Minireview: Apoptosis as seen through a lens

The lens represents an ideal model system for studying many of the cellular and molecular events of differentiation. It is composed of two ectodermally-derived cell types: the lens epithelial cells and the lens fibre cells, which are derived from the lens epithelial cells by differentiation. Programmed removal of nuclei and other organelles from the lens fibre cells ensures that an optically clear structure is created, while the morphology of the degenerating nuclei is similar to that observed during apoptosis and is accompanied by DNA fragmentation. These observations suggest the existence of biochemical parallels between the process of lens fibre cell organelle loss and classical apoptosis. For example, proteins encoded by the bcl-2 and caspase gene families are expressed in developing lenses and nuclear degeneration in lens fibre cells can be inhibited in vivo by overexpression of bcl-2 and in vitro by incubation of differentiating lens epithelial cell cultures with caspase inhibitors. Thus, the developing lens may represent a particularly useful model system for researchers interested in apoptosis. In this review, the recent literature pertaining to lens fibre cell organelle loss and its relationship to apoptosis is reviewed and possible future research directions are suggested.     

Junctions between lens cells in differentiating cultures: structure, formation, intercellular permeability, and junctional protein expression

We previously described cultures of chick embryo lens cells which displayed a marked degree of differentiation. In this report, the junctions found between the lens fiber-like cells in the differentiated "lentoids" are characterized in several ways. Thin-section methods with electron microscopy first demonstrated that numerous, large junctions between lentoid cells accompanied the other differentiated features of these cells. Freeze-fracture techniques, including quantitative analysis, then revealed that (a) junctional particles were loosely arranged as is typical of fiber cells, (b) the population of individual junctional areas in culture was indistinguishable from that found in 10- to 12-day chick embryo lenses, and (c) apparent junction formation occurred during the development of the lens cells, with lacy arrays of particles being associated with fiber-like junctions. In addition, gap junctions with hexagonally packed particles, typical of lens epithelial cells, largely disappeared during the course of differentiation. Injection of tracer dyes into lentoid cells resulted in rapid intercellular movement of dye, consistent with functional cell-to-cell channels connecting lentoid cells. During the development of the lens cells in culture, as junction formation occurred, an increase of approximately eight-fold in MP28 protein was observed within the cells. These combined results indicate that (a) extensive lens fiber junctions and functional cell-to-cell channels are found between differentiated lentoid lentoid cells in vitro, (b) lens fiber junctions appear to form during the course of lens cell differentiation in culture, (c) a significant increase occurs in the putative junctional protein before the cultures are highly developed, (d) the increased levels of MP28 and junction formation may be required for the full expression of the differentiated state in the lens fiber cell, and (e) this culture system should prove to be valuable for additional experiments on lens junctions and for other studies requiring the development of lens fiber cells in vitro.     

Chicken embryo lens cultures mimic differentiation in the lens

Embryonic chicken lenses, which had been disrupted by trypsin, were grown in culture. These cultures mimic lens development as it occurred in vivo, forming lens-like structures known as lentoids. Using a variety of techniques including electron microscopic analysis, autoradiography, immunofluorescence, and polyacrylamide gel electrophoresis, it was shown that the lentoid cells had many characteristics in common with the differentiated cells of the intact lens, the elongated fiber cells. These characteristics included a shut off of DNA synthesis, a loss of cell organelles, an increase in cell volume, an increase in δ-crystallin protein, and the development of extensive intercellular junctions. The cultures began as a simple epithelial monolayer but then underwent extensive morphogenesis as they differentiated. This morphogenesis involved three distinctive morphological types which appeared in sequence as an epithelial monolayer of polygonal shaped cells with pavement packing, elongated cells oriented end to end, and the multilayered, multicellular lentoids. These distinct morphological stages of differentiation in culture mimic morphogenesis as it occurs in the lens.     

Lens Fiber Cell Differentiation and Denucleation Are Disrupted through Expression of the N-Terminal Nuclear Receptor Box of Ncoa6 and Result in p53-dependent and p53-independent Apoptosis

