylated pRb family members. Entry into the S phase occurs when the pRb/E2F/DP complexes are disrupted by cyclin-dependent kinase (Cdk)–mediated phosphorylation of pRb, leading to release of active E2F/DP heterodimers and expression of target genes that are necessary for the G1/S cell cycle transition and DNA replication. ^{1 2 3} To date, six E2F family members (E2F1, E2F2, E2F3, E2F4, E2F5, and E2F6) have been found in mammalian cells. They can be subdivided into three subfamilies that show distinct affinities for pRb family members. The first subfamily contains E2F1, 2, and 3, which bind with high affinity to pRb; the second subfamily includes E2F4 and 5, which bind with high affinity to the pRb homologues, p107 and p130; the third subfamily contains E2F6, a novel E2F species, which binds none of the pRb family members. ^{3 4 5} Previous studies have shown that almost all the E2F family members can induce cell proliferation, and some of them can induce subsequent apoptosis when overexpressed in cultured cells. ^{6 7 8} The mechanisms by which individual E2Fs induce apoptosis when overexpressed are largely unknown. 

The embryonic lens of the eye is an attractive model system for studying the molecular mechanisms that regulate cell proliferation and differentiation. The lens is composed of a monolayer of proliferative cuboidal epithelial stem cells on the anterior surface overlaying a core of terminally differentiated, postmitotic, elongated fiber cells. ^{9 10} At the equatorial region of the lens, epithelial cells are induced to exit from the cell cycle and to differentiate into fiber cells. Almost all E2F family members are expressed in the lens epithelial cells, whereas only E2F1, E2F3, and E2F5 are expressed in the postmitotic lens fibers. ¹¹ Previous studies have shown that inactivation of pRb in lens fiber cells, either by expression of viral proteins or by targeted mutagenesis of the RB gene, results in inappropriate cell proliferation and p53-dependent cell death. ^{12 13 14} A recent study indicates that human papilloma virus (HPV) gene E7-induced proliferation in the differentiated lens fibers is partially dependent on E2F1. ¹⁵  

In the present study, we generated transgenic mice that overexpressed E2F1 and E2F2 in the fiber cells of the lens. In our experiments, both E2F1 and E2F2 induced lens fiber cells to re-enter the cell cycle. In each case, inappropriate cell cycle entry activated p53, resulting in upregulation of p21 and Bax, followed by fiber cell apoptosis. 

Methods

Generation of the Constructs and Transgenic Mice

HindIII-EcoRI fragments encoding human E2F1 and E2F2 cDNA were cloned into the αA-crystallin promoter vector CPV2. ¹⁶ The resultant plasmids (Fig. 1) were digested with NotI to release 2.5-kb fragments for microinjection. The fragments were isolated by electrophoresis through a 1% agarose gel, and purified (Geneclean; Bio 101, Vista, CA). Transgenic mice were generated by pronuclear injection of the purified fragments into one-cell-stage inbred FVB/N embryos. ^{17 18}  

Screening of Mice

Genomic DNA from mouse tails or embryonic torsos was isolated as previously described. ¹⁷ For polymerase chain reaction (PCR) screening, ¹² an upstream sense primer (5′-GTGAAGGAACCTTACTTCTGTGGTG) and a downstream antisense primer (5′-GTCCTTGGGGTCTTCTACCCTTTCTC) specific for the simian virus (SV)40 sequences in CPV2 were used to amplify a 300-bp fragment. PCR assays were performed in a 30-μl volume of 1× PCR buffer (10 mM Tris-HCl [pH 8.0]; 50 mM KCl; 1.5 mM MgCl₂; 0.1% gelatin), 1 μl of tail DNA, 0.1 mM dNTPs, 1.0 μM (each) primer, and 2.5 U Taq DNA polymerase (Promega; Madison, WI). Reactions were run for 25 cycles at 94°C for 30 seconds, 58°C for 30 seconds, and 72°C for 1 minute. 