Nuclear receptor coactivator 6 (NCOA6) is a multifunctional protein implicated in embryonic development, cell survival, and homeostasis. An 81-amino acid fragment, dnNCOA6, containing the N-terminal nuclear receptor box (LXXLL motif) of NCOA6, acts as a dominant-negative (dn) inhibitor of NCOA6. Here, we expressed dnNCOA6 in postmitotic transgenic mouse lens fiber cells. The transgenic lenses showed reduced growth; a wide spectrum of lens fiber cell differentiation defects, including reduced expression of γ-crystallins; and cataract formation. Those lens fiber cells entered an alternate proapoptotic pathway, and the denucleation (karyolysis) process was stalled. Activation of caspase-3 at embryonic day (E)13.5 was followed by double-strand breaks (DSBs) formation monitored via a biomarker, γ-H2AX. Intense terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) signals were found at E16.5. Thus, a window of ∼72 h between these events suggested prolonged though incomplete apoptosis in the lens fiber cell compartment that preserved nuclei in its cells. Genetic experiments showed that the apoptotic-like processes in the transgenic lens were both p53-dependent and p53-independent. Lens-specific deletion of Ncoa6 also resulted in disrupted lens fiber cell differentiation. Our data demonstrate a cell-autonomous role of Ncoa6 in lens fiber cell differentiation and suggest novel insights into the process of lens fiber cell denucleation and apoptosis.     

The lens epithelium and radiation cataract

The effect of radiation upon the differentiating meridional row cells of the rat lens epithelium was studied by electron microscopy. X-irradiation of rat lenses was shown to disrupt the nuclear and, in some instances, the cytoplasmic organization of these cells. Indications of damage, 4.5 h post-irradiation, are a disruption in the shape of the nucleus and its membranes. This is typified by a bulging or blebbing of the nucleus. At 12 h the affected nuclei deteriorate with obvious membrane breakdown. The nucleoplasm at this time may appear highly condensed and shrunken. Disruption of the plasma membrane is also evident. Based upon previous findings it is likely that the nuclear changes are independent of the processes involved in normal fiber denucleation.     

Junctions between lens fiber cells are labeled with a monoclonal antibody shown to be specific for MP26

A monoclonal antibody (mcAb) that recognizes an intracellular domain of the major lens membrane protein in both chicken and bovine lenses is described. Mice were immunized with chicken lens fiber cell membranes that had been washed with 7 M urea. Hybridomas were screened by means of enzyme-linked immunosorbent assays and the molecular specificities of the mcAbs were determined using electrophoretic transfer procedures, "Westerns." One of these mcAbs, an IgG designated B2, reacted with a single band of 28,000 Mr from the chicken embryo lens (MP28) and the analogous 26,000 Mr protein in the bovine lens (MP26). Monoclonal B2 was shown to be specific for these proteins, since (a) heating in SDS caused MP26 to aggregate and reduced B2 binding to the protein band at an Mr of 26,000 in Western transfer analysis; (b) apparent dimers were bound by B2 in Western transfers; (c) soluble protein fractions from the lens contained no detectable B2 antigens; and (d) a cyanogen bromide fragment of MP26 was bound by B2. Studies with several proteases indicated that the antigenic site for B2 resides on a 2-kd, protease-sensitive region at the C-terminal end of MP26 and MP28. Evidence for B2 binding on the cytoplasmic side of the membrane comes from labeling studies done at the ultrastructural level. These studies, utilizing indirect methods with peroxidase and colloidal gold markers, clearly demonstrated that B2 labels two types of junctional profiles. In our calf lens membrane preparations after tannic acid staining, the predominant type (80%) measured 16-18 nn thick, with the second type measuring only 12-14 nm. Chick embryo lens cells that had differentiated in vitro and formed groups of lens fiber-like cells (termed lentoids), fluoresced brightly only when they had been permeabilized before labeling with B2 and a fluorochrome-conjugated antibody. This binding was concentrated at the plasma membranes of cells within the lentoids, even outside areas of cell-cell contact. Surrounding epithelioid cells were not stained. Solubilized lens cultures, examined by Westerns, displayed a single immunoreactive band, which co-migrated with MP28.     

Members of the bcl-2 and caspase families regulate nuclear degeneration during chick lens fibre differentiation.

The optical clarity of the lens is ensured by the programmed removal of nuclei and other organelles from the lens fibre cells during development. The morphology of the degenerating nuclei is similar to that observed during apoptosis and is accompanied by DNA fragmentation. Proteins encoded by the bcl-2 proto-oncogene family are important in either promoting or inhibiting apoptosis, and caspases are involved in downstream proteolytic events. Here, the expression of bcl-2 family members (bcl-2, bax, bad, and bcl-x(s/l)) and caspases-1, -2, -3, -4, and -6 was investigated through a range of stages of chick lens development using immunocytochemistry, Western blotting, and affinity labelling for caspases using biotinylated caspase inhibitors. Using differentiating lens epithelial cell cultures, it was demonstrated that the addition to cultures of synthetic peptide inhibitors of caspases -1, -2, -4, -6, and -9 brought about a 50-70% reduction in the number of degenerating nuclei per unit area of culture, as assessed by image analysis. These effects were comparable to those seen when general inhibitors of caspases were added to cultures. On the other hand, inhibitors of caspases-3 and -8 were not effective in significantly reducing the number of TUNEL-labelled nuclei. Expression of the caspase substrates poly(ADP-ribose) polymerase (PARP) and the 45-kDa subunit of DNA fragmentation factor (DFF 45) was also observed in the developing lens. Western blots of cultures to which caspase inhibitors were added revealed alterations in the PARP cleavage pattern, but not in that of DFF. These results demonstrate a role for members of the bcl-2 family and caspases in the degeneration of lens fibre cell nuclei during chick secondary lens fibre development and support the proposal that this process has many characteristics in common with apoptosis.     