Lens Histology

Embryonic heads at embryonic day (E)15.5 were fixed in 10% formalin, paraffin embedded, cut into 5-μm-thick sections, and stained with hematoxylin and eosin by standard techniques. 

In Situ Hybridization

In situ hybridization was performed using ³⁵S-labeled riboprobes, as described in Fromm and Overbeek. ¹⁹ To test for transgene expression, an EcoRI/HindIII fragment of wild-type SV40 T antigen cDNA (4280–4558 bp) was subcloned into pBluescript KS− (Stratagene, La Jolla, CA), and used to generate an SV40-specific riboprobe. The probes for mouse genes involved in cell cycle regulation were generated from the following mouse cDNAs: p27 from Hideo Toyoshima (The Salk Institute, San Diego, CA); p21 and p57 from Stephen Elledge (Baylor College of Medicine, Houston, TX); cyclin E from Julie A. Deloia (Magee–Womens Research Institute, Pittsburgh, PA); cyclins A2 and B1 from Debra Wolgemuth (Columbia University, New York, NY); cyclins D1, D2, and D3 from Charles Sherr (St. Jude Children’s Hospital, Memphis, TN); p53 from Gigi Lozano (MD Anderson Cancer Center, Houston, TX); and Bax from Stanley Korsmeyer (Washington University, St. Louis, MO). Hybridization signals were initially captured as dark-field images. For some of the figures, the dark-field images were pseudocolored red, then superimposed on bright-field images of the same tissue section (counterstained by hematoxylin) using image analysis software (Photoshop; Adobe, San Diego, CA). 

BrdU Incorporation

For detection of DNA replication, pregnant female mice were injected with 5-bromo-2′-deoxyuridine (BrdU; Sigma, St. Louis, MO) and killed 1 hour later. Embryos were analyzed for BrdU incorporation by immunohistochemistry as described. ¹² For quantification, the number of BrdU-positive nuclei in lens fiber cells was counted and compared with the total number of nuclei in the same region, determined by hematoxylin staining. 

Detection of Apoptosis

DNA fragmentation was detected using an in situ apoptosis detection kit (apo TACS; Trevigen, Gaithersburg, MD). The kit incorporates biotinylated nucleotides onto free 3′ ends using terminal deoxynucleotidytransferase (TdT). This type of assay is often referred to as a TUNEL assay. ²⁰ Tissue sections from embryos at E15.5 were dewaxed, rehydrated, treated with proteinase K at 37°C for 30 minutes, peroxidase quenched, washed in H₂O, and treated with the TdT reaction mixture at 37°C for 1 hour. After washing with PBS, the slides were incubated with horseradish peroxidase–conjugated streptavidin for 1 hour at room temperature. Diaminobenzidine (DAB) was used for detection of the enzyme conjugate, as described. ¹² Slides were counterstained with methyl green, dehydrated, and mounted. For quantification, the number of apoptotic nuclei was counted and compared with the total number of nuclei. 

Results

Expression of E2F1 and E2F2 in Lenses of Transgenic Mice

Human E2F1 or E2F2 cDNAs were linked to the lens-specific mouseα A-crystallin promoter and the constructs (Fig. 1A) were used for microinjections to generate transgenic mice. Two stable transgenic families were generated for the E2F1 construct (OVE 527 and OVE 530). Because mice in both families had the same phenotype (small eyes and cataracts, see Figs. 1C 1D ), we chose to characterize embryos from one family (OVE 527) in detail (Table 1) . For the E2F2 construct, we generated three founder transgenic embryos at E15.5. Two of the embryos (B279 and B283) had lens defects and were characterized in detail (Table 1) . Stable E2F2 families were not generated. The E2F1 and E2F2 transgenic embryos showed similar defects in lens fiber cell differentiation (Figs. 2A 2B 2C 2D 2E 2F) , as discussed in more detail in the next section. In situ hybridizations showed that transcripts of both transgenes were present specifically in lens fiber cells (Figs. 3A 3B 3C ). There was no transgene expression in the lens epithelial cells or in other regions of the eye (Figs. 3B 3C) . Previous studies have similarly found that the αA-crystallin promoter is activated in transgenic mice only in lens fiber cells that have already exited from the cell cycle and begun to elongate. ^{12 16 21}  