Increased O-GlcNAc causes disrupted lens fiber cell differentiation and cataracts

Diminished proteolytic functionality in the lens may cause cataracts. We have reported that O-GlcNAc is an endogenous inhibitor of the proteasome. We hypothesize that in the lens there is a cause-and-effect relationship between proteasome inhibition by O-GlcNAc, and cataract formation. To demonstrate this, we established novel transgenic mouse models to over-express a dominant-negative form of O-GlcNAcase, GK-NCOAT, in the lens. Expression of GK-NCOAT suppresses removal of O-GlcNAc from proteins, resulting in increased levels of O-GlcNAc in the lenses of our transgenic mice, along with decreased proteasome function. We observed that transgenic mice developed markedly larger cataracts than controls and lens fiber cell denucleation was inhibited. Our study suggests that increased O-GlcNAc in the lens could lead to cataract formation and attenuation of lens fiber cell denucleation by inhibition of proteasome function. These findings may explain why cataract formation is a common complication of diabetes since O-GlcNAc is derived from glucose.     

Cellular distribution of lens epithelium-derived growth factor (LEDGF) in the rat eye: loss of LEDGF from nuclei of differentiating cells

Lens epithelium-derived growth factor (LEDGF) enhances the survival and growth of cells. To understand LEDGF's spatial localization and its putative function(s) during proliferation and differentiation, we localized LEDGF during terminal differentiation in whole rat lenses, lens epithelial cell (LEC) explants stimulated with FGF-2, and insulin, iris, human LECs with lentoids. In addition, intracellular localization of LEDGF was performed in other ocular tissues: ciliary body, retina, and cornea. We found the immunopositivity of nuclear LEDGF decreased in LECs of the equatorial region. In contrast, immunopositivity of LEDGF was detected in the cytoplasm of LECs and superficial fiber cells. After treating LEC explants with FGF-2 and insulin, which are known to be differentiating factors for LECs, the nuclei of these cells showed no LEDGF immunopositivity, but explants did express p57(kip2), a differentiation marker protein. Also, immunopositive LEDGF was not detected in the nuclei of differentiated cells, lentoid body, and corneal epithelial cells. This demonstrated that the loss of LEDGF from the nucleus may be associated with the process of terminal differentiation that might be in some way common with the biochemical mechanisms of apoptosis. The spatial and temporal distribution of LEDGF in the present study also provides a vision for further investigation as to how this protein is involved in cell fate determination.     

Conditional ablation of the Notch2 receptor in the ocular lens

Notch signaling is essential for proper lens development, however the specific requirements of individual Notch receptors have not been investigated. Here we report the lens phenotypes of Notch2 conditionally mutant mice, which exhibited severe microphthalmia, reduced pupillary openings, disrupted fiber cell morphology, eventual loss of the anterior epithelium, fiber cell dysgenesis, denucleation defects, and cataracts. Notch2 mutants also had persistent lens stalks as early as E11.5, and aberrant DNA synthesis in the fiber cell compartment by E14.5. Gene expression analyses showed that upon loss of Notch2, there were elevated levels of the cell cycle regulators Cdkn1a (p21Cip1), Ccnd2 (CyclinD2), and Trp63 (p63) that negatively regulates Wnt signaling, plus down-regulation of Cdh1 (E-Cadherin). Removal of Notch2 also resulted in an increased proportion of fiber cells, as was found in Rbpj and Jag1 conditional mutant lenses. However, Notch2 is not required for AEL proliferation, suggesting that a different receptor regulates this process. We found that Notch2 normally blocks lens progenitor cell death. Overall, we conclude that Notch2-mediated signaling regulates lens morphogenesis, apoptosis, cell cycle withdrawal, and secondary fiber cell differentiation.     