Lens Histology

At E15.5, the E2F1 and E2F2 transgenic lenses showed defects in fiber cell elongation and fiber cell alignment as well as extra nuclei at the posterior of the lens (Figs. 2A 2B 2C 2D 2E 2F) . Many nuclei in the center of the lens had condensed chromatin indicative of apoptosis (Figs. 2B 2C 2E 2F) . Newborn E2F1 transgenic mice showed a hollow lens with no fiber cells extending to the posterior surface of the epithelial cells (Fig. 2H) . The transgenic lens was smaller than normal (compare Figs. 2G 2H ) and had fewer epithelial cells due to anteriorization of the transition zone (arrowheads in Fig. 2H ). The anteriorization probably reflects the fact that the anterior retina encircles a larger portion of the smaller lens (Fig. 2H) . The adult E2F1 transgenic mice had microphakia with disorganized, vacuolated, poorly elongated fiber cells (Fig. 2J) . 

Cell Cycle Regulation

Newly induced fiber cells exit from the cell cycle and discontinue BrdU incorporation at a specific region of the equatorial zone of the lens. ¹⁹ As a result BrdU-positive cells are always epithelial cells in nontransgenic lenses (Fig. 3G) . In E2F1 (Fig. 3H) and E2F2 (Fig. 3I) transgenic lenses, the pattern of BrdU incorporation in the epithelial cells was similar to the nontransgenic pattern (Fig. 3G) . In both E2F1 and E2F2 mice, the lenses had a contiguous set of young fiber cells at the equatorial zone that had exited from the cell cycle and that had not yet begun to express the transgene (bracketed regions in Figs. 3E 3F 3H 3I ). Once transgene expression began in the more mature fiber cells (Figs. 3E 3F , posterior to the brackets), BrdU incorporation was induced. BrdU incorporation was prevalent in the fiber cells expressing either the E2F1 or the E2F2 transgenes (Figs. 3H 3I) . These results demonstrate that both E2F1 and E2F2 can induce differentiated fiber cells to re-enter the cell cycle. In E2F1 transgenic embryos an average of 24% of the lens fiber cells were BrdU positive, whereas in E2F2 transgenic embryos a lower percentage of lens fiber cells (6%) were BrdU positive (Table 1) . 

Expression of Cyclins

Progression through the cell cycle is often regulated by the activity of Cdks. ^{22 23 24} Activation of transcription of cyclins or inactivation of transcription of Cdk inhibitors (e.g., p27 or p57) could be responsible for S-phase induction in the transgenic fiber cells. ^{22 24 25} To assay for changes in gene expression induced by E2F1 or E2F2, in situ hybridization was performed using probes for cyclin family members (A2, B1, D1, D2, D3, and E), as well as Cdk inhibitors (p21, p27, and p57). The results are summarized in Table 1 . Wild-type lens epithelial cells expressed the G1/S cyclins D2, D3, and A2 as well as the G2/M cyclin B1 (Fig. 4) . Cyclins A2 and B1 were expressed in a subset of the epithelial cells (Figs. 4A 4D) , suggesting that their expression is cell cycle regulated. In contrast, the fiber cells that exited from the cell cycle (see Fig. 3G ) ceased to express cyclins A2 and B1 (Figs. 4A 4D) . The D-type cyclins (D2 and D3) were expressed constitutively by the epithelial cells and newly differentiating fiber cells (Figs. 4J 4M) . Cyclin E expression was not detected in the normal lens (Fig. 4P) . In transgenic lenses, we found that lens fiber cell–specific E2F1 or E2F2 expression induced expression of cyclin E (Figs. 4Q 4R) and cyclin A2 (Figs. 4B 4C) . The expression of cyclin D3 also appeared to be upregulated (Figs. 4N 4O) . In addition, induction of cyclin B1 expression was observed in fiber cells expressing either E2F1 or E2F2 (Figs. 4E 4F) , indicating that these cells had progressed to the G2/M phase of the cell cycle. E2F1 was slightly more active than E2F2 in inducing transcription of cyclin A2 (Figs. 4B 4C) and cyclin B1 (Figs. 4E 4F ; Table 1 ), consistent with the higher percentage of BrdU-positive fiber cells. Cyclins A2, B1, and D1 are well expressed in the proliferating cells of the wild-type and transgenic retinas (Figs. 4A 4B 4C 4D 4E 4F 4G 4H 4I) . Although the FVB mice have a mutation (rd) that causes degeneration of mature photoreceptor cells, ¹⁸ the pattern of prenatal cellular proliferation in the retina appears to be normal. 

Expression of Cdk Inhibitors

The Cdk inhibitor p57 is one of the earliest genes expressed during lens fiber cell differentiation. ²⁶ Its expression is induced when fiber cells are stimulated to differentiate (Fig. 5D ), and it is required for the cell cycle exit that accompanies fiber cell differentiation. ²⁶ In embryonic lenses, p27 upregulation was also seen to coincide with exit from the cell cycle during fiber cell differentiation (Fig. 5A) . In E2F1 and E2F2 transgenic mice, p27 (Figs. 5B 5C) and p57 (Figs. 5E 5F) were still induced at the equatorial zone (Fig. 5D , ez), and both genes were expressed in the proliferating fiber cells, suggesting that expression of p27 and p57 is not sufficient to block E2F-induced entry into the S phase. 

Apoptosis in the Lens

Previous studies in tissue culture cells have shown that overexpression of E2F1 can cause p53-dependent apoptosis. ^{8 27} E2F2 and E2F3 can also induce cell death in cultured cells. ⁷ To test whether inappropriate cell cycle entry induced in vivo by E2F1 or E2F2 expression can also result in programmed cell death, E15.5 lenses from transgenic embryos were analyzed by TUNEL assay. ²⁰ Apoptotic nuclei were prevalent in lens fiber cells expressing either E2F1 or E2F2 (Figs. 3K 3L) . The average percentage of apoptotic fiber cells was 25% for E2F1 versus 14% for E2F2 (Table 1) . 

To ascertain whether cell death in E2F1 and E2F2 transgenic lenses involves activation of p53, in situ hybridization was used to assay for expression of p53 and the p53-regulated genes, Bax and p21. ^{28 29} Transcripts of p53 and Bax were increased in fiber cells expressing either E2F1 or E2F2 (Figs. 6E 6F 6H 6I ) compared with nontransgenic lenses where transcripts were not detected (Figs. 6D 6G) . In addition, there was a significant increase in p21 expression (Figs. 6B 6C) , consistent with the prediction that inappropriate cell cycle entry leads to stabilization and activation of p53. ^{12 13 14}  

Discussion

We have used the ocular lens in transgenic mice as a model system to show that E2F1 or E2F2 expression is sufficient to induce postmitotic lens fiber cells to re-enter the cell cycle, followed by activation of p53 and subsequent programmed cell death. A model depicting the changes in gene expression and cell cycle regulation induced by E2F1 and E2F2 is provided in Figure 7 . 

Previous studies have shown that pRb inactivation causes aberrant proliferation and apoptosis in the developing nervous system and the ocular lens that is mediated in part by E2F1. ^{4 15} Our transgenic studies show that elevated E2F1 expression is sufficient to induce fiber cell proliferation and subsequent cell death, consistent with these previous results. It has been suggested that almost all E2F family members can induce cell proliferation when overexpressed in specific cell lines. ^{6 8} In one study only E2F1, not the other members of the E2F family, induced apoptosis. ⁶ In contrast, a different study indicated that E2F2 and E2F3 can also be inducers of apoptosis. ⁷ Our present transgenic study has demonstrated that E2F2 can indeed induce both cell proliferation and apoptosis. 