Bone morphogenetic protein signaling and the initiation of lens fiber cell differentiation

Previous studies showed that the retina produces factors that promote the differentiation of lens fiber cells, and identified members of the fibroblast growth factor (FGF) and insulin-like growth factor (IGF) families as potential fiber cell differentiation factors. A possible role for the bone morphogenetic proteins (BMPs) is suggested by the presence of BMP receptors in chicken embryo lenses. We have now observed that phosphorylated SMAD1, an indicator of signaling through BMP receptors, localizes to the nuclei of elongating lens fiber cells. Transduction of chicken embryo retinas and/or lenses with constructs expressing noggin, a secreted protein that binds BMPs and prevents their interactions with their receptors, delayed lens fiber cell elongation and increased cell death in the lens epithelium. In an in vitro explant system, in which chicken embryo or adult bovine vitreous humor stimulates chicken embryo lens epithelial cells to elongate into fiber-like cells, these effects were inhibited by noggin-containing conditioned medium, or by recombinant noggin. BMP2, 4, or 7 were able to reverse the inhibition caused by noggin. Lens cell elongation in epithelial explants was stimulated by treatment with FGF1 or FGF2, alone or in combination with BMP2, but not to the same extent as vitreous humor. These data indicate that BMPs participate in the differentiation of lens fiber cells, along with at least one additional, and still unknown factor.     

Disruption of Gja8 (alpha8 connexin) in mice leads to microphthalmia associated with retardation of lens growth and lens fiber maturation.

The development of the vertebrate lens utilizes a sophisticated cell-cell communication network via gap junction channels, which are made up of at least three connexin isoforms, alpha8 (Cx50), alpha3 (Cx46) and alpha1 (Cx43), and which are encoded by three different genes. In a previous study, we reported that, with a disruption of Gja3 (alpha3 connexin), mice developed nuclear cataracts with a normal sized lens. We show that Gja8tm1 (alpha8-/-) mice develop microphthalmia with small lenses and nuclear cataracts, while the alpha8 heterozygous (+/-) mice have relatively normal eyes and lenses. A comparative study of these alpha3 and alpha8 knockout mice showed that the protein levels of both alpha3 and alpha8 were independently regulated and there was no compensation for either the alpha3 or alpha8 protein from the wild-type allele when the other allele was disrupted. More interestingly, western blotting data indicated that the presence of alpha8 in the lens nucleus is dependent on alpha3 connexin, but not vice versa. The staining of the knock-in lacZ reporter gene showed the promoter activity of alpha8 connexin is much higher than that of alpha3 connexin in embryonic lenses and in adult lens epithelium. More importantly, a delayed denucleation process was observed in the interior fibers of the alpha8-/- lenses. Therefore, alpha8 connexin is required for proper fiber cell maturation and control of lens size.     

Prox1 is differentially localized during lens development

Prox1, the vertebrate cognate of Drosophila Prospero, is a homeodomain protein essential for the development of the lens, liver and lymphatic system. While it is well established that the subcellular distribution of Prospero changes during development, this had not been demonstrated for Prox1. Here, high-resolution confocal microscopy demonstrated that Prox1 protein is predominately cytoplasmic in the lens placode as well as the lens epithelium and germinative zone throughout development. However during fiber cell differentiation, Prox1 protein redistributes to cell nuclei. Finally, as lens fiber cells condense their chromatin in response to lens denucleation, Prox1 remains in the nucleus but does not appear to interact with DNA. Thus, it appears that the function of Prox1, like that of its Drosophila cognate Prospero, is at least partially controlled by changes in its subcellular distribution during development.     

Overexpression of the vimentin gene in transgenic mice inhibits normal lens cell differentiation

To investigate the role of the intermediate filament protein vimentin in the normal differentiation and morphogenesis of the eye lens fiber cells, we generated transgenic mice bearing multiple copies of the chicken vimentin gene. In most cases, the vimentin transgene was overexpressed in the lenses of these animals, reaching up to 10 times the endogenous levels. This high expression of vimentin interfered very strongly with the normal differentiation of the lens fibers. The normal fiber cell denucleation and elongation processes were impaired and the animals developed pronounced cataracts, followed by extensive lens degeneration. The age of appearance and extent of these abnormalities in the different transgenic lines were directly related to the vimentin level. Electron microscopic analysis revealed that the accumulated transgenic protein forms normal intermediate filaments.     