Previous research has shown that fiber cells exit from the cell cycle and discontinue BrdU incorporation at a specific region of the equatorial zone of the lens. ¹⁹ At the same time that BrdU incorporation is stopped, the cells stop expressing cyclin A2 (Fig. 4) and Cdk2, ¹⁹ and they upregulate expression of p57 (Fig. 5) , a Cdk inhibitor. ²⁶ These changes in gene expression occur before any fiber cell elongation at the bow region of the lens. Because normal fiber cells never again enter the cell cycle (see Fig. 3G ), these data imply that the cells have become postmitotic and have entered G0. The αA-crystallin promoter is activated in transgenic mice only in lens fiber cells that have already exited from the cell cycle and begun to elongate (Figs. 3E 3F) . ^{12 16} The E2F transgenes are not expressed in the anterior epithelial cells and are not expressed in the fiber cells until after the initial cell cycle exit has occurred (boxed regions in Fig. 3 ). 

When p57 expression is induced at the equatorial region of the normal lens, the cells that upregulate p57 initiate the pathway of fiber cell differentiation. ²⁶ The temporal and spatial correlation between cell cycle exit and p57 upregulation suggests that p57 expression is sufficient to inhibit the Cdk activity in these cells. As a result pRb should remain hypophosphorylated, and there is apparently sufficient pRb present to sequester and inhibit all the E2Fs that are normally expressed by these cells. Inhibition of E2F activity appears to be required for the cells to remain postmitotic, because inactivation of Rb is sufficient to cause the cells to re-enter the cell cycle. ^{12 13 14} In the transgenic mice, the level of E2F expression is predicted to exceed the binding capacity of the pRb that is present, so that even in the presence of p57, there is sufficient E2F activity to cause the cells to enter the S phase of the cell cycle. The two E2Fs appear to have nearly redundant activities, because overexpression of either is sufficient to induce fiber cells to re-enter the cell cycle and to upregulate expression of other cell cycle–controlling genes. 

Cell cycle progression is thought to be regulated by cyclin/Cdk complexes. In this transgenic study, we found that both E2F1 and E2F2 can induce the expression of cyclin A2 and cyclin E, consistent with previous studies. ^{30 31 32} The absence of cyclin E expression in wild-type lens epithelial cells (Fig. 4P) suggests that neither E2F1 nor E2F2 is activated during normal lens cell proliferation. Presumably, lens epithelial cell proliferation can occur in the absence of E2F activity or involves the activation of a different E2F family member that does not stimulate cyclin E expression. Although the expression of cyclin B1 was also upregulated by E2F1 or E2F2, this is presumably an indirect consequence of cell cycle re-entry. It has also been reported that E2F1 overexpression in mesangial cells can increase the expression of cyclin D1. ³³ In contrast with this, lens-specific E2F1 expression did not activate the transcription of cyclin D1. Because E2Fs function downstream from the D-type cyclins, the upregulation of cyclin D2 and D3 expression by expression of E2F1 or E2F2 in lens fiber cells (Figs. 4K 4N 4O) may reflect either the increase in cell number or the absence of fiber cell maturation and denucleation. 

E2F1 overexpression has been shown to lead to p53-dependent and -independent apoptosis in tissue culture cells and transgenic mice. ^{7 8 27} In addition, E2F1 expression has been shown to increase transcription of p53. ²⁷ Various downstream target genes of p53 have been identified, including Bax, p21, MDM2, GADD45, and cyclin G. ^{28 29 34} These genes function as regulators of diverse aspects of cell growth and cell death. ^{35 36} In the present study, transgenic lenses were assayed by in situ hybridization for expression of p53, Bax, and p21. Transcription of p53 was upregulated by lens-specific expression of E2F1 or E2F2. Bax and p21 transcription were also upregulated. Although we have not demonstrated that the Bax expression is p53-mediated, the results suggest that overexpression of either E2F1 or E2F2 induces inappropriate cell proliferation, which induces (directly or indirectly) activation of p53, which then induces Bax expression and leads to cell death (Fig. 7) . The precise mechanism by which E2F expression and cell cycle entry activate p53 and cause apoptosis in lens fiber cells remains to be determined. 

 Supported by National Institutes of Health Grants EY-10448, EY-10803, and EY-11348 (PAO).

 Submitted for publication May 1, 2000; revised July 24, 2000; accepted August 16, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

E2F transgenes. (A) Microinjected constructs, including theα A-crystallin promoter; E2F1 and E2F2 coding sequences; and the SV40 intron/polyA region. (B) Nontransgenic FVB mouse. (C) Heterozygous (OVE 530) E2F1 transgenic mouse. (D) Homozygous (OVE 527) E2F1 transgenic mouse. The transgenic mice have cataracts and microphthalmia.

Table 1.

 

Fiber Cell Characteristics in E2F1 and E2F2 Transgenic Embryos at E 15.5

Transgenic Constructs	n	Gene Expression in Fiber Cells*																			BrdU, ‡ (%)	TUNEL, ‡ (%)
				Transgene	cyc.A2	cyc.B1	cyc.D1	cyc.D2	cyc.D3	cyc.E	p21	p27	p57
;l>E2F1	2	+++	+++	+++	−	++	++	+++	+++	+† >	+++	24 ± 4.0	25 ± 4.5
E2F2	2	+++	++	++	−	+	++	+++	++	+† >	+++	6 ± 1.0	14 ± 1.6
FVB	2	−	−	−	−	−	+	−	−	+, †	++, †	0	0

*  Expression levels in lens fiber cells, assayed by in situ hybridization: −, background level of expression; +, weak expression;++ , modest expression; +++, high expression.

†  Expression in equatorial zone.

‡  Percentage of positive cells: the number ± SD of brown nuclei in lens fiber cells compared with the total number of fiber cell nuclei.

Figure 2.

 

Ocular histology. Histology sections of eyes from nontransgenic FVB (A, D, G, and I), E2F1 (B, E, H, and J), and E2F2 (C, F) transgenic mice at E15.5 (A through F), birth (G, H), and 3 months of age (I, J). (D, E, and F) Higher magnifications of (A), (B), and (C). At E15.5, the transgenic lenses showed disruption of fiber cell elongation, the presence of extra nuclei in the center of the lens, and condensation of fiber cell nuclei indicative of apoptosis (E, F). Newborn and adult transgenic eyes had small lenses (microphakia) with poorly elongated and defective fiber cells (H, J). Arrowheads indicate the anterior shift of the boundary between epithelial cells and fiber cells (H). co, cornea; le, lens epithelium; lf, lens fibers; nr, neuronal retina. Scale bars, 500 μm.

Figure 3.

 

Patterns of E2F1 and E2F2 transgene expression, BrdU incorporation, and apoptosis. Nontransgenic FVB (A, D, G, and J), E2F1 transgenic (B, E, H, and K), and E2F2 transgenic (C, F, I, and L) mice at E15.5 were assayed for transgene expression (A through F), BrdU incorporation (G, H, and I), and the presence of cells undergoing apoptosis (J, K, and L). Assays for transgene expression were performed by in situ hybridization with an SV40 riboprobe. (A, B, and C) Hybridization signals were initially captured as dark-field images, pseudocolored red, and superimposed on bright-field images of the same tissue sections counterstained by hematoxylin. (D, E, and F) Higher magnification views without pseudocoloring. Transgene expression was localized to fiber cells posterior to the equatorial zone in E2F1 (B, E) and E2F2 (C, F) transgenic mice (squares bracket the newly induced fiber cells in the equatorial region that had exited from the cell cycle but had not yet begun to express the transgene). No expression was detected in lens epithelial (le) cells or other regions of the eye. In the region of the lens where the transgenes were expressed, most but not all the fiber cells showed nuclear accumulation of the transgenic transcripts (E, F). BrdU incorporation was assayed by immunohistochemistry. In the wild-type lens, BrdU-positive cells (brown nuclear stain) are restricted to the epithelial (le) cells (G). In the E2F1 and E2F2 transgenic mice, there are BrdU-positive fiber cells (H, I), with a higher percentage of cells positive in the E2F1 lens (Table 1) . The BrdU-positive fiber cells are restricted to the region of transgene expression, which is posterior to the equatorial zone (bracketed by squares). For the TUNEL assays, brown nuclear stain indicates DNA fragmentation, a characteristic feature of apoptosis. TUNEL-positive fiber cells are present in the transgenic lenses (K, L), but not the wild-type lens (J). lf, lens fiber. Scale bars, 500 μm.