The common modification in alphaA-crystallin in the lens, N101D, is associated with increased opacity in a mouse model

To elucidate the morphological and cellular changes due to introduction of a charge during development and the possible mechanism that underlies cataract development in humans as a consequence of an additional charge, we generated a transgenic mouse model mimicking deamidation of Asn at position 101. The mouse model expresses a human αA-crystallin gene in which Asn-101 was replaced with Asp, which is referred to as αAN101D-transgene and is considered to be "deamidated" in this study. Mice expressing αAN101D-transgene are referred to here CRYAA(N101D) mice. All of the lines showed the expression of αAN101D-transgene. Compared with the lenses of mice expressing wild-type (WT) αA-transgene (referred to as CRYAA(WT) mice), the lenses of CRYAA(N101D) mice showed (a) altered αA-crystallin membrane protein (aquaporin-0 (AQP0), a specific lens membrane protein) interaction, (b) extracellular spaces between outer cortical fiber cells, (c) attenuated denucleation during confocal microscopic examination, (d) disrupted normal fiber cell organization and structure during scanning electron microscopic examination, (e) distorted posterior suture lines by bright field microscopy, and (f) development of a mild anterior lens opacity in the superior cortical region during the optical coherence tomography scan analysis. Relative to lenses with WT αA-crystallin, the lenses containing the deamidated αA-crystallin also showed an aggregation of αA-crystallin and a higher level of water-insoluble proteins, suggesting that the morphological and cellular changes in these lenses are due to the N101D mutation. This study provides evidence for the first time that expression of deamidated αA-crystallin caused disruption of fiber cell structural integrity, protein aggregation, insolubilization, and mild cortical lens opacity.     

The ability of the epithelium of diencephalic origin to differentiate into cells of the ocular lens.

After the discovery that in adult salamanders following lentectomy a new, functional lens develops by transdifferentiation (cell-type conversion) of previously depigmented epithelial cells of the iris (Wolffian lens regeneration), this phenomenon has been intensively studied by various experimental approaches. During the last two decades it was shown that pleiomorphic aggregates of atypical lens cells (lentoids) differentiated in reaggregates of dissociated cells of the chick neural retina and in spread cell cultures of the pigmented epithelium of the iris and retina, of the neural retina and the pineal gland of the chick embryo. The neural retina of human fetuses and adults also displayed this capacity. We showed that lentoids developed at a low incidence in renal isografts of rat embryonic shields or isolated embryonic ectoderm and of lentectomized eyes of rat fetuses, as well as in organ cultures of rat embryonic shields in chemically defined media. The addition of transferrin significantly increased the incidence of differentiation of lentoids in explants. In both renal isografts and explants in vitro a continuous transformation of retinal epithelial cells into atypical lens cells was observed. In renal isografts lentoids were also observed to originate from the ependyma of the brain ventricle. All tissues having the capacity to convert into lens cells belong to the diencephalon in a broad sense. Evolutionary aspects of this feature are discussed.     

Age-dependent changes in the distribution and concentration of human lens cholesterol and phospholipids

Analyses of total lipid in individual lenses 1.8-63 years of age indicate that both the cholesterol and the phospholipid concentrations have reached a high level of 10 and 14 micrograms/mg lens dry weight, respectively, after the first ten years of growth. Thereafter, the rate of phospholipid accumulation was greatly reduced to a value of 0.05 microgram/mg per year while that of cholesterol reduced to 0.19. Analyses of the distribution of lipid in successive lens fiber layers indicate that both the cholesterol and phospholipid levels increase in the entire lens between the age of 1.8 and 9 years. Older lenses showed a continuous increase in the accumulation of cholesterol in the deep cortical fibers, while little or no increase in phospholipid concentration was observed. These results indicate that the accumulation of lipids is greater than that of lens dry mass (protein) during the first decade of lens growth. Since more than 90% of lenticular lipids are associated with fiber cell membranes, these data suggest a gradual change in the differentiation of the newly formed secondary fibers from the epithelium during this period. Analyses of the phospholipid composition of the successive fiber fractions indicate that the major phospholipids of phosphatidyl ethanolamine (PE), phosphatidylserine (PS) and sphingomyelin maintained a uniform distribution in the 1.8- and 5-year-old lenses. While no change was observed with the cortical fibers, older lenses showed a gradual loss of PE and PS in the nuclear fiber up to 63 years of age. By the late teen years, nuclear PS can no longer be detected, while high levels of PE are maintained in lens nucleus. The disappearance of nuclear PE begins in the teen years and is completed by the age of 40. The decrease in PE and PS resulted in a continuous increase in the cholesterol/phospholipid ratio, a measure of membrane rigidity in the nuclear fiber in lenses 20 years of age and older. This decrease is also responsible for the exceedingly high rigidity of the nuclear fibers of lenses 60 years of age and older. Possible lamellar cholesterol organization in the lens fiber membrane is discussed.