Figure 4.

 

Changes in expression of cyclins. In situ hybridization was used to assay for cyclin expression in nontransgenic (A, D, G, J, M, and P), E2F1 (B, E, H, K, N, and Q) and E2F2 (C, F, I, L, O, and R) transgenic eyes. The in situ hybridization images were captured by dark-field illumination. In nontransgenic (FVB) lenses, the S-phase and G2/M-phase cyclins A2 and B1 were expressed in a punctuate pattern in the epithelial cells, but expression of these cyclins was turned off at the equatorial zone, and fiber cells did not express either cyclin (A, D). Cyclin E expression was not detected in the wild-type lens (P). In E2F1 and E2F2 transgenic lenses, these cyclins (A2, B1, and E) were all induced in the transgene-expressing lens fiber cells. Cyclins D2 and D3 also appeared to be upregulated in the E2F1 fiber cells (K, N). The expression of cyclins A2 and B1 in fiber cells expressing E2F2 (C, F) was weaker than E2F1 (B, E) but was still punctate, suggesting cell cycle regulation. Expression of cyclin E was more ubiquitous in the transgenic fiber cells, consistent with the notion that E2Fs can directly regulate transcription of cyclin E. Scale bar, 500 μm.

Figure 5.

 

Expression of Cdk inhibitors. In situ hybridization was used to assay for expression of p27 (A, B, and C) and p57 (D, E, and F) Cdk inhibitors in E15.5 eyes from nontransgenic FVB (A, D), E2F1 transgenic (B, E), and E2F2 transgenic (C, F) mice. Hybridization signals were captured as dark-field images. The expression of p27 in nontransgenic (A) lens was weak and localized primarily to the equatorial zone (ez). In the transgenic mice, p27 expression was maintained in the lens fiber cells (B, C) in contrast to normal fiber cells where p27 expression was turned off during fiber cell maturation (A). In the nontransgenic (D) lens, newly differentiating fiber cells at the equatorial zone (ez) showed strongly induced expression of p57 that coincided perfectly with exit from the cell cycle and loss of cyclin A2 and B1 transcription (Figs. 3 4) . In E2F1 and E2F2 transgenic mice (E, F), p57 expression was upregulated appropriately at the equatorial zone and was maintained in the proliferating lens fibers. Scale bars, 500μ m.

Figure 6.

 

Expression of p21, p53, and Bax in the lens. In situ hybridization was used to test for expression of p21, p53, and Bax in nontransgenic FVB (A, D, and G), E2F1 (B, E, and H), and E2F2 (C, F, and I) transgenic lenses. Hybridization signals were captured as dark-field images. Expression of p21, which can be induced by p53, was not detected in the wild-type lens (A) but was elevated in the transgenic fiber cells (B, C). Expression of p53 and Bax was also not detected in nontransgenic lenses (D, G), but was detected in epithelial cells and fiber cells in the transgenic lenses expressing E2F1 (E, H) or E2F2 (F, I). Elevated Bax expression could cause (or contribute to) fiber cell apoptosis. Scale bars, 500 μm.

Figure 7.

 

Consequences of E2F1 and E2F2 expression in fiber cells. Our transgenic studies showed that E2F1 or E2F2 expression in postmitotic lens fiber cells was sufficient to induce the expression of S-phase cyclins (A2 and E), helping the fiber cells to re-enter the cell cycle (S phase). By an unknown mechanism, the inappropriate cell cycle entry leads to induction of p53 expression (and presumably also to accumulation of p53 protein), followed by upregulation of the transcription of the target genes, p21 and Bax, and subsequently to cell death.

The authors thank Barbara Harris and Gaby Schuster for excellent technical assistance.